Dairy Industry: Information on Milk Prices, Factors Affecting	 
Prices, and Dairy Policy Options (29-DEC-04, GAO-05-50).	 
                                                                 
In 2003, U.S. dairy farmers marketed nearly 19.7 billion gallons 
of raw milk, one-third of which were used in fluid milk products.
Farmers, cooperatives, processors, and retailers receive a	 
portion of the retail price of milk for their part in providing  
milk to consumers. During 2002 and 2003, farm prices fell while  
retail prices did not similarly decline. This pattern raised	 
concerns about a growing spread between farm and retail prices.  
Farm prices have since increased, reaching record highs in April 
2004. As requested, GAO examined (1) the portion of retail milk  
prices received by farmers, cooperatives, processors, and	 
retailers, how this changed over time, and the relationship	 
between price changes at these levels; (2) how various factors	 
influence prices and affect the transmission of price changes	 
among levels; and (3) how federal dairy program changes and	 
alternative policy options have affected or might affect farm	 
income and federal costs, among other considerations.		 
-------------------------Indexing Terms------------------------- 
REPORTNUM:   GAO-05-50						        
    ACCNO:   A14696						        
  TITLE:     Dairy Industry: Information on Milk Prices, Factors      
Affecting Prices, and Dairy Policy Options			 
     DATE:   12/29/2004 
  SUBJECT:   Agricultural policies				 
	     Agricultural products				 
	     Cost analysis					 
	     Dairy industry					 
	     Dairy products					 
	     Prices and pricing 				 
	     Farmers						 

******************************************************************
** This file contains an ASCII representation of the text of a  **
** GAO Product.                                                 **
**                                                              **
** No attempt has been made to display graphic images, although **
** figure captions are reproduced.  Tables are included, but    **
** may not resemble those in the printed version.               **
**                                                              **
** Please see the PDF (Portable Document Format) file, when     **
** available, for a complete electronic file of the printed     **
** document's contents.                                         **
**                                                              **
******************************************************************
GAO-05-50

                 United States Government Accountability Office

                     GAO Report to Congressional Requesters

December 2004

DAIRY INDUSTRY

 Information on Milk Prices, Factors Affecting Prices, and Dairy Policy Options

                                       a

GAO-05-50

Highlights of GAO-05-50, a report to congressional requesters

In 2003, U.S. dairy farmers marketed nearly 19.7 billion gallons of raw
milk, one-third of which were used in fluid milk products. Farmers,
cooperatives, processors, and retailers receive a portion of the retail
price of milk for their part in providing milk to consumers. During 2002
and 2003, farm prices fell while retail prices did not similarly decline.
This pattern raised concerns about a growing spread between farm and
retail prices. Farm prices have since increased, reaching record highs in
April 2004. As requested, GAO examined (1) the portion of retail milk
prices received by farmers, cooperatives, processors, and retailers, how
this changed over time, and the relationship between price changes at
these levels; (2) how various factors influence prices and affect the
transmission of price changes among levels; and (3) how federal dairy
program changes and alternative policy options have affected or might
affect farm income and federal costs, among other considerations.

To continue informed decisionmaking, the Secretary of Agriculture should
build on GAO's analysis of the potential effects of various policy options
as USDA proposes future changes or provides information to the Congress.
USDA expressed concern about the practicality of some of the dairy policy
options discussed. USDA did not comment on the report's recommendation.

www.gao.gov/cgi-bin/getrpt?GAO-05-50.

To view the full product, including the scope and methodology, click on
the link above. For more information, contact Lawrence J. Dyckman, (202)
512-3841, [email protected].

December 2004

DAIRY INDUSTRY

Information on Milk Prices, Factors Affecting Prices, and Dairy Policy Options

Between October 2000 and May 2004, on average, farmers received about 46
percent, cooperatives 6 percent, wholesale processors 36 percent, and
retailers over 12 percent of the retail price of a gallon of 2 percent
milk (the most common type of milk purchased) in the 15 U.S. markets GAO
reviewed. During this period, in 12 of the 15 markets, the spread between
farm and retail prices increased. However in some markets, the price
spread between these levels increased and then moderated. Price changes at
one level were most closely reflected in changes at adjacent levels of the
marketing chain.

Farm, cooperative, wholesale, and retail milk prices are determined by the
interaction of a number of factors. For example, farm prices are affected
by the supply of raw milk and the demand for milk products such as fluid
milk, cheese, and butter, as well as by federal and state dairy programs.
At the cooperative level, prices are influenced by the cost of services
that cooperatives provide, and the relative bargaining power of
cooperatives and milk processors. At the wholesale and retail levels,
input costs such as labor and energy, and the continued consolidation of
firms influence milk prices.

Recent changes in federal dairy programs have affected farm income,
federal costs, and other considerations. For example, the Milk Income Loss
Contract program has supported some farm incomes but has exceeded initial
cost estimates because of low farm prices. A number of options have been
suggested to change federal dairy policies such as amending federal milk
marketing orders and raising or eliminating the support price. In general,
these options would have mixed effects depending upon whether milk prices
were high or low over the short or long term. For example, options that
increase farm income over the short term tend to increase milk production
and lower farm prices over the long term. These options also tend to be
costly for the federal government during periods of low prices.

Farm and Retail 2 Percent Milk Prices, October 2000 to May 2004

Contents

        Letter                                                              1 
                                     Conclusions                            8 
                         Recommendations for Executive Action               9 
                          Agency Comments and Our Evaluation                9 

Appendixes

Appendix I: Appendix II:

Appendix III: Appendix IV:

Appendix V:

Appendix VI:

Objectives, Scope, and Methodology 14

Analysis of Prices at Four Marketing Levels for 2 Percent
Milk in Selected Markets 24
Portion Received by Farmers, Cooperatives, Wholesale Milk

Processors and Retailers 24 Changes in Farm and Retail Prices and the
Price Spread 26 Correlation between Price Changes at the Four Marketing

Levels 27 Comparison of Average Annual and Monthly Prices for 2 Percent
Milk 30

Retail Prices for Four Kinds of Milk in Selected Markets 49

Average Monthly and Annual Farm, Cooperative, Wholesale and Retail Milk
Prices in Selected Markets 64

Factors That Influence the Price of Milk as It Moves from the Farm to the
Consumer 94 Market Forces, in Addition to Federal and State Policies,
Influence

Milk Prices at the Farm Level 94

Services Provided by Cooperatives, Market Structure, and Collective Action
Can Influence the Price of Milk at the Cooperative Level 108

Input Costs, Service Levels, Innovations, and Market Structure Influence
Wholesale Fluid Milk Prices 113 Retailing Costs, Consumer Demand, and
Market Structure Changes Affect Retail Prices for Fluid Milk 120

Economic Studies of Price Transmission in the U.S. Fluid
Milk Market 131
Summary of Recent Economic Studies of Price Transmission in the

U.S. Fluid Milk Market 135 Results on the Extent of Farm-to-Retail Price
Transmission and Price Transmission Asymmetry 150 Results on
Farm-to-Retail Speed of Adjustment and Price Transmission Asymmetry 153

                                    Contents

Results for Retail-to-Farm Price Transmission and Price Transmission
Asymmetry 154 Possible Causes of Asymmetry in the Extent and Speed of
Price Transmission 154

Appendix VII:	Effects of Recent Federal Dairy Program Changes and
Alternative Policy Options 159 Recent Changes in Federal Dairy Programs
Vary in Their Effects on

Policy Considerations We Identified 159 Effects of Alternative Dairy
Policies Differ under Various Scenarios 176

Appendix VIII:	Comments from the U.S. Department of Agriculture 237 GAO
Comments 242

Appendix IX:	GAO Contacts and Staff Acknowledgments 246 Contacts 246
Acknowledgments 246

Related GAO Products

Tables Table 1:

Table 2: Table 3:

Table 4:

Table 5: Table 6:

Portion of the Retail Price of a Gallon of 2 Percent Milk
Received by Farmers, Cooperatives, Wholesale
Processors, and Retailers for 15 Markets, October 2000
through May 2004 25
Changes in the Farm-to-Retail Price Spread for a Gallon of
2 Percent Milk for 15 Markets, October 2000 through May
2004 26
Correlation between Farm Price Changes and Changes in
Cooperative, Wholesale, and Retail Prices for a Gallon of 2
Percent Milk for 15 Markets, October 2000 through May
2004 28
Correlation between Cooperative Price Changes and
Changes in Wholesale and Retail Prices for a Gallon of 2
Percent Milk for 15 Markets, October 2000 through May
2004 29
Correlation between Wholesale and Retail Price Changes
for a Gallon of 2 Percent Milk for 15 Markets, October 2000
through May 2004 30
Average Annual Farm, Cooperative, Wholesale, and Retail
Prices for a Gallon of 2 Percent Milk in Selected Markets,
2000 31

Contents

Table 7:	Average Annual Farm, Cooperative, Wholesale, and Retail Prices
for a Gallon of 2 Percent Milk in Selected Markets, 2001 31

Table 8:	Average Annual Farm, Cooperative, Wholesale, and Retail Prices
for a Gallon of 2 Percent Milk in Selected Markets, 2002 32

Table 9:	Average Annual Farm, Cooperative, Wholesale, and Retail Prices
for a Gallon of 2 Percent Milk in Selected Markets, 2003 33

Table 10: Average Annual Farm, Cooperative, Wholesale, and Retail Prices
for a Gallon of 2 Percent Milk in Selected Markets, 2004 33

Table 11: Prices for a Gallon of Milk in Atlanta, Georgia, October 2000
through May 2004 64 Table 12: Prices for a Gallon of Milk in Boston,
Massachusetts, October 2000 through May 2004 66 Table 13: Prices for a
Gallon of Milk in Charlotte, North Carolina, October 2000 through May 2004
68 Table 14: Prices for a Gallon of Milk in Cincinnati, Ohio, October 2000
through May 2004 70 Table 15: Prices for a Gallon of Milk in Dallas,
Texas, October 2000 through May 2004 72 Table 16: Prices for a Gallon of
Milk in Denver, Colorado, October 2000 through May 2004 74 Table 17:
Prices for a Gallon of Milk in Miami, Florida, October 2000 through May
2004 76 Table 18: Prices for a Gallon of Milk in Milwaukee, Wisconsin,
October 2000 through May 2004 78 Table 19: Prices for a Gallon of Milk in
Minneapolis, Minnesota, October 2000 through May 2004 80 Table 20: Prices
for a Gallon of Milk in New Orleans, Louisiana, October 2000 through May
2004 82 Table 21: Prices for a Gallon of Milk in Phoenix, Arizona, October
2000 through May 2004 84 Table 22: Prices for a Gallon of Milk in Salt
Lake City, Utah, October 2000 through May 2004 86 Table 23: Prices for a
Gallon of Milk in San Diego, California, October 2000 through May 2004 88
Table 24: Prices for a Gallon of Milk in Seattle, Washington, October 2000
through May 2004 90 Table 25: Prices for a Gallon of Milk in Washington,
D.C., October 2000 through May 2004 92

                                    Contents

Table 26: Average Ownership and Operating Costs by Region and

Herd Size, 2000 96 Table 27: USDA's Milk Classes Used for Setting Milk
Prices 102 Table 28: Change in Costs for Fluid Milk Processing Plants in
Maine,

1993 through 2000 115 Table 29: Supermarket Operating Costs as a
Percentage of Sales and Gross Margin, 2003 121 Table 30: Growth of
Supermarket Sales and Expenses during the Last Decade, 1993 through 2003
122 Table 31: Market Share of the Top Four Food Retailers in Selected
Markets, 2003 127 Table 32: Overview of Fluid Milk Price Transmission
Studies and Results on Price Transmission Asymmetry/Symmetry 132 Table 33:
Recent Price Transmission Studies of the Fluid Milk Market 136

Figures	Figure 1: Figure 2: Figure 3: Figure 4: Figure 5: Figure 6: Figure
7: Figure 8:

Selected Milk Markets, Corresponding Defense Commissaries Used for Our
Analysis, and Federal Milk Marketing Order Areas Prior to April 2004 19
Farm, Cooperative, Wholesale, and Retail Prices for a Gallon of 2 Percent
Milk for Atlanta, Georgia, October 2000 through May 2004 34 Farm,
Cooperative, Wholesale, and Retail Prices for a Gallon of 2 Percent Milk
for Boston, Massachusetts, October 2000 through May 2004 35 Farm,
Cooperative, Wholesale, and Retail Prices for a Gallon of 2 Percent Milk
for Charlotte, North Carolina, October 2000 through May 2004 36 Farm,
Cooperative, Wholesale, and Retail Prices for a Gallon of 2 Percent Milk
for Cincinnati, Ohio, October 2000 through May 2004 37 Farm, Cooperative,
Wholesale, and Retail Prices for a Gallon of 2 Percent Milk for Dallas,
Texas, October 2000 through May 2004 38 Farm, Cooperative, Wholesale, and
Retail Prices for a Gallon of 2 Percent Milk for Denver, Colorado, October
2000 through May 2004 39 Farm, Cooperative, Wholesale, and Retail Prices
for a Gallon of 2 Percent Milk for Miami, Florida, October 2000 through
May 2004 40

Contents

Figure 9:	Farm, Cooperative, Wholesale, and Retail Prices for a Gallon of
2 Percent Milk for Milwaukee, Wisconsin, October 2000 through May 2004 41

Figure 10: Farm, Cooperative, Wholesale, and Retail Prices for a Gallon of
2 Percent Milk for Minneapolis, Minnesota, October 2000 through May 2004
42

Figure 11: Farm, Cooperative, Wholesale, and Retail Prices for a Gallon of
2 Percent Milk for New Orleans, Louisiana, October 2000 through May 2004
43

Figure 12: Farm, Cooperative, Wholesale, and Retail Prices for a Gallon of
2 Percent Milk for Phoenix, Arizona, October 2000 through May 2004 44

Figure 13: Farm, Cooperative, Wholesale, and Retail Prices for a Gallon of
2 Percent Milk for Salt Lake City, Utah, October 2000 through May 2004 45

Figure 14: Farm, Cooperative, Wholesale, and Retail Prices for a Gallon of
2 Percent Milk for San Diego, California, October 2000 through May 2004 46

Figure 15: Farm, Cooperative, Wholesale, and Retail Prices for a Gallon of
2 Percent Milk for Seattle, Washington, October 2000 through May 2004 47

Figure 16: Farm, Cooperative, Wholesale, and Retail Prices for a Gallon of
2 Percent Milk for Washington, D.C., October 2000 through May 2004 48

Figure 17: Retail Prices in Atlanta, Georgia, for Whole, 2 Percent, 1
Percent, and Skim Milk, October 2000 through May 2004 49

Figure 18: Retail Prices in Boston, Massachusetts, for Whole, 2 Percent, 1
Percent, and Skim Milk, October 2000 through May 2004 50

Figure 19: Retail Prices in Charlotte, North Carolina, for Whole, 2
Percent, 1 Percent, and Skim Milk, October 2000 through May 2004 51

Figure 20: Retail Prices in Cincinnati, Ohio, for Whole, 2 Percent, 1
Percent, and Skim Milk, October 2000 through May 2004 52

Figure 21: Retail Prices in Dallas, Texas, for Whole, 2 Percent, 1
Percent, and Skim Milk, October 2000 through May 2004 53

Figure 22: Retail Prices in Denver, Colorado, for Whole, 2 Percent, 1
Percent, and Skim Milk, October 2000 through May 2004 54

Contents

Figure 23: Retail Prices in Miami, Florida, for Whole, 2 Percent, 1
Percent, and Skim Milk, October 2000 through May 2004 55

Figure 24: Retail Prices in Milwaukee, Wisconsin, for Whole, 2 Percent, 1
Percent, and Skim Milk, October 2000 through May 2004 56

Figure 25: Retail Prices in Minneapolis, Minnesota, for Whole, 2 Percent,
1 Percent, and Skim Milk, October 2000 through May 2004 57

Figure 26: Retail Prices in New Orleans, Louisiana, for Whole, 2 Percent,
1 Percent, and Skim Milk, October 2000 through May 2004 58

Figure 27: Retail Prices in Phoenix, Arizona, for Whole, 2 Percent, 1
Percent, and Skim Milk, October 2000 through May 2004 59

Figure 28: Retail Prices in Salt Lake City, Utah, for Whole, 2 Percent, 1
Percent, and Skim Milk, October 2000 through May 2004 60

Figure 29: Retail Prices in San Diego, California, for Whole, 2 Percent, 1
Percent, and Skim Milk, October 2000 through May 2004 61

Figure 30: Retail Prices in Seattle, Washington, for Whole, 2 Percent, 1
Percent, and Skim Milk, October 2000 through May 2004 62

Figure 31: Retail Prices in Washington, D.C., for Whole, 2 Percent, 1
Percent, and Skim Milk, October 2000 through May 2004 63

Figure 32: Map of Federal Milk Marketing Orders 101 Figure 33: Number of
Fluid Milk Processing Plants and Volume of Milk Processed per Plant, 1980
through 2002 119 Figure 34: Percentage of Retail Food Sales by Market
Segment, 1993 and 2003 129 Figure 35: Marketing Chain and Price
Transmission for the U.S. Fluid Milk Industry 134 Figure 36: Potential
Effects of Options to Change FMMOs on Various Policy Considerations 178
Figure 37: Potential Effects of Options to Change the Price Support
Program on Various Policy Considerations 194 Figure 38: Potential Effects
of Options to Extend MILC on Various Policy Considerations 204 Figure 39:
Potential Effects of Adopting the NDEA on Various Policy Considerations
211

Contents

Figure 40: Potential Effects of Changing Trade Restrictions and

Export Incentives on Various Policy Considerations 217 Figure 41:
Potential Effects of Options to Facilitate Risk

Management on Various Policy Considerations 228 Figure 42: Potential
Effects of Options to Manage Raw Milk Supplies

on Various Policy Considerations 233

Abbreviations

AMS Agricultural Marketing Service
CCC Commodity Credit Corporation
DEIP Dairy Export Incentive Program
FMMO federal milk marketing order
FSA Farm Service Agency
MILC Milk Income Loss Contract
NDEA National Dairy Equity Act
NEDC Northeast Interstate Dairy Compact
NYC New York City
UNY Upstate New York
USDA U.S. Department of Agriculture
WTO World Trade Organization

This is a work of the U.S. government and is not subject to copyright
protection in the United States. It may be reproduced and distributed in
its entirety without further permission from GAO. However, because this
work may contain copyrighted images or other material, permission from the
copyright holder may be necessary if you wish to reproduce this material
separately.

A

United States Government Accountability Office Washington, D.C. 20548

December 29, 2004

Congressional Requesters:

In 2003, U.S. dairy farmers marketed nearly 19.7 billion gallons of raw
milk for which they received approximately $21.2 billion. Roughly
one-third of these 19.7 billion gallons were ultimately sold to consumers
as fluid milk (milk) products. The milk marketing chain that moves milk
from the dairy farm to the consumer is composed of farmers, dairy
cooperatives,1 wholesale milk processors,2 and retailers, and entities at
each level receive a portion of the retail price of a gallon of milk for
the functions they perform. These functions include producing, processing,
distributing, and selling milk. During 2002 and 2003, farm prices were the
lowest since 1979, while retail milk prices generally did not experience a
similar decline.3 This disparity raised concerns about a growing price
spread between farm and retail milk prices. However, in 2004, farm price
trends have reversed, with farm prices reaching record highs.

Since the 1930s, the U.S. Department of Agriculture (USDA) and some states
have operated programs designed to support dairy farmers by ensuring that
farm prices do not fall below minimum levels. These programs include,
among others, federal milk marketing orders (FMMOs), which use formulas to
establish minimum farm prices in different marketing areas of the country;
a price support program to sustain farm prices through a standing offer to
purchase certain milk products at set prices; and trade restrictions or
export subsidies to protect the U.S. milk industry from competition from
foreign milk products or reduce supplies.

Low farm prices have generated interest in new federal policies to provide
assistance to dairy farmers; however, the cost of these programs during
periods of low prices has raised concerns. For example, in the Farm

1Dairy cooperatives are member-owned organizations that assist producers
in marketing their milk. Cooperatives provide a variety of services for
their members such as ensuring adequate raw milk supplies to meet
processors' needs, transporting milk, quality assurance, and standardizing
milk composition.

2Wholesale milk processors include bottlers and major retail food chains
with bottling plants, and cooperatives that process, package, and
distribute fluid milk for sale to retailers. Our review did not include
other entities that market milk at the wholesale level.

3In this report, farm prices are the prices paid for raw milk supplies.
After the milk has been processed into a fluid milk beverage, references
to "milk prices" are retail milk prices.

Security and Rural Investment Act of 2002 (the 2002 Farm Bill), Congress
authorized USDA to establish the Milk Income Loss Contract (MILC) program,
which provides supplemental payments to farmers during periods of low
prices.4 Depressed farm prices triggered $1.8 billion in MILC payments to
dairy farmers in fiscal year 2003. Additionally, USDA spent over $1.5
billion in fiscal years 2001 through 2003 attempting to maintain minimum
farm prices at a legislatively set level through the price support
program.

Recent changes in milk prices have illustrated the volatility of milk
price cycles. For example, as of April 2004, farm prices reached record
highs as USDA announced a minimum price for raw milk used in manufacturing
cheese that was approximately 36 percent higher, at $19.66 per
hundredweight, than the March price of $14.49 per hundredweight, an
increase of nearly $0.45 per gallon of milk.5 This increase reflected a
number of changed market conditions that decreased the amount of raw milk
production, while also increasing the demand for some manufactured dairy
products.6 However, high farm prices create incentives for farmers to
produce more milk, which then tends to depress milk prices over the long
term, thereby increasing concerns about farm viability in the future.

At your request, this report examines (1) what portion of the retail price
of fluid milk is received by dairy farmers, dairy cooperatives, wholesale
milk processors, and retailers in selected markets throughout the United
States, how this distribution has changed over the period of our review
(October 2000 through May 2004), and the relationships among price changes
at these levels; (2) how various factors, such as costs, influence the
price of milk as it moves from the farm to the consumer, as well as how
these and other factors affect the extent to which changes in price are
transmitted among levels as milk moves from the farm to the consumer; and
(3) how changes in dairy policies and alternative policy options have
affected or

4Pub. L. No. 107-171, 116 Stat. 134 (2002).

5A hundredweight is a unit of measure equal to 100 pounds of, in this
case, raw milk. Approximately 11.6 gallons of fluid milk can be made from
100 pounds of raw milk.

6Recent evidence suggests that retail milk prices have also begun rising
with the recovering farm prices.

might affect farm income, federal costs, economic efficiency,7 and
consumer prices, among other policy considerations. It also updates other
information on milk prices included in our June 2001 report.8

To update our information on the pricing distribution among the various
levels of the milk marketing chain, we analyzed milk prices in 15 selected
markets nationwide, ensuring that (1) these markets provided national
geographic coverage; (2) at least one market was located in each of the
federal milk marketing order areas, as they existed during the majority of
the period from October 2000 through May 2004; (3) the selected markets
included both state and federally regulated markets; and (4) these areas
represented similar marketing areas for which we reported information in
our June 2001 report.9 For these 15 markets, we obtained data from USDA,
the cognizant state milk control agency, the Department of Defense's
Commissary Agency, and a private data collection company on the prices
received by farmers, cooperatives, wholesale milk processors, and
retailers.10 We limited our data collection efforts to whole, skim, 2
percent, and 1 percent milk as sales of these milk types constitute over
93 percent of fluid milk sales annually. We also confined our analysis to
the prices of these milk types sold in gallon containers as milk sold in
gallon containers

7There can be different kinds of economic efficiency effects. Government
policies may be more or less economically efficient depending upon the
extent to which they prevent the transmission of market price signals and
lead to a misallocation of resources into excess production. Policies may
also be more or less efficient depending upon the extent to which they
affect the distribution of production between farmers with high or low
costs of production.

8GAO, Dairy Industry: Information on Milk Prices and Changing Market
Structure, GAO01-561 (Washington, D.C.: June 15, 2001).

9The beginning of our period of analysis corresponds to the end of the
data presented in our previous report, GAO-01-561. The end point
represents the most recent month for which data were available from all of
our sources.

10We were unable to obtain actual data on the prices received by wholesale
milk processors because these data are considered proprietary. As a proxy
for these data we used data obtained from the Defense Commissary Agency on
the prices at which it sold milk. Commissary prices reflect a 5 percent
markup above the prices at which the commissaries purchase milk from
wholesale suppliers. We obtained data from 39 different commissary
locations near the 15 markets in our analysis.

accounts for about 65 percent of fluid milk products sold under FMMOs.11
We focused our detailed data analysis on 2 percent milk sold in gallon
containers, the largest volume of reduced-fat milk sold nationwide. As a
result, our analysis may not reflect pricing trends for all types of milk
and package sizes.

To update our information on factors influencing milk prices, explore
price transmission within the milk marketing chain, and examine the
effects of federal dairy policy changes and alternative policy options, we
contacted a number of national dairy experts in academia or representing
federal and state agencies, cooperatives, processors, retailers, or
industry groups. We also reviewed relevant legislation, studies, and other
publications. We qualitatively analyzed the effects of recent changes in
federal dairy programs, as well as alternative policy options, on various
policy considerations as identified in previous GAO reports, relevant
studies, and legislation, as well as through our conversations with dairy
policy experts. These policy considerations included farm income, milk
production, federal costs, price volatility, economic efficiency, and
consumer prices. Different stakeholders in the dairy policy arena may have
alternative views on the relative importance of these policy
considerations, as well as other considerations that we did not include,
which could lead to differing perspectives on these options. We conducted
our review from September 2003 through October 2004 in accordance with
generally accepted government auditing standards. We did not independently
verify the data we received from various sources. However, we discussed
with these sources the measures they take to ensure the accuracy of data,
and these measures seemed reasonable. Appendix I provides additional
information on the scope and methodology of our review.

In summary, we found the following:

o 	Between October 2000 and May 2004, on average, farmers received 45.9
percent, cooperatives 6.1 percent, wholesale milk processors 35.6 percent,
and retailers 12.5 percent of the retail price of a gallon of 2 percent
milk in the 15 markets we reviewed. However, these percentages varied
widely depending on the specific market. For example, the farmers' portion
ranged from 36.0 percent to about 58.6 percent.

11This estimate is based on USDA data for the month of November for
selected years between 1991 and 2001, which were the most recent data we
could obtain.

o 	Furthermore, during this time period, the price spread between farm and
retail prices increased in 12 of the 15 markets we examined. However, in
some of the 12 markets, the spread between farm and retail milk prices
increased dramatically, and then moderated. In 9 of the 15 markets, retail
prices showed a statistically significant increase. In 4 of the remaining
markets, retail prices decreased over time, while in the other 2 markets,
retail prices showed no statistically significant change. At the same
time, farm prices decreased in 12 of the 15 markets and increased in the
remaining 3 markets during the 44-month period. However, declines in farm
prices, in most cases, began to reverse during the latter months of our
period of analysis. Price changes generally correlated across levels in
the marketing chain, with the strongest correlations occurring between
adjacent levels. For example, in most of the markets we analyzed, changes
in cooperative prices correlated strongly with changes in wholesale
prices. However, changes in cooperative prices correlated less strongly
with changes in retail prices.

o 	Prices at all levels of the milk marketing chain are determined by the
interaction of a number of factors. For example, farm prices are
determined primarily by factors affecting the supply of raw milk, such as
costs of production; the demand for milk products, such as fluid milk,
cheese and butter; and the effects of federal and state dairy programs. At
the cooperative level, the difference between what cooperatives pay
farmers for raw milk and the prices at which they sell raw milk to
wholesale fluid milk processors is influenced by the types of services the
cooperatives provide, the relative bargaining power of cooperatives and
milk processors, as well as collective action taken by dairy cooperatives.
Factors that affect the difference between what processors pay for this
raw milk and the prices at which they sell fluid milk products to
retailers include input costs, such as labor, energy, transportation and
packaging; the level of services they provide to retailers; innovations in
processing technology; and changes in the structure of the fluid milk
processing industry, such as increases in the consolidation and market
share of some firms. At the retail level, costs such as labor and energy,
along with other factors such as consumer demand and the structure of the
market, help determine the difference between what retailers pay for fluid
milk products and the prices they charge consumers for those products.

o 	While price changes at one level of the milk marketing chain are
generally reflected in price changes at other levels, several factors can
influence the extent or speed at which these changes are reflected and

whether price increases and decreases are reflected differently. For
example, increasing concentration, resulting in greater market power at
successive levels of the milk marketing chain, may provide entities an
opportunity to influence how changes in prices are transmitted among the
different levels. In addition, other factors, such as federal and state
dairy policies, can affect prices in ways that can cause market
participants to alter the way they transmit price changes between
marketing levels. Some recent economic studies of the U.S. fluid milk
market have found that these and other factors may cause retail milk
prices to react more completely and quickly to farm price increases than
to decreases.

o 	Recent changes in the FMMO program, adjustments to the prices of
products purchased under the price support program, and the establishment
of the MILC program have had various impacts on policy considerations such
as dairy farm income, federal costs, and price volatility. Reforms to the
FMMO system had mixed effects on farm income depending on the geographic
location of the farmer, while the overall effects on all farmers are less
clear. Further, recent adjustments to the prices of products purchased
under the price support program have generally decreased federal costs but
also decreased farm income during periods of low prices. Finally,
government payments introduced through the MILC program have kept some
small dairy farms in business, but the program has exceeded initial
federal cost estimates because farm prices during 2002 and 2003 were lower
than anticipated.

o 	A number of options that would modify existing policies or introduce
alternative policies have been proposed or discussed. While we examined
the potential impacts of these options on a range of policy considerations
related to the dairy sector, we did not assess their overall economic or
budgetary impacts or their consistency with U.S. international trade
commitments or positions in ongoing negotiations. As a result, the purpose
of this analysis is not to take a position for or against any of these
options, but simply to discuss their likely effects on the policy
considerations we identified. Various policy options could affect dairy
policy considerations in different ways under different scenarios, such as
periods of high or low prices. Also, short-term effects may differ from
long-term effects. In general, options that increase farm income over the
short term also tend to increase milk production and thus the potential
for oversupply and lower average farm prices over the long term. For
example, the dairy support price could be raised, which would limit the
fall of farm prices and increase farm income in the short

term. However, it would spur additional production and therefore reduce
average farm prices over the long term. Such options also tend to be
costly for the federal government during periods of low prices. Thus,
extending the MILC program, which is scheduled to expire at the end of
September 2005, would allow the federal government to continue to support
farm income but could also increase federal costs if farm prices trend
downward in the future. In some cases, options that increase the economic
efficiency of federal dairy programs also increase price volatility
because they allow clearer transmission of market price signals. For
example, eliminating the price support program would increase volatility
by removing the price floor on manufactured dairy products, but it would
also increase economic efficiency by reducing incentives to allocate
resources to surplus production. Also, to the extent that price changes at
the wholesale level are passed through to the retail level, a number of
options would likely have mixed effects on consumer prices depending upon
the particular product under consideration (e.g., butter, cheese, or fluid
milk).

As a result of recent farm bills, such as those in 1996 and 2002, USDA has
studied different aspects of dairy policy, either to implement program
changes mandated by the Congress or to provide information to the Congress
on the effects of various dairy programs. For example, the 2002 farm bill
required USDA to conduct studies of dairy policy issues, including an
economic evaluation of the effects of various elements of the national
dairy policy and of terminating federal dairy programs relating to milk
price support and supply management. One study required by this
legislation, which addressed the subject of changing standards for fluid
milk, was issued in August 2003.12 A second study, which included both an
evaluation of the effects of various elements of the national dairy policy
and the termination of federal milk price support and supply management
programs, was provided to the Congress in September 2004.13

12USDA/Agricultural Marketing Service, Raising the Minimum Nonfat Solids
Standard to the National Average in Raw Milk: A Study of Fluid Milk
Identity Standards (Washington, D.C.: August 2003).

13USDA, Economic Effects of U.S. Dairy Policy and Alternative Approaches
to Milk Pricing, (Washington, D.C.: July 2004). This study combined the
results of work by more than 20 researchers from 10 universities, as well
as information provided by researchers at USDA.

This report is divided into nine appendixes. Appendix I describes in
detail our objectives, scope, and methodology. Appendix II provides
information on average milk prices at the farm, cooperative, wholesale,
and retail levels; changes in farm and retail milk prices and how they
affect the farmto-retail price spread; and the extent to which price
changes at one level of the milk marketing chain correlate with price
changes at other levels. Appendix III compares retail prices of whole, 2
percent, 1 percent, and skim milk. Appendix IV provides the average
monthly prices and annual prices of the four types of milk at each level
of the milk marketing chain. Appendix V describes the factors that
influence prices as milk moves from the farmer to the consumer. Appendix
VI provides a technical review of recent research examining price
transmission within the milk marketing chain. Appendix VII provides a
qualitative analysis of the effects of recent changes in federal dairy
programs and alternative dairy policy options that have been proposed or
discussed. Appendix VIII presents USDA's comments and our evaluation of
them. Appendix IX lists GAO contacts and contributors to this report.

Conclusions	The difference in price between what farmers receive for their
raw milk and what consumers pay for fluid milk products has increased in
recent years. This growing spread between farm and retail prices may be
attributable to a number of factors at each level of the milk marketing
chain, including supply and demand forces, changes in input costs to
processing and retailing, and the continued concentration of cooperatives,
wholesale milk processors, and retailers. A variety of federal policies
exist to influence the prices that farmers receive. However, the effects
of these policies may not be uniform; they can affect different sizes of
farms or regions of the country in different ways. Moreover, policies that
benefit farm income may adversely affect other policy considerations such
as economic efficiency and federal costs. Given the complexity of federal
dairy policy, the decision to change existing policies or introduce new
policies requires consideration of a variety of these potential effects.
Examining the effects of policy alternatives on a variety of different
policy considerations will help the Congress formulate federal dairy
policy based on comprehensive analyses that consider these alternatives in
relation to their effects on different considerations, farm sizes, and
regions of the country. In addition, although recent USDA studies have
examined some policy options, there are other potential policy options to
consider, as discussed in this report.

Recommendations for 	To continue the facilitation of informed decision
making by USDA and the Congress, we recommend that the Secretary of
Agriculture build on GAO's

Executive Action	analysis of the potential effects of various dairy policy
options as USDA proposes future changes to current dairy laws or
regulations or provides information to the Congress in response to
congressional proposals.

Agency Comments and Our Evaluation

We provided a draft of this report to USDA and DOD for their review and
comment. We received written comments from USDA's Under Secretary for Farm
and Foreign Agricultural Services and Under Secretary for Marketing and
Regulatory Programs, which are presented in appendix VIII. USDA also
provided suggested technical corrections, which we have incorporated into
this report, as appropriate. These technical corrections were offered by
several USDA agencies, including the Agricultural Marketing Service,
Economic Research Service, Farm Service Agency, and Office of the Chief
Economist. DOD had no comments on the draft report.

In its written comments, USDA said the information provided in the report
on milk prices at the farm, cooperative, and retail levels is valid.
However, USDA said it has reservations regarding our use of prices paid
for fluid milk at commissaries as an indicator of the wholesale price of
fluid milk and that we should make clear the weaknesses of using
commissary price data. USDA acknowledged, however, that there seems to be
no viable alternative. During the course of our work, we were unable to
obtain wholesale price data because these data are considered proprietary
by industry officials. After consulting with USDA officials and other
dairy experts, we determined that commissary price data were the best
surrogate because commissaries generally sell milk at a standard 5 percent
markup from cost. Based on USDA's comments, we expanded the discussion in
the report of the potential limitations of using commissary data.

USDA said it largely agrees with the report's discussion of the factors
that influence the price of milk as it moves from the farm to the consumer
and the report's characterization of economic studies of price
transmission in the U.S. fluid milk market. However, USDA expressed some
concerns regarding the report's discussion of recent federal dairy program
changes and alternative policy options. First, USDA said that this
discussion appears to be a compilation of policy recommendations that are
examined independently and qualitatively within the existing program
structure. Our discussion of dairy policy options are not policy
recommendations. As stated in the report, to identify these policy options
and their potential

impacts we relied heavily on a synthesis of the views of leading dairy
experts and the results of an extensive literature search, including our
review of more than 50 studies and other publications. Time and resource
constraints for completing our work precluded us from developing or
contracting for the use of an economic model that would have provided
quantitative estimates of these potential impacts. In addition, some of
the policy options would have been difficult to model and quantify, such
as the potential impacts of accelerating USDA's hearing and rulemaking
process for amending FMMOs. The report also notes that we compared the
policy options identified against a baseline scenario of policies in place
as of August 2004. This baseline scenario existed at the start of our work
and was needed to provide a consistent context for our analysis.

Second, USDA suggested that we make clear the caveats of this type of
analysis. As noted in the report, we examined the impact of federal dairy
program changes and policy options on six policy considerations: farm
income, milk production, federal costs, price volatility, economic
efficiency, and consumer prices. We acknowledge that other stakeholders
may have different views on the importance of these policy considerations,
or other considerations that we did not include in our analysis. The
report also states that the potential effects of policy options on these
considerations could vary depending upon economic conditions and other
policy decisions. In this regard, we did not assess the options' overall
economic or budgetary impacts, or their consistency with U.S.
international trade commitments or positions in ongoing negotiations. In
addition, the report does not identify a preferred option or combination
of options. As indicated in the report, each option has varying potential
impacts on the policy considerations used in our analysis. Despite these
caveats, we believe this analysis is informative and helpful to
congressional decision makers who must weigh competing interests in
determining dairy policy.

USDA also said that in some cases the report mischaracterizes the
operation of current programs and the effects that changes to current
programs or the introduction of new programs would have on program
outlays, producers, and consumers. For example, USDA noted that the report
offers several options for improving the operation of the Dairy Export
Incentive Program (DEIP), including expanding the use of this program.
However, USDA indicated that expanding the use of DEIP is not a legitimate
option because, under World Trade Organization (WTO) rules, DEIP is bound
by quantitative and monetary caps and product-specific restrictions that
limit its use to the current range of eligible dairy

commodities. We do not agree that we mischaracterized the operation of
this program. The report clearly states that USDA has announced and
awarded subsidies under DEIP to the limits allowed by WTO rules for nonfat
dry milk and various cheeses. Regarding expansion, the report discusses
options suggested by dairy experts for the additional use of this program
as an effective marketing tool, and does not call for expanding its use to
exceed relevant WTO caps or restrictions. However, we have adjusted the
language in the report to make this distinction clearer.

USDA also offered several comments regarding the FMMO program. Among
these, USDA said that the report is incorrect in stating that the
objective of this program is "to ensure an adequate level of milk
production." According to USDA, this objective is associated with the
Dairy Price Support Program. We have revised the report to reflect this
clarification and added language suggested by USDA to better describe the
FMMO program's objectives.

In addition, USDA raised concerns about the practicality of implementing
some of the options discussed in the report, particularly (1) adopting a
competitive pay price to establish class prices under the FMMO program and
(2) combining Class III and Class IV into a single manufacturing class.14
Regarding the first, USDA said that it and a committee of academicians
spent considerable time several years ago trying to devise a competitive
price series that could be used to establish minimum class prices.
However, this effort was unsuccessful. USDA said that our report does not
identify or indicate how to create such a price series. Similarly,
regarding combining Class III and Class IV, USDA notes that no specifics
are offered in the report as to how milk in such a class would be priced.
We acknowledge that the report does not explain how a competitive price
series could be created or how milk would be priced if the classes were
merged. However, these options were identified by stakeholders during the
course of our work. Other options discussed in the report also may present
challenging implementation issues and in many cases the report discusses
those issues.

14Under the FMMO program, a classified pricing plan provides different
classes and minimum prices for milk depending on how it is used. Milk used
in fluid products is placed in Class I. Milk used for various manufactured
products is placed in Classes II through IV. Class II includes soft
products, such as cottage cheese, ice cream, and yogurt. Class III
includes spreadable and hard cheeses. Class IV includes butter and dried
milk products, such as nonfat dry milk.

Finally, USDA said that it does not believe the hearing and rulemaking
process it uses to modify FMMOs inhibits its ability to respond to
changing market conditions or the marketing of new dairy products.
However, as discussed in the report, some stakeholders cited the slowness
of this process as a concern. In addition, the report discusses USDA's
efforts to improve this process to more quickly respond to problems or
needed changes while ensuring the promulgation of economically sound
regulation.

USDA did not comment on the report's recommendation.

As agreed with your offices, unless you publicly announce its contents
earlier, we plan no further distribution of this report until 30 days from
the date of this letter. We will then send copies of the report to the
Senate Committee on Agriculture, Nutrition, and Forestry; the House
Committee on Agriculture; other appropriate congressional committees;
interested Members of Congress; the Secretary of Agriculture; the
Secretary of Defense; the Director of the Office of Management and Budget;
and other interested parties. Copies will also be made available to others
upon request. In addition, the report will be available at no charge on
GAO's Web site at http://www.gao.gov.

If you or your staff have any questions about this report, please contact
me at (202) 512-3841. Key contributors to this report are listed in
appendix IX.

Lawrence J. Dyckman Director, Natural Resources

and Environment

List of Requesters

The Honorable Olympia J. Snowe
Chair
Committee on Small Business & Entrepreneurship
United States Senate

The Honorable Hillary Rodham Clinton United States Senate

The Honorable Susan Collins United States Senate

The Honorable Christopher J. Dodd United States Senate

The Honorable James M. Jeffords United States Senate

The Honorable Edward M. Kennedy United States Senate

The Honorable John F. Kerry United States Senate

The Honorable Patrick Leahy United States Senate

The Honorable Barbara A. Mikulski United States Senate

The Honorable John D. Rockefeller IV United States Senate

The Honorable Paul S. Sarbanes United States Senate

The Honorable Charles E. Schumer United States Senate

The Honorable Arlen Specter United States Senate

Appendix I

                       Objectives, Scope, and Methodology

In March 2003, Senator Snowe, Chair, Committee on Small Business and
Entrepreneurship, joined by Senators Clinton, Collins, Dodd, Jeffords,
Kennedy, Kerry, Leahy, Mikulski, Rockefeller IV, Sarbanes, Schumer, and
Specter requested that GAO examine a number of issues concerning the
pricing and marketing of milk in the United States. Specifically, they
asked us to update the information contained in our 2001 report, entitled
Dairy Industry: Information on Milk Prices and Changing Market Structure

(GAO-01-561, June 15, 2001), and to address other issues. This report
examines (1) what portion of the retail price of fluid milk is received by
dairy farmers, dairy cooperatives, wholesale milk processors, and
retailers in selected markets throughout the United States, how this
distribution has changed over the period of our review, and the
relationships among price changes at these levels; (2) how various
factors, such as costs, influence the price of milk as it moves from the
farm to the consumer, as well as how these and other factors affect the
extent to which changes in price are transmitted among levels as milk
moves from the farm to the consumer; and (3) how changes in dairy policies
and alternative policy options have affected or might affect farm income,
federal costs, economic efficiency, and consumer prices, among other
policy considerations.1 It also updates other information on milk prices
included in our June 2001 report.

To update our information on the price distribution among the various
levels of the milk marketing chain, we analyzed milk prices in 15 selected
markets nationwide: Atlanta, Georgia; Boston, Massachusetts; Charlotte,
North Carolina; Cincinnati, Ohio; Dallas, Texas; Denver, Colorado; Miami,
Florida; Milwaukee, Wisconsin; Minneapolis, Minnesota; New Orleans,
Louisiana; Phoenix, Arizona; Salt Lake City, Utah; San Diego, California;
Seattle, Washington; and Washington, D.C.2 In selecting these markets, we
ensured that (1) they provided national geographic coverage; (2) at least
one market was located in each of the federal milk marketing orders
(FMMOs) as they existed during most of the period from October 2000
through May 2004; (3) the selected markets included both state and
federally regulated markets; and (4) these areas represented similar
marketing areas for which we reported information in our June 2001 report.
For the 15 markets, we collected data on the prices received by

1This report does not specifically address consolidation and concentration
trends in the dairy industry and what is known about the impact of
concentration on milk prices, as the 2001 report did. However, the report
discusses market concentration as one of the factors that influence the
price of milk as it moves from the farm to the consumer. (See app. V.)

2Portions of some of these market areas lie in adjacent states.

Appendix I
Objectives, Scope, and Methodology

farmers, cooperatives, wholesale milk processors, and retailers for
October 2000 through May 2004. We limited our data collection efforts to
the prices of whole, 2 percent, 1 percent, and skim milk because sales of
these milk types constitute over 93 percent of fluid milk sales annually.
We also confined our analysis to the prices of these milk types sold in
gallon containers because milk sold in gallon containers accounts for
about 65 percent of fluid milk products sold under FMMOs.3

There is no precise method for calculating the price that farmers receive
for raw milk that is ultimately processed and sold in fluid milk products
because dairy farmers receive a blend price for their milk, which is the
average price for milk used for fluid and manufactured products.
Therefore, any calculation of the value received by farmers for raw milk
that is to be used for fluid milk products is necessarily only
approximate. To estimate a farm price for raw milk used in fluid milk
products, we used data provided by the U.S. Department of Agriculture's
(USDA) Agricultural Marketing Service (AMS). AMS developed an adjustment,
which accounts for various charges such as hauling and marketing fees,
that we subtracted from the announced cooperative Class I price to obtain
the estimated farm price for raw milk used in fluid milk products for each
of the selected markets in our review except San Diego, which is not part
of the FMMO system.4 AMS's adjustment accounts for farm-to-plant hauling
costs, cooperative dues and capital assessment, mandatory advertising and
promotion costs, competitive and receiving credits, and a representative
estimate of the value of reimbursements to cooperatives for the services
performed for handlers and for transportation costs not covered by the
order minimum price.5 Most of the items that make up the adjustment are
not available for the specific fluid milk market that we selected, but
rather are based on information collected for milk used over wider
geographical areas. Therefore, an order-wide value used for any of these
items provides an estimate rather than the actual value for this item.
Also, the values for

3This estimate is based on USDA data for the month of November for
selected years between 1991 and 2001. It was the most recent data we could
obtain.

4Under the federal milk marketing orders, a classified pricing plan
provides different classes and minimum prices for milk depending on how
the milk is used. Milk used in fluid products is placed in Class I, which
is the highest-priced class. Milk used for various manufactured products
is placed in lower-priced classes.

5Handler is the federal order term for cooperatives, processors, or
dealers of milk who commonly purchase raw milk and sell pasteurized milk
and milk products. See 7 C.F.R. S:1000.9 for a more complete definition.

Appendix I
Objectives, Scope, and Methodology

two of the adjustment items-reimbursements to cooperatives for services
performed for handlers and for transportation costs not covered by the
order minimum price-were not readily available so they were estimated
indirectly based on other reported data and, in some cases, on anecdotal
information provided by industry members. However, despite these
limitations, AMS believes that the estimated farm price is a good
representation of the price that dairy farmers receive for raw milk used
in fluid milk products. For the farm price for San Diego-a state-regulated
market-we used mailbox price data collected by the California Department
of Food and Agriculture. The mailbox price is the weighted average of the
prices received by dairy farmers in the market for all of their raw milk
sold and therefore is computed as the total net dollars received for milk
divided by the total pounds of milk marketed.6 This price is likely to be
lower than the price received for milk used for fluid purposes because the
prices for milk used for manufacturing purposes are generally lower.
However, it is the best measure we could obtain.

To determine cooperative prices, we used AMS data on announced cooperative
prices to represent prices that wholesale milk processors paid to
cooperatives. Wholesale milk processors in federally regulated markets
generally purchase milk from cooperatives and pay the federal minimum
price for milk plus premiums that are negotiated between cooperatives and
wholesale milk processors. The announced cooperative price is the Class I
milk price announced by the major cooperative in each of the markets. This
price does not apply to all Class I sales in federally regulated markets
and is not necessarily the price actually received for all of the milk
sold by the major cooperative; the announced cooperative prices have not
been verified by USDA as actually having been paid by processors. For San
Diego, we used the minimum fluid prices established by the state of
California. Data on the premiums paid in excess of these minimums were not
available for this market. (See app. V for a detailed discussion of
overorder premiums.)

To determine wholesale prices, we used the prices paid at Department of
Defense Commissary Agency locations. The Defense Commissary Agency
purchases milk under competitive and noncompetitive contracts with

6The mailbox price information collected by the California Department of
Food and Agriculture does not account for all of the various factors that
influence farm prices. As a result, it does not provide as good a measure
of the price that farmers receive based on the value of their milk as the
AMS adjustment does for the FMMO markets where it is available.

Appendix I
Objectives, Scope, and Methodology

wholesalers. We used commissary prices as surrogates for privately
established wholesale prices because (1) defense commissaries sell
groceries at a standard 5 percent markup from cost to active and retired
military personnel and (2) wholesale price data are considered proprietary
by industry officials and were not available to us. The commissary network
of stores ranks twelfth in the United States in sales volume for
supermarket chains. We selected 39 different commissary locations near the
15 markets we reviewed, and the Defense Commissary Agency provided us with
weekly prices paid by consumers at these locations for gallons of whole, 2
percent, 1 percent, and skim milk.7 We averaged these weekly prices to
obtain monthly prices. We then adjusted these monthly prices to account
for the 5 percent markup. Where we had multiple commissary locations for a
market, we averaged the adjusted monthly prices to obtain a wholesale
price for the market. We recognize that these locations may not provide an
ideal match with other price data analyzed for a given location; for
example, in some markets the available commissary locations were not in
close proximity to the selected marketing areas. Also, wholesale
processors may provide these commissary locations with different levels of
service than they do for retailers in these markets.8 In such cases, the
prices paid by these commissaries for fluid milk may have been different
than the prices that retailers in the selected markets paid to their
wholesale suppliers. However, these were the best wholesale data that we
could obtain. In those locations where commissaries sold more than one
brand of milk, we used the price for the brand that had the highest sales
volume for a particular period.

For retail prices, we contracted with Information Resources, Inc., a
private data collection and analysis company, to obtain average weekly
retail prices for whole, 2 percent, 1 percent, and skim milk sold in
gallon containers. These data represented a weighted average of prices at
supermarkets with yearly sales exceeding $2 million for the markets
included in our analysis. We then averaged these weekly prices to obtain
monthly prices. We were unable to obtain data from some types of
nonsupermarket retailers such as mass merchandisers, thus the retail

7The 39 commissary locations were selected either because they were within
the marketing area for one of our 15 selected markets (as defined by our
source for retail pricing data, Information Resources, Inc.), or because
they were the closest location available.

8In addition to shipping the products to stores, some wholesalers provide
in-store services, including unloading the milk on the store dock,
restocking the dairy case, and removing outdated or leaking containers.

Appendix I
Objectives, Scope, and Methodology

pricing data that we present may not be representative of fluid milk
prices at those locations. Figure 1 shows the locations of the 15 selected
markets, the corresponding commissaries, and the federal milk marketing
order areas.

                                   Appendix I
                       Objectives, Scope, and Methodology

Figure 1: Selected Milk Markets, Corresponding Defense Commissaries Used
for Our Analysis, and Federal Milk Marketing Order Areas Prior to April
2004

Source: GAO, based on USDA and Defense Commissary Agency information.

Appendix I
Objectives, Scope, and Methodology

Note: The Western Order was terminated as of April 1, 2004. However,
because the order existed during the majority of our period of analysis we
chose to include the data in this report. Farm and cooperative level
prices for this order for April and May 2004 were provided by USDA based
on what they would have been had the order still existed.

To determine (1) the portion of the retail price of a gallon of milk
received by farmers, cooperatives, wholesale milk processors, and
retailers; (2) how changes in retail and farm prices affect the
farm-to-retail price spread; and (3) how price changes at any level of the
marketing chain correlate to changes in prices at other levels, we limited
our analysis to 2 percent milk, which currently represents the largest
volume of reduced-fat milk sold nationwide. Therefore, our analysis of 2
percent prices may not necessarily reflect pricing patterns and trends for
the other three kinds of milk. Appendix II includes graphs that show the
relationships among the farm, cooperative, wholesale, and retail prices
for a gallon of 2 percent milk for each of the 15 markets. Because farm
and cooperative prices reflect a higher milkfat content than that in 2
percent milk, we adjusted these prices to reflect the value of removing
milkfat and replacing it with skim milk.9 This adjustment allowed us to
use farm and cooperative prices that were comparable to the wholesale and
retail prices for our analysis.

To determine the degree that farm and retail prices had changed and the
effect these changes had on the farm-to-retail price spread from October
2000 through May 2004 for each of the 15 markets, we used a statistical
procedure to estimate farm-level and retail prices at the beginning and
end of the period.10 We relied on estimated rather than actual prices to
reduce the influence of the starting and ending months and years selected
for our analysis in markets in which milk prices varied from month to
month. We

9For San Diego, we used prices that were adjusted for 2 percent milkfat
and 10 percent nonfat milk solids so that they were comparable with the
prices of retail milk sold in California.

10We used a regression procedure for each market to determine whether the
price could be reliably predicted as a function of time for both
farm-level and retail prices. This procedure allowed us to estimate
initial and final prices for farm-level and retail prices that take into
account the variability in these price series during the 44-month period.
This regression procedure was done on time and time-squared because we
observed that prices generally fell for part of the period and then rose.
Including the time-squared variable in the regression provided a better
fit for the data. A statistically significant relationship indicates that
we found a consistent association, either up or down, between price and
time. For statistically significant relationships, we calculated a final
price estimate (computed for the last month of our data series) and
compared to an estimated price calculated for the first month. In the
absence of a statistically significant relationship-when no consistent
association was found-we treated initial and final estimates of price the
same, even if actual beginning and final prices differ.

Appendix I
Objectives, Scope, and Methodology

used the differences between the estimated initial and final prices to
represent the changes during the period. When our statistical procedure
did not find a consistent association between prices and time, we treated
the difference in the estimated initial and final prices as zero. We
calculated the change in the farm-to-retail price spread as the estimated
retail price difference minus the estimated farm price difference.

To describe the relationship between price changes at any given level in
the milk marketing chain and price changes at the other levels, we tested
for correlations between price changes at the various levels for each of
the 15 markets included in our analysis. Specifically, we calculated
coefficients describing the degree of correlation between changes in farm
prices and price changes at the cooperative, wholesale, and retail levels;
price changes at the cooperative level and price changes at the wholesale
and retail levels; and price changes at the wholesale and retail levels.
In appendix II, we report those correlation coefficients and indicate
which are statistically different from zero at the 95 percent confidence
level.

To update information provided in our June 2001 report on the retail
prices for four kinds of milk, we analyzed the retail price data that we
obtained from Information Resources, Inc. We array these data in appendix
III for each of the selected 15 markets for October 2000 through May 2004.

To update information provided in our June 2001 report on average monthly
and annual farm and cooperative prices, and wholesale and retail prices
for different kinds of milk, we analyzed data obtained from USDA, the
California Department of Food and Agriculture, the Department of Defense
Commissary Agency, and Information Resources, Inc. We report these data in
appendix IV for each of the selected 15 markets for October 2000 through
May 2004.

To update our information on the major factors influencing milk prices and
explore price transmission within the milk marketing chain, we conducted
more than 50 interviews with national dairy experts working with the

Appendix I
Objectives, Scope, and Methodology

federal and state governments,11 cooperatives,12 processors,13
retailers,14 or industry groups,15 or in academia.16 We also reviewed a
number of relevant studies and publications from USDA and other sources.
Where possible, we obtained data on production costs, services provided by
cooperatives, as well as inputs to processing and retailing. We also
obtained information on concentration and market power at each level of
the milk marketing chain. We present information on the factors
influencing the price of milk in appendix V.

To compare the results and methodologies of various studies looking at the
issue of price transmission in fluid milk marketing from the farm to the
retail level, we performed a technical review of 14 academic studies
conducted over the past 10 years, looking at model descriptions,
assumptions, and results. We also spoke with the economists involved in
these studies concerning their model results and the causes of differences
in fluid milk price transmission across markets. The scope of these
studies encompassed national, regional, and city-level models of fluid
milk price transmission. Appendix VI provides a summary of our review of
price transmission and the various price transmission studies.

To identify and examine the effects of federal dairy program changes and
alternative policy options, we contacted many of the same dairy experts
previously mentioned. We also conducted an extensive literature search and
reviewed more than 50 relevant studies and other publications we
identified. We qualitatively analyzed the effects of federal dairy program

11Federal government sources included officials from USDA's AMS, Economic
Research Service, and Farm Service Agency.

12Cooperatives we contacted included Dairy Farmers of America; California
Dairies, Inc.; Agri-Mark; Land O'Lakes, Inc.; and Prairie Farms.

13Fluid milk processors and processors of manufactured dairy products that
we contacted included HP Hood LLC and Leprino Foods, as well as
cooperatives and retailers that process and manufacture these products.

14Retailers that we contacted included Wal-Mart, Albertsons, the H. E.
Butt Grocery Company, and Demoulas Super Markets, Inc.

15Industry groups that we contacted included the National Milk Producers
Federation, the International Dairy Foods Association, and the Food
Marketing Institute.

16Academic dairy experts that we contacted included professors from
Cornell University, North Carolina State University, Ohio State
University, Texas A&M University, University of Connecticut, University of
Wisconsin-Madison, and the University of California-Davis.

Appendix I
Objectives, Scope, and Methodology

changes and policy options we identified on six main policy
considerations: farm income, milk production, federal costs, price
volatility, economic efficiency, and consumer prices. We evaluated impacts
on these policy considerations under both high-and low-price scenarios,
over the short and long terms. We identified these policy considerations
by reviewing previous GAO reports, relevant studies, and legislation, as
well as through our conversations with dairy policy experts. Different
stakeholders in the dairy policy arena may have alternative views on the
relative importance of these policy considerations, as well as other
considerations that we did not include, which could lead to differing
perspectives on these options. In addition, the potential effects of
policy options on these considerations could vary depending upon economic
conditions and other policy decisions. We compared the dairy policy
options we identified against a baseline scenario of the policies in place
as of August 2004: FMMO regulations, a Milk Income Loss Contract (MILC)
program that is scheduled to expire at the end of fiscal year 2005, a
price support program at $9.90 per hundredweight, a Dairy Export Incentive
Program (DEIP), trade restrictions, and milk regulatory policies in some
states. We include a discussion of the effects of recent federal dairy
program changes and alternative policy options in appendix VII.

We conducted our review from September 2003 through October 2004 in
accordance with generally accepted government auditing standards. We did
not independently verify the data we received from various sources.
However, we discussed with these sources the measures they take to ensure
the accuracy of the data, and these measures seemed reasonable.
Additionally, we consulted with the following dairy experts concerning the
results of our analysis of price transmission within the milk marketing
chain and the effects of changes in federal dairy programs and alternative
policy options:

o 	Ed Jesse, Ph.D., Professor, Department of Agricultural and Applied
Economics, University of Wisconsin-Madison;

o 	Daniel Lass, Ph.D., Professor, College of Natural Resources and the
Environment, University of Massachusetts, Amherst;

o 	Richard Sexton, Ph.D., Professor, Department of Agricultural and
Resource Economics, University of California-Davis; and

o 	Mark Stephenson, Ph.D., Senior Extension Associate, Department of
Applied Economics and Management, Cornell University.

Appendix II

Analysis of Prices at Four Marketing Levels for 2 Percent Milk in Selected
Markets

This appendix reports on our analysis of prices at four marketing levels
for a gallon of 2 percent milk in 15 selected markets for October 2000
through May 2004. Our analysis includes information on (1) the portion of
the retail price of a gallon of milk received by farmers, cooperatives,
wholesale milk processors, and retailers; (2) how changes in farm and
retail milk prices affect the farm-to-retail milk price spread; and (3)
how price changes at any level of the marketing chain correlate with
changes in prices at other levels.

We limited our analysis to gallons of 2 percent milk because sales of milk
with reduced fat content account for nearly 52 percent of all sales of
fluid milk and sales of 2 percent milk account for about 62 percent of
these reduced-fat sales. The farm and cooperative prices used in our
analysis and presented in this appendix have been adjusted to reflect 2
percent milkfat. This analysis may not reflect pricing patterns and trends
for other kinds of milk. We present complete data for prices for all four
types of milk-whole, 2 percent, 1 percent, and skim-in appendix III.

Portion Received by Farmers, Cooperatives, Wholesale Milk Processors and
Retailers

Between October 2000 and May 2004, on average, our data suggest that
farmers received 45.9 percent, cooperatives 6.1 percent, wholesale
processors 35.6 percent, and retailers 12.5 percent of the retail price of
a gallon of 2 percent milk in the 15 markets we reviewed.1 However, these
percentages varied depending on the specific market. For example, the
farmers' portion ranged from 36.0 percent to 58.6 percent, while retailers
in 12 markets received anywhere from 3.5 percent to 44.1 percent.2 In
comparison, the average percentages we reported in 2001 for the period
March 1998 through September 2000 were 43 percent, 5 percent, 33 percent,
and 19 percent, respectively, for farmers, cooperatives, wholesale
processors, and retailers. Table 1 summarizes the price breakdown for each
market.

1These figures represent unweighted averages of the percentages for each
market.

2See explanatory note (a) to table 1. In some places (3 out of 15
markets), retailers experienced negative returns.

                                  Appendix II
                  Analysis of Prices at Four Marketing Levels
                     for 2 Percent Milk in Selected Markets

Table 1: Portion of the Retail Price of a Gallon of 2 Percent Milk
Received by Farmers, Cooperatives, Wholesale Processors, and Retailers for
15 Markets, October 2000 through May 2004

                                                                     Subtotal 
                                                                     received 
                                              Percent                      by 
                                              received              wholesale 
                         Percent Percent              by Percent   processors 
                        received received      wholesale received         and 
    Selected market   by farmers by           processors        by  retailers 
                                 cooperatives            retailers 
      Atlanta, Ga.          42.7          7.4       34.8      15.1 
     Boston, Mass.          51.2          4.5       24.0      20.4 
    Charlotte, N.C.         41.1          7.5       46.6       4.8 
    Cincinnati, Ohio        45.8          7.6       39.0       7.7 
      Dallas, Tex.          55.9          5.3       49.5   (10.7)a 
     Denver, Colo.          39.4          4.5       25.7      30.3 
      Miami, Fla.           50.8          9.5       45.0    (5.3)a 
    Milwaukee, Wisc.        52.1          7.9       36.5       3.5 
Minneapolis, Minn.       58.6          7.8       64.1   (30.5)a 
    New Orleans, La.        36.8          9.5       36.6      17.1 
     Phoenix, Ariz.         42.5          3.3       34.8      19.4 
    Salt Lake City,         44.8          3.6       43.0       8.6 
          Utah                                                     
San Diego, Calif.        36.0          6.8       28.1      29.1 
           b                                                       
     Seattle, Wash.         36.7          3.2       16.0      44.1 
    Washington, D.C.        53.4          3.3       10.2      33.1 
Average for the 15                                              
        markets             45.9          6.1       35.6      12.5 

Source: GAO's analysis of farm and cooperative price data provided by USDA
(for the San Diego market, the mailbox and Southern California Class I
price data were provided by the California Department of Food and
Agriculture), wholesale price data provided by the Defense Commissary
Agency, and retail price data provided by Information Resources, Inc.

Note: Percentages may not total 100 due to rounding.

aOur analysis found that retailers in the Dallas, Miami, and Minneapolis
markets received a negative return on a gallon of 2 percent milk for the
44-month period. Officials with one major retailer in the Dallas market
indicated that a price war was occurring in Dallas during this time
period. We were unable to obtain explanations for the negative returns in
Miami and Minneapolis. These negative returns may be the result of
problems with comparability between commissary and supermarket data.

bFigures for the San Diego market may understate the percent received by
cooperatives and overstate the percent received by wholesale processors
because the cooperative price data for this market did not include
estimates of over-order premiums paid to cooperatives. On the other hand,
figures for this market may overstate the percent received by cooperatives
and understate the percent received by farmers because unlike the farm
price data for other markets, the farm price data for San Diego represent
a weighted average of milk sold for all uses. See appendix I.

                                  Appendix II
                  Analysis of Prices at Four Marketing Levels
                     for 2 Percent Milk in Selected Markets

Changes in Farm and Retail Prices and the Price Spread

From October 2000 through May 2004, the spread between farm and retail
milk prices increased in 12 of the 15 markets.3 However, in some of the 12
markets, the spread between farm and retail milk prices increased
dramatically and then moderated. In 9 of the 15 markets, retail prices
showed a statistically significant increase. In 4 of the remaining
markets, retail prices decreased over time; in the other 2 markets, retail
prices showed no statistically significant change. At the same time, farm
prices decreased in 12 of the 15 markets and increased in the remaining 3
markets over the 44-month period. However, these declining farm prices
began to moderate, or, in most cases, began to rise during the latter
months of our period of analysis. Table 2 provides these data for selected
markets.4

Table 2: Changes in the Farm-to-Retail Price Spread for a Gallon of 2
Percent Milk for 15 Markets, October 2000 through May 2004

Change in farm Change in retail Change in the farm-to-retail Selected
market prices prices price spread

                     Atlanta, Ga.     $0.03            ($0.21)        ($0.24) 
                    Boston, Mass.    (0.06)          0.12      
                  Charlotte, N.C.    (0.05)          0.20      
                 Cincinnati, Ohio    (0.07)         (0.12)             (0.05) 
                     Dallas, Tex.     0.02           0.24      
                    Denver, Colo.    (0.06)          0.16      
                      Miami, Fla.    (0.01)          0.25      
                 Milwaukee, Wisc.    (0.03)          0.05      
               Minneapolis, Minn.    (0.02)           a        
                 New Orleans, La.    (0.10)          0.16      
                   Phoenix, Ariz.    (0.07)         (0.01)     

3The price spread of a commodity represents differences in prices between
two levels of the marketing chain. These differences reflect the costs
incurred and profits received through adding value by processing,
transporting, and distributing a commodity.

4The values used to calculate the changes in farm and retail prices in
table 2 are based on statistical estimates of initial and final prices for
these two levels, not the actual prices recorded in the first and last
month of the period for which we have data. The method for calculating
these estimates is described in appendix I. The changes in the
farm-to-retail price spread are the differences between the changes in
retail and farm prices from October 2000 to May 2004 and consequently are
determined from the statistically estimated initial and final prices for
the retail and farm levels.

                                  Appendix II
                  Analysis of Prices at Four Marketing Levels
                     for 2 Percent Milk in Selected Markets

(Continued From Previous Page)

                        Change in farm Change in retail

Change in the farm-to-retail price spread

                         Selected market prices prices

               Salt Lake City, Utah    (0.04)           a       
                  San Diego, Calif.     0.03         (0.16)            (0.19) 
                     Seattle, Wash.    (0.08)         0.10      
                   Washington, D.C.    (0.03)         0.28      

Source: GAO's analysis of farm and cooperative price data provided by USDA
(for the San Diego market, the mailbox and Southern California Class I
price data were provided by the California Department of Food and
Agriculture), wholesale price data provided by the Defense Commissary
Agency, and retail price data provided by Information Resources, Inc.

Note: Differences between changes in farm and retail prices may not add to
the change in the farm-toretail price spread due to rounding.

aNo statistically significant change was observed in the price spread over
the 44-month period.

Correlation between Price Changes at the Four Marketing Levels

We found that price changes generally correlated across levels in the
marketing chain, with the strongest correlations occurring between
adjacent levels. The values of correlation coefficients presented are
estimates of the degree that price changes at one level in the milk
marketing chain are associated with price changes at other levels. The
higher the coefficient, the closer the association between changes in
prices at different levels. Changes in cooperative prices, in general,
were strongly correlated with changes in wholesale prices.5 However,
changes in cooperative prices correlated less strongly with changes in
retail prices. As discussed in appendix V, many factors other than farm or
wholesale prices influence the retail price of fluid milk.

Correlation coefficients between prices at different marketing levels
varied across markets. For example, correlations between cooperative and
wholesale prices in individual markets range from a high of 0.982 to a low
of -0.031. We ranked the 15 markets by the extent of correlation between
cooperative and wholesale prices. The correlation coefficient for the

5As discussed in appendix I, farm prices were estimated by subtracting out
certain cost factors, provided by USDA, from the announced cooperative
Class I prices. To the extent that these factors do not change much over
time, it would be expected that price changes at the farm level would
closely correlate with price changes at the cooperative level, except for
San Diego. As discussed in appendix I, we estimated the farm and
cooperative prices differently in San Diego because California is not part
of the federal milk marketing order system. The approach that we used for
San Diego would not be expected to generate a similarly close correlation
between farm and cooperative prices. Because farm prices for raw milk used
for fluid purposes are not directly observed but are derived from
cooperative prices, correlations of wholesale and retail prices with
cooperative prices provide a better picture of the price relationship
across levels in the milk marketing chain.

Appendix II
Analysis of Prices at Four Marketing Levels
for 2 Percent Milk in Selected Markets

market that fell in the middle of this ranking was 0.788. In comparison,
the market in the middle of a similar ranking for the time period analyzed
in our 2001 report had a lower correlation coefficient between these
prices, 0.716. Similarly, correlations between cooperative and retail
prices in individual markets range from a high of 0.879 to a low of 0.214.
We did a comparable ranking of the 15 markets by the extent of correlation
between cooperative and retail prices. The correlation coefficient for the
market that fell in the middle of this ranking was 0.588. In comparison,
the market in the middle of a similar ranking for the time period
presented in our 2001 report again had a lower correlation coefficient
between these prices, 0.483. Tables 3 through 5 present data from our
correlation analysis of price changes across marketing levels.

Table 3: Correlation between Farm Price Changes and Changes in
Cooperative, Wholesale, and Retail Prices for a Gallon of 2 Percent Milk
for 15 Markets, October 2000 through May 2004

                                Correlation      Correlation      Correlation 
                           coefficients for coefficients for coefficients for 
           Selected market cooperative      wholesale prices    retail prices 
                           prices                            
              Atlanta, Ga.           0.999*           0.525*           0.537* 
             Boston, Mass.           1.000*           0.691*           0.780* 
           Charlotte, N.C.           0.995*           0.693*           0.778* 
          Cincinnati, Ohio           1.000*           0.983*           0.589* 
              Dallas, Tex.           0.998*           0.813*           0.713* 
             Denver, Colo.           0.999*           0.932*           0.428* 
               Miami, Fla.           0.999*          (0.024)           0.826* 
          Milwaukee, Wisc.           0.999*           0.683*           0.881* 
        Minneapolis, Minn.           0.999*            0.220           0.503* 
          New Orleans, La.           0.993*           0.784*           0.863* 
            Phoenix, Ariz.           1.000*           0.542*            0.217 
           Salt Lake City,                                   
                      Utah           1.000*           0.829*           0.360* 
         San Diego, Calif.           0.407*           0.432*           0.341* 
            Seattle, Wash.           1.000*           0.939*            0.249 
          Washington, D.C.           1.000*           0.962*           0.810* 

Source: GAO's analysis of farm and cooperative price data provided by USDA
(for the San Diego market, the mailbox and Southern California Class I
price data were provided by the California Department of Food and
Agriculture), wholesale price data provided by the Defense Commissary
Agency, and retail price data provided by Information Resources, Inc.

Note: In calculating the correlation coefficients for each market, we
omitted the months for which data were missing.

Appendix II
Analysis of Prices at Four Marketing Levels
for 2 Percent Milk in Selected Markets

*Indicates that the correlation coefficient estimated for this time period
is statistically significant at the 5 percent level (i.e., the probability
that the two price series are uncorrelated is less than 5 percent).

Table 4: Correlation between Cooperative Price Changes and Changes in
Wholesale and Retail Prices for a Gallon of 2 Percent Milk for 15 Markets,
October 2000 through May 2004

                             Correlation coefficients             Correlation 
                                                  for        coefficients for 
             Selected market         wholesale prices           retail prices 
                Atlanta, Ga.                   0.546*                  0.543* 
               Boston, Mass.                   0.694*                  0.775* 
             Charlotte, N.C.                   0.697*                  0.778* 
            Cincinnati, Ohio                   0.982*                  0.588* 
                Dallas, Tex.                   0.818*                  0.725* 
               Denver, Colo.                   0.939*                  0.436* 
                 Miami, Fla.                  (0.031)                  0.823* 
            Milwaukee, Wisc.                   0.682*                  0.879* 
          Minneapolis, Minn.                    0.228                  0.499* 
            New Orleans, La.                   0.788*                  0.861* 
              Phoenix, Ariz.                   0.540*                   0.214 
        Salt Lake City, Utah                   0.830*                  0.360* 
           San Diego, Calif.                   0.969*                  0.480* 
              Seattle, Wash.                   0.939*                   0.250 
            Washington, D.C.                   0.962*                  0.809* 

Source: GAO's analysis of cooperative price data provided by USDA (for the
San Diego market, the Southern California Class I price data were provided
by the California Department of Food and Agriculture), wholesale price
data provided by the Defense Commissary Agency, and retail price data
provided by Information Resources, Inc.

Note: In calculating the correlation coefficients for each market, we
omitted the months for which data were missing.

*Indicates that the correlation coefficient estimated for this time period
is statistically significant at the 5 percent level (i.e., the probability
that the two price series are uncorrelated is less than 5 percent).

                                  Appendix II
                  Analysis of Prices at Four Marketing Levels
                     for 2 Percent Milk in Selected Markets

Table 5: Correlation between Wholesale and Retail Price Changes for a
Gallon of 2 Percent Milk for 15 Markets, October 2000 through May 2004

           Selected market Correlation coefficients for retail prices

                              Atlanta, Ga. 0.319*

                              Boston, Mass. 0.348*

                             Charlotte, N.C. 0.746*

                            Cincinnati, Ohio 0.544*

                              Dallas, Tex. 0.802*

                              Denver, Colo. 0.429*

                              Miami, Fla. (0.003)

                            Milwaukee, Wisc. 0.744*

                            Minneapolis, Minn. 0.205

                            New Orleans, La. 0.734*

                             Phoenix, Ariz. 0.419*

                           Salt Lake City, Utah 0.270

                            San Diego, Calif. 0.497*

                              Seattle, Wash. 0.183

                            Washington, D.C. 0.778*

Source: GAO's analysis of wholesale price data provided by the Defense
Commissary Agency and retail price data provided by Information Resources,
Inc.

Note: In calculating the correlation coefficients for each market, we
omitted the months for which data were missing.

*Indicates that the correlation coefficient estimated for this time period
is statistically significant at the 5 percent level (i.e., the probability
that the two price series are uncorrelated is less than 5 percent).

Comparison of Average Tables 6 through 10 show the average annual price
for a gallon of 2 percent

milk in the 15 markets for each of the four marketing levels during part
ofAnnual and Monthly 2000, all of 2001, 2002, and 2003, and part of 2004.
Figures 2 through 16 Prices for 2 Percent present average monthly data for
the period October 2000 through May Milk 2004 on farm, cooperative,
wholesale, and retail prices for gallons of 2

percent milk in each of the 15 markets. Gaps in any of the lines shown in

the figures indicate that data were unavailable for those months.

Appendix II
Analysis of Prices at Four Marketing Levels
for 2 Percent Milk in Selected Markets

Table 6: Average Annual Farm, Cooperative, Wholesale, and Retail Prices
for a Gallon of 2 Percent Milk in Selected Markets, 2000

Selected market Farm price Cooperative price Wholesale price Retail price

                     Atlanta, Ga.   $1.05      $1.25      $2.04         $2.67 
                    Boston, Mass.    1.25       1.35       1.96          2.23 
                  Charlotte, N.C.    1.09       1.25       2.06          2.58 
                 Cincinnati, Ohio    1.03       1.21             a       2.24 
                     Dallas, Tex.    1.08       1.20       2.05          1.74 
                    Denver, Colo.    0.99       1.11       1.73          2.45 
                      Miami, Fla.    1.22       1.44       2.50          2.33 
                 Milwaukee, Wisc.    1.05       1.18       1.79          1.94 
               Minneapolis, Minn.    1.01       1.12       2.05          1.77 
                 New Orleans, La.    1.02       1.22       2.25          2.61 
                   Phoenix, Ariz.    1.01       1.09       1.83          2.09 
             Salt Lake City, Utah    0.98       1.06       1.97          2.25 
                San Diego, Calif.    0.92       1.16       1.90          2.54 
                   Seattle, Wash.    1.02       1.10       1.47          2.56 
                 Washington, D.C.    1.21       1.28       1.51          2.11 

Source: GAO's analysis of farm and cooperative price data provided by USDA
(for the San Diego market, the mailbox and Southern California Class I
price data were provided by the California Department of Food and
Agriculture), wholesale price data provided by the Defense Commissary
Agency, and retail price data provided by Information Resources, Inc.

Note: Averages for 2000 were calculated using data from October through
December, the portion of 2000 included in our analysis.

aWe were unable to obtain data from the Defense Commissary Agency to
represent wholesale prices for 2 percent milk in the Cincinnati market
during 2000.

Table 7: Average Annual Farm, Cooperative, Wholesale, and Retail Prices
for a Gallon of 2 Percent Milk in Selected Markets, 2001

                                        Cooperative    Wholesale    Retail 
       Selected market      Farm price          price       price    price 
        Atlanta, Ga.             $1.14          $1.34       $2.50    $2.68 
        Boston, Mass.             1.29           1.39        1.95     2.32 
       Charlotte, N.C.            1.18           1.34        2.32     2.62 
      Cincinnati, Ohio            1.15           1.32        2.24     2.28 
        Dallas, Tex.              1.17           1.28        2.15     1.98 
        Denver, Colo.             1.12           1.23        1.93     2.54 
         Miami, Fla.              1.31           1.53        2.56     2.40 

Appendix II
Analysis of Prices at Four Marketing Levels
for 2 Percent Milk in Selected Markets

(Continued From Previous Page)

                                             Cooperative   Wholesale   Retail 
               Selected market  Farm price          price       price   price 
              Milwaukee, Wisc.         1.17          1.32        2.17    2.07 
            Minneapolis, Minn.         1.14          1.27        2.21    1.73 
              New Orleans, La.         1.11          1.31        2.32    2.68 
                Phoenix, Ariz.         1.14          1.21        1.91    2.42 
          Salt Lake City, Utah         1.10          1.18        2.11    2.24 
             San Diego, Calif.         1.05          1.25        2.04    2.84 
                Seattle, Wash.         1.11          1.20        1.66    2.73 
              Washington, D.C.         1.34          1.42        1.71    2.30 

Source: GAO's analysis of farm and cooperative price data provided by USDA
(for the San Diego market, the mailbox and Southern California Class I
price data were provided by the California Department of Food and
Agriculture), wholesale price data provided by the Defense Commissary
Agency, and retail price data provided by Information Resources, Inc.

Table 8: Average Annual Farm, Cooperative, Wholesale, and Retail Prices
for a Gallon of 2 Percent Milk in Selected Markets, 2002

Selected market Farm price Cooperative price Wholesale price Retail price

                  Atlanta, Ga.    $1.04       $1.22       $2.01         $2.47 
                 Boston, Mass.    1.11        1.21        1.77           2.33 
               Charlotte, N.C.    1.01        1.22        2.55           2.59 
              Cincinnati, Ohio    0.95        1.11        1.98           2.34 
                  Dallas, Tex.    1.08        1.17        2.19           2.03 
                 Denver, Colo.    0.93        1.05        1.66           2.56 

Miami, Fla. 1.17 1.41 2.56 2.36 Milwaukee, Wisc. 0.98 1.14 1.81 2.01
Minneapolis, Minn. 0.95 1.10 2.38 1.95 New Orleans, La. 0.92 1.19 2.22
2.70 Phoenix, Ariz. 0.94 1.02 1.88 2.59 Salt Lake City, Utah 0.91 0.99
1.92 2.29 San Diego, Calif. 0.88 1.06 1.84 2.66 Seattle, Wash. 0.93 1.01
1.39 2.74 Washington, D.C. 1.16 1.23 1.44 2.28

Source: GAO's analysis of farm and cooperative price data provided by USDA
(for the San Diego market, the mailbox and Southern California Class I
price data were provided by the California Department of Food and
Agriculture), wholesale price data provided by the Defense Commissary
Agency, and retail price data provided by Information Resources, Inc.

Appendix II
Analysis of Prices at Four Marketing Levels
for 2 Percent Milk in Selected Markets

Table 9: Average Annual Farm, Cooperative, Wholesale, and Retail Prices
for a Gallon of 2 Percent Milk in Selected Markets, 2003

Selected market Farm price Cooperative price Wholesale price Retail price

                      Atlanta, Ga.   $1.07      $1.25      $2.02        $2.49 
                     Boston, Mass.    1.14       1.25       1.79         2.33 
                   Charlotte, N.C.    1.05       1.27       2.65         2.63 
                  Cincinnati, Ohio    0.98       1.16       2.00         2.34 
                      Dallas, Tex.    1.09       1.19       2.26         2.02 
                     Denver, Colo.    0.97       1.08       1.68         2.55 
                       Miami, Fla.    1.20       1.43       2.56         2.47 
                  Milwaukee, Wisc.    1.01       1.19       1.87         2.00 
                Minneapolis, Minn.    0.99       1.13       2.36         1.62 
                  New Orleans, La.    0.94       1.24       2.22         2.68 
                    Phoenix, Ariz.    0.97       1.05       1.97         2.45 
              Salt Lake City, Utah    0.95       1.03       1.99         2.14 
                 San Diego, Calif.    0.93       1.12       1.84         2.69 
                    Seattle, Wash.    0.95       1.04       1.46         2.88 
                  Washington, D.C.    1.19       1.27       1.46         2.33 

Source: GAO's analysis of farm and cooperative price data provided by USDA
(for the San Diego market, the mailbox and Southern California Class I
price data were provided by the California Department of Food and
Agriculture), wholesale price data provided by the Defense Commissary
Agency, and retail price data provided by Information Resources, Inc.

Table 10: Average Annual Farm, Cooperative, Wholesale, and Retail Prices
for a Gallon of 2 Percent Milk in Selected Markets, 2004

Selected market Farm price Cooperative price Wholesale price Retail price

           Atlanta, Ga.          $1.18      $1.35      $2.21         $2.54 
          Boston, Mass.          1.27        1.38       1.90          2.43 
         Charlotte, N.C.         1.13        1.35       2.84          2.87 
         Cincinnati, Ohio        1.09        1.27       2.21          1.95 
           Dallas, Tex.          1.20        1.30       2.47          2.13 
          Denver, Colo.          1.06        1.20       2.04          2.72 
           Miami, Fla.           1.30        1.52       2.56          2.63 
         Milwaukee, Wisc.        1.15        1.31       2.13          2.13 
        Minneapolis, Minn.       1.14        1.29       2.43          1.93 
         New Orleans, La.        1.03        1.33       2.15          2.87 
          Phoenix, Ariz.         1.08        1.17       2.17          2.25 

                                  Appendix II
                  Analysis of Prices at Four Marketing Levels
                     for 2 Percent Milk in Selected Markets

(Continued From Previous Page)

Selected market Farm price Cooperative price Wholesale price Retail price

              Salt Lake City, Utah    1.07       1.14       2.16         2.19 
                 San Diego, Calif.    1.12       1.22       1.92    
                    Seattle, Wash.    1.06       1.15       1.74    
                  Washington, D.C.    1.31       1.39       1.66    

Source: GAO's analysis of farm and cooperative price data provided by USDA
(for the San Diego market, the mailbox and Southern California Class I
price data were provided by the California Department of Food and
Agriculture), wholesale price data provided by the Defense Commissary
Agency, and retail price data provided by Information Resources, Inc.

Note: The 2004 averages were calculated using data from January through
May of that year, the portion of 2004 included in our analysis.

Figure 2: Farm, Cooperative, Wholesale, and Retail Prices for a Gallon of
2 Percent Milk for Atlanta, Georgia, October 2000 through May 2004

Dollars per gallon

                                      Dec.

.gAu

                                      Jan.

Jan.

                           Feb. Dec. Sept.Oct.No Feb.

.gAu

                                      Feb.

.NoDec.

                              Sept. Jan.Feb. Sept.

Oct.

.MarApr

.gAu

                                      Oct.

.Apr

                                   . Oct.No.

.MarApr

.yMa

                                       .

.No

.Mar

                                    Jan.Dec.

.Mar

.Apr

en

                                     yMaJu

enJu

yl

en

lyJu

                                     yMaJu

lyJu

                                     yMaJu

2000 2001 2002 2003 2004

Retail price

Wholesale price

Cooperative price

Farm price

Source: GAO analysis of data provided by USDA, the Defense Commissary
Agency, and Information Resources, Inc.

Note: For Atlanta, Georgia, the farm price is the USDA estimated farm
Class I price adjusted to 2 percent milkfat content; the cooperative price
is the announced cooperative Class I price for Atlanta adjusted to 2
percent milkfat content; the wholesale price is the average of the prices
paid by the commissaries at Fort Gillem and Fort McPherson, adjusted for a
5 percent markup; and the retail price is the price collected by
Information Resources, Inc., for the Atlanta market.

                                  Appendix II
                  Analysis of Prices at Four Marketing Levels
                     for 2 Percent Milk in Selected Markets

Figure 3: Farm, Cooperative, Wholesale, and Retail Prices for a Gallon of
2 Percent Milk for Boston, Massachusetts, October 2000 through May 2004

Dollars per gallon Oct.

                                     Dec. .

Dec.

                                Feb. Jan. Sept.

.yAprMa

                        Oct. Feb. Dec. Feb. Dec.Jan.Feb.

Jan.

.Apr

.MarApr

                                   Sept.Oct.

.No

.MarApr

.gAu

                                       .

.No

                                     Jan. .

.gAu

                                     Oct.No

.Mar

.Mar

.gAu

                                     Sept.

.No

en

en

en

                                     yMaJu

yl

Ju

yl

Ju

                                   yMaJuyMaJu

lyJu2000 2001 2002 2003 2004

Retail price

Wholesale price

Cooperative price

Farm price

Source: GAO analysis of data provided by USDA, the Defense Commissary
Agency, and Information Resources, Inc.

Note: For Boston, Massachusetts, the farm price is the USDA estimated farm
Class I price adjusted to 2 percent milkfat content; the cooperative price
is the effective cooperative Class I price for Boston adjusted to 2
percent milkfat content during the period that the Northeast Interstate
Dairy Compact was effective (July 1997 through September 2001); the
cooperative price after September 2001 is the announced cooperative Class
I price for Boston adjusted to 2 percent milkfat content; the wholesale
price is the price paid by the commissary at Hanscom Air Force Base,
adjusted for a 5 percent markup; and the retail price is the price
collected by Information Resources, Inc., for the Boston market.

                                  Appendix II
                  Analysis of Prices at Four Marketing Levels
                     for 2 Percent Milk in Selected Markets

Figure 4: Farm, Cooperative, Wholesale, and Retail Prices for a Gallon of
2 Percent Milk for Charlotte, North Carolina, October 2000 through May
2004

Dollars per gallon

                                   Jan. Sept.

Oct.No

                               Dec.Oct. Feb. Dec.

Oct.

.gAu

                               Sept. Dec. . Feb.

.No

                                      Feb.

.Mar

                                   Sept. Jan.

.gAu

                                     Oct.No

.Mar

.Apr

                                      Jan.

.Apr

.gAu

Dec.

Jan.

.Mar

.Apr

.MarApr

                                      Feb.

.yMa

.No

                                       .

en

en

                                       e

ynMaJu

                                     yMaJu

yl

Ju

yl

Ju

lyJu

                                     yMaJu

2000 2001 2002 2003 2004

Retail price

Wholesale price

Cooperative price

Farm price

Source: GAO analysis of data provided by USDA, the Defense Commissary
Agency, and Information Resources, Inc.

Note: For Charlotte, North Carolina, the farm price is the USDA estimated
farm Class I price adjusted to 2 percent milkfat content; the cooperative
price is the announced cooperative Class I price for Charlotte adjusted to
2 percent milkfat content; the wholesale price is the price paid by the
Fort Bragg North Post Store, adjusted for a 5 percent markup; and the
retail price is the price collected by Information Resources, Inc., for
the Charlotte market.

                                  Appendix II
                  Analysis of Prices at Four Marketing Levels
                     for 2 Percent Milk in Selected Markets

Figure 5: Farm, Cooperative, Wholesale, and Retail Prices for a Gallon of
2 Percent Milk for Cincinnati, Ohio, October 2000 through May 2004

Dollars per gallon Dec.

                                   Jan. Dec.

Oct.

                                   Dec. Jan.

.Apr

                         Sept. Feb.MarFeb. Sept.Oct.No

.Mar

                                  Jan. Oct. .

Jan.

.Mar

.gAu

                                     Sept.

.No

                                  Feb. .AprMa

.AprMa

                                     . Feb.

.gAu

.Mar

.g

.yAprMa

.No

                                    Dec.Oct.

.No

enJu

                                       e

yl

en

yJu

yl

Ju

yJu

ynMaJu

                                     lyJuAu

2000 2001 2002 2003 2004

Retail price

Wholesale price

Cooperative price

Farm price

Source: GAO analysis of data provided by USDA, the Defense Commissary
Agency, and Information Resources, Inc.

Note: For Cincinnati, Ohio, the farm price is the USDA estimated farm
Class I price adjusted to 2 percent milkfat content; the cooperative price
is the announced cooperative Class I price for Cincinnati adjusted to 2
percent milkfat content; the wholesale price is the price paid by the
commissary at Wright-Patterson Air Force Base, adjusted for a 5 percent
markup; and the retail price is the price collected by Information
Resources, Inc., for the Cincinnati market. We were unable to obtain data
from the Defense Commissary Agency to represent wholesale prices for 2
percent milk in the Cincinnati market during 2000 and for January and
February 2001.

                                  Appendix II
                  Analysis of Prices at Four Marketing Levels
                     for 2 Percent Milk in Selected Markets

Figure 6: Farm, Cooperative, Wholesale, and Retail Prices for a Gallon of
2 Percent Milk for Dallas, Texas, October 2000 through May 2004

Dollars per gallon

                             Sept.Dec.Oct. Jan.Feb.

.gAu

                                      Dec.

.gAu

                                      Feb.

Dec.

                             Sept. Feb. Oct.No Jan.

.gAu

.Mar

                                      Dec.

.

                                   Oct. Jan.

..MarApr

.yAprMa

                                      Feb.

.Apr

.Apr

                                     Sept.

Oct.No

.Mar

.No

.No

.MarJan.

                                     .yMaJu

en

en

en

yl

Ju

                                     yMaJu

lyJu

                                     yMaJu

yl

Ju2000 2001 2002 2003 2004

Retail price

Wholesale price

Cooperative price

Farm price

Source: GAO analysis of data provided by USDA, the Defense Commissary
Agency, and Information Resources, Inc.

Note: For Dallas, Texas, the farm price is the USDA estimated farm Class I
price adjusted to 2 percent milkfat content; the cooperative price is the
announced cooperative Class I price for Dallas adjusted to 2 percent
milkfat content; the wholesale price is the average of the prices paid by
the two commissaries at Fort Hood, adjusted for a 5 percent markup; and
the retail price is the price collected by Information Resources, Inc.,
for the Dallas market.

                                  Appendix II
                  Analysis of Prices at Four Marketing Levels
                     for 2 Percent Milk in Selected Markets

Figure 7: Farm, Cooperative, Wholesale, and Retail Prices for a Gallon of
2 Percent Milk for Denver, Colorado, October 2000 through May 2004

Dollars per gallon Oct. Jan.

                                      Feb.

.MarApr

                                   Dec. Feb.

.g

                                      Dec.

.NoDec.

                                      Jan.

.gAu

                                Sept. Jan.Feb..

.g

                         Sept.Oct. Jan. Sept. Feb. Oct.

..MarApr

.Mar

.Apr

                                   Oct. Dec.

.No

.No

.Mar

.Apr

en

.No

en

                                     lyJuAu

                                     yMaJu

                                     lyJuAu

en

                                     yMaJu

lyJu

yMa

                                     yMaJu

2000 2001 2002 2003 2004

Retail price Wholesale price Cooperative price Farm price

Source: GAO analysis of data provided by USDA, the Defense Commissary
Agency, and Information Resources, Inc.

Note: For Denver, Colorado, the farm price is the USDA estimated farm
Class I price adjusted to 2 percent milkfat content; the cooperative price
is the announced cooperative Class I price for Denver adjusted to 2
percent milkfat content; the wholesale price is the average of the prices
paid by the commissaries at Fort Carson, Peterson Air Force Base, Buckley
Air Force Base, and the U.S. Air Force Academy, adjusted for a 5 percent
markup; and the retail price is the price collected by Information
Resources, Inc., for the Denver market.

                                  Appendix II
                  Analysis of Prices at Four Marketing Levels
                     for 2 Percent Milk in Selected Markets

Figure 8: Farm, Cooperative, Wholesale, and Retail Prices for a Gallon of
2 Percent Milk for Miami, Florida, October 2000 through May 2004

Dollars per gallon Oct.No

                                   Feb. Dec.

.Mar

                                 Dec.Jan. Jan.

.

                                   Feb. Feb.

.gAu

.No

                                   Sept.Oct.

.MarApr

                                   Sept. Dec.

.Apr

.No

.gAu

.No

.Mar

.yAprMaJan.

.gAu

.Apr

                                     . Oct.

Dec.

                                 Sept. Jan.Feb.

.Mar

                                      Oct.

en

en

lyJu

en

yl

Ju

                                     yMaJu

                                   yMaJuyMaJu

yl

Ju2000 2001 2002 2003 2004

Retail price Wholesale price Cooperative price Farm price

Source: GAO analysis of data provided by USDA, the Defense Commissary
Agency, and Information Resources, Inc.

Note: For Miami, Florida, the farm price is the USDA estimated farm Class
I price adjusted to 2 percent milkfat content; the cooperative price is
the announced cooperative Class I price for Miami adjusted to 2 percent
milkfat content; the wholesale price is the price paid by the commissary
at the Naval Air Station, Key West, adjusted for a 5 percent markup; and
the retail price is the price collected by Information Resources, Inc.,
for the Miami market.

                                  Appendix II
                  Analysis of Prices at Four Marketing Levels
                     for 2 Percent Milk in Selected Markets

Figure 9: Farm, Cooperative, Wholesale, and Retail Prices for a Gallon of
2 Percent Milk for Milwaukee, Wisconsin, October 2000 through May 2004

Dollars per gallon

                                   Jan. Sept.

Jan.

..MarApr

                              Dec. Oct. Dec. Feb.

.NoDec.

                                 Feb. Oct. Dec.

.Mar

.gAu

                                      Feb.

.Apr

.gAu

                                   Sept.Oct.

.No

                                    Jan.Feb.

.Mar

.Apr

.No

                                   Jan.Sept.

.No

.gAu

.Mar

.yAprMaOct.

en

en

                                     yMaJu

en

yl

Ju

                                     yMaJu

                                     yMaJu

yl

Ju

yl

Ju2000 2001 2002 2003 2004

Retail price Wholesale price Cooperative price Farm price

Source: GAO analysis of data provided by USDA, the Defense Commissary
Agency, and Information Resources, Inc.

Note: For Milwaukee, Wisconsin, the farm price is the USDA estimated farm
Class I price adjusted to 2 percent milkfat content; the cooperative price
is the announced cooperative Class I price for Milwaukee adjusted to 2
percent milkfat content; the wholesale price is the price paid by the
commissary at the Naval Station Great Lakes, adjusted for a 5 percent
markup; and the retail price is the price collected by Information
Resources, Inc., for the Milwaukee market.

                                  Appendix II
                  Analysis of Prices at Four Marketing Levels
                     for 2 Percent Milk in Selected Markets

Figure 10: Farm, Cooperative, Wholesale, and Retail Prices for a Gallon of
2 Percent Milk for Minneapolis, Minnesota, October 2000 through May 2004

Dollars per gallon

.gAu

.gAu

Jan.

                         Sept. Jan. Dec. Feb. Jan.Feb.

Dec.

                              Oct. Dec. Sept.Oct.

.No

.Mar

.Apr

                                Sept.Oct.NoDec.

..MarApr

                                   Feb. Feb.

.Mar

                                      Jan.

Oct.

.gAu

.No

.Apr

                                      . .

en

.No

en

.MarApr

en

yl

Ju

yl

Ju

                                     yMaJu

yMa

                                     yMaJu

lyJu

                                     yMaJu

2000 2001 2002 2003 2004

Retail price Wholesale price Cooperative price Farm price

Source: GAO analysis of data provided by USDA, the Defense Commissary
Agency, and Information Resources, Inc.

Note: For Minneapolis, Minnesota, the farm price is the USDA estimated
farm Class I price adjusted to 2 percent milkfat content; the cooperative
price is the announced cooperative Class I price for Minneapolis adjusted
to 2 percent milkfat content; the wholesale price is the price paid by the
commissary at Fort McCoy, adjusted for a 5 percent markup; and the retail
price is the price collected by Information Resources, Inc., for the
Minneapolis market.

                                  Appendix II
                  Analysis of Prices at Four Marketing Levels
                     for 2 Percent Milk in Selected Markets

Figure 11: Farm, Cooperative, Wholesale, and Retail Prices for a Gallon of
2 Percent Milk for New Orleans, Louisiana, October 2000 through May 2004

Dollars per gallon Oct.NoDec.Jan.

                         Feb. Jan. Jan. Sept.Oct.NoDec.

.

.AprMa

.gAu

                                   Dec. Feb.

.gAu

                              Sept.Oct.NoDec. Feb.

.gAu

                                    Jan.Feb.

.Mar

                                  Sept.Oct. .

..MarApr

.MarApr

.No

                                       .

.MarApr

                                       .

.yMa

en

en

yl

Ju

en

                                     yMaJu

yJu

yl

Ju

lyJu

                                     yMaJu

2000 2001 2002 2003 2004

Retail price Wholesale price Cooperative price Farm price

Source: GAO analysis of data provided by USDA, the Defense Commissary
Agency, and Information Resources, Inc.

Note: For New Orleans, Louisiana, the farm price is the USDA estimated
farm Class I price adjusted to 2 percent milkfat content; the cooperative
price is the announced cooperative Class I price for New Orleans adjusted
to 2 percent milkfat content; the wholesale price is the average of the
prices paid by the Naval Support Activities, New Orleans; the Naval
Construction Battalion Center, Gulfport; and Keesler Air Force Base
commissaries, adjusted for a 5 percent markup; and the retail price is the
price collected by Information Resources, Inc., for the New Orleans
market.

                                  Appendix II
                  Analysis of Prices at Four Marketing Levels
                     for 2 Percent Milk in Selected Markets

Figure 12: Farm, Cooperative, Wholesale, and Retail Prices for a Gallon of
2 Percent Milk for Phoenix, Arizona, October 2000 through May 2004

Dollars per gallon

Oct.No

Jan.

                     Dec.Jan.Feb. Sept.Oct. Dec. Feb. Feb.

Dec.

                                 Feb.Mar Sept.

.gAu

                            Jan. Sept.Oct.NoDec.Jan.

.gAu

                                     Oct. .

.gAu

.MarApr

.Apr

                                       .

.Apr

.No

.No

.Mar

.MarApr

.yMa

.v

                                       .

en

en

en

                                     yMaJu

                                   yMaJuyMaJu

lyJu

yl

Ju

lyJu2000 2001 2002 2003 2004

Retail price Wholesale price Cooperative price Farm price

Source: GAO analysis of data provided by USDA, the Defense Commissary
Agency, and Information Resources, Inc.

Note: For Phoenix, Arizona, the farm price is the USDA estimated farm
Class I price adjusted to 2 percent milkfat content; the cooperative price
is the announced cooperative Class I price for Phoenix adjusted to 2
percent milkfat content; the wholesale price is the average of the prices
paid by the commissaries at Luke Air Force Base and Davis-Monthan Air
Force Base, adjusted for a 5 percent markup; and the retail price is the
price collected by Information Resources, Inc., for the Phoenix market.

                                  Appendix II
                  Analysis of Prices at Four Marketing Levels
                     for 2 Percent Milk in Selected Markets

Figure 13: Farm, Cooperative, Wholesale, and Retail Prices for a Gallon of
2 Percent Milk for Salt Lake City, Utah, October 2000 through May 2004

Dollars per gallon

                                    Dec.Jan.

.gAu

Dec.

                                   Oct. Feb.

.gAu

             Sept.Oct.NoDec.Jan. Sept. Dec.Jan.Feb. Sept. Feb. Feb.

Jan.

.MarApr

                                      Oct.

.MarOct.No.

.gAu

.No

.Mar

.Apr

                                       .

.yAprMa

                                      . .

.MarApr

.No

en

en

en

                                   yMaJuyMaJu

yl

Ju

yl

Ju

                                     yMaJu

lyJu2000 2001 2002 2003 2004

Retail price

Wholesale price

Cooperative price

Farm price

Source: GAO analysis of data provided by USDA, the Defense Commissary
Agency, and Information Resources, Inc.

Note: For Salt Lake City, Utah, the farm price is the USDA estimated farm
Class I price adjusted to 2 percent milkfat content; the cooperative price
is the announced cooperative Class I price for Salt Lake City adjusted to
2 percent milkfat content; the wholesale price is the price paid by the
commissary at Hill Air Force Base, adjusted for a 5 percent markup; and
the retail price is the price collected by Information Resources, Inc.,
for the Salt Lake City market.

                                  Appendix II
                  Analysis of Prices at Four Marketing Levels
                     for 2 Percent Milk in Selected Markets

Figure 14: Farm, Cooperative, Wholesale, and Retail Prices for a Gallon of
2 Percent Milk for San Diego, California, October 2000 through May 2004

Dollars per gallon Oct.No

                                      Feb.

Dec.

Jan.

Feb.

          Sept.Oct. Dec.Jan. Sept.Oct. Dec.Jan. Sept.Oct. Dec.Jan.Feb.

.gAu

                                      Feb.

.gAu

.Mar

.yAprMa.

                                       .

.Apr

.Apr

.gAu

.Mar

.Mar

.No

.MarApr

.No

.No

en

enJu

                                    yMaJuly

                                     yMaJu

en

lyJu

                                     yMaJu

lyJu2000 2001 2002 2003 2004

Retail price Wholesale price Cooperative price Farm price

Source: GAO analysis of data provided by the California Department of Food
and Agriculture, the Defense Commissary Agency, and Information Resources,
Inc.

Note: For San Diego, California, the farm price is the California
Department of Food and Agriculture mailbox price adjusted to 2 percent
milkfat content; the cooperative price is the Southern California Class I
price adjusted to 2 percent milkfat content; the wholesale price is the
average of the prices paid by the commissaries at the Naval Base, San
Diego; Marine Corps Base, Camp Pendleton; Naval Outlying Landing Field,
Imperial Beach; Marine Corps Air Station, Miramar; Naval Air Station,
North Island; and by the San Onofre Commissary, Camp Pendleton, adjusted
for a 5 percent markup; and the retail price is the price collected by
Information Resources, Inc., for the San Diego market.

                                  Appendix II
                  Analysis of Prices at Four Marketing Levels
                     for 2 Percent Milk in Selected Markets

Figure 15: Farm, Cooperative, Wholesale, and Retail Prices for a Gallon of
2 Percent Milk for Seattle, Washington, October 2000 through May 2004

Dollars per gallon

                                   Oct. Dec.

Oct.

Dec.

Jan.

                             Feb. Oct. Dec.Jan.Feb.

.gAu

.No

.gAu

                              Sept. Sept. Jan.Feb.

.gAu

                               Sept. Dec.Jan.Feb.

.Mar

.Apr

.No

.Mar

.Apr

.No

.Mar

.Apr

.No

.MarApr

.yMa

en

                                      Oct.

en

en

                                     yMaJu

yl

Ju

                                     yMaJu

yl

Ju

                                     yMaJu

lyJu2000 2001 2002 2003 2004

Retail price Wholesale price Cooperative price Farm price

Source: GAO analysis of data provided by USDA, the Defense Commissary
Agency, and Information Resources, Inc.

Note: For Seattle, Washington, the farm price is the USDA estimated farm
Class I price adjusted to 2 percent milkfat content; the cooperative price
is the announced cooperative Class I price for Seattle adjusted to 2
percent milkfat content; the wholesale price is the average of the prices
paid by the commissaries at the Naval Station Everett, Smokey Point
Support Center; Fort Lewis; and McChord Air Force Base, adjusted for a 5
percent markup; and the retail price is the price collected by Information
Resources, Inc., for the Seattle market.

                                  Appendix II
                  Analysis of Prices at Four Marketing Levels
                     for 2 Percent Milk in Selected Markets

Figure 16: Farm, Cooperative, Wholesale, and Retail Prices for a Gallon of
2 Percent Milk for Washington, D.C., October 2000 through May 2004

Dollars per gallon Oct. Dec.Jan.

                 Feb. Oct. Dec.Jan.Feb. Sept.Oct. Jan.Feb. Feb.

.Mar

.gAu

                            Dec. Sept.Oct. Dec.Jan.

.No

.Apr

.gAu

                                     Sept.

.No

.Mar

.Apr

.No

.Mar

.Apr

.gAu

.No

.yMa

.MarApr

en

en

en

                                yMaJuyMaJuyMaJu

lyJu

lyJu

yl

Ju2000 2001 2002 2003 2004

Retail price Wholesale price Cooperative price Farm price

Source: GAO analysis of data provided by USDA, the Defense Commissary
Agency, and Information Resources, Inc.

Note: For Washington, D.C., the farm price is the USDA estimated farm
Class I price adjusted to 2 percent milkfat content; the cooperative price
is the announced cooperative Class I price for Washington, D.C., adjusted
to 2 percent milkfat content; the wholesale price is the average of the
prices paid by the commissaries at Bolling Air Force Base; Walter Reed
Army Medical Center; Fort Myer; Fort Belvoir; the Marine Corps Base,
Quantico; Andrews Air Force Base; Aberdeen Proving Ground; the Naval
Station, Annapolis; Fort Meade; and Fort Detrick, adjusted for a 5 percent
markup; and the retail price is the price collected by Information
Resources, Inc., for the Washington, D.C., market.

Appendix III

Retail Prices for Four Kinds of Milk in Selected Markets

This appendix updates information provided in our June 2001 report on the
average retail prices for whole, 2 percent, 1 percent, and skim milk in 15
selected markets for October 2000 through May 2004. We found that retail
pricing patterns varied significantly across markets. For example, in the
Boston market from October 2000 through May 2004, the average price for 2
percent milk was generally the same as the average price for 1 percent
milk; however, whole and skim milk prices were generally lower. On the
other hand, for this period in the San Diego market, the average price of
2 percent milk was generally lower than the prices of whole and 1 percent
milk, but higher than skim milk prices. Figures 17 through 31 provide
information on the average retail price for the four kinds of milk in the
15 selected markets for October 2000 through May 2004.

Figure 17: Retail Prices in Atlanta, Georgia, for Whole, 2 Percent, 1
Percent, and Skim Milk, October 2000 through May 2004 Dollars per gallon

0.00

                                   Feb. Dec.

Dec.

Jan.

                              Sept. Dec. Dec. Feb.

.gAu

                          Oct.No Jan.Feb.Oct.No. Jan.

.gAu

                                 Sept.Oct. Jan.

.Apr

                                     Sept.

.gAu

Feb.

.Mar

.Mar

.Apr

.No

.Mar

.Mar

.yAprMa

.No

.AprOct.

                                     .yMaJu

en

yl

Ju

enJu

                                     yMaJu

yl

yl

                                     yMaJu

enJu2000 2001 2002 2003 2004

Whole milk 2 percent milk 1 percent milk Skim milk

Source: GAO analysis of data provided by Information Resources, Inc.

Note: The retail price is the price collected by Information Resources,
Inc., for the Atlanta market.

     Appendix III Retail Prices for Four Kinds of Milk in Selected Markets

Figure 18: Retail Prices in Boston, Massachusetts, for Whole, 2 Percent, 1
Percent, and Skim Milk, October 2000 through May 2004

Dollars per gallon

                       Jan. Sept. Feb.Feb. Feb. Sept.Oct.

.No

                                      Dec.

.gAu

                                     Oct.No

Jan.

.Apr

.gAu

                        Sept. Dec.Jan. . Feb. .Dec.Jan.

Dec.

.Mar

.MarApr

.gAu

.MarApr

                                       .

.Mar

.yAprMa.

                                      Oct.

.vNo

enJu

Oct.No

enJu

en

                                     yMaJu

yl

Juyl

                                     yMaJu

yMa

lyJu2000 2001 2002 2003 2004

Whole milk 2 percent milk 1 percent milk Skim milk

Source: GAO analysis of data provided by Information Resources, Inc.

Note: The retail price is the price collected by Information Resources,
Inc., for the Boston market.

     Appendix III Retail Prices for Four Kinds of Milk in Selected Markets

Figure 19: Retail Prices in Charlotte, North Carolina, for Whole, 2
Percent, 1 Percent, and Skim Milk, October 2000 through May 2004

Dollars per gallon

                                      Feb.

Oct.No

                               Oct.No Oct.NoDec.

Jan.

.gAu

                                Dec. Sept. Jan.

.gAu

                         Sept. Jan.Feb.Feb. Sept. Jan.

.gAu

                                      Dec.

..MarApr.Dec.

                                   Oct. Feb.

.Mar

.Apr

.Mar

.Apr

.Mar

                                      . .

.No

.Apr

en

en

enJu

                                     yMaJu

                                    yMaJuly

lyJu

yMa

yl

Ju

                                     yMaJu

2000 2001 2002 2003 2004

Whole milk 2 percent milk 1 percent milk Skim milk

Source: GAO analysis of data provided by Information Resources, Inc.

Note: The retail price is the price collected by Information Resources,
Inc., for the Charlotte market.

     Appendix III Retail Prices for Four Kinds of Milk in Selected Markets

Figure 20: Retail Prices in Cincinnati, Ohio, for Whole, 2 Percent, 1
Percent, and Skim Milk, October 2000 through May 2004 Dollars per gallon

                             Feb. Jan. Dec.Feb.Jan.

.gAu

                                   Sept.Oct.

Oct.No

Dec.

Jan.

.gAu

                                 Dec. Feb. Dec.

Oct.No

.gAu

.No

.Apr

                                      Feb.

.

                                     Sept.

.Mar

.Apr

.No

.Mar

.yAprMa

.Mar

.Apr

                                      Oct.

.Mar

                                . Jan.Sept.yMaJu

en

en

                                     yMaJu

en

                                     yMaJu

lyJu

yl

Ju

yl

Ju2000 2001 2002 2003 2004

Whole milk 2 percent milk 1 percent milk Skim milk

Source: GAO analysis of data provided by Information Resources, Inc.

Note: The retail price is the price collected by Information Resources,
Inc., for the Cincinnati market.

     Appendix III Retail Prices for Four Kinds of Milk in Selected Markets

Figure 21: Retail Prices in Dallas, Texas, for Whole, 2 Percent, 1 Percent, and
                    Skim Milk, October 2000 through May 2004

Dollars per gallon

                                   Dec. Sept.

Oct.No

                                   Jan. Dec.

Dec.

.gAu

.gAu

                                 Oct. Dec.Jan.

.g

                                     Sept.

.No

                                      Jan.

.

.No

                                      Feb.

.No

                                      Feb.

.Apr

.Apr

.MarApr

.yMaJan.

                                      Feb.

.Mar

                               Sept. . Oct. Feb.

.MarApr

.Mar

                                      Oct.

en

                                     yMaJu

en

yl

Ju

yl

JuAu

                                     yMaJu

lyJu

                                     yMaJu

en2000 2001 2002 2003 2004

Whole milk 2 percent milk 1 percent milk Skim milk

Source: GAO analysis of data provided by Information Resources, Inc.

Note: The retail price is the price collected by Information Resources,
Inc., for the Dallas market.

     Appendix III Retail Prices for Four Kinds of Milk in Selected Markets

 Figure 22: Retail Prices in Denver, Colorado, for Whole, 2 Percent, 1 Percent,
                  and Skim Milk, October 2000 through May 2004

Dollars per gallon

                                      Feb.

Oct.

Jan.

                                   Oct. Jan.

Dec.

                                      Oct.

.gAu

                                      Jan.

.gAu

                                 Feb.Sept. Dec.

.No

                                      Dec.

.MarApr

                                 Oct. Jan.Feb..

.No

                                  Feb. . Sept.

.gAu

                                   Sept. Dec.

.Mar

.MarApr

.Mar

.Apr

.No

.yAprMa

.vNo

en

en

lyJu

                                     yMaJu

                                     yMaJu

en

yl

Ju

                                     yMaJu

lyJu2000 2001 2002 2003 2004

Whole milk 2 percent milk 1 percent milk Skim milk

Source: GAO analysis of data provided by Information Resources, Inc.

Note: The retail price is the price collected by Information Resources,
Inc., for the Denver market.

     Appendix III Retail Prices for Four Kinds of Milk in Selected Markets

Figure 23: Retail Prices in Miami, Florida, for Whole, 2 Percent, 1 Percent, and
                    Skim Milk, October 2000 through May 2004

Dollars per gallon Oct.

                                   Sept. Dec.

.gAu

                                      Dec.

Dec.

Jan.

.gAu

                     Oct. Feb. Dec.Jan.Feb. Feb.Feb. . Jan.

.gAu

                                Oct. Sept. Jan.

.No

.MarApr

.No

.MarApr

                                    Sept. .

.Mar

.yAprMa

.MarApr

.No

.No

                                     Oct..

en

lyJu

en

                                   yMaJuyMaJu

yl

Ju

enJu

                                     yMaJu

yl2000 2001 2002 2003 2004

Whole milk 2 percent milk 1 percent milk Skim milk

Source: GAO analysis of data provided by Information Resources, Inc.

Note: The retail price is the price collected by Information Resources,
Inc., for the Miami market.

     Appendix III Retail Prices for Four Kinds of Milk in Selected Markets

Figure 24: Retail Prices in Milwaukee, Wisconsin, for Whole, 2 Percent, 1
Percent, and Skim Milk, October 2000 through May 2004

Dollars per gallon

                                   Dec. Feb.

Oct.

.gAu

                           Oct.No Jan.Feb. Sept. Jan.

Dec.

                            Jan.Feb. Sept.Oct. Oct.

.Mar

                                   Sept. Dec.

.Apr

.gAu

                                      Dec.

Jan.

.Apr

                                       .

.Mar

.gAu

.Mar

.Mar

.yAprMa

.Apr

.No

.No

.vNo

                                      Feb.

enJu

en

                                     yMaJu

en

lyJu

                                     yMaJu

lyJu

                                     yMaJu

yl2000 2001 2002 2003 2004

Whole milk 2 percent milk 1 percent milk Skim milk

Source: GAO analysis of data provided by Information Resources, Inc.

Note: The retail price is the price collected by Information Resources,
Inc., for the Milwaukee market.

     Appendix III Retail Prices for Four Kinds of Milk in Selected Markets

Figure 25: Retail Prices in Minneapolis, Minnesota, for Whole, 2 Percent,
1 Percent, and Skim Milk, October 2000 through May 2004

Dollars per gallon

                                      Jan.

Jan.

                            Feb. Dec. Feb. Jan. Dec.

Oct.No

                              Dec. Sept. Jan.Feb.

.Dec.

.gAu

                                     Sept.

.gAu

.Mar

.Apr

.Apr

                                   Oct. Feb.

.Mar

.Mar

.Mar

.Apr

                                       .

.yAprMa

.gAu

                                     .Sept.

.No

                                  Oct.NoOct.No

                                     yMaJu

en

                                     yMaJu

en

en

yl

Ju

                                     yMaJu

lyJu

yl

Ju2000 2001 2002 2003 2004

Whole milk 2 percent milk 1 percent milk Skim milk

Source: GAO analysis of data provided by Information Resources, Inc.

Note: The retail price is the price collected by Information Resources,
Inc., for the Minneapolis market.

     Appendix III Retail Prices for Four Kinds of Milk in Selected Markets

Figure 26: Retail Prices in New Orleans, Louisiana, for Whole, 2 Percent,
1 Percent, and Skim Milk, October 2000 through May 2004

Dollars per gallon

                                Sept. Dec. Feb.

Oct.

Dec.

.gAu

                              Jan.Feb. Sept. Dec.

Jan.

                                      Feb.

.gAu

                               Oct. Oct. Jan.Feb.

.Apr

                                      Jan.

.Mar

                                   Sept. Dec.

.Mar

.gAu

.Mar

.No

                                      . .

.No

.Apr

.yAprMa

.No

.MarApr

                                      Oct.

                                       No

en

enJu

enJu

                                   yMaJuyMaJu

yl

                                     yMaJu

yl

yl

Ju2000 2001 2002 2003 2004

Whole milk 2 percent milk 1 percent milk Skim milk

Source: GAO analysis of data provided by Information Resources, Inc.

Note: The retail price is the price collected by Information Resources,
Inc., for the New Orleans market.

     Appendix III Retail Prices for Four Kinds of Milk in Selected Markets

Figure 27: Retail Prices in Phoenix, Arizona, for Whole, 2 Percent, 1
Percent, and Skim Milk, October 2000 through May 2004 Dollars per gallon

                               Jan.Jan. Dec. Feb.

.gAu

Dec.

                                   Sept. Dec.

.Apr

                                Sept. Feb. Sept.

Oct.

.NoJan.

                                      Oct.

                                  No. Jan.Feb.

.gAu

.Apr

.gAu

                                 Oct.Oct. Dec.

.MarApr

.yMa

.Feb.Mar

.Apr

.Mar

.No

.Mar

.No

en

en

en

yl

Ju

                                     yMaJu

yl

Ju

                                     yMaJu

lyJu

                                     yMaJu

2000 2001 2002 2003 2004

Whole milk 2 percent milk 1 percent milk Skim milk

Source: GAO analysis of data provided by Information Resources, Inc.

Note: The retail price is the price collected by Information Resources,
Inc., for the Phoenix market.

     Appendix III Retail Prices for Four Kinds of Milk in Selected Markets

Figure 28: Retail Prices in Salt Lake City, Utah, for Whole, 2 Percent, 1
             Percent, and Skim Milk, October 2000 through May 2004

Dollars per gallon Oct.No

                                 Jan.Feb. Dec.

.gAu

                                   Sept. Feb.

.Dec.

                                Sept. Jan..Dec.

.AprJan.

                                      Feb.

.MarApr

                                     Oct.No

.No

.Mar

.Apr

                                Oct. Dec.. Oct.

.gAu

                                     Sept.

.No

.Mar

.gAu

                                      Jan.

.MarApr

                                      Feb.

en

.yMa

yl

Ju

enJu

lyJu

yl

en

                                   yMaJuyMaJu

                                     yMaJu

2000 2001 2002 2003 2004

Whole milk 2 percent milk 1 percent milk Skim milk

Source: GAO analysis of data provided by Information Resources, Inc.

Note: The retail price is the price collected by Information Resources,
Inc., for the Salt Lake City market.

     Appendix III Retail Prices for Four Kinds of Milk in Selected Markets

Figure 29: Retail Prices in San Diego, California, for Whole, 2 Percent, 1
Percent, and Skim Milk, October 2000 through May 2004 Dollars per gallon

                                      Feb.

Oct.No

Dec.

                                   Dec. Jan.

Jan.

.gAu

.gAu

                                   Sept. Jan.

.gAu

                                 Feb.Feb. Sept.

.MarApr

                             Oct. Jan. . Oct. Dec.

.yMa

                               . Sept. Dec. Oct.

                                     NoFeb.

.MarApr

                                       .

.No.

.v

.MarApr

.No

.MarApr

en

                                       e

en

ynMaJu

lyJu

yl

Ju

                                     yMaJu

lyJu

                                     yMaJu

2000 2001 2002 2003 2004

Whole milk 2 percent milk 1 percent milk Skim milk

Source: GAO analysis of data provided by Information Resources, Inc.

Note: The retail price is the price collected by Information Resources,
Inc., for the San Diego market.

     Appendix III Retail Prices for Four Kinds of Milk in Selected Markets

    Figure 30: Retail Prices in Seattle, Washington, for Whole, 2 Percent, 1
             Percent, and Skim Milk, October 2000 through May 2004

Dollars per gallon Oct. Jan.

                    Feb. Sept.Sept.Oct.No Dec.Jan.Feb. Feb.

.gAu

                                      Oct.

Dec.

.gAu

.gAu

                                     Oct.No

.No

                              Dec.Jan.Feb.Dec.Jan.

.MarApr

                                    . Sept.

.No

.Apr

.Mar

                                       .

.Mar

                                       .

.Apr

..MarApr

enJu

en

                                     yMaJu

en

yl

Ju

                                    ly yMaJu

lyJu

yMa

                                     yMaJu

2000 2001 2002 2003 2004

Whole milk 2 percent milk 1 percent milk Skim milk

Source: GAO analysis of data provided by Information Resources, Inc.

Note: The retail price is the price collected by Information Resources,
Inc., for the Seattle market.

     Appendix III Retail Prices for Four Kinds of Milk in Selected Markets

 Figure 31: Retail Prices in Washington, D.C., for Whole, 2 Percent, 1 Percent,
                  and Skim Milk, October 2000 through May 2004

Dollars per gallon

                                    Feb.Mar

Oct.No

Jan.

                                    Dec.Jan.

.gAu

Oct.No

                                  Dec.Jan.Feb.

Dec.

                         Sept.Oct. Feb. Oct.No Jan.Feb.

.

                                       .

.Apr

.gAu

                                   Sept. Dec.

.gAu

                                     Sept.

.Mar

.MarApr

.Apr

                                       .

.MarApr

                                       .

.yMa

.No

                                       .

en

en

enJu

                                     yMaJu

yl

Ju

lyJu

                                     yMaJu

                                     yMaJu

yl2000 2001 2002 2003 2004

Whole milk 2 percent milk 1 percent milk Skim milk

Source: GAO analysis of data provided by Information Resources, Inc.

Note: The retail price is the price collected by Information Resources,
Inc., for the Washington, D.C., market.

Appendix IV

Average Monthly and Annual Farm, Cooperative, Wholesale and Retail Milk
Prices in Selected Markets

This appendix updates information provided in our June 2001 report on
average monthly and annual farm and cooperative prices of raw milk and on
the average monthly and annual wholesale and retail prices for a gallon of
whole, 2 percent, 1 percent, and skim milk. Tables 11 through 25 provide
these data for 15 selected markets over the period October 2000 through
May 2004.

Table 11: Prices for a Gallon of Milk in Atlanta, Georgia, October 2000 through
                           May 2004 Wholesale Retail

  Year Month  Farm Cooperative Whole  2%      1% Skim  Whole    2%    1%  Skim 
  2000 Oct.  $1.20       $1.41 $2.16 $2.04 $2.01 $1.99 $2.64 $2.77 $2.87 $2.75 
       Nov.   1.20        1.40  2.16 2.04  2.01  1.99   2.48 2.53  2.59   2.51 
       Dec.   1.21        1.41  2.16 2.04  2.01  1.99   2.63 2.71  2.76   2.69 
       Avg.   1.20        1.41  2.16 2.04  2.01  1.99   2.58 2.67  2.74   2.65 
  2001 Jan.   1.37        1.57  2.21 2.42    N/A 2.04   2.66 2.77  2.86   2.74 
       Feb.   1.21        1.41  2.17 2.45    N/A 2.07   2.23 2.33  2.37   2.32 
       Mar.   1.25        1.45  2.17 2.45    N/A 2.07   2.65 2.73  2.81   2.71 
       Apr.   1.32        1.52  2.17 2.45    N/A 2.07   2.53 2.69  2.71   2.65 
        May   1.37        1.58  2.17 2.45  2.05  2.07   2.58 2.75  2.76   2.69 
       June   1.41        1.61  2.20 2.47  2.05  2.10   2.50 2.62  2.64   2.57 
       July   1.45        1.66  2.22 2.52  2.05  2.13   2.55 2.67  2.71   2.64 
       Aug.   1.46        1.66  2.22 2.54  2.05  2.13   2.55 2.68  2.68   2.63 
       Sept.  1.47        1.68  2.22 2.56  2.05  2.13   2.54 2.70  2.70   2.65 
       Oct.   1.50        1.71  2.22 2.56  2.05  2.13   2.64 2.77  2.81   2.71 
       Nov.   1.49        1.69  2.22 2.56  2.08  2.13   2.60 2.73  2.74   2.68 
       Dec.   1.20        1.41  2.22 2.56  2.08  2.13   2.57 2.68  2.72   2.66 
       Avg.   1.38        1.58  2.20 2.50  2.06  2.10   2.55 2.68  2.71   2.64 
  2002 Jan.   1.22        1.40  2.22 2.25  2.08  2.13   2.54 2.64  2.68   2.62 
       Feb.   1.22        1.40  2.22 2.18  2.08  2.13   2.48 2.62  2.63   2.60 
       Mar.   1.22        1.40  2.13 2.06  1.99  1.99   2.51 2.67  2.72   2.64 
       Apr.   1.22        1.40  2.04 1.95  1.89  1.84   2.38 2.55  2.55   2.51 
        May   1.22        1.40  2.04 1.95  1.89  1.84   2.32 2.51  2.54   2.49 
       June   1.20        1.38  2.04 1.95  1.89  1.84   2.30 2.43  2.46   2.42 
       July   1.17        1.35  2.04 1.95  1.89  1.84   2.12 2.26  2.22   2.23 
       Aug.   1.16        1.34  2.04 1.95  1.89  1.84   2.29 2.51  2.50   2.47 
       Sept.  1.15        1.34  2.04 1.95  1.89  1.84   2.21 2.40  2.35   2.37 
       Oct.   1.15        1.32  2.04 1.95  1.89  1.84   2.08 2.25  2.21   2.22 

 Appendix IV Average Monthly and Annual Farm, Cooperative, Wholesale and Retail
                        Milk Prices in Selected Markets

                         (Continued From Previous Page)

                                Wholesale Retail

Year  Month   Farm Cooperative  Whole   2%    1% Skim Whole  2%   1%  Skim 
          Nov.   1.17         1.35  2.06  1.97 1.92 1.86 2.13  2.30 2.27 2.26 
          Dec.   1.16         1.34  2.08  1.99 1.94 1.88 2.32  2.49 2.51 2.48 
          Avg.   1.19         1.37  2.08  2.01 1.94 1.91 2.31  2.47 2.47 2.44 
2003   Jan.   1.16         1.34  2.08  1.99 1.94 1.88 2.21  2.38 2.38 2.35 
          Feb.   1.16         1.34  2.08  1.99 1.94 1.88 2.31  2.47 2.50 2.45 
          Mar.   1.13         1.31  2.08  1.99 1.94 1.88 2.28  2.47 2.45 2.43 
          Apr.   1.11         1.29  2.08  1.99 1.94 1.88 2.23  2.41 2.43 2.38 
          May    1.12         1.30  2.08  1.99 1.94 1.88 2.23  2.40 2.38 2.37 
          June   1.12         1.30  2.08  1.99 1.94 1.88 2.28  2.47 2.47 2.44 
          July   1.12         1.30  2.08  1.99 1.94 1.88 2.23  2.45 2.44 2.40 
          Aug.   1.19         1.38  2.08  1.99 1.94 1.88 2.28  2.44 2.42 2.40 
         Sept.   1.39         1.57  2.08  1.99 1.94 1.88 2.44  2.67 2.68 2.62 
          Oct.   1.41         1.60  2.21  2.11 2.06 2.02 2.44  2.62 2.66 2.60 
          Nov.   1.41         1.60  2.21  2.12 2.07 2.02 2.46  2.66 2.67 2.62 
          Dec.   1.40         1.58  2.21  2.12 2.07 2.02 2.27  2.49 2.51 2.46 

           Avg.   1.23  1.41  2.11  2.02  1.97  1.92  2.31  2.49  2.50   2.46 
    2004   Jan.   1.25  1.43  2.21  2.12  2.07  2.02  2.24  2.42  2.38   2.36 
           Feb.   1.25  1.43  2.21  2.12  2.07  2.02  2.25  2.39  2.37   2.34 
           Mar.   1.26  1.43  2.21  2.12  2.07  2.02  2.28  2.45  2.43   2.40 
           Apr.   1.40  1.57  2.22  2.12  2.07  2.03  2.29  2.49  2.38   2.35 
           May    1.87  2.05  2.64  2.55  2.50  2.45  2.69  2.93  2.85   2.78 
           Avg.   1.41  1.58  2.30  2.21  2.16  2.11  2.35  2.54  2.48   2.45 

Source: GAO's analysis of price data provided by USDA, the Defense
Commissary Agency, and Information Resources, Inc.

Note: For the Atlanta market, the farm price is the USDA-estimated farm
Class I price for 3.5 percent milkfat content; the cooperative price is
the announced cooperative Class I price for Atlanta for 3.5 percent
milkfat content; the wholesale price is the average of the prices paid by
the commissaries at Fort Gillem and Fort McPherson, adjusted for a 5
percent markup; and the retail price is the price collected by Information
Resources, Inc., for the Atlanta market. Prices may not average due to
rounding. "N/A" indicates data not available.

Appendix IV Average Monthly and Annual Farm, Cooperative, Wholesale and
Retail Milk Prices in Selected Markets

  Table 12: Prices for a Gallon of Milk in Boston, Massachusetts, October 2000
                       through May 2004 Wholesale Retail

  Year Month  Farm Cooperative Whole  2%      1% Skim  Whole    2%    1%  Skim 
  2000 Oct.  $1.41       $1.50 $2.00 $1.96 $1.91 $1.90 $2.19 $2.21 $2.23 $2.02 
       Nov.   1.41        1.50  2.00 1.96  1.91  1.90   2.20 2.22  2.24   2.03 
       Dec.   1.41        1.50  2.00 1.96  1.91  1.90   2.23 2.26  2.27   2.05 
       Avg.   1.41        1.50  2.00 1.96  1.91  1.90   2.21 2.23  2.25   2.03 
  2001 Jan.   1.43        1.53  2.00 1.96  1.91  1.90   2.23 2.25  2.25   2.05 
       Feb.   1.41        1.50  2.00 1.96  1.91  1.90   2.25 2.29  2.29   2.08 
       Mar.   1.41        1.50  2.00 1.96  1.91  1.90   2.26 2.31  2.32   2.12 
       Apr.   1.41        1.50  2.00 1.96  1.91  1.90   2.26 2.30  2.31   2.12 
        May   1.45        1.55  2.00 1.96  1.91  1.90   2.26 2.31  2.32   2.13 
       June   1.55        1.65  2.00 1.96  1.91  1.90   2.26 2.31  2.32   2.14 
       July   1.59        1.68  2.00 1.96  1.91  1.90   2.27 2.32  2.33   2.14 
       Aug.   1.65        1.75  2.00 1.96  1.91  1.90   2.27 2.32  2.33   2.15 
       Sept.  1.66        1.76  2.00 1.96  1.91  1.90   2.28 2.34  2.36   2.18 
       Oct.   1.69        1.79  2.00 1.96  1.91  1.90   2.30 2.35  2.37   2.19 
       Nov.   1.68        1.78  2.00 1.96  1.93  1.90   2.32 2.38  2.40   2.21 
       Dec.   1.36        1.45  2.07 1.85  1.71  1.71   2.27 2.36  2.38   2.20 
       Avg.   1.52        1.62  2.01 1.95  1.90  1.88   2.27 2.32  2.33   2.14 
  2002 Jan.   1.35        1.45  2.07 1.85  1.71  1.71   2.27 2.33  2.35   2.17 
       Feb.   1.35        1.45  2.07 1.85  1.71  1.71   2.28 2.36  2.38   2.20 
       Mar.   1.32        1.42  2.04 1.82  1.68  1.68   2.28 2.37  2.38   2.21 
       Apr.   1.30        1.41  2.03 1.81  1.67  1.67   2.27 2.34  2.35   2.18 
        May   1.29        1.39  2.01 1.81  1.67  1.67   2.27 2.34  2.36   2.19 
       June   1.27        1.37  2.00 1.78  1.63  1.63   2.26 2.32  2.34   2.17 
       July   1.23        1.33  1.96 1.74  1.60  1.60   2.26 2.31  2.33   2.16 
       Aug.   1.20        1.30  1.95 1.73  1.59  1.59   2.24 2.28  2.31   2.14 
       Sept.  1.20        1.30  1.95 1.73  1.59  1.59   2.26 2.32  2.33   2.17 
       Oct.   1.17        1.28  1.92 1.70  1.56  1.56   2.25 2.32  2.33   2.18 
       Nov.   1.21        1.31  1.96 1.74  1.60  1.60   2.26 2.32  2.33   2.18 
       Dec.   1.20        1.31  1.95 1.73  1.59  1.59   2.24 2.30  2.29   2.15 
       Avg.   1.26        1.36  1.99 1.77  1.63  1.63   2.26 2.33  2.34   2.18 
  2003 Jan.   1.20        1.31  1.95 1.73  1.59  1.59   2.27 2.33  2.31   2.16 
       Feb.   1.17        1.28  1.90 1.70  1.56  1.55   2.28 2.33  2.33   2.19 
       Mar.   1.14        1.25  1.87 1.66  1.52  1.52   2.27 2.33  2.33   2.17 
       Apr.   1.12        1.23  1.86 1.65  1.51  1.51   2.25 2.29  2.28   2.15 

 Appendix IV Average Monthly and Annual Farm, Cooperative, Wholesale and Retail
                        Milk Prices in Selected Markets

                         (Continued From Previous Page)

                                Wholesale Retail

Year  Month   Farm Cooperative  Whole   2%   1%  Skim Whole  2%   1%  Skim 
          May    1.13         1.24  1.87  1.66 1.52 1.52 2.26  2.31 2.31 2.16 
          June   1.13         1.24  1.87  1.66 1.52 1.52 2.26  2.29 2.30 2.15 
          July   1.16         1.27  1.88  1.67 1.53 1.53 2.25  2.29 2.29 2.13 
          Aug.   1.26         1.37  1.98  1.76 1.62 1.62 2.25  2.29 2.30 2.14 
         Sept.   1.50         1.61  2.15  1.94 1.79 1.79 2.31  2.35 2.37 2.21 
          Oct.   1.54         1.66  2.25  2.04 1.89 1.89 2.33  2.38 2.40 2.24 
          Nov.   1.55         1.67  2.26  2.05 1.90 1.90 2.33  2.39 2.40 2.24 
          Dec.   1.51         1.62  2.21  2.00 1.85 1.85 2.34  2.40 2.42 2.26 
          Avg.   1.28         1.40  2.00  1.79 1.65 1.65 2.28  2.33 2.34 2.18 
2004   Jan.   1.34         1.45  2.08  1.87 1.72 1.72 2.35  2.40 2.42 2.26 
          Feb.   1.31         1.43  2.03  1.81 1.67 1.67 2.34  2.37 2.39 2.24 
          Mar.   1.34         1.46  2.03  1.81 1.67 1.67 2.32  2.37 2.38 2.22 
          Apr.   1.49         1.60  2.20  1.98 1.83 1.83 2.32  2.37 2.38 2.26 
          May    2.01         2.12  2.26  2.05 1.90 1.90 2.58  2.63 2.66 2.51 
          Avg.   1.50         1.61  2.12  1.90 1.76 1.76 2.38  2.43 2.45 2.30 

Source: GAO's analysis of price data provided by USDA, the Defense
Commissary Agency, and Information Resources, Inc.

Note: For the Boston market, the farm price is the USDA-estimated farm
Class I price for 3.5 percent milkfat content; the cooperative price is
the effective cooperative Class I price for Boston for 3.5 percent milkfat
content during the period that the Northeast Interstate Dairy Compact was
effective (July 1997 through September 2001); the cooperative price after
September 2001 is the announced cooperative Class I price for Boston for
3.5 percent milkfat content; the wholesale price is the price paid by the
commissary at Hanscom Air Force Base, adjusted for a 5 percent markup; and
the retail price is the price collected by Information Resources, Inc.,
for the Boston market. Prices may not average due to rounding.

Appendix IV Average Monthly and Annual Farm, Cooperative, Wholesale and
Retail Milk Prices in Selected Markets

Table 13: Prices for a Gallon of Milk in Charlotte, North Carolina, October 2000
                       through May 2004 Wholesale Retail

  Year Month  Farm Cooperative Whole  2%      1% Skim  Whole    2%    1%  Skim 
  2000 Oct.  $1.25       $1.41 $2.09 $2.05 $2.03 $2.00 $2.59 $2.59 $2.64 $2.55 
       Nov.   1.24        1.40  2.08 2.06  2.03  2.00   2.56 2.56  2.59   2.52 
       Dec.   1.25        1.41  2.11 2.08  2.03  2.00   2.57 2.59  2.62   2.55 
       Avg.   1.25        1.41  2.09 2.06  2.03  2.00   2.57 2.58  2.62   2.54 
  2001 Jan.   1.40        1.57  2.25 2.18  2.08  2.00   2.61 2.62  2.64   2.58 
       Feb.   1.25        1.41  2.10 2.07  2.03  1.98   2.56 2.57  2.60   2.53 
       Mar.   1.29        1.45  2.14 2.11  2.08  2.03   2.61 2.63  2.65   2.57 
       Apr.   1.36        1.52  2.21 2.19  2.15  2.10   2.58 2.61  2.63   2.56 
        May   1.41        1.58  2.28 2.25  2.21  2.17   2.59 2.62  2.64   2.57 
       June   1.44        1.61  2.35 2.32  2.28  2.23   2.58 2.61  2.63   2.57 
       July   1.49        1.66  2.38 2.35  2.31  2.26   2.59 2.62  2.64   2.57 
       Aug.   1.49        1.66  2.52 2.42  2.36  2.30   2.58 2.61  2.63   2.56 
       Sept.  1.51        1.68  2.57 2.45  2.38  2.32   2.57 2.60  2.62   2.55 
       Oct.   1.54        1.71  2.57 2.46  2.39  2.33   2.58 2.60  2.63   2.55 
       Nov.   1.53        1.69  2.60 2.50  2.46  2.42   2.62 2.64  2.68   2.62 
       Dec.   1.24        1.41  2.60 2.55  2.54  2.52   2.62 2.65  2.68   2.63 
       Avg.   1.41        1.58  2.38 2.32  2.27  2.22   2.59 2.62  2.64   2.57 
  2002 Jan.   1.18        1.40  2.60 2.55  2.54  2.52   2.63 2.65  2.67   2.60 
       Feb.   1.18        1.40  2.60 2.55  2.54  2.52   2.59 2.60  2.65   2.58 
       Mar.   1.18        1.40  2.60 2.55  2.54  2.52   2.59 2.61  2.65   2.59 
       Apr.   1.18        1.40  2.60 2.55  2.54  2.52   2.58 2.60  2.63   2.56 
        May   1.19        1.40  2.60 2.55  2.54  2.52   2.57 2.59  2.63   2.57 
       June   1.17        1.38  2.60 2.55  2.54  2.52   2.57 2.59  2.60   2.56 
       July   1.14        1.35  2.60 2.55  2.54  2.52   2.54 2.56  2.61   2.55 
       Aug.   1.12        1.34  2.60 2.55  2.54  2.52   2.56 2.58  2.62   2.56 
       Sept.  1.12        1.34  2.60 2.55  2.54  2.52   2.51 2.55  2.59   2.53 
       Oct.   1.10        1.31  2.60 2.55  2.54  2.52   2.51 2.53  2.59   2.52 
       Nov.   1.14        1.35  2.63 2.57  2.56  2.54   2.59 2.60  2.65   2.58 
       Dec.   1.12        1.34  2.65 2.58  2.57  2.56   2.58 2.58  2.63   2.56 
       Avg.   1.15        1.37  2.61 2.55  2.54  2.53   2.57 2.59  2.63   2.56 
  2003 Jan.   1.13        1.34  2.65 2.58  2.57  2.56   2.52 2.52  2.58   2.50 
       Feb.   1.13        1.34  2.64 2.57  2.56  2.54   2.59 2.58  2.64   2.56 
       Mar.   1.10        1.31  2.62 2.56  2.55  2.52   2.59 2.59  2.64   2.56 
       Apr.   1.08        1.29  2.62 2.56  2.55  2.52   2.58 2.57  2.62   2.54 

 Appendix IV Average Monthly and Annual Farm, Cooperative, Wholesale and Retail
                        Milk Prices in Selected Markets

                         (Continued From Previous Page)

                                Wholesale Retail

Year  Month   Farm Cooperative  Whole   2%   1%  Skim Whole  2%   1%  Skim 
          May    1.09         1.30  2.62  2.56 2.55 2.52 2.56  2.56 2.61 2.53 
          June   1.09         1.30  2.62  2.56 2.55 2.52 2.59  2.58 2.62 2.55 
          July   1.09         1.30  2.62  2.56 2.55 2.52 2.56  2.56 2.63 2.53 
          Aug.   1.16         1.38  2.62  2.56 2.55 2.52 2.59  2.61 2.68 2.58 
         Sept.   1.35         1.57  2.77  2.70 2.69 2.66 2.62  2.64 2.71 2.65 
          Oct.   1.38         1.60  2.95  2.85 2.87 2.85 2.75  2.76 2.83 2.73 
          Nov.   1.38         1.60  2.98  2.85 2.92 2.91 2.78  2.80 2.88 2.80 
          Dec.   1.37         1.58  2.98  2.85 2.92 2.91 2.80  2.82 2.90 2.81 
          Avg.   1.20         1.41  2.72  2.65 2.65 2.63 2.63  2.63 2.70 2.61 
2004   Jan.   1.21         1.43  2.98  2.90 2.92 2.90 2.78  2.79 2.87 2.78 
          Feb.   1.21         1.43  2.88  2.82 2.81 2.80 2.76  2.77 2.84 2.77 
          Mar.   1.21         1.43  2.78  2.75 2.69 2.66 2.74  2.75 2.83 2.73 
          Apr.   1.35         1.57  2.70  2.69 2.57 2.52 2.83  2.83 2.92 2.81 
          May    1.83         2.05  3.10  3.05 2.85 2.84 3.19  3.21 3.24 3.21 
          Avg.   1.36         1.58  2.89  2.84 2.77 2.74 2.86  2.87 2.94 2.86 

Source: GAO's analysis of price data provided by USDA, the Defense
Commissary Agency, and Information Resources, Inc.

Note: For the Charlotte market, the farm price is the USDA-estimated farm
Class I price for 3.5 percent milkfat content; the cooperative price is
the announced cooperative Class I price for Charlotte for 3.5 percent
milkfat content; the wholesale price is the price paid by the Fort Bragg
North Post Store, adjusted for a 5 percent markup; and the retail price is
the price collected by Information Resources, Inc., for the Charlotte
market. Prices may not average due to rounding.

Appendix IV Average Monthly and Annual Farm, Cooperative, Wholesale and
Retail Milk Prices in Selected Markets

Table 14: Prices for a Gallon of Milk in Cincinnati, Ohio, October 2000 through
                           May 2004 Wholesale Retail

  Year Month  Farm Cooperative Whole  2%     1% Skim  Whole    2%    1%  Skim 
  2000 Oct.  $1.20       $1.38  N/A  N/A  $1.96 $1.80 $1.88 $2.28 $2.32 $2.25 
       Nov.   1.19        1.37  N/A  N/A  1.96  1.80   1.94 2.18  2.25   2.19 
       Dec.   1.22        1.40  N/A  N/A  1.96  1.80   1.93 2.27  2.31   2.25 
       Avg.   1.20        1.38  N/A  N/A  1.96  1.80   1.92 2.24  2.29   2.23 
  2001 Jan.   1.38        1.55  N/A  N/A  1.96  1.80   2.09 2.37  2.46   2.37 
       Feb.   1.20        1.37  N/A  N/A  1.96  1.80   1.86 2.20  2.27   2.22 
       Mar.   1.26        1.43 2.29  2.12 1.96  1.80   1.91 2.06  2.08   2.04 
       Apr.   1.33        1.50 2.34  2.15 1.99  1.81   1.88 2.21  2.28   2.21 
        May   1.39        1.56 2.43  2.22 2.03  1.84   1.80 2.07  1.93   1.88 
       June   1.46        1.63 2.49  2.26 2.05  1.84   1.98 2.27  2.31   2.26 
       July   1.50        1.67 2.53  2.29 2.07  1.85   1.98 2.28  2.32   2.25 
       Aug.   1.51        1.67 2.53  2.29 2.07  1.85   1.94 2.38  2.47   2.41 
       Sept.  1.52        1.69 2.54  2.30 2.08  1.86   2.00 2.42  2.46   2.40 
       Oct.   1.55        1.72 2.56  2.30 2.06  1.82   1.95 2.23  2.29   2.15 
       Nov.   1.52        1.70 2.39  2.36 2.15  1.96   2.02 2.56  2.67   2.55 
       Dec.   1.21        1.37 2.28  2.09 2.19  2.00   1.87 2.29  2.36   2.27 
       Avg.   1.40        1.57 2.44  2.24 2.05  1.85   1.94 2.28  2.33   2.25 
  2002 Jan.   1.19        1.36 2.26  2.06 2.19  2.00   1.96 2.38  2.44   2.34 
       Feb.   1.19        1.36 2.26  2.06 2.19  2.00   1.91 2.37  2.46   2.38 
       Mar.   1.17        1.33 2.23  2.03 2.19  2.00   1.93 2.36  2.50   2.36 
       Apr.   1.15        1.32 2.23  2.03  N/A   N/A   1.95 2.28  2.29   2.22 
        May   1.13        1.30 2.20  2.00  N/A   N/A   1.88 2.44  2.48   2.38 
       June   1.11        1.28 2.19  1.99  N/A   N/A   1.81 2.21  2.27   2.18 
       July   1.08        1.24 2.16  1.96  N/A   N/A   1.81 2.30  2.33   2.24 
       Aug.   1.07        1.23 2.14  1.94  N/A   N/A   1.82 2.33  2.36   2.27 
       Sept.  1.07        1.23 2.14  1.94  N/A   N/A   1.81 2.34  2.38   2.28 
       Oct.   1.03        1.20 2.11  1.91  N/A   N/A   1.79 2.28  2.32   2.23 
       Nov.   1.07        1.24 2.15  1.95  N/A   N/A   1.83 2.43  2.48   2.37 
       Dec.   1.07        1.23 2.14  1.94 1.75  1.56   1.85 2.37  2.40   2.32 
       Avg.   1.11        1.28 2.18  1.98 2.08  1.89   1.86 2.34  2.39   2.30 
  2003 Jan.   1.07        1.24 2.14  1.94 1.75  1.56   1.72 2.35  2.39   2.30 
       Feb.   1.06        1.24 2.11  1.90 1.75  1.56   1.82 2.29  2.33   2.24 
       Mar.   1.02        1.20 2.07  1.86 1.71  1.52   1.84 2.30  2.34   2.24 
       Apr.   1.01        1.19 2.06  1.85 1.71  1.52   1.74 2.21  2.24   2.17 

 Appendix IV Average Monthly and Annual Farm, Cooperative, Wholesale and Retail
                        Milk Prices in Selected Markets

                         (Continued From Previous Page)

                                Wholesale Retail

Year  Month   Farm Cooperative  Whole   2%   1%  Skim Whole  2%   1%  Skim 
          May    1.02         1.19  2.07  1.86 1.68 1.49 1.84  2.38 2.42 2.32 
          June   1.02         1.20  2.07  1.86 1.68 1.49 1.80  2.47 2.54 2.42 
          July   1.02         1.20  2.08  1.87 1.69 1.50 1.82  2.30 2.33 2.25 
          Aug.   1.09         1.27  2.17  1.97 1.78 1.59 1.86  2.43 2.48 2.37 
         Sept.   1.33         1.51  2.39  2.19 2.01 1.81 1.83  2.38 2.42 2.32 
          Oct.   1.38         1.56  2.44  2.24 2.06 1.86 1.86  2.36 2.42 2.31 
          Nov.   1.38         1.56  2.45  2.25 2.07 1.87 1.88  2.40 2.46 2.34 
          Dec.   1.34         1.52  2.39  2.20 2.06 1.86 1.87  2.15 2.18 2.08 
          Avg.   1.15         1.32  2.20  2.00 1.83 1.64 1.82  2.34 2.38 2.28 
2004   Jan.   1.18         1.36  2.27  2.07 1.89 1.70 1.71  2.08 2.07 2.00 
          Feb.   1.15         1.33  2.22  2.02 1.83 1.64 1.63  1.89 1.88 1.82 
          Mar.   1.18         1.36  2.25  2.05 1.86 1.67 1.63  1.87 1.85 1.80 
          Apr.   1.33         1.51  2.41  2.21 2.03 1.83 1.46  1.70 1.60 1.56 
          May    1.85         2.03  2.88  2.68 2.50 2.31 1.87  2.19 2.14 2.05 
          Avg.   1.34         1.52  2.41  2.21 2.02 1.83 1.66  1.95 1.91 1.85 

Source: GAO's analysis of price data provided by USDA, the Defense
Commissary Agency, and Information Resources, Inc.

Note: For the Cincinnati market, the farm price is the USDA-estimated farm
Class I price for 3.5 percent milkfat content; the cooperative price is
the announced cooperative Class I price for Cincinnati for 3.5 percent
milkfat content; the wholesale price is the price paid by the commissary
at Wright-Patterson Air Force Base, adjusted for a 5 percent markup; and
the retail price is the price collected by Information Resources, Inc.,
for the Cincinnati market. Prices may not average due to rounding. "N/A"
indicates data not available.

Appendix IV Average Monthly and Annual Farm, Cooperative, Wholesale and
Retail Milk Prices in Selected Markets

Table 15: Prices for a Gallon of Milk in Dallas, Texas, October 2000 through May
                             2004 Wholesale Retail

  Year Month  Farm Cooperative Whole  2%      1% Skim  Whole    2%    1%  Skim 
  2000 Oct.  $1.24       $1.35 $2.06 $2.06 $2.02 $1.95 $1.79 $1.64 $1.75 $1.67 
       Nov.   1.23        1.35  2.06 2.06  2.02  1.95   1.87 1.77  2.01   1.80 
       Dec.   1.24        1.35  2.06 2.03  2.02  1.95   2.00 1.82  2.02   1.81 
       Avg.   1.24        1.35  2.06 2.05  2.02  1.95   1.89 1.74  1.93   1.76 
  2001 Jan.   1.35        1.46  2.06 2.03  2.02  1.95   2.01 1.89  2.12   1.87 
       Feb.   1.25        1.36  2.15 2.12  2.11  2.04   2.02 1.91  2.07   1.91 
       Mar.   1.29        1.40  2.16 2.13  2.12  2.04   1.96 1.85  2.03   1.84 
       Apr.   1.35        1.46  2.16 2.13  2.12  2.04   2.00 1.92  2.06   1.92 
        May   1.40        1.51  2.16 2.13  2.12  2.04   2.05 1.94  2.07   1.94 
       June   1.43        1.55  2.16 2.13  2.12  2.04   2.07 2.00  2.05   1.98 
       July   1.50        1.61  2.16 2.13  2.12  2.04   2.08 2.00  2.09   1.96 
       Aug.   1.50        1.62  2.21 2.18  2.15  2.10   2.08 2.06  2.18   2.05 
       Sept.  1.50        1.61  2.21 2.19  2.18  2.10   2.06 2.02  2.09   1.99 
       Oct.   1.54        1.64  2.21 2.19  2.18  2.10   2.10 2.04  2.12   2.05 
       Nov.   1.53        1.63  2.21 2.19  2.18  2.10   2.08 2.05  2.19   2.06 
       Dec.   1.25        1.36  2.21 2.19  2.18  2.10   2.16 2.10  2.14   2.10 
       Avg.   1.41        1.52  2.17 2.15  2.13  2.06   2.06 1.98  2.10   1.97 
  2002 Jan.   1.25        1.36  2.21 2.19  2.18  2.10   2.03 2.03  2.20   2.01 
       Feb.   1.26        1.36  2.21 2.19  2.18  2.10   2.09 2.03  2.15   2.02 
       Mar.   1.27        1.37  2.21 2.19  2.18  2.10   2.02 2.00  2.11   1.97 
       Apr.   1.27        1.37  2.21 2.19  2.18  2.10   2.05 2.01  2.14   2.00 
        May   1.27        1.37  2.21 2.19  2.18  2.10   2.08 2.02  2.17   2.00 
       June   1.27        1.37  2.21 2.19  2.18  2.10   2.10 2.06  2.21   2.07 
       July   1.20        1.29  2.21 2.19  2.18  2.10   2.07 2.06  2.23   2.06 
       Aug.   1.18        1.28  2.21 2.19  2.18  2.10   2.08 2.06  2.23   2.06 
       Sept.  1.18        1.28  2.21 2.19  2.18  2.10   2.07 2.05  2.18   2.05 
       Oct.   1.15        1.25  2.21 2.19  2.18  2.10   2.06 2.06  2.18   2.06 
       Nov.   1.20        1.28  2.21 2.19  2.18  2.10   2.05 1.99  2.08   1.99 
       Dec.   1.18        1.28  2.21 2.19  2.18  2.10   2.02 1.97  2.05   1.97 
       Avg.   1.22        1.32  2.21 2.19  2.18  2.10   2.06 2.03  2.16   2.02 
  2003 Jan.   1.19        1.29  2.21 2.19  2.18  2.10   2.02 1.98  2.07   1.96 
       Feb.   1.16        1.26  2.21 2.19  2.18  2.10   2.01 1.95  2.03   1.93 
       Mar.   1.12        1.22  2.21 2.19  2.18  2.10   2.05 1.97  2.04   1.96 
       Apr.   1.11        1.21  2.21 2.19  2.18  2.10   2.07 2.00  2.07   2.00 

 Appendix IV Average Monthly and Annual Farm, Cooperative, Wholesale and Retail
                        Milk Prices in Selected Markets

                         (Continued From Previous Page)

                                Wholesale Retail

Year  Month   Farm Cooperative  Whole   2%   1%  Skim Whole  2%   1%  Skim 
          May    1.11         1.22  2.21  2.19 2.18 2.10 2.06  1.96 2.05 1.96 
          June   1.11         1.22  2.21  2.19 2.18 2.10 2.07  2.01 2.08 1.99 
          July   1.11         1.22  2.21  2.19 2.18 2.10 2.06  1.98 2.07 1.96 
          Aug.   1.19         1.30  2.21  2.19 2.18 2.10 2.14  2.01 2.12 1.99 
         Sept.   1.41         1.52  2.39  2.37 2.26 2.27 2.18  2.05 2.14 2.02 
          Oct.   1.43         1.54  2.39  2.37 2.43 2.27 2.27  2.17 2.27 2.16 
          Nov.   1.43         1.54  2.39  2.37 2.45 2.27 2.24  2.11 2.21 2.10 
          Dec.   1.42         1.53  2.49  2.46 2.44 2.37 2.19  2.05 2.06 2.02 
          Avg.   1.23         1.34  2.28  2.26 2.25 2.17 2.11  2.02 2.10 2.00 
2004   Jan.   1.27         1.38  2.41  2.38 2.37 2.28 2.16  2.01 2.01 1.95 
          Feb.   1.27         1.38  2.39  2.37 2.35 2.26 2.15  1.96 1.98 1.92 
          Mar.   1.28         1.38  2.39  2.37 2.35 2.26 2.14  1.97 1.96 1.94 
          Apr.   1.42         1.52  2.43  2.40 2.38 2.29 2.24  2.03 2.03 1.95 
          May    1.89         2.00  2.87  2.84 2.82 2.74 2.79  2.67 2.77 2.52 
          Avg.   1.43         1.53  2.50  2.47 2.45 2.37 2.30  2.13 2.15 2.06 

Source: GAO's analysis of price data provided by USDA, the Defense
Commissary Agency, and Information Resources, Inc.

Note: For the Dallas market, the farm price is the USDA-estimated farm
Class I price for 3.5 percent milkfat content; the cooperative price is
the announced cooperative Class I price for Dallas for 3.5 percent milkfat
content; the wholesale price is the average of the prices paid by the two
commissaries at Fort Hood, adjusted for a 5 percent markup; and the retail
price is the price collected by Information Resources, Inc., for the
Dallas market. Prices may not average due to rounding.

Appendix IV Average Monthly and Annual Farm, Cooperative, Wholesale and
Retail Milk Prices in Selected Markets

Table 16: Prices for a Gallon of Milk in Denver, Colorado, October 2000 through
                           May 2004 Wholesale Retail

  Year Month  Farm Cooperative Whole    2%    1% Skim  Whole    2%    1%  Skim 
  2000 Oct.  $1.18       $1.31 $1.94 $1.71 $1.62 $1.52 $2.55 $2.43 $2.45 $2.23 
       Nov.   1.19        1.30  1.94 1.72  1.63  1.53   2.55 2.44  2.46   2.25 
       Dec.   1.21        1.33  1.97 1.75  1.65  1.56   2.58 2.49  2.45   2.27 
       Avg.   1.19        1.31  1.95 1.73  1.63  1.54   2.56 2.45  2.45   2.25 
  2001 Jan.   1.38        1.49  2.10 1.88  1.79  1.70   2.62 2.50  2.56   2.28 
       Feb.   1.21        1.31  1.97 1.75  1.54  1.57   2.57 2.48  2.54   2.27 
       Mar.   1.27        1.37  1.99 1.77  1.55  1.59   2.49 2.45  2.48   2.15 
       Apr.   1.33        1.44  2.07 1.85  1.63  1.67   2.59 2.50  2.50   2.30 
        May   1.40        1.51  2.14 1.93  1.70  1.74   2.59 2.51  2.61   2.28 
       June   1.46        1.57  2.20 2.00  1.76  1.81   2.65 2.56  2.58   2.34 
       July   1.49        1.60  2.23 2.02  1.79  1.83   2.65 2.55  2.48   2.31 
       Aug.   1.49        1.61  2.24 2.03  1.80  1.84   2.65 2.55  2.52   2.31 
       Sept.  1.51        1.62  2.25 2.04  1.81  1.85   2.60 2.53  2.49   2.31 
       Oct.   1.55        1.65  2.28 2.07  1.82  1.88   2.64 2.60  2.62   2.35 
       Nov.   1.53        1.64  2.27 2.06  1.83  1.87   2.65 2.63  2.72   2.39 
       Dec.   1.22        1.33  1.96 1.74  1.64  1.56   2.64 2.65  2.64   2.40 
       Avg.   1.40        1.51  2.14 1.93  1.72  1.74   2.61 2.54  2.56   2.31 
  2002 Jan.   1.21        1.32  1.96 1.74  1.64  1.56   2.67 2.65  2.73   2.39 
       Feb.   1.21        1.32  1.95 1.74  1.64  1.56   2.66 2.64  2.72   2.39 
       Mar.   1.18        1.29  1.92 1.71  1.62  1.53   2.64 2.63  2.72   2.34 
       Apr.   1.17        1.28  1.91 1.70  1.61  1.52   2.69 2.71  2.73   2.39 
        May   1.15        1.26  1.89 1.68  1.59  1.50   2.60 2.61  2.71   2.40 
       June   1.13        1.24  1.87 1.66  1.57  1.48   2.59 2.54  2.71   2.34 
       July   1.09        1.21  1.83 1.62  1.53  1.44   2.50 2.43  2.64   2.30 
       Aug.   1.08        1.20  1.82 1.62  1.52  1.43   2.50 2.43  2.63   2.33 
       Sept.  1.08        1.19  1.82 1.62  1.52  1.43   2.52 2.51  2.70   2.30 
       Oct.   1.05        1.17  1.80 1.59  1.49  1.41   2.55 2.52  2.70   2.34 
       Nov.   1.09        1.21  1.83 1.62  1.53  1.44   2.43 2.50  2.68   2.32 
       Dec.   1.09        1.20  1.82 1.62  1.52  1.43   2.49 2.49  2.64   2.30 
       Avg.   1.13        1.24  1.87 1.66  1.57  1.48   2.57 2.56  2.69   2.35 
  2003 Jan.   1.09        1.20  1.82 1.61  1.52  1.43   2.57 2.52  2.71   2.32 
       Feb.   1.06        1.17  1.79 1.58  1.49  1.40   2.61 2.56  2.70   2.35 
       Mar.   1.03        1.14  1.77 1.55  1.46  1.38   2.56 2.50  2.65   2.34 
       Apr.   1.01        1.12  1.75 1.54  1.44  1.36   2.51 2.44  2.61   2.31 

 Appendix IV Average Monthly and Annual Farm, Cooperative, Wholesale and Retail
                        Milk Prices in Selected Markets

                         (Continued From Previous Page)

                                Wholesale Retail

Year  Month  Farm  Cooperative  Whole   2%   1%  Skim Whole  2%   1%  Skim 
          May    1.02         1.13  1.76  1.55 1.45 1.37 2.55  2.47 2.66 2.34 
          June   1.02         1.13  1.76  1.55 1.45 1.37 2.54  2.45 2.64 2.31 
          July   1.02         1.14  1.77  1.56 1.46 1.38 2.56  2.47 2.68 2.35 
          Aug.   1.12         1.24  1.85  1.64 1.55 1.46 2.58  2.50 2.71 2.35 
         Sept.   1.36         1.47  2.03  1.81 1.72 1.63 2.80  2.69 2.97 2.53 
          Oct.   1.41         1.52  2.13  1.92 1.82 1.73 2.87  2.73 2.99 2.55 
          Nov.   1.42         1.53  2.15  1.94 1.84 1.75 2.81  2.70 2.93 2.45 
          Dec.   1.37         1.49  2.14  1.96 1.82 1.72 2.73  2.62 2.72 2.36 
          Avg.   1.16         1.27  1.89  1.68 1.59 1.50 2.64  2.55 2.75 2.38 
2004   Jan.   1.17         1.31  2.10  1.91 1.78 1.68 2.72  2.65 2.70 2.34 
          Feb.   1.15         1.29  2.07  1.89 1.75 1.65 2.69  2.55 2.58 2.20 
          Mar.   1.18         1.32  2.07  1.89 1.75 1.64 2.80  2.69 2.56 2.42 
          Apr.   1.33         1.47  2.21  2.04 1.89 1.78 2.88  2.79 2.59 2.49 
          May    1.85         1.99  2.66  2.48 2.34 2.23 3.00  2.90 2.87 2.64 
          Avg.   1.34         1.48  2.22  2.04 1.90 1.80 2.82  2.72 2.66 2.42 

Source: GAO's analysis of price data provided by USDA, the Defense
Commissary Agency, and Information Resources, Inc.

Note: For the Denver market, the farm price is the USDA-estimated farm
Class I price for 3.5 percent milkfat content; the cooperative price is
the announced cooperative Class I price for Denver for 3.5 percent milkfat
content; the wholesale price is the average of the prices paid by the
commissaries at Fort Carson, Peterson Air Force Base, Buckley Air Force
Base, and the U.S. Air Force Academy, adjusted for a 5 percent markup; and
the retail price is the price collected by Information Resources, Inc.,
for the Denver market. Prices may not average due to rounding.

Appendix IV Average Monthly and Annual Farm, Cooperative, Wholesale and
Retail Milk Prices in Selected Markets

 Table 17: Prices for a Gallon of Milk in Miami, Florida, October 2000 through
                           May 2004 Wholesale Retail

  Year Month Farm  Cooperative Whole  2%    1%  Skim  Whole    2%    1%  Skim 
  2000 Oct.  $1.40       $1.62 $2.46 $2.46 N/A  $2.46 $2.47 $2.36 $2.38 $2.25 
       Nov.   1.39        1.62  2.47 2.47  N/A  2.47   2.47 2.31  2.34   2.25 
       Dec.   1.40        1.62  2.56 2.56  N/A  2.56   2.47 2.33  2.35   2.24 
       Avg.   1.40        1.62  2.50 2.50  N/A  2.50   2.47 2.33  2.36   2.25 
  2001 Jan.   1.56        1.78  2.56 2.56  N/A  2.56   2.48 2.34  2.35   2.24 
       Feb.   1.40        1.63  2.56 2.56  N/A  2.56   2.47 2.32  2.32   2.25 
       Mar.   1.43        1.66  2.56 2.56  N/A  2.56   2.49 2.34  2.35   2.26 
       Apr.   1.51        1.74  2.56 2.56  N/A  2.56   2.49 2.37  2.36   2.27 
        May   1.57        1.79  2.56 2.56  N/A  2.56   2.48 2.37  2.39   2.28 
       June   1.60        1.83  2.56 2.56  N/A  2.56   2.57 2.45  2.47   2.34 
       July   1.65        1.87  2.56 2.56  N/A  2.56   2.56 2.44  2.45   2.32 
       Aug.   1.65        1.88  2.56 2.56  N/A  2.56   2.58 2.45  2.47   2.34 
       Sept.  1.67        1.89  2.56 2.56  N/A  2.56   2.58 2.45  2.47   2.35 
       Oct.   1.70        1.92  2.56 2.56  N/A  2.56   2.56 2.44  2.45   2.33 
       Nov.   1.68        1.91  2.56 2.56  N/A  2.56   2.56 2.44  2.45   2.34 
       Dec.   1.40        1.62  2.56 2.56  N/A  2.56   2.56 2.43  2.44   2.33 
       Avg.   1.57        1.79  2.56 2.56  N/A  2.56   2.53 2.40  2.41   2.30 
  2002 Jan.   1.38        1.62  2.56 2.56  N/A  2.56   2.53 2.40  2.40   2.31 
       Feb.   1.38        1.62  2.56 2.56  N/A  2.56   2.49 2.38  2.38   2.30 
       Mar.   1.38        1.62  2.56 2.56  N/A  2.56   2.49 2.38  2.39   2.30 
       Apr.   1.39        1.62  2.56 2.56  N/A  2.56   2.52 2.39  2.40   2.30 
        May   1.39        1.62  2.56 2.56  N/A  2.56   2.47 2.36  2.36   2.27 
       June   1.37        1.60  2.56 2.56  N/A  2.56   2.48 2.36  2.39   2.28 
       July   1.33        1.57  2.56 2.56  N/A  2.56   2.46 2.37  2.35   2.27 
       Aug.   1.32        1.55  2.56 2.56  N/A  2.56   2.34 2.33  2.31   2.24 
       Sept.  1.32        1.55  2.56 2.56  N/A  2.56   2.37 2.29  2.27   2.22 
       Oct.   1.29        1.52  2.56 2.56  1.86 2.56   2.38 2.28  2.26   2.19 
       Nov.   1.32        1.56  2.56 2.56  1.88 2.56   2.45 2.35  2.36   2.28 
       Dec.   1.31        1.55  2.56 2.56  1.88 2.56   2.49 2.38  2.39   2.30 
       Avg.   1.35        1.58  2.56 2.56  1.87 2.56   2.46 2.36  2.36   2.27 
  2003 Jan.   1.31        1.55  2.56 2.56  1.90 2.56   2.50 2.38  2.39   2.30 
       Feb.   1.29        1.52  2.56 2.56  1.90 2.56   2.50 2.36  2.37   2.28 
       Mar.   1.26        1.49  2.56 2.56  1.90 2.56   2.51 2.38  2.39   2.29 
       Apr.   1.25        1.48  2.56 2.56  1.90 2.56   2.49 2.38  2.40   2.30 

 Appendix IV Average Monthly and Annual Farm, Cooperative, Wholesale and Retail
                        Milk Prices in Selected Markets

                         (Continued From Previous Page)

                                Wholesale Retail

Year  Month  Farm  Cooperative  Whole   2%   1%  Skim Whole  2%   1%  Skim 
          May    1.26         1.49  2.56  2.56 1.90 2.56 2.54  2.42 2.43 2.32 
          June   1.26         1.49  2.56  2.56 1.90 2.56 2.54  2.42 2.43 2.32 
          July   1.25         1.49  2.56  2.56 1.91 2.56 2.51  2.40 2.41 2.31 
          Aug.   1.34         1.57  2.56  2.56 2.08 2.56 2.58  2.48 2.50 2.39 
         Sept.   1.53         1.77  2.56  2.56 2.27 2.56 2.69  2.59 2.62 2.50 
          Oct.   1.56         1.79  2.56  2.56 2.27 2.56 2.71  2.61 2.62 2.51 
          Nov.   1.56         1.79  2.56  2.56 2.27 2.56 2.69  2.59 2.57 2.49 
          Dec.   1.55         1.78  2.56  2.56 2.27 2.56 2.71  2.61 2.56 2.51 
          Avg.   1.37         1.60  2.56  2.56 2.04 2.56 2.58  2.47 2.47 2.38 
2004   Jan.   1.41         1.63  2.56  2.56 2.18 2.56 2.67  2.54 2.50 2.44 
          Feb.   1.41         1.63  2.56  2.56 2.08 2.56 2.63  2.49 2.46 2.41 
          Mar.   1.41         1.63  2.56  2.56 2.08 2.56 2.64  2.51 2.47 2.42 
          Apr.   1.55         1.77  2.56  2.56 2.09 2.56 2.74  2.62 2.58 2.53 
          May    2.02         2.25  2.56  2.56 2.66 2.56 3.07  2.97 2.92 2.86 
          Avg.   1.56         1.78  2.56  2.56 2.22 2.56 2.75  2.63 2.59 2.53 

Source: GAO's analysis of price data provided by USDA, the Defense
Commissary Agency, and Information Resources, Inc.

Note: For the Miami market, the farm price is the USDA-estimated farm
Class I price for 3.5 percent milkfat content; the cooperative price is
the announced cooperative Class I price for Miami for 3.5 percent milkfat
content; the wholesale price is the price paid by the commissary at the
Naval Air Station, Key West, adjusted for a 5 percent markup; and the
retail price is the price collected by Information Resources, Inc., for
the Miami market. Prices may not average due to rounding. "N/A" indicates
data not available.

Appendix IV Average Monthly and Annual Farm, Cooperative, Wholesale and
Retail Milk Prices in Selected Markets

  Table 18: Prices for a Gallon of Milk in Milwaukee, Wisconsin, October 2000
                       through May 2004 Wholesale Retail

  Year Month  Farm Cooperative Whole    2%    1% Skim  Whole    2%    1%  Skim 
  2000 Oct.  $1.20       $1.34 $1.87 $1.79 $1.76 $1.72 $1.97 $1.93 $2.01 $1.87 
       Nov.   1.20        1.33  1.87 1.79  1.76  1.72   1.98 1.95  2.03   1.89 
       Dec.   1.21        1.34  1.87 1.79  1.76  1.72   1.99 1.95  2.00   1.89 
       Avg.   1.20        1.34  1.87 1.79  1.76  1.72   1.98 1.94  2.01   1.88 
  2001 Jan.   1.39        1.54  2.24 2.15  2.10  2.04   2.12 2.09  2.15   1.99 
       Feb.   1.17        1.32  2.09 2.06  2.05  2.04   2.05 2.03  2.07   1.94 
       Mar.   1.24        1.39  2.07 2.05  2.05  2.04   2.08 2.03  2.10   1.96 
       Apr.   1.31        1.46  2.14 2.09  2.07  2.04   2.10 2.05  2.10   1.96 
        May   1.38        1.53  2.23 2.15  2.12  2.07   2.11 2.07  2.12   1.96 
       June   1.46        1.61  2.29 2.19  2.15  2.09   2.10 2.06  2.08   1.94 
       July   1.49        1.64  2.32 2.21  2.16  2.09   2.14 2.07  2.12   1.96 
       Aug.   1.50        1.65  2.34 2.23  2.17  2.11   2.15 2.09  2.13   1.97 
       Sept.  1.53        1.67  2.42 2.31  2.25  2.18   2.16 2.09  2.13   1.97 
       Oct.   1.55        1.70  2.44 2.30  2.23  2.15   2.18 2.12  2.16   1.99 
       Nov.   1.55        1.69  2.40 2.30  2.26  2.21   2.20 2.13  2.18   2.01 
       Dec.   1.22        1.36  2.19 2.05  2.00  1.98   2.11 2.04  2.08   1.94 
       Avg.   1.40        1.55  2.26 2.17  2.13  2.09   2.13 2.07  2.12   1.97 
  2002 Jan.   1.20        1.36  1.97 1.86  1.84  1.84   2.07 2.04  2.04   1.91 
       Feb.   1.18        1.35  1.92 1.81  1.81  1.81   2.08 2.03  2.03   1.90 
       Mar.   1.16        1.32  1.92 1.81  1.81  1.81   2.08 2.02  2.02   1.90 
       Apr.   1.16        1.32  1.92 1.81  1.81  1.81   2.08 2.05  2.06   1.92 
        May   1.13        1.30  1.92 1.81  1.81  1.81   2.06 2.00  2.03   1.89 
       June   1.13        1.29  1.92 1.81  1.81  1.81   2.08 2.02  2.05   1.92 
       July   1.09        1.25  1.92 1.81  1.81  1.81   2.08 2.00  2.04   1.90 
       Aug.   1.08        1.25  1.92 1.81  1.81  1.81   2.05 2.00  2.02   1.89 
       Sept.  1.08        1.25  1.92 1.81  1.81  1.81   2.05 1.99  2.03   1.89 
       Oct.   1.05        1.22  1.92 1.81  1.81  1.81   2.04 1.98  2.02   1.88 
       Nov.   1.09        1.26  1.92 1.81  1.81  1.81   2.06 1.99  2.03   1.88 
       Dec.   1.08        1.25  1.92 1.81  1.81  1.81   2.06 1.99  2.02   1.88 
       Avg.   1.12        1.29  1.92 1.81  1.81  1.81   2.07 2.01  2.03   1.90 
  2003 Jan.   1.09        1.27  1.92 1.81  1.81  1.81   2.06 1.98  2.01   1.87 
       Feb.   1.05        1.22  1.92 1.81  1.81  1.81   2.06 1.98  1.99   1.86 
       Mar.   1.01        1.18  1.92 1.81  1.81  1.81   2.06 1.98  2.01   1.87 
       Apr.   1.00        1.18  1.92 1.81  1.81  1.81   2.06 1.98  2.00   1.87 

 Appendix IV Average Monthly and Annual Farm, Cooperative, Wholesale and Retail
                        Milk Prices in Selected Markets

                         (Continued From Previous Page)

                                Wholesale Retail

Year  Month   Farm Cooperative  Whole   2%   1%  Skim Whole  2%   1%  Skim 
          May    1.01         1.18  1.92  1.81 1.81 1.81 2.05  1.95 1.97 1.85 
          June   1.01         1.18  1.92  1.81 1.81 1.81 2.03  1.93 1.96 1.83 
          July   1.01         1.18  1.92  1.81 1.81 1.81 2.04  1.94 1.98 1.84 
          Aug.   1.13         1.31  1.92  1.81 1.81 1.81 2.06  2.01 2.03 1.90 
         Sept.   1.37         1.54  1.92  1.81 1.81 1.81 2.14  2.06 2.09 1.94 
          Oct.   1.39         1.57  2.16  2.05 2.04 2.04 2.15  2.07 2.11 1.96 
          Nov.   1.40         1.58  2.17  2.06 2.05 2.05 2.16  2.10 2.12 1.98 
          Dec.   1.36         1.53  2.17  2.06 2.05 2.05 2.17  2.05 2.11 1.92 
          Avg.   1.15         1.33  1.98  1.87 1.87 1.87 2.09  2.00 2.03 1.89 
2004   Jan.   1.18         1.33  2.17  2.06 2.05 2.05 2.16  2.07 2.04 1.91 
          Feb.   1.16         1.32  2.17  2.06 2.05 2.05 2.14  2.06 2.01 1.89 
          Mar.   1.18         1.34  2.17  2.06 2.05 2.05 2.12  2.01 1.99 1.85 
          Apr.   1.37         1.52  2.17  2.06 2.05 2.05 2.17  2.08 2.03 1.88 
          May    2.01         2.17  2.53  2.42 2.28 2.19 2.52  2.45 2.41 2.18 
          Avg.   1.38         1.54  2.24  2.13 2.10 2.08 2.22  2.13 2.10 1.94 

Source: GAO's analysis of price data provided by USDA, the Defense
Commissary Agency, and Information Resources, Inc.

Note: For the Milwaukee market, the farm price is the USDA-estimated farm
Class I price for 3.5 percent milkfat content; the cooperative price is
the announced cooperative Class I price for Milwaukee for 3.5 percent
milkfat content; the wholesale price is the price paid by the commissary
at the Naval Station Great Lakes, adjusted for a 5 percent markup; and the
retail price is the price collected by Information Resources, Inc., for
the Milwaukee market. Prices may not average due to rounding.

Appendix IV Average Monthly and Annual Farm, Cooperative, Wholesale and
Retail Milk Prices in Selected Markets

 Table 19: Prices for a Gallon of Milk in Minneapolis, Minnesota, October 2000
                       through May 2004 Wholesale Retail

  Year Month  Farm Cooperative Whole  2%      1% Skim  Whole    2%    1%  Skim 
  2000 Oct.  $1.16       $1.27 $2.07 $2.05 $2.03 $2.01 $1.84 $1.74 $1.86 $1.68 
       Nov.   1.15        1.26 2.07  2.05  2.03  2.01   1.82 1.79  1.89   1.71 
       Dec.   1.16        1.27 2.07  2.05  2.03  2.01   1.79 1.79  1.92   1.74 
       Avg.   1.16        1.27 2.07  2.05  2.03  2.01   1.82 1.77  1.89   1.71 
  2001 Jan.   1.35        1.47 2.24  2.16  2.10  2.04   1.78 1.73  1.85   1.68 
       Feb.   1.14        1.27 2.08  2.06  2.06  2.04   1.68 1.67  1.80   1.64 
       Mar.   1.21        1.33 2.07  2.05  2.05  2.04   1.66 1.69  1.85   1.70 
       Apr.   1.28        1.41 2.14  2.09  2.07  2.04   1.67 1.72  1.86   1.76 
        May   1.35        1.48 2.23  2.15  2.12  2.07   1.65 1.70  1.88   1.75 
       June   1.43        1.56 2.29  2.19  2.15  2.09   1.68 1.71  1.91   1.78 
       July   1.46        1.58 2.32  2.21  2.16  2.09   1.68 1.73  1.94   1.80 
       Aug.   1.47        1.59 2.34  2.23  2.18  2.11   1.69 1.75  1.96   1.82 
       Sept.  1.50        1.62 2.42  2.30  2.25  2.18   1.69 1.76  1.99   1.83 
       Oct.   1.53        1.64 2.44  2.30  2.23  2.15   1.69 1.74  1.99   1.82 
       Nov.   1.51        1.64 2.43  2.37  2.34  2.30   1.74 1.80  2.05   1.88 
       Dec.   1.19        1.31 2.40  2.37  2.34  2.29   1.66 1.76  2.00   1.83 
       Avg.   1.37        1.49 2.28  2.21  2.17  2.12   1.69 1.73  1.92   1.77 
  2002 Jan.   1.17        1.32 2.40  2.37  2.34  2.29   1.68 1.74  1.99   1.81 
       Feb.   1.16        1.31 2.40  2.37  2.34  2.29   1.76 1.87  2.10   1.86 
       Mar.   1.13        1.28 2.40  2.38  2.35  2.30   1.86 2.01  2.20   1.95 
       Apr.   1.13        1.28 2.40  2.38  2.35  2.30   1.85 2.00  2.19   1.94 
        May   1.11        1.26 2.40  2.38  2.35  2.30   1.84 1.97  2.23   1.92 
       June   1.09        1.24 2.40  2.38  2.35  2.30   2.04 2.09  2.33   2.02 
       July   1.05        1.20 2.40  2.38  2.35  2.30   2.06 2.10  2.33   2.02 
       Aug.   1.04        1.19 2.40  2.38  2.35  2.30   2.04 2.05  2.31   1.97 
       Sept.  1.04        1.19 2.40  2.38  2.35  2.30   1.97 1.98  2.27   1.92 
       Oct.   1.02        1.16 2.40  2.38  2.35  2.30   1.78 1.85  2.14   1.83 
       Nov.   1.05        1.20 2.40  2.38  2.35  2.30   1.72 1.76  2.00   1.78 
       Dec.   1.05        1.19 2.40  2.38  2.35  2.30   1.83 1.95  2.17   1.90 
       Avg.   1.09        1.24 2.40  2.38  2.35  2.30   1.87 1.95  2.19   1.91 
  2003 Jan.   1.07        1.21 2.40  2.38  2.35  2.30   1.87 1.96  2.18   1.92 
       Feb.   1.02        1.16 2.40  2.38  2.35  2.30   1.75 1.79  2.00   1.78 
       Mar.   0.98        1.13 2.38  2.35  2.32  2.27   1.57 1.59  2.24   1.54 
       Apr.   0.97        1.12 2.35  2.32  2.29  2.24   1.52 1.53  2.50   1.46 

 Appendix IV Average Monthly and Annual Farm, Cooperative, Wholesale and Retail
                        Milk Prices in Selected Markets

                         (Continued From Previous Page)

                                Wholesale Retail

Year  Month   Farm Cooperative  Whole   2%   1%  Skim Whole  2%   1%  Skim 
          May    0.98         1.12  2.35  2.32 2.29 2.24 1.44  1.44 2.75 1.34 
          June   0.98         1.13  2.35  2.32 2.29 2.24 1.43  1.45 2.82 1.31 
          July   0.98         1.13  2.35  2.32 2.29 2.24 1.45  1.43 2.72 1.32 
          Aug.   1.10         1.25  2.35  2.32 2.29 2.24 1.52  1.46 2.84 1.35 
         Sept.   1.34         1.49  2.35  2.32 2.29 2.24 1.51  1.49 2.98 1.39 
          Oct.   1.37         1.51  2.41  2.37 2.34 2.27 1.53  1.55 2.81 1.48 
          Nov.   1.38         1.53  2.49  2.43 2.38 2.31 1.66  1.88 2.55 1.86 
          Dec.   1.33         1.48  2.49  2.43 2.38 2.31 1.57  1.88 2.57 1.91 
          Avg.   1.13         1.27  2.39  2.36 2.32 2.27 1.57  1.62 2.58 1.56 
2004   Jan.   1.15         1.30  2.49  2.43 2.38 2.31 1.65  1.94 2.57 1.94 
          Feb.   1.14         1.29  2.49  2.43 2.38 2.31 1.64  1.91 2.55 1.93 
          Mar.   1.16         1.31  2.49  2.43 2.38 2.31 1.60  1.80 2.54 1.81 
          Apr.   1.35         1.49  2.49  2.43 2.38 2.31 1.52  1.78 2.52 1.63 
          May    1.99         2.14  2.49  2.43 2.38 2.31 2.09  2.21 2.56 2.06 
          Avg.   1.36         1.51  2.49  2.43 2.38 2.31 1.70  1.93 2.55 1.87 

Source: GAO's analysis of price data provided by USDA, the Defense
Commissary Agency, and Information Resources, Inc.

Note: For the Minneapolis market, the farm price is the USDA-estimated
farm Class I price for 3.5 percent milkfat content; the cooperative price
is the announced cooperative Class I price for Minneapolis for 3.5 percent
milkfat content; the wholesale price is the price paid by the commissary
at Fort McCoy, adjusted for a 5 percent markup; and the retail price is
the price collected by Information Resources, Inc., for the Minneapolis
market. Prices may not average due to rounding.

Appendix IV Average Monthly and Annual Farm, Cooperative, Wholesale and
Retail Milk Prices in Selected Markets

 Table 20: Prices for a Gallon of Milk in New Orleans, Louisiana, October 2000
                       through May 2004 Wholesale Retail

  Year Month  Farm Cooperative Whole    2%    1% Skim  Whole    2%    1%  Skim 
  2000 Oct.  $1.22       $1.42 $2.36 $2.25 $2.25 $2.25 $2.64 $2.62 $2.62 $2.60 
       Nov.   1.22        1.41  2.32 2.25  2.25  2.25   2.59 2.60  2.54   2.56 
       Dec.   1.22        1.42  2.32 2.25  2.25  2.19   2.63 2.60  2.56   2.56 
       Avg.   1.22        1.42  2.33 2.25  2.25  2.23   2.62 2.61  2.57   2.57 
  2001 Jan.   1.38        1.58  2.40 2.32  2.32  2.25   2.68 2.67  2.63   2.62 
       Feb.   1.23        1.42  2.34 2.29  2.29  2.19   2.59 2.58  2.56   2.58 
       Mar.   1.27        1.46  2.34 2.29  2.29  2.19   2.57 2.59  2.60   2.60 
       Apr.   1.34        1.53  2.38 2.32  2.32  2.19   2.59 2.63  2.61   2.62 
        May   1.38        1.58  2.42 2.35  2.35  2.25   2.57 2.62  2.58   2.58 
       June   1.41        1.62  2.38 2.32  2.32  2.19   2.64 2.68  2.65   2.64 
       July   1.46        1.67  2.41 2.35  2.35  2.22   2.67 2.71  2.70   2.67 
       Aug.   1.47        1.67  2.44 2.38  2.38  2.25   2.70 2.74  2.70   2.68 
       Sept.  1.48        1.69  2.38 2.32  2.32  2.19   2.68 2.72  2.71   2.69 
       Oct.   1.51        1.72  2.38 2.32  2.32  2.19   2.70 2.73  2.74   2.71 
       Nov.   1.50        1.71  2.38 2.31  2.31  2.19   2.75 2.78  2.79   2.74 
       Dec.   1.21        1.42  2.29 2.23  2.23  2.19   2.73 2.75  2.80   2.74 
       Avg.   1.39        1.59  2.38 2.32  2.32  2.21   2.66 2.68  2.67   2.66 
  2002 Jan.   1.15        1.41  2.29 2.23  2.23  2.18   2.73 2.74  2.79   2.73 
       Feb.   1.15        1.41  2.27 2.21  2.21  2.15   2.74 2.74  2.77   2.71 
       Mar.   1.14        1.41  2.27 2.21  2.21  2.15   2.73 2.76  2.76   2.71 
       Apr.   1.15        1.41  2.27 2.21  2.21  2.15   2.72 2.74  2.74   2.71 
        May   1.15        1.42  2.25 2.21  2.21  2.17   2.71 2.72  2.73   2.69 
       June   1.13        1.40  2.30 2.23  2.23  2.19   2.70 2.71  2.72   2.68 
       July   1.09        1.36  2.31 2.23  2.24  2.19   2.70 2.71  2.72   2.69 
       Aug.   1.08        1.35  2.31 2.23  2.24  2.19   2.71 2.70  2.73   2.69 
       Sept.  1.08        1.35  2.31 2.23  2.24  2.19   2.64 2.64  2.67   2.64 
       Oct.   1.07        1.34  2.30 2.23  2.24  2.18   2.63 2.65  2.64   2.63 
       Nov.   1.10        1.36  2.27 2.19  2.22  2.16   2.61 2.65  2.60   2.62 
       Dec.   1.09        1.35  2.27 2.19  2.22  2.16   2.61 2.63  2.63   2.64 
       Avg.   1.12        1.38  2.29 2.22  2.23  2.17   2.69 2.70  2.71   2.68 
  2003 Jan.   1.06        1.36  2.27 2.19  2.22  2.16   2.60 2.62  2.61   2.60 
       Feb.   1.06        1.36  2.27 2.19  2.22  2.16   2.56 2.60  2.61   2.60 
       Mar.   1.02        1.32  2.27 2.19  2.22  2.16   2.56 2.58  2.60   2.60 
       Apr.   1.01        1.31  2.27 2.19  2.22  2.16   2.56 2.59  2.62   2.60 

 Appendix IV Average Monthly and Annual Farm, Cooperative, Wholesale and Retail
                        Milk Prices in Selected Markets

                         (Continued From Previous Page)

                                Wholesale Retail

Year  Month   Farm Cooperative  Whole   2%   1%  Skim Whole  2%   1%  Skim 
          May    1.01         1.31  2.27  2.19 2.22 2.16 2.57  2.61 2.64 2.61 
          June   1.02         1.31  2.27  2.19 2.22 2.16 2.64  2.67 2.70 2.64 
          July   1.01         1.32  2.28  2.19 2.22 2.16 2.58  2.62 2.66 2.62 
          Aug.   1.09         1.39  2.32  2.25 2.24 2.19 2.62  2.68 2.74 2.67 
         Sept.   1.28         1.59  2.38  2.32 2.31 2.22 2.72  2.80 2.89 2.80 
          Oct.   1.31         1.61  2.44  2.38 2.37 2.25 2.75  2.81 2.86 2.80 
          Nov.   1.31         1.61  2.34  2.25 2.24 2.10 2.74  2.79 2.83 2.79 
          Dec.   1.29         1.60  2.25  2.16 2.14 1.97 2.76  2.81 2.82 2.78 
          Avg.   1.12         1.42  2.30  2.22 2.24 2.15 2.64  2.68 2.72 2.68 
2004   Jan.   1.15         1.45  2.20  2.13 2.11 1.91 2.75  2.78 2.79 2.75 
          Feb.   1.15         1.45  2.14  2.08 2.05 1.85 2.71  2.74 2.74 2.70 
          Mar.   1.15         1.45  2.14  2.08 2.05 1.85 2.70  2.74 2.75 2.70 
          Apr.   1.31         1.60  2.18  2.11 2.07 1.86 2.80  2.85 2.88 2.80 
          May    1.78         2.08  2.44  2.33 2.27 2.04 3.18  3.25 3.28 3.18 
          Avg.   1.31         1.61  2.22  2.15 2.11 1.90 2.83  2.87 2.89 2.83 

Source: GAO's analysis of price data provided by USDA, the Defense
Commissary Agency, and Information Resources, Inc.

Note: For the New Orleans market, the farm price is the USDA-estimated
farm Class I price for 3.5 percent milkfat content; the cooperative price
is the announced cooperative Class I price for New Orleans for 3.5 percent
milkfat content; the wholesale price is the average of the prices paid by
the Naval Support Activities, New Orleans; the Naval Construction
Battalion Center, Gulfport; and Keesler Air Force Base commissaries,
adjusted for a 5 percent markup; and the retail price is the price
collected by Information Resources, Inc., for the New Orleans market.
Prices may not average due to rounding.

Appendix IV Average Monthly and Annual Farm, Cooperative, Wholesale and
Retail Milk Prices in Selected Markets

Table 21: Prices for a Gallon of Milk in Phoenix, Arizona, October 2000 through
                           May 2004 Wholesale Retail

  Year Month  Farm Cooperative Whole  2%      1% Skim  Whole    2%    1%  Skim 
  2000 Oct.  $1.17       $1.24 $1.97 $1.83 $1.73 $1.63 $2.25 $2.27 $2.44 $2.16 
       Nov.   1.16        1.23  1.98 1.83  1.74  1.63   2.20 2.17  2.39   2.10 
       Dec.   1.19        1.26  1.98 1.83  1.74  1.63   2.16 1.82  2.29   1.91 
       Avg.   1.17        1.24  1.98 1.83  1.74  1.63   2.20 2.09  2.37   2.06 
  2001 Jan.   1.35        1.42  1.99 1.84  1.74  1.62   2.60 2.58  2.60   2.49 
       Feb.   1.17        1.24  2.04 1.89  1.76  1.64   2.70 2.59  2.59   2.50 
       Mar.   1.23        1.31  2.15 1.97  1.81  1.66   2.48 2.44  2.47   2.37 
       Apr.   1.30        1.37  2.18 1.99  1.81  1.66   2.62 2.56  2.57   2.46 
        May   1.37        1.44  2.08 1.96  1.80  1.64   2.46 2.38  2.43   2.27 
       June   1.43        1.51  2.04 1.85  1.72  1.60   2.28 2.35  2.46   2.26 
       July   1.46        1.54  2.04 1.85  1.72  1.60   2.38 2.44  2.49   2.34 
       Aug.   1.47        1.54  2.05 1.86  1.72  1.60   2.20 2.23  2.31   2.20 
       Sept.  1.48        1.56  2.07 1.89  1.74  1.62   2.27 2.31  2.38   2.27 
       Oct.   1.52        1.59  2.10 1.92  1.74  1.61   2.41 2.47  2.53   2.39 
       Nov.   1.50        1.57  2.11 1.92  1.74  1.60   2.08 2.08  2.20   2.06 
       Dec.   1.18        1.25  2.11 1.97  1.85  1.76   2.47 2.56  2.54   2.40 
       Avg.   1.37        1.45  2.08 1.91  1.76  1.63   2.41 2.42  2.46   2.33 
  2002 Jan.   1.18        1.25  2.07 1.91  1.79  1.70   2.60 2.66  2.66   2.51 
       Feb.   1.17        1.25  2.09 1.88  1.78  1.69   2.52 2.52  2.58   2.51 
       Mar.   1.14        1.22  2.09 1.88  1.78  1.69   2.73 2.73  2.70   2.65 
       Apr.   1.13        1.20  2.06 1.88  1.78  1.69   2.53 2.64  2.64   2.50 
        May   1.11        1.19  1.97 1.88  1.78  1.69   2.71 2.73  2.72   2.69 
       June   1.09        1.17  1.97 1.88  1.78  1.69   2.58 2.58  2.66   2.55 
       July   1.06        1.13  1.97 1.88  1.78  1.69   2.68 2.73  2.78   2.65 
       Aug.   1.04        1.12  1.97 1.88  1.78  1.69   2.46 2.51  2.75   2.38 
       Sept.  1.03        1.12  1.97 1.88  1.78  1.69   2.36 2.25  2.37   2.28 
       Oct.   1.00        1.09  1.97 1.88  1.78  1.69   2.63 2.58  2.63   2.60 
       Nov.   1.05        1.13  1.97 1.88  1.78  1.69   2.48 2.57  2.60   2.51 
       Dec.   1.04        1.12  1.98 1.90  1.80  1.71   2.52 2.59  2.62   2.60 
       Avg.   1.09        1.17  2.01 1.88  1.78  1.69   2.57 2.59  2.64   2.54 
  2003 Jan.   1.04        1.13  1.99 1.92  1.81  1.73   2.49 2.58  2.61   2.55 
       Feb.   1.01        1.10  1.99 1.92  1.81  1.73   2.46 2.53  2.59   2.56 
       Mar.   0.98        1.06  1.99 1.92  1.81  1.73   2.43 2.43  2.52   2.48 
       Apr.   0.97        1.05  1.99 1.92  1.81  1.73   2.44 2.49  2.54   2.42 

 Appendix IV Average Monthly and Annual Farm, Cooperative, Wholesale and Retail
                        Milk Prices in Selected Markets

                         (Continued From Previous Page)

                                Wholesale Retail

Year  Month   Farm Cooperative  Whole   2%   1%  Skim Whole  2%   1%  Skim 
          May    0.97         1.05  1.94  1.88 1.81 1.73 2.43  2.43 2.51 2.44 
          June   0.98         1.06  1.89  1.84 1.80 1.73 2.51  2.55 2.59 2.52 
          July   0.97         1.06  1.89  1.84 1.80 1.73 2.30  2.32 2.45 2.27 
          Aug.   1.08         1.16  1.89  1.84 1.80 1.73 2.09  2.34 2.59 2.06 
         Sept.   1.31         1.40  2.08  1.99 1.88 1.79 2.44  2.45 2.63 2.45 
          Oct.   1.36         1.45  2.26  2.17 2.06 1.97 2.67  2.70 2.85 2.67 
          Nov.   1.37         1.45  2.27  2.18 2.07 1.98 2.33  2.36 2.59 2.38 
          Dec.   1.33         1.41  2.27  2.18 2.07 1.98 2.13  2.16 2.33 2.14 
          Avg.   1.11         1.20  2.04  1.97 1.88 1.80 2.39  2.45 2.57 2.41 
2004   Jan.   1.15         1.24  2.27  2.18 2.07 1.98 1.97  2.11 2.27 2.02 
          Feb.   1.13         1.21  2.27  2.18 2.07 1.98 2.07  2.10 2.22 2.05 
          Mar.   1.16         1.24  2.27  2.18 2.07 1.98 2.14  2.25 2.24 2.16 
          Apr.   1.31         1.39  2.23  1.94 1.91 1.76 2.00  2.10 2.18 2.03 
          May    1.83         1.91  2.77  2.35 2.33 2.14 2.62  2.67 2.65 2.55 
          Avg.   1.32         1.40  2.36  2.17 2.09 1.97 2.16  2.25 2.31 2.16 

Source: GAO's analysis of price data provided by USDA, the Defense
Commissary Agency, and Information Resources, Inc.

Note: For the Phoenix market, the farm price is the USDA-estimated farm
Class I price for 3.5 percent milkfat content; the cooperative price is
the announced cooperative Class I price for Phoenix for 3.5 percent
milkfat content; the wholesale price is the average of the prices paid by
the commissaries at Luke Air Force Base and Davis-Monthan Air Force Base,
adjusted for a 5 percent markup; and the retail price is the price
collected by Information Resources, Inc., for the Phoenix market. Prices
may not average due to rounding.

Appendix IV Average Monthly and Annual Farm, Cooperative, Wholesale and
Retail Milk Prices in Selected Markets

  Table 22: Prices for a Gallon of Milk in Salt Lake City, Utah, October 2000
                       through May 2004 Wholesale Retail

  Year Month  Farm Cooperative Whole  2%      1% Skim  Whole    2%    1%  Skim 
  2000 Oct.  $1.13       $1.21 $2.03 $1.96 $1.84 $1.80 $2.60 $2.18 $2.17 $2.44 
       Nov.   1.13        1.20  2.04 1.97  1.85  1.81   2.59 2.38  2.39   2.44 
       Dec.   1.15        1.23  2.06 1.98  1.86  1.81   2.58 2.18  2.24   2.44 
       Avg.   1.14        1.21  2.04 1.97  1.85  1.81   2.59 2.25  2.27   2.44 
  2001 Jan.   1.31        1.39  2.14 2.03  1.89  1.81   2.68 2.27  2.17   2.42 
       Feb.   1.13        1.21  2.15 2.04  1.89  1.82   2.58 2.21  2.19   2.31 
       Mar.   1.19        1.28  2.09 2.00  1.87  1.82   2.50 2.23  2.21   2.21 
       Apr.   1.26        1.34  2.15 2.04  1.89  1.82   2.34 2.28  2.27   2.09 
        May   1.33        1.41  2.22 2.08  1.91  1.82   2.13 2.09  2.03   2.18 
       June   1.39        1.48  2.28 2.12  1.93  1.83   2.38 2.18  2.16   2.21 
       July   1.42        1.51  2.34 2.15  1.96  1.84   2.39 2.19  2.12   2.00 
       Aug.   1.43        1.51  2.36 2.17  1.97  1.84   2.39 2.13  2.13   2.00 
       Sept.  1.44        1.53  2.37 2.18  1.97  1.85   2.46 2.20  2.20   2.06 
       Oct.   1.48        1.56  2.39 2.19  1.97  1.83   2.39 2.35  2.37   2.21 
       Nov.   1.47        1.55  2.41 2.22  2.02  1.91   2.41 2.36  2.33   2.22 
       Dec.   1.14        1.22  2.28 2.14  1.98  1.90   2.37 2.35  2.34   2.15 
       Avg.   1.33        1.42  2.27 2.11  1.94  1.84   2.42 2.24  2.21   2.17 
  2002 Jan.   1.14        1.22  2.16 2.04  1.89  1.83   2.25 2.24  2.24   2.06 
       Feb.   1.14        1.22  2.16 2.04  1.89  1.83   2.46 2.41  2.42   2.18 
       Mar.   1.11        1.19  2.15 2.04  1.89  1.83   2.41 2.33  2.34   2.17 
       Apr.   1.10        1.18  2.02 1.96  1.88  1.75   2.37 2.33  2.29   2.13 
        May   1.08        1.16  1.89 1.88  1.86  1.67   2.34 2.31  2.30   2.13 
       June   1.06        1.14  1.88 1.88  1.85  1.66   2.39 2.38  2.42   2.17 
       July   1.03        1.11  1.85 1.87  1.86  1.68   2.38 2.30  2.26   2.17 
       Aug.   1.01        1.09  1.84 1.86  1.85  1.67   2.37 2.33  2.37   2.10 
       Sept.  1.01        1.09  1.84 1.85  1.85  1.68   2.16 2.01  1.94   1.91 
       Oct.   0.99        1.06  1.84 1.84  1.84  1.68   2.46 2.42  2.40   2.17 
       Nov.   1.02        1.10  1.85 1.85  1.85  1.70   2.38 2.44  2.41   2.17 
       Dec.   1.02        1.10  1.87 1.87  1.86  1.71   2.10 1.97  1.99   1.84 
       Avg.   1.06        1.14  1.95 1.92  1.86  1.72   2.34 2.29  2.28   2.10 
  2003 Jan.   1.02        1.10  1.87 1.86  1.85  1.70   2.10 2.06  2.01   1.83 
       Feb.   0.99        1.07  1.86 1.85  1.84  1.68   2.09 2.10  2.08   1.88 
       Mar.   0.96        1.04  1.85 1.84  1.82  1.66   2.07 2.05  2.03   1.83 
       Apr.   0.94        1.02  1.88 1.87  1.85  1.69   2.10 2.09  2.07   1.85 

 Appendix IV Average Monthly and Annual Farm, Cooperative, Wholesale and Retail
                        Milk Prices in Selected Markets

                         (Continued From Previous Page)

                                Wholesale Retail

Year  Month   Farm Cooperative  Whole   2%    1% Skim Whole  2%   1%  Skim 
          May    0.95         1.03  1.90  1.89 1.87 1.71 2.12  2.11 2.09 1.89 
          June   0.95         1.03  1.90  1.89 1.87 1.71 1.96  2.00 2.01 1.83 
          July   0.95         1.03  1.90  1.89 1.87 1.71 2.04  2.03 2.01 1.86 
          Aug.   1.05         1.14  1.96  1.95 1.94 1.77 2.32  2.33 2.23 2.11 
         Sept.   1.29         1.37  2.13  2.10 2.09 1.92 2.37  2.51 2.28 2.32 
          Oct.   1.34         1.42  2.27  2.24 2.23 2.07 2.22  2.29 2.28 2.18 
          Nov.   1.35         1.43  2.29  2.28 2.26 2.09 2.00  2.05 2.06 1.88 
          Dec.   1.30         1.38  2.28  2.26 2.24 2.07 2.03  2.10 2.10 1.95 
          Avg.   1.09         1.17  2.01  1.99 1.98 1.82 2.12  2.14 2.10 1.95 
2004   Jan.   1.14         1.21  2.21  2.18 2.16 1.99 2.00  2.06 2.08 1.93 
          Feb.   1.12         1.19  2.12  2.08 2.04 1.86 1.96  2.03 2.09 1.90 
          Mar.   1.14         1.22  2.13  2.06 2.00 1.80 1.99  2.06 2.13 1.95 
          Apr.   1.29         1.37  2.20  2.10 2.01 1.78 2.19  2.22 2.21 2.02 
          May    1.81         1.88  2.52  2.38 2.28 2.03 2.50  2.58 2.63 2.38 
          Avg.   1.30         1.37  2.24  2.16 2.10 1.89 2.13  2.19 2.23 2.04 

Source: GAO's analysis of price data provided by USDA, the Defense
Commissary Agency, and Information Resources, Inc.

Note: For the Salt Lake City market, the farm price is the USDA-estimated
farm Class I price for 3.5 percent milkfat content; the cooperative price
is the announced cooperative Class I price for Salt Lake City for 3.5
percent milkfat content; the wholesale price is the price paid by the
commissary at Hill Air Force Base, adjusted for a 5 percent markup; and
the retail price is the price collected by Information Resources, Inc.,
for the Salt Lake City market. Prices may not average due to rounding.

Appendix IV Average Monthly and Annual Farm, Cooperative, Wholesale and
Retail Milk Prices in Selected Markets

  Table 23: Prices for a Gallon of Milk in San Diego, California, October 2000
                       through May 2004 Wholesale Retail

  Year Month  Farm Cooperative Whole    2%    1% Skim  Whole    2%    1%  Skim 
  2000 Oct.  $0.99       $1.21 $1.90 $1.90 $1.87 $1.69 $2.55 $2.45 $2.66 $2.29 
       Nov.   1.04        1.21  1.88 1.88  1.85  1.66   2.51 2.54  2.54   2.23 
       Dec.   1.06        1.28  1.95 1.91  1.88  1.67   2.72 2.63  2.79   2.40 
       Avg.   1.03        1.23  1.91 1.90  1.87  1.67   2.59 2.54  2.66   2.31 
  2001 Jan.   1.03        1.43  2.03 1.98  1.95  1.68   2.81 2.72  2.93   2.52 
       Feb.   1.05        1.21  1.92 1.91  1.91  1.66   2.69 2.75  2.71   2.36 
       Mar.   1.11        1.26  1.98 1.97  1.94  1.67   2.68 2.57  2.77   2.40 
       Apr.   1.18        1.33  2.05 2.01  1.96  1.67   2.81 2.74  2.84   2.39 
        May   1.25        1.39  2.11 2.06  1.98  1.67   2.85 2.71  2.81   2.30 
       June   1.31        1.46  2.13 2.07  1.98  1.67   2.91 2.85  2.80   2.34 
       July   1.30        1.49  2.15 2.08  1.99  1.67   3.00 2.94  2.99   2.50 
       Aug.   1.34        1.49  2.19 2.11  2.00  1.67   3.10 2.99  3.04   2.55 
       Sept.  1.37        1.49  2.19 2.11  2.00  1.67   3.22 2.95  3.24   2.62 
       Oct.   1.21        1.55  2.23 2.13  2.00  1.65   3.27 2.95  3.24   2.61 
       Nov.   1.11        1.41  2.11 2.09  2.02  1.71   3.18 2.93  2.81   2.52 
       Dec.   1.06        1.19  2.01 2.00  2.00  1.64   3.06 2.94  3.08   2.52 
       Avg.   1.19        1.39  2.09 2.04  1.98  1.67   2.97 2.84  2.94   2.47 
  2002 Jan.   1.07        1.18  1.99 1.98  1.99  1.64   3.05 2.91  2.94   2.46 
       Feb.   1.01        1.20  1.97 1.97  1.97  1.64   2.95 2.76  2.72   2.40 
       Mar.   0.99        1.16  1.92 1.93  1.96  1.63   2.85 2.71  2.50   2.32 
       Apr.   0.98        1.16  1.88 1.89  1.88  1.61   2.76 2.47  2.63   2.40 
        May   0.94        1.14  1.82 1.81  1.79  1.58   2.88 2.67  2.73   2.35 
       June   0.90        1.10  1.79 1.80  1.78  1.58   2.87 2.62  2.85   2.40 
       July   0.87        1.10  1.77 1.78  1.77  1.57   2.85 2.49  2.77   2.37 
       Aug.   0.89        1.08  1.75 1.75  1.76  1.57   2.69 2.58  2.63   2.28 
       Sept.  0.91        1.08  1.75 1.75  1.76  1.57   2.72 2.64  2.72   2.29 
       Oct.   0.94        1.05  1.75 1.76  1.77  1.58   2.75 2.71  2.74   2.30 
       Nov.   0.92        1.16  1.83 1.85  1.86  1.65   2.80 2.70  2.80   2.30 
       Dec.   0.92        1.08  1.77 1.78  1.78  1.58   2.73 2.71  2.74   2.23 
       Avg.   0.95        1.12  1.83 1.84  1.84  1.60   2.83 2.66  2.73   2.34 
  2003 Jan.   0.92        1.11  1.77 1.77  1.77  1.57   2.68 2.59  2.84   2.28 
       Feb.   0.89        1.06  1.74 1.71  1.71  1.50   2.74 2.63  2.94   2.29 
       Mar.   0.87        1.05  1.72 1.71  1.69  1.44   2.70 2.58  2.78   2.27 
       Apr.   0.88        1.02  1.70 1.70  1.67  1.43   2.72 2.60  2.80   2.27 

 Appendix IV Average Monthly and Annual Farm, Cooperative, Wholesale and Retail
                        Milk Prices in Selected Markets

                         (Continued From Previous Page)

                                Wholesale Retail

Year  Month   Farm Cooperative  Whole   2%   1%  Skim Whole  2%   1%  Skim 
          May    0.87         1.03  1.68  1.69 1.65 1.42 2.81  2.67 2.80 2.29 
          June   0.87         1.04  1.68  1.69 1.65 1.42 2.80  2.68 2.78 2.30 
          July   0.99         1.03  1.70  1.71 1.68 1.44 2.82  2.70 2.84 2.32 
          Aug.   1.10         1.27  1.86  1.89 1.87 1.60 3.00  2.78 2.98 2.36 
         Sept.   1.14         1.41  2.03  2.06 2.04 1.73 3.13  2.93 3.07 2.50 
          Oct.   1.16         1.42  2.03  2.08 2.07 1.76 3.19  2.86 3.23 2.43 
          Nov.   1.10         1.42  2.03  2.07 2.06 1.75 3.12  2.80 3.29 2.44 
          Dec.   1.08         1.32  1.94  1.98 1.97 1.65 2.72  2.49 2.93 2.23 
          Avg.   0.99         1.18  1.82  1.84 1.82 1.56 2.87  2.69 2.94 2.33 
2004   Jan.   1.04         1.22  1.85  1.87 1.85 1.55 2.64  2.45 2.82 2.19 
          Feb.   1.09         1.17  1.77  1.77 1.74 1.44 2.62  2.43 2.72 2.17 
          Mar.   1.26         1.21  1.80  1.77 1.71 1.40 2.55  2.34 2.76 2.13 
          Apr.   1.48         1.38  1.99  1.87 1.76 1.43 2.76  2.49 2.63 2.15 
          May    1.52         1.85  2.41  2.34 2.25 1.83 3.03  2.78 2.93 2.40 
          Avg.   1.28         1.37  1.96  1.92 1.86 1.53 2.72  2.50 2.77 2.21 

Source: GAO's analysis of price data provided by the California Department
of Food and Agriculture, the Defense Commissary Agency, and Information
Resources, Inc.

Note: For the San Diego market, the farm price is the California mailbox
price for 3.5 percent milkfat content; the cooperative price is the
Southern California Class I price for 3.5 percent milkfat content; the
wholesale price is the average of the prices paid by the commissaries at
the Naval Base, San Diego; Marine Corps Base, Camp Pendleton; Naval
Outlying Landing Field, Imperial Beach; Marine Corps Air Station, Miramar;
Naval Air Station, North Island; and by the San Onofre Commissary, Camp
Pendleton, adjusted for a 5 percent markup; and the retail price is the
price collected by Information Resources, Inc., for the San Diego market.
Prices may not average due to rounding.

Appendix IV Average Monthly and Annual Farm, Cooperative, Wholesale and
Retail Milk Prices in Selected Markets

Table 24: Prices for a Gallon of Milk in Seattle, Washington, October 2000
                       through May 2004 Wholesale Retail

  Year Month  Farm Cooperative Whole  2%      1% Skim  Whole    2%    1%  Skim 
  2000 Oct.  $1.20       $1.28 $1.59 $1.46 $1.38 $1.31 $2.71 $2.50 $2.50 $2.37 
       Nov.   1.14        1.22  1.58 1.46  1.37  1.30   2.83 2.62  2.58   2.49 
       Dec.   1.18        1.27  1.61 1.48  1.40  1.33   2.73 2.57  2.57   2.45 
       Avg.   1.17        1.26  1.59 1.47  1.38  1.31   2.76 2.56  2.55   2.44 
  2001 Jan.   1.32        1.41  1.76 1.63  1.55  1.48   2.62 2.39  2.40   2.25 
       Feb.   1.17        1.25  1.60 1.48  1.39  1.33   2.98 2.77  2.71   2.62 
       Mar.   1.21        1.29  1.64 1.51  1.43  1.36   2.77 2.52  2.52   2.29 
       Apr.   1.27        1.36  1.71 1.59  1.50  1.43   2.95 2.72  2.67   2.54 
        May   1.34        1.42  1.78 1.66  1.57  1.51   2.95 2.65  2.67   2.44 
       June   1.41        1.49  1.84 1.72  1.63  1.57   2.80 2.77  2.70   2.49 
       July   1.44        1.52  1.87 1.75  1.66  1.60   3.14 3.01  2.85   2.71 
       Aug.   1.44        1.53  1.88 1.76  1.67  1.61   3.15 2.99  2.85   2.71 
       Sept.  1.46        1.54  1.89 1.77  1.68  1.62   2.84 2.65  2.62   2.43 
       Oct.   1.49        1.57  1.92 1.80  1.71  1.64   2.92 2.72  2.69   2.46 
       Nov.   1.47        1.56  1.91 1.78  1.70  1.63   3.01 2.76  2.73   2.50 
       Dec.   1.15        1.23  1.60 1.48  1.39  1.32   2.93 2.75  2.68   2.47 
       Avg.   1.35        1.43  1.78 1.66  1.57  1.51   2.92 2.73  2.67   2.49 
  2002 Jan.   1.15        1.23  1.60 1.47  1.39  1.32   3.18 2.93  2.82   2.67 
       Feb.   1.25        1.33  1.60 1.47  1.39  1.32   3.13 2.86  2.74   2.62 
       Mar.   1.12        1.20  1.57 1.44  1.36  1.29   3.16 2.91  2.78   2.68 
       Apr.   1.11        1.19  1.56 1.43  1.35  1.28   2.99 2.66  2.59   2.42 
        May   1.09        1.17  1.54 1.41  1.33  1.26   3.08 2.70  2.68   2.48 
       June   1.07        1.15  1.52 1.40  1.31  1.24   3.08 2.56  2.54   2.32 
       July   1.03        1.12  1.48 1.36  1.27  1.21   3.07 2.67  2.69   2.49 
       Aug.   1.02        1.10  1.47 1.35  1.26  1.20   2.81 2.62  2.42   2.43 
       Sept.  1.02        1.10  1.47 1.35  1.26  1.20   3.01 2.70  2.71   2.47 
       Oct.   0.99        1.07  1.45 1.32  1.24  1.17   3.10 2.77  2.73   2.56 
       Nov.   1.03        1.11  1.47 1.34  1.26  1.19   2.98 2.73  2.69   2.41 
       Dec.   1.02        1.11  1.47 1.35  1.26  1.20   2.98 2.76  2.75   2.50 
       Avg.   1.08        1.16  1.52 1.39  1.31  1.24   3.05 2.74  2.68   2.50 
  2003 Jan.   1.03        1.11  1.47 1.35  1.26  1.19   3.12 3.04  2.91   2.78 
       Feb.   1.00        1.08  1.44 1.32  1.23  1.17   3.07 2.92  2.73   2.53 
       Mar.   0.96        1.05  1.42 1.29  1.21  1.14   2.97 2.75  2.66   2.42 
       Apr.   0.95        1.03  1.40 1.27  1.19  1.12   3.00 2.93  2.72   2.49 

 Appendix IV Average Monthly and Annual Farm, Cooperative, Wholesale and Retail
                        Milk Prices in Selected Markets

                         (Continued From Previous Page)

                                Wholesale Retail

Year  Month   Farm Cooperative  Whole   2%   1%  Skim Whole  2%   1%  Skim 
          May    0.95         1.04  1.44  1.32 1.23 1.16 2.94  2.81 2.66 2.45 
          June   0.96         1.04  1.44  1.32 1.24 1.17 3.00  2.82 2.76 2.54 
          July   0.95         1.04  1.45  1.33 1.24 1.18 3.07  2.86 2.82 2.55 
          Aug.   1.05         1.15  1.55  1.43 1.36 1.28 3.11  2.93 2.90 2.65 
         Sept.   1.29         1.38  1.81  1.69 1.63 1.54 3.22  3.00 3.01 2.72 
          Oct.   1.34         1.43  1.85  1.72 1.88 1.58 3.06  2.98 2.97 2.69 
          Nov.   1.35         1.44  1.86  1.73 1.82 1.59 3.10  2.89 2.93 2.63 
          Dec.   1.30         1.39  1.82  1.69 1.99 1.55 2.75  2.62 2.49 2.39 
          Avg.   1.09         1.18  1.58  1.46 1.44 1.31 3.03  2.88 2.80 2.57 
2004   Jan.   1.13         1.22  1.69  1.56 1.92 1.42 2.59  2.56 2.47 2.37 
          Feb.   1.10         1.20  1.65  1.53 1.72 1.38 2.81  2.50 2.45 2.33 
          Mar.   1.13         1.23  1.73  1.60 1.67 1.44 2.44  2.46 2.32 2.16 
          Apr.   1.28         1.38  1.90  1.76 1.68 1.64 2.56  2.53 2.46 2.32 
          May    1.80         1.89  2.38  2.24 1.96 2.13 3.06  2.88 2.82 2.61 
          Avg.   1.29         1.38  1.87  1.74 1.79 1.60 2.69  2.59 2.50 2.36 

Source: GAO's analysis of price data provided by USDA, the Defense
Commissary Agency, and Information Resources, Inc.

Note: For the Seattle market, the farm price is the USDA-estimated farm
Class I price for 3.5 percent milkfat content; the cooperative price is
the announced cooperative Class I price for Seattle for 3.5 percent
milkfat content; the wholesale price is the average of the prices paid by
the commissaries at the Naval Station Everett, Smokey Point Support
Center; Fort Lewis; and McChord Air Force Base, adjusted for a 5 percent
markup; and the retail price is the price collected by Information
Resources, Inc., for the Seattle market. Prices may not average due to
rounding.

Appendix IV Average Monthly and Annual Farm, Cooperative, Wholesale and
Retail Milk Prices in Selected Markets

Table 25: Prices for a Gallon of Milk in Washington, D.C., October 2000 through
                           May 2004 Wholesale Retail

  Year Month  Farm Cooperative Whole    2%    1% Skim  Whole    2%    1%  Skim 
  2000 Oct.  $1.32       $1.40 $1.66 $1.51 $1.49 $1.42 $2.10 $2.10 $2.09 $1.94 
       Nov.   1.33        1.41  1.66 1.50  1.49  1.42   2.09 2.10  2.09   1.94 
       Dec.   1.36        1.44  1.68 1.53  1.51  1.43   2.11 2.12  2.11   1.97 
       Avg.   1.34        1.42  1.67 1.51  1.50  1.42   2.10 2.11  2.10   1.95 
  2001 Jan.   1.54        1.62  1.83 1.67  1.62  1.54   2.21 2.24  2.22   2.05 
       Feb.   1.36        1.44  1.67 1.52  1.53  1.44   2.22 2.25  2.23   2.04 
       Mar.   1.43        1.50  1.73 1.57  1.57  1.48   2.21 2.23  2.22   2.05 
       Apr.   1.49        1.57  1.79 1.64  1.61  1.52   2.25 2.27  2.27   2.10 
        May   1.56        1.63  1.86 1.70  1.66  1.58   2.29 2.31  2.32   2.15 
       June   1.63        1.70  1.93 1.77  1.72  1.65   2.32 2.31  2.30   2.13 
       July   1.63        1.71  1.96 1.80  1.74  1.68   2.29 2.27  2.25   2.10 
       Aug.   1.64        1.71  1.96 1.80  1.74  1.68   2.29 2.28  2.26   2.10 
       Sept.  1.65        1.72  1.97 1.80  1.75  1.69   2.31 2.31  2.30   2.13 
       Oct.   1.68        1.76  1.99 1.83  1.78  1.71   2.37 2.35  2.34   2.17 
       Nov.   1.69        1.77  1.99 1.83  1.79  1.74   2.38 2.39  2.40   2.25 
       Dec.   1.37        1.44  1.73 1.57  1.60  1.60   2.35 2.36  2.35   2.19 
       Avg.   1.56        1.63  1.87 1.71  1.68  1.61   2.29 2.30  2.29   2.12 
  2002 Jan.   1.37        1.44  1.71 1.54  1.53  1.55   2.36 2.36  2.35   2.18 
       Feb.   1.37        1.44  1.68 1.52  1.52  1.52   2.33 2.33  2.32   2.15 
       Mar.   1.34        1.41  1.65 1.49  1.50  1.50   2.30 2.31  2.30   2.15 
       Apr.   1.33        1.40  1.64 1.48  1.43  1.49   2.30 2.30  2.30   2.14 
        May   1.31        1.38  1.63 1.46  1.42  1.48   2.32 2.31  2.32   2.16 
       June   1.29        1.36  1.61 1.44  1.40  1.46   2.29 2.27  2.27   2.12 
       July   1.25        1.32  1.57 1.41  1.37  1.44   2.31 2.29  2.29   2.13 
       Aug.   1.24        1.31  1.56 1.40  1.36  1.43   2.29 2.27  2.28   2.13 
       Sept.  1.24        1.31  1.56 1.40  1.36  1.43   2.27 2.27  2.29   2.13 
       Oct.   1.21        1.28  1.54 1.37  1.34  1.41   2.25 2.25  2.27   2.12 
       Nov.   1.25        1.32  1.57 1.41  1.37  1.44   2.19 2.20  2.22   2.08 
       Dec.   1.24        1.32  1.56 1.40  1.36  1.43   2.24 2.25  2.24   2.09 
       Avg.   1.29        1.36  1.61 1.44  1.41  1.47   2.29 2.28  2.29   2.13 
  2003 Jan.   1.24        1.32  1.56 1.40  1.36  1.43   2.20 2.22  2.20   2.06 
       Feb.   1.20        1.28  1.53 1.37  1.33  1.41   2.19 2.23  2.22   2.08 
       Mar.   1.17        1.24  1.50 1.33  1.30  1.38   2.18 2.19  2.20   2.06 
       Apr.   1.17        1.24  1.48 1.32  1.29  1.38   2.22 2.22  2.21   2.07 

 Appendix IV Average Monthly and Annual Farm, Cooperative, Wholesale and Retail
                        Milk Prices in Selected Markets

                         (Continued From Previous Page)

                                Wholesale Retail

Year  Month   Farm Cooperative  Whole   2%   1%  Skim Whole  2%   1%  Skim 
          May    1.17         1.25  1.49  1.33 1.30 1.38 2.23  2.22 2.22 2.07 
          June   1.18         1.25  1.49  1.33 1.30 1.38 2.25  2.22 2.20 2.06 
          July   1.19         1.26  1.50  1.34 1.30 1.39 2.28  2.23 2.20 2.07 
          Aug.   1.29         1.36  1.58  1.43 1.38 1.45 2.34  2.29 2.27 2.14 
         Sept.   1.52         1.60  1.74  1.61 1.55 1.59 2.50  2.47 2.42 2.28 
          Oct.   1.57         1.65  1.86  1.70 1.65 1.71 2.58  2.53 2.51 2.35 
          Nov.   1.58         1.66  1.87  1.71 1.66 1.72 2.59  2.53 2.50 2.34 
          Dec.   1.53         1.61  1.83  1.67 1.63 1.70 2.59  2.55 2.52 2.37 
          Avg.   1.32         1.39  1.62  1.46 1.42 1.49 2.35  2.33 2.31 2.16 
2004   Jan.   1.36         1.44  1.69  1.53 1.49 1.58 2.59  2.54 2.51 2.36 
          Feb.   1.34         1.42  1.66  1.50 1.46 1.55 2.56  2.50 2.48 2.33 
          Mar.   1.37         1.45  1.66  1.50 1.45 1.54 2.53  2.48 2.45 2.31 
          Apr.   1.51         1.59  1.82  1.66 1.61 1.66 2.58  2.49 2.43 2.25 
          May    2.03         2.11  2.28  2.12 2.05 2.04 2.90  2.78 2.72 2.51 
          Avg.   1.52         1.60  1.82  1.66 1.61 1.67 2.63  2.56 2.52 2.35 

Source: GAO's analysis of price data provided by USDA, the Defense
Commissary Agency, and Information Resources, Inc.

Note: For the Washington, D.C., market, the farm price is the
USDA-estimated farm Class I price for 3.5 percent milkfat content; the
cooperative price is the announced cooperative Class I price for
Washington, D.C., for 3.5 percent milkfat content; the wholesale price is
the average of the prices paid by the commissaries at Bolling Air Force
Base; Walter Reed Army Medical Center; Fort Myer; Fort Belvoir; the Marine
Corps Base, Quantico; Andrews Air Force Base; Aberdeen Proving Ground; the
Naval Station, Annapolis; Fort Meade; and Fort Detrick, adjusted for a 5
percent markup; and the retail price is the price collected by Information
Resources, Inc., for the Washington, D.C., market. Prices may not average
due to rounding.

Appendix V

Factors That Influence the Price of Milk as It Moves from the Farm to the
Consumer

The prices that farmers, cooperatives, wholesale processors, and retailers
receive are determined by the interaction of many factors, such as forces
affecting the supply of raw milk and manufactured and fluid milk products,
consumer demand for manufactured and fluid milk products, federal and
state dairy programs, the level of services provided by dairy
cooperatives, market structure at various levels of the marketing chain,
and other input costs of processing and retailing. Dairy farmers receive a
price for raw milk, and each entity involved in the processing and
marketing of fluid milk adds value to the product and retains a portion of
the difference between the farm and retail prices. (This difference is
known as the price spread.) This appendix examines the key factors that
influence milk prices at the different levels of the marketing chain.

Market Forces, in Addition to Federal and State Policies, Influence Milk
Prices at the Farm Level

Supply and demand forces, which in turn are influenced by federal and
state dairy programs, determine farm prices for the raw milk that is sold
for use in fluid milk and other dairy products.1 For example, in recent
months, a variety of supply and demand forces have come together to
significantly increase farm milk prices. On the supply side, the available
supply of raw milk has been reduced by farmers cutting back production due
to a previous period of low prices and by the closing of the Canadian
border to replacement cows as a result of concerns about mad cow disease.2
On the demand side, consumer demand for nonfluid dairy products has
increased as consumers resumed eating out following the attacks of
September 11, 2001, and as dietary trends, such as the rising popularity
of lowcarbohydrate diets, have changed. While these forces have been
driving recent price trends, federal and state dairy programs continue to
influence milk prices. For example, major domestic programs such as
federal milk marketing orders (FMMOs) and price supports help individual
farmers who lack market power compared to other entities such as wholesale
processors and retailers and help to ensure that farm prices do not fall
below a minimum level. At the same time, U.S. import restrictions maintain
domestic dairy prices at levels higher than average international market

1Each year the United States processes about 7 billion gallons of the
approximately 20 billion gallons of raw milk into fluid milk products,
such as flavored milks, buttermilk, whole, 2 percent, 1 percent, and skim
milk. The rest of the raw milk supply is used to produce manufactured
products, such as butter, ice cream, yogurt, powdered milk, and cheese.

2Bovine spongiform encephalopathy (BSE), commonly referred to as mad cow
disease, is a degenerative neurological disease affecting the central
nervous system in cattle.

                                   Appendix V
                 Factors That Influence the Price of Milk as It
                      Moves from the Farm to the Consumer

prices by limiting the quantities of milk products that are imported into
the country.

Supply and Demand Forces Affect Farm Milk Prices

The quantity of raw milk that dairy farmers supply (production) is
determined by the operating costs of producing that milk, such as feed and
fuel,3 ownership costs for dairying equipment,4 land costs, and labor
costs, as well as the price that farmers expect to receive for milk (as
based on demand). Of these costs, the 2002 annual report on the costs of
milk production in California, the largest milk producing state, showed
that the highest cost is feed, at 44 percent of milk production costs.
Other major costs include replacement cows (14 percent), other operating
expenses (13 percent), and labor (11 percent).5 Milk production can also
vary seasonally, according to weather, and is affected by farmers'
management practices.

In February 2004, the U.S. Department of Agriculture (USDA) published a
report on the characteristics and costs of milk production in the United
States, which found that dairy farmers in the West had a cost advantage
over farmers in other regions because western operations were appreciably
larger.6 Farms with 500 or more milk cows had substantially lower total
operating and ownership costs, averaging $11.60 per hundredweight of milk
sold. This cost advantage arises because as herd size increases,
associated increases in fixed costs, such as capital investments, are
spread proportionally over a larger amount of production, thereby lowering
the fixed costs per hundredweight of milk produced. USDA found that the
average herd size of low-cost operations was more than three times the
size

3Operating costs include major inputs such as feed, veterinary services
and medicine, bedding and litter, marketing, fuel, lubricant, electricity,
repairs, and interest on operating expenses.

4Ownership costs include the annualized cost of maintaining the capital
investment (depreciation and interest) in dairy facilities and equipment,
and costs for non-real estate property taxes and insurance.

5This labor estimate includes salary, benefits, and all employer taxes for
hired labor. If family labor were also included, labor would represent
over 12 percent of total production costs.

6S.D. Short, U.S. Department of Agriculture, Characteristics and
Production Costs of U.S. Dairy Operations. Resource Economics Division,
Economic Research Service, Statistical Bulletin Number 974-6, February
2004.

Appendix V
Factors That Influence the Price of Milk as It
Moves from the Farm to the Consumer

of high-cost operations.7 Table 26 shows the average ownership and
operating costs by region and herd size in 2000.

Table 26: Average Ownership and Operating Costs by Region and Herd Size,
2000

Ownership and operating costs per Region hundredweight

Arizona/Las Vegas, Pacific Northwest,
Californiaa $11.58

                             Upper Midwestb $14.40

                               Northeastc $14.44

                                Centrald $18.09

                           Mideast/Southeaste $18.23

Herd size

                               Small (<50) $17.98

                             Medium (50-199) $15.16

                             Large (200-499) $13.32

                            Industrial (>500) $11.60

Source: USDA.

aStates in this survey region included parts of Arizona, California,
Idaho, Oregon, and Washington.

bStates in this survey region included Michigan and Wisconsin and part of
Minnesota.

cStates in this survey region included Connecticut, Maine, Massachusetts,
New Hampshire, New Jersey, New York, Rhode Island, and Vermont and parts
of Maryland, Ohio, and Pennsylvania.

dStates in this survey region included Illinois, Indiana, and Iowa and
parts of Kentucky, Minnesota, Missouri, Nebraska, Ohio, and South Dakota.

eStates in this survey region included West Virginia and parts of Alabama,
Arkansas, Georgia, Kentucky, Missouri, North Carolina, Ohio, Oklahoma,
Pennsylvania, Tennessee, and Virginia.

The USDA study also reported that while milk is produced in all 50 states,
the top 5 milk-producing states in 2000-California, Wisconsin, New York,
Pennsylvania, and Minnesota-accounted for 53 percent of total milk
produced. Growth in the importance of western regions as major sources of
milk over the past 25 years is a significant feature of the United States
dairy industry. For example, in 1975, midwestern states such as Iowa,
Ohio, and Missouri were prominent among the top 10 dairy producing states.
By 2000, production in Idaho, New Mexico, and Washington State surpassed

7The USDA study defined low-cost producers as the 25 percent of milk
producers with costs of $12.17 or less per hundredweight of milk sold.
High-cost producers were defined as the 25 percent of milk producers with
costs of $18.79 or more per hundredweight of milk sold.

Appendix V
Factors That Influence the Price of Milk as It
Moves from the Farm to the Consumer

production in these traditional dairy states. Milk production is
consolidating so that farms with larger numbers of cows account for a
growing share of production. In 1993, farms with 100 or more cows
accounted for 14 percent of all U.S. dairy farms, over half of all cows,
and over 55 percent of the milk produced. As of 2000, these numbers had
grown to 20 percent of all dairy farms, 66 percent of all cows, and more
than 70 percent of milk produced. Significantly, farms with 500 or more
cows accounted for 3 percent of all dairy farms, but 35 percent of cows
and 31 percent of milk produced in 2000.

The production of milk has unique characteristics that distinguish it from
other agricultural products and that cause relatively small changes in
supply or demand to result in relatively large changes in price,
particularly at the farm level. Farmers employ specialized assets or
equipment to produce milk, and they have limited ability to use their
farms, cows, and equipment for other purposes during periods of low
prices. This limited flexibility creates a relatively inelastic supply of
milk with respect to price.8

Demand for raw milk is mainly derived from consumer demand for fluid milk
and manufactured milk products.9 Consumer demand for different fluid or
manufactured milk products affects the price of raw milk used for other
products because increased consumer demand for one particular
product-causing more of that product to be produced--reduces the supply of
raw milk available for other products, thus increasing the price that
manufacturers of other products must pay to acquire raw milk. Over the
long term, per capita demand for fluid milk products has been steadily
declining, in large part because consumers have substituted carbonated
soft drinks and other beverages for fluid milk. On the other hand,
consumer demand for milk products has varied based on dietary
considerations, such as the rising popularity of low-carbohydrate diets
and changes in food consumption patterns, such as an increase in the
amount of food consumed away from the home. Despite some evidence that
consumer demand for

8Price elasticity of supply refers to the percentage change in the
quantity supplied relative to a percentage change in the price received
for a product, in this case raw milk. Supply is price elastic when a
proportional change in the quantity supplied of a product exceeds the
proportional change in its price; supply is price inelastic when a
proportional change in the quantity supplied of a product is less than a
proportional change in its price.

9Demand for raw milk is also influenced by the costs of transporting,
processing, and marketing raw milk to produce dairy products for retail
sale.

Appendix V
Factors That Influence the Price of Milk as It
Moves from the Farm to the Consumer

fluid milk has become more price elastic, it remains relatively price
inelastic compared to the demand for many other products.10

Given the relative inelasticities of milk supply and demand with respect
to price, a number of sources indicated that recent changes in the amount
of raw milk produced, when combined with changes in demand, have affected
farm prices. After late 1999, farm prices began falling in response to the
production surplus that existed at that time. In addition, a number of
dairy experts indicated that following the terrorist attacks of September
11, 2001, prices began to fall even further as people stopped eating out
as much, thereby reducing the demand for manufactured milk products such
as cheese.11 This reduced consumption compounded the long-term decline in
demand for fluid milk products. USDA reported that the combination of
these supply and demand factors was responsible for the low farm prices
that occurred during 2002 and 2003.

More recently, however, supply and demand conditions have changed to
produce record high farm prices in 2004. For example, in response to low
prices during 2002 and 2003, some farmers began to cut back on production
by reducing the sizes of their herds. However, with the relative price
inelasticity of milk supplies, one academic expert noted that it can take
12 to 18 months to achieve a supply response to low prices. During this
time, the identification of a cow infected with mad cow disease in
Alberta, Canada, in May 2003, led to a temporary U.S. ban on imports of
Canadian beef and cattle. This ban included live animals, some of which
would have been used as replacement cows in U.S. dairy herds. While some
beef imports have resumed, USDA has not lifted restrictions on imports of
live cattle.12 Consequently, in June 2004, a report by USDA's Economic
Research Service noted that with relatively few expansions in late 2003
and the tight supplies of replacement cows, few dairy farmers could
increase production

10Price elasticity of demand refers to the percentage change in the
quantity demanded relative to a percentage change in the price of a
product, in this case fluid milk products. Demand is price elastic when a
proportional change in the quantity demanded of a product exceeds the
proportional change in its price; demand is price inelastic when a
proportional change in the quantity demanded of a product is less than a
proportional change in its price.

11According to USDA officials, a decline in the restaurant business began
in mid-2001 and may have been related to other economic factors, such as
falling stock market values and corporate scandals.

12Canadian feedlot operators have filed a class-action lawsuit against the
United States under the North American Free Trade Agreement.

Appendix V
Factors That Influence the Price of Milk as It
Moves from the Farm to the Consumer

in response to rising milk prices, a response that usually limits price
increases.13

Another factor in reducing milk production has been the lower amount of
bST-a hormone used in milk production-available to U.S. farmers.14 USDA
reported that about 2 percent of the U.S. milk supply can be attributed to
the use of bST.15 However, in January 2004, Monsanto, the maker of the
hormone, announced that its customers would receive only half their normal
supply. This reduced availability began March 1 and is expected to
continue through the end of 2004. Additionally, drought conditions in
recent years have led to higher feed costs and have negatively affected
the quality of the feed.16 Finally, some sources identified the National
Milk Producers Federation's Cooperatives Working Together program as
another factor leading to reduced raw milk supplies.17 Since the program
began in July 2003, cooperatives have tried to reduce raw milk supplies by
eliminating some dairy herds, decreasing production, and increasing
exports. USDA estimates indicate that from January through June 2004, milk
production in the top 20 dairy producing states averaged about 1 percent
below production levels during the same period in 2003.

While these factors combined to reduce the available supply of raw milk,
dairy experts indicated that demand for manufactured milk products has
recovered during 2004. In part, they cited a general economic recovery as
contributing to this increased demand. They also indicated that people
have returned to consuming more food away from home, as they did prior

13"Dairy Markets Adjust, But Are Expected to Remain Tight," Livestock,
Dairy, and Poultry Outlook, ERS, LDP-M-120, June 2004.

14USDA's Animal and Plant Health Inspection Service indicated that use of
bovine somatotropin (bST) typically increases average milk yield by 10
pounds per cow per day over the course of the cow's entire lactation, or
as long as bST continues to be used.

15"Meat Markets Roiled by Disease Outbreaks," Livestock, Dairy, and
Poultry Outlook, ERS, LDP-M-116, February 2004.

16Poor quality feed can reduce milk production and may require farmers to
supplement their feed, thereby raising overall feed costs.

17This initiative is funded by an assessment of $0.05 per hundredweight of
milk that participating farmers market. The National Milk Producers
Federation estimates a $0.59 per hundredweight increase in farm income
from the program through September 2004. However, while some dairy experts
indicated that the program has reduced raw milk supplies and thus
benefited farmers' income, they also indicated that this estimate could be
overstated.

                                   Appendix V
                 Factors That Influence the Price of Milk as It
                      Moves from the Farm to the Consumer

to September 11, 2001. This recovery in demand, coming at a time of
reduced milk supplies, pushed farm prices to record high levels in April
and May of 2004. For example, USDA's market-based announced minimum price
for milk to be used in manufactured products, such as cheese, was $19.66
and $20.58 per hundredweight in April and May of 2004, respectively. These
prices compare with $9.41 and $9.71 per hundredweight for April and May of
2003. More recently, these high prices have started to moderate; the
comparable announced minimum price for June 2004 was $17.68. However, USDA
has estimated that average 2004 farm-level prices will be more than $3 per
hundredweight higher than they were in 2003.

Federal and State Dairy Programs and Policies Influence Farm Prices

Federal and State Milk Marketing Orders

A complex system of programs and policies influences the price of raw milk
used to produce fluid milk and manufactured products. USDA's milk
marketing orders, as well as some states' dairy programs, attempt to
stabilize milk marketing conditions by establishing minimum raw milk
prices and other marketing rules, thus, these programs assist individual
farmers and dairy cooperatives, which lack the market power of other
entities such as wholesale milk processors.18 USDA's price support program
attempts to ensure that farm prices do not fall below a minimum level,
and, together with the Milk Income Loss Contract (MILC) program, provides
a safety net for individual farmers during periods of low prices. These
programs and other federal dairy policies operate in a broader context of
trade restrictions, which can limit competition from imported dairy
products and maintain U.S. prices above average international market
prices.

In 2003, the price of about 67 percent of the fluid grade milk marketed by
dairy farmers in the United States was regulated under the FMMO program,
created in 1933 and administered by USDA.19 Under this program, USDA uses
national dairy market price information to set the minimum prices that
must be paid by processors for raw fluid grade milk in specified

18According to USDA, the objectives of FMMOs are to promote orderly
marketing conditions in fluid milk markets, to improve the income
situation of dairy farmers by establishing minimum milk prices, to
supervise the terms of trade in milk markets in such manner as to achieve
more equality of bargaining between farmers and processors, and to assure
consumers adequate supplies of good quality milk at reasonable prices.

19The 1933 Agricultural Adjustment Act, as amended and reenacted in the
Agricultural Marketing Agreement Act of 1937, established the federal
marketing order agreements.

                                   Appendix V
                 Factors That Influence the Price of Milk as It
                      Moves from the Farm to the Consumer

marketing areas, or orders.20 Figure 32 shows a map of the current 10
FMMOs.

Figure 32: Map of Federal Milk Marketing Orders

Source: GAO, based on USDA Agricultural Marketing Service Dairy Programs.

20Dairy farmers sell two grades of milk. Grade A may be used for fluid
consumption or in manufactured products. Grade B may be used only for
manufactured products. Farmers producing Grade A milk must adhere to
higher sanitation requirements than for Grade B milk. Prior to World War
I, much of the milk marketed by farmers did not meet Grade A standards.
However, by 1999, only about 3 percent of all milk marketed in the United
States did not meet Grade A standards.

Appendix V
Factors That Influence the Price of Milk as It
Moves from the Farm to the Consumer

Under the FMMO program, USDA has a classified pricing system for setting
minimum prices for milk on a monthly basis, based upon the intended use of
the milk, as shown in table 27. While there is some variation among the
methods used for setting prices in different orders, in general, FMMO
class prices are determined by formulas with milk component values derived
from wholesale dairy product prices. For example, Class III formulas use
weekly average butter, cheese, and dry whey prices to determine values on
a monthly basis for butterfat, protein, and nonfat solids. The Class IV
formulas use weekly average butter and nonfat dry milk prices to determine
values on a monthly basis for butterfat and nonfat solids, respectively.
The Class II price is determined by adding a fixed amount-a Class II
differential of $0.70 per hundredweight-to the advanced Class IV skim milk
value, while the Class I price is determined by adding a Class I
differential to the higher of the advanced Class III or IV skim milk
values.21 The Class I differentials vary by order. These differentials
were, and to some extent remain, designed to represent the cost of
transporting milk from areas with a surplus-traditionally the Upper
Midwest region-to areas with a deficit, when necessary to meet the demands
for fluid milk products. Because these differentials vary among orders,
Class I prices differ from one marketing order to another.

Table 27: USDA's Milk Classes Used for Setting Milk Prices

Class Usage (examples)

Fluid milk for drinking purposes

Soft manufactured products, such as cream products, cottage cheese, ice
cream, and yogurt

Cream cheese, other spreadable cheeses, and hard cheese

Butter and dried milk products, such as nonfat dry milk

Source: GAO presentation of USDA data.

Dairy farmers selling raw milk to manufactured or fluid milk processors
regulated by an FMMO receive an average, or "blend," price that is the
weighted average of the prices of Class I through IV milk. The weights are
determined by the amount of milk sold in each class in the marketing
order.

21Basing Class I and II prices on the advanced Class III and IV skim milk
values ensures that minimum prices of raw milk used in these products will
be known in the month preceding the month to which they apply.

Appendix V
Factors That Influence the Price of Milk as It
Moves from the Farm to the Consumer

The average price farmers receive, therefore, depends in part on the
extent to which the total raw milk supply in a specific order is used to
make fluid milk, as opposed to the three classes of manufactured products.
Dairy farmers located in a milk marketing order sometimes ship their milk
to another order to obtain a higher price.22 If the farmer meets the
receiving milk marketing order's shipping requirements, all of that
farmer's milk, not only the shipped milk, can qualify for that order's
blend price.23 However, farmers must consider whether the benefit of
receiving a higher blend price outweighs the cost of transporting a
sufficient amount of milk to qualify for the receiving order's blend
price.

To generate the money paid to farmers, processors pay into, or draw from,
a federal order "pool" based on the value of the use for which they are
buying the raw milk.24 Fluid milk processors are required to participate
in the federal order pool if they are covered by one of the federal milk
marketing orders. Processors of manufactured products are not required to
participate in the pool. Under the classified pricing system, raw milk
used in fluid products is valued more highly. Therefore, the fluid milk
processors typically pay money to the pool, while those producing other
products typically draw money from the pool. This draw represents a
benefit to processors of manufactured milk products for serving as a
reserve supply plant for that order's Class I market. In part, a
processor's payment or draw depends on the producer price differential, a
measure of the difference between the value of that processor's use of raw
milk as determined by the market and the value if all of that processor's
raw milk were used in Class III products. In times of significant price
volatility, it is possible for the producer price differential to be
negative, so that some processors of manufactured products would have to
pay into the pool. In such cases,

22Farmers not located in a federal order can also ship their milk into a
federal order. While farmers can ship milk independently, generally it is
dairy cooperatives acting on behalf of their farmer members that assemble
and ship milk to processors.

23Each order has its own requirements, such as the minimum amount of raw
milk required to be shipped to processing plants participating in that
order to qualify for its blend price. Entities regulated by the FMMO
program have the ability to petition USDA for a hearing to amend order
regulations, including the requirements that outside farmers must meet
when shipping their milk to another order.

24The term "pool" is used interchangeably to refer to both the amount of
money generated by applying the minimum federal order class prices to the
amount of milk used in each class within an order and the raw milk
associated with the order for which farmers and processors are able to
share payments from the amount of money generated by the order system.

                                   Appendix V
                 Factors That Influence the Price of Milk as It
                      Moves from the Farm to the Consumer

some of these processors choose not to participate in the pool, or de-pool
their milk, because they would be required to pay into the pool instead of
receiving a draw.

Some states, such as California, Maine, Nevada, New York, Pennsylvania,
and Virginia, have established their own minimum farm-level milk pricing
programs that cover all or portions of their states. These states have
established commissions or boards to perform functions similar to those of
USDA. For example, Virginia's milk commission, created in 1934,
establishes monthly farm prices to ensure dairy farmers an adequate return
on their investment and to preserve market stability. Similarly, Nevada's
dairy commission, established in 1955, sets minimum prices for raw milk
sold to processing facilities located within that state.

Dairy Price Support Program	The dairy price support program, established
in 1949, supports farm prices by providing a standing offer from USDA's
Commodity Credit Corporation (CCC) to purchase butter, cheese, and nonfat
dry milk at specified prices.25 The prices offered for these dairy
products are intended to provide sufficient revenue so that processors of
these products can pay farmers, on average, a legislatively set support
price for raw milk. Since 1999, the support price has been set at $9.90
per hundredweight.26

By offering to purchase as much butter, cheese, and nonfat dry milk as
processors offer to sell at specified prices, the price support program
sets a floor on the price of these commodities and, thus, indirectly on
the raw milk used to produce them. Because processors are not required to
sell to the CCC and milk processing costs vary, farmers may receive prices
that are either above or below the support price. However, manufactured
product prices generally will not fall below the floor for very long.
Also, because the price for raw milk used for fluid purposes under the
FMMO program is based in part on the price of raw milk used for
manufacturing purposes, the price support program indirectly influences
the price that farmers receive for raw milk used for fluid purposes as
well.

25The dairy price support program was established by the Agricultural Act
of 1949, legislation that also expanded the CCC's role as a
government-owned entity to carry out price support activities for a
variety of agricultural commodities.

26The Federal Agriculture Improvement and Reform Act of 1996 required that
the support price be reduced $0.15 per year from $10.35 in 1996 to $9.90
per hundredweight in 1999.

                                   Appendix V
                 Factors That Influence the Price of Milk as It
                      Moves from the Farm to the Consumer

The Secretary of Agriculture can adjust-or tilt-the related CCC purchase
prices for butter and nonfat dry milk and still achieve the target support
price for raw milk used in manufactured products. These products are
considered joint products manufactured from the same 100 pounds of milk.27
Therefore, by increasing the support price of butter while lowering the
support price of nonfat dry milk, or vice versa, USDA is able to adjust
the CCC purchase prices, while maintaining the overall support price. The
ability to adjust the relative purchase prices of these products is
important for correcting imbalances in the CCC's purchases of milkfat
(butter) and nonfat solids (nonfat dry milk). Failure to correct for such
imbalances can create an incentive for farmers to expand production and
may alter the flow of milk to alternative uses. The 1990 Farm Bill
authorized the Secretary of Agriculture to adjust the tilt twice annually
to limit the accumulation of significant government stocks of certain
commodities.28

As market prices rise, the support program allows the CCC to release its
commodity stocks if the market price for a particular commodity exceeds
that commodity's purchase price. In this respect, the program helps to
decrease volatility in milk prices with regard to high-price periods as
well as low-price periods.

Milk Income Loss Contract In 2002, the MILC program began to provide
countercyclical payments

Program	directly to farmers during periods of low prices.29 The MILC
program provides support to farmers when the price of Class I milk in
Boston falls

27In fact, most of the butter manufactured in the United States is
produced independent of nonfat dry milk; however, the assumption correctly
implies that the relative price relationship between butter and nonfat dry
milk must conform to relative yields from raw milk.

28Food, Agriculture, Conservation and Trade Act of 1990, Pub. L. No.
101-624 (codified at 7 U.S.C. S:7251(d)). Butter/nonfat dry milk tilts
were common in the early 1990s-several tilts were made between April 1990
and July 1993-to account for market conditions in which butter was in
surplus relative to nonfat dry milk.

29The MILC program, also known as the National Dairy Market Loss
Assistance Payment program, was authorized in the Farm Security and Rural
Investment Act of 2002. USDA began accepting applications for the MILC
program in August 2002 (payments were made retroactively to December 1,
2001) and will continue to do so until September 30, 2005.

                                   Appendix V
                 Factors That Influence the Price of Milk as It
                      Moves from the Farm to the Consumer

below $16.94.30 MILC payments are equal to 45 percent of the difference
between $16.94 and the lower Boston Class I price. Farmers in all regions
of the country have access to payments under this program, but only 2.4
million pounds of milk per farm are eligible for payments during each
federal fiscal year. Farmers may choose the month that they begin
accepting their payments.31 This discretion may enable farmers producing
more than 2.4 million pounds of milk per year to target their MILC
payments during the lowest-price periods of the year to maximize the MILC
payments they receive before reaching the cap on eligible production.

Trade Restrictions and Export According to some government and academic
experts, trade restrictions

Incentives for Dairy Products	have the greatest effect of any federal
policy on farm milk prices. Trade restrictions maintain U.S. prices above
average international market prices by restricting the amount of imports,
particularly of manufactured dairy products, that enter the country.32 In
other countries, costs of production may be lower, or exports may be more
heavily subsidized, possibly allowing these countries to export products
to the United States at competitive prices. Thus, without trade
restrictions, manufactured products from these other countries might enter
the United States in greater quantities. This increased supply of
manufactured products would be expected to decrease the demand for
domestic raw milk and lead to lower farm prices. Without these trade
restrictions, other dairy programs, such as the price support program,
might not be feasible because lower manufactured product prices resulting
from international competition

30This price is the same price that was in place when the Northeast
Interstate Dairy Compact was effective. The compact included the six New
England states-Connecticut, Maine, Massachusetts, New Hampshire, Rhode
Island, and Vermont. Legislative authority for the compact expired at the
end of September 2001. MILC serves as an alternative to regional dairy
compacts and ad hoc emergency payments to farmers.

31However, this discretion is limited. USDA regulations prohibit a farmer
from selecting a month to receive payments if the month has already begun,
if the month has already passed, or during which no milk was produced. A
farmer also cannot change a previously selected start month after the 15th
of the month before the month selected. Furthermore, once monthly payments
begin, the farmer has no discretion in determining which month or months
to receive payments.

32Between January 1998 and June 2004, U.S. prices remained above average
international market prices reported by USDA's Foreign Agricultural
Service for four dairy commodities whose prices are tracked
internationally: butter, cheese, nonfat dry milk, and whole milk powder.
During this time, U.S. prices averaged between 103 to 125 percent higher
for butter, 44 to 58 percent higher for cheese, 24 to 35 percent higher
for nonfat dry milk, and 62 to 74 percent higher for whole milk powder.

Appendix V
Factors That Influence the Price of Milk as It
Moves from the Farm to the Consumer

could trigger an increase in purchases by the CCC, which could render the
program prohibitively expensive.

The primary U.S. international trade restriction is the tariff-rate quota,
which is the primary international trade restriction allowed under current
international agreements. USDA's Foreign Agricultural Service uses
licensing to administer a tariff-rate quota system for most dairy
products. Under tariff-rate quotas, a low tariff rate applies to imports
up to a specified quantity, and a higher tariff rate applies to any
imports exceeding that amount. These higher over-quota tariff rates
generally limit trade to within quota levels. Quota rates and quantities
vary by product.

Another aspect of U.S. trade policy that affects farm prices is the Dairy
Export Incentive Program (DEIP), an initiative that aims to help exporters
of certain U.S. dairy products-specifically, nonfat dry milk, butterfat,
and various cheeses-meet prevailing world prices for targeted dairy
products and destinations.33 A major objective of the program is to
develop export markets for dairy products where U.S. products are not
currently competitive. Under the program, the Foreign Agricultural Service
pays cash to exporters as bonuses, allowing them to buy dairy products at
U.S. prices and then sell them abroad at lower international prices. DEIP
could affect farm prices primarily by increasing demand for dairy products
through export subsidies.34 According to a 2002 report by the
Congressional Research Service, past studies have indicated that DEIP
subsidies have at times enhanced farm prices; for example, these studies
indicated that DEIP subsidies enhanced farm prices by $0.30 to $0.50 per
hundredweight in 1992.35 Additionally, in May 2003, the National Milk
Producers Federation

33The Dairy Export Incentive Program was introduced in the Food Security
Act of 1985, Pub.

L. No. 99-198, Title I, S:153 (1985) (codified at 15 U.S.C. S:713a-14) and
has been reauthorized since then in successive farm bills. Most recently,
the Farm Security and Rural Investment Act of 2002 extended the program to
2007.

34As is the case with other export subsidy programs, the degree to which
demand is increased under a program such as DEIP depends on the degree to
which the exports under the program are additional to those that would
have occurred in the absence of the program.

35However, the Congressional Research Service study also noted that while
DEIP may help to raise farm prices when milk markets are relatively
balanced, the program would not be likely to have such impacts when
supplies are relatively tight because the industry would not be willing to
give up products for export that were needed to supply the domestic
market. Geoffrey S. Becker, "Agricultural Export Programs: The Dairy
Export Incentive Program (DEIP)," Congressional Research Service Report
for Congress, September 18, 2002.

                                   Appendix V
                 Factors That Influence the Price of Milk as It
                      Moves from the Farm to the Consumer

testified that the subsidies for 5,000 metric tons of butterfat provided
by DEIP in March 2003 increased wholesale butter prices by an estimated
$0.06 per pound.36 This price increase boosted farm income by between $20
million and $30 million. DEIP can also help lower government costs by
reducing the amount of product purchased under the price support program
to the extent that savings in the price support program exceed the costs
of subsidies. Given recent market conditions, DEIP has primarily been used
to encourage exports of nonfat dry milk and cheese,37 and for the most
part, from 1998 through 2002 the program supported exports of these
products to the maximum extent allowable under international trade
commitment limits.38

Services Provided by Cooperatives, Market Structure, and Collective Action
Can Influence the Price of Milk at the Cooperative Level

Milk reaches the consumer through a variety of pathways; however, most
milk produced by dairy farmers in the United States is marketed through
dairy cooperatives.39 Dairy cooperatives can either sell, or arrange the
sale of, raw milk purchased from farmers to wholesale milk processors, or
they can process it into fluid and manufactured milk products and
distribute them to retail outlets.40 As part of sales to wholesale milk
processors, cooperatives negotiate with the processors for over-order
premiums, which represent the difference between the prices charged to the
wholesalers and regulated minimum prices, in areas with federal or state
marketing orders or regulations. The difference between the price at which
cooperatives sell raw milk to wholesale fluid milk processors and the farm
price for fluid

36This testimony was for the U.S. House of Representatives, Committee on
Agriculture, Subcommittee on Department Operations, Oversight, Nutrition,
and Forestry.

37DEIP assisted exports of butterfat have varied depending on market
conditions.

38In some years, such as 1998 and 1999, actual shipments appeared to
exceed the World Trade Organization (WTO) maximum. This excess occured
because of provisions that enabled unused quantities from previous years
to be rolled over into future years. Limits established by U.S. trade
agreements are explained in greater detail in appendix VII.

39Dairy farmers who produce the raw milk used in fluid and manufactured
products can (1) market it through dairy cooperatives, (2) sell it
directly to wholesale milk processors, or (3) process it for direct sale
to consumers.

40See GAO, Dairy Industry: Information on Prices for Fluid Milk and the
Factors That Influence Them, GAO/RCED-99-4 (Washington, D.C.: Oct. 8,
1998) and GAO, Dairy Industry: Information on Marketing Channels and
Prices for Fluid Milk, GAO/RCED-9870 (Washington, D.C.: Mar. 16, 1998) for
more information on the role of cooperatives and other entities in the
marketing of fluid milk.

                                   Appendix V
                 Factors That Influence the Price of Milk as It
                      Moves from the Farm to the Consumer

milk is influenced by the costs of services that cooperatives provide to
their members and to their buyers, the relative market power of
cooperatives and fluid milk processors, and the effects of collective
action taken by dairy cooperatives in marketing their members' milk.

Costs of Services Provided by Cooperatives Influence Milk Prices

Over-order premiums, in part, compensate cooperatives for the services
they provide to their members and on behalf of their members to
wholesalers. Some distinctive features of cooperatives include member
ownership and control, at-cost services for members, and distribution of
income to members on the basis of patronage. Farmers join dairy
cooperatives to guarantee a market outlet for their milk, to gain
bargaining power to obtain the best price in the market, to have their
milk marketed efficiently, with the assurance that their milk will be
accurately weighed and tested, and to be effectively represented in
legislative, regulatory, and public relations matters. Most dairy
cooperatives require farmers to sign a 1-year membership agreement that
commits them to market all their milk through the cooperative.41

Cooperatives operate like corporate businesses to perform services for
their members. For example, Dairy Farmers of America, the largest dairy
cooperative in the country, serves almost 23,000 members, producing about
21 percent and marketing about 33 percent of the milk in the United
States. According to the cooperative's Web site, the cooperative provides
a variety of services to its members, including the following:

o 	insurance-medical programs, dental/vision plans, and life insurance
available to members via a milk check deduction;

o 	direct deposit-direct deposit of members' milk checks, ensuring that
farmers' pay checks will be available within 24 hours of the pay date;

o 	forward contracting-a marketing service that allows members to protect
themselves against price volatility by locking in the future sale price of
their milk several months before it is produced; and

o 	financing services-loan packages for cattle, equipment, and operating
expenses.

41Typically, these agreements are self-renewing.

                                   Appendix V
                 Factors That Influence the Price of Milk as It
                      Moves from the Farm to the Consumer

In some cases, dairy farmers pay on a per-use basis for the services they
receive. However, cooperatives may also try to offset the costs of their
services through negotiations with wholesale milk processors for overorder
premiums.

Over-order premiums also compensate dairy cooperatives for a number of
services that they provide to fluid milk processors on behalf of their
members. Generally, these services include (1) transporting milk from
different milk-producing areas, (2) scheduling-or balancing-milk
deliveries to coincide with demand, and (3) standardizing the component
content of milk deliveries. Different cooperatives also provide additional
services for fluid milk processors. For example, one cooperative we
contacted noted the rigorous quality control procedures it performs on its
members' milk. According to the cooperative official, these efforts allow
the cooperative to market its members' milk as better quality, potentially
helping the cooperative negotiate higher over-order premiums. Officials
from another cooperative said that a major component of the costs of
services provided by cooperatives is balancing the delivery of raw milk
supplies to processors' plants. At certain times processors' plants have
surging demand for raw milk, while at other times the plants are empty. In
addition, supply disruptions, such as labor strikes, create significant
balancing disruptions. In this environment, few, if any, fluid milk
processing firms have the capital (plants to make cheese and other
products during periods of low fluid demand) to assume the risks inherent
in balancing, and so in most cases this responsibility is met by the
cooperatives.

Relative Market Power of Cooperatives and Processors Influences Milk
Prices

Historically, farmers produced and distributed fluid milk as well as some
manufactured products. Milk is a highly perishable product that is bulky
to transport. Traditionally, this left farmers dependent on local markets
for the sale of their milk. The role of dairy cooperatives developed as
farmers faced greater demand for fluid milk and dairy products and the
number of farmers who processed and distributed their own milk products
declined. Instead, specialized firms began taking on the role of
processing fluid and manufactured milk products and marketing them for
sale to consumers. However, in this environment, there were many more
farmers than processors, so processors had the opportunity to bargain with
different farmers to obtain a lower price for their raw milk supplies. In
this situation, farmers were at a disadvantage. Consequently, cooperatives
took on the role of collecting raw milk from farmers and distributing it
to processors. By doing so, cooperatives helped to balance the bargaining
power between

Appendix V
Factors That Influence the Price of Milk as It
Moves from the Farm to the Consumer

farmers and processors. The 1922 Capper-Volstead Act provides limited
antitrust immunity to cooperatives that meet certain requirements under
certain conditions and gives farmers an opportunity to work together in
setting raw milk prices, including bargaining for market premiums. Thus,
over-order premiums, in part, reflect market power acquired by
cooperatives relative to processors.

Since our June 2001 report, the concentration of dairy cooperatives has
increased, with the potential effect of enhancing their market power in
negotiations with processors. In 2001, we reported that 83 percent of the
milk produced in the United States was marketed by cooperatives. However,
USDA recently reported that in 2002, the share of milk sold to processors
and other distributors by cooperatives reached 86 percent of all the milk
produced in the United States.42 Cooperatives attained this market share
despite a 13 percent decrease in the number of dairy cooperatives between
1997 and 2002. During this time the amount of member-produced milk
marketed by the eight largest dairy cooperatives grew from 52 to 63
percent of the total volume of milk marketed by cooperatives. This
translated into an increase from 42 to 52 percent of the total volume of
milk produced in the United States.

A number of dairy experts cited the need to offset gains in market power
made by increasingly concentrated firms at the wholesale processor and
retail levels of the milk marketing chain as a key factor in the continued
concentration of cooperatives. The greater the percentage of the milk
supply that a cooperative markets, the greater its ability might be to
obtain higher over-order premiums in negotiations with wholesale
processors. On the other hand, one academic source questioned the extent
to which increased concentration is enhancing the market power of dairy
cooperatives, particularly over the long term. He noted that although
Dairy Farmers of America has been consolidating its control over milk
supplies in some regions, farmers and cooperatives have been able to
command larger over-order premiums in the East and Upper Midwest
regions-where the cooperative's presence is not as strong-than in the
West, where milk supplies have been increasing. Other sources noted that
competition still exists among cooperatives and independent dairy farmers
and that this competition prevents even larger cooperatives from obtaining
excessively high over-order premiums.

42K. Charles Ling, Marketing Operations of Dairy Cooperatives, 2002,
USDA/Rural Business-Cooperative Service, Research Report 201, February
2004.

                                   Appendix V
                 Factors That Influence the Price of Milk as It
                      Moves from the Farm to the Consumer

Collective Action by Cooperatives Influences Milk Prices

Another factor in determining the over-order premiums received by
cooperatives for raw milk is collective action taken by cooperatives.
Cooperatives work together to try to set prices by coordination allowed
under the protection afforded by the Capper-Volstead Act. For example,
officials with Dairy Farmers of America said that a major factor in the
price of milk at the cooperative level is the action of marketing agencies
composed of cooperatives. Marketing agencies behave like cartels and
announce prices for their cooperative members. In most cases these
agencies set prices for raw milk used in fluid milk and other products.
Most of the prices announced by the marketing agencies represent the
minimum federal order prices; additional charges may be added representing
the costs of services provided by the cooperatives to the processors.

Representatives of Dairy Farmers of America said that there are marketing
agency agreements in most major markets except in the Pacific Northwest
and that for the most part, cooperatives participate in marketing
agencies. They further stated that the use of marketing agencies has
become more common in recent years. The marketing agencies may also market
milk for independent farmers. The officials noted that while cooperatives
and independent farmers can choose not to participate in the marketing
agencies, experience has shown that as more producers choose to market
milk outside the system, the marketing agencies face significant
competition and prices fall. Eventually, if the prices get low enough, the
producers have an incentive to work together again.

In an alternative type of collective action, three cooperatives-the
Dairylea Cooperative, Dairy Farmers of America, and St. Albans Cooperative
Creamery-established a milk marketing organization called Dairy Marketing
Services. According to a Dairy Marketing Services official, the
organization was formed because the cooperatives realized that they needed
more market power to compete with increasingly concentrated processors and
retailers. Cooperatives such as Dairylea, or individual farmers, establish
contracts with Dairy Marketing Services to market their milk. Dairy
Marketing Services markets about 16 billion pounds of milk annually for
farmers in the Northeast area that extends from Maine to Maryland, and
includes a small area in Ohio. The official estimated that this quantity
represents about 45 percent of the milk marketed in the Northeast and is
produced by some 10,000 to 11,000 farmers.

The Dairy Marketing Services official stated that the organization has
been able to carve a niche for itself in the milk marketing chain by
convincing processors that it is more efficient for them to have Dairy
Marketing

                                   Appendix V
                 Factors That Influence the Price of Milk as It
                      Moves from the Farm to the Consumer

Services arrange to have raw milk transported from the farm to the plant
and allow the processors to focus on processing milk. As a result, Dairy
Marketing Services has been able to obtain contracts from a number of
major processors in the Northeast, including Dean Foods, Crowley Foods,
and Kraft, to ensure an adequate supply of milk for their plants.
Additionally, Dairy Marketing Services provides specialized services for
farmers such as health insurance and workmen's compensation, a livestock
purchasing service, and risk management operations for farmers engaged in
forward contracting. Although we were unable to confirm the effects that
Dairy Marketing Services' efforts have had, the official stated that the
organization has provided higher over-order premiums and lower
transportation charges for its participating cooperatives and farmers than
would have otherwise been the case.

Input Costs, Service Levels, Innovations, and Market Structure Influence
Wholesale Fluid Milk Prices

The difference between the price at which wholesale fluid milk processors
sell fluid milk products to retail firms and the price they pay for raw
milk is influenced by changes in input costs, such as fuel, labor,
packaging, transportation, and capital expenses. These costs, in turn, are
affected by recent innovations that have increased efficiency and lowered
costs of fluid milk processing, as well as by the level of service that
fluid milk processors provide to retailers. For example, in addition to
shipping the products to retailers, some wholesalers provide in-store
services, including unloading the milk on the store dock, restocking the
dairy case, and removing outdated or leaking containers. The difference
between what fluid milk processors pay for raw milk and the wholesale
price they charge their retail customers is also influenced by continued
structural change in the fluid milk processing industry, including a
steady increase in firm consolidation and the market share of some firms.
While there have been many reasons for these trends, the effects on the
market and fluid milk prices at this level are unclear.

Input Costs, Service Levels, Several fluid milk processors stated that the
cost of raw milk, and, in

and Innovations Influence particular, the federal order minimum price, was
the single most important

Wholesale Fluid Milk Prices	influence on wholesale milk prices. We
estimate that the price of raw milk ranges from about 60 to 70 percent of
the wholesale price of 2 percent milk. As such, the wholesale price that
processors charge would be directly linked to the Class I federal order
price on a year-to-year basis, with adjustments for over-order premiums
and other inputs. However, a variety of other input costs can also affect
the price at which fluid milk processors

Appendix V
Factors That Influence the Price of Milk as It
Moves from the Farm to the Consumer

sell fluid milk products to retailers. Some sources indicated that costs
of inputs other than raw milk have been increasing in recent years.43 As
one executive of a milk-processing firm explained, the primary input
costs, apart from raw milk, include labor and energy. A 2002 study
examining changes in fluid milk processing plants located in the state of
Maine found that total processing costs rose at an annual rate of about
2.4 percent (adjusted for inflation) from 1993 through 2000.44 The study
indicated that economywide wage inflation plus a dramatic increase in
health care premiums paid by employers drove labor costs above the costs
of other inputs, such as land and building expenses and plant supplies.
Equipment costs increased 10.9 percent per year with investments in plant
automation and greater reliance on information technologies. Also, fuel
costs increased by 4.6 percent per year, reflecting economywide trends in
energy costs. Moreover, while the cost of operating capital constituted
only 1.0 percent of processing costs, it increased substantially during
this period due to an increase in the short-term lending rate. Table 28
displays the percentage change in fluid milk processing costs in Maine
reported in this study for each cost category from 1993 through 2000.

43We could not identify any recent studies that have examined milk
processing and distribution costs on a national level.

44T.J. Dalton, G.K. Criner, and J. Halloran, "Fluid Milk Processing Costs:
Current State and Comparisons," Journal of Dairy Science, Vol. 85, No. 4
(2002): 984-991. The study used costengineering methods to look at four
state-of-the-art fluid milk processing plants in Maine ranging from
335,000 gallons per week to 600,000 gallons per week.

Appendix V
Factors That Influence the Price of Milk as It
Moves from the Farm to the Consumer

Table 28: Change in Costs for Fluid Milk Processing Plants in Maine, 1993
through 2000

                                                            Total change from 
                            Percent of total Annual rate of      1993 through 
             Cost category  cost in 2000 (%)   change (%)            2000 (%) 
               Electricity               5.0            1.6              12.0 
                 Equipment              20.0           10.9             106.0 
                  Fuel oil               1.0            4.6              37.0 
                     Labor              31.0            3.7              29.0 
         Land and building              13.0            3.8              30.0 
         Operating capital               1.0           15.5             175.0 
              Product loss               1.0            0.4 
                  Supplies              27.0          (3.0)            (19.0) 
           Water and sewer               1.0          (0.3)             (2.0) 
                     Total               100            2.4              18.0 

Source: GAO analysis of data from Dalton et al. (2002).

Note: Percentages calculated based on 2000 dollars. Percent of total cost
in 2000 was calculated by GAO and is rounded to the nearest percent.
Annual rates of change and total change from 1993 through 2000 are from
Dalton et al. (2002).

Changes in the level of service that some fluid milk processors provide
their retail customers have increased the efficiency of the dairy supply
chain, thus potentially influencing wholesale milk prices. For example,
some fluid milk processors have begun to undertake supply-chain management
for their retail customers.45 According to a number of retailers and
processors, supply-chain management commonly involves shared computer
systems, which, in the vertical marketing chain, allow processors to more
efficiently manage the processing and transporting of fluid milk products.
One processor indicated that it uses an electronic data transfer system to
manage supplies for certain retailers. In particular, this system allows
the processor to contract for a certain number of loads of milk per day.
Further, according to a recent presentation given by company officials,
Dean Foods' national, refrigerated, direct-store-delivery system allows it
to deliver fluid milk to its customers with increased route network
efficiency

45Supply-chain management is an attempt to coordinate processes involving
producing, shipping, and distributing products, generally performed only
by large corporations with large suppliers.

Appendix V
Factors That Influence the Price of Milk as It
Moves from the Farm to the Consumer

and without customer disruption.46 Dean Foods operates 129 fluid
processing plants in 39 states, servicing more than 150,000 customers
coast to coast via its direct-store-delivery system of more than 6,000
routes.47

By allowing fluid milk products to move more efficiently from the
processor to the retailer, these kinds of services help to ensure quality
and reduce waste and costs along the supply chain. To the extent that
processors benefit from the reduced costs of supplying retailers with
fluid milk products, the provision of these services could have a downward
effect on wholesale prices. On the other hand, these services could
provide value to retailers for which they might be willing to pay a higher
price when acquiring fluid milk products. Therefore, the net effect on
wholesale prices of the level of service that processors provide to
retailers is uncertain.

Additionally, innovations in technology can affect prices at the wholesale
milk processing level. For example, changes in processing technology, such
as more automated equipment, can improve the efficiency of processing
operations and, to the extent that processing firms are successful at
reducing their costs through innovative practices, they may be able to
reduce their prices. A representative of one fluid milk processor
explained that improvements in processing and packaging technology have
doubled and tripled output. Also, a representative of one firm that
processes milk for sale in its own retail stores stated that the firm has
dedicated a large contingent of people toward the goal of reducing milk
losses at its processing plants and has been successful at cutting these
losses in half. He noted that a driving force behind these efforts is to
try to alleviate increases in other input costs, such as labor.

With innovations in technology, the fluid milk processing industry has
also invested in innovative new products. By developing products with
extended shelf lives, processors can potentially save shipping costs,
leading to lower wholesale prices. For instance, the dairy processing
industry's collective investment in extended shelf life, ultra high
temperature, and aseptic packaging technology allows fluid milk products

46Dean Foods Company, Presentation at the Consumer Analyst Group of New
York Conference, Scottsdale, Arizona, February 17, 2004. Available online
at http://www.deanfoods.com/investors.

47Dean Foods Annual Report (2001), 18.

                                   Appendix V
                 Factors That Influence the Price of Milk as It
                      Moves from the Farm to the Consumer

to reach the end user more efficiently while maintaining quality.48 The
benefits to processors and their retailers include the ability to ship
these products longer distances because they are able to endure more
stress than traditionally processed milk.

Market Structure in the Fluid Milk Processing Industry Can Influence
Wholesale Prices

Since the 1960s, there has been long-term structural change in the
wholesale fluid milk processing industry as a continuously declining
number of firms have processed an increasing average volume of milk.49
Structural change in the processing industry has been driven by economies
of size, technological changes, high concentration at other levels of the
milk marketing chain, and rapid consolidation into fewer and fewer
firms.50 While structural change can lead to lower prices due to cost
reduction from greater efficiency in production, it can also lead to
higher market concentration, particularly in individual markets. In
general, high and increasing market concentration can result in greater
market power, potentially allowing firms to increase prices above
competitive levels. Accordingly, the net impact of increased market
concentration on wholesale prices can be either positive or negative.

In recent years, through aggressive acquisitions of independent dairy
processing plants, a handful of fluid milk processing firms have changed
the market structure of the dairy industry at the wholesale level. These
companies have generally pursued the business strategy of acquiring strong
regional dairy processing plants so that they can strengthen their
presence in existing markets, while expanding their geographic coverage to
a national level. The acquisition and consolidation trend at the wholesale
level has affected market structure by leading to higher market

48Aseptic packaging, commonly called "drink boxes," is the result of a
beverage and liquid food system that allows perishable food products to be
distributed and stored without refrigeration for periods up to six months
or more. It is used to preserve and package everything from milk, juice,
and drinks of all kinds to scrambled egg mix, tomato sauce, soups, and
other liquid foods.

49Structural change is characterized by broad and long-term changes in key
industry characteristics, frequently including consolidation and changes
in concentration, methods of vertical coordination, and the mix of
products and services offered by firms in an industry.

50Don P. Blayney and James J. Miller, "Concentration and Structural Change
in Dairy Processing and Manufacturing," Paper presented at the 10th Annual
Workshop for Dairy Economists and Policy Analysts, Memphis, Tennessee,
April 2003.

Appendix V
Factors That Influence the Price of Milk as It
Moves from the Farm to the Consumer

concentration for fluid milk processors in some markets. One common
measure of market concentration is the four-firm concentration ratio-the
percentage of sales by the top four firms in a market. According to the
1997 Census of Manufacturers, the market share for the top four fluid milk
processors in the nation was about 21 percent. However, the market share
for top fluid milk processors at the local level was significantly higher.
For example, in our June 2001 report, we found that in Boston,
Massachusetts, the market share of the top four fluid milk processors
increased from 66 percent in December 1997 to 88 percent in December
1999.51

Since our last report in 2001 on fluid milk prices, this trend has
continued, and there have been several significant mergers, acquisitions,
and joint ventures that have further consolidated the industry. For
example, in late 2001, Dean Foods merged with Suiza, Inc., bringing
together the number one and two firms in terms of market share in the
processing industry. Then, in July 2002, the Land O' Lakes dairy
cooperative sold its fluid milk operations to Dean Foods. We estimate that
these acquisitions and mergers gave Dean Foods about a 27 percent market
share nationally in fluid milk products in 2002. Others have estimated
that Dean Foods' market share is about 35 percent nationally and
approximately 70 percent in New England.52 As of 2002, we estimate that
the market share of the top four fluid milk processors has increased to
approximately 47 percent.53 As seen in figure 33, with increased
concentration, the number of fluid milk processing plants has gone from
1,066 plants producing an average of 50.1 million pounds of milk per year
in 1980 to 385 plants producing an average of 154.2 million pounds per
year in 2002.

51See GAO-01-561, appendix VI, for more information on consolidation at
the wholesale milk processor level.

52As a result of a divestiture plan stemming from the Dean-Suiza merger,
the second largest fluid milk processor, the National Dairy Holdings
Group, was created.

53We estimated the four-firm concentration ratio using expert opinion from
conversations with dairy experts at USDA on the leading four fluid milk
processing firms, estimates of company sales from the August 2003 Dairy
Foods Magazine, "The Dairy 100: The List," and the total value of
shipments from the 2002 Economic Census, Industry Series (September 2004)
for fluid milk manufacturing. We also adjusted company sales listed in
Dairy Foods, per company annual statements and expert opinion, for the
percent of sales that would account for products such as ice cream and
cheese. To make our estimate consistent with the 1997 Economic Census
four-firm concentration ratio, we also included the same categories for
fluid milk manufacturing, such as cottage cheese and yogurt, along with
fluid milk processing, that went into this calculation. This figure does
not include the sales that firms may have from various joint ventures.

                                   Appendix V
                 Factors That Influence the Price of Milk as It
                      Moves from the Farm to the Consumer

Figure 33: Number of Fluid Milk Processing Plants and Volume of Milk
Processed per Plant, 1980 through 2002 Average volume (million lbs.)

Number of plants 1,200

160 1,000

140

120 800

100 600

80

60 400 40 20

                                      200

                                       00

1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994
1995 1996 1997 1998 1999 2000

                                   2001 2002

Year

Average volume (in millions of pounds) Number of plants

Source: GAO presentation of USDA data.

Such increased concentration of fluid milk processing firms, particularly
in individual markets, can increase the price at which fluid milk is sold
to retailers because market concentration can provide these firms greater
market power. Thus, some analysts viewed the trend toward greater
concentration in the wholesale market as a means toward greater dominance
and market power in selling fluid milk to retailers. Further, they noted
that increased market power can also benefit processors in their
negotiations for raw milk supplies from cooperatives and independent
farmers. For example, the exercise of market power could allow processors
to negotiate more favorable supply contracts, which could drive down input
prices and increase the spread between wholesale and retail prices. Other
economists who study the causes of market concentration described a
phenomenon called the "replication hypothesis"-as concentration grows at
one marketing level, it is likely to be replicated at other marketing
levels. For instance, high market concentration at the retail level can
lead to greater concentration at the fluid milk processor level, and
higher concentration in fluid milk processing can, in turn, lead to higher
concentration at the cooperative level. One fluid milk processor that

                                   Appendix V
                 Factors That Influence the Price of Milk as It
                      Moves from the Farm to the Consumer

we spoke with stated that retail concentration has resulted in retailers
preferring only one supplier, requiring a processor to have multiple
plants in order to supply a retailer who serves many markets.

On the other hand, increasing concentration can lead to cost savings
through efficiency gains, which may be passed on to retailers in the form
of lower wholesale prices. For example, some economists viewed
consolidation of processing firms as a result of increasing economies of
scale and excess plant capacity. That is, processors decrease their costs
per gallon for items like packaging or processing costs, as they increase
the amount of milk they process.54 One dairy analyst reported that in
plants ranging from a monthly volume of 90,000 pounds per month to 30
million pounds per month processing costs decreased from about $1 per
gallon to about $0.50 per gallon. In the end, the impact of market
concentration on wholesale prices, either positive or negative, depends on
whether market power or efficiency dominates.55

Retailing Costs, Consumer Demand, and Market Structure Changes Affect
Retail Prices for Fluid Milk

Three key factors that influence fluid milk prices at the retail level are
retailing costs, consumer demand, and market structure. Recent increases
in input costs such as labor and energy have been substantial. In an
effort to hold down their retailing costs and remain competitive, some
retailers are implementing supply-chain management and other innovations
that increase efficiency. At the same time, consumers are purchasing a
declining amount of traditional fluid milk and are increasing consumption
of other beverages, such as soft drinks and bottled water. Market
structure changes include continued consolidation in recent years through
mergers and acquisitions among large food retailers at the national level
and in many local markets, along with an increasing number of outlets that
are competing with traditional supermarkets to sell fluid milk.

54See T.J. Dalton, G.K. Criner, and J. Halloran, "Fluid Milk Processing
Costs: Current State and Comparisons," Journal of Dairy Science, Vol. 85,
No. 4, 2002, 984-991. Also, Day 3 of the Hearing on Pacific Northwest and
Arizona/Las Vegas Producer-Handlers, September 25, 2003, including the
exhibit on "Cost Structure of Fluid Milk Plants of Various Sizes."

55One study that we identified, however, that separated market power and
cost-efficiency effects on the prices in 33 food processing industries
from 1972 to 1992, found that the market power effect dominated in fluid
milk manufacturing. Rigoberto A. Lopez, Azzedine M. Azzam, and Carmen
Liron-Espana, "Market Power and/or Efficiency: A Structural Approach"
Review of Industrial Organization, Vol. 20, March 2002: 115-126.

                                   Appendix V
                 Factors That Influence the Price of Milk as It
                      Moves from the Farm to the Consumer

Costs of Retailing Influence the Price of Fluid Milk

Representatives of the Food Marketing Institute stated that after the
wholesale costs of the milk, the primary costs that influence the retail
price of fluid milk are related to labor and energy.56 They added that all
of these costs have been rising recently. According to the Bureau of Labor
Statistics, the average hourly earnings for nonsupervisory food store
employees went from $7.56 per hour in 1992 to $10.20 per hour in 2002.
These payroll costs are the largest percentage of retail operating costs,
followed by the second largest single category, employee benefits such as
health insurance.57 Table 29 shows the breakdown of supermarket operating
costs in 2003 as a percentage of total sales and gross margin.58

Table 29: Supermarket Operating Costs as a Percentage of Sales and Gross
Margin, 2003

                       Expenses Percentage of sales       Percentage of gross 
                                                                       margin 
               Depreciation and                     
                   amortization                 1.3 
              Employee benefits                 3.5 
                      Insurance                 0.3 
        Maintenance and repairs                 0.7 
          Other operating costs                 3.8 
                        Payroll                11.4 
               Property rentals                 1.8 
                       Supplies                 1.1 
             Taxes and licenses                 0.4 
                      Utilities                 1.3 
                 Total expenses                25.7 

Source: GAO analysis of Food Marketing Institute data.

56The Food Marketing Institute is a trade association that represents
segments of the retail food industry such as supermarkets. The Food
Marketing Institute does not represent convenience stores.

57As table 29 shows, the category "other operating costs" is somewhat
larger than "employee benefits," but according to the Food Marketing
Institute, it is a catch-all category including items such as travel,
equipment rentals, communication, and services purchased.

58The gross margin is a firm's sales revenue minus the costs of acquiring
products, expressed as a percentage of sales.

Appendix V
Factors That Influence the Price of Milk as It
Moves from the Farm to the Consumer

Table 30 displays the sales and expense growth as a percentage of sales
for the supermarket industry during the last decade, from 1993 through
2003. During this time, total employment costs increased by 12.0 percent,
including a 10.7 percent increase in payroll expenses; the cost of
supplies also increased by 10.0 percent.

Table 30: Growth of Supermarket Sales and Expenses during the Last Decade,
1993 through 2003

Percent change 1993 2003 1993 through 2003

                    Supermarket sales (dollars in billions)

Sales from supermarkets with more than
$2 million in annual sales $292.0 $432.8 48.2

                    Retailing costs as a percentage of sales

                      Payroll expenses         10.3         11.4         10.7 
                 Total employment cost         13.3         14.9         12.0 
                              Supplies          1.0     1.1              10.0 
                             Utilities          1.2     1.3      

Source: GAO analysis of Food Marketing Institute data.

A 2003 study that was more specific to retailer costs related to fluid
milk sales noted that these costs include both direct and indirect
costs.59 Direct costs are those for electricity, labor, store equipment,
and fluid milk. Indirect costs include corporate, division, and store
overhead. While the study found variation in the indirect costs, such as
store overhead, there was less variation across retail stores in direct
costs.

Increasing per unit costs have led some retailers to try to improve
efficiency and reduce total costs. As mentioned in the discussion of
factors influencing fluid milk prices at the wholesale level, some
retailers are reducing costs by working with their wholesale suppliers to
achieve supply-chain management. For example, officials with Wal-Mart
noted that the firm has tried to reduce its costs and maintain its
everyday-low-pricing strategy for consumers through

59George Criner, "Milk Retail Costs and Margins." Paper presented at the
Northeast Dairy Policy Summit Meeting, November 17-18, 2003, University of
Connecticut, Storrs.

                                   Appendix V
                 Factors That Influence the Price of Milk as It
                      Moves from the Farm to the Consumer

o 	a computerized system called Collaborative Planning Forecast
Replenishment that allows processors to track stock levels at Wal-Mart
locations and schedule deliveries to specific locations;

o 	direct-store-delivery of the majority of its fluid milk products to
increase the efficiency of its supply chain;60 and

o 	changes to its shipping practices, such as not putting labels on its
cases, that have allowed Wal-Mart to save time and money.

Another retailer indicated that it is trying to improve the way it stocks
its shelves to cut costs. A representative indicated that the retailer has
invested in retrofitting its stores to use a device called a "bossy cart,"
which allows store employees to move 80 gallons of milk into the milk case
in one shelf-stocking.

Consumer Demand Influences Retail Fluid Milk Pricing

Consumer demand, driven by factors such as taste, convenience, and health,
influences the retail price of milk. Moreover, since fluid milk represents
approximately 3 percent of total supermarket sales, it is an important
category for store performance, and retailers have an incentive to price
their products competitively. However, over time, fluid milk consumption
has gradually declined, with per capita demand for milk trending downward
at a rate of 2 to 3 percent per year. This downward trend stems from
several key factors including increasing consumption of substitute drinks
such as carbonated soft drinks, juice drinks, coffee, teas, soy products,
and bottled water. Also, there has been an increasing trend toward more
eating outside the home, reducing the demand for fluid milk sold in food
stores. Within the fluid milk category, whole milk has gone from being 92
percent of fluid milk consumed in 1960 to about 35 percent in 2001.
Private labels represent the largest portion of the market, about 60
percent.61 More recently, however, there has been growth in the
development of innovative value-added dairy products. These new
innovations include dairy products for medicine/health (such as low

60For most other products, Wal-Mart officials said that the retailer
maintains a network of approximately 30 food distribution centers.

61Private label refers to products that are "store brands" and have labels
that represent the retailer. These products may be produced by
manufacturers that also produce national brands or by retailers who own
their own manufacturing operations.

Appendix V
Factors That Influence the Price of Milk as It
Moves from the Farm to the Consumer

carbohydrate products), multipack drinks such as single-serve and vending
drinks, and organic dairy products.

In response to trends in consumer demand for fluid milk products,
retailers from high-end supermarkets to mass merchandisers use diverse
pricing strategies, and no single approach applies to any group. However,
according to some retail executives, one method that retailers are
currently using is category management. Using this strategy, a retailer
would not focus on how much 1 percent, 2 percent, or whole milk it sells,
but rather on how much is sold from the entire dairy case.62 Accordingly,
category managers would view product assortment strategically, evaluating
the performance of entire groups of related dairy products. The goal is to
maximize the sales for the entire category, which requires continual
adjustment to match consumer demand. To accomplish this goal, managers may
feed scanner data and other market information into computer models that
make product assortment decisions.

A related issue influencing the retail price of fluid milk is the price
elasticity of demand, that is, the sensitivity of fluid milk consumption
to changes in price. For many years, empirical studies have indicated that
milk prices were very price inelastic, meaning that there was little
change in demand in response to a change in price. Most studies suggest
that overall, the demand for milk is still price inelastic. However, some
recent studies suggest that the demand for milk is not as price inelastic
as it was previously.

Moreover, some researchers have found that for many fluid milk products,
demand is elastic, or that there is a greater change in demand relative to
a change in price for certain types of milk. One study reported that price
elasticities varied considerably by container size, type (such as white or
flavored), and fat content of the milk.63 For instance, the study found
that the demand for whole milk, skim milk, and low fat milk in half-gallon
containers was price elastic. This research also suggested that carbonated

62Different retailers may define the dairy case differently depending on
the market that they serve. The dairy case may include products such as
fluid milk, cottage cheese, creamers, yogurt, butter, and eggs as well as
a variety of new "value added" dairy products. A recent study by Cornell
University on dairy case management found that, on average, stores used
approximately 37 different milk products in their dairy cases.

63Oral Capps, Jr., "Demand and Marketing Analyses for Fluid Milk Products
by Type and by Package Size." Presentation at the 2004 Dairy Forum, Boca
Raton, Florida, January 2004.

                                   Appendix V
                 Factors That Influence the Price of Milk as It
                      Moves from the Farm to the Consumer

soft drinks are the chief substitute or competitor for fluid milk
products, while water is a complement in consumption.64 In another study,
researchers found that the elasticities of demand for skim/low fat and
whole milk brands are different.65 Demand for skim/low fat milk was found
to be more price elastic than demand for whole milk, suggesting that
retailers could increase overall fluid milk sales by lowering the prices
of skim/low fat milk relative to prices for whole milk.

            Retail Market Structure Can Influence Fluid Milk Prices

As with the fluid milk processing industry, there have been trends of
increasing consolidation and concentration at the retail level during
recent years, especially among retail firms in some individual markets.
Structural change and increased consolidation at the retail level of the
milk marketing chain could lead to lower retail prices as individual
retailers experience increased efficiencies in their operations. On the
other hand, high levels of concentration can result in greater market
power, potentially allowing firms to increase market prices above
competitive levels.66 Also, greater market concentration at this level
could increase a retailer's buying power with fluid milk processors,
potentially lowering costs. Depending upon whether these lower costs are
passed on to consumers, this can either lower retail milk prices or
increase the spread between wholesale and retail prices.

According to USDA, since 1996, almost 4,700 supermarkets, representing
$75.5 billion in sales, were acquired by other firms. Major mergers and
acquisitions that have occurred in the retail food market in recent years
include the following:

64If products are substitutes, an increase in the price of one would
increase the demand for the other. A complement in consumption means that
the products are consumed in conjunction with one another. An increase in
price for one product would decrease the demand for the complementary
product.

65Tirtha Dhar and Tom Cox, "Strategic Implications of Retail Pricing in
the U.S. Fluid Milk Market." Paper prepared for presentation at the
American Agricultural Economics Association Annual Meeting, Montreal,
Canada, July 27-30, 2003.

66Economic studies of the relationship between price and market
concentration in food retailing have found mixed results. Carolyn Dimitri,
Abebayehu Tegene, and Phil R. Kaufman, U.S. Fresh Produce Markets:
Marketing Channels, Trade Practices, and Retail Price Behavior,
Agricultural Economic Report Number 825, Economic Research Service, USDA,
September 2003.

Appendix V
Factors That Influence the Price of Milk as It
Moves from the Farm to the Consumer

o 	In 1998, Kroger, the nation's largest supermarket chain, acquired Fred
Meyer, and Albertsons acquired American Stores, the second-largest at that
time.

o 	In 2000, Delhaize America, operator of the Food Lion chain of stores,
purchased Hannaford Brother's Shop 'n Save supermarkets in New England to
become the eighth-largest food retailer at that time.

o 	In 2001, Kroger purchased supermarkets in Oklahoma and Texas from
Winn-Dixie.

o 	In 2001, Safeway made several acquisitions including Genuardi's Family
Market stores (Pennsylvania, New Jersey, and Delaware), Randall's food
markets (Houston, Texas), and Dominick's supermarkets (Chicago
metropolitan area).

o 	In 2004, Albertsons, the third-largest U.S. food retailer, purchased
Shaw's, the eleventh-largest at the time.

In June 2001, we reported that, for the 100 largest U.S. cities, the
combined average market share of the top four firms increased from 69
percent in 1992 to 72 percent in 1998, with some variation depending upon
the particular market area.67 An official of one large supermarket chain
noted that because of Wal-Mart's large presence in the market, other
companies' slices of the "demand pie" got thinner, providing an incentive
to expand and buy out other companies. According to USDA data, the top
four firms among all food retailers in 2003 were Kroger, Wal-Mart,
Albertsons, and Safeway.68

Consolidation in food retail chains has led to high levels of
concentration in individual metropolitan markets. Table 31 displays market
concentration, as measured by the four-firm concentration ratio, in the 15
markets that we used in this report to analyze the spread between retail
and farm milk

67These were defined by the U.S. Census Bureau as Metropolitan Statistical
Areas. See GAO01-561, appendix VI, for more information on concentration
among retail firms.

68Sales by Wal-Mart and other super centers and mass merchandisers are not
included in the supermarket category. Super centers are defined as a large
combination supermarket and discount general merchandise store, with
grocery products accounting for up to 40 percent of selling area.

Appendix V
Factors That Influence the Price of Milk as It
Moves from the Farm to the Consumer

prices.69 While this threshold varies, some economists have characterized
a market with a four-firm concentration ratio of 60 percent or greater as
a "tight oligopoly" or highly concentrated.70 In 2003, the levels of
concentration varied by metropolitan market, with the percentage of the
market held by the four largest firms ranging from 62.8 percent in the
Minneapolis/St. Paul area to 84.9 percent in Denver, with an overall
unweighted average of 73.9 percent. Moreover, for the 15 markets that we
analyzed, the overall average four-firm concentration ratio for 1998 that
we reported in 2001-74 percent-is comparable to the 2003 average.71

Table 31: Market Share of the Top Four Food Retailers in Selected Markets,
2003

             Retail markets Four-firm market share, 2003 (percent)

                               Atlanta, Ga. 78.2

                               Boston, Mass. 70.1

                              Charlotte, N.C. 82.0

                             Cincinnati, Ohio 79.4

                               Dallas, Tex. 67.3

                               Denver, Colo. 84.9

                                Miami, Fla. 81.6

                             Milwaukee, Wisc. 63.2

                            Minneapolis, Minn. 62.8

                             New Orleans, La. 74.7

                              Phoenix, Ariz. 80.2

69The four-firm concentration ratio in a market is the share of sales made
by the four largest sellers.

70William G. Shepherd, The Economics of Industrial Organization, 3rd ed.
Prentice Hall, 1990. While a useful tool for categorizing the degree of
competition in a market, a numerical cut-off of a concentration index,
such as the four-firm concentration ratio, is not generally considered a
precise demarcation between various categories of the degree of
competition in a market. The Federal Trade Commission and Department of
Justice use an index of market concentration, the Herfindahl-Hirschman
Index, in combination with a number of other market factors, in
determining the potential anticompetitive effects of a proposed merger.
Economic studies of various markets, however, have found that the
four-firm concentration ratio and the Herfindahl-Hirschman Index are
highly correlated.

71While very similar, some of the 15 market areas for our 2001 report for
retail concentration in 1998 did not exactly match those in our present
data for 2003 due to a realignment of MarketScope's metropolitan market
area definitions.

Appendix V
Factors That Influence the Price of Milk as It
Moves from the Farm to the Consumer

(Continued From Previous Page)

             Retail markets Four-firm market share, 2003 (percent)

                           Salt Lake City, Utah 67.6

                             San Diego, Calif. 68.8

                              Seattle, Wash. 71.8

                             Washington, D.C. 76.5

                                  Average 73.9

Source: GAO presentation of USDA/ERS data.

Note: Four-firm market shares are for supermarkets and super centers only.
Smaller grocery stores and convenience stores are excluded.

At the same time, the traditional dominance of supermarkets in food sales
has been challenged by competition from new mass merchandisers and super
centers such as Wal-Mart, K-Mart, and Target. These retailers tend to
offer lower prices and often purchase their inventories in large
quantities to pass on low prices to consumers. According to a recent USDA
report, even the larger conventional food stores do not have the same
buying power as these large general merchandisers. They also tend to grow
by new investment in stores rather than through mergers and acquisitions,
in contrast to traditional supermarkets. Figure 34 displays the change in
food sales by market segment of food retailers between 1993 and 2003.
Sales from supermarkets decreased from 63 percent to 58.3 percent during
this time period. However, sales from warehouse clubs and super centers
increased from 3.6 percent to 9 percent, while those of convenience stores
and drug stores also increased-from 11.9 percent to 13.6 percent. Overall,
sales from nontraditional food retailers-warehouse clubs and super
centers, mass merchandisers, and convenience and drugstores-went from 17.2
percent in 1993 to 24.4 percent in 2003.

                                   Appendix V
                 Factors That Influence the Price of Milk as It
                      Moves from the Farm to the Consumer

Figure 34: Percentage of Retail Food Sales by Market Segment, 1993 and
2003

                                      2003

1.7% 1.8%

Mass merchandisers

3.6%

Warehouse clubs	Warehouse clubs and super centers

Convenience stores and drugstores and drugstores

Other grocery Other grocery stores stores

Supermarkets

Source: GAO presentation of USDA data.

Within the fastest-growing segment, warehouse clubs and super centers, the
largest food retailer is Wal-Mart, followed by Target and Meijer's, while
the second fastest growing segment includes the major drug chains, such as
Walgreen's and CVS. As of 2003, Wal-Mart super center sales reached $103.2
billion, with estimated grocery sales of $41.3 billion.72 According to a
recent ACNielsen study, while all U.S. households still shop in
traditional grocery stores, the annual number of trips to such stores
continues to decline.73 In contrast, super centers have shown strong gains
in household penetration as well as gains in the number of trips per year.
In dairy,

72Wal-Mart super center sales for 2003 were taken from Supermarket News,
SN's Top 75 (supplement), January 12, 2004. The estimate of grocery sales
was provided by USDA. These grocery sales represent about 40 percent of
total super center sales.

73"ACNielsen Study Finds Grocery Stores Continuing to Lose Share of
Customer Shopping Trips," news release, ACNielsen, May 4, 2002,
http://www.acnielsen.com/news/american/us/2002/200220504.htm (accessed
Sept. 9, 2004).

Appendix V
Factors That Influence the Price of Milk as It
Moves from the Farm to the Consumer

however, conventional food stores still offer a larger selection of milk
products. A recent study by researchers at Cornell and Oklahoma State
Universities on dairy case management found that the number of milk
products offered was highest in supermarkets (74) and lowest in drug
stores (16).74 While the volume of milk products was highest for mass
merchandisers, the number of products (24) was similar to convenience
stores (22). The authors explained that historically, mass merchandisers
concentrated on moving a large volume of product with a limited variety.

74Todd M. Schmit, Harry M. Kaiser, and Chanjin Chung, The Dairy Case
Management Program: Does It Mooove More Milk? National Institute for
Commodity Promotion Research and Evaluation, Department of Applied
Economics and Management, Cornell University, Ithaca, New York, January
2004. Specifically, this study evaluated the Dairy Case Management Program
operated by the American Dairy Council in the Northwestern Hudson Valley
Market in New York State.

Appendix VI

Economic Studies of Price Transmission in the U.S. Fluid Milk Market

This appendix summarizes the findings of 14 economic studies of price
transmission in U.S. fluid milk markets. These studies estimated the
extent to which price changes at one level, such as the farm level, are
transmitted to other levels, such as the retail level, and the time in
which these price changes are transmitted. Many of the studies found a
difference, or asymmetry, in either the extent or speed of price
transmission, depending on whether the initial price change was an
increase or a decrease (see table 32).1 Some of the studies analyzed
possible causes for price asymmetry and often identified the presence of
noncompetitive markets as a contributing factor.2 Although most studies
estimated how prices are transmitted from the farm to the retail level, a
few also estimated how price changes are transmitted from the retail level
back to the farm level.

1More generally, a recent article in the economics literature looking at
77 consumer and 165 producer goods suggested that asymmetric price
transmission is a broad phenomenon, and that prices rise faster than they
fall in about two-thirds of the markets that were examined. See S.
Peltzman, "Prices Rise Faster than they Fall," Journal of Political
Economy, Vol. 108, No. 3, 466-502.

2Other factors identified included the costs of changing prices,
government policies, spatial market competition, asymmetric information,
economies of scale, and differentiated products.

                                  Appendix VI
                   Economic Studies of Price Transmission in
                           the U.S. Fluid Milk Market

Table 32: Overview of Fluid Milk Price Transmission Studies and Results on Price
                        Transmission Asymmetry/Symmetry

                                                      Results on  
                                                        price     
                                                     transmission 
                                   Period             asymmetry/  Additional  
     Author    Year      Model     of     Geographic symmetry       results   
                                   study  area       (long run)a  
Carman and Forth- Econometric   1999-     Nine    Asymmetry in Asymmetry   
                     model based           western     certain    in          
     Sexton   coming   on Houck     2003     U.S.    milk types    timing in  
                        (1977)             citiesb   and markets      CA      
               2005                                                 markets   
    Sexton,    2004    Hotelling   1999-     Nine                 Asymmetry   
    Xia, and          framework;           western    Asymmetry   in          
     Carman           Econometric   2003     U.S.                   timing    
                         model             citiesc                
                     Houck                Seven U.S.              
               2004  econometric   1994-  citiesd    Asymmetry in 
Capps, Jr.        model                           majority     
                     and error                                    
                     correction     2002              of cities   
                     model                                        
                     Asymmetric            Ten U.S.               
    Linkow,    2004  friction      1997-   citiese                
     Gould           model                            Asymmetry   
      and                           2004                          
    Stiegert                                                      
    Chidmi,    2004  New Empirical 1996-    Boston                
     Lopez,           Industrial                         N/A      
      and            Organization   2000                          
Cotterill             model                                    
      Lass     2004  Econometric   1982-     New                              
                     model based           England   Post-compact Pre-compact
                      on Kinnucan   2002              asymmetryf  same as     
                      and Forker                                  2001        
                        (1987)                                    
               2003   Structural   1971-   National               Asymmetry   
      Wang            econometric                     Asymmetry   in          
                      model based   1997                          short run;  
                      on Emerick                                  
                        (1994)                                       Symmetry 
                                                                       retail 
                                                                    to farm   
    Romain,    2002    Marketing   1980-  Upstate     Asymmetry          Same 
     Doyon,             margin,           New York     UNY and     results in 
and Frigon         econometric   1997  (UNY) and  NYC prior to  short run  
                         model               New        price-    
                                          York City  gouging law; 
                                            (NYC)    only NYC     
                                                      after the   
                                                         lawg     
                     Two-stage                                    
    Dhar and   2002  market        1996-    Boston                
                     channel                             N/A      
Cotterill             model      1998                          

Lass, Adanu, and Allen

2001	Econometric model based on Kinnucan and Forker

                                    19821997

New England Symmetry	Asymmetry in short run

Frigon, Doyon,

  1999 Marketing margin, 1980-Northeast U.S.; Symmetry in all Asymmetry in all

and Romain

                       econometric model 1997 UNY and NYC

markets except asymmetry in NYC after lawg  markets in short run

     Carman   1998 Econometric model  1985- Three       Symmetry    Asymmetry 
                   based                    California                     in 
                        on Houck      1997    citiesh              short run  
    Emerick   1994     Structural     1971-  National                         
                      econometric                       Asymmetry Symmetry in
                   model based on     1991                         short run  
                   Houck                                          
    Hansen,   1994 Econometric model  1983-  National             
     Hahn,                                              Asymmetry 
and Weimar                         1990                        

Source: GAO analysis using various sources (see table 33).

aThese results indicate farm-to-retail price asymmetry or symmetry.

bThese cities include the four California markets of Sacramento, San
Francisco, Los Angeles, and San Diego and the five non-California markets
of Seattle, Portland, Salt Lake City, Denver, and Phoenix.

Appendix VI
Economic Studies of Price Transmission in
the U.S. Fluid Milk Market

cThe cities are the same as in Carman and Sexton (forthcoming 2005).
dThese metropolitan markets include Atlanta, Boston, Chicago, Dallas,
Hartford, St. Louis, and Seattle.
eThese metropolitan markets could not be disclosed.
fThe Compact refers to the Northeast Interstate Dairy Compact.
gThis law refers to the New York milk price-gouging law of 1991 (codified
at NY Gen. Bus. S:396-rr).
hThese three California cities include Los Angeles, San Francisco, and
Sacramento.

How prices are transmitted within the milk marketing chain is important to
policy makers because it affects both farmers and consumers. Farmers may
be concerned with price transmission because they may believe increases in
retail prices are not fully passed back to the farm level, while decreases
are passed on. Consumers and farmers may believe that decreases in farm
prices are not fully passed along to the retail level, while increases are
passed on. Figure 35 illustrates price transmission through the vertical
milk marketing chain; the arrows show how price signals are transmitted in
both directions between marketing levels.

Appendix VI
Economic Studies of Price Transmission in
the U.S. Fluid Milk Market

Figure 35: Marketing Chain and Price Transmission for the U.S. Fluid Milk
                                    Industry

Source: GAO.

The first section of this appendix is a detailed table summarizing the
models, data, and key assumptions used in each study, and each study's
results. The second and third sections discuss the farm-to-retail results,
including evidence of price asymmetry, with respect to the extent of price
transmission and the speed of price transmission. The fourth section
discusses the retail-to-farm results on price transmission. The last
section discusses the studies' findings regarding factors that might cause
price asymmetry.

                                  Appendix VI
                   Economic Studies of Price Transmission in
                           the U.S. Fluid Milk Market

Summary of Recent Economic Studies of Price Transmission in the U.S. Fluid
Milk Market

While most of the studies summarized in table 33 use, as a basis for their
models, the standard Houck (1977)3 and Kinnucan and Forker (1987) models
to identify price asymmetry,4 others use newer methods such as the error
correction model, which some researchers believe provides a more
appropriate specification for examining asymmetric price transmission.5
Most of the studies estimated only "forward" price transmission, or price
transmission from the farm to the retail level, but we also report on one
study that estimated "backward" price transmission, from the retail level
to the farm. The studies also differed in whether they estimated short-run
or long-run price transmission asymmetry or both.6 We take all of these
differences into account in interpreting the studies' overall conclusions
and discussing their results.

3James P. Houck, "An Approach to Specifying and Estimating Nonreversible
Functions," American Journal of Agricultural Economics 59 (1977): 570-572.
In the Houck model (1977), changes in the retail price are linked to
increases and decreases in the farm level price as well as changes in
other marketing costs. Also, the estimated parameters on the farm-level
price increases and decreases can be tested to determine if retail price
movements in response to farm price changes are symmetric or asymmetric.

4H.W. Kinnucan and O.D. Forker, "Asymmetry in Farm-Retail Price
Transmission for Major Dairy Products," American Journal of Agricultural
Economics 69 (1987): 285-292. Kinnucan and Forker combine the mark-up
model proposed by Heien (1980) and the approach used by Houck (1977) to
estimate asymmetric functions using a dynamic approach for fluid milk,
cheese, butter, and ice cream.

5A recent study by von Cramon-Taubadel and Fahlbusch (1994) points out
that in the case of cointegration between two time series, an error
correction model, extended by the incorporation of asymmetric adjustment
terms, provides for a more appropriate specification for testing for
asymmetric price transmission. See S. von Cramon-Taubadel and S.
Fahlbusch, "Identifying Asymmetric Price Transmission with Error
Correction Models." Poster session at the European Association of
Agricultural Economists European Seminar in Reading, UK, 1994.

6Short-run asymmetry occurs when increasing and decreasing prices have
different immediate responses when the farm price is changed. Long-run
asymmetry occurs when increasing and decreasing farm prices have different
responses over the full adjustment period. For most of the studies that we
examined, both the short run and long run represented fairly short time
horizons. For example, the short run typically extended from the current
month to one month out, while the long run was usually from 2 to 4 months.
Nevertheless, in most of the studies, the long run was the time required
for the lag structure to run its course.

Appendix VI
Economic Studies of Price Transmission in
the U.S. Fluid Milk Market

Table 33: Recent Price Transmission Studies of the Fluid Milk Market Model
                     Description Price Transmission Results

Author/ Study/Year Type of economic model Scope of model and price data
used Additional assumptions/ variables/ limitations Timing/symmetry and
degree of price transmissiona Other model results

Carman and Econometric Average Market power Sexton, model based monthly
prices based on "Supermarket on the farm-to-for nine hypothesis Fluid Milk
retail price California and tests of Pricing transmission non-California
parameter Practices in model of Houck metro markets, size, the Western
(1977) for 1999-2003. parameter United States," whole, skim, 1 tests for
Agribusiness, percent, and 2 Farm prices: asymmetry, forthcoming percent
milk. California: and a 2005. Estimated time Class I departure from

lags using an California milk cost Almon prices. differences of
distributed lag Other states: fluid milk model. FMMO Class I products due

prices.	to different fat contents.

Retail prices:

ACNielsen average monthly retail prices.

Timing/price asymmetry: California and non-California: Contemporaneous to
3-month period.

California markets: Some evidence of price symmetry and some of price
asymmetry, depending on market and product. Non-California markets: Price
transmission in almost all markets was asymmetric.

Degree of price transmission:

California:

Sacramento: Increases: 72%-122% Decreases: 52%-97% Los Angeles: Increases:
60%-115% Decreases: 75%-106% San Francisco: Increases: 65%-115% Decreases:
49%-94% San Diego: Increases: 56%-109% Decreases: 58%-110%

Non-California:

Seattle: Increases: 57%-72% Decreases: 39%-64% Portland: Increases: 3%-25%
Decreases: 4%-30% Phoenix: Increases: 5%-59% Decreases: 6%-63% Salt Lake
City: Increases: 39%-56% Decreases: 54%-83% Denver Increases: 43%-56%
Decreases: 40%-55% Analysis revealed evidence of market power in fluid
milkpricing in each of the nine metro markets analyzed.

In Portland, estimated retail price coefficients did not respond to farm
prices.

In non-California markets, only 3 of 40 estimated coefficients were
consistent with perfect competition in pricing.

Addition of retail skim milk, 1 percent, and 2 percent led to different
results than previous analysis of only whole milk.

Appendix VI
Economic Studies of Price Transmission in
the U.S. Fluid Milk Market

                         (Continued From Previous Page)

                  Model Description Price Transmission Results

Author/ Study/Year Type of economic model Scope of model and price data
used Additional assumptions/ variables/ limitations Timing/symmetry and
degree of price transmissiona Other model results

Sexton, Xia, Theoretical Average Three and Carman, model based monthly
prices competition "Horizontal on Hotelling for nine scenarios
Differentiation framework of California and examined for with horizontally
non-California retail fluid milk Differential differentiated cities.
markets: Input Costs: products with California data, perfect Retail Prices
differential 1999-2003; competition, for Milk by Fat costs. Non-California
monopoly, and Content." Econometric data, 2000- oligopoly. Paper
estimation - 2003. presented at seemingly Skim milk the annual unrelated
Farm prices: (low-cost) and AAEA regression-of a Class I prices whole milk
meetings, differentiated from California (high-cost) Denver, product,
market Dairy represent Colorado, competition Information products with
2004.b model using Bulletin and differential

metropolitan FMMO Class I costs retail markets component transmitted for
whole and prices. differently in skim milk. different

Retail prices: competition
California Dairy scenarios.
Information
Bulletin. Farm prices do

not contain over-order premiums.

Timing/price asymmetry:

California:

Los Angeles and San Diego have longer transmission periods than Sacramento
and San Francisco.

Non-California:

Prices are transmitted gradually for Denver, Salt Lake City, Portland, and
Phoenix, while prices in Seattle are transmitted quickly, within 1 month.

Degree of price transmission:

N/A Competition results:

California:

Results suggest oligopoly scenario for Sacramento, San Francisco, and Los
Angeles markets. Some evidence of oligopoly scenario for San Diego market.

Non-California:

Results suggest oligopoly scenario for Phoenix, Salt Lake City, and
Seattle markets. Model also suggests Denver market is closer to oligopoly,
while Portland is closer to monopoly scenario.

Appendix VI
Economic Studies of Price Transmission in
the U.S. Fluid Milk Market

                         (Continued From Previous Page)

                  Model Description Price Transmission Results

Author/ Study/Year Type of economic model Scope of model and price data
used Additional assumptions/ variables/ limitations Timing/symmetry and
degree of price transmissiona Other model results

Capps, Jr., The standard Monthly data Elasticities of
"Asymmetry in Houck from January price
Farm-Retail econometric 1994 to transmission
Price price October 2002 are calculated
Transmission transmission for seven U.S. for each
Associated approach is cities: Atlanta, product and
with Whole compared to Boston, market.
Milk and Two the error Chicago,
Percent Milk in correction Dallas, The model
Selected U.S. model using Hartford, St. does not
Cities: A co-integration Louis, and include any
Comparison of procedures. Seattle. other
the Model marketing
Conventional estimates Farm prices: input costs.
Houck results for USDA
Approach and whole milk and Agricultural
the Error 2 percent milk. Marketing
Correction Service.
Model
Approach." Retail prices:
Working USDA
paper, Agricultural
Department of Marketing
Agricultural Service.
Economics,
Texas A&M
University,
2004.

Houck method Error correction model

Price asymmetry:

(Whole milk) Asymmetric-Atlanta, Boston, Chicago, Atlanta, Chicago, and
Dallas, and Hartford and Dallas Symmetric-St. Louis and Seattle St. Louis

(2 percent) Asymmetric-Atlanta, Boston, Chicago, Atlanta, Chicago, and
Hartford and Seattle

Symmetric-

Dallas, St. Louis, and St. Louis and Seattle Dallas

Elasticity of price transmission (ranges): (Whole milk)

Increases: Increases:

(+0.23)-(+0.58) (+0.24)-(+0.35) Decreases: Decreases: (-0.02)-(+0.12)
(-0.03)-(+0.06)

                                  (2 percent)

Increases: Increases:

(+0.11)-(+0.46) (+0.05)-(+0.29)

Decreases: Decreases:

Author notes that results suggest there is a possible market failure due
to a highly concentrated retail sector, with the exception of St. Louis
and Seattle.

                        (-0.06)-(+0.25) (-0.006)-(+0.12)

Appendix VI
Economic Studies of Price Transmission in
the U.S. Fluid Milk Market

                         (Continued From Previous Page)

                  Model Description Price Transmission Results

Scope of model and price data used Additional assumptions/ variables/
limitations Type of economic model Author/ Study/Year Timing/symmetry and
degree of price transmissiona Other model results

Linkow, Gould, and Stiegert, "Retail Consolidation, Market Concentration,
and Farm-Retail Price Asymmetry in the U.S. Fluid Milk Market." Working
paper, Department of Agricultural Economics, University of
Wisconsin-Madison, 2004.

Authors use an asymmetric friction model that accounts for the existence
of menu costs (the costs of repricing products). Estimated model uses the
price of private label whole milk.

Weekly panel data from 1997--2004 for 10 metropolitan fluid milk markets.

Farm prices:

FMMO Class I announced cooperative price.

Retail prices:

IRI retail price data; private label (store brand) whole

c

milk.

Price asymmetry and menu costs are allowed to vary by city.

Includes a marketing cost index variable based on a paper by Romain et
al., (2002).

Includes a measure of market concentration and a measure of spatial market
concentration.

Timing/price asymmetry:

Asymmetric-Farm price decreases are passed on less completely than
increases.

Responses to price increases do not differ considerably across cities.
Responses to price decreases do differ across cities.

Degree of price transmission:

N/A

Appendix VI
Economic Studies of Price Transmission in
the U.S. Fluid Milk Market

                         (Continued From Previous Page)

                  Model Description Price Transmission Results

Author/ Study/Year Type of economic model Scope of model and price data
used Additional assumptions/ variables/ limitations Timing/symmetry and
degree of price transmissiona Other model results

Chidmi, The New Weekly data Assesses Lopez, and Empirical from1996-2000
retail prices Cotterill, Industrial for Boston under four "Dairy
Organization metropolitan scenarios: Compact, structural fluid milk NEDC,
no Market Power, model of market. NEDC, with and Milk Appelbaum perfect
Prices in (1982) that Farm prices: competition, Boston." includes The
higher of and oligopoly; Working measures of the FMMO estimates paper,
market power Class I prices market power Department of and price or
Compact and market Agriculture transmission. Class I prices, conduct and
Resource The model plus over-order variables. Economics, tests for values
premiums. University of of these Assumes cost-Connecticut, variables with
Retail prices: plus pricing 2004. and without the Computed from between

Northeast IRI retail data processors Interstate Dairy by dividing total
and retailers in Compact sales by total the Boston (NEDC).d volume.c area.

Model does not estimate the timing of price transmission.

Timing/price asymmetry:

N/A

Degree of price transmission: Increases: 68% Transmission elasticities
were estimated at 0.331 pre-Compact and 0.2911 post-Compact.

Retail price increases due to market power outweighed impact of NEDC
bynearly 7 times.

Both the preand post-NEDC retail price markup over marginal cost (the
competitive benchmark) is about 25% of the retail fluid milk price.

Market conduct- supermarkets do not ignore each other's actions.

MILC program provides benefits to farmers, consumers, and retailers at the
expense of

e

taxpayers.

Appendix VI
Economic Studies of Price Transmission in
the U.S. Fluid Milk Market

                         (Continued From Previous Page)

                  Model Description Price Transmission Results

Author/ Study/Year Type of economic model Scope of model and price data
used Additional assumptions/ variables/ limitations Timing/symmetry and
degree of price transmissiona Other model results

Lass, "Impacts Update of the Monthly time-Longer data of the previous 2001
series price series than Northeast econometric data for New 2001 paper
Dairy mark-up model England with time Compact on of Kinnucan (Boston and
periods split New England and Forker Hartford), 1982 into pre and Prices
(1987) allowing to 2002. post-Compact; Revisited: for rising and a
pre-1997 New Data, falling farm Farm prices: period and a New Lessons."
prices and FMMO Class I 1997-2002 Working different price for the period.
paper, speeds of New England Department of adjustment. market. Parameters
of Resource the pre- and Economics, Retail prices: post-Compact University
of USDA periods were Massachusett Agricultural tested to s, Amherst,
Marketing determine 2004.d Service retail changes in

price series for price Boston and transmission. Hartford.

Assumes constant returns to scale and a competitive market beyond the farm
gate.

Included variable to measure changes in processing costs using USDA's Food
Marketing Cost Index.

Timing/price asymmetry:

Pre-Compact period:

Results were consistent with original findings (Lass et al., 2001); price
asymmetry in the short run and symmetry in the long run.

Post-Compact period: Short run: Price asymmetry

Long run: Price asymmetry.

Degree of price transmission:

N/A Boston: There was greater variation infarm prices without the Compact,
leading to higher retail prices.

Hartford: There was greater variation infarm prices without the Compact,
with greater increases in retail prices.

Appendix VI
Economic Studies of Price Transmission in
the U.S. Fluid Milk Market

                         (Continued From Previous Page)

                  Model Description Price Transmission Results

Author/ Study/Year Type of economic model Scope of model and price data
used Additional assumptions/ variables/ limitations Timing/symmetry and
degree of price transmissiona Other model results

Wang, Price Structural National in Competitive Transmission econometric
scope; uses fluid milk and the Role model using a national market and of
Federal simultaneous monthly constant Dairy Policy in system of average
data returns to U.S. Dairy equations of from 1971- scale. Markets.
national dairy 1997. Master's prices, Policy variable thesis, Cornell
including farm Farm prices: for changes in University, and retail fluid
Announced price support 2003. milk. Models cooperative in 1980s.

both farm-to-Class I price.
retail and retail-Price lags
to-farm price Retail prices: chosen from
transmission. U.S. retail price dairy market
Model based from Bureau of literature.
on a previous Labor
one by Emerick Statistics
(1994). USDA.

1988 - 1997 Period Farm-to-retail: Retail-to-farm:

                            Timing/price asymmetry:

Short run: Short run: (1 month) (1 month) Asymmetric Asymmetric

Long run: Long run:

(3 months) (3 months) Asymmetric Symmetric

                         Degree of price transmission:

Short run: Short run:

Increases: 94% Increases: 94% Decreases: 31% Decreases: 2%

Long run: Long run:

Increases: 83% Increases: 40% Decreases: 64% Decreases: 34% Decreases in
dairy price supports in the mid-1980s led to greater asymmetric price
transmission and more volatility in the post-1988 period than in the
pre-1988 period for fluid milk and nonfat dry milk.

Appendix VI
Economic Studies of Price Transmission in
the U.S. Fluid Milk Market

                         (Continued From Previous Page)

                  Model Description Price Transmission Results

Author/ Study/Year Type of economic model Scope of model and price data
used Additional assumptions/ variables/ limitations Timing/symmetry and
degree of price transmissiona Other model results

Romain, Marketing Monthly data Model allows
Doyon, and margin model from 1980 - for testing of
Frigon, that includes 1997 for constant
"Effects of quantity of Upstate New returns to
State commodity York (UNY) and scale.
Regulations marketed and New York City
on Marketing marketing costs (NYC) markets. Model allows
Margins and combined with for asymmetry
Price the Houck Farm prices: of other
Transmission (1977) model. It UNY and NYC marketing
Asymmetry: includes prices are the costs.
Evidence from variables FMMO Class I
the New York measuring prices. Model includes
City and cumulative state policy
Upstate New increases and Retail prices: variables-such
York Fluid Milk decreases in UNY and NYC as
Markets," farm-level prices are from deregulation of
Agribusiness, prices. The the New York milk
2002. Akaike State distribution in

Information Department of NYC and the
Criterion was Agriculture. NYC price
used to gouging law.f
determine the
appropriate lag Model does
structure. not account for

over-order premiums in farm prices.

Timing/price asymmetry:

Short run:

Asymmetric-NYC and UNY before price-gouging law. Asymmetric-NYC after
price-gouging law.

Long run:

Asymmetric-UNY and NYC, before price-gouging law. Asymmetric-NYC after
price-gouging law.

Price transmission elasticities:

NYC UNY Before the price-gouging law:

                       Increasing elasticities: 0.70 0.62

                       Decreasing elasticities: 0.30 0.49

After the price-gouging law:

                       Increasing elasticities: 0.52 0.52

                       Decreasing elasticities: 0.43 0.51

Model results indicate constant returns to scale.

NYC experienced price asymmetry in the long run prior to 1991, before the
price-gouging law came into effect, but much less thereafter.

The deregulation of milk distribution to NYC in 1987 allowed Farmland
Dairies' entry into the NYC market and significantly reduced marketing
margins.

Appendix VI
Economic Studies of Price Transmission in
the U.S. Fluid Milk Market

                         (Continued From Previous Page)

                  Model Description Price Transmission Results

Author/ Study/Year Type of economic model Scope of model and price data
used Additional assumptions/ variables/ limitations Timing/symmetry and
degree of price transmissiona Other model results

Dhar and A two-stage Monthly Model allows
Cotterill, structural average data for both
"Price market channel for Boston vertical and
Transmission model of the market from horizontal
in Boston milk March 1996 to market
Differentiated market using a July 1998. channel
Product nonlinear effects.
Market model of costs Farm prices:
Channels: A and demand. Boston FMMO Model allows
Study of the The model Class I fluid for the
Boston Fluid looks at the milk price. identification of
Milk Market Boston fluid cross-firm
and the milk processor Retail prices: pass-through
Northeast and retailer IRI data for rates.
Dairy market aggregate retail
Compact," channels. and top 4 Authors
Food supermarket specify three
Marketing chains in the different
Policy Center, Boston market.c oligopoly
Department of games.
Agricultural
and Resource Study does not
Economics, test for the
University of timing of price
Connecticut, transmission
2002. or price

transmission asymmetry.

Timing/price asymmetry:

N/A

Degree of price transmission (ranges):

Processor-to-wholesale: 55%-65% Wholesale-to-retail: 54%-62%

Firm-specific transmission: 32%-47% Industrywide transmission: 88%-100%
Competition between retailers was lessened dueto focal point pricing and
the Compact.

Industrywide and firmspecific cost shocks were not identical, nor did the
latter aggregate to the former.

Appendix VI
Economic Studies of Price Transmission in
the U.S. Fluid Milk Market

                         (Continued From Previous Page)

                  Model Description Price Transmission Results

Author/ Study/Year Type of economic model Scope of model and price data
used Additional assumptions/ variables/ limitations Timing/symmetry and
degree of price transmissiona Other model results

Lass, Adanu, Econometric Monthly time-Uses national and Allen model based
series price data to "Impacts of on the basic data for New determine the
the Northeast markup model England lag structure Dairy of Kinnucan (Boston
and for fluid milk- Compact on and Forker Hartford), 1982 uses lags of 1
Retail Prices," (1987) allowing to 1997. and 2 months. Agricultural for
rising and and Resource falling farm Farm prices: Model uses Economics
prices and Class I FMMO impacts of Review, 2001. different price for the
price

speeds of New England transmission adjustment. market. to determine

the impact of Retail prices: the Compact. USDA Agricultural Assumes
Marketing constant Service retail returns to price series for scale and a
Boston and competitive Hartford. market beyond

the farm gate.

Includes variable to measure changes in processing costs using USDA's Food
Marketing Cost Index.

Timing/price asymmetry:

Boston: Retail prices increased most rapidly in the current period; for
declines, the greatest decreases occurred after a one-period lag.
Hartford: Current period effects of rising farm prices were greater than
current period falling farm prices.

Boston and Hartford: Short run: Price asymmetry. Long run: Price symmetry.

Degree of price transmission and elasticities of price transmission:

Degree of Boston Hartford price transmission: 58% average 48% average

Elasticities: SR LR SR LR Rising elasticity 0.46 0.35 0.30 0.33

Falling elasticity 0.14 0.35 0.15 0.25

In the long run, increases in margins were primarily accounted for by
other factors, such as processing costs, rather than the Compact.

Impact of the Compact on retail prices was less than the over-order
premium, suggesting that less than the full amount of the premium was
passed on to consumers.

Appendix VI
Economic Studies of Price Transmission in
the U.S. Fluid Milk Market

                         (Continued From Previous Page)

                  Model Description Price Transmission Results

Author/ Study/Year Type of economic model Scope of model and price data
used Additional assumptions/ variables/ limitations Timing/symmetry and
degree of price transmissiona Other model results

Frigon, Doyon, Hybrid of the Northeast Competitive and Romain, marketing
cost United States, market and "Asymmetry in model, which Upstate New
constant Farm-Retail includes the York (UNY), returns to Transmission
quantity of and New York scale in the commodity City (NYC) assumptions
Northeastern marketed and markets using are relaxed. Fluid Milk costs,
with the monthly data, Market," Food Houck model, 1980-1997. A variable
for Marketing which includes market Policy Center, variables of Farm
prices: concentration, Department of cumulative UNY and NYC the four-firm
Agricultural increases and prices are the concentration and Resource
decreases in Class I FMMO ratio, included Economics, farm-level prices;
for for UNY and University of prices. Northeast NYC. Connecticut, United
States- 1999. FMMO Class I Model allows

average prices for marketing
of the four costs.
FMMO regions
at that time. Model includes

policy Retail prices: variables such UNY and NYC as the NYC prices from
price-gouging New York State law. Department of Agriculture; Model does
Northeast not account for United States over-order prices from the
premiums in Bureau of farm prices. Labor Statistics.

Timing/price asymmetry:

Short run:

Asymmetry existed for UNY after 2 months, and still existed in NYC and the
Northeast United States after 3 months.

Long run:

After 4 months, the Northeast United States, UNY, and NYC (postgouging
law) fully adjusted. Only NYC (pregouging law) had not adjusted.

Symmetry existed in all regions in the long run except NYC prior to 1991.

Degree of price transmission:

N/A NYC experienced price asymmetry in the long run prior to 1991, before
the price-gouging law came into effect.

Variable for market concentration was found insignificant for both UNY and
NYC. Authors believed that the problem of consumer information was the
predominant cause of asymmetric price transmission.

Appendix VI
Economic Studies of Price Transmission in
the U.S. Fluid Milk Market

                         (Continued From Previous Page)

                  Model Description Price Transmission Results

Author/ Study/Year Type of economic model Scope of model and price data
used Additional assumptions/ variables/ limitations Timing/symmetry and
degree of price transmissiona Other model results

     Carman, Response of Three Uses   Timing/price asymmetry: For five of the 
               lags of 1              
     "California retail-level milk                             six equations, 
       California month for price     
    Milk Marketing prices to market                Short run: the one-for-one 
          areas: increases and        
Margins," changes in Los Angeles,    Asymmetric-Unlike price increases,    
              2 months for                       there was a price            
       Journal of farm prices San      significant 1-month lag between farm   
            Francisco, price                    price transmission            
    Food using Houck's and deceases.     decreases and total retail price     
                                           decreases for each process is      
       Distribution (1977) model      city. consistent with                   
              Sacramento.             
    Research, for estimating Monthly                          constant dollar 
          data, Uses an index         
    1998. nonreversible January 1985                Long run: markup pricing. 
              of marketing            
        functions. The to March 1997.   Symmetric-Retail milk price changes   
                       costs from the                were not                 
                  model is USDA/Econo  significantly different for increases  
                                               and decreases For Los          
     estimated using Farm prices: mic of farm prices. Angeles, tests          
                             Research 
    prices for whole Class I prices                               showed that 
                Service               
    milk. for Northern marketing cost Degree of price transmission: retailers 
                                      were                                    
                  and Southern index.         Sacramento: maintaining         
            California from              Increases: (111%-112%) prices by     
            California Dairy            Decreases: (110%-117%) absorbing some 
              Information             San Francisco: cost increases,          
               Bulletin.              Increases: (117%-118%) and reducing     
                                       Decreases: (94%-103%) prices less than 
             Retail prices:                  Los Angeles: farm prices         
            California Dairy              Increases: (76%-88%) decreased.     
              Information                      Decreases: (62%-94%)           
               Bulletin.              

Appendix VI
Economic Studies of Price Transmission in
the U.S. Fluid Milk Market

                         (Continued From Previous Page)

                  Model Description Price Transmission Results

Author/ Study/Year Type of economic model Scope of model and price data
used Additional assumptions/ variables/ limitations Timing/symmetry and
degree of price transmissiona Other model results

Emerick, An Structural National scope-Price lags Econometric econometric
national chosen from Analysis of model based monthly knowledge of Dairy
Market on Houck's average data the dairy Price method (1977) series,
1971-market-lag Transmission using a 1991. length 2 Processes.
simultaneous months for Master's system of 32 Farm prices: farm price
thesis, Cornell equations and Announced increases and University, a
distributed cooperative decreases. 1994. lag formulation. Class I price.

The model Models both estimates price Retail prices: farm-to-retail
transmission U.S. average and retail-tofor five dairy city retail fluid
farm price products, milk price from transmission including fluid Consumer
results. milk. Price Index,

Bureau of Model uses Labor slope and Statistics. intercept

dummy variables used to separate time periods into 1977- 1988 and
1988-1991 to account for government support price decreases and greater
price volatility.

Farm-to-retail:

Timing/price asymmetry

Short run: Symmetric (Immediate)

Long run: Asymmetric (2 months)

Degree of price transmission Short run:

Increases: 45% Decreases: 41%

Long run:

Increases: 107% Decreases: 55% Author estimates causality for fluid milk
using data from 1971- 1991 and finds it multidirectional.

Since 1988, different retail price responses (asymmetric) occur 1 and 2
months after the initial farmlevel price change. Author notes this may be
due to greater volatility causing difficulties in determining the
"appropriate" price at wholesale or retail.

Asymmetry is not a short-run phenomenon in most other dairy products
including retail butter, wholesale nonfat dry milk, and retail ice cream.

                                  Appendix VI
                   Economic Studies of Price Transmission in
                           the U.S. Fluid Milk Market

                         (Continued From Previous Page)

                  Model Description Price Transmission Results

Author/ Study/Year Type of economic model Scope of model and price data
used Additional assumptions/ variables/ limitations Timing/symmetry and
degree of price transmissiona Other model results

Hansen, Statistical National Unlike other
Hahn, and techniques average studies,
Weimar, similar to those quarterly data, includes the
Determinants used by Hahn's 1983-1990 for wholesale
of the Farm-(1989, 1990) farm, market level in
to-Retail Milk markup model wholesale, and the model.
Price Spread, in his study of retail levels.
Economic price
Research transmission. Farm prices:
Service, Estimates price Class I FMMO
USDA, 1994. transmission prices from

for whole milk.	AMS/USDA plus over-order premiums.

Wholesale prices:

Firm-level private costaccounting company data for 30 companies across the
country.

Retail prices:

Bureau of Labor Statistics data based on retail prices in 91 areas.

Timing/price asymmetry: Farm-to-wholesale:

Asymmetric
Farm price increases: wholesale price adjusted in
three quarters.
Farm price decreases: wholesale price adjusted in
one quarter.

Farm-to-retail:
Asymmetric
Farm price increases:
retail price adjusted in one quarter.
Farm price decreases: retail price adjusted in 10
quarters.

Degree of price transmission:
Retail price adjustment:
Asymmetric
Farm price increases: 173%
Farm price decreases: 92%

Wholesale prices exhibit short-run asymmetric responses to farm prices,
while retail prices exhibit short-run and long-run, or irreversible, price
adjustments.

In this study, wholesale milk price adjustments were more rapid for farm
price decreases than for farm price increases.

aThe degree or level of price transmission is represented by the
percentage of full price pass-through from farm to retail. Some studies,
however, measured instead the elasticity of price transmission, which is
defined as the percentage change in the retail price of a product due to a
1 percent change in the corresponding farm price. As noted in the table,
some studies do not measure the degree of price transmission or the
elasticity of price transmission and only report on the timing and
symmetry of transmission.

bThe AAEA is the American Agricultural Economics Association.

cIRI stands for Information Resources, Inc.

dThe Northeast Interstate Dairy Compact was a pricing program established
by the 1996 Farm Bill that set a minimum price for raw milk to be used for
and sold as fluid milk in the New England states of Connecticut, Maine,
Massachusetts, New Hampshire, Rhode Island, and Vermont. This program
expired on September 30, 2001.

                                  Appendix VI
                   Economic Studies of Price Transmission in
                           the U.S. Fluid Milk Market

eThe Milk Income Loss Contract program (MILC) is similar to the Compact,
in that it is based on the Boston raw milk price and provides a partial
subsidy to dairy farmers; however, it is funded through taxpayers rather
than processors.

fThe New York City price-gouging law (codified at NY Gen. Bus. S:396-rr)
imposed a duty on the New York Commissioner of Agriculture and Markets to
determine if prices of fluid milk are unconscionably excessive whenever
the retail price of fluid milk exceeds 200 percent of the price for Class
I fluid milk. The New York State Legislature passed this law, in part,
because of a perceived lack of response in retail milk prices to decreases
in farm prices.

Results on the Extent On both national and regional or citywide levels,
the majority of the fluid

milk studies that we identified found evidence of farm-to-retail price of
Farm-to-Retail Price transmission asymmetry in price levels. While the
studies estimated a wide Transmission and Price range of price
transmission levels, in general, the estimates of price Transmission
transmission for initial farm price increases were greater than for farm

price decreases.

Asymmetry

National Level Results	The fluid milk price transmission studies that we
identified using national aggregate data estimated a wide range of price
transmission levels and generally found evidence indicating asymmetric
price transmission-farm price increases were more fully transmitted to
retail prices than farm price decreases. Using national average farm and
retail prices for whole milk, Emerick (1994) and Wang (2003) developed two
similar studies that identified the degree of price transmission
nationally. While both studies used similar models, the Wang study used
somewhat more recent data. Taken together, both studies' short-run results
suggest that about 45 percent to 94 percent of farm price increases were
passed along to the retail level, while only 31 percent to 41 percent of
farm price decreases were similarly passed along. These studies estimated
that in the long run, transmission levels for price increases ranged from
83 percent to 107 percent, while transmission levels for price decreases
ranged from 55 percent to 64 percent. Both researchers found price
asymmetry in the long run, but Wang also found price asymmetry in the
short run. In his study of seven metropolitan markets across the country
using data from 1994 to 2003, Capps Jr. (2004) also identified price
asymmetry in a majority of selected fluid milk markets for whole and 2
percent milk. Measuring the level of price transmission by using the
elasticity of price transmission, he estimated elasticities for farm price
increases ranging from 0.23 to 0.58 for whole milk and 0.11 to 0.46 for 2
percent milk. For farm price decreases, these elasticities were much
lower, ranging from -0.02 to 0.12 for whole

                                  Appendix VI
                   Economic Studies of Price Transmission in
                           the U.S. Fluid Milk Market

and -0.06 to 0.25 for 2 percent, respectively.7, 8 Similarly, in Linkow et
al. (2004), using an asymmetric friction model of 10 metropolitan markets
across the country, the authors found evidence of price asymmetry-retail
prices were more responsive to cooperative farm price increases than
decreases.

Regional and City-Level Results

Many studies that used regional or city-level data also estimated a wide
range of price transmission levels as well as asymmetric price
transmission. In an econometric model for whole, skim, 1 percent, and 2
percent milk, Carman and Sexton (forthcoming 2005) estimated farm-toretail
price transmission for nine metropolitan markets in the Western United
States. Estimated levels of price transmission for all types of fluid milk
combined for the California markets ranged from 56 percent in San Diego to
122 percent in Sacramento for price increases, and from 49 percent in San
Francisco to 110 percent in San Diego for price decreases. For the
non-California markets, for both increases and decreases, the estimated
levels of price transmission were much lower, ranging from 3 percent in
Portland to 72 percent in Seattle for price increases and 6 percent in
Phoenix to 83 percent in Salt Lake City for price decreases. The authors
also noted that none of their price transmission parameters for all types
of fluid milk for Portland was statistically different from zero,
indicating no evidence that retail prices responded to farm price changes
in this market.

Several researchers have also estimated farm-to-retail price transmission
for markets in the Northeast United States. Lass et al. (2001) estimated
price transmission in reference to the Northeast Interstate Dairy Compact
for the Boston and Hartford metropolitan areas. For price increases, they
found that the level of price transmission for a farm price increase was
58 percent for Boston and 48 percent for Hartford. While Lass et al. only
found price asymmetry in the short run, in a subsequent study, using more
recent data, Lass (2004) found price asymmetry in the short run and in the
long run. Using the New Empirical Industrial Organization approach,
Chidmi, Lopez, and Cotterill (2004) found a similar result: a degree of
price

7The elasticity of price transmission from the farm to retail level is the
percentage change in the retail price of a product due to a 1 percent
change in the corresponding farm price.

8While these elasticity estimates are ones obtained from using the Houck
method, the author found similar results, although smaller, using a
time-series, cointegration approach: the error correction model (see table
33).

Appendix VI
Economic Studies of Price Transmission in
the U.S. Fluid Milk Market

transmission of 68 percent for farm price increases for the Boston
market.9 Within the same study, the authors also note that Cotterill
(2003), in other research for Boston, estimated a pass-through rate of
between 20 and 26 percent for price decreases, suggesting price asymmetry
in this market. In another study of the Boston market, using a two-stage
market channel model, Dhar and Cotterill (2002) found that the
firm-specific price pass through rate was 32 to 47 percent, while
industrywide pass through was 88 to 100 percent for price increases.10

In yet another study of the Northeast market, Romain et al. looked at
price asymmetry before and after the imposition of the New York
price-gouging law in 1991 and tested for price asymmetry using the
elasticity of price transmission. Before the price-gouging law, they found
that a 1 percent increase in the farm price translated into a 0.70 percent
and 0.62 percent increase in retail prices in New York City and Upstate
New York, respectively, while a 1 percent decrease translated into a 0.30
percent and 0.49 percent decrease in these markets. After the law went
into effect, they found that a 1 percent increase in the farm price
translated into a 0.52 percent increase in retail prices in both New York
City and Upstate New York, while a 1 percent decrease translated into a
0.43 percent and 0.51 percent decrease in these markets, respectively.
Therefore, long-run price asymmetry was significant in both regions prior
to the price-gouging law, but remained statistically significant only in
New York City afterwards, though at a much lower level.

9The New Empirical Industrial Organization is an approach that focuses
primarily on the relationship between prices and marginal costs that has
been used to identify and estimate the degree of oligopoly market power in
a market. Specifically, the Chidmi et al. study is a structural oligopoly
model that measures market power, demand, marginal cost, and the
farm-to-retail price transmission.

10This approach is a cost-pass through model consisting of a two-stage
vertical market system where there are two processors at the first stage
and two retailers at the second stage. Within these stages the authors
assume different conduct and game theoretic assumptions. This study,
however, did not address price asymmetry because it did not estimate the
extent to which price decreases were passed through.

                                  Appendix VI
                   Economic Studies of Price Transmission in
                           the U.S. Fluid Milk Market

Results on Farm-to-Retail Speed of Adjustment and Price Transmission
Asymmetry

We identified fewer studies that examined the speed of adjustment and
related price asymmetry, or differences in the time required for farm and
wholesale level price increases and decreases to be passed through to the
retail level. Hansen et al. (1994), using national aggregate price data
for whole milk, estimated that it took 3 months after the wholesale price
increased for the retail milk price to increase, but that it took 30
months for the retail price to adjust to wholesale price decreases. In
another study using national data, Wang (2003), using a structural model,
found that, for farm price increases, retail prices adjusted more quickly
in the first month after the increase and more slowly in subsequent
periods. Conversely, for farm price decreases, the speed of price
adjustment was slower in the initial month and increased in the following
months, implying speed of adjustment asymmetry.

Nearly all of the studies of regional or metropolitan price transmission
found asymmetry in timing-the price adjustment process for price decreases
much exceeded that for farm price increases. Carman (1998) found a 1-month
lag for price decreases and no lag for price increases in the California
markets. In a later study, Carman and Sexton (forthcoming 2005) found that
for the majority of cities they analyzed, the time lags estimated for
price decreases generally exceeded those for price increases. For the four
types of fluid milk, Carman and Sexton found that farm price decreases
generally took from 1 to 3 months to be transmitted to the retail level,
while price increases took no more than 1 month. In the California
markets, the authors found that, in general, retail prices responded more
quickly to farm price increases than to decreases. Lass et al. (2001)
found that for the Boston and Hartford markets, retail price adjustments
to rising farm prices were much more rapid than similar adjustments to
falling farm prices. Lass (2004) also found evidence of slower price
transmission to the retail level when farm prices were falling. For
markets in the Northeast, Frigon et al. (1999) reported short-run
asymmetry in price adjustment for several markets: price adjustment was
complete in Upstate New York after 2 months and in New York City and the
Northeast United States after 3 months. The authors concluded that
short-run asymmetry seemed to be milder in Upstate New York, because it
lasted for only 2 months. In the long run, after 4 months, the Northeast
United States, Upstate New York, and New York City (after enactment of the
gouging law) markets had fully adjusted, with only the New York City
(before enactment of the gouging law) market not fully adjusting.

                                  Appendix VI
                   Economic Studies of Price Transmission in
                           the U.S. Fluid Milk Market

Results for Retail-to-Farm Price Transmission and Price Transmission
Asymmetry

Using data from 1971 through 1991, Emerick (1994) tested for causality in
fluid milk pricing between the farm and retail levels and found that it
was bidirectional. That is, the author found that for these data, in
addition to farm price changes affecting retail prices, retail price
changes also affect prices at the farm level. However, this specification
resulted in some parameter values for retail-to-farm price transmission
that were inconsistent with economic expectations. In a later study, Wang
(2003) estimated results for retail-to-farm price transmission. In the
short run, which was estimated to be 1 month, the author found that price
increases were immediately passed on to the farm level with a level of
transmission of 94 percent, while price decreases were passed on at a
level of 2 percent. However, in the longer run, which was estimated as 3
months, price increases were passed through at a level of 40 percent,
while price decreases were passed through at a level of 34 percent. In the
long-run specification, neither increases nor decreases in price were
fully passed through. Thus, while the author found price asymmetry in the
short run, she found price symmetry in the long run for retail-to-farm
level price changes. As in Wang's farm-to-retail analysis, increases in
the retail price were passed through nearly fully in the initial month,
and then decreased substantially. However, although decreases in the
retail price were not passed on initially, they were passed on at an
increasing rate in the following months.

Possible Causes of Asymmetry in the Extent and Speed of Price Transmission

Even fewer economic studies have provided evidence on what causes price
transmission and price transmission asymmetry. For the U.S. fluid milk
market in particular, we found few studies that examined factors affecting
the extent of price transmission and price transmission asymmetry. While
two major explanations are cited in the economic literature as central to
explaining price transmission and transmission asymmetry- noncompetitive
markets and adjustment costs-there are several others cited, including the
role of government policies, spatial market competition, substitution in
processing technology, asymmetric information, economies of scale, and
differentiated products. However, only the presence of noncompetitive
markets and the effects of government policies were examined in the
studies of price transmission that we identified.

                                  Appendix VI
                   Economic Studies of Price Transmission in
                           the U.S. Fluid Milk Market

Noncompetitive Markets	In general, economic research has found that a
higher degree of market power can reduce the degree of price
transmission.11 Particularly relevant to the fluid milk market,
researchers have also shown that the number of vertical stages and the
extent to which a market varies from the competitive norm both influence
the degree of price pass through.12 Of the 14 studies that we examined,
only 5 explored the role of competition in combination with the degree of
price transmission and price transmission asymmetry in the milk marketing
chain. In particular, 2 of these studies examined market power stemming
from product differentiation of different milk types. The evidence from
these studies is somewhat mixed. While 1 study did not find a linkage
between market concentration and price transmission, other studies using a
variety of methods did find evidence of either a lack of price
transmission or price transmission asymmetry in markets that also
possessed a degree of market power.

Carman and Sexton (forthcoming 2005), using multiple analytical
techniques, found that fluid milk markets in the Western United States
that displayed noncompetitive pricing also tended to lack price
transmission and show price asymmetry. Using monthly data from 1999
through 2003, the authors (1) analyzed the effects of horizontal
differentiation among fluid milk types by ranking milk with different fat
contents for different markets based on the costs that would be predicted
under perfect competition, (2) performed correlation analysis between
changes in the monthly farm and retail prices of milk with different fat
contents, with a lack of correlation indicating the exercise of market
power, and (3) analyzed price transmission, along with the estimated price
transmission coefficients, to determine competition in the market.

In the first analysis, rank values for all milk types did not conform with
price expectations, with the exception of whole milk for all the months in
Seattle. Moreover, except in Portland, they found that the rankings of
retail milk prices for whole, skim, 1 percent, and 2 percent milk provided

11S. McCorriston, C. W. Morgan, and A. J. Rayner, "Processing Technology,
Market Power, and Price Transmission," Journal of Agricultural Economics,
Vol. 49, No. 2, 1998, 185-201. Steve McCorriston, "Why Should Imperfect
Competition Matter to Agricultural Economists?" European Review of
Agricultural Economists, Vol. 29, No. 3, 2002, 349-371.

12Steve McCorriston and Ian M. Sheldon, "The Effect of Vertical Markets on
Trade Policy Reform." Working paper, September 1994. Also, S. M.
McCorriston and I. M. Sheldon, "Agricultural Policy Reform in
Successive-Oligopolistic Markets: A General Framework," University of
Exeter and Ohio State University, December 1995.

Appendix VI
Economic Studies of Price Transmission in
the U.S. Fluid Milk Market

evidence of not being based on costs, as would be expected in perfect
competition. For the price correlations, the results indicated that only a
few product pairs in the nine markets have a high degree of
interdependence, as one would expect for close substitutes. Low
correlations ranging from nearly complete independence to moderate
independence for at least one pair of products were evident in each
market. For instance, retail price changes for skim milk appeared
independent of other milk prices in Sacramento, Seattle, Portland, Salt
Lake City, and Denver. The authors explained that these correlations all
indicated pricing that was inconsistent with competitive pricing.

For the price transmission analysis, the estimated results differed among
the California markets, depending on the city and type of milk. For
instance, in certain California markets, such as Los Angeles and San
Diego, farm price decreases lagged farm price increases by 2 to 3 months,
depending on the product, indicating price asymmetry in timing. Some of
these price coefficients were consistent with competitive pricing and
others were not. However, price transmission estimates for other
metropolitan regions in the West (Seattle, Portland, Phoenix, Salt Lake
City, and Denver) provided stronger evidence of noncompetitive pricing,
and some also indicated price adjustment asymmetry, such as Salt Lake City
and Phoenix. For these markets, only 3 of the 40 estimated price
transmission coefficients were consistent with perfect competition.

Using another model of horizontal product differentiation, a subsequent
study by Sexton, Xia, and Carman (2004) econometrically estimated the
timing of fluid milk price transmission and tested for market power for
four California and five non-California cities from 1999 to 2003. While
the results were somewhat mixed, hypothesis tests for the cities
indicating oligopoly or monopoly scenarios also displayed more gradual
price transmission results than those indicating more competitive
scenarios, suggesting a link between noncompetitive market structures and
a lack of price transmission.

Chidmi et al. (2004) estimated price transmission and market power for the
Boston fluid milk market using the New Empirical Industrial Organization
approach and data from 1996 through 2000. The empirical results of the
model, in particular the conjectural variation elasticity, suggest that
participants in this market may possess market power and that

                                  Appendix VI
                   Economic Studies of Price Transmission in
                           the U.S. Fluid Milk Market

supermarkets do not ignore each other's actions.13 Although the model did
not account for speed of adjustment, the authors estimated a price
transmission level of 68 percent, suggesting that market power is
associated with incomplete price transmission. A study by Frigon et al.
(1999) includes a measure of market power, the four-firm concentration
ratio, in its model of price transmission for Upstate New York and New
York City. However, this variable did not prove to be significant. Later,
in a similar study (Romain et al., 2002), the authors explain that their
results of price asymmetry prior to 1991 in New York City were evidence
that middlemen in the fluid milk market were exercising market power prior
to the price-gouging law. The authors found that price asymmetry decreased
after the law went into effect. They acknowledged, however, that to
rigorously address the issue of a noncompetitive market, an alternative
market power model would have to be developed.

Government Policies	Two studies looked at the effects of national
government intervention on price transmission. Emerick (1994) and Wang
(2003) both examined the question of whether changes in dairy policy,
especially the reduction in the dairy price support level that began in
the mid-1980s, had changed the nature of price transmission for dairy
products. Both authors basically came to the same conclusions. Emerick
noted that asymmetry is more likely to have occurred since 1988, adding
that the greater price volatility may have caused some difficulties for
retailers and wholesalers in determining the "appropriate" price. Wang,
using additional data through 1997, found that reductions in the price
support level tend to have a large impact on the fluid milk and nonfat dry
milk price transmission relationships. In the fluid milk market, the
farm-to-retail price transmission process became asymmetric, with the
greater price volatility in the post1988 period. While the degree of price
transmission increased for both increases and decreases in price, it
increased proportionately more for increases than for decreases.

Six of the studies examined price transmission in conjunction with other
state and federal policies and programs, such as the New York pricegouging
law and the Northeast Interstate Dairy Compact. Four studies, Lass et al.
(2001), Lass (2004), Chidmi et al. (2004), and Dhar and Cotterill

13For this model, the conjectural variation elasticity, a measure of
market conduct and firm interdependence within the market, was
significantly different from zero at the 5 percent level.

Appendix VI
Economic Studies of Price Transmission in
the U.S. Fluid Milk Market

(2002), estimated price transmission while the Northeast Interstate Dairy
Compact was in effect. Lass et al. found that processors and/or retailers
did not fully pass through their price increases and, in fact, may have
absorbed part of the cost of the Compact's over-order premium. In his 2004
study, Lass explained that the greater variation in farm prices that
occurs without the Compact would actually lead to higher retail prices
because of the larger estimated impacts on retail prices of increasing
farm prices than decreasing farm prices. Dhar and Cotterill disagreed and
contended in their study that the risk reduction benefit from the Compact
was completely overpowered by a shift toward tacit collusion in the
post-Compact period. In the studies of New York markets (Frigon et al.,
1999, and Romain et al., 2002) that looked at the effect of the
price-gouging law, researchers found that after the law took effect, price
asymmetry was not present or was present at much lower levels.

Appendix VII

                    Effects of Recent Federal Dairy Program
                     Changes and Alternative Policy Options

Recent changes in federal dairy programs vary in their effects on policy
considerations that we identified, such as farm income, milk production,
federal costs, price volatility, economic efficiency, and consumer
prices.1 A number of options have been proposed or discussed to further
modify existing programs or introduce alternative policies, all of which
could affect these policy considerations in different ways. The likely
effects of these program modifications or alternative policies are
influenced by prevailing conditions, such as high and low dairy prices,
and may be different in the short and long terms.

Recent Changes in Federal Dairy Programs Vary in Their Effects on Policy
Considerations We Identified

Since 2000, three major changes have taken place in federal dairy
programs. First, in response to legislative requirements, the U.S.
Department of Agriculture (USDA) reformed the federal milk marketing order
(FMMO) system. Second, USDA adjusted the relative purchase prices of
butter and nonfat dry milk under the price support program. Finally,
Congress authorized and USDA established the Milk Income Loss Contract
(MILC) program. These changes had mixed effects on the policy
considerations included in our analysis. Reforms to the FMMO system had
mixed effects on farm income, depending on the geographic location of the
farmer, while the overall effects on all farmers are less clear. Because
of their effect on fluid milk prices, changes in the price support program
tended to reduce the level of support for farm income and reduce federal
costs, but increase economic efficiency. Introduction of the MILC program
typically had the opposite effects, while maintaining production.

Federal Milk Marketing Order Reforms

In carrying out requirements in the Federal Agriculture Improvement and
Reform Act of 1996 to reform FMMOs, USDA conducted extensive research and
held public hearings. Agricultural Marketing Service (AMS) officials
indicated that as a result of this process USDA implemented reforms to the
FMMO system in January 2000 that were consistent with the findings of its
research. Its major reforms included

1There can be different kinds of economic efficiency effects. Government
policies may be more or less economically efficient depending upon the
extent to which they prevent the transmission of market price signals and
lead to a misallocation of resources into excess production. Policies may
also be more or less efficient depending upon the extent to which they
affect the distribution of production between farmers with high or low
costs of production.

Appendix VII Effects of Recent Federal Dairy Program Changes and
Alternative Policy Options

o  consolidating the number of marketing orders from more than 30 to 11;2

o 	changing the classified pricing structure by creating a new class for
manufactured milk products, Class IV, with the "higher of" the advanced3
Class III or Class IV skim milk values as the basis-or mover4-for Class I
prices; reducing the lag between the Class I and Class III and IV price
announcements; and establishing a fixed differential of $0.70 per
hundredweight to be added to the advanced Class IV skim milk value in
determining the price to be paid for milk used in Class II products;

o  introducing a new product formula pricing system; and

o  relaxing restrictions on pooling milk in some marketing orders.5

2The Federal Agriculture Improvement and Reform Act of 1996 required USDA
to consolidate the number of marketing orders. Two marketing orders were
discontinued between 1996 and 1999, so 31 orders were actually
consolidated.

3Basing Class I and II prices on the advanced Class III and IV skim milk
values ensures that minimum prices of raw milk used in these products will
be known in the month preceding the month to which they apply.

4The mover links the dairy price support program to the FMMO system, so
that changes in one program are reflected in the other. The dairy price
support program maintains the prices of manufactured products throughout
the United States regardless of whether a particular area is part of the
FMMO system. By using these manufactured product prices as the basis for
other milk class prices, the FMMO system aims to ensure that its minimum
prices will reflect the level of support provided by the price support
program.

5In addition to these major reforms, USDA made other changes, including
modifying the system of differentials used to determine Class I prices.
USDA increased Class I differentials in 21 markets, ranging from $0.01 per
hundredweight in New England (Boston) and New York/New Jersey (New York
City), to $0.50 per hundredweight in the Upper Midwest (Minneapolis).
Class I differentials in 4 markets were not changed. Class I differentials
in 8 markets were reduced, ranging from $0.04 per hundredweight in the
Ohio Valley (Columbus) to $0.18 in Eastern Colorado (Denver). In general,
these changes retained the existing Class I pricing surface in markets
east of the Rocky Mountains. USDA originally proposed an alternative set
of differentials, but the Consolidated Appropriations Act, 2000 required
adoption of the differentials noted above.

  Appendix VII Effects of Recent Federal Dairy Program Changes and Alternative
                                 Policy Options

USDA implemented additional reforms to the classified pricing system in
April 2003 that modified aspects of the Class III and IV pricing
formulas.6

Federal Order Consolidation	In response to the legislative requirement,
USDA reduced the number of FMMOs to 11, which were typically combinations
of pre-existing orders.7 For example, the Central Order is a combination
of several smaller marketing orders in the central part of the United
States. According to USDA's final regulatory impact analysis for the order
reforms, these consolidation decisions were based on structural factors
such as milk movement, the number of market participants, and natural
boundaries.8

USDA officials and other dairy experts told us that nationally, the prices
received by farmers for their raw milk did not change much as a result of
FMMO consolidation. One academic study reported that order consolidation
probably increased the economic efficiency of the FMMO system by more
closely aligning areas where raw milk is marketed by dairy farmers with
areas where it is distributed as fluid milk products. Additionally, the
study noted that consolidation helped to reduce the amount of market
distortion created by order regulation.

However, the magnitude of the effects on farm income varied among orders
because, in some cases, the consolidation combined orders that had
substantially different raw milk utilization rates for the manufactured
and fluid products in the various milk classes, particularly Class I
(fluid milk). As a result, some dairy farmers experienced higher or lower
utilization of their raw milk in Class I products than they had in the
past. Changes in utilization rates are significant because farmers receive
a blend price for their milk based on the utilization rates for the
different milk classes within an order; thus, farmers in orders where
Class I utilization rates increased generally saw their incomes increase,
while farmers in orders where the Class I utilization rates decreased
generally had their incomes reduced.

6The Consolidated Appropriations Act, 2000 also required USDA to
reconsider the Class III and IV pricing formulas that were implemented
with the 2000 reforms. USDA conducted a hearing on the issue in May 2000.
The reforms implemented in April 2003 stemmed from the changes USDA
proposed as a result of this hearing.

7In some cases the new orders included areas previously unregulated by the
FMMO system.

8U.S. Department of Agriculture, Federal Milk Marketing Order Reform: New
England et al., Final Decision, Regulatory Impact Analysis, Agricultural
Marketing Service, Dairy Programs, March 1999.

Appendix VII Effects of Recent Federal Dairy Program Changes and
Alternative Policy Options

The changes in utilization rates associated with FMMO consolidation were
particularly evident in the Western Order. When USDA created the Western
Order, it combined the Great Basin and the Southwestern Idaho-Eastern
Oregon Orders. These orders had substantially different Class I
utilization rates. In 1999, the Great Basin Order had a Class I
utilization rate of 51 percent, while the Southwestern Idaho-Eastern
Oregon Order had a Class I utilization rate of 8 percent. When these
orders were combined into the Western Order, the resulting Class I
utilization rate was estimated to be about 23 percent, lowering income for
the farmers in the Great Basin Order who had previously received much
higher blend prices. To address this and other concerns, Dairy Farmers of
America, a cooperative representing a number of farmers in the Western
Order, requested that USDA hold a hearing to reform the order's
provisions. USDA made some changes based on the concerns presented at the
hearing; however, the revised order provisions did not receive the
two-thirds approval necessary to be adopted, and USDA terminated the order
as of April 1, 2004, stating that the continuation of the existing Western
Order would not be in conformance with declared policy.9

Elimination of the Western Order has raised concerns that increased
amounts of Idaho milk, which had been pooled on the Western Order, would
be pooled on the Upper Midwest Order. Based on past experience, this would
reduce the Class I utilization rate and lower the blend price for Upper
Midwest farmers. However, dairy experts had mixed views on whether
additional orders would be terminated. In particular, one industry expert
noted that it remains unclear whether farmers in the former Western Order
will be able to receive higher prices for their milk without their order.
Some of these farmers, particularly those that had been in the Great Basin
Order, could benefit by not having to pool their Class I milk. On the
other hand, one source stated that these farmers could face increased
shipping requirements to pool their milk on a remaining order.
Additionally, some experts stated that without FMMOs farmers and
cooperatives do not

9After USDA holds a hearing, it may decide to issue proposed amended
regulations for notice and comment. At the end of the administrative
process, USDA issues a final rule, and those covered by the order must
vote to adopt the order in its amended form. The amended order must be
approved by at least two-thirds of the affected dairy farmers, or dairy
farmers who produce at least two-thirds of the milk produced in that
order. In addition, the amended order will not become effective until the
handlers of at least 50 percent of the milk covered by the order have
signed a marketing agreement, or until a hearing has been held and the
Secretary of Agriculture makes particular determinations that the
marketing order should nonetheless be effective.

  Appendix VII Effects of Recent Federal Dairy Program Changes and Alternative
                                 Policy Options

have the market power to obtain high prices for their raw milk in
negotiations with processors.

Classified Pricing Structure FMMO reform changed the structure of the
classified pricing system by

Changes

creating a new Class IV, representing the minimum price that processors
pay for raw milk used in butter, nonfat dry milk, and other dry milk
powders.10 Additionally, the new mover of Class I prices became the
"higher of" the advanced Class III or Class IV skim milk values. Use of
the "higher of" mover was intended to enable fluid milk processors to
attract milk from butter, nonfat dry milk, and cheese processors by
helping to ensure that the blend price would exceed both the Class III and
IV prices. USDA also reduced the lag period-the time between when the
Class I price is announced and the Class III and IV prices are
announced-from approximately 8 weeks, to 6 weeks. Class I prices are
announced in the month preceding the month to which they apply, based on
the "higher of" the advanced Class III and IV skim milk values.11 However,
the Class III and IV prices that determine the price of raw milk used to
manufacture these products are not announced until the Friday on or before
the 5th of the month following the month to which they apply.
Consequently, there is a 6week lag between these two price announcements.
Further, USDA established that the minimum prices paid for skim milk used
in Class II products would be the advanced Class IV skim milk and
butterfat values, plus a fixed differential of $0.70 per hundredweight.12
In its March 1999 regulatory impact analysis, USDA concluded that these
changes would help to eliminate situations in which prices of milk used in
manufactured products rise above the price of milk used in fluid milk
products and thus make the Class I mover more representative of current
market conditions.

10Over time, the number of classes has varied by order, from at least two,
to seven or eight. Class I use has consistently been defined as the milk
going into fluid milk products. Class II includes soft manufactured
products such as ice cream and yogurt, Class III includes cheese, and
Class IV includes butter and nonfat dry milk.

11Advanced values are derived from formulas that use weighted average
prices of butter, cheese, dry whey, and nonfat dry milk reported by the
National Agricultural Statistics Service on the Friday on or before the
23rd of the month preceding the month to which the prices apply. There is
a 1-week lag in reporting by the Service, so the advanced values are
usually based on the commodity prices for the first two weeks of the
preceding month.

12Class II butterfat is priced by adding $0.007 per pound to the monthly
Class III/IV butterfat price. Consequently, Class II skim milk is advanced
priced but Class II butterfat is not.

Appendix VII Effects of Recent Federal Dairy Program Changes and
Alternative Policy Options

Academic researchers and an industry official indicated that the creation
of Class IV continued disincentives that were present prior to the 2000
reforms to shift milk to its highest-valued use. Previously, separate
minimum prices were established for raw milk used in manufactured products
that are now included in the new Class IV. Raw milk used in the production
of butter was priced under Class III, which also included cheese and other
products. However, nonfat dry milk was priced in a separate Class III-A.
According to AMS officials, the development of Class III-A was necessary
because manufacturers were unable to sell nonfat dry milk at market prices
that would allow them to pay the Class III minimum price for their raw
milk. They noted that for a classified pricing system to work, the minimum
class prices must be below the market clearing prices for products
produced with that raw milk (taking into account the cost of other inputs
to these products). However, one study found that the creation of Class IV
institutionalized separate pricing for nonfat dry milk.13 By separating
out the price for nonfat dry milk (the lowest-valued use), the classified
pricing system might maintain production of nonfat dry milk even when
market signals indicate that raw milk should be used to manufacture cheese
as the higher-valued use.

Additionally, a 2004 study sponsored by the American Farm Bureau
Foundation for Agriculture (American Farm Bureau) reported that creating a
separate Class IV and then basing Class I prices on the "higher of" the
advanced Class III or IV skim milk values, has reduced the influence that
cheese prices traditionally had over other prices in the FMMO system, and
thus partially isolated Class I prices from market forces.14 For example,
in every month from January 2000 through July 2001, advanced Class IV skim
milk values were higher than advanced Class III skim milk values. However,
as of 2000, utilization of milk for Class IV products across all federal
orders averaged 7 to 8 percent, while Class III products accounted for
about 45 percent of milk utilization.15 According to the American Farm
Bureau study, without the advanced Class III skim milk value as the mover

13Bob Cropp and Ed Jesse, "The Butter-Powder Tilt," Marketing and Policy
Briefing Paper Number 72, Department of Agricultural and Applied
Economics, University of Wisconsin- Madison, June 2001.

14David Anderson, Bob Cropp, Wilson Gray, Joe Outlaw, and Mark Stephenson,
"Milk Pricing Policy Options and Consequences," an Analysis for the
American Farm Bureau Foundation for Agriculture, February 2004.

15According to a USDA official, Class IV use averages about 12 percent
across all federal orders and is much higher than 12 percent in several
orders.

  Appendix VII Effects of Recent Federal Dairy Program Changes and Alternative
                                 Policy Options

for Class I prices, when the Class IV price exceeds the Class III price,
similar price signals are no longer received by farmers in relatively high
Class I utilization markets and in high Class III utilization markets.
This difference occurs because during these times farmers in high Class I
utilization markets are receiving their price signals based on the high
Class IV prices, which are heavily influenced by the price support program
during periods of excess production and low manufacturing product prices.
Therefore, farmers in high Class I utilization areas receive higher farm
prices than would otherwise be the case, and higher prices encourage
increased production by these farmers. However, the higher production
levels of these farmers puts downward pressure on the Class III prices and
causes regional inequities in farm income.

Furthermore, because Class I prices are now more closely related to the
level at which the price support program sustains nonfat dry milk prices,
proposed changes to the price support program have become much more
controversial. Prior to the 2000 FMMO reforms, Class I prices were based
on the Class III price, which, as noted previously, did not include nonfat
dry milk. However, with the 2000 reforms, the level of support provided by
the price support program for nonfat dry milk prices directly influences
the Class IV price. During periods when the Class IV price is higher than
the Class III price, changing the price support program in such a way that
Class IV prices are reduced will cause the Class I price to similarly
fall, thus having a greater impact on the overall blend prices received by
farmers.

Introduction of a New Product As part of FMMO reform, USDA introduced a
new product formula pricing

Formula Pricing System	system that established minimum prices for raw milk
based on milk component values for butterfat, protein, nonfat solids, and
other solids. These values are derived from the wholesale prices of
cheddar cheese, butter, nonfat dry milk, and dry whey as announced in
weekly surveys conducted by the National Agricultural Statistics
Service.16 The minimum prices also factor in allowances based on estimates
of manufacturing costs for these products and product yield factors
representing the amount of a particular product that can be manufactured
from specified quantities of the underlying components. Seven of the 11
orders (primarily the Northern orders) adopted the new product formula
pricing system, while the other

16Whey is the water and solid components of milk that remain after the
curd is removed in cheese-making. It contains about 93.5 percent water and
6.5 percent lactose, protein, minerals, enzymes, water-soluble vitamins,
and traces of fat.

Appendix VII Effects of Recent Federal Dairy Program Changes and
Alternative Policy Options

four orders (primarily the Southern orders) use a pricing system that
bases milk prices on skim milk and butterfat.17

During much of the time that classified pricing has been part of the
federal order system, the formulas used to set minimum prices paid to
farmers were based on competitive pay prices. The pay price was known as
the Minnesota-Wisconsin price, and it represented the results of state
surveys of competitive market prices for Grade B milk paid by
manufacturing plants in Minnesota and Wisconsin. However, with a reduction
in Grade B milk production, this milk was very thinly traded and the
pricing series became less representative of the value of Grade A milk
used for manufacturing.18

According to a number of dairy experts, the change from a competitive pay
to a product formula pricing system that incorporates fixed manufacturing
allowances has enhanced the effects of price volatility on dairy
farmers.19 As noted in appendix V, there are a variety of input costs to
the manufacturing process, including labor, energy, and capital. With
product formula pricing, manufacturing allowances, which are supposed to
compensate for these other input costs, and product yield factors are
fixed. To the extent that changes in these other input prices are
reflected in the prices at which manufacturers sell their products, fixed
manufacturing allowances will allow changes in other input costs to more
readily affect the minimum raw milk prices paid to farmers.

17Raw milk quality is also factored into the pricing formulas in all but
two orders.

18As a result of declining Grade B milk production, the
Minnesota-Wisconsin price was replaced by the basic formula price in the
mid-1990s. The basic formula price established the minimum prices for raw
milk by updating the Minnesota-Wisconsin competitive pay price series with
wholesale product price information on the value of cheese, butter, and
nonfat dry milk. This system was in place from June 1995 through December
1999.

19This is not to say that the FMMO classified pricing system is causing
additional volatility in milk prices. As noted by AMS officials, an
historical review of milk prices shows that price volatility existed well
before the product pricing formulas were introduced in 2000. AMS officials
indicated that there are several causes of price volatility including: the
level of support provided by the price support program, exposure to
international trade, and the responses to changes in prices by farmers and
consumers. However, according to a number of dairy experts, the fixed
manufacturing allowances in the new product formula pricing system have
more readily transmitted price volatility within the classified pricing
system and thus enhanced its effects.

Appendix VII Effects of Recent Federal Dairy Program Changes and
Alternative Policy Options

Dairy experts also indicated that the fixed manufacturing allowances in
the product pricing formulas reduced economic efficiency by reflecting raw
milk supply and demand conditions less clearly. Moreover, one large
processor stated that the manufacturing allowances in the pricing formulas
are too low and do not adequately represent the costs of manufacturing.20
Regardless of the market price of cheese, butter, or nonfat dry milk, the
fixed manufacturing allowances provide manufacturing plants with the same
net returns from 100 pounds of raw milk. Therefore, when market conditions
reflect higher prices for one of these products, relative to the others,
manufacturers have less of an incentive to shift production to the
higher-valued use because any gains they might have realized from selling
a higher-priced product would be negated by the fact that their
manufacturing allowance is fixed.21 The 2004 American Farm Bureau study
noted that prior to the introduction of the new product formula pricing
system, manufacturers that produced butter, cheese, and nonfat dry milk
competed more aggressively for raw milk. The study found that if the
prices of nonfat dry milk and butter, for example, were depressed relative
to cheese prices, cheese manufacturers would attract milk away from the
manufacturers of these other products. Therefore, raw milk would more
readily move to its highest-valued use.22

Further, some dairy experts noted that the additional volatility
introduced by the fixed manufacturing allowances in the new product
formula pricing system, when combined with the disincentives these
allowances and separate manufacturing classes create against shifting milk
to its highestvalued use, might have contributed to negative producer
price

20AMS officials indicated that USDA has been working with researchers at
Cornell University to collect manufacturing plant costs, which these
officials believe will assist them in establishing the proper level of
manufacturing allowances should the dairy industry propose a change.

21Other factors that could affect the incentive to shift milk to a
higher-valued use include transportation costs and changes in processing
technology. In addition, when cooperatives own the capacity to produce
butter, cheese, and nonfat dry milk, they may still have an incentive to
shift milk to the higher-valued use despite fixed manufacturing
allowances, in order to provide greater returns to their members.

22One source also noted that by not providing incentives to shift milk to
its highest-valued use, the product formula pricing system could
discourage the development of innovative dairy products.

Appendix VII Effects of Recent Federal Dairy Program Changes and
Alternative Policy Options

differentials23 and de-pooling.24 Negative producer price differentials
can occur because with the 6-week lag between the Class I and Class III
and IV price announcements, rapid increases in the manufactured product
prices from which Class III and IV prices are derived can raise these
prices above the Class I price.25 USDA officials noted that the change to
a "higher of" mover for Class I prices and the reduction of the lag period
were designed to reduce the frequency of negative producer price
differentials. However, to the extent that the fixed manufacturing
allowances have introduced additional volatility into the pricing system,
and the disincentives created by these fixed manufacturing allowances and
separate manufacturing classes have prevented raw milk supplies from
moving to their highestvalued use, negative producer price differentials
and de-pooling have continued.

During times when the producer price differential is negative, some
processors of manufactured products who normally receive a draw from their
federal order pool to pay farmers instead have to pay into the pool. In
such circumstances, many of these processors choose to de-pool because by
doing so, they gain a competitive advantage over those that remain and
have to pay into the pool.26 One study on FMMO pooling issues reported
that since June 2003, negative producer price differentials and de-pooling
have become more common.27 The study noted that the producer price
differential in the Upper Midwest Order was negative from July through
November 2003 and reached a record low level in April 2004 of $4.11 per

23The producer price differential represents the difference between the
value of milk established by the classified pricing system as paid by
processors and the weighted average value of milk as paid to farmers for
milk components (butterfat, protein, and other solids). The producer price
differential is calculated by multiplying the volume of each class of milk
used in the order by its announced price to obtain a total classified
value of milk for that order, and then subtracting the total value of milk
paid to farmers for butterfat, protein, and other solids on a per
hundredweight basis.

24This presumes that minimum prices are the effective prices.

25Rapidly rising Class IV prices are a remote possibility at this time due
to the surplus of nonfat dry milk.

26Processors may remain in a federal order pool despite negative producer
price differentials for a variety of reasons. For example, Class I
processors are not allowed to depool under FMMO regulations.

27Bob Cropp and Ed Jesse, "Federal Milk Marketing Order Pooling,
Depooling, and Distant Pooling: Issues and Impacts," Marketing and Policy
Briefing Paper Number 85, Department of Agricultural and Applied
Economics, University of Wisconsin-Madison, June 2004.

  Appendix VII Effects of Recent Federal Dairy Program Changes and Alternative
                                 Policy Options

hundredweight. For example, cheese prices in the Upper Midwest Order began
to rise sharply in mid-July 2003; thus the advanced Class III skim milk
value that served as the Class I mover for August did not include these
higher prices. However, the Class III price that was announced in August
did include these higher prices, creating a negative producer price
differential. As a result, a number of cheese processors de-pooled in
August, reducing the order's Class III utilization, which is usually
around 75 to 77 percent, to just 8.4 percent. Nationally, negative
producer price differentials were reported for this month in the 7 FMMOs
that used the new product formula pricing system. For the 11 FMMOs
existing at the time, de-pooling resulted in 33 percent less milk being
pooled compared to the same month in the prior year.

De-pooling reduces the overall value of the federal order pool and
increases differences in the abilities of processors to pay for raw milk.
As a result, dairy farmers do not receive uniform prices. Farmers
marketing milk with processors who are able to de-pool may receive higher
prices and thus an increase in farm income, while farmers marketing milk
with processors who do not de-pool may receive lower prices. According to
the June 2004 University of Wisconsin study, this situation represents an
inequity and is contrary to one of the stated purposes of the FMMO system:
orderly marketing conditions.

Relaxed Pooling Provisions	In some cases, the 2000 FMMO reforms resulted
in more relaxed pooling provisions. AMS officials noted that when USDA
consolidated the marketing orders, the pre-existing orders each had its
own pooling provisions, such as minimum amounts of raw milk required to be
shipped to processing plants participating in that order's pool to qualify
for its blend price or restrictions on rejoining the pool after
de-pooling.28 AMS officials indicated that where two orders were combined
that had different pooling provisions, USDA applied the more liberal
pooling standard to the combined order to prevent farmers from being shut
out of the consolidated order pool.

28Because these provisions are established on an order-by-order basis, the
minimum shipping percentages vary by order. In federal milk orders with
relatively high Class I use, such as the Southeast and Florida Orders, the
shipping percentage requirements are higher than in orders with relatively
low Class I use, such as the Upper Midwest Order. Shipping requirements
may also vary by months of the year. In orders where milk production tends
to vary a lot seasonally, the minimum shipping requirements may be greater
during low production months.

Appendix VII Effects of Recent Federal Dairy Program Changes and
Alternative Policy Options

According to AMS reports from 2000 and 2001, relaxed pooling provisions
contributed to the pooling of more distant raw milk to receive other
orders' attractive blend prices. Pooling was easier because most of the
milk that was pooled from outside individual federal orders was not
required to actually be shipped to those orders. Therefore, distant
farmers were able to share in an order's blend price without incurring
substantial transportation costs for shipping milk. For example, under the
Upper Midwest Order's pooling provisions, an Idaho dairy cooperative could
choose to ship raw milk from some of its Idaho farmers to a processing
plant that participates in the Upper Midwest Order. All subsequent milk
deliveries of those designated farmers would be priced under the Upper
Midwest Order, even if only one day's production was actually shipped to
the participating processing plant. Other deliveries from these farmers
would stay in Idaho for processing. In 2001, USDA reported that raw milk
from California was pooled on the Central, Upper Midwest, and Western
Orders. However, most of this 4 billion pounds of milk was actually
processed in California plants that are not regulated by the federal order
system. Also during 2001, large volumes of raw milk from Minnesota and
Wisconsin were pooled on the Central, Mideast, and Northeast Orders, while
increasing amounts of raw milk from Idaho were pooled on the Upper Midwest
Order.

According to some dairy experts, the increased pooling of milk across
orders has had mixed effects on farm income. Those farmers who were able
to have their milk pooled on distant orders that had higher blend prices
received an increase in their farm income after accounting for the costs
of transporting their milk. However, farmers in the receiving orders had
their farm income reduced as milk from outside the orders reduced the
value of the pool that could be shared among farmers in the receiving
orders. The combination of more milk pooled on these orders and constant
sales of higher-valued Class I and II products decreased the weighted
average value of the orders' pools by decreasing the utilization rates of
the higher-valued classes.

In response to this loss in value, participants in some orders petitioned
USDA to hold hearings to address relaxed pooling provisions. For example,
through the hearing process, the Central and Mideast Orders tightened
their pooling provisions to control the large quantities of milk from
Minnesota and Wisconsin that were being pooled on these two orders. With
these tightened provisions, those seeking to pool milk on the Central or
Mideast Orders have to ship more milk per year or meet other requirements
to become eligible to share in the receiving orders' blend prices. A
number

  Appendix VII Effects of Recent Federal Dairy Program Changes and Alternative
                                 Policy Options

of dairy experts indicated that these changes have significantly reduced
the incentive for distant pooling on these orders.

2003 Classified Pricing Reforms	USDA made additional reforms to the
classified pricing system that went into effect in April 2003, modifying
aspects of the Class III and IV pricing formulas. The principal changes in
2003 were increasing the manufacturing allowance in the formula that
established a price for the other solids component of milk used in Class
III products; eliminating the lower bound of zero on the Class III other
solids component price;29 reducing the product yield for the nonfat solids
components of milk used in Class IV products; and altering the Class III
protein formula to prevent Class III prices from being lowered by rising
butter prices. The 2003 changes were partly the result of a court-ordered
injunction against the implementation of other changes that USDA had
proposed based on the 2000 requirement that USDA reconsider the Class III
and IV pricing formulas.

An analysis by researchers at the University of Wisconsin before
implementation of the changes indicated that the differences from many of
these changes would not be dramatic.30 However, the researchers estimated
that the changes would increase Class III prices by as much as $0.57 per
hundredweight. More specifically, the study found that the 2003 changes
would eliminate the negative effect that rising butter prices were having
on the Class III price. Under the prior protein price formula, a $0.10 per
pound increase in the butter price would lower the Class III price by
$0.04 per hundredweight. The researchers found that the revised formula
instead yields about a $0.04 per hundredweight increase in the Class III
price for a $0.10 per pound increase in the butter price. The study also
reported that the new protein price formula would make it somewhat less

29The Class III other solids component price (per pound) is calculated by
taking the difference between the dry whey price reported by the National
Agricultural Statistics Service and a manufacturing allowance of $0.159,
multiplied times a product yield factor of 1.03. The lower bound of zero,
or "snubber," was designed to prevent the price of other solids components
from being negative should the market price for dry whey fall below the
manufacturing allowance. Removing this lower bound created the possibility
that the other solids component price could actually lower the Class III
price during periods of low market prices for dry whey, which in fact,
occurred from April to June of 2003. However, because other solids
comprise a small percentage of the overall component value of Class III
milk, the effect of this change was minor.

30Bob Cropp, Brian Gould, and Ed Jesse, "Federal Milk Marketing Order
Reform: November 2002 Final Decision on Class III/IV Formulas," Marketing
and Policy Briefing Paper Number 79, Department of Agricultural and
Applied Economics, University of Wisconsin-Madison, November 2002.

  Appendix VII Effects of Recent Federal Dairy Program Changes and Alternative
                                 Policy Options

likely that the advanced Class IV skim milk value rather than the advanced
Class III skim milk value would consistently serve as the mover for Class
I prices.31 Increased Class III prices would most likely benefit farmers
in areas where cheese is an important commodity and where processors do
not typically pay premiums in excess of federal order minimum prices.

Dairy Price Support Program Adjustments

Since 2000, USDA twice adjusted-or tilted-the purchase prices of butter
and nonfat dry milk as part of its efforts to administer the dairy price
support program. The first tilt occurred in May 2001, when USDA reduced
the nonfat dry milk purchase price by approximately $0.10 per pound (to
$0.90 per pound) and increased the butter purchase price by about $0.20
per pound (to approximately $0.85 per pound). USDA adjusted the tilt again
in November 2002 by reducing the nonfat dry milk purchase price another
$0.10 per pound (to $0.80 per pound) and increasing the butter purchase
price approximately $0.20 per pound (to $1.05 per pound). USDA took these
actions because the Commodity Credit Corporation (CCC) was accumulating
large stocks of nonfat dry milk, leading to high purchase and storage
costs for USDA, as well as significant market distortions.

As a result of the 2000 FMMO reforms, the federal order class prices and
the level of support provided by the price support program for nonfat dry
milk were tied more closely. Many dairy experts noted that, subsequently,
tilts became more politically controversial because they can have a
greater negative effect on the FMMO class prices. Lowering the purchase
price of nonfat dry milk while raising the purchase price of butter
decreases the overall Class IV price when market prices for nonfat dry
milk are at the level of the purchase price and market prices for butter
are above the level of the purchase price. When the advanced Class IV skim
milk value is serving as the mover for the Class I price, this reduction
in Class IV prices also reduces Class I prices. Further, because the
advanced Class IV skim milk value serves as the basis for Class II prices,
Class II prices are similarly reduced. This scenario occurred during both
the May 2001 and November 2002 tilts. A representative of one dairy
cooperative stated that these impacts were particularly pronounced in
areas with high Class I utilization rates, such as the Northeast and
Southeast.

31However, the study noted that Class III prices would have been the
driver in only one additional month between January 2001 and November
2002.

Appendix VII Effects of Recent Federal Dairy Program Changes and
Alternative Policy Options

According to a report published by the International Trade Commission in
May 2004, estimates of the actual impacts of these tilts on farm prices
varied.32 The study presented a USDA estimate that the November 2002 tilt
reduced fiscal year 2003 average milk prices from $12.10 to $11.90 per
hundredweight. While USDA reported that this decrease lowered the amount
of raw milk produced and thus was partially offset by an increase in
butter prices from reduced production, it still led to a loss in net farm
income of $192 million.33 Alternatively, the study reported that the
National Milk Producers Federation estimated that the two tilts ultimately
lowered farm prices by $0.19 per hundredweight in 2001, $0.48 per
hundredweight in 2002, and $0.76 per hundredweight in 2003. With these
price reductions, the organization projected that farm income would fall
by $156 million in 2001, $816 million in 2002, and about $1.3 billion in
2003. Another study cited by the International Trade Commission's report
estimated that the 2002 tilt could have decreased average milk prices by
$0.16 per hundredweight, reducing production by 814 million pounds and
farm income by $371 million. However, that study also found that these
impacts varied substantially, depending upon the assumption of high or low
prices and the effects of other government programs.

While the tilts reduced farm income and raw milk production, a number of
dairy experts indicated that USDA's tilts have increased economic
efficiency and reduced federal costs associated with the dairy price
support program. Additionally, some experts noted that by maintaining
nonfat dry milk prices at artificially high levels, the price support
program was inducing surplus production of nonfat dry milk. In some cases,
nonfat dry milk was produced specifically for sale to the government at
the CCC purchase price. From the beginning of October 2000 through the end
of May 2001, the CCC purchased approximately 330 million pounds of nonfat
dry milk (more than 40 percent of national production), and government
purchase costs exceeded $340 million. Furthermore, in 2002, the CCC stocks
of nonfat dry milk were equivalent to two-thirds of domestic production
and exceeded annual domestic consumption by more than 30

32U.S. International Trade Commission, Conditions of Competition for Milk
Protein Products in the U.S. Market, Publication 3692 (Washington, D.C.:
May 2004). The U.S. International Trade Commission is an independent
federal agency. Its purpose is to assist in the administration of trade
laws and to protect U.S. industry, as well as to provide advice and
research on trade issues.

33The USDA estimate also factored in the benefits provided by the
introduction of the Milk Income Loss Contract program.

  Appendix VII Effects of Recent Federal Dairy Program Changes and Alternative
                                 Policy Options

percent. By reducing the purchase price of nonfat dry milk, USDA reduced
the incentive to produce surplus nonfat dry milk. The International Trade
Commission study reported that while production of nonfat dry milk
continued to rise between 2001 and 2002, after the second tilt production
declined by about 5 percent between 2002 and 2003. The tilts also helped
to reduce federal costs associated with purchasing and storing nonfat dry
milk.

Some sources also indicated that the tilts affected the balance of trade
in dairy products between the United States and its trade partners. The
International Trade Commission study reported that during the majority of
the period from January 1998 to November 2002, U.S. prices for nonfat dry
milk exceeded international market prices by more than $500 per metric
ton. Consequently, domestic manufacturers had an incentive to import
alternative dairy protein products such as milk protein concentrates. By
lowering the purchase price of nonfat dry milk through the tilts, USDA
decreased this incentive because domestic manufacturers could obtain
nonfat dry milk more cheaply.

Introduction of the Milk Income Loss Contract (MILC) Program

With the introduction of the MILC program in 2002, dairy farmers began
receiving payments on milk production up to 2.4 million pounds annually
when the Class I price in Boston dropped below $16.94 per hundredweight.
MILC payments are equal to 45 percent of the difference between $16.94 and
the lower Boston Class I price. From the program's inception, MILC
payments were made every month from the retroactive start date of December
2001 through August 2003 because there was an extended period of depressed
farm prices, which reached a 25-year low in early 2003. Prices temporarily
recovered from September through December 2003, so no MILC payments were
made in those months; however, payments resumed during January and
continued through April 2004. During the spring of 2004, farm milk prices
reached record highs and remained strong through the fall of 2004, so no
MILC payments were required for the remainder of fiscal year 2004.

As a result of depressed farm milk prices during 2002 and 2003, federal
costs associated with MILC payments exceeded original estimates. Based on
market conditions in March 2002, the Congressional Budget Office estimated
total federal costs of the MILC program at $963 million over the life of
the program (i.e., about 4 years). However, 1 year later, the
Congressional Budget Office revised its total cost estimate for the MILC
program to $4.2 billion. USDA distributed approximately $1.8 billion in

Appendix VII Effects of Recent Federal Dairy Program Changes and
Alternative Policy Options

MILC payments to dairy farmers in fiscal year 2003.34 Thus, the cost of
MILC through fiscal year 2003 alone exceeded the previously estimated
total costs for the entire 4-year period through 2005 by about $800
million. The Congressional Budget Office's March 2004 estimate for total
MILC program costs was $3.8 billion, somewhat lower than the 2003 estimate
because of higher farm milk prices in 2004.35

Many dairy experts indicated that by providing income support during
lowprice periods, the MILC program has helped keep some farmers,
particularly smaller farmers, in business.36 For example, some academic
experts noted that some farmers received MILC checks of about $20,000 to
$25,000, and others said that despite low prices, fewer farmers exited the
market in 2003 than in previous years. USDA officials indicated that these
payments delayed the supply response to low prices and maintained
depressed milk prices over a longer period of time. By providing direct
payments when prices were low, MILC obscured market signals that would
normally cause farmers to decrease production, and continued high levels
of production retained downward pressure on milk prices. To the extent
that these lower farm prices were passed on through the retail level,
consumers may have experienced lower prices for dairy products.

Despite this effect, dairy experts stated that smaller farmers receive a
net benefit from MILC because those with about 100 to 130 cows can have
all of their production covered under the 2.4 million-pound annual cap. In
contrast, larger farmers do not receive net benefits from MILC because the
negative farm income effects of reduced milk prices are greater than the
payments they receive under the production cap. A couple of dairy experts
noted that the break-even size, at which MILC payment benefits just offset
the negative farm income effects of prolonged low prices, is about 400
cows. Because the effects of MILC vary by producer size, they also vary

34This included retroactive payments made for production in fiscal year
2002.

35The $3.8 billion estimate comprises actual outlays of about $1.8 billion
in fiscal year 2003 and forecasts of $935 million in fiscal year 2004,
$963 million in fiscal year 2005, and $77 million in fiscal year 2006.
Some MILC payments will extend into fiscal year 2006 for milk produced in
fiscal year 2005. USDA noted that its estimate of MILC payments for fiscal
year 2004 is $300 million, much less than the Congressional Budget
Office's estimate.

36While MILC has helped some farmers to weather low price periods, USDA
officials stated that the program has not necessarily changed the
long-term strategy of farmers to exit production. When making a decision
to remain in production, farmers have to weigh a number of considerations
in addition to MILC payments, particularly costs and market conditions.

  Appendix VII Effects of Recent Federal Dairy Program Changes and Alternative
                                 Policy Options

regionally. States with many small dairy farmers, such as Pennsylvania,
Wisconsin, and Vermont, have received greater proportional benefits from
MILC. However, the MILC program has disadvantaged states with larger
producers, such as western states.

Effects of Alternative Dairy Policies Differ under Various Scenarios

A number of options have been proposed or discussed to further modify
existing programs and policies or introduce alternative ones, all of which
could affect the policy considerations we identified in different ways.
These options span a range of existing and potential federal dairy
programs and policies, including FMMOs, price supports, MILC, target price
deficiency payments, the proposed National Dairy Equity Act, trade
restrictions and export incentives, risk management, and supply
management. Current international trade agreements and ongoing
negotiations can have implications for certain of these policy options,
such as price supports, export incentives, and trade restrictions.37 The
purpose of this analysis is not to take a position for or against any of
these options or to analyze them in terms of their overall economic
impacts, but simply to discuss their likely effects on the policy
considerations we identified. The likely effects of these alternatives
sometimes differ under various scenarios, such as high or low prices, and
may be different in the short and long terms.

In general, options that increase farm income over the short term also
tend to increase milk production and thus the potential for oversupply and

37In the 1994 Uruguay Round Agreement on Agriculture of the General
Agreement on Tariffs and Trade, World Trade Organization (WTO) members,
including the United States, made commitments to improve market access,
reduce export subsidies, and limit and in some cases reduce
trade-distorting domestic agricultural supports. For example, Uruguay
Round commitments on domestic agricultural subsidies include annual
aggregate ceiling levels for certain types of direct government support,
which include price supports for dairy products. The United States and
other WTO members are currently engaged in another round of multilateral
trade negotiations that began in Doha, Qatar, in 2001, where they
reaffirmed their commitment to agricultural trade liberalization, and in
July 2004, WTO members agreed on a framework to guide negotiations on
agriculture. The U.S. negotiating position has emphasized harmonizing,
reducing, and further disciplining agricultural subsidies. In the interim,
existing Uruguay Round commitments will continue at established levels.
WTO members found to be in violation of these commitments can be subject
to retaliatory measures, such as punitive tariffs on exports.

  Appendix VII Effects of Recent Federal Dairy Program Changes and Alternative
                                 Policy Options

lower average farm prices over the long term.38 These options also tend to
be costly for the federal government during periods of low prices. In some
cases options that increase the economic efficiency of federal dairy
programs also increase price volatility because they allow clearer
transmission of market price signals.39 Further, to the extent that price
changes at the wholesale level are passed through to the retail level, a
number of options would likely have mixed effects on consumer prices
depending upon the particular product under consideration (e.g., butter,
cheese, or fluid milk). The potential impacts of the options also vary
according to the size of the producer and region of the country. In
general, options that affect farm income without respect to farm size or
cost of production could further shift production towards larger, western
farms. Production shifts toward larger farms could increase the potential
for oversupply in the market, because such farms have a greater capacity
to increase production in response to policy incentives. In some cases,
options that reduce support for farm income could have disproportionately
negative impacts on smaller farmers, who often have higher costs of
production.40

Change the Federal Milk Marketing Order Program

Dairy experts have cited a number of concerns with FMMOs, including that
with the increasing ability to transport milk products longer distances,
the differences in Class I differentials provide incentives for
overproduction in some regions; that recent revisions to the classified
pricing system enhanced the effects of price volatility on dairy farmers;
that changes in pooling restrictions increased the flow of milk between
different regions of the country, which disrupts the market; that
consolidation of the FMMOs combined some areas of the country that were
not part of the same natural "milksheds;" and that it takes too long to
change the federal order system

38Small surpluses or shortages in the short term may be the result of
market corrections and may not represent inefficiencies in the allocation
of production resources.

39The level of price volatility could influence both the supply and demand
for milk products. For example, manufacturers of products using dairy
ingredients might purchase larger quantities of these ingredients if
prices are more stable.

40To evaluate the impacts of various options on policy considerations such
as farm income, milk production, and federal costs, we compared these
options against a baseline scenario of the policies in place as of August
2004: FMMO regulations, a MILC program that is scheduled to expire at the
end of fiscal year 2005, a price support program at $9.90 per
hundredweight, a Dairy Export Incentive Program (DEIP), trade
restrictions, and milk regulatory policies in some states.

  Appendix VII Effects of Recent Federal Dairy Program Changes and Alternative
                                 Policy Options

through the USDA hearing process. A number of options have been proposed
or discussed to modify the FMMO program, including revising the classified
pricing system, making administrative changes such as tightening pooling
provisions, or eliminating FMMOs altogether. Figure 36 shows the effects
of various options to change the FMMO program over the short and long
terms.

Figure 36: Potential Effects of Options to Change FMMOs on Various Policy
                                 Considerations

Increase Decrease Mixed impacts ? No/unclear impacts

Source: GAO analysis based on interviews with dairy experts and reviews of
relevant studies.

  Appendix VII Effects of Recent Federal Dairy Program Changes and Alternative
                                 Policy Options

Note: The figure does not indicate the degree of increase or decrease for
each policy consideration as it relates to a policy option. In addition,
as discussed in this appendix, limitations affecting some policy options
could affect the degree of an option's potential impact on a policy
consideration.

Revise Classified Pricing	Academic, industry, and government sources cited
a variety of concerns with recent reforms in the classified pricing
system, including that these changes enhance the effects of price
volatility on farmers; lessen the transmission of market price signals;
reduce incentives for market participants to shift milk to its
highest-valued use; discourage innovation; and contribute to de-pooling
during periods of price volatility. A number of options exist to revise
the classified pricing system, including basing the class pricing formulas
on competitive pay prices instead of product prices, combining Class III
and IV into a single manufacturing class of milk, and changing the Class I
mover from being the "higher of" the advanced Class III or IV skim milk
values to a weighted average of these prices.

Use Competitive Pay Prices Instead of Product Formula Pricing

One option to reform the classified pricing system would be to return to a
system of competitive pay prices as the basis for the minimum prices that
manufacturers pay for raw milk. As indicated by AMS officials, the change
to a product formula pricing system has enhanced the effects of price
volatility on the prices that farmers receive for their raw milk. The AMS
officials, as well as academic and industry sources, noted that
competitive pay prices for raw milk would be a better basis for the
minimum class prices. However, the officials noted that the lack of good
data is a major challenge to developing a competitive pay system that is
more broadly based than the old Minnesota-Wisconsin price series.41

In an earlier report, we compared the concept of a competitive pay system
to a product formula pricing system, among other options.42 We found that
while a product formula pricing system would be superior to other
mechanisms in reflecting national prices of manufactured dairy products,

41With declining Grade B milk production, any alternative competitive pay
system would have to include Grade A milk. Consequently, a key challenge
in developing a good competitive pay price series is obtaining data that
are not already influenced by the FMMO classified pricing system. AMS
officials noted that during FMMO reform and consolidation, AMS
commissioned a blue-ribbon committee of academics to evaluate various milk
pricing methods. Although the committee's preference was for a competitive
pay price series, they were unable to develop or suggest a way of creating
such a series.

42GAO, Milk Pricing: New Method for Setting Farm Milk Prices Needs to Be
Developed, GAO/RCED-90-8 (Washington, D.C.: Nov. 3, 1989).

Appendix VII Effects of Recent Federal Dairy Program Changes and
Alternative Policy Options

the accuracy of price levels under the system would depend on a number of
factors, including whether the manufacturing allowances-deductions made
for the costs of manufacturing different dairy products-are accurate.
Setting accurate manufacturing allowances is difficult because individual
plants could have different cost structures. As noted earlier, we heard
from several academic, industry, and government sources that the fixed
manufacturing allowances in the current product pricing formulas have
negative impacts on the economic efficiency of the classified pricing
system by reflecting supply and demand conditions less clearly.43

On the other hand, we found that while a competitive pay system for Grade
A milk was similar to a product formula pricing system in that it would
generally reflect national prices of manufactured dairy products, it would
more readily reflect national supply and demand conditions for raw milk
used for manufacturing. Also, it would more accurately reflect competitive
pressures from the fluid milk market because Grade A milk is used to meet
shortages in areas of the country with high Class I utilization.
Furthermore, we reported that a competitive pay system would be better
than a product formula pricing system at self-adjusting automatically
because the competitive system would be based on actual reported prices.
Therefore, a competitive pay system could improve the economic efficiency
of the classified pricing system by providing clearer market price
signals.

AMS officials stated that if it were possible to obtain adequate data,
returning to a competitive pay system would introduce greater stability
into farm milk prices because basing the price formulas on competitive pay
prices allows manufacturers' margins to be set outside the federal order
system. According to the officials, when manufacturing costs increase,
manufacturers tend to decrease their margins; they then increase their
margins when costs go back down. The AMS officials also stated that to the
extent the effects of price volatility are reduced by eliminating fixed
manufacturing allowances, raw milk production would increase, holding
average milk prices constant.

43Another alternative modification to the classified pricing system would
be to raise the manufacturing allowances; however, the accuracy of any
adjustment that does not change according to market forces would
eventually decrease. Further, the problem of trying to develop an
allowance that adequately represents costs of production in an environment
of widely varying cost structures would still exist.

Appendix VII Effects of Recent Federal Dairy Program Changes and
Alternative Policy Options

Combine Class III and IV

A second option discussed by some of the academic and industry sources we
contacted would be to combine Class III and IV into a single manufacturing
class.44 While an objective of USDA's FMMO reforms, including developing
separate manufacturing classes on which to base fluid milk prices, was to
avoid situations in which the price of milk used in manufactured products
rises above the price of milk used in fluid products; as we noted earlier,
this change has muted market price signals and has reduced incentives to
move milk to its highest-valued use. However, an academic source indicated
that a potential challenge in combining Class III and IV is identifying an
appropriate formula that considers the products in an expanded class. In
addition, AMS officials cited this issue noting that one barrier would be
finding a way to price the lowest-valued use, nonfat dry milk, so that
manufacturers of this product would be able to afford to pay the minimum
class price to farmers.

A number of industry and academic sources said that combining Class III
and IV into a single manufacturing class would allow milk to move to its
highest-valued use. With a separate Class IV, processors of butter and
nonfat dry milk can pay less for their raw milk supplies under certain
market conditions, which can stimulate additional allocation of raw milk
into Class IV products. Under such conditions, by reducing these market
distortions, combining Class III and IV might help to increase the
economic efficiency of the classified pricing system.45 It could also help
to limit the decline in market prices caused by overproduction of nonfat
dry milk and thus reduce the federal costs of CCC purchases of this
commodity. Over the long term, limiting overproduction incentives will
help all farmers by maintaining higher average farm prices. However, these
benefits could be limited by constraints on the extent to which raw milk
is free to move between different uses. Some USDA officials have indicated
that manufacturing capacity for different products varies by region and
that manufacturing plants are often specialized. In addition, fixed supply
agreements within the dairy industry may not allow manufacturers

44Class I includes fluid milk products, Class II includes soft
manufactured products such as ice cream and yogurt, Class III includes
cheese, and Class IV includes butter and nonfat dry milk.

45As discussed later in this appendix, this distortion could also be
reduced by changing the price support program.

Appendix VII Effects of Recent Federal Dairy Program Changes and
Alternative Policy Options

significant freedom to shift raw milk between uses in response to price
signals in the short term.

Combining Class III and IV could have mixed impacts by region on farm
income because with a combined manufactured product price, high prices for
one particular use of milk would offset low prices for other uses in
creating a weighted average price. Utilization rates for the different
manufacturing uses vary among orders, and farmers whose average blend
prices might be higher based on utilization of their raw milk for a
highervalued use under a separate class price scenario could experience
lower average returns under a combined class price scenario if the price
of their higher-valued use were weighted down by the inclusion of the
lower-valued use. Conversely, farmers whose average blend prices might be
lower based on utilization of their raw milk for a lower-valued use under
a separate class price scenario could experience higher average returns
under a combined class price scenario. For example, in regions with higher
Class IV utilization, the loss of a separate Class IV could result in
lower average farm income during periods when Class IV prices would
otherwise be higher than Class III prices. In regions with higher Class
III utilization, however, the loss of a separate Class IV could result in
higher average farm income during periods when Class III prices would
otherwise be lower than Class IV prices.

Change Class I Mover

If Class III and IV prices are kept separate, a third option to modify
classified pricing would be to use a weighted average of Class III and IV
prices as the mover for Class I prices. This option would tie fluid milk
prices more closely to market related manufacturing prices, particularly
when prices of Class III products are depressed relative to Class IV
products. One academic study that modeled the price volatility impacts of
using a weighted average of Class III and IV prices to set Class I prices
reported that using the weighted average, rather than the "higher of" the
advanced Class III or IV skim milk values, slightly decreases the
volatility of farm prices, largely through a more substantial decrease in
the volatility of the Class I price.46 As a result of this reduced
volatility, the study estimated that the average Class I price would
decline by roughly $0.40 per

46Charles F. Nicholson and Thomas Fiddaman, "Dairy Policy and Price
Volatility." Paper presented at the 10th Annual Workshop for Dairy
Economists and Policy Analysts, Memphis, Tennessee, April 2003.

  Appendix VII Effects of Recent Federal Dairy Program Changes and Alternative
                                 Policy Options

hundredweight, causing an average farm price decline of $0.09 per
hundredweight. The study found that there would be essentially no change
in the volatility of other class prices or product prices.

These effects are likely to vary by region. As the Class I price is
expected to decrease more than the average farm price, the effects of this
option could be more significant in regions with high Class I utilization
of raw milk. The reduction in farm prices could cause a marginal downward
supply adjustment over the long term, the effects of which would also be
stronger in high Class I utilization areas. The effects of reduced prices
on farm income could be partially offset by increased MILC payments as
long that program remains in existence. Increased MILC payments would
raise federal costs, particularly during periods of low prices. However,
this option could also help USDA minimize the costs of the price support
program by making it less controversial to adjust the tilt between butter
and nonfat dry milk purchase prices because these prices would no longer
exert as great an influence on the Class I price.

AMS officials cautioned, however, that this option could increase the
likelihood that manufactured product prices would rise above blend prices,
leading to more frequent negative producer price differentials and
depooling.47 The officials noted that the purpose of implementing the
"higher of" the advanced Class III or IV skim milk values provision was to
reduce the frequency of negative producer price differentials. They said
that implementing an option that reduces blend prices makes it more likely
that the value of one of the manufacturing class prices will rise above
the blend price. The officials also indicated that consumers may benefit
from lower prices to the extent that the Class I and blend price decreases
are transmitted through the milk marketing chain. One AMS official
questioned the extent to which this option would reduce the volatility of
Class I prices, noting that since January 2000 the volatility of Class I
prices would have been about the same under a weighted average mover.

Change the Administration of Another set of options to modify FMMOs
involves changing their

FMMOs	administration. Pooling provisions could be tightened by increasing
the amount of milk that must be shipped to an order to qualify for that
order's blend price or by placing restrictions on de-pooling.
Alternatively, federal

47AMS officials noted that other alternatives, not discussed in this
appendix, that could prevent negative producer price differentials include
eliminating the advanced Class I price announcement and setting minimum
prices for all classes in advance.

Appendix VII Effects of Recent Federal Dairy Program Changes and
Alternative Policy Options

order reform could be reconsidered by splitting up some of the
consolidated orders. Finally, USDA may be able to shorten the time between
a hearing request and implementation of a final decision.

Tighten Pooling Provisions with Increased Minimum Shipment Requirements

Federal order reforms relaxed the pooling provisions of many orders, which
negatively affected some farmers by diluting their Class I utilization
rates and lowering their blend prices as increased amounts of distant milk
were pooled on their orders. In response, participants in some orders
called for hearings to tighten their pooling provisions, often by
increasing the minimum amount of milk that must be delivered to processing
plants participating in their pool to qualify for the blend price. With
the end of the Western Order, the concern that some milk formerly pooled
on that order could be pooled on the Upper Midwest Order led two groups of
dairy cooperatives operating there to request a hearing to tighten the
order's pooling provisions.48

According to some AMS officials, restricting the pooling of raw milk
through increased shipping requirements would not have a significant
national impact. However, they said there could be mixed regional effects
on farm income and production. Farmers in areas seeking to pool milk to
other orders would generally see a negative impact on their farm incomes
because they would incur greater transportation costs trying to share in
the value of another order's pool and would, therefore, pool less milk in
other orders. However, farmers in the receiving order could see an
increase in their farm income because with less milk pooled from outside
the order, their pool would retain more of its value and they would
receive higher blend prices. In each case, there could be localized
production effects depending upon whether farmers experience an increase
or a decrease in their farm income. The AMS officials stated that
tightening pooling provisions would have minimal effects on price
volatility because the tighter provisions would not change overall supply
and demand conditions. However, the officials also indicated that the
economic efficiency of the FMMO system would increase to the extent that
reducing the amount of milk pooled on distant orders would reduce the
amount of money spent on transporting milk. Because impacts on national
production levels are likely to be limited, impacts on federal costs would
be minimal.

48This hearing took place in August 2004.

Appendix VII Effects of Recent Federal Dairy Program Changes and
Alternative Policy Options

Tighten Pooling Provisions with Restrictions on De-Pooling

A second option for tightening federal order pooling provisions is to
place additional restrictions on those who choose to de-pool. As noted
earlier, price volatility leading to negative producer price differentials
and depooling can negatively affect those who remain in a federal order
pool because the overall value of the pool is reduced.49 Some orders
restrict depooling by preventing milk handlers who choose to de-pool from
re-pooling for a specific period of time.50 One such restriction recently
proposed by cooperatives in the Upper Midwest Order would limit a
processor's pooled milk in any month to a specified percentage of that
processor's pooled milk in the previous month. Under that restriction, if
a processor partially depooled in one month, it could only partially
re-pool in the subsequent month. If it fully de-pooled, it would have to
wait a month before it could re-pool.

Restricting de-pooling would have mixed effects on farm income. Because
de-pooling allows some processors to pay farmers higher prices, some
farmers would be harmed if de-pooling were restricted. Conversely, those
farmers who are harmed by de-pooling could benefit if more of the pool
value were retained during periods of volatile prices. While nonuniform
farm prices do not help to achieve orderly marketing, restricting
de-pooling could make this problem worse if it encourages some processors
to leave the order system permanently. In that case, the reserve supply of
milk for fluid production would shrink, and orders would have to increase
minimum shipping requirements for remaining pooled processors and dairy
cooperatives. If restricted de-pooling actually caused fewer processors to
be associated with the federal order system, volatility in fluid milk
prices could increase because with less milk available in reserve to
supply the Class I market, seasonal or episodic fluctuations in milk
supply and demand could have greater price impacts. However, an AMS
official indicated that in general processors benefit over the long term
from being pooled, and so restrictions on de-pooling would not necessarily
decrease the supply of milk available for fluid milk products. If the
supply of milk

49Changes in federal dairy policy that reduce price volatility, such as
increasing the level of support provided by the price support program,
would help to reduce instances of negative producer price differentials
and de-pooling.

50Handler is the federal order term for cooperatives, processors, or
dealers of milk who commonly purchase raw milk and sell pasteurized milk
and milk products. See 7 C.F.R. S:1000.9 for a more complete definition.

Appendix VII Effects of Recent Federal Dairy Program Changes and
Alternative Policy Options

available for fluid milk products did not decrease, then the volatility of
fluid milk prices would not increase. The effects of restricted de-pooling
on other policy considerations, such as federal costs or consumer prices,
are unclear.

Split Up Consolidated Federal Orders

A third option to revise the administration of FMMOs is to split up some
of the consolidated orders where the combination of orders with different
utilization rates of the different classes of milk has created problems.
Many industry, academic, and government sources stated that federal order
consolidation has had some significant regional effects, including the
demise of the Western Order. One source stated that because of order
consolidation, some orders have become so large that they include milk
that would not normally be in the milkshed for particular marketing areas.
This increases the potential that some farmers' blend prices will decrease
with lower utilization rates of raw milk in Class I products. To address
this problem, USDA is currently considering a proposal to split up the
Southeast Order and create a "Mississippi Valley" Order.51

Splitting up certain orders could have mixed effects on farm income. In
cases in which experts have cited problems with federal order
consolidation, the problems developed because two previously existing
orders that had largely different Class I utilization rates were combined.
With the additional milk pooled under the consolidated order, farmers that
had been in the order with the higher Class I utilization rate saw a
decrease in their blend price. However, farmers that had been in the order
with the lower Class I utilization rate experienced an increase in their
blend price. Therefore, splitting up the consolidated orders in these
instances would affect farmers differently depending on the Class I
utilization rates of the new orders. AMS officials indicated that
splitting up orders would affect the distribution of farm income but would
not affect overall production and therefore would have minimal impacts on
federal costs. They also noted that splitting up orders could increase the
movement of milk because having smaller orders makes farm prices more
closely reflect local supply and demand conditions. Therefore, to the
extent that smaller orders would increase the blend prices in some areas,
this option could create incentives to transport more milk to those areas.

51Also under consideration are alternative proposals that would merge the
Southeast and Appalachian Orders, as well as add areas to the merged
orders.

Appendix VII Effects of Recent Federal Dairy Program Changes and
Alternative Policy Options

Accelerate USDA's Hearing Process

Some experts suggested that USDA's federal order hearing system is too
slow to effectively respond to problems and changing market conditions.
The American Farm Bureau study reported that it can take 2 or more years
from the time USDA receives a request for a hearing or direction from the
Congress before USDA implements the rules of a final decision. The study
noted that within this long time frame, either the industry may have
struggled under faulty rules or, by the time final rules are effective,
industry changes may have occurred rendering the final rule obsolete. For
example, a couple of industry sources stated that the dairy industry is
changing rapidly, with a number of new products coming onto the market.
They indicated that USDA would have difficulty regulating these products
because it takes too long to get a decision on the class under which they
would be priced.

USDA faces a number of challenges in shortening the time between receiving
a hearing request and implementing the rules of a final decision, while
still ensuring the promulgation of economically sound regulation.

o 	The USDA hearing process is set forth by law.52 Before issuing or
amending marketing orders, the Secretary of Agriculture must conduct a
formal on-the-record rulemaking proceeding. USDA must notify the public
and provide an opportunity for a public hearing and comments. Before an
order regulation or amendment to a milk marketing order can become
effective, it must meet certain requirements including that it be approved
by at least two-thirds of the affected dairy farmers in the order, or
dairy farmers who produce at least two-thirds of the milk produced in that
order.

o 	If individual parties do not agree with the USDA decision, they can
seek review of that decision in federal court. For example, after the 2000
FMMO reforms, several industry groups received an injunction that
prevented USDA from implementing new pricing rules that would have
established separate Class III and IV butterfat prices. In response, USDA
issued a revised decision, but these new rules were not implemented until
April 2003.

527 U.S.C. S:608(c).

  Appendix VII Effects of Recent Federal Dairy Program Changes and Alternative
                                 Policy Options

o 	Making informed decisions about changes in complex federal dairy policy
can be time-consuming. For example, such decisions require thorough
analysis and possibly modeling. In addition, AMS officials indicated that
hearing participants are not always ready on time, and keeping a stricter
schedule could result in an incomplete hearing record. Furthermore, they
stated that it is difficult to compile the hearing transcripts quickly and
accurately. However, they noted that USDA has recently begun evaluating
its transcript contracts using an approach that considers timeliness and
accuracy.

o 	The politicized nature of dairy policy makes it difficult to agree on
proposed changes to FMMOs. Because FMMO provisions affect cooperatives and
processors in different ways, these entities may not always agree on a
proposed change. Moreover, given regional differences in production and
utilization, farmers in different regions may not agree on changes in
federal dairy policy.

AMS officials said that delays are also caused by the lack of available
judges and attorneys who deal with milk pricing issues. The officials said
that increasing the speed of the decision-making process is likely to
increase federal costs because more of these resources would be required.
USDA officials indicated that they do not believe the hearing process
inhibits the ability of FMMOs to respond to changing market conditions or
the marketing of new dairy products.

Eliminate FMMOs or Classified Rather than trying to reform the FMMO
system, some dairy experts have

Pricing	considered the possibility of eliminating FMMOs and thus the
classified pricing system.53 To the extent that manufactured product
prices stay above the level of the price support program, market forces
would set prices for all uses of milk. In a 1988 report on FMMOs,54 we
found that the production and marketing conditions used to justify
federally guaranteed milk prices under marketing orders no longer existed
because most milk

53This would be such a major change in federal dairy policy that its
impacts are uncertain. This uncertainty is reflected in the differences in
the estimated impacts of various empirical studies that considered this
alternative. AMS officials indicated that without the FMMO system of
minimum prices set for milk based on wholesale dairy product prices, it is
unclear whether or not a structure would evolve that similarly transmits
wholesale price signals to farmers.

54GAO, Milk Marketing Orders: Options for Change, GAO/RCED-88-9
(Washington, D.C.: Mar. 21, 1988).

Appendix VII Effects of Recent Federal Dairy Program Changes and
Alternative Policy Options

being produced is now Grade A and is eligible to serve the fluid milk
market during periods of supply and demand imbalance.55 Also, our study
noted that improvements in refrigeration and the transportation system
have made it less expensive to rely on milk supplies from other markets.
Further, we reported that the differences in Class I differentials-which
were, in part, intended to represent the costs of producing and
transporting milk from areas with a surplus to areas with a
deficit-actually bear little relationship to differences in either
production or transportation costs for milk, thereby providing incentives
for overproduction in certain regions. A 2002 University of Wisconsin
study argued that this overproduction has hurt farmers in areas with low
Class I utilization through an overall reduction in the price of milk used
for manufacturing purposes.56

Academic and USDA studies have generally concluded that without the
classified prices established by FMMOs, fluid milk processors would likely
pay lower average prices to farmers, which would decrease farm income in
high fluid milk utilization areas, especially in the short term. Estimates
of impacts on farm prices varied among studies. For example, the American
Farm Bureau study estimated that average farm prices for raw milk would
fall by about $0.50 per hundredweight during the first couple of years
following federal order elimination.57 Another study, published by USDA,
estimated that eliminating the federal order system would decrease Class I
prices an average of $0.95 per hundredweight over the period from 2002

55Dairy farmers sell two grades of milk. Grade A may be used for fluid
consumption or in manufactured products. Grade B may be used only for
manufactured products. Farmers producing Grade A milk must adhere to
higher sanitation requirements than for Grade B milk. The FMMO system was
developed to ensure an adequate supply of Grade A milk to meet the fluid
milk needs of consumers.

56Tom Cox, Bob Cropp, Randy Fortenbery, and Ed Jesse, "Rethinking
Dairyland Chapter 3: The Effects of Federal Dairy Programs on the
Competitiveness of Dairying in Wisconsin," Marketing and Policy Briefing
Paper Number 78C, Department of Agricultural and Applied Economics,
University of Wisconsin-Madison, September 2002.

57This analysis assumed that a fluid milk premium of $0.50 per
hundredweight would remain nationally and that California would make no
changes to its state milk marketing system. If the assumption of a fluid
milk premium is dropped, the study showed a farm price decrease of closer
to $0.70 per hundredweight during the first couple of years following FMMO
elimination.

Appendix VII Effects of Recent Federal Dairy Program Changes and
Alternative Policy Options

through 2007.58 Further, the 2002 University of Wisconsin study estimated
that farm prices would decrease around $0.05 to $0.10 per hundredweight.
AMS officials indicated that farmers would likely reduce production to the
extent they receive lower average farm prices.

The effects of FMMO elimination could be different based on farm size.
Some of the academic and industry sources we contacted noted that farmers
are at a disadvantage in terms of market power within the dairy industry.
Without the pooling of milk proceeds and the payment of uniform blend
prices, larger farms would have increasing incentives to establish
contracts directly with processors, and processors would increase their
efforts to procure milk directly from larger farms closer to their own
plants. Smaller and more distant farms could be more likely to be
bypassed. Also, to the extent that dairy cooperatives are unable to cover
the costs of balancing and other services they provide, processors may be
able to deflect the costs of these operations back to farmers.

Studies also indicated that the magnitude of these effects could vary by
region. Without classified pricing, the prices of raw milk used in fluid
milk products are likely to fall, while the prices of raw milk used in
manufactured products are likely to rise. As a result, farmers in regions
with higher utilization of raw milk for fluid purposes, such as the
Northeast, would be worse off without classified pricing, while farmers in
regions with high utilization of raw milk for manufacturing purposes, such
as the Upper Midwest, could be better off without classified pricing. For
example, the American Farm Bureau study reported that states with less
than 20 percent fluid utilization of raw milk would have higher average
farm prices with the elimination of federal orders, while those states
with fluid utilization of raw milk in excess of 35 percent have higher
farm prices with the federal order system in place.

In the short term, to the extent that lower farm prices paid for raw milk
used in fluid products are passed on to consumers, fluid milk consumption
could marginally increase. The American Farm Bureau study estimated a

58J. Michael Price, U.S. Department of Agriculture, Effects of U.S. Dairy
Policies on Markets for Milk and Dairy Products, Market and Trade
Economics Division, Economic Research Service, Technical Bulletin Number
1910, May 2004. This analysis assumed a national fluid milk premium of
$1.30 per hundredweight. According to the USDA study, this change is in
the effective Class I price. Since Class I prices and prices for raw milk
used for manufacturing purposes would move in opposite directions with
elimination of the FMMO program, the change in average prices would be
less than $0.95 per hundredweight.

Appendix VII Effects of Recent Federal Dairy Program Changes and
Alternative Policy Options

2.5 percent increase in fluid milk demand, while the USDA study estimated
a 2 percent increase. On the other hand, the University of Wisconsin study
reported that the combination of less milk production and more fluid milk
consumption would reduce the amount of raw milk available for manufactured
products and increase manufactured product prices accordingly.59 To the
extent that manufactured product prices increase, consumers may buy less
of these products.

Over the long term, increased prices for milk used in manufactured
products could limit reductions in both farm income and production
resulting from elimination of FMMOs and classified pricing. In fact, the
American Farm Bureau study found that once a supply adjustment occurs,
average milk prices would return to levels similar to those prior to FMMO
elimination. However, to the extent that farmers in areas with high
manufacturing use experience higher prices for their milk, the incentive
to produce more milk could limit potential increases in manufacturing
product prices. In the end, some decline in production could be expected
over the long term because the overproduction incentive resulting from the
classified pricing system would be removed.

Eliminating FMMOs and classified pricing could also affect other federal
dairy policies and, therefore, affect federal costs. If MILC payments were
still based on the relationship between what fluid milk processors pay to
acquire milk in Boston and a target price of $16.94, MILC payments would
likely increase in size and frequency.60 At the same time, elimination of
FMMOs could decrease federal costs related to the price support program
because an increase in manufactured dairy product prices resulting from
eliminating classified pricing could reduce the need for dairy commodity
purchases by the CCC. However, it is unclear whether this effect would be
large enough to offset additional payments under the MILC program.

Some AMS officials also indicated that farmers could experience increased
price volatility without FMMOs. In the absence of minimum class prices,
greater price volatility could result, in part, from seasonal production
variation or short-term factors, such as holidays or weather events.
Further, while some sources questioned the extent to which a state

59The May 2004 USDA study estimated that manufactured product prices would
increase by 2 percent above the study's baseline estimate for the period
2002 through 2007.

60Should the MILC program not be extended beyond fiscal year 2005, this
scenario would only exist in the short term.

Appendix VII Effects of Recent Federal Dairy Program Changes and
Alternative Policy Options

regulatory system, such as California's, could continue to exist in the
absence of the federal classified pricing system, others indicated that
state regulation of milk could increase if the federal system were
eliminated. A couple of industry sources that we contacted indicated that
an increase in the number of states that regulate milk could make it more
difficult for them to do business and would be a less efficient system of
regulation.

On the other hand, eliminating FMMOs and classified pricing could also
provide greater incentives for product innovation. Without classified
pricing, the market would price products more openly based on supply and
demand and would increase the incentives for processors to develop
alternative dairy products. To the extent that alternative products
generate new demand for milk, this innovation could benefit farmers. For
example, a recent study by researchers at Cornell University on the
assignment of new products under a classified pricing system found
significant difficulties.61 The study reported that the assignment of a
new product to a higher-priced class increases farm income in the short
run; however, the incentive to increase production provided by the use of
raw milk in this higher-priced class and reduced demand for raw milk
stemming from these higher prices can offset farm income gains in the long
run. Furthermore, the study found that whether the new product detracts
from sales of existing fluid milk products could also affect whether
assignment to a higher-priced class increases net revenues to farmers. One
USDA representative stated that a number of new products, such as
low-carbohydrate milk beverages, have recently entered the market. In some
cases, according to the representative, these products are intended to
compete with Class I products but are formulated to avoid regulation as
Class I products. Processors seek to avoid having these products regulated
as Class I products because they would then be required to pay more for
raw milk. Conversely, farmers want these products classified as Class I
products so that their raw milk used in these items will be priced at the
higher level.

Some sources suggested that classified pricing could be eliminated without
eliminating FMMOs altogether. In this case, FMMOs might continue to
perform functions such as pooling revenue, auditing, verifying weights and
milk components, and collecting statistical information. A few dairy
experts indicated that these particular aspects of the FMMO system benefit

61Andrew M. Novakovic, Charles F. Nicholson, and Mark W. Stephenson,
"Assignment of New Products Under Classified Pricing: A Conceptual Dynamic
Model of Class Assignment Outcomes," Cornell University, January 2004.

  Appendix VII Effects of Recent Federal Dairy Program Changes and Alternative
                                 Policy Options

the dairy industry. While some of these functions might be picked up by
the private sector if FMMOs were eliminated, they would come at a cost to
dairy farmers.

Retaining FMMOs while eliminating classified pricing would probably lessen
the impacts of deregulating raw milk prices but would be unlikely to
change the direction of most effects or who benefits. For example, while
farm income would still fall without classified pricing, continuing to
pool revenues through FMMOs, if possible, could help cooperatives
negotiate over-order premiums because pooling could help cooperatives
maintain their market power relative to processors. Thus, farm income
might fall by less than it would if FMMOs were eliminated entirely.
Moreover, by continuing to pool revenues, retaining FMMOs could limit
increases in price volatility resulting from the elimination of classified
pricing. In addition, the orders would continue to aim to ensure equitable
treatment for producers and processors. Maintaining orders is not likely
to change the fact that farmers in regions with high fluid milk
utilization would experience greater reductions in farm income from
eliminating classified pricing than farmers in regions with high
manufacturing utilization. Production would still adjust downward in
response to lower milk prices, and retail fluid milk prices would also
decrease to the extent that lower prices for raw milk used in fluid milk
products are passed on to the consumer. However, to the extent that
cooperatives are able to maintain higher over-order premiums by retaining
FMMOs, these production and consumer price effects might be less than they
would be if orders were eliminated entirely. Effects on federal costs
would still be mixed.

Change the Dairy Price Support Program

Dairy experts have raised several concerns about the price support program
in recent years, including that the support level is too low to adequately
support farmers; that the program provides incentives to overproduce milk
and certain commodities purchased by the CCC; that USDA has not managed
the program to maintain the established support price during periods of
low market prices; that there are additional costs of selling dairy
products to the government, which diminish the effectiveness of the
support price; and that the program stifles innovation in the industry.
Accordingly, a number of options have been discussed to modify the price
support program, including raising the overall level of the support price
(and thus the related commodity purchase prices), making administrative
changes such as allowing the CCC to purchase a wider range of dairy

  Appendix VII Effects of Recent Federal Dairy Program Changes and Alternative
                                 Policy Options

products, and eliminating the program altogether.62 Figure 37 shows the
effects of various options to change the dairy price support program under
low- and high-price scenarios over the short and long terms.

                 Raise support price                   
                 Raise support price               LP                        
                                                   HP  ?  ? ? ? ? ?          
    Change administration of price support program                           
    Allow CCC to purchase wider range of products  LP  ?  ?   ?      ? ?  ?  
                                                   HP       ?   ? ?          
             Tilt prices based on market           LP         ?           ?  
                                                   HP  ?  ? ?   ? ?          
     Reflect cost differences in selling to CCC    LP         ?           ?  
                                                   HP  ?  ? ?   ? ?          
           Eliminate price support program                                   
           Eliminate price support program         LP                        
                                                   HP  ?  ?   ? ? ?          

Increase

Decrease

Mixed impacts ? No/unclear impacts LP: Low price scenario HP: High price
scenario Source: GAO analysis based on interviews with dairy experts and
reviews of relevant studies.

62Changes in the dairy price support program could potentially conflict
with the U.S.'s WTO commitments.

  Appendix VII Effects of Recent Federal Dairy Program Changes and Alternative
                                 Policy Options

Note: The figure does not indicate the degree of increase or decrease for
each policy consideration as it relates to a policy option. In addition,
as discussed in this appendix, limitations affecting some policy options
could affect the degree of an option's potential impact on a policy
consideration.

Raise the Level of the Support One option for modifying the dairy price
support program is to raise the

Price

support price and the related commodity purchase prices. Many dairy
experts indicated that the support price has fallen below the costs of
production for most farmers and, therefore, is not providing an effective
safety net during periods of low prices. Additionally, sources cited the
reduction in the support price as a factor in increasingly volatile milk
prices. For example, one academician we contacted stated that recent
volatility in milk prices has resulted from the virtual elimination of the
price support system as an effective price floor during periods of low
milk prices. The price support program also worked to reduce price
volatility during periods of high milk prices by releasing CCC stocks of
purchased dairy commodities when market prices reached 110 percent of the
support price.63 However, without purchasing sufficient quantities of
manufactured dairy products, the program does not perform this balancing
function.

According to an economist with USDA's Farm Service Agency (FSA), raising
the support price would increase farm income and, thus, raw milk
production. The economist indicated that this option, by raising the floor
for milk prices, would also reduce price volatility; however, increased
purchases of dairy products would mean higher federal costs for the price
support program. Similarly, the American Farm Bureau study found that
while raising the support price would reduce price volatility, it could
create a situation in which the CCC purchases surplus dairy products in
most years. The FSA economist noted that these federal costs might be
offset, at least in part, by a reduction in payments under the MILC
program. The economist also said that consumer costs would likely be
higher, on average, because increasing the support price would limit the
fall of prices for manufactured products.

The FSA economist also noted that increasing the support price would
decrease the economic efficiency of federal dairy policies, particularly
to the extent that a higher support price stimulates increased production.
Under this scenario, increased production would represent an allocative
inefficiency because resources would go into producing milk that is not

63USDA indicated that currently, the price support program allows the CCC
to release its commodity stocks if the market price for a particular
commodity exceeds that commodity's purchase price.

  Appendix VII Effects of Recent Federal Dairy Program Changes and Alternative
                                 Policy Options

needed to supply the market. Some sources cited examples of how high
support prices under the program led to misallocation of resources into
surplus milk production. For example, from 1977 to 1981, the support price
for Grade A milk rose from $8.26 to $13.10 per hundredweight, and annual
government expenditures on dairy price supports increased from a few
hundred million dollars to over $2 billion.

Change the Administration of the A number of options have been proposed or
discussed to change the

                             Price Support Program

administration of the dairy price support program. These options include
allowing the CCC to purchase a wider range of products, adjusting
commodity purchase prices based on market conditions, and setting
commodity purchase prices to reflect cost differences between selling to
the CCC and selling in the marketplace.

Allow the CCC to Purchase a Wider Range of Products

One potential modification to the administration of the price support
program would be to allow the CCC to purchase a wider range of products
than butter, cheese, and nonfat dry milk. Some dairy experts and studies
indicated that by focusing on the purchase of a few specific commodities,
the price support program distorts the market by providing incentives to
overproduce these commodities, while at the same time dampening incentives
for innovation in the dairy industry. Manufacturers that develop
innovative products incur more risk because they will not be able to sell
their products to the government if they cannot obtain a market price high
enough to cover their costs. One cooperative representative said that
while nonfat dry milk contains protein and calcium-both valuable
components that could be used in other products, such as protein
bars-manufacturers continue to produce nonfat dry milk in excess
quantities because that is what the government is buying. A number of
other sources, including industry representatives, academicians, and a
report by the International Trade Commission, noted that by purchasing
nonfat dry milk, the price support program may be impeding development of
a domestic milk protein concentrate industry by creating disincentives to
shift raw milk supplies to innovative products.

USDA officials cautioned that in order for a product to function well as a
price support product, it must (1) represent a major use of milk; (2) have
enough extra capacity to absorb a substantial amount of milk; (3) be
storable for long periods; and (4) have an active, liquid wholesale
market. Given these conditions, it is questionable whether some
alternative products, such as protein bars, would be effective as price
support

Appendix VII Effects of Recent Federal Dairy Program Changes and
Alternative Policy Options

products. Furthermore, the officials argued that requiring the price
support program to incur the risk of product innovation through this
approach would alter the fundamental purpose of the program-supporting
farm prices.

If the CCC were to purchase a wider range of products, manufacturers would
have greater incentives to use milk in alternative ways because the price
support program would decrease the risks of trying to produce and market a
greater number of products. Consumers could benefit as innovative products
gained easier access to the market, causing consumer prices to fall.
However, one FSA economist stated that this option would greatly increase
the complexity of the price support program, potentially increase federal
costs, and require new legislation. Moreover, he noted that this change
would require close coordination between trade policies and the price
support program. For example, without tariff-rate quotas on some products
such as milk protein concentrates, the CCC could end up supporting
additional imports if the purchase prices were set too high. The economist
stated that this option could reduce production of nonfat dry milk but
would be less likely to affect price volatility.

Adjust Commodity Purchase Prices Based on Market Conditions

A second potential modification would be to adjust-or tilt-commodity
purchase prices based on market conditions. Some dairy experts, as well as
academic studies, reported that because USDA does not tilt CCC purchase
prices frequently enough to maintain a balance between butter and nonfat
dry milk purchase prices that is based on current economic conditions, the
price support program has distorted the market with unclear price signals
and induced surplus production of certain goods (notably nonfat dry milk).
Therefore, some experts indicated that tilting prices based on established
criteria would be better. For example, one dairy processor recommended
changing the balance of butter and nonfat dry milk purchase prices
automatically if the ratio of CCC purchases of butter and nonfat dry milk
falls outside a certain range. Similarly, the American Farm Bureau study
suggested that to achieve the support program's objectives without
distorting the market and increasing government costs, changes to
commodity purchase prices should be based on market conditions so that
they would not be subject to political pressure.

Basing the tilt of commodity purchase prices on market conditions would
increase the economic efficiency of the price support program by reducing
the price distortions that lead to surplus production of goods that are
not

Appendix VII Effects of Recent Federal Dairy Program Changes and
Alternative Policy Options

required to supply the market. Falling market prices for a particular
commodity suggest that the quantity supplied temporarily, at least,
exceeds the quantity demanded. If the CCC continues to buy the commodity
in significant amounts while the market price remains low, it distorts the
market by providing an incentive to produce that commodity purely for the
purpose of selling it to the government. This incentive not only delays a
market response to the lower commodity price but also prevents that milk
from going toward a higher-valued use. Moreover, a study by researchers at
the University of Wisconsin found that tilts make it more likely that
fluid milk prices will be driven by cheese prices because the price
support program would no longer be supporting the manufacturing prices of
Class IV products above market levels.64 This study reported that
tightening the relationship between cheese prices and Class I prices
improves market signals to dairy farmers because nationally, as noted
earlier, a greater percentage of raw milk is used in Class III products
(cheese) than in Class IV products (butter and nonfat dry milk).

According to one FSA economist, lowering the purchase price of a
particular commodity in the short term could reduce farm income; however,
by encouraging production levels to respond more quickly to low price
periods, tilting CCC purchase prices to reflect market conditions could
maintain higher farm prices over the long term. The economist also
indicated that this change could decrease federal costs for purchasing and
storing dairy products. The consumer price effects of market-based tilts
are less clear. For example, if market butter prices are high and market
nonfat dry milk prices are low, a market-based response would indicate
that USDA should raise the CCC butter price and lower the CCC nonfat dry
milk price. However, further price adjustments could result in the CCC
purchase price for butter exceeding the market price, which could trigger
CCC purchases of butter and raise the market price even higher. In such
cases, assuming that price changes at the wholesale level are passed on
through the retail level, consumers would benefit from lower prices on
some commodities, while potentially experiencing higher prices on others.
The net effect to consumers would then depend on the relative price
changes for these products and the quantities of each that were purchased.

64Bob Cropp and Ed Jesse, "The Butter-Powder Tilt," Marketing and Policy
Briefing Paper Number 72, Department of Agricultural and Applied
Economics, University of Wisconsin- Madison, June 2004.

Appendix VII Effects of Recent Federal Dairy Program Changes and
Alternative Policy Options

Reflect Cost Differences in Selling to the CCC

A third option to change the administration of the price support program
would be to set CCC purchase prices to reflect cost differences in selling
to the CCC versus selling in the marketplace. In recent years market
prices have fallen below the support price level in some months. For
instance, between July 2002 and June 2003, the Class III milk price was
below the $9.90 target level in 9 months. Although FSA officials indicated
that USDA is required to set product purchase prices in such a way that
only the average annual farm milk price, not the monthly price, is at the
support price, the Farm Security and Rural Investment Act of 2002 (2002
Farm Bill) simply states that the price of milk should be supported at
$9.90 per hundredweight.65 Concerns over USDA's management of the price
support program led to language in the Consolidated Appropriations Act,
2004 requiring the Secretary of Agriculture to more diligently support the
farm price of milk.66 Some dairy experts and studies indicated that one
reason that market prices sometimes fall below the CCC support price is
because there are additional costs of selling to the government that are
not reflected in the commodity purchase prices. Therefore, the effective
support price is actually below $9.90 because with higher costs of selling
to the government, the market price has to fall below the CCC purchase
price before processors are better off selling to the CCC than to the
market. Some of these additional costs include packaging for longer-term
storage, meeting stricter grading standards, and a time lag between when
the product is made and when it is approved for sale to the federal
government. An FSA economist estimated that these cost differences amount
to about $0.04 to $0.05 per pound for cheese.67

One option for reflecting the differences in cost between selling dairy
products to the government and selling in the marketplace would be to
raise the CCC purchase prices of these products to reflect the additional

65USDA also averages the Class III and Class IV prices in determining if
the $9.90 per hundredweight support price has been met.

66The Consolidated Appropriations Act, 2004 required the Secretary of
Agriculture to more diligently support the farm price of milk at the farm
bill-mandated support price of $9.90 per hundredweight or lose funding for
the administration of the program. Pub. L. No. 108-199 S:780.

67The FSA economist also stated that there are minimal cost differences
for Class IV products (butter and nonfat dry milk). An academic expert
indicated that with the exception of April to June 2003, the Class IV
price has not dropped below the support price.

Appendix VII Effects of Recent Federal Dairy Program Changes and
Alternative Policy Options

costs of manufacturing product for sale to the government.68 This would
help ensure that manufacturers receive a price for their products that
allows them to return at least $9.90 per hundredweight (the support price)
to farmers. One FSA economist indicated that this change could cause
farmers to marginally increase production, leading to increased CCC
purchases. Increased CCC purchases, in addition to higher purchase prices,
would increase the federal costs of the price support program and, at the
same time, higher manufactured product prices could translate into higher
consumer prices. However, it is difficult to estimate the added costs of
selling to the CCC because these costs are likely to vary widely among
different manufacturers. Thus, raising the product purchase prices could
provide unwarranted benefits to some manufacturers while still being
insufficient to induce sales to the government by some others. In
addition, according to one academic study, there is no clear evidence that
higher selling costs are the major barrier in selling to the government.
Some experts have put forth the possibility that fixed contracts between
dairy product manufacturers and their buyers may prevent manufacturers
from selling to the government.

An alternative way to reflect the differences in cost between selling
dairy products to the government and selling them in the market place
would be to require the CCC to alter product specifications and payment
terms to conform to those used on the Chicago Mercantile Exchange (the
Exchange).69 An FSA economist stated that some changes to product
specifications are already being considered and have been put out for
comment to the dairy industry.70 The economist stated that while these
proposed changes will help bring CCC product specifications into greater
conformance with market standards, some differences would remain. Most
notably, the CCC requires that products be storable for up to 3 years, a
longer period than is generally required in the market. To the extent that
this proposal reduces additional costs of selling manufactured dairy

68The National Milk Producers Federation has asked the CCC to increase its
purchase prices for cheese by $0.056 per pound.

69The Chicago Mercantile Exchange is a central marketplace with
established rules and regulations where buyers and sellers trade
agreements known as futures contracts to, for example, buy or sell a
commodity at a future date.

70USDA put the proposed draft Announcement Dairy 6 out to the dairy
industry for comment on April 5, 2004. The proposal includes a number of
changes such as increasing lot quantities for manufactured dairy product
purchases, as well as changing commodity, packaging, and marking
specifications.

  Appendix VII Effects of Recent Federal Dairy Program Changes and Alternative
                                 Policy Options

products to the CCC by more closely aligning product specifications with
market standards, it could induce greater manufactured product sales to
the CCC and would keep market prices higher. Therefore, aligning these
specifications could increase farm income, provide a marginal production
stimulus, and raise federal costs related to additional CCC purchases.
Also, this option would not necessarily prevent the Exchange prices from
falling below CCC purchase prices because if there are barriers other than
costs (such as contractual obligations) that prevent manufacturers from
selling to the government, these barriers would still exist.71

Eliminate Dairy Price Supports	Another policy option would be to eliminate
the price support program altogether and rely on alternatives available to
farmers to assist them in managing risk of low and/or highly volatile
prices for their milk. A number of dairy experts have argued that the
support price has been set so low at $9.90 per hundredweight that it is
not having significant impacts. Thus some academicians, as well as USDA,
have studied the potential effects of eliminating the program.

In the short term, eliminating the program would have a greater impact if
market prices were at or below the level of the support price. The May
2004 USDA study estimated that eliminating the price support program would
cause wholesale prices of nonfat dry milk to decline by 15 to 20 percent
over the first couple of years.72 For subsequent years, the study
estimated that prices would recover somewhat to 10 percent below baseline
levels. Further, the study estimated that the decline in nonfat dry milk
prices would encourage diversion of this milk to alternative uses, leading
to lower prices for these alternative uses. Generally, lower prices would
reduce farm income and potentially lead to lower consumer prices. Farmers
would likely respond to these lower prices by producing less milk.

71Another proposed alternative would be to have the CCC place bids for
commodity purchases on the Exchange as a better way to provide a price
floor for manufactured products. However, because the CCC cannot entirely
change its product standards to conform to industry requirements, an FSA
economist indicated that this alternative is not feasible.

72The study established a baseline assumption whereby CCC purchases of
nonfat dry milk averaged approximately 17 percent of total U.S. production
over the period 2002 through 2007. Purchases of other commodities were
assumed to be minimal. However, this assumption was based on a nonfat dry
milk purchase price of $1.01 per pound. The actual purchase price was
lowered to $0.80 per pound and, thus, the study may overstate the effects
of eliminating this program.

  Appendix VII Effects of Recent Federal Dairy Program Changes and Alternative
                                 Policy Options

According to an FSA economist, eliminating the price support program would
increase economic efficiency by allowing market price signals to be
transmitted more clearly. The economist also stated that volatility in
milk prices would increase. He noted that the combination of increased
volatility and reduced farm income would force less efficient farmers to
exit production. This exit would increase the economic efficiency of
national resource allocation by enhancing current shifts in production to
more efficient dairy farms. He added that in the absence of a price
support program, new entrants to dairy production would likely be larger,
more efficient operations. The FSA economist also said that eliminating
the price support program would provide a cost savings for the federal
government because the CCC would no longer have to purchase or store dairy
commodities. These savings would be greater during periods when farm
prices are low because when they are high, even if the program remains in
effect, the government purchases fewer dairy products and incurs less
cost. However, when farm milk prices are low, savings from eliminating the
price support program could be partially offset by increased payments
under the MILC program for as long as that program continues.

Over the long term, reduced production would mitigate some of the impacts
of eliminating the price support program, because reduced supplies lead to
increased prices (assuming demand stays the same). However, the USDA study
estimated that even with the positive price effects resulting from reduced
production, farm income would still decrease by approximately $3.5 billion
over the long term. Additionally, without CCC purchases of dairy
commodities, USDA would be unable to balance high market prices by
releasing these stocks, thereby contributing to increased price volatility
over the long term.

Extend the Milk Income Loss Contract Program

The MILC program has benefited many smaller dairy farmers during the most
recent period of low farm prices by providing them income support.
However, by providing support to some farmers who otherwise might have
exited the dairy industry, the program has slowed the normal downward
supply response to lower farm prices and kept aggregate production higher
during this period than it otherwise would have been. The MILC program is
scheduled to expire at the end of fiscal year 2005. If it is not extended
in some form, aggregate production is likely to respond more rapidly to
future low-price periods because smaller farmers are likely to exit
production at greater rates than they did during the most recent period of
low farm prices. With this more rapid production response, farm prices
would likely start rising again sooner than in the recent past when the
MILC program

Appendix VII Effects of Recent Federal Dairy Program Changes and
Alternative Policy Options

has been in place. However, although production levels in the short term
would likely decrease more during low-price periods, in the long term
aggregate production might not decrease substantially because higher
average farm prices would stimulate additional production from the dairy
farmers that stay in business. In addition, by allowing the MILC program
to expire, the government can avoid the costs of the payments to farmers
that the program provides.

There are several options to maintain the benefits of the MILC program to
some dairy farmers by extending it beyond 2005. One option is to extend
MILC at its current target price and eligible production limit. A second
option, a proposal introduced in the Senate in the 108th Congress, would
extend MILC through fiscal year 2007 with an increase in the eligible
production cap from 2.4 million pounds to 4.8 million pounds.73 A third
option would extend MILC with a lower target price but a higher or no
eligible production limit. Figure 38 shows the effects of various policy
options to extend the MILC program under low-and high-price scenarios over
the short and long terms.

73The draft legislation is known as the Milk Income Loss Contract
Extension, S. 2609 (2004). A bill that would extend authorization for the
MILC program to 2007, but would not increase the eligible production cap,
has been introduced in the House, H.R. 3990 (2004). Additionally, a
separate bill, S. 2825 (2004), was introduced in the Senate. Section 426
of this proposed legislation would extend authorization for the MILC
program to 2007, increase the target price for the MILC program from
$16.94 per hundredweight to $17.10, and eliminate a provision for
determining whether producers are on separate dairy operations for
purposes of determining payment eligibility. However, the bill would not
change the MILC program's eligible production cap of 2.4 million pounds.

  Appendix VII Effects of Recent Federal Dairy Program Changes and Alternative
                                 Policy Options

Figure 38: Potential Effects of Options to Extend MILC on Various Policy
Considerations

Increase

Decrease

Mixed impacts ? No/unclear impacts LP: Low price scenario HP: High price
scenario Source: GAO analysis based on interviews with dairy experts and
reviews of relevant studies.

Note: The figure does not indicate the degree of increase or decrease for
each policy consideration as it relates to a policy option. In addition,
as discussed in this appendix, limitations affecting some policy options
could affect the degree of an option's potential impact on a policy
consideration.

Extend MILC in Its Current Form	One analysis conducted by a researcher at
the University of Missouri estimated that compared to a baseline estimate
in which the MILC program expires in 2005, extending the MILC program
through 2012 in its current form would result in greater milk production
and lower farm prices.74

74Scott Brown, "The Effect on the United States Dairy Industry of Removing
Current Federal Regulations," Report Number 03-03, Food and Agricultural
Policy Research Institute, University of Missouri, April 2003. Both the
baseline estimate and the estimate for the MILC extension scenario are
based on the assumption that the price support program, federal milk
marketing orders, import restrictions, and the Dairy Export Incentive
Program remain in effect.

Appendix VII Effects of Recent Federal Dairy Program Changes and
Alternative Policy Options

Production was estimated to be 0.8 billion pounds higher in 2006 and to
average 1.4 billion pounds per year higher from 2008 through 2012. The
estimated price difference was also greater in the longer term than
initially.75 Greater production and lower farm prices are consistent with
the expectation that extending the MILC program would keep some smaller
dairy farmers in the industry who otherwise might exit after 2005 if the
MILC program is allowed to expire then. This study also estimated that due
to the MILC payments, farm income would increase if the program were
extended, despite lower farm prices. Initially, in 2006, the estimated
increase in farm income was $0.43 per hundredweight, and even in 2012,
when the estimated farm price was $0.50 per hundredweight below what it
would be if MILC expires after 2005, the estimated increase in farm income
was $0.08 per hundredweight. In addition, extending MILC in its current
form would increase federal costs. This study estimated that, on average,
annual government costs from 2006 through 2012 would be about $1.2 billion
higher than if the MILC program expires after 2005.76

According to an FSA economist, lower farm prices resulting from extending
MILC in its current form could be passed on as lower retail prices for
consumers. However, the economist indicated that the effects on price
volatility are less clear. He added that the extension of the MILC program
as currently designed would continue to favor smaller farmers over larger
farmers, because a greater percentage of smaller farmers' production is
eligible for MILC payments. Thus, in general, the MILC program would
continue to benefit farmers in the eastern and upper midwestern states
over farmers in the western states. Therefore, for a given level of milk
production and to the extent that larger farmers in the West are more
efficient than smaller farmers in the East, extension of the MILC program
as currently designed would reduce the economic efficiency of the
allocation of dairy production resources nationally compared to the
allocation that would occur if MILC expires.

75Although, as discussed above, one might expect that the price and
production differences would become smaller over time, according to the
author of this study the model assumes competitive markets and, therefore,
does not fully capture long-term adjustments.

76However, the baseline assumption for farm milk prices was lower than the
prevailing prices of the past few months and, according to the study's
author, with higher assumed farm prices, the effect on government costs of
extending the MILC program would be less. The accuracy of the other
estimates also depends on the accuracy of the market prices assumed in the
baseline.

  Appendix VII Effects of Recent Federal Dairy Program Changes and Alternative
                                 Policy Options

Extend MILC with a Higher Cap or with No Cap

An alternative proposal has been introduced in the Senate in the 108th
Congress that would extend the MILC program at its current target price
but increase the cap on eligible production. The University of Missouri
study examined the impacts of a similar option in which the cap on
eligible production is removed in 2003 and MILC is extended through 2012
without a cap.77 This study estimated that extending MILC without a cap
would result in a larger increase in production and a larger decrease in
farm prices than extending MILC with a cap. For example, production in
2006 was estimated to be 2.3 billion pounds per year higher than if MILC
is allowed to expire or 1.5 billion pounds greater than if MILC is
continued in its current form. Greater production and lower farm prices
are consistent with the expectation that making all milk eligible for MILC
payments would provide farmers who might otherwise exit the industry after
2005 (if MILC is allowed to expire) an even greater incentive not to leave
than is provided by extending MILC in its current form. As with the
previous option, this study estimated that due to MILC payments, farm
income would increase if MILC were extended without a cap despite lower
farm prices. Initially, in 2006, the estimated increase in farm income
would be $0.63 per hundredweight, and even in 2012, when the farm price
was estimated to be $1.05 per hundredweight below what it would be if MILC
expires after 2005, the estimated increase would be $0.18 per
hundredweight. Moreover, this study estimated that farm income would be
higher with this option than with extending MILC in its current form
because the additional payments to farmers due to eliminating the cap
would more than offset the additional farm price reduction resulting from
greater production. Extending MILC without a cap would increase federal
costs even more than extending MILC in its current form. This study
estimated that, on average, annual government costs from 2006 through 2012
would be about $2.5 billion higher than if the MILC program expires after
2005, more than $1 billion per year more than was estimated if MILC is
extended in its current form.78

77Because the University of Missouri study's baseline included a MILC
program through 2005, its results can only be used as estimates of the
impact of extending MILC with a higher cap compared to letting MILC expire
for years beyond 2005. The study's results comparing this option with the
study's baseline for 2003 through 2005 are estimates of the difference
between having a MILC program with no cap and one with the current cap.

78However, the baseline assumption for farm prices was lower than the
prevailing prices of the past few months and, according to the study's
author, with higher assumed farm milk prices, the effect on government
costs of extending the MILC program would be less. The accuracy of the
other estimates also depends on the accuracy of the market prices assumed
in the baseline.

  Appendix VII Effects of Recent Federal Dairy Program Changes and Alternative
                                 Policy Options

An FSA economist indicated that consumers would likely benefit from
reduced retail prices under this scenario. Additionally, without the cap
on eligible production levels, the equity concerns about MILC would be
eliminated because the program would no longer favor smaller farmers over
larger farmers. However, the FSA economist also indicated that in addition
to the inefficiencies of increased surplus milk production, this
alternative would reduce economic efficiency of milk production compared
to allowing MILC to expire after 2005 to the extent that it provided
incentives for production by smaller producers with higher costs of
production.

Extend with Lower Target Price A third alternative, considered in the
American Farm Bureau study, is to

and Higher or No Cap	extend MILC with a lower target price and a higher or
no cap on eligible production. The study argued that the $16.94 per
hundredweight target price is too high because MILC payments are triggered
any time the Class I mover is less than $13.69 per hundredweight; at this
target price, MILC payments could be expected in most months given that
from 1990 through 1999 the Class I mover averaged $12.28 per hundredweight
and was below the $13.69 threshold in 104 of 120 months. The analysis
concluded that a target level for MILC payments of a Class I mover at
$12.00 per hundredweight or lower would help to make the program more
marketbased. Further, expanding the cap or eliminating it completely would
help to make the program more equitable among farmers of different sizes.

The effects of extending MILC in the absence of a cap on eligible
production depend significantly on the level of the target price. If the
target price were set too high, it would stimulate surplus production.
During periods of low prices, the government would have to contribute
additional MILC payments to counteract the effects of lower prices.
Additional production would also decrease manufactured product prices,
potentially increasing the costs of the dairy price support program
because these prices would be more likely to reach the support level. At
the same time, these lower prices could provide benefits for consumers.
While a lower target price without a cap would treat farmers with
different sized herds the same way, it might not provide enough payments
to keep farmers in areas with high costs of production from exiting during
periods of low prices. Therefore, this option could increase the economic
efficiency of national dairy production, while also accelerating shifts in
dairy production to the West.

  Appendix VII Effects of Recent Federal Dairy Program Changes and Alternative
                                 Policy Options

Introduce a New Target Price Deficiency Payment Program

Some dairy experts indicated that in lieu of maintaining both the dairy
price support and MILC programs, operating one new target price deficiency
payment program could be a better alternative.79 USDA and other dairy
experts stated that having both the price support and MILC programs is
problematic, particularly during periods of low prices. As noted in the
American Farm Bureau study, the idea behind establishing a target price
deficiency payment program is to allow markets to work to clear dairy
products at market prices and then, when the market price is below the
target price, to pay farmers based on the difference between these prices.
But with both dairy programs in force, MILC maintains and encourages
surplus milk production that must then be purchased by the CCC under the
price support program. This market distortion adds to the costs of both
programs.

Under a new target price deficiency payment program, dairy farmers would
receive a payment when the market price of Class III milk products drops
below a specific target level. Thus, the program would establish a floor
on farm income through a countercyclical payment to dairy farmers instead
of a floor on manufactured product prices with purchases by the CCC.80
Instead of providing an incentive for manufacturers to continue producing
a particular product that the CCC is purchasing, the program would allow
the market to clear a wholesale product price and then pay farmers the
difference if the price were too low. This option would provide
manufacturers the incentive to shift raw milk supplies to their
highestvalued use, further promoting the development of new and innovative
products. This option could also potentially reduce federal costs,
depending on the level of the target price; one expert estimated that with
a target price of $10.50 per hundredweight for Class III milk, the
government would have spent $300 million less than under the MILC program
since its inception in December 2001. With a target price of $10.00 per
hundredweight for Class III milk, these savings would have reached $1.2
billion.

Notwithstanding these potential benefits, dairy experts indicated that a
new target price deficiency payment program could have its own challenges
depending upon how it is designed. In particular, such a program would
require some key decisions regarding price level and

79The MILC program is a target price deficiency payment program. 80A
countercyclical payment is a payment provided during periods of low
prices.

  Appendix VII Effects of Recent Federal Dairy Program Changes and Alternative
                                 Policy Options

regional differences, and whether to cap program benefits based on
payments or quantities of production.

Set Price Level and Consider Some dairy experts said that a problem with
the target price deficiency

Regional Differences

Place Cap on Payments or Production

payment approach is that it would be hard to determine the appropriate
target price without creating distortions in production incentives. If the
target price were set too high, it would have the same effect as setting
the support price too high: it would lead to excess production by
supporting farm income at higher levels than would be available if farmers
received market prices. Long-term overproduction would place additional
downward pressure on market prices and increase federal costs for the
program. If the target price were set at a low level to avoid stimulating
production and increasing government costs, it might not maintain adequate
support for farm income, and could increase price volatility.
Additionally, by influencing domestic production levels, which in turn
influence U.S. market prices, the level of the target price can affect the
incentives of other countries to export manufactured products to the
United States. However, these incentives are also affected by the export
subsidies and lower production costs of some other countries.

The difficulty in setting the appropriate target price is exacerbated by
regional differences in costs of production. A certain Class III target
price might provide adequate support during periods of low prices based on
the costs of production in one region, but not in another. However,
increasing the target price to provide adequate support for higher-cost
regions would not only support production in areas where it is less
economically efficient to do so, but would also provide greater benefits
to farmers in lower-cost regions. These higher benefits would increase the
incentives to overproduce in those areas. Given current trends in the
United States, this scenario would encourage the western shift in dairy
production.

Another challenge in designing a target price deficiency payment program
would be to determine whether to cap the program's benefits either by
limiting the payment a farmer could receive or by limiting the quantity of
milk production on which a farmer would be eligible to receive payments.
The MILC program calls for payments equal to 45 percent of the difference
between $16.94 and the Boston Class I per hundredweight price and has a
production cap (2.4 million pounds of production per dairy operation each
fiscal year). These controls have helped to keep federal costs lower than
they otherwise would have been during periods of depressed prices by
limiting incentives for overproduction. However, the production cap has

  Appendix VII Effects of Recent Federal Dairy Program Changes and Alternative
                                 Policy Options

targeted the program's benefits primarily to smaller farmers, raising
questions of equity.

If a target price deficiency payment program were implemented without any
controls, the risk of market distortions and increased federal costs from
establishing a target price level that is too high could increase
substantially. At the same time, a couple of researchers noted that
establishing the target price deficiency payment program without a
production cap could encourage farmers to enhance the efficiency of their
dairy operations. With a cap limiting eligible production, farmers have
less incentive to adopt new technologies that would increase production.
However, whether or not a cap is placed on eligible production, the
program is likely to confer benefits to farmers in varying degrees. If the
program did not cap eligible production, farmers who could increase their
production efficiency with new technology might have more of an incentive
to do so. But the farmers who would be most likely to take advantage of
this incentive would be larger farmers who may be more efficient, and
might have the resources and access to capital to undertake such an
investment. Conversely, capping eligible production would target benefits
to smaller farmers in the same way as MILC.

Adopt the National Dairy Equity Act

To address concerns about the pending expiration of the MILC program and
provide additional support to dairy farmers, the National Dairy Equity Act
of 2004 (NDEA) has been introduced in the House and the Senate.81 This
proposed legislation would change the federal regulation of milk marketing
through the establishment of regional dairy marketing areas in which
boards created to administer these areas would set minimum prices that
processors would have to pay for raw milk used to make fluid milk products
sold in those areas.82 The NDEA would have a similar effect as the MILC
program in that it might lead to higher incomes for some dairy farmers.
However, concerns have been raised about its impact on farm

81This proposed legislation was introduced in the 108th Congress as S.
2525 and H.R. 4597.

82In earlier legislation, the Federal Agriculture Improvement and Reform
Act of 1996, Congress authorized the six New England states to enter into
an agreement called the Northeast Interstate Dairy Compact, which was
implemented in July 1997. The Compact functioned similarly to the proposed
dairy marketing boards of the NDEA in that it supplemented the FMMOs and
state programs by setting the monthly minimum price to be paid for raw
milk used for fluid milk marketed in the six-state area. The authorization
for the Compact expired on September 30, 2001.

  Appendix VII Effects of Recent Federal Dairy Program Changes and Alternative
                                 Policy Options

incomes in some regions, retail fluid milk prices, coordination of milk
prices across regions, and existing trade agreements. Figure 39 shows the
effects of adopting the NDEA over the short and long terms.

      Figure 39: Potential Effects of Adopting the NDEA on Various Policy
                                 Considerations

Increase Decrease Mixed impacts ? No/unclear impacts Source: GAO analysis
based on interviews with dairy experts and reviews of relevant studies.

Note: The figure does not indicate the degree of increase or decrease for
each policy consideration as it relates to a policy option. In addition,
as discussed in this appendix, limitations affecting some policy options
could affect the degree of an option's potential impact on a policy
consideration.

  Appendix VII Effects of Recent Federal Dairy Program Changes and Alternative
                                 Policy Options

Elements of the NDEA	The NDEA would create five marketing areas that
together would encompass the entire nation. States in the Northeast,83
Southern,84 and Upper Midwest85 regions would automatically be
participating in the marketing area program established by the NDEA upon
enactment of the legislation. States in the Intermountain86 and Pacific87
regions could become participating states by providing written notice to
the Secretary of Agriculture. The NDEA would authorize each region's board
to set an "overorder" minimum price for Class I sales that exceeded the
FMMO Class I price in that region, with an initial maximum of $17.50 per
hundredweight, subject to approval by farmers within the region in a
referendum.88 Although the boards would have discretion in setting the
over-order price, the legislation directs the boards to consider several
factors including the

o 	balance between production and consumption of milk and milk products in
the regulated area;

o  costs of milk production in the regulated area;

o  prevailing price for milk outside the regulated area;

o  purchasing power of the public; and

o  price necessary to yield a reasonable return to an eligible farmer.

The NDEA would establish a fund in the U.S. Treasury to carry out the
program. During months in which a region's over-order price exceeded the

83States in this region include Connecticut, Delaware, Maine, Maryland,
Massachusetts, New Hampshire, New Jersey, New York, Ohio, Pennsylvania,
Rhode Island, and Vermont.

84States in this region include Alabama, Arkansas, Florida, Georgia,
Kansas, Kentucky, Louisiana, Mississippi, Missouri, Nebraska, North
Carolina, Oklahoma, South Carolina, Texas, Tennessee, Virginia, and West
Virginia.

85States in this region include Illinois, Indiana, Iowa, Michigan,
Minnesota, North Dakota, South Dakota, and Wisconsin.

86States in this region include Arizona, Colorado, Idaho, Montana, Nevada,
New Mexico, Utah, and Wyoming.

87States in this region include Alaska, California, Hawaii, Oregon, and
Washington.

88The NDEA would allow the maximum to increase over time in accordance
with changes in the Consumer Price Index for All Urban Consumers published
by the Bureau of Labor Statistics in the Department of Labor.

Appendix VII Effects of Recent Federal Dairy Program Changes and
Alternative Policy Options

FMMO Class I minimum price in Boston, processors would be required to pay
the Secretary of Agriculture an amount equal to the difference between
those two prices, known as the over-order premium, times the quantity of
milk purchased for use in Class I products. The Secretary would deposit
these amounts into the U.S. Treasury fund, and the fund would make
payments to each board, which would distribute the payments to eligible
farmers in its region.89 Each month the board would receive at a minimum
an amount equal to the over-order premium times 50 percent of the milk
produced in that region.90 The proposed legislation would require CCC
funds to be transferred to the fund when necessary to allow the fund to
make the required payments; according to one analysis, such contributions
would typically be necessary because only one of the proposed dairy
marketing areas has a Class I utilization rate equal to 50 percent or
higher.

The NDEA also would require boards to compensate the CCC for any
additional costs of CCC purchases of milk products resulting from
increases in milk production that exceed the national average growth rate.
To manage overproduction of milk that could result from the NDEA, the NDEA
would authorize boards to take action, including developing and
implementing incentive-based supply management programs.

In addition, the NDEA would link participation in the dairy marketing
areas with participation in the current MILC program. Farmers who
participate in the new program would not be able to continue to receive
MILC payments. If states in the Intermountain and Pacific regions chose
not to participate in the NDEA program, farmers in those states could
continue to receive MILC payments, and the NDEA would extend the
authorization period for the MILC program until the end of September 2007.
States in the other regions, which would become participants upon
enactment of the NDEA, could withdraw their participation. If they did,
farmers in those states could also continue their MILC payments through
the end of September 2007. Individual farmers in states that participated
in the NDEA program could

89Other payments from the fund would include compensation for
administrative costs that USDA or the boards incur in carrying out the
program, increased costs of specified federal nutrition programs that
might result from the program, and increased costs that the states might
occur due to the program in carrying out the Child Nutrition Act of 1996.
Boards could also collect assessments from processors in their region not
to exceed $0.03 per hundredweight to cover the boards' administrative
costs.

90Each month the board would receive the greater of (1) this amount or (2)
the amount of payments made by processors to the fund for purchases of
Class I milk that will be sold in the region during the month.

  Appendix VII Effects of Recent Federal Dairy Program Changes and Alternative
                                 Policy Options

choose not to participate in the new program and would then be able to
continue receiving MILC payments. However, those farmers would not be able
to extend their payments beyond September 2005 and would not be eligible
for subsequent participation in the NDEA program.

Impacts of the NDEA	Although the NDEA may lead to higher incomes for some
dairy farmers, academicians and industry participants have raised many
concerns about the proposed legislation's impacts. These concerns include
regional divisiveness due to lower incomes for dairy farmers in some
regions, higher retail prices, reduced coordination of dairy prices across
regions, and potential conflict with World Trade Organization (WTO) rules.

The NDEA could lead to higher incomes for farmers in participating states
when milk prices are relatively low because processors would have to pay
the over-order prices set by the boards for Class I milk; however, the
total effect on farmers would depend on what happens to the price of milk
used for manufacturing purposes as well.91 To the extent that higher blend
prices stemming from higher Class I prices lead farmers to increase milk
production, the result could be lower prices for Class III and IV milk
both in participating and nonparticipating states. Even if all states
participated in the dairy marketing areas, the large differences in Class
I utilization rates across regions imply that different regions would be
affected differently by the combination of higher Class I prices and lower
Class III and IV prices, and farmers in regions with low Class I
utilization rates might see a decline in their incomes. If some states do
not participate, their farmers would be even worse off because unless
these farmers pool milk in states that are included in the marketing
areas, they would not receive any benefits of higher Class I prices. In a
report that we issued in September 200192 in which we analyzed the
inter-regional impacts of various scenarios in which some states were
grouped in dairy compacts that functioned like the NDEA's dairy marketing
areas, we reported that one effect of compacts was to reduce farm income
in noncompact regions. We estimated this effect to be minimal when we
examined the impact of the Northeast Interstate Dairy Compact because the
six New England states included in that Compact

91Even the extent to which income from Class I sales might rise from the
NDEA is hard to estimate because it is unclear to what extent the
over-order premiums that processors would be required to pay for Class I
milk might be replacements for market-driven overorder premiums that would
exist without the NDEA.

92GAO, Dairy Industry: Estimated Economic Impacts of Dairy Compacts,
GAO-01-866 (Washington, D.C.: Sept. 14, 2001).

Appendix VII Effects of Recent Federal Dairy Program Changes and
Alternative Policy Options

produced only 3 percent of the nation's milk. However, we estimated that
the effect was somewhat greater in a scenario in which states producing 27
percent of the nation's milk supply were included in compacts.

To the extent that the NDEA would result in fluid milk processors paying
more to buy their milk from farmers, the NDEA would also lead to increases
in retail fluid milk prices. In our report on compacts, we reported that
several studies concluded that the Northeast Interstate Dairy Compact
resulted in higher retail prices for fluid milk in New England, with
estimated impacts ranging from $0.03 to $0.20 per gallon. Higher retail
prices could have a greater effect on retail sales in upcoming years than
occurred in the past, as some dairy experts believe that the demand for
fluid milk has become more responsive to price changes, given the
increasing number of beverages that are considered substitutes for fluid
milk, among other reasons. Consequently, retailers with whom we spoke
generally opposed the NDEA. Furthermore, declines in fluid milk sales
would cause more milk to be available for manufacturing purposes, which
would further depress the prices for Class III and IV milk.

Several academicians told us that they believed the NDEA would also create
regional distortions because price-setting in each dairy marketing area
would be controlled by its board, and prices for raw milk used in fluid
products would no longer be closely linked to prices for raw milk used in
manufactured products. This would be a major change from the current
system in which Class I prices are set based on differentials added to the
"higher of" the advanced Class III or IV skim milk values, with the
differentials still somewhat reflective of the costs of transporting milk
from the Upper Midwest, a key dairy surplus region. Before the 1960s,
Class I prices in different orders were not coordinated, and the resulting
disorderly marketing system led to the coordinated system that we now
have. Adopting the decentralized price-setting system of the NDEA risks
losing the advantage of more orderly marketing that the coordination of
the 1960s brought to the dairy industry.

Concerns about whether the NDEA would make U.S. dairy policy less
consistent with existing agreements under the WTO arise because of the
effects of the NDEA on milk production and, hence, U.S. milk prices. As
indicated previously, recent U.S. commitments under the General Agreement
on Tariffs and Trade are leaning in the direction of more liberalized
trade. To the extent that the NDEA provides a subsidy for U.S. milk
production and reduces the prices of manufactured dairy products, the act
would reduce the competitiveness of imported products.

  Appendix VII Effects of Recent Federal Dairy Program Changes and Alternative
                                 Policy Options

We identified one study that estimates the effects of the NDEA on milk
production, farm prices, and government costs compared to a baseline
scenario that did not include the NDEA.93 This study estimated that the
NDEA would increase milk production compared to the baseline by an average
of about 7.6 billion pounds per year during the period from 2006 through
2013. The increased production would result in estimated declines in Class
III and IV prices from the baseline such that average milk prices would be
$1.17 per hundredweight below the baseline estimate. The estimated annual
average increase in federal costs for payments to the boards was $1.7
billion.94

Change Trade Restrictions and Export Incentives

Recent concerns about the effects of imported dairy products, most notably
milk protein concentrates, on U.S. dairy prices have highlighted the
importance of U.S. trade policy-trade restrictions and subsidy programs-as
a foundation for domestic dairy policies. Several policy options related
to international trade in dairy products have been suggested. As noted
previously, current international trade agreements and ongoing
negotiations can have implications for certain policy options that have
been suggested. These options include (1) increasing trade restrictions,
specifically for imports of milk protein concentrates;95 (2) relaxing
trade restrictions; (3) introducing domestic subsidies for products
significantly affected by international trade competition, specifically
establishing a subsidy program for domestic production of milk protein
concentrates; and (4) changing the Dairy Export Incentive Program (DEIP),
either by using it more effectively or by eliminating it. Those options
that succeed in limiting imports or encouraging exports of manufactured
dairy products could support higher farm income, production levels, and
consumer prices. These options may also reduce

93Scott Brown, "An Introductory Examination of the National Dairy Equity
Act," in Dairy Policy and Product Innovation, 11th Annual Workshop for
Dairy Economists and Policy Analysts, April 15-16, 2004.

94These estimates were developed under a particular set of assumptions,
including that all states would participate in the NDEA and that
over-order prices were set relatively high. These estimates, as
acknowledged in the study, are likely to cause the estimated effects of
the NDEA to be larger than if, alternatively, it had been assumed that few
states participated and the over-order prices were relatively low.

95Milk protein concentrate is a concentrated milk protein product that
contains both of the major forms of protein found in milk: casein and
whey. The protein content of this product can vary considerably from 42 to
over 90 percent.

  Appendix VII Effects of Recent Federal Dairy Program Changes and Alternative
                                 Policy Options

federal costs if higher farm prices reduce costs to the price support
program to a greater degree than the cost of these trade options. Figure
40 shows the effects of options to change trade restrictions and export
incentives over the short and long terms.

Figure 40: Potential Effects of Changing Trade Restrictions and Export
Incentives on Various Policy Considerations

Increase

Decrease

Mixed impacts ? No/unclear impacts LP: Low price scenario HP: High price
scenario Source: GAO analysis based on interviews with dairy experts and
reviews of relevant studies.

Note: The figure does not indicate the degree of increase or decrease for
each policy consideration as it relates to a policy option. In addition,
as discussed in this appendix, limitations affecting some policy options
could affect the degree of an option's potential impact on a policy
consideration.

Increase Trade One international trade policy that has been proposed is to
increase trade Restrictions/Establish a Tariff-restrictions within the
constraints of existing international trade Rate Quota for Milk Protein
agreements. More specifically, a bill entitled the Milk Import Tariff
Equity Concentrates

Appendix VII Effects of Recent Federal Dairy Program Changes and
Alternative Policy Options

Act96 has been introduced in the Congress that would impose tariff-rate
quotas on dairy protein products such as milk protein concentrates and
certain casein products. Import quotas prior to the WTO Uruguay Round
Agreement on Agriculture did not cover milk protein concentrates and
casein.97 Therefore, no tariff-rate quota was established for these
products after the agreement was implemented. The May 2004 report by the
International Trade Commission showed that U.S. imports of some dairy
proteins increased significantly from 1998 through 2000 and then declined.
Some of these protein imports displaced domestic dairy proteins,
particularly those used in making processed cheese products not covered by
the Food and Drug Administration's standards of identity.98

The International Trade Commission report concluded that imports of milk
protein concentrates to the United States have the effect of lowering U.S.
farm prices either directly or indirectly, depending upon whether U.S.
market prices for manufactured dairy products are above or at the level of
the support price. If U.S. prices are above the support price, then
imports of dairy proteins could directly lower the market prices of nonfat
dry milk, butter, and cheese to the extent these proteins can be imported
at lower prices than proteins available in the domestic market. In turn,
lower product prices could reduce the prices received by U.S. farmers for
their raw milk. If U.S. prices are at the support price, then imports of
dairy proteins could indirectly affect U.S. market prices. Increasing
imports of proteins when U.S. market prices for manufactured dairy
products are at the support price will cause the CCC to purchase more
nonfat dry milk as this alternative protein source is displaced in the
market. Eventually, these increasing stocks could cause USDA to lower the
purchase price of nonfat dry milk in an attempt to reduce federal costs.
This adjustment

96This proposed legislation was introduced in the 108th Congress as S. 560
and H.R. 1160.

97As a result of the agreement, the United States committed to convert
quotas and fees that had been established pursuant to section 22 of the
Agricultural Adjustment Act of 1933 into tariffs in the form of
tariff-rate quotas. Further, the United States agreed not to use section
22 in the future against imports from other WTO members. Virtually all
section 22 quotas were converted to tariff-rate quotas in 1995.

98Standards of identity were introduced under the 1938 Federal Food, Drug,
and Cosmetic Act and are enforced by the Food and Drug Administration.
Standards of identity define food products by establishing what
ingredients must or may be used in the food manufacturing process, as well
as the quantity of such ingredients. In most cases, milk protein
concentrate is not accepted for use in cheeses covered by the standards of
identity.

  Appendix VII Effects of Recent Federal Dairy Program Changes and Alternative
                                 Policy Options

would then lower prices for manufactured dairy products, in turn
decreasing the prices received by U.S. farmers for their raw milk.99

Because introducing tariff-rate quotas on milk protein concentrate and
casein would likely reduce imports of these products, federal costs due to
CCC purchases and storage of nonfat dry milk under the price support
program would likely decrease. When U.S. market prices are above the
purchase prices established by the price support program, a reduction in
imports could maintain prices of some products, such as cheese, at higher
levels due to reduced domestic supply of dairy proteins used in these
products. Higher product prices could increase farm prices and thus
stimulate additional production. These effects could be more significant
in regions of the United States where raw milk is used to a greater extent
in the manufacturing of Class III products. (If U.S. trading partners
successfully challenge tariff-rate quotas on dairy proteins, compensation
such as increased market access for other products could be required,
reducing the benefits to U.S. dairy farmers.) Because production
adjustments tend to lag behind price changes, this additional production
could delay adjustments to lower market prices in the future. Therefore,
to the extent that the exclusion of imports masks market price signals
that would exist without the exclusion, this policy option would decrease
economic efficiency. Additionally, to the extent that changes in the
manufacturing prices of dairy products are passed on through the retail
level, consumers could experience higher prices for some products (such as
cheese) and lower prices for other products (such as butter).100

Relax Trade Restrictions	A second option is to relax trade restrictions by
reducing or eliminating tariffs on dairy products. Trade restrictions such
as tariff-rate quotas support domestic programs such as the FMMO
classified pricing system and the dairy price support program by limiting
the available supply of dairy products to take advantage of higher U.S.
market prices. Unilaterally

99Despite its conclusion that imports of dairy proteins to the United
States could lower manufactured product prices, the International Trade
Commission was unable to conclude that these imports actually caused USDA
to tilt commodity purchase prices in May 2001 and November 2002. The
International Trade Commission found that between 1996 and 2002, CCC
stocks of nonfat dry milk grew from zero to 1 billion pounds. However, of
this amount, the commission estimated that only 25 to 35 percent of the
stocks came from nonfat dry milk that had been displaced by imports of
dairy proteins. Therefore, 66 percent of the increase in CCC stocks was
due to factors other than imports.

100Butter prices could decrease as higher production levels, among other
things, increase the supply of butterfat.

Appendix VII Effects of Recent Federal Dairy Program Changes and
Alternative Policy Options

relaxing U.S. trade restrictions would likely increase imports of
manufactured products as foreign producers seek to take advantage of the
higher prices available in the U.S. market. Despite higher transportation
costs for imported products, some manufacturers might be able to import
certain dairy products from U.S. trading partners at prices below U.S.
market prices either because those partners provide export subsidies (as
the European Union does) or because they have lower milk production costs
(as Australia and New Zealand do). Relaxing trade restrictions such as
tariff-rate quotas is unlikely to increase imports of fluid milk products
because of health restrictions, transportation costs, and the perishable
nature of these products.

Increased imports could put pressure on the CCC to purchase larger
quantities of manufactured dairy products, thereby increasing federal
costs.101 At some point these pressures could become unsustainable,
leading to a reduction in the support price or the end of the price
support program and to a decline in the price of milk used for
manufacturing purposes. Moreover, because the price of milk used in fluid
milk products is based on the price of milk used in manufactured products,
a decline in the prices of manufactured products from increased imports
could lower average farm prices. As long as MILC is authorized, these
lower prices could trigger additional MILC payments to farmers, further
increasing government costs.

In the short term, increased imports could cause U.S. prices to fall,
resulting in a decline in farm income for U.S. dairy farmers. With reduced
farm income, production would also decrease as less efficient farmers exit
production. Over the long term, the decline in production could cause farm
prices to rebound toward the levels that existed before trade restrictions
were relaxed. Similarly, in the short term, lower farm prices are likely
to lead to lower consumer prices for fluid milk and other dairy products,
but as farm prices rise toward their previous levels due to production
decreases, consumer prices for fluid milk and other dairy products would
likely rise as well. Economic efficiency will increase with relaxed trade
restrictions because such relaxation will allow market price signals to be
more visible than with restrictions in place. These signals will lead to
increased imports when it is cheaper to substitute increased imports for

101The extent to which increased imports put pressure on the CCC to
increase its purchases depends on the extent to which increased imports
replace commercial sales of domestically produced dairy products.

  Appendix VII Effects of Recent Federal Dairy Program Changes and Alternative
                                 Policy Options

some domestic dairy production. Additionally, the economic efficiency of
U.S. dairy production resource allocation could increase with relaxed
trade restrictions as the reduction in domestic production is more likely
to come from less efficient domestic farmers.

Subsidize Domestic Milk Protein Another proposed option is to support the
development of domestic casein

Concentrate Production	and milk protein concentrate production. Under the
proposed U.S. Dairy Proteins Incentive Program, the CCC would make subsidy
payments, on a bid basis, to entities that produce and market dairy
proteins from skim milk.102 The proposed legislation would provide, among
other things, that receipt of a payment is contingent upon the end use of
the dairy proteins produced; that no applicant receives a payment if the
contract submitted for review would undercut domestic prices for milk,
nonfat dry milk, or dairy proteins; and that the sale of the dairy
proteins represents a new use of domestically produced dairy proteins.

This program's potential impact on domestic dairy protein production
depends on the relative profitability of these proteins, which, in turn
depends on production costs and demand. The International Trade
Commission's May 2004 study on milk protein products found that, given
disincentives inherent in the price support program and constraints on
U.S. demand for dairy proteins other than nonfat dry milk, the
profitability of domestic protein production could be limited. For
example, the study reported that the price support program creates a
disincentive for U.S. processors to produce dairy proteins other than
nonfat dry milk because by purchasing nonfat dry milk the price support
program reduces the financial risk of manufacturing that product.
Processors of other proteins would need to invest in production facilities
and then market their product without the benefit of a standing government
offer of support. The study found that only under the most favorable
conditions (high skim milk protein yield103 and low variable costs) would
it be beneficial for U.S. processors to begin producing milk protein
concentrate instead of nonfat

102A bill to create this program was introduced in the U.S. House of
Representatives as H.R. 4223 on April 27, 2004. The bill defines dairy
proteins as whey, whey protein concentrate, casein, or milk protein
concentrate.

103The protein content of raw milk can change with the age of the cow,
seasonally, and with feeding and management practices.

Appendix VII Effects of Recent Federal Dairy Program Changes and
Alternative Policy Options

dry milk. Even then, positive returns were only for milk protein
concentrates with protein concentrations above 70 or 80 percent.104

The classified pricing system could create an impediment to the
development of a domestic protein industry depending upon how milk protein
concentrate is classified. Based on analysis presented in the
International Trade Commission's report, classification under a
highervalued class would require producers of milk protein concentrate to
pay more for their raw milk supplies, thus reducing their profits.105 In
addition, the report noted that since May 2002, the CCC has had a program
to provide incentives to convert nonfat dry milk held in its stocks to
casein. Under this program, the CCC accepts competitive bids for CCC-owned
nonfat dry milk stocks for the manufacture of casein. However, while USDA
has accepted some bids, in many cases processors' bids have been so low
that USDA has rejected them.

Finally, the International Trade Commission's report found that while milk
protein concentrate is considered a useful additive to standardize protein
content, the limitation on its use inherent in the Food and Drug
Administration's standards of identity further restrict domestic milk
protein concentrate production. This limitation keeps the market for milk
protein concentrates relatively small in comparison to the market for
other dairy proteins. Given these restrictions, the International Trade
Commission estimated that the total U.S. market for milk protein
concentrate is 40,000 to 50,000 metric tons per year. A new production
facility in New Mexico reportedly is capable of producing 16,000 tons
annually. Therefore, barring a large drop in imports or changes to the
standards of identity, the demand for milk protein concentrate would have
to increase substantially to induce additional domestic production.

To the extent that the proposed program can overcome these challenges, it
could provide incentives for manufacturers to produce alternative dairy
proteins domestically. Should these proteins replace some nonfat dry milk
production, they could reduce federal costs for the price support program.

104The International Trade Commission reached these conclusions by
estimating the price, cost, and returns for U.S. producers who produce
nonfat dry milk and sell it to the CCC. The Commission then compared the
result with what producers in its survey had earned from producing and
selling milk protein concentrate.

105However, USDA officials noted that this may be a moot issue because
they believe milk protein concentrates will be classified as Class IV
products.

  Appendix VII Effects of Recent Federal Dairy Program Changes and Alternative
                                 Policy Options

However, the overall impact on federal costs would depend on whether these
reductions are offset by the cost of the subsidy program itself. Various
dairy experts disagree as to whether subsidizing domestic dairy protein
production would result in a net increase or decrease in federal costs.
For example, a study that was published in May 2004 by researchers at
Cornell University concluded that reduced costs to the price support
program would not be great enough to offset the cost of the subsidy
program.106 Conversely, an analysis by the National Milk Producers
Federation found that a protein subsidy program providing assistance up to
$2.30 per hundredweight of skim milk would result in a net cost savings
for the federal government. The study by Cornell University researchers
also estimated that a subsidy program for casein and milk protein
concentrate would raise average milk prices by $0.40 per hundredweight,
yielding an increase in farm income of $913 million. These increases would
have greater impacts on farm income and milk production in areas of the
country with higher Class IV utilization, such as the West.

Also, should these proteins replace some nonfat dry milk production, the
resulting effects on the prices that consumers pay for products made with
dairy proteins could be mixed. If dairy protein prices increase with
reduced imports, consumer prices for those products for which they are an
ingredient (such as cheese), could also increase to the extent that these
price changes are passed through. However, additional production resulting
from higher farm prices could lower consumer prices for butter. Over the
long term, increased domestic production of casein and milk protein
concentrate could lower their production costs and also the costs of other
products for which they are ingredients. With the subsidy program in
place, the economic efficiency of resource allocation would likely
decrease as the government provides incentives for the production of
proteins that could potentially be supplied more cheaply through imports.

Change the Dairy Export Some dairy experts indicated that DEIP can help
reduce government

Incentive Program	expenditures by allowing USDA to subsidize exports
rather than purchase products and maintain high stock levels through the
price support program. However, in some cases, academic experts indicated
that under current market conditions the impact of DEIP on U.S. prices is
limited. For

106Phillip M. Bishop and Charles F. Nicholson, "Dairy Market Impacts of
U.S. Milk Protein Imports and Trade Policy Alternatives," Department of
Applied Economics and Management, Cornell Program on Dairy Markets and
Policy, R.B. 2004-08, Cornell University, Ithaca, N.Y. (May 2004).

Appendix VII Effects of Recent Federal Dairy Program Changes and
Alternative Policy Options

example, one source noted that with CCC purchases of nonfat dry milk at
over 800 million pounds in 2002, and with over 1 billion pounds of CCC
stocks on hand, DEIP currently has no impact on U.S. prices. Also, WTO
commitments have limited the scope of DEIP. For example, in the Uruguay
Round Agreement on Agriculture the United States committed to reducing the
quantity of subsidized exports by 21 percent and the value of these
exports by 36 percent over the period from 1995 to 2000. Therefore, a
couple of alternative policy options have been discussed with regard to
DEIP, using it more effectively or eliminating it entirely.

Use DEIP More Effectively

Some dairy experts suggested that the government could make greater use of
DEIP.107 The American Farm Bureau study concluded that DEIP may be
underutilized, noting that it is difficult to develop foreign markets
unless a commitment is made to serving the market. Such a commitment is
more difficult if a given product is made available only when it is in
surplus. The study criticized USDA for (1) being slow to invite and accept
bids and (2) concentrating on products in surplus. Other dairy experts
indicated that invitations for DEIP bids may be announced too late in the
year for potential exporters to participate in the program due to seasonal
sales patterns. To improve the effectiveness of DEIP, the American Farm
Bureau study recommended three potential changes:

o 	Exporters should be encouraged to submit bids for products and
countries that offer the greatest potential for longer-term market
development. USDA should use DEIP in conjunction with the Foreign
Agricultural Service to coordinate export assistance programs to fully
develop markets.

o 	USDA should consider DEIP bids for any eligible products and not base
acceptance primarily on removing surplus products from the domestic
market. Bids should be accepted for products that may have the greatest
market development potential and do not violate WTO subsidization volume
limits.

107As an alternative to government export subsidies, the National Milk
Producers Federation's Cooperatives Working Together program includes,
among other things, an export subsidy element.

Appendix VII Effects of Recent Federal Dairy Program Changes and
Alternative Policy Options

o 	USDA needs to act under shorter time frames in reviewing and accepting
DEIP bids to maximize the volume allowable under WTO rules.

USDA indicated that it has announced and awarded subsidies under the DEIP
program up to the limits allowed by WTO rules for nonfat dry milk and
cheese and that DEIP-assisted exports of butterfat have varied depending
on market conditions. USDA also noted that the Foreign Agricultural
Service has worked closely with the dairy industry to ensure that national
and annual DEIP assistance optimizes longer-term market development
prospects and minimizes any potential detrimental effects on the U.S.
market. For example, a Foreign Agricultural Service official stated that
the Service generally tries to wait until the market for a particular
product, such as butterfat, is in surplus before inviting bids for DEIP
export subsidies. The official said that if the market is not in surplus,
subsidized exports would increase U.S. market prices for products
manufactured with butterfat, such as ice cream. USDA further indicated
that expanding the use of DEIP is not possible as the program is bound by
quantitative and monetary caps under WTO rules. Finally, USDA noted that
the bid review and acceptance process occurs within a time span equivalent
to less than one working day. Specifically, another Foreign Agricultural
Service official stated that USDA responds to all bid proposals by 10:00
a.m. on the next business day after the proposals are submitted.

To the extent that USDA could identify ways to make greater use of DEIP,
the effects on the dairy industry of increased exports of U.S. dairy
products could be similar to the effects of an increase in demand.108 In
the short term, greater demand for dairy products would increase wholesale
prices, raising farm prices and, to the extent that these changes are
passed on to consumers, retail prices. While higher retail prices could
cause marginal declines in domestic consumption, higher farm prices over
the long term could stimulate additional production, which could put
downward pressure on wholesale and retail prices. Under such a scenario,
price volatility could decrease as the government balances swings in
domestic manufactured product prices by adjusting its level of support for
DEIP. However, economic efficiency would decrease to the extent that
increased use of DEIP induces additional production that would not have
occurred

108As is the case with other export subsidy programs, the degree to which
demand is increased under a program such as DEIP depends on the degree to
which the exports under the program are additional to those that would
have occurred in the absence of the program.

Appendix VII Effects of Recent Federal Dairy Program Changes and
Alternative Policy Options

without the government program. Finally, in relation to federal costs,
while increased use of DEIP could increase federal costs for the program,
these increases might be offset by decreases in the costs of purchasing
and storing commodities under the price support program. Further, greater
competition for milk supplies used in manufactured products could increase
Class I prices, decreasing MILC payments while the MILC program remains in
existence.

Eliminate DEIP

Eliminating DEIP could lower U.S. market prices and increase government
costs, specifically for nonfat dry milk.109 For example, the May 2004 USDA
study estimated that exports of nonfat dry milk under DEIP account for
approximately 9 percent of total U.S. production. The study further found
that in comparison to just eliminating the price support program,
eliminating DEIP as well would reduce wholesale nonfat dry milk prices
another 5 percent below baseline levels. Under this scenario, the study
reported that for 2002 through 2007, farm income would decline by
approximately $5.3 billion and payments required under the MILC program
would increase by approximately $900 million. Therefore, while ending DEIP
would eliminate the costs of the subsidies provided by the program, these
cost savings could be offset by increased MILC expenditures. Also, should
DEIP be eliminated without eliminating the price support program, there
would be an increase in federal costs for purchases of commodities that
would otherwise have been exported.

The American Farm Bureau study reported that in the short term eliminating
DEIP would cause dairy products formerly exported under the program to be
commercially exported. However, the study found that over the long term
dairy product prices in the U.S. market would be too high relative to
world prices to allow formerly subsidized products to move as commercially
exported products. In the short term, a decline in the prices of some
manufactured products formerly exported under DEIP, such as nonfat dry
milk and cheese, would lower the blend prices farmers receive which, in
turn, would cause a decline in milk production. This reduced production
could cause butter prices to increase. However, the overall effect is a
decrease in both average milk prices and milk supplies. With lower prices
for some manufactured dairy products and higher prices for

109The degree to which the U.S. market would be affected depends on the
degree to which the exports under DEIP would be eliminated if the program
were eliminated.

  Appendix VII Effects of Recent Federal Dairy Program Changes and Alternative
                                 Policy Options

others, consumer prices may rise or fall depending on the extent to which
these price changes are passed on by retailers. In other respects, the
elimination of DEIP could increase price volatility because USDA would
lose an outlet to help balance surplus production. However, the supply
adjustment could marginally increase economic efficiency as excess supply
is wrung out of the market. Further, to the extent that DEIP is providing
incentives for nonfat dry milk production rather than production of other
dairy products, elimination of these incentives would also marginally
increase economic efficiency.

Facilitate Risk Management for Dairy Farmers

The increased volatility in farm milk prices has increased dairy farmers'
interest in managing the risk that low prices will reduce farm incomes.
Risk management alternatives to stabilize dairy farmers' incomes can take
many forms. These include, among others, increased use of forward
contracting to guarantee prices, such as through USDA's Dairy Forward
Pricing Pilot Program;110 revenue insurance policies that pay farmers when
dairy proceeds fall below specified levels; and tax-deferred savings
incentives that encourage setting money aside during higher income years
that could be withdrawn during lower income years. Figure 41 shows the
effects of options to facilitate risk management under low-and high-price
scenarios over the short and long terms.

110Trading in milk futures and options represents another alternative for
farmers to reduce the risks associated with price volatility by locking in
prices, or setting price floors, for future production. This alternative
would remain available to farmers even if some dairy programs, such as the
dairy price support program, were eliminated.

  Appendix VII Effects of Recent Federal Dairy Program Changes and Alternative
                                 Policy Options

Figure 41: Potential Effects of Options to Facilitate Risk Management on
Various Policy Considerations

    Dairy forward pricing program           ?  ?     ?  ?      ?  ?     ?  ?  
          Revenue insurance         LP            ?  ?               ?  ?  
                                    HP   ?  ?                              
        Tax-deferred savings        LP            ?  ?               ?  ?  
                                    HP   ?  ?                              

Increase

Decrease

Mixed impacts ? No/unclear impacts LP: Low price scenario HP: High price
scenario Source: GAO analysis based on interviews with dairy experts and
reviews of relevant studies.

Note: The figure does not indicate the degree of increase or decrease for
each policy consideration as it relates to a policy option. In addition,
as discussed in this appendix, limitations affecting some policy options
could affect the degree of an option's potential impact on a policy
consideration.

Continue Dairy Forward Pricing Forward contracting of milk-entering into a
contract with a processor or

Program	cooperative to sell milk in the future at a guaranteed price-is
one way for farmers to manage the risk that volatile prices create for
their income. Although forward contracting may prevent farmers from
benefiting from price increases, this risk management tool stabilizes
their income and ensures that the price they receive does not fall below
the contracted price. In this respect, forward contracting, like the MILC
program, limits the decline in farm income from a fall in farm milk
prices.

Most, but not all, dairy cooperatives offer their members the ability to
enter into forward contracts to guarantee the prices members will receive
for

Appendix VII Effects of Recent Federal Dairy Program Changes and
Alternative Policy Options

their milk. One option that could allow dairy farmers to make more use of
forward contracting to manage their price risk would be to extend and
expand the Dairy Forward Pricing Pilot Program to cover Class I milk. The
program was mandated by the Congress in 1999 and is scheduled to expire at
the end of 2004.111 Under the pilot program, farmers are not allowed to
enter into forward contracts for Class I milk. However, if the program
were extended or made permanent, it could also be expanded to allow
forward contracting on all classes of milk.

The forward contracting pilot program offers farmers who sell to
proprietary processors an option to enter into fixed-price forward
contracts by providing processors with an exemption from paying the
otherwise required FMMO minimum prices. (Dairy cooperatives are already
exempt from paying the FMMO minimum prices.)112 Normally, when either
cooperatives or proprietary processors buy milk from farmers under forward
contracts, they can offset their risk that farm prices might be lower at
the end of the contract period (in which case they would pay more for milk
than their competitors who buy later at the lower price) with a futures
market transaction in which they would gain an amount equivalent to the
decrease in the farm price of milk.113 However, when prices rise during
the contract, the cooperatives or processors will lose money on the
futures market transactions and cannot afford to pay farmers more than the
contracted price. The pilot program encourages proprietary processors to
enter into forward contracts with farmers by exempting these processors
from having to pay farmers the relevant order minimum

111Legislation has been introduced in the U.S. House of Representatives
and the U.S. Senate to make permanent the authority for the Dairy Forward
Pricing program, H.R. 3308 (2003) and S. 2565 (2004).

112As a practical matter, it could be difficult for cooperatives to
continue paying their members less than the regulated minimum price for an
extended period of time because farmers could choose to market their milk
with a different cooperative. However, farmers' ability to switch
cooperatives could be limited by the extent to which individual
cooperatives dominate milk marketing over a wide area.

113Most forward contracts for milk are for a base Class III price, which
in multiple component pricing orders, is the sum of milk component values
per hundredweight of milk at standardized test. Class III milk futures and
options contracts are actively traded on the Chicago Mercantile Exchange.
To offset their price risk from the forward contracts to buy milk,
proprietary processors and cooperatives can take an equal but opposite
position in the futures market by selling futures contracts for an
equivalent amount of Class III or IV milk. Then, if the price of farm milk
falls during the life of the contract, when the contract period ends,
their gain in the futures market will equal their loss from buying milk
under a fixedprice forward contract rather than buying after the price
fell.

Appendix VII Effects of Recent Federal Dairy Program Changes and
Alternative Policy Options

price for the portion of their milk that is under forward contract.
Without this program, proprietary firms might not readily enter into
forward contracts because if they offset their risk of a price decline in
the futures market and the prices were to rise above the contract price,
they might have insufficient funds-including their futures market loss-to
pay their forward-contracted farmers the minimum price.

Some cooperatives have opposed proposed legislation that would make
permanent the authority for forward contracting by proprietary processors
on the grounds that allowing those processors to pay farmers less than the
minimum price could undermine the federal order pricing system. However,
one academic analysis of options to address price volatility indicated
that this exemption would not undermine the federal order pricing system
because while proprietary processors could pay farmers less than the
minimum price, these processors would still have to pay minimum class
prices into their federal order pools.114 Another academician told us that
even if allowing forward contracting on Class I milk caused farmers to
receive lower minimum prices in return for reduced risk, neither that
price reduction nor any other rationale would be a good reason for not
making fluid milk eligible for forward contracting by those farmers who
want to enter into such contracts.

In October 2002,115 USDA released a report that examined the performance
of the pilot program from its inception, in September 2000, through March
2002. The report found that the average monthly price received for milk
sold under forward contracts authorized by the pilot program was lower
than the average monthly price that would have been received for that same
milk if it had not been under contract by about $0.50 per hundredweight,
but that the variation in price for milk sold under forward contracts was
much less. At times the contract price exceeded the price that would have
been received without the contracts, and at times it was lower. More
recent USDA data show that the contract price exceeded the price that
would have been received without the contracts in each month

114Bob Cropp, "Innovations To Address Price Risk Management and Price
Volatility" in Dairy Policy and Product Innovation, 11th Annual Workshop
for Dairy Economists and Policy Analysts, April 15-16, 2004.

115"A Study of the Dairy Forward Pricing Pilot Program and Its Effect on
Prices Paid Producers for Milk," prepared for the Senate Committee on
Agriculture, Nutrition and Forestry and the House Committee on Agriculture
by Dairy Programs, Agricultural Marketing Service, U.S. Department of
Agriculture.

  Appendix VII Effects of Recent Federal Dairy Program Changes and Alternative
                                 Policy Options

from April 2002 through July 2003, but for the remainder of 2003 the
contract price was lower. On average, for the entire period from April
2002 through December 2003, the contract price was about $1.40 per
hundredweight higher.

Participation in the program during the period covered by the USDA report
was relatively small and was concentrated most heavily in the Upper
Midwest Order, which has a low Class I utilization rate. More recent USDA
data show that participation remained relatively low through the end of
2003. Farmers in orders with high Class I utilization rates had less
opportunity to participate because Class I milk was ineligible, so it is
uncertain whether participation rates in those orders would have been
higher if farmers were allowed to enter into forward contracts for Class I
milk. One academic source also noted that even if the program were
expanded to include Class I milk, fluid milk processors might be reluctant
to engage in forward contracting for this milk because there is no good
hedge in the futures market for Class I prices. In addition, the USDA
study reported that participating farmers were generally more accustomed
to using risk management tools than were nonparticipants. Farmer education
on using forward contracting may be important to increase participation.

Provide Revenue Insurance	Another option to help dairy farmers manage
their price risk is revenue insurance. Revenue insurance allows farmers to
protect themselves against loss of revenue from, for example, low market
prices, high feed prices, or reduced production due to natural disasters.
Revenue insurance can stabilize farm income, reducing the need for direct
payments such as under the MILC program, during periods of low prices.
Whether there would be a savings to the government would depend on whether
the subsidies required to induce farmers to participate in the insurance
program would offset the savings from reduced direct income support.
Overall production could increase with this type of option because the
revenue insurance would step in when prices are low representing in
effect, a countercyclical payment. With this type of income support,
downward supply adjustments during periods of low prices might not happen
as quickly. Consumers would then benefit from prolonged periods of low
prices to the extent that price changes are passed through the retail
level.

USDA's Risk Management Agency currently operates several pilot programs in
selected states, using three different approaches to revenue insurance.
Although none of these programs applies to dairy farming, they could, in
theory, be extended to cover dairy farmers. Doing so could be difficult,
however, because the complexity of the dairy industry and the

  Appendix VII Effects of Recent Federal Dairy Program Changes and Alternative
                                 Policy Options

variation in management expertise among farmers would make it hard to
estimate the probability of losses, a calculation that is necessary for
pricing the insurance.

Institute Tax-Deferred Savings	Tax-deferred savings accounts allow farmers
to manage fluctuations in farm income by accumulating cash reserves during
higher income years with deferral of some tax liability. Farmers could
then withdraw from these accounts in lower income years and, in essence,
receive tax benefits if they accumulated funds in these accounts. One
study suggested that farmers might be more comfortable with this risk
management tool than with forward contracting because these accounts
resemble individual retirement accounts and other familiar tax-deferred
savings vehicles.

Similar to revenue insurance, this option has the potential for reducing
direct income support payments from the government, such as MILC payments,
by stabilizing farm income during periods of low prices. However, whether
these accounts worked as a risk management tool would depend on whether
the authorization of these new accounts-with their tax benefits-led to
substantial additional savings by farmers in higher income years, or
whether the new accounts were simply funded with savings that would have
been made anyway in other, not tax-favored, accounts. Also, like revenue
insurance, withdrawals from tax-deferred savings would represent, in
effect, countercyclical payments during periods of low prices. Thus
withdrawals from tax-deferred savings accounts could maintain overall
production by dampening supply adjustments during periods of low prices.
Similarly, consumers would benefit from the prolonged periods of low
prices to the extent that price changes are passed through the retail
level.

Many issues would have to be resolved to start this type of account, such
as the amount that farmers would be allowed to deposit in any year,
whether the government would match any funds deposited, and whether there
would be restrictions on farmers' ability to withdraw funds from the
accounts based on price drops or income losses. Canada and Australia both
offer these accounts, and they were first proposed in the United States in
1996. Since then, there have been several proposals to adopt them here,
but none has been implemented.

Manage Raw Milk Supplies	According to one dairy expert, given the
long-term declining demand for fluid milk as well as the increasing
productivity of dairy farmers, the best way to maintain farm income is
through some form of effective supply

  Appendix VII Effects of Recent Federal Dairy Program Changes and Alternative
                                 Policy Options

management. Following periods of excess supply in the 1980s, the U.S.
government introduced supply management initiatives, such as the Dairy
Termination Program.116 Other options, such as production quota systems,
have been tried by different dairy-producing nations. Thus, a number of
options have been discussed to try to manage dairy supplies, including
reintroducing a program similar to the earlier Dairy Termination Program,
or establishing mandatory supply controls through quota allocations, as
has been done in other countries. Figure 42 shows the effects of options
to manage raw milk supplies under low-and high-price scenarios over the
short and long terms.

 Figure 42: Potential Effects of Options to Manage Raw Milk Supplies on Various
                             Policy Considerations

    Reintroduce Dairy Termination Program               ?                ?  
     Establish a production quota system                                    

Increase Decrease Mixed impacts ? No/unclear impacts Source: GAO analysis
based on interviews with dairy experts and reviews of relevant studies.

Note: The figure does not indicate the degree of increase or decrease for
each policy consideration as it relates to a policy option. In addition,
as discussed in this appendix, limitations affecting some policy options
could affect the degree of an option's potential impact on a policy
consideration.

116Another initiative was the Milk Diversion Program. Both of these
programs were in effect for only a short time.

  Appendix VII Effects of Recent Federal Dairy Program Changes and Alternative
                                 Policy Options

                   Reintroduce the Dairy Termination Program

One option that has been discussed to manage milk supplies is to
reintroduce the Dairy Termination Program.117 This program was first tried
as part of the Food Security Act of 1985. Under the Dairy Termination
Program, farmers submitted competitive bids for the minimum raw milk price
per hundredweight for which they would be willing to comply with the
program requirements. If their bids were accepted, farmers were required
to sell for slaughter or export their entire herds and not participate in
dairying for the next 5 years. The program was in effect from April 1986
through September 1987 and resulted in the removal of more than 1 million
cows, or about 9 percent of the national dairy herd in 1985. In total,
this culling of the national dairy herd was estimated to decrease milk
supplies by about 39.4 billion pounds between 1986 and 1990 at a cost of
more than $1.8 billion.118 California farmers accounted for the largest
portion of this reduced production, but farmers in southeastern states had
the highest rates of participation.

In a 1989 report looking at the effects of the Dairy Termination Program,
we indicated that it was unlikely to have a lasting effect on milk
production, given that some participants would likely return to production
after the 5-year waiting period.119 In the short term, high market prices
resulting from lower levels of production reduced federal purchases of
surplus dairy products. In 1989, we estimated that these reduced purchases
provided a net cost savings to the government of $2.4 billion for fiscal
years 1986 through 1990. However, over the long term, we predicted that
increased production would bring the return of excess milk supply.

The effects of reintroducing this supply management alternative would
depend on a variety of factors. In particular, the effects on federal
costs and milk production would depend heavily on how much farmers needed
to be paid to terminate their herds and agree not to produce for a
specific period of time. Farmers' decisions about whether to participate
in a new Dairy Termination Program at a certain price would rest on the
individual profitability of dairy farms, the long-term production outlook
of the

117The National Milk Producers Federation's Cooperatives Working Together
program, started in July 2003, is an industry-led supply management
initiative. Among other things, it includes a herd retirement element
similar to the Dairy Termination Program.

118Over one-third of this cost was funded by an industry assessment.

119GAO, Dairy Termination Program: An Estimate of Its Impact and
Cost-Effectiveness, GAO/RCED-89-96 (Washington, D.C.: July 6, 1989).

  Appendix VII Effects of Recent Federal Dairy Program Changes and Alternative
                                 Policy Options

individual farmer, and the expectation of certain market conditions. For
example, with today's high market prices, farmers may be unlikely to agree
to stop production except at payment levels that could make the program
prohibitively expensive. During low-price periods, the program could
potentially reduce supply to a level that reduces overall government costs
for both the price support program and the MILC program, as long as the
latter program remains in existence. In either case, the program would
likely increase price volatility and consumer prices, although higher
consumer prices for manufactured products could be mitigated by additional
imports. Given the fact that farmers in southeastern states had the
highest rate of participation in the Dairy Termination Program, a
reintroduction of this program is likely to strengthen western shifts in
production. As compared to western farmers, eastern farmers tend to have
lower profitability and higher costs of production. Consequently, they are
more likely to participate in this kind of program and less likely to
return to production after the program, because their re-entry costs are
higher. Thus, a new Dairy Termination Program could have disparate
regional effects.

Establish a Production Quota A second option that has been discussed for
supply management is to

System	implement a quota system, as has been done in Canada and in the
European Community.120 Under this option, production would be controlled
by allocating production shares, or quota shares, limiting how much milk
each dairy farmer could market. Such quota shares could be set based on a
farmer's historical marketing level. Any milk marketed over the allocated
quota shares would be priced far below the cost of production. Quota
shares could be traded and increased over time as additional supplies are
needed.

A quota system would help manage supply by taking away incentives to
increase production based on the benefits provided by government

120California also has a quota system; however, it does not function as a
supply management program. In California, milk receipts are pooled, and
those farmers who own quota shares get $1.70 per hundredweight more than
the California butter/nonfat dry milk/cheese price for the number of quota
shares they own. The original allocation of quota shares was made in 1969
based on Class I sales, and because not many additional quota shares have
been allocated, quota shares currently cover about 25 percent of
California's milk production. Most farmers have some quota shares, but
some larger newer operations have come into the system without a quota.
While the California quota shares are traded through brokers, people are
not allowed to purchase them for investment purposes. Purchasers have to
be farmers, and the quota shares come with minimum shipping requirements.
One representative of a major California dairy cooperative said that the
effect of the quota system is purely a way to distribute the proceeds of
California's classified pricing system.

Appendix VII Effects of Recent Federal Dairy Program Changes and
Alternative Policy Options

programs, such as MILC or the price support program. Moreover, compared to
other supply management alternatives, a quota system has a greater
likelihood of achieving long-term supply management, because production
incentives would continue to be limited by the number of quota shares
available in the system. With more effective supply management, federal
costs for other programs such as MILC and the price support program would
be reduced. Additionally, federal costs for administering a quota system
are relatively low. In the short term, price volatility might increase,
but as the market adjusts to a stabilized production level, longterm price
volatility could be reduced.

Nonetheless, there are some drawbacks to implementing a quota system. As
quota shares reduced production, consumer prices could increase. While
demand for fluid milk products is relatively price inelastic, higher
prices could reduce long-term consumption by providing incentives to
purchase substitute goods. Also, the distribution of quotas would provide
a substantial benefit to current farmers to the detriment of farmers who
might try to enter the system in the future, entrenching geographical
production patterns and stifling incentives for technological
enhancements. Given the high production costs in some areas and the
greater efficiency of larger, newer dairy operations, this would represent
an economic inefficiency because milk would not be produced and marketed
as cheaply as possible. However, this drawback could be limited to the
extent that a well-functioning market is established to trade quota
shares. If participation costs in this market were kept low, farmers would
still realize incentives to adopt technology enhancements. The most
efficient dairy farmers' willingness to pay for additional quota shares
would represent their cost advantage over less efficient farmers plus some
assessment of risk. To the extent that this willingness to pay was greater
than the profits realized by less efficient farmers, these less efficient
farmers would have an incentive to sell their quota shares to more
efficient farmers. Thus, in the long term, the quota system might not
hamper increased economic efficiency if trade is relatively easy.

Appendix VIII

Comments from the U.S. Department of Agriculture

Note: GAO comments supplementing those in the report text appear at the
end of this appendix.

See comment 1.

See comment 2.

See comment 3.

See comment 4.

Appendix VIII
Comments from the U.S. Department of
Agriculture

                                 See comment 5.

                         See comment 6. See comment 7.

                                 See comment 8.

                                 See comment 9.

                                See comment 10.

Appendix VIII
Comments from the U.S. Department of
Agriculture

                                See comment 11.

                                See comment 12.

                                See comment 13.

                                See comment 14.

Appendix VIII
Comments from the U.S. Department of
Agriculture

                                See comment 15.

                                See comment 16.

                                See comment 17.

                                See comment 18.

Appendix VIII
Comments from the U.S. Department of
Agriculture

                                See comment 19.

                                 Appendix VIII
                      Comments from the U.S. Department of
                                  Agriculture

The following are GAO's comments on the Department of Agriculture's letter
dated October 22, 2004.

GAO Comments 1.

2.

3.

4.

5.

We revised the report to reflect that farm prices during 2002 and 2003
were the lowest since 1979.

We revised the report to reflect that USDA has estimated that average 2004
farm prices will be more than $3 per hundredweight higher than they were
in 2003.

We agree that there are limitations on the use of commissary data as a
proxy for proprietary wholesale data. However, as noted in USDA's
comments, there seems to be no viable alternative. We revised the report
to further discuss the limitations of using these data.

During the course of our work, USDA declined to provide us with a draft of
its study. Although USDA indicates that it submitted its study to the
Congress on September 10, 2004, we were unable to obtain a copy until
early October 2004, well after we had provided a draft of our report to
USDA for review and comment. Because of this timing, we were unable to
fully consider and analyze the results of USDA's study and related
documents. Furthermore, although USDA notes that it developed quantitative
estimates of the effects on producers (dairy farmers) and consumers and
the cost of various federal programs under various policy scenarios, the
scope of USDA's analysis was more limited than the range of policy options
discussed in our report.

We agree that the potential effects of various policy options in appendix
VII are examined independently and qualitatively within the existing
program structure. Our discussion of dairy policy options are not policy
recommendations. As stated in the report, to identify these policy options
and their potential impacts we relied heavily on a synthesis of the views
of leading dairy experts and the results of an extensive literature
search, including our review of more than 50 studies and other
publications. Time and resource constraints for completing our work
precluded us from developing or contracting for the use of an economic
model that would have provided quantitative estimates of these potential
impacts. In addition, some of the policy options would have been difficult
to model and quantify, such as the potential impacts of accelerating
USDA's hearing and rulemaking process for amending FMMOs. The report also
notes that we compared the policy options

Appendix VIII
Comments from the U.S. Department of
Agriculture

identified against a baseline scenario of policies in place as of August
2004. This baseline scenario existed at the start of our work and was
needed to provide a consistent context for our analysis.

6.	Regarding caveats, as noted in the report, we examined the impact of
federal dairy program changes and policy options on six policy
considerations: farm income, milk production, federal costs, price
volatility, economic efficiency, and consumer prices. We acknowledge in
the report that other stakeholders may have different views on the
importance of these policy considerations or other considerations that we
did not include in our analysis. The report also states that the potential
effects of policy options on these considerations could vary depending
upon economic conditions and other policy decisions. In this regard, we
did not assess the options' overall economic or budgetary impacts, or
their consistency with U.S. international trade commitments or positions
in ongoing negotiations. As indicated in the report, each option has
varying potential impacts on the policy considerations used in our
analysis. Despite these caveats, we believe this analysis is informative
and helpful to congressional decision makers who must weigh competing
interests in determining dairy policy.

7.	We have made some technical corrections and clarifications in light of
USDA's comments, but we do not agree that we mischaracterized the
operation of current programs or the effects that changes to current
programs or the introduction of new programs would have on program
outlays, producers (dairy farmers), and consumers. See also our responses
in comments 8 through 19 below.

8.	Although during the course of our work, USDA officials suggested that
it was possible that dairy farmers might divide their holdings to make
more of their milk eligible for compensation through the MILC program, we
deleted this discussion from the report in light of USDA's comment that it
has no evidence that farmers have done this.

9.	We correctly state in the report that farmers may choose the month that
they begin accepting their payments. However, in response to USDA's
comment we revised the report to clarify that farmers' discretion on when
they receive MILC payments is limited by USDA's regulations for
implementing this program. Specifically, these regulations prohibit a
farmer from selecting a month to receive payments if the month has already
begun, if the month has already passed, or during which no

Appendix VIII
Comments from the U.S. Department of
Agriculture

milk was produced. A farmer also cannot change a previously selected start
month after the 15th of the month before the month selected. Once monthly
payments begin, a farmer has no discretion in determining in which month
or months to receive payments.

10. The discussion of expanding the use of DEIP in the draft report did
not suggest that WTO rules, including quantitative and monetary caps and
product-specific restrictions, be violated. Rather, this discussion
identified ways in which dairy experts suggested that DEIP might be used
more effectively as a marketing tool. However, we have revised the
language in the report to more fully reflect USDA's views and to minimize
confusion as to what we mean by increasing the use of DEIP as a marketing
tool without exceeding WTO caps.

11. We revised the language in the report to clarify that USDA has not
proposed or considered any proposal to eliminate the Dairy Price Support
Program. However, USDA analyzed the potential effects of eliminating this
program in the study it prepared in response to the 2002 Farm Bill
mandate.

12. We agree with USDA that ensuring an adequate level of milk production
is not an objective of the FMMO program. We revised the report accordingly
and added language suggested by USDA to better describe the FMMO program's
objectives.

13. We agree that price volatility contributes to disorderly market
conditions, and we revised the report to better explain the potential
causes of price volatility. We also agree that FMMOs cannot directly
address price volatility in wholesale dairy markets. However, by setting
minimum prices that must be paid to farmers for raw milk, the FMMOs can
affect the extent to which price volatility is reflected in the prices
that farmers receive.

14. We revised the report to reflect that cooperatives owning the capacity
to produce multiple products may still have an incentive to shift milk to
the higher-valued use, in order to provide greater returns to their
members. We also expanded our discussion of other factors that might
influence how milk is used, such as transportation costs and changes in
processing technology.

15. We acknowledge that the report does not explain how a competitive pay
price series could be created for use in the FMMO program.

Appendix VIII
Comments from the U.S. Department of
Agriculture

However, this option was identified by stakeholders during the course of
our work. Other options discussed in the report also may present
challenging implementation issues and in many cases the report discusses
those issues. Nonetheless, we revised the report to reflect that USDA and
a panel of academicians attempted to devise a competitive pay price series
but ultimately were unsuccessful. We also revised the report to note that
a key difficulty in developing a competitive pay price series is the need
for data that are not already influenced by the FMMO classified pricing
system.

16. We acknowledge that the report does not explain how, after combining
Class III and Class IV, milk would be priced in the expanded class.
However, this option was identified by stakeholders during the course of
our work. Other options discussed in the report also may present
challenging implementation issues and in many cases the report discusses
those issues. Nonetheless, we revised the report to reflect that a barrier
to combining Class III and IV is identifying an appropriate pricing
formula that considers the products in an expanded class.

17. We revised the report to reflect this clarification and added language
suggested by USDA to better describe the FMMO program's objectives.

18. We agree with USDA that any area's supply of fluid milk can come from
local or distant farmers. USDA noted that the Class I price surface,
generated from different minimum Class I prices in different locations,
reflects the cost of moving milk from surplus to deficit markets. However,
as we pointed out in a 1988 report, when the price surface does not also
account for regional differences in production costs, it can result in
incentives for overproduction in certain regions. Furthermore, the Class I
price surface that resulted from the 2000 federal order reform differs
from USDA's recommended option.

19. The report accurately reflects the views of some stakeholders that the
slowness of USDA's hearing and rulemaking process used to modify FMMOs
inhibits the agency's ability to respond to changing market conditions or
the marketing of new products. The report also discusses challenges USDA
faces to improving this process while ensuring the promulgation of
economically sound regulation. Further, the report notes that USDA has
made efforts to improve the hearing process, particularly in the way it
evaluates its contracts for hearing transcripts.

Appendix IX

                     GAO Contacts and Staff Acknowledgments

Contacts	Lawrence J. Dyckman, (202) 512-3841 James R. Jones, Jr., (202)
512-9839

Acknowledgments	In addition to the individuals named above, Jay Cherlow,
Barbara El Osta, Joshua Habib, Eileen Harrity, Christopher Murray, and
Cynthia Norris made key contributions to this report. Important
contributions were also made by Beverly Ross, Jena Sinkfield, and Amy
Webbink.

Related GAO Products

Dairy Industry: Estimated Economic Impacts of Dairy Compacts. GAO01-866.
Washington, D.C.: September 14, 2001.

Dairy Industry: Information on Milk Prices and Changing Market Structure.
GAO-01-561. Washington, D.C.: June 15, 2001.

Fluid Milk: Farm and Retail Prices and the Factors That Influence Them.
GAO-01-730T. Washington, D.C.: May 14, 2001.

Dairy Products: Imports, Domestic Production, and Regulation of
Ultrafiltered Milk. GAO-01-326. Washington, D.C.: March 5, 2001.

Dairy Industry: Information on Prices for Fluid Milk and the Factors That
Influence Them. GAO/RCED-99-4. Washington, D.C.: October 8, 1998.

Dairy Industry: Information on Marketing Channels and Prices for Fluid
Milk. GAO/RCED-98-70. Washington, D.C.: March 16, 1998.

Dairy Programs: Effects of the Dairy Termination Program and Support Price
Reductions. GAO/OCE-93-1. Washington, D.C.: June 15, 1993.

Federal Dairy Programs: Insights Into Their Past Provide Perspectives on
Their Future. GAO/RCED-90-88. Washington, D.C.: February 28, 1990.

Milk Pricing: New Method for Setting Farm Milk Prices Needs to Be
Developed. GAO/RCED-90-8. Washington, D.C.: November 3, 1989.

Dairy Termination Program: An Estimate of Its Impact and
Cost-Effectiveness. GAO/RCED-89-96. Washington, D.C.: July 6, 1989.

Milk Marketing Orders: Options for Change. GAO/RCED-88-9. Washington,
D.C.: March 21, 1988.

Overview of the Dairy Surplus Issue-Policy Options for Congressional
Consideration. GAO/RCED-85-132. Washington, D.C.: September 18, 1985.

Effects and Administration of the 1984 Milk Diversion Program.
GAO/RCED-85-126. Washington, D.C.: July 29, 1985.

GAO's Mission	The Government Accountability Office, the audit, evaluation
and investigative arm of Congress, exists to support Congress in meeting
its constitutional responsibilities and to help improve the performance
and accountability of the federal government for the American people. GAO
examines the use of public funds; evaluates federal programs and policies;
and provides analyses, recommendations, and other assistance to help
Congress make informed oversight, policy, and funding decisions. GAO's
commitment to good government is reflected in its core values of
accountability, integrity, and reliability.

Obtaining Copies of The fastest and easiest way to obtain copies of GAO
documents at no cost

is through GAO's Web site (www.gao.gov). Each weekday, GAO postsGAO
Reports and newly released reports, testimony, and correspondence on its
Web site. To Testimony have GAO e-mail you a list of newly posted products
every afternoon, go to

www.gao.gov and select "Subscribe to Updates."

Order by Mail or Phone	The first copy of each printed report is free.
Additional copies are $2 each. A check or money order should be made out
to the Superintendent of Documents. GAO also accepts VISA and Mastercard.
Orders for 100 or more copies mailed to a single address are discounted 25
percent. Orders should be sent to:

U.S. Government Accountability Office 441 G Street NW, Room LM Washington,
D.C. 20548

To order by Phone:	Voice: (202) 512-6000 TDD: (202) 512-2537 Fax: (202)
512-6061

To Report Fraud, Contact:
Waste, and Abuse in Web site: www.gao.gov/fraudnet/fraudnet.htm

E-mail: [email protected] Programs Automated answering system: (800)
424-5454 or (202) 512-7470

Congressional	Gloria Jarmon, Managing Director, [email protected] (202)
512-4400 U.S. Government Accountability Office, 441 G Street NW, Room 7125

Relations Washington, D.C. 20548

Public Affairs	Susan Becker, Acting Manager, [email protected] (202)
512-4800 U.S. Government Accountability Office, 441 G Street NW, Room 7149
Washington, D.C. 20548
*** End of document. ***