Sugar Program: Supporting Sugar Prices Has Increased Users' Costs While
Benefiting Producers (Letter Report, 06/09/2000, GAO/RCED-00-126).

Pursuant to a congressional request, GAO estimated the Department of
Agriculture's (USDA) sugar program's: (1) costs to domestic sweetener
users; (2) benefits to domestic sugar and high-fructose corn syrup
producers; and (3) net effects on the U.S. economy--the differences
between the costs to users and the benefits to producers that result
from artificially high sweetener prices.

GAO noted that: (1) GAO estimates that the sugar program cost domestic
sweetener users about $1.5 billion in 1996 and about $1.9 billion in
1998; (2) sweetener users included: (a) sugarcane refiners that bought
raw cane sugar; (b) food manufacturers that bought refined sugar and
other sweeteners; and (c) final consumers who bought sweeteners and
sweetener-containing products; (3) the program's costs to U.S. sweetener
users depend on the world price of sugar and can vary from year to
year--they will be higher, other things being equal, when the difference
between the domestic and the world price is greater; (4) in 1998, for
example, the program's costs to users were higher than in 1996 because
the world price dropped while the domestic price remained about the
same; (5) the primary beneficiaries of the sugar program's higher prices
are domestic sugar beet and sugarcane producers who, GAO estimates,
received benefits of about $800 million in 1996 and about $1 billion in
1998; (6) about 70 percent of the benefits went to sugar beet growers
and processors; (7) sugarcane producers received about 30 percent of the
benefits; (8) GAO estimates that the sugar program resulted in net
losses to the U.S. economy of about $700 million in 1996 and about $900
million in 1998; (9) GAO net loss estimates include economic
inefficiencies and transfers to foreign producers; (10) economic
inefficiencies occurred, for example, when the sugar program's
artificially high domestic prices encouraged farmers to grow sugar beets
instead of another crop, such as wheat, that, without the sugar program,
might have been relatively more profitable; (11) inefficiencies also
occurred when artificially high sugar prices discouraged consumers from
purchasing sugar; (12) the cost of these inefficiencies totalled about
$300 million in 1996 and about $500 million in 1998; (13) transfers from
the U.S. economy to foreign producers occurred because foreign producers
received artificially high prices for the raw sugar they exported to the
United States; (14) GAO estimates that these transfers amounted to about
$400 million in both 1996 and 1998; and (15) the transfers were about
the same in each year despite the larger difference between domestic and
world prices in 1998 because the United States imported less sugar in
1998.

--------------------------- Indexing Terms -----------------------------

 REPORTNUM:  RCED-00-126
     TITLE:  Sugar Program: Supporting Sugar Prices Has Increased
	     Users' Costs While Benefiting Producers
      DATE:  06/09/2000
   SUBJECT:  Projections
	     Sugar
	     Agricultural production
	     Agricultural industry
	     Importing
	     Prices and pricing
	     Foreign trade agreements
	     Price supports
	     Agricultural programs
	     Economic analysis
IDENTIFIER:  USDA Sugar Program

******************************************************************
** This file contains an ASCII representation of the text of a  **
** GAO Testimony.                                               **
**                                                              **
** 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/RCED-00-126

Appendix I: The U.S. Sugar Program and Recent Price and
Production Trends

12

Appendix II: Estimating Economic Gains and Losses From
the U.S. Sugar Program

20

Appendix III: Structure of the CARD International Sugar Model

46

Appendix IV: Economic Analysis of a Tariff-Rate Quota

50

Appendix V: Estimated Net Losses to the U.S. Economy Accruing
to Foreign Sugar Importers, Fiscal Year 1998

53

Appendix VI: Comments From the U.S. Department of Agriculture

55

Appendix VII: Comments From the American Sugar Alliance

78

Appendix VIII: Comments From the U.S. Cane Sugar Refiners'
Association

102

Appendix IX: GAO Contacts and Staff Acknowledgments

104

Table 1: U.S. and World Prices for Raw and Refined Sugar,
1985-98 14

Table 2: Acres of Sugarcane and Sugar Beets Harvested, by State,
Crop Years 1995-99 16

Table 3: Cane Sugar and Beet Sugar Production and USDA Loans,
by State, Crop Years 1997 and 1999 18

Table 4: U.S. Prices for HFCS, 1985-98 19

Table 5: Estimated Economic Gains and Losses Resulting From
the Sugar Program, 1996 and 1998 21

Table 6: Estimated Distribution Among User Groups of Benefits of Eliminating
the Sugar Program Under Different Pass-Through Assumptions, 1996 and 1998 23

Table 7: Estimated Effect of Eliminating the Sugar Program on
Prices and Production 25

Table 8: Estimated Longer Term Economic Gains and Losses if
the Sugar Program Were Eliminated 26

Figure 1: U.S. Sugar Production (in thousands of short tons),
Fiscal Years 1990-2000 15

Figure 2: Effects of Removing the TRQ and USDA's Loan Program
on U.S. Prices and Quantities of Raw Sugar 30

Figure 3: The TRQ's Effect on Imports 51

ASA American Sugar Alliance

CARD Center for Agricultural and Rural Development at Iowa State University

EV equivalent variation

HFCS high-fructose corn syrup

NAFTA North American Free Trade Agreement

SIC standard industrial classification

TRQ tariff-rate quota

USDA U.S. Department of Agriculture

Resources, Community, and
Economic Development Division

B-285072

June 9, 2000

The Honorable Dianne Feinstein
United States Senate

The Honorable George Miller
The Honorable Dan Miller
House of Representatives

The sugar program, administered by the U.S. Department of Agriculture
(USDA), guarantees domestic cane sugar and beet sugar producers (growers and
processors) a minimum price for sugar, which at times during the past year
was about three times the world market price. The sugar program supports
domestic sugar prices by offering loans to sugar processors at a rate
established by law: 18 cents per pound for raw cane sugar and 22.9 cents per
pound for refined beet sugar, with the sugar serving as collateral for these
loans. (See app. I.) The program has allowed processors to forfeit their
sugar to the federal government instead of repaying their loans--which some
processors might do if domestic sugar prices fall below the level of the
loan rate plus certain costs that processors would no longer incur if they
forfeited. To minimize the likelihood of forfeitures, a direct cost to
taxpayers, the sugar program has maintained artificially high sugar prices
by restricting the amount of sugar that can be imported at a low tariff
rate.

Our April 1993 report,1 requested by then Representative Charles E. Schumer,
estimated that the sugar program cost the domestic users of sugar and other
sweeteners about $1.4 billion (in 1991 dollars) annually from 1989 through
1991 and found that the program primarily benefited U.S. sugar producers and
manufacturers of high-fructose corn syrup (HFCS), a sweetener primarily used
in soft drinks. We recommended that the Congress gradually lower the loan
rate for sugar and direct USDA to adjust import quotas accordingly to
achieve a lower U.S. market price. The 1996 Farm Act did not revise the
sugar program along the lines that we had recommended.

To account for changing conditions in U.S. and world sweetener markets in
recent years, you requested that we update the analysis in our 1993 report.
Specifically, you asked us to estimate the U.S. sugar program's (1) costs to
domestic sweetener users, (2) benefits to domestic sugar and HFCS producers,
and (3) net effects on the U.S. economy--that is, the differences between
the costs to users and the benefits to producers that result from
artificially high sweetener prices.

Because USDA does not have a comprehensive economic model of the U.S. and
world sweetener markets, we developed a model for analyzing the costs,
benefits, and net effects of the U.S. sugar program. Our model used data on
sweetener prices and quantities for the U.S. and world markets and estimated
sugar prices for both markets in the absence of the U.S. sugar program. (See
apps. II and III for descriptions of the model's U.S. and world markets.)
Specifically, we used our model to compare--for both 1996 and 1998--the
domestic and world prices for sugar, HFCS, and other corn sweeteners with
estimated domestic and world prices if there were no sugar program.2 Our
model then used these price comparisons to estimate the program's costs,
benefits, and net effects. As in any modeling effort, there is uncertainty
associated with the estimates that we developed because of uncertainty about
the model's structure, data, and assumptions. These estimates, therefore,
should be interpreted as indicative of the order of magnitude of benefits
and costs, rather than as precise estimates. Despite this uncertainty, we
believe that the process for developing our model has been rigorous and that
our model is both comprehensive and methodologically sound. As agreed with
your offices, we did not analyze how the gradual reduction of U.S. tariffs
for Mexican sugar through 2008 under the North American Free Trade Agreement
might affect domestic sugar prices. However, our model could be used for
this analysis as well as for other types of analyses in the future.

This letter summarizes our findings. More detailed information on our
model's results is presented in appendix II.

We estimate that the sugar program cost domestic sweetener users about $1.5
billion in 1996 and about $1.9 billion in 1998.3 Sweetener users included
(1) sugarcane refiners that bought raw cane sugar, (2) food manufacturers
that bought refined sugar and other sweeteners, and (3) final consumers who
bought sweeteners and sweetener-containing products. The program's costs to
U.S. sweetener users depend on the world price of sugar and can vary from
year to year--they will be higher, other things being equal, when the
difference between the domestic and the world price is greater. In 1998, for
example, the program's costs to users were higher than in 1996 because the
world price dropped while the domestic price remained about the same.

If the sugar program were eliminated and the domestic price of sugar fell,
it would be difficult to predict the extent to which and the speed with
which sugar refiners and manufacturers of sugar-containing foods would pass
their cost reductions through to final consumers. This would depend on,
among other things, the degree of competition in their markets. For example,
the market for table sugar is likely to be more price-competitive than the
markets for sugar-containing foods for several reasons, including that table
sugar varies little from brand to brand while most sugar-containing foods
have greater product differentiation. As a result, sugar refiners that
market table sugar might be more likely than manufacturers of
sugar-containing foods (such as candy makers) to lower their prices, and
this pass-through might occur more quickly. Assuming that, in absence of a
sugar program, competition among sugar refiners had caused them to pass all
of their cost reductions for table sugar through to consumers and the
manufacturers of sugar-containing foods had not passed through any of their
cost reductions, we estimate that final consumers could have benefited
generally by about $800 million using 1998 data and by about $600 million
using 1996 data. If both sugar refiners and the manufacturers of
sugar-containing foods passed through all of their cost reductions, the
maximum annual benefit to final consumers would be about $1.9 billion using
1998 data and about $1.4 billion using 1996 data. Table 6 in appendix II
provides additional information on the benefits accruing to all sweetener
users (refiners, food manufacturers, and final consumers) under these
alternative assumptions about the degree to which cost reductions are passed
through.

The primary beneficiaries of the sugar program's higher prices are domestic
sugar beet and sugarcane producers who, we estimate, received benefits of
about $800 million in 1996 and about $1 billion in 1998.4 About 70 percent
of the benefits went to sugar beet growers and processors. Sugarcane
producers received about 30 percent of the benefits.

HFCS producers received little, if any, benefit from the sugar program in
either 1996 or 1998, according to our current model's estimates. This result
contrasts with our finding in 1993. At that time, HFCS cost a few cents per
pound less than domestic sugar, and both products cost about twice as much
as sugar on the world market. Thus, if the domestic price of sugar had
fallen in 1993, HFCS producers would have had to reduce their prices to
remain competitive. Since 1993, the price of HFCS has fallen, and today it
is much lower than the wholesale price of sugar in the United States.
Furthermore, the possibilities for substitution between sugar and HFCS are
more limited than in prior years because technological advances have
improved HFCS products and created more specialized sweetener markets. As a
result, even if the sugar program were removed and the price of domestic
sugar fell substantially, the impact on the price of HFCS would be
limited--HFCS producers might no longer need to lower their prices to remain
competitive. Executives from the Corn Refiners' Association, which
represents HFCS manufacturers, agreed with our model's results as they
pertained to HFCS, stating that HFCS producers do not benefit from the sugar
program because domestic HFCS prices are no longer linked to sugar prices.

We estimate that the sugar program resulted in net losses to the U.S.
economy of about $700 million in 1996 and about $900 million in 1998. Our
net loss estimates include economic inefficiencies and transfers to foreign
producers. Economic inefficiencies occurred, for example, when the sugar
program's artificially high domestic prices encouraged farmers to grow sugar
beets instead of another crop, such as wheat, that, without the sugar
program, might have been relatively more profitable. Inefficiencies also
occurred when artificially high sugar prices discouraged consumers from
purchasing sugar. The cost of these inefficiencies totaled about $300
million in 1996 and about $500 million in 1998. Transfers from the U.S.
economy to foreign producers occurred because foreign producers received
artificially high prices for the raw sugar they exported to the United
States. (See app. V.) We estimate that these transfers amounted to about
$400 million in both 1996 and 1998. The transfers were about the same in
each year despite the larger difference between domestic and world prices in
1998 because the United States imported less sugar in 1998.

We provided the U.S. Department of Agriculture; the American Sugar Alliance,
which represents sugarcane and sugar beet growers; and the U.S. Cane Sugar
Refiners' Association with a draft of this report for review and comment.

The U.S. Department of Agriculture stated that our model's results were
suspect and should not be quoted authoritatively, citing three broad areas
of concerns. First, the Department stated that the report's methodology was
not adequately developed or justified. We disagree. We took several actions
to help ensure that our model was methodologically sound. Initially, we
contracted with a well-known expert in modeling the international trade of
agricultural commodities and with a prominent agricultural economist to work
with us in developing the model. Then, in December 1999, we sent our
proposed model to four outside academicians who specialize in agricultural
economics and international trade economics and revised the model in
response to their comments. We also sent the proposed model to the
Department at that time, but the Department did not provide comments.
Finally, our process for obtaining agency comments on draft reports served
as a last check of our methodology. In response to these comments, we
adjusted our model to more fully account for certain transportation costs in
our final estimates. Second, the Department stated that documentation of the
economic model was inadequate. We disagree. We included two appendixes in
the draft report that provided detailed information about the model. While
we believe that our model was adequately documented, the Department
suggested some useful clarifications that we incorporated. Third, the
Department stated that in numerous places the model's results are
inconsistent with our description of the model or alternative data sources.
For example, the Department disagrees with our model's finding that HFCS
producers would experience few economic losses if the sugar program were
eliminated. We continue to believe that this finding is accurate. Recent
price data show that while HFCS producers benefited from the sugar program
in the 1980s and early 1990s, the domestic HFCS and sugar markets have since
been decoupled because of domestic price, cost, and market conditions. After
thoroughly examining this and each of the Department's other concerns about
our model's results and discussing them with experts outside GAO, we
continue to believe that the model's results are reasonable. See appendix VI
for the Department's written comments and our responses.

The American Sugar Alliance also disagreed with the draft report's analysis.
Its concerns also fell into three general areas. First, the Alliance stated
that our methodology for estimating the sugar program's cost to users was
flawed because our model used world prices of raw sugar that did not reflect
each country's cost of production--the Alliance asserted that our model's
world prices were distorted by, among other things, subsidies provided by
other countries and differences in labor and environmental standards. We
believe that our methodology is appropriate. Our model estimated the impact
of the U.S. sugar program in the world as it exists today--not in a
hypothetical "free market" environment, as the Alliance suggested. To
isolate the sugar program's effects, we made the analytical assumption that
all factors, other than the U.S. sugar program, would remain unchanged.
Second, the Alliance stated that our model incorrectly assumed that food
manufacturers and retailers would pass through 100 percent of their savings
to final consumers if the sugar program were eliminated. In fact, the draft
report noted that the extent to which cost reductions would be passed
through to consumers would depend on the degree of competition in their
markets. We revised the report, however, to further clarify the different
assumptions we used in discussing the extent to which final consumers could
benefit if the sugar program were eliminated. Third, the Alliance stated
that we did not sufficiently assess the benefits and costs of the sugar
program, noting among other things that rural areas would suffer if the
sugar program were eliminated. As stated in our study objectives, this
report focuses on the U.S. sugar program's (1) costs to domestic sweetener
users, (2) benefits to domestic sugar and HFCS producers, and (3) net
effects on the U.S. economy. The American Sugar Alliance also provided
comments to clarify what it perceived to be misleading statements and
analytical errors in the draft report. We address each of these comments in
appendix VII, which contains the American Sugar Alliance's complete written
comments and our responses.

The U.S. Cane Sugar Refiners' Association stated that we had done a thorough
job of illuminating the very significant impact of the sugar program on
different interests. The Association also provided comments to improve the
report's technical accuracy, which we incorporated as appropriate. See
appendix VIII for the written comments of the U.S. Cane Sugar Refiners'
Association.

Estimating the total cost of the sugar program to users is controversial
because the total cost is not a simple difference between the current U.S.
and world sugar prices. Instead, the cost estimate depends in part on
assumptions about how much the world price would rise if the United States
did not have a sugar program. In developing our model, we:

ï¿½ Contracted with Professor John C. Beghin, who heads the Trade and
Agricultural Policy Division at Iowa State University's Center for
Agricultural and Rural Development, to work with our staff economists to
integrate an economic model of the U.S. sweetener market with the Center's
international trade model for agricultural commodities. Professor Beghin is
a recognized expert in agricultural trade policy and has extensive
experience in developing econometric models of the U.S. and world markets
for agricultural commodities. We also contracted with Professor Bruce
Gardner, of the University of Maryland and a former USDA Assistant Secretary
for Economics, to consult with Professor Beghin and our economists in
developing our model, assessing the reasonableness of its results, and
responding to comments from outside consultants and organizations.

ï¿½ Provided our proposed model for review and comment in December 1999 to
Professor James Anderson of Boston College, Professor Andrew Schmitz of the
University of Florida, Professor Daniel Sumner of the University of
California at Davis, and Professor Michael Wohlgenant of North Carolina
State University. We incorporated many of their comments as we revised the
proposed model. We also provided our proposed model to USDA for review and
comment in December 1999. However, USDA did not provide comments.

ï¿½ Obtained extensive data on sweetener prices and quantities for the U.S.
and world markets in 1996 and 1998. These included data on (1) production
and prices for sugar, HFCS, and other corn sweeteners and (2) the use of
sweeteners by U.S. industries, disaggregated to the fourth level of the U.S.
Bureau of the Census' Standard Industrial Classification. We obtained these
data from USDA's Economic Research Service; the Bureau of the Census and the
Bureau of Economic Analysis, within the Department of Commerce; the Bureau
of Labor Statistics, within the Department of Labor; and private industry
organizations. We selected 1998 because it is the most current year for
which price and quantity data were available when we developed our model and
1996 because it was a recent year that provided a contrast in the world
price of sugar and the U.S. prices of corn, wheat, and other crops.

ï¿½ Used the model to estimate the effect of terminating the U.S. sugar
program on (1) the U.S. and world prices of sugar and (2) U.S. sweetener
producers and users. The U.S. sweetener part of the model is the U.S.
component of the CARD world sugar model, extended on the supply side to
include linkages with the corn, HFCS, and wheat markets. In addition,
Professor Beghin, in collaboration with our economists, extended the demand
side of the CARD domestic sweetener model to include the demand for
sweeteners from food processors, as well as the direct demand for sweeteners
by final consumers. We interpreted the sugar program's estimated costs in
1996 and 1998 as the opposite of estimated benefits that would be derived by
the program's elimination. While our model primarily examined the short-run
effects of eliminating the sugar program, we also assessed longer-term
effects by using alternative elasticities for the supply of sweeteners.5

ï¿½ Obtained the comments of Professor Gardner and Professor Wohlgenant on the
final design of our model and its preliminary results before we provided a
draft of our report to USDA for comment.

We conducted our work between August 1999 and May 2000 in accordance with
generally accepted government auditing standards. We did not independently
verify the data that we obtained from USDA, Commerce, Labor, or industry
sources and used in our economic model. However, these are the best data
available on U.S. and world prices and production. The estimates in this
report apply only to the 2 years that we analyzed. Estimates for other years
might be larger or smaller than our 1996 and 1998 estimates, depending on
the difference between the domestic and world prices of sugar.

We are sending copies of this report to the Senate Committee on Agriculture,
Nutrition, and Forestry; the House Committee on Agriculture; and other
appropriate congressional committees; the Honorable Dan Glickman, Secretary
of Agriculture; the Honorable Jacob Lew, Director, Office of Management and
Budget; and other interested parties. We will also make copies available to
others upon request.

Please contact me at (202) 512-5138 if you or your staff have any questions
about this report. Key contributors to this report are listed in appendix
IX.

Robert E. Robertson
Associate Director, Food
and Agriculture Issues

The U.S. Sugar Program and Recent Price and Production Trends

The United States was the fifth largest producer and the fourth largest
consumer of sugar in the world in 1998. Historically, the United States and
other countries have protected their domestic growers and processors of cane
sugar and beet sugar6 from lower world prices through quotas and/or high
tariffs that restrict sugar imports.

The U.S. sugar program guarantees domestic sugar producers (growers and
processors) a minimum price for sugar by (1) offering loans to sugar
processors at a rate established by law and (2) using a tariff-rate quota
(TRQ) to restrict the supply of sugar that can be imported at a low tariff
rate. The Agriculture and Food Act of 1981 allowed sugar processors to
obtain loans from the U.S. Department of Agriculture's (USDA) Commodity
Credit Corporation by pledging their sugar as collateral.7 The act also gave
processors the option to forfeit the sugar that secures their loans to the
federal government rather than repay their loans in cash, effectively
establishing the loan rate as a floor for domestic sugar prices. Since 1990,
USDA has used a TRQ for raw sugar to restrict low-priced imports and
maintain domestic prices at levels that are high enough to prevent producers
from forfeiting on their sugar loans.8 Under the TRQ, imported sugar up to
the quota is either assessed no tariff or a 0.63-cent-per-pound tariff,
while imports above the quota are assessed a 15.82-cents-per-pound tariff,
making them prohibitively expensive.9 In 1994, the United States agreed to
set the TRQ for imported sugar at 1.26 million tons or more each year and to
administer the TRQ in a manner consistent with its commitments under the
World Trade Organization Agreement on Agriculture. USDA normally sets the
size of the TRQ at the beginning of a fiscal year. The U.S. Trade
Representative then allocates shares of it among 40 designated countries on
the basis of their exports to the United States from 1975 through 1981.

The Federal Agriculture Improvement and Reform Act of 1996, commonly known
as the 1996 Farm Act, modified the sugar program by (1) legislatively
establishing USDA's loan rate at 18 cents per pound for raw cane sugar and
22.9 cents per pound for refined beet sugar, (2) assessing a 1-cent penalty
on each pound of raw cane sugar and a 1.07-cent penalty on each pound of
refined beet sugar forfeited to the government, (3) eliminating a
requirement that the sugar program operate at no net cost to U.S. taxpayers,
(4) limiting processors' opportunities to forfeit their sugar to the
Commodity Credit Corporation by not allowing such forfeitures if the TRQ is
1.5 million tons or less, (5) eliminating USDA's authority to impose
marketing allotments for sugar, and (6) increasing the assessment on
processors to 0.2475 cents per pound for of raw cane sugar and 0.2654 cents
per pound for beet sugar.10

USDA has established a TRQ greater than 1.5 million tons each year since the
1996 Farm Act was enacted. As a result, processors that have obtained
Commodity Credit Corporation loans have had the option to forfeit their
pledged sugar instead of repaying their loans. This option becomes important
to processors if domestic sugar prices drop below USDA's loan rate plus
transportation and interest costs but minus the 1-cent-per-pound penalty. If
sugar is forfeited, USDA can choose to immediately resell the sugar on the
domestic market, sell it for such restricted uses as ethanol manufacture,
store it, or donate it for humanitarian purposes.

Table 1 shows that, since 1985, the sugar program has resulted in U.S. raw
sugar prices that have been more than 8 cents per pound higher than world
raw sugar prices.11 In 1998, the U.S. raw sugar price was more than 10 cents
per pound higher than the world price.

 Prices in cents per pound
                   U.S. wholesale                              World
 Year U.S. raw     refined beet      U.S. retail   World raw   refined
      cane sugara                    refined sugar sugarb
                   sugar                                       sugarc
 1985 20.34        23.18             35.34         4.04        6.79
 1986 20.95        23.38             35.08         6.05        8.47
 1987 21.83        23.60             35.28         6.71        8.75
 1988 22.12        25.44             36.60         10.17       12.01
 1989 22.81        29.06             40.03         12.79       17.16
 1990 23.26        29.97             42.78         12.55       17.32
 1991 21.57        25.65             42.80         9.04        13.41
 1992 21.31        25.44             41.53         9.09        12.39
 1993 21.62        25.15             40.54         10.03       12.79
 1994 22.04        25.15             39.99         12.13       15.66
 1995 22.96        25.83             39.83         13.44       17.99
 1996 22.40        29.20             41.79         12.24       16.64
 1997 21.96        27.09             43.26         12.06       14.33
 1998 22.06        26.12             42.98         9.68        11.59

Note: U.S. and world prices are in nominal dollars.

a U.S. prices are based on futures contract prices for number 14 raw cane
sugar on the New York Coffee, Sugar, and Cocoa Exchange.

b World prices are based on bulk spot contracts for number 11 raw cane sugar
on the New York Coffee, Sugar, and Cocoa Exchange (free on board stowed
Caribbean port, including Brazil). To compare the world and U.S. prices, 1.5
cents per pound needs to be added to the world price to account for the cost
of transporting raw sugar from the Caribbean to New York.

c World prices are based on spot contracts for number 5 refined sugar,
London Daily Price (free on board Europe).

Source: Sugar and Sweetener Situation and Outlook, Economic Research
Service, USDA (Sept. 1999), and the New York Coffee, Sugar, and Cocoa
Exchange.

Figure 1 shows that domestic sugar production has grown by about 25 percent
in recent years--from 7.2 million tons in fiscal year 1997 to a projected 9
million tons in fiscal year 2000.12

1990-2000

This growth reflects an 8-percent increase in the estimated acres of
sugarcane and sugar beets harvested in 2000 compared with the actual acres
harvested in 1997. (See table 2.) Many farmers have increased the size of
their sugarcane and sugar beet crops because these crops have offered better
returns than cotton, wheat, or other crops that the farmers grew in the
past. In addition, sugarcane and sugar beet farmers have increased their
crop yields per acre, and sugarcane farmers, in particular, have improved
their sugar-per-acre recovery rates.

 State        1995       1996       1997       1998      1999
 Sugarcane
 Florida      417,000    417,000    421,000    426,000   436,000
 Hawaii       48,500     42,900     32,000     30,300    32,600
 Louisiana    368,000    335,000    380,000    400,000   441,800
 Texas        41,200     34,600     27,300     32,000    30,700
 Subtotal     874,700    829,500    860,300    888,300   941,100
 Sugar beets
 California   114,000    82,000     99,000     100,000   105,000
 Colorado     41,100     51,100     66,400     57,300    67,900
 Idaho        197,000    184,000    197,000    203,000   210,000
 Michigan     188,000    130,000    160,000    173,000   187,000
 Minnesota    420,000    438,000    446,000    458,000   469,000
 Montana      55,500     57,500     58,300     62,400    61,800
 Nebraska     72,300     51,200     60,300     47,400    66,800
 New Mexico   a          900        1,600      0         0
 North Dakota 204,200    225,300    227,500    242,600   253,000
 Ohio         15,300     4,600      900        1,100     1,200
 Oregon       17,800     16,300     17,400     17,500    19,700
 Texas        19,300     12,600     15,000     0         0
 Washington   a          13,000     18,000     36,000    27,000
 Wyoming      61,500     56,800     60,900     53,300    57,000
 Subtotal     1,406,000  1,323,300  1,428,300  1,451,600 1,525,400
 Total        2,280,700  2,152,800  2,288,600  2,339,900 2,466,500

Note: USDA's World Agricultural Outlook Board estimates that 939,000 acres
of sugarcane and 1,527,000 acres of sugar beets will be harvested in crop
year 2000.

a In 1995, the acreage harvested in New Mexico and Washington was included
in 14,100 acres listed as other states.

Source: Sugar and Sweetener Situation and Outlook, Economic Research
Service, USDA (Sept. 1999).

In November 1999, USDA announced that the TRQ for raw sugar imports in
fiscal year 2000 would be slightly higher than 1.5 million tons, giving
sugar processors that obtained USDA loans the option to forfeit their sugar
to the government if domestic prices drop below the loan rate plus certain
costs that processors would no longer incur if they forfeit.13 In February
2000, U.S. raw sugar prices on the New York Coffee, Sugar, and Cocoa
Exchange dropped to 16.5 cents per pound as a result of the increasing
supply of sugar. As of May 1, 2000, raw sugar prices had recovered to about
19.7 cents per pound14--around the minimum price necessary to provide an
economic incentive to U.S. sugar processors to repay their USDA loans rather
than forfeit the sugar that secures their loans to the government.15

In May 2000, USDA announced that it would seek to purchase 150,000 tons of
domestic sugar to reduce the cost of expected sugar program loan forfeitures
later this fiscal year. While data on the total amount of USDA loans to
sugar processors in 2000 were not available, table 3 shows that sugarcane
processors substantially increased their use of USDA loans in 1999, as
compared with 1997. U.S. sugar processors have not forfeited sugar to the
government since 1994, when small amounts were forfeited.

                       Sugar production in short tons (raw value)

                                USDA loans made for                USDA loans made for
                                       1997                               1999
                   Estimated  Production              Estimated  Production
   State   Farmsa     1997     covered   Dollar value    1999      covered  Dollar value
                  production  by loans               production   by loans
 Cane sugar
 Florida   152    1,925,000   56,000     $20,000,000 2,125,000   470,000    $167,798,000
 Hawaii    13     363,000     0          0           380,000     0          0
 Louisiana 705    1,262,000   126,000    46,171,000  1,615,000   354,000    129,975,000
 Texas     103    80,000      18,000     5,851,000   100,000     24,000     8,659,000
 Subtotal  973    3,630,000   200,000    $72,022,000 4,220,000   848,000    $306,432,000
 Beet sugar
 California449    436,000     0          $0          450,000     0          $0
 Colorado  530    192,000     0          0           203,000     0          0
 Idahob    921    765,000     422,000    189,506,000 734,000     320,000    142,302,000
 Michigan  1,182  446,000     175,000    83,245,000  465,000     146,000    69,237,000
 Minnesota 1,622  1,212,000   0          0           1,361,000   180,000    82,008,000
 Montana   415    180,000     0          0           201,000     0          0
 Nebraska  367    149,000     0          0           183,000     0          0
 New Mexico5      7,000       0          0           0           0          0
 North
 Dakota    873    618,000     0          0           752,000     18,000     8,201,000
 Ohio      33     2,000       0          0           3,000       0          0
 Oregon    167    73,000      0          0           68,000      0          0
 Texas     119    40,000      0          0           0           0          0
 Washington58     87,000      0          0           119,000     28,000     12,499,000
 Wyoming   356    182,000     0          0           162,000     0          0
 Subtotal  7,097c 4,389,000   597,000    $272,751,0004,701,000   692,000    $314,247,000
 Total     8,070  8,019,000   797,000    $344,773,0008,921,000   1,540,000  $620,679,000

Note: USDA loans are in nominal dollars.

a USDA's 1997 census of agriculture estimate of the farms that grew
sugarcane and sugar beets. In contrast, USDA's Farm Service Agency estimated
that about 11,800 farms grew sugar beets nationwide in 1997. The census of
agriculture uses a more restrictive definition of a farm that, for example,
counts as a single farm land that has been subdivided into several smaller
units that, in some cases, are operated by different family members.

b USDA loans were with a sugar beet processor that has its corporate
headquarters in Utah and processing facilities mainly in Idaho.

c Subtotal excludes two farms in Illinois and three farms in Kansas that
USDA's 1997 census of agriculture identified because sugar beet production
data for these states were not available.

Source: USDA's 1997 census of agriculture, Economic Research Service, and
Farm Service Agency.

High-fructose corn syrup (HFCS) competes with sugar for leadership in the
U.S. sweetener market. Each year from 1985 through 1994, the weighted
average HFCS price was consistently at least 3 cents per pound lower than
the wholesale refined beet sugar price. Since 1992, the weighted average
HFCS price dropped from 22 cents per pound to 12.4 cents per pound in
1998--13.7 cents per pound less than the wholesale refined beet sugar price.
(See table 4; prices are in nominal dollars.) Because of this cost
advantage, the domestic supply of HFCS grew by almost 40 percent, from 6.8
million tons (dry weight) in 1992 to 9.5 million tons (dry weight) in 1999.
HFCS now accounts for more than half of the total U.S. sweetener output,
with approximately 75 percent of the entire HFCS supply going to the
beverage industry. However, HFCS represented only about 6 percent of total
U.S. corn usage in 1998.

 Prices in cents per pound (dry weight)
                          Weighted average
 Year                                       U.S. wholesale refined beet
      HFCS-42a  HFCS-55b  HFCS price        sugar price
 1985 17.75     19.95     19.18             23.18
 1986 18.07     19.96     19.30             23.38
 1987 16.50     17.46     17.11             23.60
 1988 16.47     18.68     17.80             25.44
 1989 19.24     21.41     20.54             29.06
 1990 19.69     21.88     20.99             29.97
 1991 20.93     23.32     22.34             25.65
 1992 20.70     23.00     22.03             25.44
 1993 18.83     20.95     20.08             25.15
 1994 18.77     22.51     21.01             25.15
 1995 15.63     19.00     17.67             25.83
 1996 14.46     20.60     18.28             29.20
 1997 10.70     13.98     12.78             27.09
 1998 10.58     13.43     12.40             26.12

Note: U.S. prices for HFCS and wholesale refined beet sugar are in nominal
dollars.

a HFCS-42 is used in canned fruits, condiments, and other processed foods
that need mild sweetness.

b HFCS-55 is used to make soft drinks, ice cream, and frozen desserts.

Source: GAO's analysis of data from USDA and the sweetener industry.

Estimating Economic Gains and Losses From the U.S. Sugar Program

The sugar program has used farm commodity and trade policy instruments to
maintain domestic sugar prices at levels that exceed world prices without
requiring the government to buy large quantities of domestic sugar. Our
economic model analyzed the effects of eliminating the sugar program on
prices and production by linking a multimarket domestic sweetener model to a
multicountry world sweetener model. We estimated the economic welfare16
effects of the program--the gains and losses to the most heavily affected
producer and consumer groups--by interpreting the estimated welfare loss
(gain) resulting from the elimination of the sugar program as an estimate of
the gain (loss) accruing to each group from the presence of the program.17
Our analysis included the markets for sugar beet and sugarcane production,
corn and HFCS production, sugar refining, food processing, and the final
consumption of sugar and food products containing sweeteners. In addition,
we estimated the net loss to the U.S. economy (economic welfare gains minus
losses) resulting from artificially high sweetener prices. This net loss
includes economic inefficiencies (known as deadweight losses) and economic
rent transfers to foreign sugar exporters.

This appendix provides (1) our model's estimates of the sugar program's
costs and benefits; (2) an overview of our modeling process; (3) a
description of the policy simulations used in our analysis; (4) a more
detailed discussion of the theoretical economic framework of our U.S. sugar
model, including the methods used to estimate welfare gains and losses for
participants in the various affected markets; and (5) a description of the
data and data sources used.

Our model estimated the costs and benefits of the sugar program by
comparing--for both 1996 and 1998--the actual domestic and world prices for
sugar, HFCS, and other sweeteners with the estimated domestic and world
prices if the sugar program were eliminated. Both the estimated costs of the
sugar program to sweetener users and the estimated benefits to sugar beet
and sugarcane producers were higher in 1998 when the difference between the
domestic and world prices for sugar was greater.

As shown in table 5, we estimated that the sugar program cost domestic
sweetener users--sugarcane refiners, food manufacturers, and final
consumers--about $1.5 billion in 1996 and about $1.9 billion in 1998. In our
analysis, the distribution of these welfare losses resulting from the sugar
program among the sweetener user groups depends on assumptions about the
extent to which refiners' and manufacturers' cost reductions from
eliminating the sugar program would be passed on to consumers.

 1999 dollars in millions
 Category                                   1996      1998
 Welfare gains accruing to producers        $788      $1,045
 Sugarcane producers                        241       307
 Sugar beet growers                         490       650
 Sugar beet processors                      58        89
 HFCS manufacturers and corn growers        (1)       (1)
 Welfare losses accruing to sweetener users ($1,471)  ($1,938)
 Net loss to the U.S. economy               ($683)    ($893)
 Economic inefficiencies                    (273)     (532)
 Transfers to foreign suppliers             (410)     (361)

Note: Numbers in parentheses are economic losses.

We estimated that the total welfare gains by domestic sugar beet and
sugarcane producers were about $800 million in 1996 and about $1 billion in
1998. About 70 percent of these benefits went to sugar beet growers and
processors, while the remaining 30 percent went to sugarcane producers.

We estimated that HFCS producers did not receive welfare gains from the
sugar program in either 1996 or 1998 primarily because (1) HFCS prices have
been much lower than the wholesale price of sugar in the United States since
1995 and (2) the possibilities for substitution between sugar and HFCS are
more limited than in prior years because technological advances have
improved the HFCS product and created more specialized sweetener markets.18
Thus, HFCS producers would not need to lower their price further to remain
competitive if the sugar program were eliminated. In contrast, our 1993
model of the sweetener markets in the late 1980s and early 1990s--when the
price of HFCS was considerably higher--found that a fall in the domestic
price of sugar would have put pressure on HFCS producers to lower the price
of HFCS to remain competitive. As a result, our 1993 model found that HFCS
producers received substantial welfare gains from the sugar program.
Executives from the Corn Refiners' Association, which represents HFCS
manufacturers, agreed with our results as they pertained to HFCS. They told
us that the domestic HFCS market was "decoupled" from the domestic sugar
market--HFCS prices are no longer linked to sugar prices, and the soft drink
industry has relied on competition among HFCS manufacturers to minimize its
sweetener costs.

We estimated that the sugar program resulted in net losses to the U.S.
economy of about $700 million in 1996 and about $900 billion in 1998 because
total welfare losses exceeded gains. These net losses included (1)
production and consumption inefficiencies of $300 million in 1996 and $500
million in 1998 and (2) transfers of $400 million in 1996 and in 1998 to
foreign countries allocated a portion of the TRQ for sugar imports to the
United States.

Eliminated

If the sugar program were eliminated, it would be difficult to predict the
extent to which or the speed with which intermediate users of sweeteners
would pass through lower sugar costs to final consumers. Table 6 presents
two estimates of how the benefits of eliminating the sugar program might be
distributed based on two different sets of pass-through assumptions.

                          1999 dollars in millions
                               1996                          1998

 Distribution of      Partial          Full            Partial       Full
    benefits       pass-through    pass-through     pass-through     pass-
                                                                   through
 Final consumers  $587            $1,434           $769            $1,960
 Food
 manufacturers    715             (60)             999             (85)
 Sugarcane
 refiners         95              97               61              63
 Total            $1,397          $1,471           $1,829          $1,938

Note: The partial pass-through results represent a full pass-through of cost
reductions by sugar refiners and no pass-through by food processors to
consumers with the elimination of the sugar program. The full pass-through
results assume all cost reductions are passed through to final consumers.
Numbers in parentheses are economic losses.

The first set of estimates is based on the assumption that (1) competition
would lead sugar refiners to pass cost reductions associated with
eliminating the sugar program through to final consumers in the form of
lower prices for table sugar but (2) manufacturers of sugar-containing foods
would retain their cost savings. Under this "partial pass-through"
assumption, final consumers would have gained about $600 million using 1996
data and about $800 million using 1998 data if the sugar program had been
eliminated.19 We chose to present estimates based on this assumption because
refined white sugar is more homogeneous than sweetener-containing food
goods. With a homogeneous product such as sugar, each brand is almost
perfectly substitutable for another. When substitutability between products
is nearly perfect, it is more difficult for sellers to insulate their
products from the price competition of rivals. In contrast, when products
are highly differentiated, as many sweetener-containing food products are,
firms may be less able to attract customers from competitors by offering
lower prices, so there is less incentive to compete by lowering prices.
Instead, firms may use other nonprice forms of competition, such as greater
advertising. In addition, in sugar refining, the cost of raw sugar is a much
larger share of the total cost of production compared with its share in the
production of other food products. Therefore, a change in the cost of raw
sugar would be likely to have a larger effect on the price of table sugar
than on the prices of sugar-containing products.

Our second set of estimates is based on the assumption that competition
among both food manufacturers and sugar refiners is such that all cost
reductions that would result from eliminating the sugar program would be
passed on to final consumers. This "full pass-through" assumption yields an
upper bound to the potential savings to consumers. Under this assumption, we
estimate that the benefits to final consumers of eliminating the sugar
program would have been about $1.5 billion in 1996 and about $1.9 billion in
1998.

Table 7 compares actual sugar prices and production in 1996 and 1998 with
our model's estimates, which assume the termination of the sugar program. In
particular, our model shows that if the sugar program had been eliminated,
the domestic price of raw sugar would have dropped from about 22 cents per
pound to about 14.9 cents per pound in 1996 and to about 12.5 cents per
pound in 1998, with comparable declines in the wholesale price of domestic
refined sugar. We estimated that raw sugar imports would have increased by
about 1.1 million tons in 1996 and by about 1.6 million tons in 1998 if the
sugar program had been eliminated, reflecting both the increased domestic
demand for sugar and the decreased domestic production of sugar beets and
sugarcane.

  Price in cents per pound and production and imports in millions of short
                              tons (raw value)
                                 1996                        1998

                         Actual     Without the      Actual    Without the
                                   sugar program              sugar program
 U.S. raw sugar price   22.40     14.91            22.06      12.46
 World raw sugar pricea 12.24     13.41            9.68       10.96
 U.S. wholesale refined
 sugar price            29.20     21.77            26.12      16.12
 World wholesale
 refined sugar priceb   16.64     19.77            11.59      14.12
 Sugarcane
 Acres harvestedc       953,700   941,300          931,500    916,200
 Production             29.1      28.7             30.0       29.5
 Sugar beets
 Acres harvestedc       1,420,100 1,350,300        1,428,300  1,338,600
 Production             28.1      26.7             29.9       28.0
 Raw sugar imports      2.2       3.3              1.7        3.3

Note: Our model's estimates of prices without the sugar program are
expressed in 1999 dollars, while the acutal prices are expressed in nominal
dollars.

a The world price for raw sugar is based on a Caribbean location. As
compared with the U.S. price, the world price for raw sugar does not include
1.5 cents per pound in cost to transport the sugar to New York.

b As compared with the U.S. price, the world price for refined sugar does
not include 2 cents per pound for transportation.

cAcreage harvested during the previous crop year.

Table 8 presents our estimates of the welfare changes that would have
resulted from eliminating the sugar program in 1998, using larger supply
elasticities than the ones we used to obtain our primary estimates to
simulate shorter-term changes. Our supply elasticity estimates are arc
elasticities evaluated for 1998 between historical and post-reform values.
In particular, our short-run domestic supply elasticities were 0.05 for
sugarcane and 0.10 for sugar beets, and our short-run import supply
elasticity was 7.26. To obtain longer-term welfare estimates, we used a
double Nerlovian domestic supply response with supply elasticities of 0.20
for cane and 0.26 for sugar beets and an import supply elasticity of 10.17.
These larger elasticities can be interpreted as longer-term elasticities.
Therefore, the resulting estimates from our simulation can be interpreted as
the welfare gains and losses after more time has passed for the economy to
adjust to the lower sugar prices that would result from eliminating the
sugar program.

 Dollars in millions
 Category                                  Gain or (loss)
 Welfare losses accruing to producers      ($1,017)
 Sugarcane producers                       (301)
 Sugar beet growers                        (530)
 Sugar beet processors                     (187)
 HFCS manufacturers and corn growers       1
 Welfare gains accruing to sweetener users $1,947
 Final consumer                            1,953
 Food manufacturers                        (84)
 Sugarcane refiners                        78
 Net gain to the U.S. economy              $930
 Economic inefficiencies                   572
 Transfers to foreign suppliers            358

Note: These results assume a double Nerlovian supply response and a full
pass-through of program costs to final consumers. Numbers in parentheses are
economic losses.

Estimating the welfare gains and losses from the U.S. sugar program requires
several steps. First, we simulate the elimination of the program to
determine price and production responses in both domestic and international
sugar markets. This simulation involves specifying complete U.S. and world
sweetener models in the presence of the U.S. sugar TRQ and commodity loan
program. To do this, we used Iowa State University's Center for Agricultural
and Rural Development (CARD) world sweetener model (see app. III), which
contains the U.S. domestic sweetener economy as one of its component
countries. However, we extended and modified the U.S. domestic sweetener
model to include a more detailed, multimarket approach, including such
markets as corn and feed, sugar, and HFCS. In the U.S. domestic model, we
simulated the sugar program's elimination by removing the TRQ for raw and
refined sugar and allowing more domestic demand to be satisfied by
lower-priced world imports. Simultaneously, as the U.S. demand for sugar
increased, the world sugar prices rose somewhat. We also removed USDA's loan
program for sugar processors and allowed the domestic market prices of sugar
to fall below the loan rate levels. After these program changes, U.S.
domestic raw sugar prices approximated the world prices.

On the supply side of the domestic market, we used the domestic component of
the CARD sugar model to estimate the welfare changes due to the change in
the price of sugar. The new U.S. raw sugar price filtered through the
domestic U.S. sugarcane, sugar beet, corn, and HFCS markets, leading to new
production quantities. For each of these producing industries, we measured
the changes in realized quasi-profits, or producer surplus, that would
result from a change in the quantity demanded and/or the price if the sugar
program were eliminated.

Within the domestic sweetener model, we estimated welfare changes for a
comprehensive demand sector, including sugar processors and refiners,
sweetener-using industries, and the final consumer. We estimated the changes
in realized quasi-profits resulting from higher sweetener prices for
sweetener-using food industries, at the 4-digit Standard Industrial
Classification (SIC)20 level, using economic methods to derive industry cost
and demand functions for sugar and HFCS. As part of this analysis, we
integrated different assumptions about the market power of these industries.

Furthermore, we developed two estimates of the welfare effects on consumers
by using different assumptions about the extent to which the benefits of
eliminating the sugar program would be passed along to the final consumers
of sugar and sweetener-containing products. Consumers are directly affected
by the sugar program's elimination through changes in the prices of both
refined sugar and food items containing a significant amount of sweetener
that they purchase. We applied an incomplete demand system approach called
LINQUAD21 and used an exact welfare measure, equivalent variation (EV),22 to
estimate changes in consumers' expenditures for these items.

Finally, we aggregated all welfare gains and losses from these groups by
treating the welfare loss (gain) experienced from eliminating the sugar
program as an estimate of the gain (loss) accruing to each group from the
presence of the program. The difference between welfare gains and losses
accruing from the program is the net loss to the U.S. economy, which
consists of (1) transfers to foreign producers that resulted from
artificially high prices for the raw sugar they exported to the United
States and
(2) economic inefficiencies, known as deadweight losses. These
inefficiencies resulted from the use of higher-cost domestic resources to
produce sweeteners (instead of importing lower-priced sugar) and reduced
total sugar consumption. These losses did not redistribute income from
consumers to producers.

Using this approach, we specified a subset of closely related agricultural
markets--sugar, HFCS, and corn--that are important in estimating the welfare
gains and losses from the sugar program. We also included the influence of
wheat prices on planted sugar beet acreage. One possible limitation of our
model is that a more general equilibrium approach of modeling the entire
agricultural sector may have been able to give us more long-term effects by,
for example, identifying what alternative crops would be produced in the
absence of the program or how many producers would leave the industry
entirely. However, general equilibrium models take a more broad-brushed
approach, often leaving out important market details. Our approach is
designed to represent the most important sweetener market relationships with
the available data, while keeping the model tractable.

The policy scenario for this analysis removes the TRQs for imported raw and
refined sugar and USDA's loan program for sugar processors that supports the
price of domestic sugar (see app. IV for a discussion of the economics of
the TRQ).23 Figure 2 shows the effects of removing both the raw sugar TRQ
and USDA's loan program: the first panel (a) represents the domestic raw
sugar market, while the second panel (b) represents the world raw sugar
market.

In panel (b), we show two world excess supply situations, ES1 and ES2,
corresponding to different trade scenarios. As discussed in appendix IV, in
the presence of a TRQ, the United States faces a kinked world excess supply
function, as in the bold line, ES1. The vertical line segment, BC, on ES1
represents the level of the TRQ, below and beyond which there is a supply
response to price by foreign exporters. Moreover, at import levels below the
quota, QTRQ, the in-quota tariff applies, and beyond that level, the
out-of-quota tariff applies. The other excess supply curve, ES2, corresponds
to the world excess supply in the absence of import restrictions in the
United States.

The effect of the TRQ on U.S. imports and prices depends on the location of
the U.S. excess demand for imports relative to the excess supply. In panel
(b), we display three potential U.S. import demand situations, ED1, ED2, and
ED3. The excess demand curve, ED1, represents an import demand schedule that
intersects the excess supply schedule below the level of the TRQ, while the
excess demand curve, ED3, represents an import demand schedule that
intersects the excess supply schedule above the level of the TRQ. With
excess demand schedule ED2, the TRQ is binding. Price and quantity reach
equilibrium at the intersection of the U.S. excess demand curve, ED2 and the
kinked excess supply curve, ES1, on its vertical segment, BC. With the
removal of the TRQ, increased imports of raw cane sugar drive down domestic
prices. With the same excess demand schedule, this situation corresponds to
a new equilibrium level: the point where ED2 intersects ES2, the excess
supply curve without import restrictions in the United States, with imports
rising to demand of QFM.

Because of USDA's loan program for sugar processors, however, domestic
prices would still not be free to drop to the world price level. Under the
loan program, producers would still be eligible to forfeit their sugar to
the government and receive the loan rate, PLR. The loan rate mechanism
provides a price floor for domestic sugar producers, maintaining domestic
sugar prices at the loan level, PLR, as in panel (a). However, with the
simultaneous elimination of the TRQ and the sugar loan program, the domestic
sugar price is free to fall below the loan rate level. In panel (b) of
figure 2, this situation corresponds to a new price and trade equilibrium
level. In the domestic market in panel (a), this corresponds to imports
increasing from Q2Q3 to Q1Q4. These increased imports lead to a drop in the
domestic price from P1 to P2. However, P2 is higher than the original world
price of PW.

Similarly, we removed the TRQ for imported refined, or "white," sugar. World
trade in refined sugar has increased because of policies in the European
Community, the entry of toll refiners,24 and a decrease in freight and
refining costs. In general, removing the TRQ for refined sugar would have
the same effect as removing the TRQ for raw sugar-the U.S. price for refined
sugar would decrease with an increase in imports of refined sugar. A lower
U.S. refined sugar price would then cause a decrease in the quantity of
domestic refined sugar supplied and a decline in refiners' demand for
domestic raw sugar.

the U.S. Sugar Program

This section describes our framework for modeling the economic gains and
losses to the various groups affected by the removal of the sugar program.
First, we describe the agricultural markets that transform sugar beets and
sugarcane into white sugar and corn into HFCS, an alternative to sugar in
soft drinks and other food products. The removal of import restrictions
under the TRQ would affect the raw cane and refined beet and cane sugar
markets by allowing free imports in these domestic markets. Second, we
estimate the welfare effects to sweetener processors, such as cane refiners,
HFCS refiners, and beet processors. Since a lower price for refined sugar
would increase the demand for it and could decrease the use of HFCS, the
price of HFCS could also decrease. The lower price for refined sugar would
also lead to a decrease in the quantity supplied by refiners, which in turn
would decrease the demand for sugar beets and sugarcane and, thus, the price
received by their producers. Third, lower prices for refined sugar and HFCS
would, other things being equal, lower the cost of production to
sweetener-intensive food goods industries. As a result, final consumers
would gain from lower prices for these foods as well as a lower price for
white table sugar.

Using the CARD sweetener model, we estimated the welfare effects to sugar
beet producers by specifying the supply of sugar beets, BS, as a function of
its price, Pb, the price of competing crops, Pg, and the price of an
aggregate input, Pf. The CARD domestic sugar model simultaneously solves for
sugar beet prices as well as acreage, yield, and production levels. Assuming
a quadratic form for profit in beet production, we can obtain a linear sugar
beet supply by taking the first derivative of profit with respect to the
price of beets:

with λ and μ summarizing the information on parameters bi and
prices Pg and Pf .25

Similarly, the supply of cane, CS, is a function of the price of cane, Pc,
and the price of an aggregate input, Pf . As in the case of beets, the
extended CARD domestic sugar model simultaneously solves for sugarcane
prices, acreage, yield, and production levels. Using these parameter
estimates, we again assume quadratic profits in cane production, and we
obtain a linear supply of cane:

If we assume a constant extraction margin ace and marginal cost pricing in
cane extraction, the cost function of raw cane sugar production is
Crcs=[(1/γc)Pc+ace]RCSS, where RCSS equals raw cane sugar supply.
Prices in raw cane sugar production, Prcs , obey the following arbitrage
condition to express marginal cost pricing in the extraction of raw sugar
from cane, with γc denoting the exogenous extraction rate:

We assume constant returns to scale in sugarcane processing, which implies
that there will be no welfare changes to sugarcane processors from price
changes due to an elimination of the program. Thus, we estimate changes in
economic welfare by the changes in quasi-profit or producer surplus realized
by cane and beet producers, Δ∏c and Δ∏b, defined as:

Parameter estimates of α, β, λ, μ, γc and ace are
available from the CARD sugar model.

With the lower price of sugar, the demand for corn might decrease if food
processors replace HFCS with sugar in production. We assume that the supply
of corn, COS, is determined by the maximum of the loan rate for corn,
LRcorn, or the market price of corn, Pcorn:

If the market price were higher than the loan rate, as it was earlier in the
1990s, then corn farmers would respond to this price. If there were a
decrease in the demand for corn caused by reduced HFCS production, corn
farmers would lose through a decrease in corn price and production.
Therefore, corn farmers would lose:

However, when the loan rate is higher than the market price, as it currently
is, the price signal perceived by corn farmers is the fixed loan rate. In
this case, farmers are eligible to receive loan deficiency payments from the
government.

Domestic sugar beet processors would also experience changes in economic
welfare from the extraction process. The domestic supply of white sugar,
WSS, comes from two sources that are perfect substitutes in supply: beet
sugar supply, WBS, and refined cane sugar supply. The supply of white sugar
from beets is a totally inelastic derived demand that comes from the
extraction of white sugar from sugar beets. With γb denoting the
exogenous rate of extraction of sugar from beets, prices in beet production
obey the following condition to express marginal cost pricing:

where abe denotes the extraction margin parameter in beet sugar extraction.

Finally, the welfare change for beet processors is then estimated as:

Because HFCS is an alternative to sugar in many foods, a change in the price
of sugar may affect the demand for HFCS, and in turn the price of HFCS,
translating into a change in economic welfare for HFCS producers. The supply
of HFCS, HFCSS, comes from extracting fructose from corn production with an
increasing marginal cost of extraction, Phfcs=(1/γcorn)
Pcorn+awmHFCSS, leading to the supply:

with awm denoting the marginal margin parameter in HFCS extraction and
γ being the actual extraction rate for HFCS from corn. Since we
extended the CARD sugar model to include linkages to the corn and HFCS
markets, we obtained all parameter estimates for these markets, as well as
extraction rates and margin parameters, from CARD. Rendleman and Hertel26
argue that because of feedback through by-product prices, HFCS supply is not
very price responsive. Equation 11 estimates the change in the welfare of
HFCS suppliers as captured by the change in the industry's producer surplus:

Cane sugar refiners experience changes in economic welfare with the
elimination of the TRQs for raw and refined sugar. Domestic refined cane
sugar comes from refining domestic and imported raw cane sugar. For cane
sugar refining, we assume that the supply of white cane sugar, WCS, is
competitive. Assuming the cost of producing refined cane sugar increases in
output, this cost consists of the cost of raw cane sugar and the refining
cost characterized by the margin parameter, arm. We also assume a fixed
proportion, γrc, between raw cane sugar and refined sugar. The
marginal cost of refining is equated to the output price to obtain a supply
schedule:

There is competitive price arbitrage between domestic sources of white
sugar, which equates the marginal cost of white cane sugar and beet sugar to
the white sugar price. Using this arbitrage condition and equations
(3) and (5), we have:

Therefore, the welfare change for domestic cane sugar refiners is obtained
by looking at the change in their quasi-profit, or producer surplus,
Δ∏wcs, resulting from the change in policy via Prcs (P0rcs to
P1rcs), output price (from P0ws to P1ws), and output change (from WCS0 to
WCS1):

We then estimate the economic welfare effects from changes in sweetener
prices for food-processing industries under two scenarios. The first
scenario assumes a constant markup, and thus a full pass-through of benefits
to consumers of lower input prices and thus output prices. The second
scenario holds output prices constant but allows food processors to absorb
the lower sweetener costs from eliminating the program in their marginal
cost function and thus in their profit margin.

The derived demand for refined sugar comes from food-processing industries
producing sweetener-intensive food goods. For food processing, we describe
the total cost function of each industry i in food processing as:

where C represents total cost, c~ represents the cost of a composite
sweetener (sugar and HFCS), "a" is an intercept term, b is a scaling term,
Ps is the price of white sugar, PHFCS is the price of high-fructose corn
syrup, Qfgi is output produced by food sector i, and A represents
information from the prices of other inputs. The derivative of the cost
function of the composite sweetener (15), with respect to the price of
sugar, is a nonhomothetic transformation of a Cobb-Douglas functional form27
and represents the output-constant industrial white sugar demand:

for all i in sweetener-using food goods where "a" is an intercept term and b
is a scaling factor to calibrate the own-price elasticity of demand
between -0.1 and -0.2.28 Similarly, the derivative of (15) with respect to
HFCS in each food-processing industry represents HFCS demand, HFCSD:

for all i in sweetener-using food goods where c is an intercept term and
once again b is a scaling factor to calibrate the own-price elasticity of
demand between -0.1 and -0.2.

Using this type of specification implies constant returns to scale in the
cost structure. Starting from each industry's cost function (15), we derive
the marginal cost underlying supply decisions:

where d is an intercept term that reflects the cost of other inputs.

From profit maximization with market power and conjectural variation,29
food-processing firms set price above marginal cost with markup coefficient,
θ, such that

Therefore, the price schedule of industry i is:

Equations (16), (17), (18), (19), and (20) determine the transmission of
lower sweetener prices into lower prices Pfgi to consumers of
sweetener-containing food goods. Several factors have a role in the price
transmission: the cost share of sweeteners in the cost of food processing,
the substitution possibilities within sweeteners (fructose and sugar) and
between sweeteners and other inputs, and finally the markup and its
evolution as prices change (McCorriston30 et al.).31 In equation (18), we
calibrated to replicate a "historical" marginal cost for the industry using
data from the Bureau of Economic Analysis on output price indexes for
4-digit SIC industries 2023 to 2099. The historical marginal cost is
estimated to be the historical price divided by 1.2, or a constant
20-percent markup of price over marginal cost. We use a 20-percent markup of
price over marginal cost as this figure is well within the estimates of
other analysts for the food manufacturing sector.32

Trade in food industries is ignored because net trade is a very small share
of total consumption or production in all food industries; these industries
tend to produce differentiated products, which do not face a strict price
discipline from the world market; and trade data are scarce and are only
available up to 1994. Hence, the equilibrium condition in the
food-processing markets is found by equating the price schedule (20) of each
industry to the corresponding Marshallian demand (24) as:

The welfare effect of the sugar program on each food industry is estimated
by the change in its profit, Δ∏fgi, resulting from the price and
output changes induced by the policy reform (from P1 to P0 and from FGS0 to
FGS1):

We estimated the welfare cost to the final sugar and HFCS consumer by
assuming a representative consumer with expenditure function E(P, U). In
this expenditure function, P is a vector of relevant consumer prices and U
denotes utility. We are interested in two types of goods: white sugar, WS,
and a vector of sweetener-containing food goods, FG. In addition, we have a
third aggregate, other goods, OG, for completeness. We use an incomplete
demand system approach-LINQUAD-as specified in LaFrance33 and Agnew.34 This
approach allows us to derive an exact welfare measure from an incomplete
demand system. In addition, we impose restrictions on the structure of
cross-price responses to reduce the number of parameters to be calibrated.
The price vector P is decomposed into P=(Pws , Pfg , Pog ), and income is
denoted by M, with subscripts indicating the respective commodities. The
subvector Pog is then dropped from the incomplete system. The Marshallian
demands for the two types of goods of interest, white sugar and
sweetener-containing food goods, denoted WSD and FGD, are:

for all i industries containing sweeteners. We use a system of consensus
estimates of own- and cross-price responses and income responses based on
Devadoss and Kropf,35 Bhuyan and Lopez,36 and Wohlgenant37 to derive
parameters ξ, ν, and χ. We solve the following system of
equations:

Equations (23) and (24) lead to an equivalent variation, EV, equal to:

Thus, we compute the change in expenditure, that would produce a change in
utility equivalent to the price changes, with superscripts 0 and 1 denoting
initial and final prices.

Finally, we list the welfare gains and losses from the presence of the sugar
program accruing to the various groups represented in the model.

Welfare Losses From the Presence of the Sugar Program

Losses to consumers (all prices higher as a result of the program):

Net losses to sweetener-using food processors (higher sweetener input prices
and higher output price):

Changes in quasi-profits to cane refiners (higher output price but
significantly higher input prices):

Welfare Gains From the Presence of the Sugar Program

Gains to beet producers (higher output price):

Gains to beet processors (higher white sugar price partly offset by higher
beet input prices):

Gains to cane producers (higher output price):

Net gains to HFCS producers (higher output price, net of slightly higher
corn input price):

Changes in quasi-profits to corn farmers:

Net gains to the foreign suppliers of raw sugar that have been given quota
rights consist of the unit rent times the total amount of the TRQ:

The net welfare loss is the difference between the additional costs of the
sugar program to the users of sweeteners and the gains to domestic sweetener
producers and processors:

This net welfare loss that results from the sugar program consists of
(1) production and consumption inefficiencies and (2) the transfer of rents
to foreign suppliers.

On the supply side, all data, parameters, and extraction rates used in the
U.S. component of the world sugar model are from CARD. To estimate welfare
effects for the food-processing industry, we identified a subset of 21
sweetener-using industries at the 4-digit SIC level (2023 to 2099). We took
these industries from the major categories of (1) dairy and frozen desserts,
(2) canned and preserved fruits and vegetables, (3) bread and bakery
products, (4) confectionery and chocolate products, (5) beverages, and (6)
miscellaneous food products industries. To calculate the demand and marginal
cost for sweeteners from these industries, we used data on the value of
shipments for each industry, the price of sugar and HFCS, and the total
quantities of sugar and HFCS sold for the years 1996 and 1998. For 1998 HFCS
cost data, we scaled each industry proportionately, using 1997 Bureau of the
Census data, to reproduce the exact total disappearance of HFCS in 1998. For
1996 HFCS data, we used cost data on corn sweeteners from the Bureau of
Economic Analysis. We obtained price data for sugar, HFCS-42, and HFCS-55
from USDA and industry sources.

For the LINQUAD model of final consumer demand, we used producer price index
data from the Bureau of Labor Statistics for each of our 21
4-digit SIC industries. We rebased the index numbers such that 1992=100. We
then divided each index number by the consumer price index number for that
year, also based with 1992=100, to capture how prices for each food group
compared with prices in the overall economy. USDA provided data on total
deliveries of sugar and HFCS. As previously noted, we obtained data on
income elasticities and own- and cross-price elasticities from several
sources in the economics literature.

Structure of the CARD International Sugar Model

The CARD International Sugar Model is a nonspatial, partial equilibrium
econometric world sugar model consisting of 29 countries/regions, including
a component for the rest of the world, to close the model.38 Major sugar
producing, exporting, and importing countries are included in the CARD
International Sugar Model. The model specifies only raw sugar trade between
countries/regions and does not disaggregate refined sugar trade from raw
sugar trade. Consequently, importers are not categorized as refiners or toll
refiners because the countries that specialize in these roles are well known
and stable over time. The model covers the following countries/regions:

 Algeria   Colombia             Guatemala  Mexico        South Korea
 Argentina Cuba                 India      Morocco       Thailand
 Australia Eastern Europe       Indonesia  Pakistan      Turkey
 Brazil    Egypt                Iran       Peru          Venezuela
 Canada    European Union       Japan      Philippines   Rest of the World
 China     Former Soviet Union  Malaysia   South Africa

The general structure of the submodel for each country includes behavioral
equations for the area harvested, yield, sugarcane and sugar beet production
on the supply side, and per-capita consumption and ending stocks on the
demand side. Equilibrium prices, quantities, and net trade are determined by
equating excess supply and excess demand across countries and regions. The
domestic price for each country or region is linked with a representative
world price through exchange rates and other policy wedges, such as tariffs
and transfer-service margins. Because of the overall scope of the model, it
is not feasible to include the complete empirical model in the text of this
report. The general framework for the submodel for each country consists of
the following:

AHt = f (AHt-1, RSPPt-1, RGPt-1, Trend),

Yield = f (Yieldt-1, Trend),

Cane and beet production = f (AH, Yield),

Per-capita sugar consumption = f (RSP, PCRGDP), and

Ending stocks = f (ESt-1, SC, RSP),

with AH denoting acreage, RSPP the cane or beet price, RGP the price of
alternative crops, PCRGDP real income per capita, ES ending stocks, SC sugar
consumption, and RSP the real raw sugar price. In many countries, beet or
cane prices are set by policy and can be treated as predetermined. Because
some countries lack information on the agricultural price, the model uses
the raw cane sugar price, RSP, instead of the agricultural price in
specifying the acreage response. In some countries, yield improvements are
captured by a time trend.

Although the CARD model in general can be used in either a dynamic or static
framework, we used its static version. In this way, the CARD international
sugar model uses contemporaneous price responses in supply, consumption and
inventory demand. Lagged acreage is fixed at the previous year level in the
acreage equations. Short-term supply elasticities are based on the
short-term price responses of the acreage equations, which reflect the cost
of adjustment between desired acreage and actual acreage. Long-term
responses are obtained by using the Nerlovian long-term response, which is
the short-term response divided by the partial adjustment coefficient of the
corresponding acreage equation. In addition, the elasticity values our
analysis uses are arc elasticities evaluated for 1998 between the historical
and post-reform levels.

We obtained data for the area harvested, yield, and sugarcane and sugar beet
production from the Food and Agricultural Organization of the United Nations
and data for sugar production, consumption, and ending stocks from USDA's
Production, Supply, and Distribution database. Cane and beet production is
tied to sugar production through the extraction rate. We obtained
macroeconomic data, such as the real gross domestic product, consumer price
index, population, and exchange rate from sources including the WEFA Group;
Project Link; and S&P/DRI.

We estimated the model for the period 1980 through 1998. We used simple
linear specifications and ordinary least squares in the estimation of these
equations to save degrees of freedom, given the short time series used. This
estimation approach overlooks the potential endogeneity of each country's
domestic sugar prices and treats them as exogenous in estimating acreage
response equations. The Caribbean raw sugar price is generally considered to
be the world market price.

Most elasticities in the CARD model are comparable to those used by Devadoss
and Kropf (1996), 39 Hafi et al. (1993),40 and Wohlgenant (1999).41 The
lagged own-price supply elasticities for sugarcane for Australia (0.02),
Brazil (0.07), Colombia (0.05), Cuba (0.01), Guatemala (0.02), Mexico
(0.002), South Africa (0.005), and Thailand (0.014) are highly inelastic in
the short run. This is consistent with the fact that these countries are
large producers of sugarcane and exporters of sugar and can harvest several
annual crops, called ratoons, from one planting of sugarcane. Therefore,
there is limited acreage adjustment to price fluctuations in the short run.
These low elasticities are largely the result of the historical policy of
acreage allotment. The own-price supply elasticities for sugar beet
production are generally not as inelastic as they are for sugarcane, except
for the former Soviet Union (0.002).

The own-price demand and income elasticities reflect the fact that in many
developing countries, sugar is considered a staple in the diet and consumers
look to sugar to fulfill basic caloric requirements. The elasticities
implied in the CARD model are very comparable to ones reported in the
literature. For several countries, when more recent data were not available
for the econometric estimation, we borrowed elasticities from Hafi et al.
(1993) and from Devadoss and Kropf (1996).

Although the CARD international sugar model does not disaggregate raw and
refined sugar, we complemented the existing model with an additional
equation to endogenize the world price of refined sugar following the
removal of the refined sugar TRQ in the United States. Consistent with Hafi
et al. (1993), we specified a reduced form to approximate the
rest-of-the-world supply faced by the United States. However, there is no
explicit aggregation of excess supply in the various countries to derive
this equation, as is the case for the raw sugar market. This equation is of
the form

IWSrow =a(Pwws)αws (Pwrcs)βcs

where IWS is imports of white sugar and the elasticities of white sugar and
raw cane sugar are αws = 0.83 and βcs = -0.44, respectively.

These elasticity values come from Hafi et al. (1993) and are medium-run
estimates. Parameter a is chosen to calibrate the IWS to the existing
refined sugar TRQ level in the United States, prior to the TRQ's removal.

This equation is treated as the rest-of-the-world supply, which underlies
the import supply faced by the United States. The latter is then equated to
the U.S. excess demand for white sugar to close the white sugar market in
our analysis, once the TRQ is removed.

Economic Analysis of a Tariff-Rate Quota

The sugar program includes the use of a TRQ to limit imports of raw and
refined sugar that come into the United States. Under a TRQ, a lower,
"in-quota" tariff applies to a limited quantity of imports, while a higher,
often prohibitive, "over-quota" tariff applies to any imports that exceed
the quota. Our model follows Moschini,42 as well as Meilke and Lariviere,43
in its representation of the import supply under a TRQ.

Figure 3 illustrates the effects of the TRQ on the excess supply curve and
shows how this type of policy affects the incentive to import. Our model
describes three excess supply conditions-ESf , ES1 , and ES2. ESf represents
the excess supply curve facing the United States if it did not have import
restrictions, while ES2 is the excess supply curve under the in-quota
tariff. However, under the current U.S. sugar TRQ, producers from very few
quota-holding countries pay the in-quota tariff and most pay none because of
preferential arrangements. Therefore, without loss of generality, we will
assume that ES2 is actually the free trade excess supply curve facing the
United States in the absence of import restrictions.

Overall, the TRQ introduces a discontinuity in the excess supply curve
facing an importing country at the level of the quota, as shown in the bold
curve ES1.44 As figure 3 illustrates, the TRQ can result in three potential
trade and price outcomes, depending on the U.S. excess demand for foreign
sugar. First, if the excess demand schedule intersects the segment AB on the
lower portion of the excess supply curve, such as ED1 , then there is a
supply response to shifts in the excess demand schedule and both quantity
and price will change. In this area of the curve--the "in-quota" portion of
excess supply--the domestic price equals the world price because the
importing country is actually not imposing any border measures. Second, if
the excess demand schedule intersects the excess supply schedule at the
level of the TRQ, such as ED2 , the domestic price will equal P1 and the
world price PW . Rents are generated for the holders of the quota rights
corresponding to the binding TRQ. In this case, the quota rent will be equal
to the difference between the domestic price and the world price. This
situation corresponds to the current U.S. sugar import situation. Third, if
the excess demand schedule intersects the excess supply schedule
sufficiently beyond the TRQ, such as ED3 , both the price and the quantity
of imports supplied will respond to shifts in the excess demand for the
imports and the rest of the world will satisfy this demand.

Estimated Net Losses to the U.S. Economy Accruing to Foreign Sugar
Importers, Fiscal Year 1998

              Allocation and imports in short tons (raw value

        Country        TRQ allocationa  Actual imports  Estimated economic
                                                               loss
 Argentina             72,300           72,200         $14,600,000
 Australia             139,500          140,100        28,400,000
 Barbados              8,600            0              0
 Belize                18,500           18,500         3,800,000
 Bolivia               13,400           12,600         2,600,000
 Brazil                243,700          242,900        49,300,000
 Colombia              40,300           37,200         7,500,000
 Congo                 8,000            8,000          1,600,000
 Costa Rica            25,200           25,200         5,100,000
 Cote d'Ivoire         8,000            30             6,000
 Dominican Republic    295,800          294,500        59,700,000
 Ecuador               18,500           18,500         3,800,000
 El Salvador           43,700           44,000         8,900,000
 Fiji                  15,100           11,900         2,400,000
 Gabon                 8,000            0              0
 Guatemala             80,700           80,400         16,300,000
 Guyana                20,200           20,200         4,100,000
 Haiti                 8,000            0              0
 Honduras              16,800           16,900         3,400,000
 India                 13,400           13,800         2,800,000
 Jamaica               18,500           18,300         3,700,000
 Madagascar            8,000            8,100          1,600,000
 Malawi                16,800           13,200         2,700,000
 Mauritius             20,200           20,400         4,100,000
 Mexico                27,600           27,600         5,600,000
 Mozambique            21,800           22,100         4,500,000
 Nicaragua             35,300           35,400         7,200,000
 Panama                48,700           48,700         9,900,000
 Papua New Guinea      8,000            100            20,000
 Paraguay              8,000            5,500          1,100,000
 Peru                  68,900           69,000         14,000,000
 Philippines           226,900          222,800        45,200,000
 St. Christopher-Nevis 8,000            8,000          1,600,000
 South Africa          38,700           38,800         7,900,000
 Swaziland             26,900           27,000         5,500,000
 Taiwan                20,200           20,200         4,100,000
 Thailand              23,500           23,500         4,800,000
 Trinidad-Tobago       11,800           12,100         2,400,000
 Uruguay               8,000            8,200          1,700,000
 Zimbabwe              20,200           20,100         4,100,000
 Total                 1,763,700        1,706,030      $346,026,000

Note: Each country supplying sugar to the United States under the TRQ is
limited to exporting sugar that solely originated within that country.

a Allocations are based on countries' exports to the United States from 1975
through 1981.

Source: Sugar and Sweetener Situation and Outlook, Economic Research
Service, USDA (Sept. 1999).

Comments From the U.S. Department of Agriculture

The following are GAO's comments on the U.S. Department of Agriculture's
written response to our draft report dated May 16, 2000. Based on agency and
industry comments, we adjusted our model to more fully account for certain
transportation costs. As a result, cost and benefit estimates referenced in
USDA's comments do not reflect those contained in the final report.

1. After thoroughly examining each of the Department's concerns and
discussing them with experts outside GAO, we continue to believe that our
analytical approach and model for estimating the costs and benefits of the
sugar program are comprehensive and methodologically sound and that the
estimates yielded by our model are reasonable. In developing the model, we
took a number of actions to ensure that it was methodologically sound.
First, we contracted with a well-known expert in modeling the international
trade of agricultural commodities and with a prominent agricultural
economist to work with us in developing the model. In December 1999, we sent
our proposed model to four outside academicians specializing in agricultural
economics and international trade economics and revised the model in
response to their comments. We also sent our proposed model to USDA for
review at that time. However, USDA did not provide any comments.
Furthermore, we asked two of the agricultural economists to review our final
model and results before we sent our draft report to USDA for comment.

2. Our approach is not strictly based on a traditional "gains-to-trade"
argument, and its validity does not depend on world prices being free of
market distortions. We did not distinguish between the effects of
comparative advantage and export subsidies or other forms of market
intervention in determining the price at which U.S. sugar users would be
able to purchase sugar. To estimate the costs and benefits of the sugar
program, our model compared baseline domestic and world prices for sugar in
1996 and 1998 with the domestic and world prices that we estimated would
occur if the sugar program were eliminated, other things being equal. These
estimated prices were higher than the baseline world prices because our
model accounted for the impact on world prices of the higher U.S. sugar
imports that would have occurred if the program had been eliminated.
Comparing these estimated prices with the baseline prices was appropriate
for estimating the effects of the sugar program because the estimated world
prices reflect what domestic sugar users would have been expected to pay for
sugar in those years if the sugar program had been eliminated. This
comparison remains appropriate regardless of whether the world price is
influenced by subsidies by countries with surplus production.

3. While the world sugar price that would result from multilateral reform is
uncertain, we believe that several elements of the Schmitz and Vercammen
analysis (University of California, Berkeley, 1990) led to a high price
estimate that is unlikely to be observed. In particular, they assume a
relatively high (in absolute value) excess demand elasticity of −0.5
and a very inelastic excess supply schedule from the rest of the world. That
is, they assume that a small increase in the world price will lead to a
large reduction in the quantity of sugar demanded but only a small increase
in the quantity of sugar that producers in other countries would want to
supply. Both assumptions tend to cause the world price to rise more
following multilateral reform than would happen if assumptions of less
elastic excess demand and more elastic excess supply were used. Schmitz and
Vercammen also overstate the world price that would follow multilateral
reform because they do not account for the supply responses of all exporting
countries.

4. On the basis of our review of the literature, we believe that the effect
of multilateral reform of sweetener markets on world sugar prices is highly
uncertain. Even if world prices were to rise, this increase would likely be
gradual because the reforms would likely be phased in over a substantial
period of time. Therefore, we regard as speculative USDA's view that
multilateral reform following the elimination of the U.S. sugar program
would erase the welfare gains that we attributed to the program's
elimination.

5. We disagree with USDA's assertion that our model is not well documented
and includes incomplete references. However, we added language to the scope
and methodology section of our report to explain further the relationship
between the model we used in our analysis and the CARD model.

6. We clarified where on the linear supply schedule the supply elasticities
that we use were estimated.

7. We disagree with USDA that the import supply elasticities are
inconsistent with our model. We clarified our report to indicate that these
elasticities are derived explicitly from the CARD model, which includes the
sugar sectors of the countries listed in appendix III.

8. We disagree with USDA's assertion that our model presumes instantaneous
and costless adjustments within the United States. In fact, our short-run
elasticities are smaller than our longer-term elasticities because they take
adjustment costs into account. However, we added clarifying language to
appendix III.

9. We disagree. Our draft report defined each of the symbols describing the
variables used in the LINQUAD model.

10. We disagree with USDA's assertion that the LINQUAD model is experimental
and not well documented. The LINQUAD approach is the latest development of
the incomplete demand system approach pioneered by Professor Jeffrey
LaFrance at the University of California, Berkeley. The LINQUAD approach is
fully documented in several working papers that are publicly available, and
Professor LaFrance's work on incomplete demand systems has appeared in many
leading economics journals. Professor LaFrance first published his results
in 1990 in an article of the Australian Journal of Agricultural Economics.
The method was first empirically implemented in an application to dairy in a
1991 article of the American Journal of Agricultural Economics. The model
was also documented in a thesis that received the American Agricultural
Economics Association's Outstanding Thesis Award in 1999. The incomplete
demand system model was appropriate for our work in modeling sugar demand
because it allowed us to compute a consistent welfare measure that included
direct and indirect sugar consumption in food.

11. We disagree that our model's treatment of the sugar-processing sector
was inadequate. Although no equations for sugarcane and sugar beet acreage
are presented in the report, our model solves for acreage as well as prices
and production levels. We clarified this point in appendix II.

12. We disagree with USDA's characterization of equation 15 (now equation
13). This equation establishes a condition that must be met for the model to
solve--at the margin, the cost of producing more white cane sugar and more
beet sugar must be equal to each other and to the price of white sugar.
Although we do not explicitly model the decision to exit the industry, the
parameter representing the refining margin takes into account changes in
output.

13. Our draft report inadvertently reported slopes rather than elasticities.
We revised the report to present the correct elasticities, which are much
lower. The supply response reported in table 7 is consistent with these
lower elasticities.

14. We agree with USDA that the revenues that producers receive are based on
contracts with processors. However, we believe that the contractual
arrangements, and the sugar program itself, lead to stable price
expectations for producers that are captured in equations 3 and 4 (now
equations 1 and 2) -- producers' supply functions. In our model, prices at
all stages are determined simultaneously by the satisfaction of equilibrium
conditions in all markets.

15. We do not think that producers' ownership of processing plants is an
important issue in our analysis. If producers' ownership were an important
influence on the supply responses of producers to price changes, the
expected direction of that influence would be to make supply less elastic
(that is, producers might reduce supply less in response to a price decline
than if they did not own the processing plants). We do not believe that
incorporating this possibility into our model would have much effect because
the supply elasticities we use are already low.

16. The requirement for minimum payments from processors to producers is not
a concern for our model because the supply decisions in our model are
calibrated using historical data that include minimum payments when
relevant.

17. We added language to the report to clarify that the model's assumption
of constant returns to scale in sugarcane processing implies that there are
no welfare changes to sugarcane processors from price changes due to the
elimination of the sugar program. Accordingly, we did not report any welfare
effects.

18. We disagree with USDA that conducting the analysis at the national level
implies any misspecification of our model, although we agree that more
detail about regional production might be interesting. Our model does not
specifically account for regional differences in cost structure or
alternative crops and we agree with USDA that producers in different regions
would be affected differently. Nevertheless, we remain confident that our
model measures the total effects of the U.S. sugar program appropriately.

19. We revised our discussion of equation 20 (now equation 18) to say that
the intercept term represents the cost to food manufacturers of inputs other
than sweeteners (rather than a reflection of market power). As shown in
equation 22 (now equation 20), market power results in a markup of price
above marginal cost.

20. We added a definition of "conjectural variation" and explained how this
parameter of market interdependence is related to industry markup.

21. We disagree with USDA's assertion that our model did not include some of
the factors cited by McCorriston et al. For example, our model includes the
influence of the retail price elasticity of demand on price pass-through. It
also includes the influences of substitution between sugar and HFCS and the
cost share of sweeteners in food products, but it does not include the role
of marketing inputs.

22. We clarified that the ES2 supply curve in figure 2 (and the ESf supply
curve in figure 3) refer to the excess supply curves that would exist in the
absence of import restrictions in the United States. They were not meant to
represent excess supply curves that would result from free trade in all
world sugar markets.

23. USDA's conjecture that the domestic sugar industry could successfully
pursue an antidumping action that would raise the price of foreign sugar for
U.S. sugar users is speculative. Accordingly, we believe that it is
appropriate to compare baseline prices with estimated world prices in the
absence of the program, other things being equal, as we have done. See
comment 2.

24. See comment 3.

25. We do not question the objective of pursuing international trade reform
and free world markets through trade negotiation. To estimate the effects of
the U.S. sugar program in this report, we needed to make the analytical
assumption that all other factors, including other countries' sugar
programs, would remain unchanged. If we had not done so, we would not have
been able to separate the effects of the U.S. program from those of other
countries' programs. Accordingly, when we say that domestic sugar prices are
artificially high, we mean in comparison with the prices that would prevail
if the U.S. sugar program were eliminated and other conditions remained
unchanged. See also comment 37.

26. We disagree with USDA and stand behind our finding that eliminating the
sugar program would have a limited impact on the price of HFCS. While HFCS
producers benefited from the sugar program in the 1980s and early 1990s, we
believe that the domestic HFCS and sugar markets have since been decoupled
because of domestic price, cost, and market conditions. The president of the
U.S. Corn Refiners' Association has also expressed that view. As shown in
table 4 in appendix I, the weighted average wholesale price of HFCS dropped
from 22 cents per pound (dry weight) in 1992 to 12.4 cents per pound (dry
weight) in 1998, while the U.S. wholesale price of refined beet sugar was
above 25 cents per pound throughout this period. The likelihood of
substitution between sugar and HFCS is less than in prior years because
technological advances have improved HFCS products and created more
specialized sweetener markets.

27. We revised appendix II to clarify that the functional form we used does
not imply that the cost shares of sugar and HFCS need to remain constant as
the price of sugar changes.

28. We added references to the literature on the price elasticity of demand
for sugar and HFCS.

29. We added language to the report to indicate that the lack of
substitution between sugar and HFCS is due not just to the relative prices
of the sweeteners but also to the increasingly specialized uses of
sweeteners, which have limited the substitutability between sugar and HFCS.

30. In our model, the relationship between raw and refined sugar prices
depends on the refining margin as well as on technical coefficients involved
in the refining process. Since the refining margin will be smaller at lower
levels of output, our model takes into account the lower margins for
domestic cane refiners that will result from increased imports of refined
cane sugar.

31. We disagree with USDA's assertion that we have not adequately analyzed
the increased competition that the domestic refining industry could face as
a result of the elimination of the refined sugar TRQ. If imports of refined
sugar increased following the elimination of the refined sugar TRQ, the
refining margin for domestic refiners would decline, according to our model.
See comment 30.

32. Our report does not predict the degree to which cost reductions
resulting from the elimination of the sugar program would be passed through
to consumers. We present two estimates of potential benefits to consumers
based on different assumptions about the pass-through of cost reductions
from food manufacturers to consumers; both estimates are based on a full
pass-through of cost reductions for the production of table sugar. For
reasons given in the report, we believe that a full pass-through is more
likely for table sugar than for sugar-containing products. In estimating
potential benefits, we assumed "other things being equal" (no other changes
in input costs) while the data that USDA presents on changes in wholesale
and retail prices do not. Changes in the prices of marketing inputs could
have offset the reduction in wholesale prices in determining retail prices.

33. It is unclear what effect accounting for marketing inputs would have had
on our results.

34. See comment 32.

35. We disagree with USDA's assertion that a comparison of our estimate of
the benefits that sugar beet growers receive from the program with data on
growers' revenues and expenses implies that our estimate is clearly
overstated. The benefits of a program that keeps price substantially higher
than it otherwise would be can be very high in the short run, when supply
elasticities are low. In the long run, the benefits of maintaining the sugar
program are less because in the absence of the program a wider set of
adjustment opportunities would be available to producers, implying larger
supply elasticities.

36. We revised the report to address USDA's concern.

37. We disagree that our characterization of domestic sugar prices as
artificially high is inflammatory and unprofessional. The sugar program
keeps domestic sugar prices above the level at which they would be if the
program were eliminated (other things being equal), which is what we mean by
artificially high. Furthermore, USDA has some flexibility in its
implementation of the sugar program, and our report entitled Sugar Program:
Changing the Method for Setting Import Quotas Could Reduce Cost to Users
(GAO/RCED-99-209, July 26, 1999) found that, from 1996 through 1998,
domestic sugar prices were over 2 cents per pound higher than necessary to
avoid sugar loan forfeitures.

38. We agree that the current conditions in sugar markets differ from those
observed in 1996 and 1998. However, 1998 was the most recent year for which
price and quantity data were available for our analysis when we began our
study. As stated in the report, our estimates apply only to 1996 and 1998.

39. We agree that the North American Free Trade Agreement has the potential
to significantly influence the operation of the sugar program and the
domestic sugar market. We note in appendix I that USDA will spend
substantial federal funds this fiscal year because the sugar program has
resulted in significantly increased domestic sugar production--USDA recently
announced that it will seek to purchase 150,000 tons of domestic sugar to
reduce the cost of expected sugar program forfeitures. Because of our need
to establish a base case for our economic model, we examined 1998, the most
recent year for which sweetener price and quantity data were available.
Accordingly, we state on page 4 of our report, "we did not analyze how the
gradual reduction of U.S. tariffs for Mexican sugar through 2008 under the
North American Free Trade Agreement might affect domestic sugar prices.
However, our model could be used for this analysis as well as other types of
analyses in the future."

40. The comments of former Under Secretary Moos on our 1993 report contrast
with USDA's official written comments in 1993 (included as an appendix to
the 1993 report), which state, "Overall, this is a reasonable report with no
major data problems. The costs and benefits, derived using assumptions of
hypothetical policy alternatives, are well within the range of most
research." We continue to believe that our 1993 report provided a reasonable
estimate of the cost of the sugar program to U.S. sugar users for the period
analyzed. More important, we believe that although the precise level of the
price premium is subject to debate, the program and policy problems we
identified in 1993 are still relevant.

41. In preparing this report, we developed a more comprehensive model than
the one we used for our 1993 report. Under our current approach, the
estimated prices that would exist in the absence of the sugar program were
derived by our model rather than assumed from outside the model, as was done
for the 1993 report. Comparing these estimated prices with the baseline
prices was appropriate for estimating the effects of the sugar program
because the estimated world price reflects what domestic sugar users would
have been expected to pay for sugar in those years if the sugar program had
been eliminated. To estimate the effects of the U.S. sugar program by
itself, we needed to make the analytical assumption that all other factors,
including other countries' sugar programs, would remain unchanged. Without
this assumption, we would not have been able to isolate the effects of the
U.S. program from the effects of other countries' programs. In preparing the
1993 report, if we had chosen to use a price that seemed more consistent
with the elimination of the U.S. program alone, we might have selected a
lower price for the "no program" scenario, which would have led to higher
estimates of the cost of the program to sugar users and the benefits to
producers.

42. We revised the report to address USDA's concern.

43. We revised the report to indicate that while the minimum requirement of
the refined sugar TRQ is 30,900 tons, USDA has set the TRQ above this level
in recent years.

44. We revised the report to address USDA's concern. The Agriculture
Appropriations Act for fiscal year 2000 suspended the sugar marketing
assessment for fiscal years 2000 and 2001.

45. We revised the report to address USDA's concern.

46. We revised table 1 to note that the basis point for the world raw sugar
contract price is the Caribbean. According to a sugar market expert, the
cost to transport raw sugar from the Caribbean to the United States is about
1.5 cents per pound. We also refined our model to more fully account for
these transportation costs in our welfare analysis.

47. We revised the report to address USDA's concern.

48. We did not include a line in table 5 for the revenues that the federal
government derives from sugar marketing assessments because those revenues
represent only part of the sugar program's financial effects on the federal
government. For example, other effects include increases in the government's
costs of purchasing food and conducting food assistance programs because the
prices for sugar and sugar-containing foods are higher under the sugar
program. Several years ago, we estimated that the government's additional
cost of purchasing food and providing the level of food assistance that it
delivered in 1994 was approximately $90 million. At that time, the
government collected about $30 million annually in marketing assessments.

49. We did not include a line for the benefits of the sugar program to
sugarcane processors because they would see no welfare change from the
elimination of the sugar program. See comment 17. The cost reported in table
5 for HFCS manufacturers and corn growers is very small; in general, the
sugar program has little effect on these groups because of the decoupling of
the sugar and HFCS markets. See comment 26.

50. We adjusted our model to more fully account for certain transportation
costs. As a result, the 1996 estimated U.S. wholesale refined sugar price
without the sugar program increased from 20.5 cents per pound in the draft
report to 21.77 cents per pound in the final report. This revised price is
higher than the 1996 average price of HFCS-55 of 20.6 cents per pound. To
further clarify our model's results, the final report states that
specialization in sweetener use was also important in estimating the effect
of the sugar program on HFCS producers. See comment 29.

51. We continue to believe that transfers to foreign sugar producers that
receive quota allocations are appropriately treated as a component of
estimated net losses to the U.S. economy due to the sugar program. The
additional revenues that they receive from higher prices as a result of the
program represent a transfer outside the United States and, therefore, are a
net loss to the U.S. economy. By suggesting that these transfers are more of
a foreign policy issue than a sugar program issue, USDA implies that in the
absence of the program the United States would provide the quota-holding
countries with other forms of assistance equal to the benefits their
producers currently derive from the program. We believe such a conclusion is
speculative.

52. Although the estimated refining margin for 1996 is high, it is not as
high relative to other years once adjustments are made for price level
differences (see table 1 in app. I). Our estimate of 6.89 cents per pound
for 1996 is in 1999 dollars. Expressed in 1999 dollars, the 14-year high
(achieved in 1990) was 5.73 cents per pound, not 4.77 cents. Nonetheless,
one limitation of our model is that the world refined and raw sugar markets
are not fully linked. If our estimated price for refined sugar without the
program is too high relative to the estimated no-program price for raw sugar
for 1996, the implication is that we have overestimated the benefits to
refiners and underestimated the benefits to other sugar users (food
manufacturers and consumers) from eliminating the sugar program.

53. See comment 13.

54. We revised the report to address USDA's concern.

55. Our model does not conduct a general equilibrium analysis to determine
the effects of eliminating the sugar program on the prices of other crops
because of acreage being transferred from sugar to these other crops.
However, we believe that any transfers out of sugar would likely have
minimal effects on the total production of most other crops because of the
relatively small acreage involved compared with the acreage already planted
in those crops.

56. We added language to the report to show that the 20-percent markup we
use is consistent with what others have used in their analyses.

Comments From the American Sugar Alliance

The following are GAO's comments on the American Sugar Alliance's (ASA)
written response to our draft report dated May 5, 2000. Based on USDA and
industry comments, we revised our model's final estimates to more fully
account for certain transportation costs. As a result, cost and benefit
estimates referenced in ASA's comments do not reflect those contained in the
final report.

1. We disagree that the methodology used in our 1993 report on the sugar
program was flawed. Nonetheless, we developed a more comprehensive economic
model for our current analysis, and while we acknowledge that no economic
model completely depicts reality, we are convinced that our current model is
methodologically sound and that the estimates yielded by our model are
reasonable. In developing the model, we took a number of actions to ensure
that it was methodologically sound. First, we contracted with a well-known
expert in modeling the international trade of agricultural commodities and
with a prominent agricultural economist to work with us in developing the
model. In December 1999, we sent our proposed model to four outside
academicians specializing in agricultural economics and international trade
economics and revised the model in response to their comments. We also sent
our proposed model to USDA for review at that time. However, USDA did not to
provide any comments. Furthermore, we asked two of the agricultural
economists to review our final model and results before we sent our draft
report to USDA, ASA, and the U.S. Cane Sugar Refiners' Association for
comment.

2. We disagree with ASA's assertion that our findings are based on
comparisons with a meaningless world price. In estimating the costs and
benefits of the sugar program, our model compared baseline domestic and
world sugar prices with an estimate of the domestic and world prices that
would have been observed if the sugar program had been eliminated, other
things being equal. Regarding the extent to which cost reductions would be
passed through to consumers in the absence of the sugar program, the report
presents two estimates showing how the benefits might be distributed based
on two different sets of pass-through assumptions. We did not predict the
extent to which cost reductions would be passed through to final consumers.
See comments 4 and 5.

3. We disagree that our report is biased toward the views of the sugar
program's critics. We used a standard economic welfare analysis methodology
that many economists have applied to evaluate how different groups within a
society are affected by specific economic policies.45 Our report provides a
quantitative analysis of the U.S. sugar program's effects, but we do not
take a position on what modifications, if any, need to be made. Those
decisions rest with the Congress. See comment 5 for a discussion of our
assumptions about manufacturers' and retailers' behavior in passing through
cost reductions.

4. As discussed in comment 2, we disagree with ASA's assertion that our
findings are based on comparisons with a meaningless world price. Our model
estimated domestic and world sugar prices in the absence of the program.
These estimated prices were higher than the baseline world prices because
our model accounted for the impact on world prices of higher U.S. sugar
imports that would have occurred if the program had been eliminated.
Comparing these estimated prices with the baseline prices was appropriate
for estimating the effects of the sugar program because the estimated world
price reflects what domestic sugar users would have been expected to pay for
sugar in those years if the sugar program had been eliminated. This
comparison remains appropriate regardless of whether the world price is
influenced by subsidies by countries with surplus production or whether
sugar is produced in countries with lower labor and environmental standards.

5. As discussed in comment 2, we disagree with ASA's assertion that our
findings are based on the assumption that food manufacturers and retailers
would pass all of their cost reductions through to final consumers. Our
report estimated the total cost of the sugar program to sweetener
users--sugarcane refiners, manufacturers of sugar-containing foods, and
final consumers. We did not predict the extent to which cost reductions
would be passed through to final consumers. Instead, we presented two
estimates of the benefits to consumers if the sugar program were eliminated,
using different assumptions about the pass-through of cost reductions. One
estimate assumed that (1) cost reductions for table sugar were passed
through to final consumers (because it is a homogeneous product and likely
to be more price competitive) but (2) cost reductions for sugar-containing
food were not passed through. The other estimate, which shows the maximum
benefit that consumers could receive, is based on the assumption that all
cost reductions were passed through to consumers. We revised our discussion
on page 4of the pass-through of cost reductions if the sugar program were
eliminated to further clarify the effects under alternative pass-through
scenarios.

6. We agree with ASA that other countries intervene in the world sugar
market. However, to estimate the effects of the U.S. sugar program by
itself, we needed to make the analytical assumption that all other factors,
including other countries' sugar programs, would remain unchanged. Without
this assumption, we could not isolate the effects of the U.S. program from
the effects of other countries' programs. Our estimates represent the actual
cost of the sugar program to the U.S. economy under conditions as they were
in 1996 and 1998.

7. As discussed in comment 6, we did not attempt to estimate the free trade
price of sugar. In addition, as discussed in comment 4, we believe that the
appropriate price comparison is between baseline domestic and world sugar
prices in 1996 and 1998 and an estimate of the prices that would have been
observed if the sugar program had been eliminated, other things being equal.

8. As stated in the objectives, this report focuses on the U.S. sugar
program's (1) costs to domestic sweetener users, (2) benefits to domestic
sugar and HFCS producers, and (3) net effects on the U.S. economy. Both the
resport's letter and appendix II present the results of our model, including
the estimated benefits to sugarcase, sugar beet and HFCS producers. However,
our model does not attempt to quantify every indirect cost and benefit
associated with the program. See comments 13 and 17.

9. We disagree with ASA that our estimates of the net costs of the sugar
program are "misleadingly high." See comments 10 and 11.

10. We disagree that transfers from the U.S. economy to foreign sugar
producers should not be considered as a component of the estimated net
losses associated with the sugar program. From the perspective of the U.S.
economy, higher prices paid by domestic sugar users and received by domestic
producers and processors as a result of the program are internal transfers
with no net effect. However, foreign producers selling sugar in the United
States also receive higher prices because of the program. The additional
revenues they receive as a result of the program represent a transfer
outside the United States and, therefore, are a net loss to the U.S.
economy. Whether the United States would incur similar costs if the sugar
program did not exist in the form of either direct aid to countries whose
producers are benefiting from the program or costs resulting from
instability, as suggested by ASA, is speculative.

11. We disagree with ASA's assertion that our model assumes that U.S. sugar
producers are inefficient. The types of efficiency losses we discuss occur
when efficient producers produce too much of the wrong crop, or consumers
consume too little of certain foods, because government programs maintain
artificially high prices.

12. The report does not state that sugar production is a less efficient use
of resources than any other agricultural enterprise.

13. We agree that we did not estimate the effects on other crops if acreage
were transferred out of sugar, but we disagree that not doing so was
egregious. Our model does not conduct a general equilibrium analysis to
determine the effects on acreage and price of all other commodities.
However, we believe that any transfers out of sugar into other field crops
would likely have minimal effects on total production because of the
relatively small acreage involved compared with the acreage already planted
in those crops. For example, a transfer of 33 percent of sugar beet acreage
into hard red spring wheat acreage, would have increased the hard red spring
wheat acreage by only about 3 percent in 1998. See comment 17.

14. We disagree. Using generally accepted economic principles, total U.S.
consumption of sugar and sweetener-containing products would rise, other
things being equal, as U.S. sugar prices were reduced.

15. In the absence of the sugar program, we believe that it is unlikely that
no cost reductions would be passed through to final consumers, particularly
in the case of table sugar. In our analysis, we assumed "other things being
equal" (no changes in other input costs), while the data that ASA presents
comparing prices at different levels--especially for sugar-containing
foods--do not. Changes in the prices of other food inputs between 1990 and
1999 could have been more important and could have overwhelmed the effect of
declining wholesale refined sugar prices on retail prices. Moreover,
wholesale refined sugar prices did not stay down continuously over the
entire decade, but increased again in 1996, 1997, and 1998. See comment 5.

16. We recognize that the demand for table sugar and many sugar-related
products is relatively inelastic with respect to price (that is, price
increases will lead to only relatively small declines in consumption), and
we included these assumptions in our model. However, we disagree with ASA's
assertion that there would be no effect on consumption at higher prices.

17. We recognize that we did not consider indirect effects associated with
the elimination of the sugar program in our analysis. These effects could
include losses to some groups, as suggested by ASA, but could also include
gains by other groups. For example, if sugar users paid lower prices for
sugar, they would be able to spend more on other items, which could lead to
increased economic activity in some other sectors. Our approach estimated
the program's major costs and benefits, while still maintaining tractability
as well as appropriate market detail. A general equilibrium analysis would
take both these gains and losses into account. We added to the report a
discussion of the differences between the partial equilibrium approach we
used and a general equilibrium analysis, indicating that using the latter
approach would require sacrificing considerable detail.

18. We did not assess the competitiveness of U.S. sugar producers relative
to producers in other countries. We assessed the costs and benefits of the
sugar program in the context of the world market in 1996 and 1998. See also
comment 6.

19. See comment 11.

20. We disagree that comparing U.S. and foreign retail sugar prices would
have been the most direct way to assess the effect of the U.S. sugar program
on consumers. Retail sugar prices in other countries do not represent the
price at which U.S. sugar users could have obtained sugar in the absence of
the program, which is the appropriate price to compare with current U.S.
prices (see comment 4). As a result, we did not compare U.S. retail sugar
prices with other countries' retail sugar prices. However, ASA's chart C
shows that countries whose retail sugar prices were higher than the U.S.
prices are mostly countries from the European Union and Japan, where support
for sugar producers is also high. (The Organization for Economic Cooperation
and Development estimates that sugar consumer support in the European Union
was even higher than in the United States in 1998.) Chart C also shows that
retail sugar prices in Australia and Canada, which have relatively free
trade in sugar, were lower than prices in the United States and the other
countries shown.

21. The minutes worked analysis in chart D may be a misleading indicator of
low sugar prices for U.S. consumers in comparison with consumers in other
countries. This analysis appears to be more dependent on a country's average
wage rate than on the retail price of sugar in explaining the differences in
the minutes of work required to buy a pound of sugar among countries. See
also comment 20.

22. We agree that U.S. sugar processors have not forfeited sugar to the
government since 1994. However, USDA recently announced plans to purchase
150,000 tons of sugar to reduce the cost of expected sugar program
forfeitures this fiscal year.

23. We revised the report by adding these increased assessments to the list
of the 1996 Farm Act's modifications. According to a USDA official, the
assessment on processors raised revenues of about $30 million in fiscal year
1998 but has been suspended for fiscal years 2000 and 2001. We note,
however, that these revenues represent only part of the sugar program's
financial effects on the federal government. The higher domestic cost of
sugar increases the government's expenditures for the military's food
purchases and for funding domestic food assistance programs. Several years
ago, we estimated that the sugar program increased the government's
expenditures for these purposes by about $90 million in 1994.

24. As discussed in comment 31, we disagree with ASA's statement that we
underestimated the number of U.S. sugar producers. As discussed in comment
28, we also disagree with ASA's assertion that the corn sweetener industry
continues to benefit from the sugar program.

25. We disagree with ASA that the benefits received by producers are derived
"theoretical" benefits. Gains that result from higher prices paid by sugar
users--rather than from direct government payments to producers--are
nonetheless real gains in income to producers and are a direct result of the
sugar program.

26. The production and acreage levels we report were the results of our
simulation of the impact of eliminating the sugar program, as estimated by
our model. We agree with ASA that over time, larger production responses are
likely to result than in the short run.

27. We revised page 25 of the draft report (now on page xx) to clarify that
our model analyzed the HFCS market to determine whether lower refined sugar
prices resulting from the program's elimination would cause a substitution
from HFCS to sugar in certain markets for sugar-containing products. We
found that eliminating the sugar program had a limited effect on the HFCS
market. See comment 28.

28. We disagree with ASA's assertion that refined sugar from the world
market would replace HFCS if the sugar program were eliminated. While HFCS
producers benefited from the sugar program in the 1980s and early 1990s, we
believe that the domestic HFCS and sugar markets have since been decoupled
because of domestic price, cost, and market conditions. The president of the
U.S. Corn Refiners' Association has also expressed this view. As shown in
table 4 in appendix I, the weighted average wholesale price of HFCS in the
United States dropped from 22 cents per pound (dry weight) in 1992 to 12.4
cents per pound (dry weight) in 1998, while the U.S. wholesale refined beet
sugar price was above 25 cents per pound throughout this period. In
particular, for 1998, the higher-priced HFCS-55 (13.43 cents per pound) was
below our model's estimated world refined sugar price (14.12 cents per
pound) if the sugar program were eliminated. Furthermore, the likelihood of
substitution between sugar and HFCS is more limited than in prior years
because technological advances have improved the HFCS product and created
more specialized sweetener markets. See also comments 24 and 27.

29. Although the 1996 Farm Act made some changes to the sugar program, we
disagree that it substantially pared back the U.S. sugar program because the
basic structure of USDA's loan program and the TRQ remained unchanged. In
addition, we disagree with ASA's implication that a fall in the U.S. price
for refined sugar should necessarily result in a lower estimate of the cost
of the program. Because the cost of the program depends on the difference
between domestic and world sugar prices, changes in the world price are also
important in determining the program's cost.

30. We disagree with ASA's argument for using a world cost-of-production
estimate as a benchmark. Because of a scarcity of data, estimating the
international cost of production would require numerous assumptions about
other countries' costs of production as well as the effects of their
policies to support producers. See also comments 4 and 7.

31. We used the 1997 census of agriculture, conducted by USDA's National
Agricultural Statistics Service, to identify the number of sugarcane farms
and sugar beet farms for table 3 in appendix 1. USDA's census of agriculture
provided a consistent and rigorous methodology for estimating the number of
sugarcane and sugar beet farms. However, as ASA mentioned, table 3 also
notes that the census of agriculture's estimate of 7,102 beet farms varied
substantially from the USDA Farm Service Agency's estimate of 11,800 sugar
beet farms in 1997. This difference reflects differences in the definition
of what constitutes a farm. In any case, the draft and final reports include
both estimates.

32. We disagree with ASA's conclusion. See comments 1 and 3 on our
methodology and objectivity.

Comments From the U.S. Cane Sugar Refiners' Association

The following are GAO's comments on the letter of the U.S. Cane Sugar
Refiners' Association to our draft report.

1. We revised the report to address the concern of the U.S. Cane Sugar
Refiners' Association.

2. We revised table 1 to note that the basis point for the world raw sugar
contract price is the Caribbean. According to a sugar market expert, the
cost to transport raw sugar from the Caribbean to the United States is about
1.5 cents per pound. We also refined our model to more fully account for
certain transportation costs in our welfare analysis.

3. We revised the report to clarify that USDA's census of agriculture and
the Farm Service Agency used different definitions of a "farm".

GAO Contacts and Staff Acknowledgments

Robert E. Robertson (202) 512-5138

Richard Cheston (202) 512-5138

In addition to those named above, Jay Cherlow and Barbara J. El-Osta made
key contributions to this report.

(150155)

Table 1: U.S. and World Prices for Raw and Refined Sugar,
1985-98 14

Table 2: Acres of Sugarcane and Sugar Beets Harvested, by State,
Crop Years 1995-99 16

Table 3: Cane Sugar and Beet Sugar Production and USDA Loans,
by State, Crop Years 1997 and 1999 18

Table 4: U.S. Prices for HFCS, 1985-98 19

Table 5: Estimated Economic Gains and Losses Resulting From
the Sugar Program, 1996 and 1998 21

Table 6: Estimated Distribution Among User Groups of Benefits of Eliminating
the Sugar Program Under Different Pass-Through Assumptions, 1996 and 1998 23

Table 7: Estimated Effect of Eliminating the Sugar Program on
Prices and Production 25

Table 8: Estimated Longer Term Economic Gains and Losses if
the Sugar Program Were Eliminated 26

Figure 1: U.S. Sugar Production (in thousands of short tons),
Fiscal Years 1990-2000 15

Figure 2: Effects of Removing the TRQ and USDA's Loan Program
on U.S. Prices and Quantities of Raw Sugar 30

Figure 3: The TRQ's Effect on Imports 51
  

1. Sugar Program: Changing Domestic and International Conditions Require
Program Changes (GAO/RCED-93-84 , Apr. 16, 1993).

2. To estimate the effect of no sugar program, we simulated the elimination
of USDA's loan program for sugar processors and the tariff-rate quota for
both raw sugar and refined sugar imports. (The tariff-rate quota is the
amount of sugar that can be imported during a fiscal year at a low tariff
rate.) The most current year for which sweetener price and quantity data are
available is 1998.

3. All estimates are in 1999 dollars. See app. II for our assumptions about
short-run and longer term supply elasticities.

4. Producers include both growers and processors. According to USDA's 1997
census of agriculture, 973 farms grew sugarcane and 7,097 farms grew sugar
beets. USDA's Farm Service Agency, using a different definition of a farm,
estimated that 11,800 farms grew sugar beets in 1997.

5. These elasticities measured the extent to which the quantity of sugar
supplied would respond to changes in price.

6. Sugar comes from sugarcane and sugar beet plants, both of which are
processed to extract the sugar. Sugarcane typically is milled into raw cane
sugar and is then sent to a refinery, which further processes it into
refined sugar for consumption. Beet sugar is transformed directly into
refined sugar by beet processors. Almost all sugar imported into the United
States is raw cane sugar.

7. Sugar processors are required to pay growers a government-specified
minimum price, equivalent to about 60 percent of the loan.

8. USDA also administers a TRQ for imported refined sugar of at least 30,900
tons each year. The refined sugar TRQ was set at 66,000 tons for fiscal year
2000.

9. Under the North American Free Trade Agreement, the tariff for Mexican
sugar imported outside the TRQ will gradually be reduced from 15.6 cents per
pound in 1994 to zero cents per pound in 2008. The high-tier tariff for
Mexican sugar is 12.1 cents per pound in 2000.

10. The Agriculture Appropriations Act for fiscal year 2000 suspended the
sugar marketing assessment for fiscal years 2000 and 2001.

11. The basis point for world raw sugar prices is a Caribbean port,
including Brazil, while the basis for domestic prices is New York City. To
compare these prices, we added 1.5 cents per pound to the world price to
cover transportation costs from the Caribbean.

12. All ton measurements in this report are short tons--a short ton equals
2,000 pounds.

13. USDA also announced plans to allow 1.25 million tons of raw sugar to be
imported during fiscal year 2000, sufficient to fulfill its trade agreement
obligations.

14. World raw sugar prices similarly recovered to about 6.53 cents per pound
as of May 1, 2000, after dropping below 5 cents per pound in Feb. 2000.

15. See Sugar Program: Changing the Method for Setting Import Quotas Could
Reduce Cost to Users (GAO/RCED-99-209 , July 26, 1999).

16. Welfare analysis uses quantitative measures to analyze how an
intervention in a market redistributes economic rents among various groups
in the economy. In the context of agricultural markets, income
redistribution to all or some farmers and processors could be measured as
their gains in economic rents at the expense of losses to consumers' or
taxpayers' incomes.

17. Our primary estimates of the welfare effects of eliminating the program
are based on supply elasticities of sugar that can be interpreted as
short-run elasticities. We also estimated these effects using higher
elasticities that can be interpreted as longer-term elasticities. These
effects would show the welfare losses and gains after more time had passed
for the economy to adjust to the lower sugar prices that would result from
eliminating the program.

18. Edward A. Evans and Carlton G. Davis, "Dynamics of the United States
High Fructose Corn Sweetener Market," presented at the Conference on
Sweetener Markets in the 21st Century, Miami, Fla. (Nov. 1999).

19. Our different pass-through assumptions also resulted in slightly
different estimates of the total gains to sweetener users if the sugar
program were eliminated, primarily because consumers would be expected to
increase their sweetener consumption somewhat if manufacturers of
sugar-containing foods lowered their prices. We estimated that the total
welfare gains from eliminating the sugar program would have been about $1.4
billion in 1996 and $1.8 billion in 1998 if only sugarcane refiners had
passed cost reductions through to consumers.

20. The SIC system classifies each industrial establishment according to its
primary activity. This system, last revised in 1987, was established to
promote more uniform and comparable data.

21. The LINQUAD is a functional form within the incomplete demand system
approach that provides a practical model for estimation that reflects
theoretically sound preference ordering. In particular, the LINQUAD
quasi-expenditure function produces demand functions that are linear in
deflated income and linear and quadratic in deflated prices.

22. EV is the amount of money that, when paid to the consumer, allows the
consumer to achieve the same level of utility before the change that the
consumer would enjoy with the economic change. EV represents the minimum
amount that a consumer would require to willingly forgo the change.

23. Our presentation follows Moschini: (Agricultural Economics (5), 1991),
Meike and Lariviere (International Trade Research Consortium, Mar. 1999),
and Morath and Sheldon (MIMEO, Feb. 1999.)

24. Toll refiners export refined sugar processed from imported raw sugar.

25. We have the following definitions: λ=bs +bg (Pg /Pf ) and μ=bb
/Pf . If the price of a competing crop changes, then λ will change as
well.

26. C.M. Rendleman and T.W. Hertel. "Do Corn Farmers Have Too Much Faith in
the Sugar Program?" Journal of Agricultural and Resource Economics Vol. 18
(1993), pp. 86-95.

27. In the nonhomothetic transformation of a Cobb-Douglas functional form,
the cost shares of inputs are not held constant.

28. We assumed that the derived demand for sugar, as well as HFCS, is price
inelastic and small. References to the price elasticities of demand for
sugar and other agricultural inputs include Lopez and Sepulveda
(Northeastern Journal of Agricultural and Resource Economics, Oct. 1985);
Shui, Beghin, and Wohlgenant (American Journal of Agricultural Economics,
Aug. 1993), Devadoss and Kropf (Agricultural Economics, Jan. 1996), and
Wohlgenaut (FAO, 1999).

29. Conjectural variation is a parameter that measures how firms with market
power recognize their mutual interdependence. Specifically, it is the
percentage change in all other firms' output that a firm expects in response
to a 1 percent change in its own output. This variable can also be defined
in terms of price behavior.

30. S. McCorriston, C.W. Morgan, and A.J. Rayner, "Processing Technology,
Market Power, and Price Transmission," Journal of Agricultural Economics,
Vol. 49, No. 2 (Spring 1998),
pp. 185-201.

31. The parameter θ can be defined as the firm's conjectural variation
elasticity divided by its own price elasticity of demand (see Bhuyan and
Lopez (American Journal of Agricultural Economics, Aug. 1997)).

32. See Bhuyan and Lopez (American Journal of Agricultural Economics, Aug.
1997) and Morrison (NBER Working Paper #3355, 1990).

33. Jeffrey LaFrance, "The LINQUAD Almost Complete Demand Model,"
Unpublished manuscript, Department of Agricultural and Resource Economics,
University of California, Berkeley.

34. Gates Kennedy Agnew, LINQUAD: An Incomplete Demand System Approach to
Demand Estimation and Exact Welfare Measures. Master's Thesis, Agricultural
and Resource Economics, University of Arizona (1998).

35. Stephen Devadoss and Jurgen Kropf. "Impacts of Trade Liberalization
Under the Uruguay Round on the World Sugar Market," Agricultural Economics,
Vol. 15 (1996), pp. 83-96.

36. Sanjib Bhuyan and Rigoberto A. Lopez, "Oligopoly Power in the Food and
Tobacco Industries," American Journal of Agricultural Economics (Aug. 1997),
pp. 1035-1043.

37. Michael K. Wohlgenant, Effects of Trade Liberalization on the World
Sugar Market, prepared for the Food and Agriculture Organization of the
United Nations, Rome (1999).

38. Our analysis replaced the U.S. sugar module of the CARD model with the
model described in app. II.

39. S. Devadoss and J. Kropf. Cited in app. II.

40. A. Hafi, P. Connell, and R. Sturgiss, "Market Potential for Refined
Sugar Exports From Australia," Australian Bureau of Agricultural and
Resource Economics Research, Report 93.17, Canberra, (1993).

41. M.K. Wohlgenant. Cited in app. II.

42. G. Moschini, "Economic Issues in Tariffication: An Overview,"
Agricultural Economics, Vol. 5 (1991), pp. 101-120.

43. Karl Meilke and Sylvain Lariviere, "The Problems and Pitfalls in
Modeling International Dairy Trade Liberalization," Working Paper #99-3,
International Agricultural Trade Research Consortium (Mar. 1999).

44. We employed the "large country" assumption in our analysis of excess
supply under a TRQ, in the sense that changes in U.S. imports would be
sufficient to affect international prices. Thus, the excess supply curve is
upward sloping at import volumes in ranges other than where the TRQ is
binding.

45. For example, see La France (American Journal of Agricultural Economics,
1991), La France and de Gorter (American Journal of Agricultural Economics,
1985), Thurman (American Journal of Agricultural Economics, 1991), Rucker
and Thurman (Journal of Law and Economics, 1990).
*** End of document. ***