Livestock Market Reporting: USDA Has Taken Some Steps to Ensure  
Quality, but Additional Efforts Are Needed (09-DEC-05,		 
GAO-06-202).							 
                                                                 
Livestock producers, with gross income of $63 billion in 2004,	 
depend on USDA's daily, weekly, and monthly livestock market news
reports. These reports provide them and others in the industry	 
with livestock and meat prices and volumes, which are helpful as 
they negotiate sales of cattle, hogs, lamb and meat products.	 
Packers also use the average prices in these reports as a basis  
for paying some producers with whom the packers have contracts.  
In 1999, the Livestock Mandatory Reporting Act was passed to	 
substantially increase the volume of industry sales transactions 
covered by USDA's market news reports and thereby encourage	 
competition in the industry. In the context of ongoing		 
discussions about the renewal of this act, GAO reviewed (1)	 
USDA's efforts to ensure the quality of its livestock market news
reports and (2) the coordination between two USDA agencies that  
are responsible for promoting competition in livestock markets.  
-------------------------Indexing Terms------------------------- 
REPORTNUM:   GAO-06-202 					        
    ACCNO:   A42918						        
  TITLE:     Livestock Market Reporting: USDA Has Taken Some Steps to 
Ensure Quality, but Additional Efforts Are Needed		 
     DATE:   12/09/2005 
  SUBJECT:   Cattle						 
	     Competition					 
	     Prices and pricing 				 
	     Reporting requirements				 
	     Sales						 
	     Swine						 
	     Government information dissemination		 
	     Government publications				 
	     Food industry					 
	     Livestock						 
	     Transparency					 

******************************************************************
** This file contains an ASCII representation of the text of a  **
** GAO Product.                                                 **
**                                                              **
** No attempt has been made to display graphic images, although **
** figure captions are reproduced.  Tables are included, but    **
** may not resemble those in the printed version.               **
**                                                              **
** Please see the PDF (Portable Document Format) file, when     **
** available, for a complete electronic file of the printed     **
** document's contents.                                         **
**                                                              **
******************************************************************
GAO-06-202

Report to Congressional Requesters

December 2005

LIVESTOCK MARKET REPORTING

USDA Has Taken Some Steps to Ensure Quality, but Additional Efforts Are
Needed

Contents

Tables

Figures

December 9, 2005Letter

The Honorable Tom Harkin Ranking Democratic Member Committee on
Agriculture, Nutrition, and Forestry United States Senate

The Honorable Charles Grassley United States Senate

Livestock producers' sales of cattle, hogs, and sheep to meatpackers are a
significant component of the U.S. agricultural industry. In 2004,
producers' gross income from these livestock totaled about $63 billion.
For producers to be well prepared to negotiate a fair price for their
livestock, they need to be able to track changes in livestock prices. For
many years, producers have relied on the U.S. Department of Agriculture's
(USDA) livestock market news reports, which until 2001, were based on
livestock sales information obtained voluntarily from producers, packers,
feedlot operators, and others. However, over the past several decades,
producers and packers increasingly have made livestock sales through
contracts that were not covered in USDA's livestock market news reports.
By 1999, about 35 percent of cattle and 60 percent of hogs were sold
through such contracts. Because of these gaps in the coverage of USDA's
livestock market news reports, many producers believed that they could no
longer effectively assess livestock prices, negotiate with packers, or
obtain a competitive price when they sold livestock. Some producers have
been concerned that there have been "sweetheart" deals between packers and
certain cattle feeders, that unreported livestock sales among packers have
lowered livestock prices, and that reduced competition in the packing
industry has adversely affected their livelihood. In 1999, Congress
enacted the Livestock Mandatory Reporting Act to provide livestock market
information that could be readily understood by producers, packers, and
other market participants and to encourage competition in the marketplace
for livestock and livestock products. Under the act and USDA regulations,
packers with large packing plants1 were required to report to USDA all of
their purchases of cattle, lamb, and hogs, as well as their sales of beef
and lamb meat. Based on packers' reports of their transactions, USDA
publishes about 100 various livestock market news reports, which are
available over the Internet.2

In December 2004, when it was due to terminate, the Livestock Mandatory
Reporting Act was extended until September 30, 2005. As of September 30,
2005, USDA has continued the program as a voluntary, instead of mandatory,
reporting program. Congress, as of December 2005, is considering proposed
legislation to extend the mandatory program.3

USDA's Agricultural Marketing Service (AMS) administered the Livestock
Mandatory Reporting Act and produced the livestock mandatory market news
reports, among other market reports.4 Specifically, about 20 AMS market
news reporters located in Des Moines, Iowa and St. Joseph, Missouri,
reviewed the packers' livestock purchases and meat sales transaction data
each day and completed AMS's reports. In addition, AMS staff visited
packers' facilities several times a year to audit the accuracy of
information that packers reported to AMS. Another USDA agency-the Grain
Inspection, Packers and Stockyards Administration (GIPSA)-has a related
responsibility. GIPSA is responsible, under the Packers and Stockyards
Act, for addressing unfair and anti-competitive practices in the marketing
of livestock. Among other things, GIPSA monitors competition in livestock
markets and investigates complaints of unfair and anti-competitive
practices under the act.

In the context of the ongoing congressional discussions about the
Livestock Mandatory Reporting Act and the quality of information in the
reports it provides to producers, we reviewed (1) AMS efforts to ensure
the quality of its livestock mandatory market news reports, and (2) the
extent to which AMS and GIPSA coordinate their efforts to promote
competition in livestock markets.

To review AMS's efforts to ensure the quality of its livestock mandatory
market news reports, among other things, we observed AMS's Market News
Branch (Market News) reporters as they checked packers' transaction data
prior to publication and prepared livestock mandatory market news reports
for cattle, hogs, lamb, beef, and lamb meat. We reviewed AMS reporters'
instructions for preparing livestock mandatory market news reports and
assessed how AMS reporters used their instructions for cattle, hogs, lamb,
beef, and lamb meat. We evaluated AMS's efforts to audit packers' records
to ensure that packers were accurately reporting their transactions to
AMS. In addition, we reviewed AMS's computer system documentation and
conducted some limited tests of its performance. We found AMS
computer-processed data to be sufficiently reliable for our purposes. To
determine the extent of coordination between GIPSA and AMS, we reviewed
their legislative authority, identified activities and investigations
involving both agencies, reviewed investigation case file information, and
discussed them with GIPSA and AMS officials. Our review focused on the
livestock mandatory reporting program that terminated on September 30,
2005. Appendix I provides additional details on our objectives, scope, and
methodology. We conducted our review between February and November 2005 in
accordance with generally accepted government auditing standards.

Results in Brief

AMS took several steps for producing quality livestock mandatory market
news reports. Among other things, AMS developed a near real-time Web-based
reporting system with automated and manual screening of packers'
transaction data, and established an audit surveillance program to ensure
that packers report accurately. While important steps have been taken, AMS
has not yet fully assured the transparency and accuracy of these reports.
Concerning transparency, AMS has not informed readers of livestock
mandatory market news reports that it regularly excluded certain
transactions in an effort to present prevailing market conditions. Our
analysis of AMS market news data shows, for example, that from April
through June 2005, AMS reporters excluded about 9 percent of the cattle
transactions that packers reported to AMS. AMS officials explained that,
in general, AMS excluded transactions that were outside the prevailing
market price ranges to avoid reporting price ranges that would be too
broad to be useful to market participants. They also said that AMS market
news reports were intended to convey overall market conditions rather than
precise statistics. According to our interviews with some agricultural
economists, they and other readers of AMS market news reports were unaware
of the extent of this practice and the effect this practice has on the
prices that AMS reports. In addition, we found that AMS's instructions to
reporters for excluding transactions lacked clarity and precision, and
differed for cattle, hogs, lamb, beef, and lamb meat. AMS officials said
that as a result of the information that we brought to their attention,
they started improving AMS's instructions to reporters. AMS officials also
said they would consider providing market news report readers with
additional information about AMS's reporting practices.

Concerning accuracy, the quality of AMS reports depends on the extent to
which packers submit correct transaction information. However, AMS audits
have frequently identified instances when packers incorrectly reported
transactions. Of 844 AMS audits of packers, which were conducted over the
36 months ending in April 2005, 540-64 percent-identified instances when
packers did not report data to AMS correctly. AMS officials said that
packers' reporting errors were of concern, but that considering the
hundreds of thousands of pieces of transaction data that packers report
each day, the errors identified by AMS audits were proportionately few.
Nevertheless, a closer look at 86 AMS audits from June through September
2004 shows that AMS identified 46 instances when 22 packers submitted
incorrect transaction data that AMS classified as possibly affecting the
accuracy of AMS reports. Moreover, some packers did not promptly correct
the problems that AMS identified. Between 2002 and September 2005, AMS
sent 11 packers 21 warning letters because these packers delayed making
corrections in their reporting of transactions. Eight of these warning
letters were sent to six packers from January 2004 through September 2005;
six letters involved cattle and two involved hogs. AMS twice levied
$10,000 fines on packers, but suspended these fines provided the packers
went a year without additional violations. Despite this record, AMS
officials said that they have seen improvement in packers' reporting of
transactions over the past 4 years, and that they believed that most
packer transactions were accurately reported to AMS. However, AMS has not
developed a method for evaluating the overall accuracy of the transaction
data. To lend greater reliability to the reports, AMS officials said they
would consider (1) auditing a statistical sample of transactions as a
basis for assessing the overall quality of the transaction data, and (2)
further assessing their audit results to develop strategies for improving
packers' reporting of transactions.

Coordination between GIPSA and AMS has been limited, primarily due to the
legal authority within which each operates. AMS implemented and enforced
the Livestock Mandatory Reporting Act while GIPSA implements and enforces
the Packers and Stockyards Act. The Livestock Mandatory Reporting Act did
not provide authority for AMS to share confidential packer transaction
data within USDA unless the Secretary of Agriculture or the Attorney
General directed AMS to disclose the information for enforcement purposes.
GIPSA monitors cattle and hog markets by analyzing publicly available AMS
livestock reports. This approach to market monitoring has limitations
because it does not include company-specific transaction data that would
be useful for detecting anti-competitive behavior. AMS provided packers'
transaction data to GIPSA for two investigations after formal GIPSA
requests for those data. Both investigations were closed by GIPSA without
finding potential violations of the Packers and Stockyards Act.
Nevertheless, in one of those investigations, which GIPSA closed in 2005,
GIPSA found that a packer reported transactions to AMS that were not fully
supported by documentation. Since our review of that case indicates there
may be a further role for both agencies, we suggested, and GIPSA and AMS
officials agreed, that both should consider further investigation on the
propriety of certain low-price and other transactions. In November 2005,
an AMS official also told us that AMS started to obtain information from
the packer to address these transactions.

In anticipation of the extension of the Livestock Mandatory Reporting Act,
we are making a number of recommendations to the Secretary of Agriculture
to improve the transparency and accuracy of AMS livestock mandatory market
news reports.

Background

The Livestock Mandatory Reporting Act of 1999 amended the Agricultural
Marketing Act of 1946. The act established a livestock marketing
information program to (1) provide producers, packers and other industry
participants with market information that can be readily understood; (2)
improve USDA price and supply reporting services; and (3) encourage more
competition in these markets. Under the act, packers were required to
report livestock market information that had previously been voluntarily
reported5 and new information not previously reported to the public-such
as information about contract livestock purchases. Under the voluntary
program, USDA employees, referred to as reporters, gathered information
daily by talking directly with producers, packers, feedlot operators,
retailers, and other industry participants; by attending public livestock
auctions, visiting feedlots and packing plants; and taking other actions.
Under the Livestock Mandatory Reporting Act, packers were instead required
to report on their cattle and hog purchases, and their sales of beef.6 The
act also authorized USDA to require that packers report on lambs.

USDA implemented the Livestock Mandatory Reporting Act by establishing a
livestock mandatory reporting program to collect packers' marketing
information and disseminate it to the public through daily, weekly,
monthly, and annual reports. Packers were required to electronically
report hog purchases three times each day, cattle purchases twice each
day, lamb purchases once daily, domestic and export sales of beef cuts
twice daily, and sales of lamb carcasses and lamb cuts once daily. As of
June 2005, 116 packers and importers were required to provide information
under the Livestock Mandatory Reporting Act.

Two branches of USDA's AMS administered the livestock mandatory reporting
program-Market News and the Audit, Review, and Compliance Branch (ARC).
Market News was responsible for collecting and generating market news
reports from information supplied by packers. Market News reporters
gathered and reviewed this data, contacted packers to resolve any
questions they had, and prepared reports. Reporters were required to
ensure that they did not breach the confidentiality of packers by
providing information that would allow the public to identify an
individual packer. In addition to preparing reports, Market News personnel
interacted with any packers that AMS believed needed to make changes in
reporting to comply with the Livestock Mandatory Reporting Act.

To identify compliance problems, ARC personnel audited the transaction
data of packing plants three times a year. When ARC found packers that
were reporting incorrectly, ARC notified the Market News reporters, who
were responsible for notifying and following up with packers until the
packers reported correctly. The Secretary of Agriculture was authorized to
assess a civil penalty of up to $10,000 a day per violation on a packer
that violated the act.

AMS Striving for Quality, but Has Faced Some Challenges

AMS designed its livestock mandatory market news reporting program with
elements intended to ensure the quality of its news reports. USDA
officials, for example, developed a Web-based reporting system with
automated and manual screening of packer transaction data and established
an audit surveillance program to ensure packers reported accurately.
However, we found that while AMS had made progress, its livestock market
news program fell short of ensuring reliability because AMS reporting was
not fully transparent, and AMS audits of packers revealed some problems
with the quality of packers' transaction data.

Steps AMS Took to Ensure Quality

AMS developed a mandatory livestock market news reporting program
incorporating a number of features to ensure quality. More specifically,
AMS took the following steps to ensure the quality of its livestock
mandatory market news reports:

o AMS hired two contractors to assist in developing a rapid and reliable
reporting system: Computer & Hi-Tech Management, Inc. was hired to assess
the capability of the packing companies to provide electronic data; and
PEC Solutions developed the computer software processes upon which the
mandatory livestock reporting system is now based. AMS and PEC Solutions
developed a software system that allows packers to provide their
transaction data on web-based forms or to upload completed files into the
reporting system data base. PEC Solutions prepared an industry guide to
give packers instructions for correctly submitting transaction data.

o PEC Solutions used programmers who did not participate in developing the
systems to test the functioning of the system. AMS further tested the
system using simulated production data, because packers had not started
reporting actual data. As a further validation step, AMS staff manually
calculated data for several reports and compared that data with data
generated by the system.

o AMS established computer based data security controls and computerized
screening of packer transaction data to ensure it is being correctly
reported.

o AMS established an audit function to periodically test the accuracy of
transaction data that packers submit to AMS by visiting packer facilities,
checking documentation in support of reported transactions and testing the
completeness of packers' reports.

In addition, in May 2001, the Secretary of Agriculture appointed a top
level USDA team-the Livestock Mandatory Price Reporting Review Team7-to
review problems in its calculations of certain boxed-beef prices.8 In
addition to reviewing that problem and making related recommendations,
most of which AMS adopted, the team assessed the overall integrity and
accuracy of the program. This team found that for the most part, AMS had
succeeded in gathering and reporting accurate data in a timely fashion.
The team's major criticism was that AMS had not adequately tested its
system to ensure it was accurately calculating data that packers had
reported. Subsequently, AMS initiated further testing to ensure the
accuracy of its reports. The team also found that AMS's plan for audit
surveillance of packers was behind schedule due to difficulties in hiring
qualified auditors. At that time AMS had conducted audits at only 19 of
the 119 packer facilities it planned to reach. Since then, AMS has
overcome these problems and conducted over 1,100 audits at packers'
facilities.

Extent of Transactions Excluded from AMS Reports Is Not Transparent

The Livestock Mandatory Reporting Act was intended to provide producers
with readily understandable market information on a significant portion of
livestock industry transactions. The quality of this information is
especially important because livestock transactions negotiated each day
may be influenced by AMS reported prices, and some contracts between
packers and producers rely on the weighted average prices that AMS
reports. AMS was authorized to make reasonable adjustments in information
reported by packers to reflect price aberrations or other unusual or
unique occurrences that the Secretary determined would distort the
published information to the detriment of producers, packers, or other
market participants.9 In addition, AMS should have adhered to the Office
of Management and Budget and USDA guidelines for disseminating influential
statistical and financial information with a high degree of transparency
about the data sources and methods, while maintaining the confidentiality
of the underlying information.10 In addition, AMS has recognized the
usefulness of providing the public with information about the preparation
of its market reports.

We found that AMS reporters adjusted the transaction data that packers
report in an effort to report market conditions, but this practice has not
been made transparent. We observed that AMS reporters sometimes eliminated
small numbers of apparent erroneous transactions, as would be expected.
Significantly, however, we found that AMS reporters eliminated numerous
low- and some high-priced transactions that they believed did not reflect
market conditions, particularly when reporting on cattle. Our analysis
shows that from April through June 2005, when livestock prices were
declining somewhat, AMS reporters excluded about 9 percent of the cattle
transactions that packers had reported to AMS, about 3 percent of the
reported beef transactions, and 0.2 percent of the reported hog
transactions. Excluding small percentages of livestock or meat
transactions may have had a small effect on the range of prices that AMS
reported and a negligible effect on weighted average prices. However, as
the percent of transactions excluded increased, so too did the possibility
that AMS weighted average prices would be changed from what AMS would
otherwise report. Table 1 provides more details about the transactions
excluded during this period.

Table 1: Cattle, Hog, and Beef Transactions Excluded from AMS Reports,
April through June, 2005

                                        

                 Cattle           Hogs             Beef   
                   Number Percent          Number Percent         Number Percent 
  Transactions 
  Reported by     207,459                  97,419                831,622 
  packers                                                                
  Excluded by      18,286     8.8             153     0.2         22,436     2.7 
  AMS                                                                    
  Head count   
  Reported by   9,043,064              11,140,099                     --      -- 
  packers                                                                
  Excluded By     136,086     1.5          17,431     0.2             --      -- 
  AMS                                                                    
  Weight       
  (lbs.)       
  Reported by          --      --              --      --  4,096,490,216 
  packers                                                                
  Excluded by          --      --              --      --    111,567,704     2.7 
  AMS                                                                    

Source: GAO analysis of AMS data reported by packers on certain daily
reporting forms.

In addition, our analysis shows that from May through October 2003, when
cattle prices were rising and changing to greater extents, AMS reporters
excluded about 23 percent of cattle transactions packers reported to
AMS.11 Concerning hogs, during a period of rising prices between October
2003 and March 2004, we found that 0.1 percent of hog transactions were
excluded from AMS reports. Because AMS reports excluded significantly more
cattle transactions, we performed further analyses on them. Tables 2 and 3
show (1) information about the cattle transactions that AMS excluded from
certain livestock mandatory market news reports from May through October
2003, and (2) examples of 12 days from this period showing the effects of
the transactions that AMS excluded on the reported price ranges, and
weighted average prices. During the period, AMS reporters' decisions to
exclude transactions had some effect on the cattle data we analyzed in AMS
reports on about one third of the days and almost no effects on the
others.

Table 2: Cattle Transactions Excluded From AMS Reports, May through
October, 2003

                                        

                                                Cattle        
           Transactions                                Number         Percent 
Reported by packers                                399,471 
Excluded by AMS                                     90,998            22.8 
Head count                   
Reported by packers                             15,792,982 
Excluded by AMS                                  2,921,956            18.5 

Source: GAO analysis of all AMS cattle data reported by packers on the
Live Cattle Daily Report (Current Established Prices).

Table 3: Examples of Changes in Cattle Prices AMS Reported, May through
October, 2003

                                        

    Dollars per  
hundredweight 
                              GAO price      AMS                    Change in 
                       AMS  ranges with reported       GAO weighted  weighted 
Date           reported transactions weighted average price with   average 
                     price  excluded by  average       transactions     price 
                    ranges          AMS  price a    excluded by AMS  reported 
                                                                       by AMS 
May 5            $123 -  $122 - $126  $123.30            $123.25     $0.05 
                      $126                                          
May 16        124 - 125    105 - 125   124.63             123.12      1.50 
May 19        125 - 128    124 - 128   126.16             125.86      0.30 
June 12       121 - 125    113 - 125   123.37             123.32      0.05 
July 11        116.50 -    115 - 120   119.05             118.98      0.07 
                       120                                          
August 1      127 - 127 122.50 - 127   127.00             125.21      1.79 
August 11     125 - 128    123 - 128   126.28             125.65      0.63 
Sept. 10      135 - 145    131 - 145   141.63             141.37      0.26 
Sept. 29      140 - 141    138 - 141   140.25             138.84      1.41 
Oct. 14           170 - 150 - 182.75   176.30             175.83      0.47 
                    182.75                                          
Oct. 15       175 - 187    167 - 187   180.48             180.04      0.44 
Oct. 29       155 - 158    155 - 180   157.52             160.27    (2.75) 

Source: Selected dates from AMS's Five Area Daily Weighted Average Direct
Slaughter Cattle Report, Negotiated; and GAO analysis of AMS data for 35 -
65 percent choice steers, dressed weight as reported by packers on the
Live Cattle Daily Report (Current Established Prices).

aGAO replicated these AMS-reported average prices, ensuring that our
calculations using AMS data produced similar results. For our calculations
of average weighted price and price ranges, we included those transactions
that AMS had excluded for market reasons, such as high or low price, or
size of lots, but we did not include those transactions that AMS had
removed because they thought them to be base prices.

Further details of our analyses are discussed in appendix I and shown in
appendix II.

AMS guidance for its reporters on eliminating transactions is limited,
lacking clarity and precision. These instructions advise AMS reporters to
review transactions which packers have reported each day, and to eliminate
certain low- and high-priced transactions. AMS's varying instructions for
reporters are described in table 4.

Table 4: AMS Market News Reporters' Instructions

                                        

               o Narrow up price spreads, if possible, depending on how trade 
                took place the previous week. Keep price ranges within $5 of  
                      the bulk of the trade and narrower, if possible.        
                                                                              
               o Remove weights that are unusually high or low for an animal. 
                                                                              
     Cattle    o For cattle sold on formula, a market news reader should not  
               see the lowest or highest cattle prices packers' reported. The 
                 lower prices should be the lower end of marketable cattle    
                prices within explainable limits; and on the high side, the   
                   best price that would be reasonable and obtainable by      
                  marketing desirable type cattle. (GAO note: "explainable    
               limits," "reasonable," and "desirable type" are not defined.)  
               o Scan reports to determine whether prices fell in the         
               expected range and eliminate small trades that are outside the 
               expected price range for each item. Contact packers when large 
               quantities of beef fall outside expected price range.          
                                                                              
Beef        o The expected price range for a beef item is to be based on   
               historical and current price ranges for the item, current      
               market conditions for the item, and overall beef market        
               conditions.                                                    
                                                                              
               o Consider excluding trades when prices are not within close   
               proximity of another packer.                                   
Hogs        o If there are problems with prices for barrows, gilts, or     
               sows, contact the packer to verify the information.            
               o Review the data to determine if there are any high, low or   
Lamb (live) questionable prices that need to be removed.                   
                                                                              
               o Separate the outliers.                                       
               o Edit extremes for all price ranges in question.              
Lamb (meat)                                                                
               o Review the report and check any price ranges that appear out 
               of line. (GAO note: "out of line" is not defined.)             

Source: AMS guidance for reporters.

Senior AMS supervisors review reporters' decisions to eliminate
transactions, and AMS headquarters officials monitor the number of-and
reasons why-transactions are being excluded by reporters. AMS officials
explained, in general, their reviews and adjustments are intended to
exclude transactions that are outside the prevailing market price ranges,
and to avoid reporting ranges of prices that appear overly broad.
Furthermore, Market News officials explained that this process is
conducted because they believe that livestock market reports are intended
to convey overall market conditions rather than precise statistics. Also,
an AMS official noted that AMS Market News reporters mostly exclude
low-price transactions involving small quantities, because those
transactions often are lower quality animals or products. Concerning hogs,
AMS's reporters of hog transactions said that they were verbally
instructed to exclude few hog transactions by headquarters officials soon
after the start of the program. AMS headquarters officials said that these
verbal instructions were provided after one or more large packers
complained that it appeared AMS was excluding transactions because of
price alone.12

Given that AMS reporters' decisions to exclude transactions modified the
prices they reported, AMS has not well-explained this practice to readers
of AMS livestock market news. AMS's Web site does not address the subject,
and AMS livestock mandatory market news reports are unqualified. Some
agricultural economists who study the livestock market and other industry
experts we interviewed said that they were not aware of the extent of
adjustments that AMS made. An AMS official explained that AMS has not
previously provided public information on this process because it would be
difficult to capture the nuances of AMS's report preparation in a public
document. Nevertheless, AMS previously acknowledged that it may be useful
to provide information to the public about types of adjustments that it
makes to its livestock mandatory market news reports.13 AMS officials also
recognized that it would be desirable for AMS to improve its instructions
for reporters and disclose more about its reporting practices to livestock
market news report readers. Our review of AMS's database indicates that
further analyses could provide AMS with more information about the reasons
why reporters eliminate transactions, the consistency of reporting, as
well as the extent of changes in AMS's presentation of prices. AMS's
Livestock and Seed Program Deputy Administrator said that, as a result of
the information we brought to his attention, he had started to improve the
reporters' instructions.

Since AMS reports help provide the industry with signals about when,
where, and at what price to buy and sell livestock and meats, some
industry participants may have been guided to somewhat different decisions
on certain days if they had a greater understanding of AMS report content.
In addition, the lack of transparency over the content and preparation of
the livestock mandatory market reports may have also limited the
confidence that some readers place in AMS reports.

ARC Audits Identified Instances When Packers Reported to AMS Inaccurately

ARC regularly audited packers to provide assurance that the packers
reported all of their transactions accurately and in compliance with AMS's
regulations.14 The quality of AMS reports depends on packers submitting
correct transaction information. Once every 4 months, ARC auditors visited
each of the 116 packers' plants, or associated company headquarters, to
review livestock transaction data.  15 These audits usually included: (1)
a test of the completeness of the packer's reports, and (2) a detailed
review of a sample of transactions to determine that each transaction in
the sample was reported accurately and was supported by appropriate
documentation. ARC has conducted over 1,100 audits at packers' facilities
since 2001. Detailed information was available for 844 of these audits
conducted over the 36 months ending in April 2005.16 Table 5 contains
additional information about the content of ARC audits.

Table 5: Content of ARC Audits of Packers

                                        

                         Review of a full day of     Review of a sample of    
      Type of audit        packer transactions        packer transactions     
                                                        reported to AMS       
                        To ensure packers report  To ensure packers report    
AMS's audit goal     complete information to   accurate information to AMS 
                        AMS                       in a timely manner          
                        o Plant slaughter records o Plant slaughter records   
                                                                              
                        o Food Safety and         o Sales invoices            
                        Inspection Service                                    
                        slaughter records         o Bills of lading           
                                                                              
                        o Sales invoices          o Original contracts,       
Information reviewed                           agreements, and receipts    
                        o Bills of lading                                     
                                                  o Other records relating to 
                        o Documentation about why the purchase, sale,         
                        transactions not reported pricing, transportation,    
                        to AMS are not subject to delivery, weighing,         
                        the Livestock Mandatory   slaughter, or carcass       
                        Reporting Act             characteristics of          
                                                  livestock                   
                        o Three times a year for                              
                        plants whose records are  
                        reviewed at the plant     
Frequency of review                            Three times a year
                        o Once a year for plants  
                        whose records are         
                        reviewed at company       
                        headquarters              

Source: AMS records.

Of the 844 AMS audits for which data were available, 540-64
percent-identified one or more instances when it appeared that packers did
not meet AMS reporting standards. The other 304 audits, or about 36
percent, did not identify any such instances. AMS audits detected a wide
variety of packer reporting inaccuracies such as the omission of livestock
slaughtered, underreporting of purchases, delayed reporting of livestock
purchases and meat sales, price inaccuracies, and the misclassification of
transactions. While noting the frequency of AMS audit findings, AMS
officials commented that packers' reporting errors were of concern. AMS
officials also said that its audit results should be considered in the
context of the volume of transactions that AMS reports-compared to the
hundreds of thousands of pieces of transaction data that packers reported
daily, the errors identified by AMS audits were relatively few. However,
our review shows that AMS findings are based on audits of a small portion
of packers' transactions, and it is likely that there have also been
errors in packers' unaudited transactions. Furthermore, a closer look at
86 AMS audits completed from June through September 2004 shows that AMS
identified 46 instances when 22 packers submitted incorrect transaction
data that AMS classified as possibly affecting the accuracy of AMS
reports. Table 6 provides examples of AMS audit findings.17

Table 6: Examples of ARC Audit Findings

      A packer incorrectly reported the weekly premiums it paid for its hogs  
      on a live weight basis instead of a carcass weight basis. As a result,  
1. the weekly premium report was not accurate. An AMS reporter said this   
      finding was significant at the time because only a small number of      
      packers were reporting this weekly data and because the error could     
      occur repeatedly.                                                       
      A packer did not report the purchase of 5 lots of cattle, totaling 117  
      head. The packer explained that some of the cattle were unscheduled     
      purchases from local producers but could not provide an explanation for 
2. the remainder. An AMS reporter said that unscheduled cattle purchases   
      can present reporting difficulties for packers. AMS verified that the   
      packer was reporting all its cattle purchases about 2 months after the  
      audit was completed.                                                    
3. A packer did not report slaughter information on 780 hogs that had been 
      purchased on a live weight basis.                                       
      A packer did not report two large loads of beef totaling 83,000 pounds. 
4. An AMS reporter explained that this beef product-boneless beef          
      trimmings-was sold in high volume, so that this omission was relatively 
      insignificant provided that it rarely occurred.                         
      A packer was submitting its weekly data on slaughtered cattle early-on  
5. Thursday or Friday-instead of the following Monday as required, thereby 
      not reporting the data in a manner consistent with other packers.       
6. A packer could not provide supporting documentation to verify the time  
      that the price was established for its hog purchases.                   

Source: GAO review of ARC audit reports and interviews with AMS staff.

AMS officials said many ARC audit findings were minor and usually had
little effect, if any, on the accuracy of AMS reports. In addition, they
also said that since 2001, packers had clearly improved their reporting of
transactions. AMS officials said that because of the overall improvement
in packers' reporting of transactions, they reduced the frequency of
audits at each packer from four to three times a year. Our review provides
some support for AMS officials' view that packers were reporting better
than at the outset of the program. From May 2002 through April 2005, the
number of AMS audits with findings as a percent of total audits decreased
each year, from 76 percent in 2002 to 55 percent in 2005. In addition, the
average number of audit findings per audit decreased from 1.8 to 1.4 over
that period. Moreover, in the first quarter of 2005, AMS audits did not
identify any problems that rose to its highest level of concern.18
Nevertheless, AMS classified 22 percent of the problems it identified in
the first quarter of 2005 as possibly having some adverse effect on the
accuracy of its reports. In addition, follow-up was sometimes lengthy on
problems ARC auditors identified. Our analysis of follow-up efforts by AMS
on the 86 audits it conducted between June through September 2004, showed
that, on average, about 85 days19 elapsed between the date of an AMS audit
and the date AMS recorded that the packer had made the needed
corrections.20

AMS reporters frequently contacted packers to convey information about the
correct way for packers to report. Their outreach was prompted both by
audit findings and by reporters' reviews of the packers' data. When
recurring reporting problems arose, headquarters officials issued internal
guidance to clarify proper reporting procedures for both auditors and
reporters. On at least two occasions, AMS reporters provided information
from this internal guidance to packers to clarify proper reporting
procedures.21

However, some packers, including three of the largest packers, did not
promptly correct reporting problems that AMS identified. Since 2002, AMS
sent 11 packers 21 letters to call to the packers' attention apparent
delays in correcting reporting issues and warning the packers that
penalties might be applied should there be further delays in addressing
these issues. Of these, AMS sent 8 letters to 6 packers between January
2004 and September 2005, with 6 letters involving cattle and 2 involving
hogs. In addition, twice AMS levied fines on packers of $10,000, although
these fines were suspended provided these packers went a year without
additional violations of the Livestock Mandatory Reporting Act. As of
September 2005, AMS had continuing issues with 2 of 11 the packers that
received AMS warning letters. Appendix III contains additional information
on the issues leading to AMS warning letters to packers.

While AMS audit reports identified many problems in packers' reporting of
transactions, there are two reasons why the reports do not provide a clear
basis for assessing the overall accuracy of packers' data which underlie
AMS livestock mandatory market news reports. First, AMS did not select
transactions for audit in a manner that would enable AMS to project the
overall accuracy of packers' transaction data. Second, AMS did not develop
analyses that demonstrate the overall accuracy of information in its
reports. We explored two approaches with AMS officials to (1) obtain
better indications of the overall accuracy of packers' transaction data,
and (2) better direct future AMS audits.

o First, AMS audits did not provide a basis for projecting the overall
accuracy of packers' transaction data. Another approach, in which AMS
would periodically audit a statistical sample of transactions, might
provide a basis for projecting the overall accuracy of the transactions.

o Second, AMS could analyze its audit results, focusing on findings of
consequence and its follow-up efforts to address those findings. Such
analyses could be useful for identifying the relative frequency of
concerns with packers' transaction data, the types of recurring errors,
the timeliness and consistency of auditor and market news follow-up on
packer's actions to address reporting issues, and the overall
effectiveness of AMS efforts to quickly resolve reporting issues.

AMS officials indicated that these suggestions appeared to be reasonable
and that they would consider taking both steps.

Packers Sometimes Reported to AMS After Deadlines

AMS data show that from April through June 2005, 4 percent, 5 percent and
7 percent of selected cattle, beef and hog data, respectively, were
received from packers by AMS after the deadlines set by the Livestock
Mandatory Reporting Act. Nevertheless, AMS officials said that while some
packers missed the reporting deadlines, most usually submitted their
transaction data within minutes thereafter-giving AMS reporters enough
time to include almost all transaction data in market news reports. In
addition, AMS officials said that if some reporting deadlines and
publication times set in the Livestock Mandatory Reporting Act were
changed, this would help packers working on the west coast meet the
reporting schedule and help AMS meet changing market conditions.

GIPSA and AMS Coordination Has Been Limited

GIPSA and AMS coordination has been limited, primarily due to the legal
authority within which each operates.

o AMS implemented and enforced the Livestock Mandatory Reporting Act.
While the Livestock Mandatory Reporting Act called for the establishment
of a mandatory reporting program, it required information be made
available to the public in a manner that ensured the confidentiality of
the identity of persons and proprietary business information. Such
information could not be disclosed except (1) to USDA agents or employees
in the course of their duties under the Livestock Mandatory Reporting Act,
(2) as directed by the Secretary or the Attorney General for enforcement
purposes, or (3) by a court. AMS officials said that they have shared
packer transaction data with GIPSA when requested for specific
investigations.

o GIPSA implements and enforces the Packers and Stockyards Act. GIPSA
monitors livestock markets and investigates when it has reason to believe
there have been violations of the act.22

Since 1999 when the Livestock Mandatory Reporting Act was adopted, there
have been two cases where GIPSA formally requested access to a packer's
transaction data from AMS for specific investigations. AMS provided access
as GIPSA requested. One investigation involved hogs, and the other, lamb.
In one case, opened in October 2002, GIPSA investigated whether a packer
was manipulating reported prices in AMS's livestock mandatory reporting
program to reduce its procurement costs. GIPSA did not identify a
violation of the Packers and Stockyards Act, and closed this investigation
in 2005. However, GIPSA identified instances in which the packer's reports
of negotiated livestock purchases met the documentation standards of the
Packers and Stockyards Act, but may not have met the standards of the
Livestock Mandatory Reporting Act.23 In September 2005, GIPSA officials
briefed AMS officials on their investigation, and suggested that AMS
consider whether the packer was complying with the Livestock Mandatory
Reporting Act. In response to our further questions about this case,
officials of AMS and GIPSA said that they would consider additional
inquiry or investigation under both statutes to determine if there have
been repeated transactions reported to AMS for which the packer lacks
certain documentation. In the second case, GIPSA investigated the
possibility that a packer paid less for livestock as a result of providing
undue preference to a select group of producers. GIPSA initiated this case
in May 2002 and closed this case in September 2005.

GIPSA officials said that individual packer transaction data held by AMS
would be useful for monitoring competitive behavior in livestock markets.
However, because GIPSA could not obtain that confidential information
unless the Attorney General or the Secretary directed disclosure of the
information for enforcement purposes, GIPSA is making due with the
publicly available AMS livestock market report data. This monitoring
effort is limited because AMS reports do not include the company-specific
transaction data that might reveal anti-competitive behavior. More
specifically, GIPSA uses publicly available AMS report data in cattle and
hog price monitoring programs to forecast market prices for comparison
with actual prices. If there are notable differences, GIPSA officials
attempt to assess whether economic conditions could be responsible. Should
GIPSA find that a difference was not readily explained by economic
conditions, then GIPSA would further investigate to determine if
anti-competitive behavior of individual firms were involved. At such a
point, GIPSA may request that AMS provide company specific livestock
transaction data for GIPSA's investigation. GIPSA officials said that
while this monitoring effort is less informative than one that would rely
on company specific transaction data, their monitoring programs are
relatively new and they have not identified better alternatives at this
point.

Conclusions

AMS has not achieved the level of transparency needed for establishing the
reliability of its livestock market news reports-a level that would more
fully disclose to market participants and observers its practices in
reviewing packers' transactions, and the effects on AMS reports. Without
further disclosure of its reporting practices, market participants are
less informed than they should be about (1) AMS reporters' reviews, (2)
AMS decisions on presenting prevailing prices, and (3) the results of AMS
audits of packers' transactions. Also, the lack of precision and clarity
in AMS's varying instructions for its reporters has led to inconsistent
reporting approaches, which could adversely affect readers' confidence in
AMS reports. AMS market news readers should have information that enables
them to understand AMS's approach to reporting prices, and to have
confidence that the approaches are based on sound statistical, economic,
and reporting guidance. In addition, the problems which AMS audits
identified in packers' transaction information warrant continued vigilance
if the mandatory reporting program is renewed. Unless AMS takes some
additional steps, it will not have information to (1) assess the overall
accuracy of packers' transaction data, (2) focus its audit efforts on
recurring significant problems, and (3) ensure that prompt and consistent
action on audit findings is being taken. Concerning the GIPSA
investigation in which GIPSA raised questions about a packer's
documentation of its transactions, unless AMS and GIPSA complete further
investigative work, neither agency can have assurance of the accuracy and
propriety of the packers' transactions.

Recommendations

Should Congress extend the Livestock Mandatory Reporting Act, we recommend
that the Secretary of Agriculture direct the Administrator, Agricultural
Marketing Service to:

o Increase transparency by (1) reporting to market news readers on its
reporters' instructions for making reporting decisions that reflect
prevailing market conditions, (2) periodically reporting on the effects of
reporters' decisions on AMS reported prices, and (3) reporting the results
of its audit efforts.

o Clarify AMS reporter's instructions to make them more specific and
consistent by (1) consulting with packers, producers, agricultural
economists, and other interested stakeholders, and (2) undertaking
revisions that consider economic analyses of past reporting trends,
livestock and meat market variations, and federal statistical and
information reporting guidance.

o Develop information about the overall accuracy of packers' transaction
data by auditing a statistical sample of packers' transactions.

o Further develop AMS audit strategies to identify recurring significant
problems.

o Address the timeliness and consistency of AMS reporters' efforts to
follow-up on audit findings.

We also recommend that the Secretary of Agriculture direct the
Administrators of the Agricultural Marketing Service and the Grain,
Inspection, and Packers and Stockyards Administration to further
investigate the reporting practices of one packer's low-price purchases of
livestock.

Agency Comments and Our Evaluation

We provided USDA with a draft of this report for review and comment. In a
memorandum dated November 18, 2005, we received formal comments from
USDA's Acting Under Secretary for Marketing and Regulatory Programs. These
comments are reprinted in appendix IV. We also received oral technical
comments from AMS and GIPSA officials, which we incorporated into the
report as appropriate.

USDA generally agreed with our findings and recommendations, and discussed
the actions it has taken, is taking, or plans to take to address our
recommendations. Among other things, USDA stated that AMS would (1)
prepare publicly available reports on the volume of transactions excluded
by reporters and their effect on reported prices, and take steps to
increase public awareness of reporting methods and processes; (2) clarify
AMS reporters' instructions while following federal and departmental
statistical and information reporting guidance; (3) post quarterly audit
information to its website and identify additional audit information to
add in the future; (4) develop auditing methods to allow conclusions to be
drawn about overall data accuracy; (5) review its auditing methods to
increase the overall effectiveness of the compliance program; and (6)
conduct further inquiry into the issues raised during one of GIPSA's
investigations.

Concerning the transactions that AMS excluded from its market news
reports, USDA agreed that 22.8 percent of cattle transactions were
excluded from May to October 2003. USDA added that AMS reporters excluded
some transactions during that period because its computer system could not
differentiate between the base and net prices for certain cattle sales.
Our review indicates that AMS exclusions for that reason were part of the
story. More specifically, AMS reporters' log entries showed that of the
transactions AMS excluded from May to October 2003, about 24 percent were
excluded for reasons relating to base prices, while about 34 percent of
the transactions were excluded to narrow the range of prices that AMS
reported, and the remainder were excluded for a variety of other reasons
such as small head count, small lots, low weight, mixed lots, or grade of
cattle. In addition, AMS suggested that its programming change to
differentiate base and net prices led to fewer exclusions-8.8
percent--during the April through June 2005 period. While we agree that is
part of the explanation, we believe, if the livestock mandatory program is
renewed, that AMS needs to focus on the bases and methods for excluding
transactions, and especially the extent to which AMS will be excluding
transactions when prices are again rapidly changing, such as they did in
2003.

AMS also stated that care should be exercised when drawing conclusions
about packer compliance because packers' errors are relatively few
compared to the 500,000 data elements packers may have submitted on some
days. We believe insufficient information is available to assess the
overall quality of packer data. AMS audits only focused on a small portion
of the data submitted by packers, and it is likely that packers' unaudited
transactions contain errors as well. We continue to believe that packer
reporting problems that AMS identified warrant continued vigilance should
the program be renewed and recommend that AMS develop auditing methods to
allow conclusions to be drawn about overall accuracy of packer's data.

As agreed with your staffs, unless you publicly announce the contents of
this report earlier, we plan no further distribution until 30 days from
the report date. At that time, we will send copies of this report to
interested congressional committees; the Secretary of Agriculture; the
Under Secretary for Marketing and Regulatory Programs; the Administrators
of the Agricultural Marketing Service and the Grain Inspection, Packers
and Stockyards Administration; and other interested parties. We will also
make copies available to others on request. In addition, the report will
be available at no charge at GAO's Web site at http://www.gao.gov.

If you or your staff have any questions about this report, please contact
me at (202) 512-3841 or [email protected]. Contact points for our Offices
of Congressional Relations and Public Affairs may be found on the last
page of this report. Key contributors to this report are listed in
appendix V.

Robert A. Robinson Managing Director, Natural   Resources and Environment

Objectives, Scope, and Methodology Appendix I

Our objectives were to review the extent to which (1) the U.S. Department
of Agriculture's (USDA) Agricultural Marketing Service (AMS) takes
sufficient steps to ensure the quality of its livestock mandatory market
news reports, and (2) AMS and the Grain Inspection, Packers and Stockyards
Administration (GIPSA) coordinate efforts to encourage competition in
livestock markets.

To review AMS's steps to ensure the quality of its reports, we visited the
two Market News Branch (Market News) field offices in Des Moines, IA, and
St. Joseph, MO, and spoke with AMS reporters about their responsibilities
related to mandatory price reporting and observed them as they prepared
livestock mandatory reports for cattle, beef, hogs, lamb and lamb meat. To
test AMS's computerized reporting system, we obtained and analyzed
unpublished data from AMS's livestock mandatory reporting database for
beef, cattle, and swine. For this analysis, we used data reported by
packers through the Live Cattle Daily Report (Current Established Prices)
(LS-113), Swine Daily Report (LS-119), and Boxed Beef Daily Report
(LS-126) contained in AMS's livestock mandatory reporting database. We
reviewed USDA documents on the report preparation and data storage system
and analyzed the flow of data into and through the system. We performed
electronic testing and validation of system data developed for us from
data available in the AMS system. We found the data were sufficiently
reliable to support our analyses. We also replicated elements of certain
reports-the Five Area Daily Weighted Average Direct Slaughter Cattle
Report and the National Daily Direct Morning Hog Report-that livestock
experts told us were important to livestock producers.

In addition, we examined transactions reporters excluded from AMS reports.
First, we examined transactions made between April and June 2005.1 More
specifically, we reviewed data packers submitted on the Live Cattle Daily
Report (Current Established Prices) (LS-113), Swine Daily Report (LS-119),
and Boxed Beef Daily Report (LS-126) and compared it with the reports
published during this period. Second, we examined transactions AMS
excluded from its reports during periods of rapidly rising cattle and hog
prices-for cattle, transactions excluded by reporters for a key category
of live and dressed cattle prices from May through October 2003; for hogs,
those excluded from October 2003 to March 2004. To determine which
transactions were eliminated for market reasons,2 we reviewed the reporter
log field in the database. The logs identify transactions eliminated for
various reasons, such as price, low price, high price, or lot size.3 We
analyzed data from all days reported for this time period in the 35 to 65
percent choice steer grade of the Five Area Weighted Average Direct
Slaughter Cattle Report.4 We then calculated the weighted average prices
with and without the excluded transactions and the difference between
these prices.5 In addition, we performed a statistical test to determine
whether the difference between the prices, as a group, was statistically
significant.6

We discussed how AMS performed audits to ensure packers were complying
with the Livestock Mandatory Reporting Act provisions with AMS's Audit
Review and Compliance (ARC) officials in USDA headquarters, and with
auditors in both Des Moines and St. Joseph. As part of this effort, we
obtained and reviewed the mandatory price reporting audit reports that ARC
conducted from May 2002 through April 2005. In particular, we used ARC's
database of audit reports to analyze the number of audits conducted over
the time period, the number of findings related to those audits, and other
information. ARC officials and our analysis indicated that the number of
audit reports in the database closely approximated the number of audits
conducted. We found this database to be sufficiently reliable for this
purpose. Because this database did not provide specifics on the reasons
AMS believed some companies were out of compliance, we performed a
detailed review of all audit reports during one 4-month audit cycle from
June through September 2004. We also obtained information from AMS
Headquarters officials regarding the formal warning letters they sent
packers and the penalties they assessed.

We analyzed ARC's audit methodology for sampling transactions and the
extent to which that sample of transactions could provide information on
packer compliance and the accuracy of the reported prices. In addition, we
reviewed ARC policy and procedures, the audit report database, and had
discussions with ARC officials and auditors. Specifically, we interviewed
ARC officials regarding their audit methodology with emphasis on their
sampling methodology, and we reviewed their documentation on sample
selection. Furthermore, to analyze the agency's sampling procedure, we
compared the time between the audit field visit and the days selected for
the audit of a full day's transactions, and the audit of a sample of
transactions over the 4-month audit cycle from June through September
2004.

To determine the extent of coordination between GIPSA and AMS, we reviewed
their legislative authority, identified activities and investigations
involving both agencies, and reviewed GIPSA case file documentation from
the competition-related investigations in which GIPSA obtained packers'
transaction data from AMS. We met with USDA Headquarters officials from
AMS and GIPSA. In Des Moines, we met with GIPSA's Packers and Stockyards
Programs regional officials, and on separate occasions, spoke with GIPSA's
Denver Regional Office officials regarding GIPSA and AMS coordination.

During the course of our review, we identified and obtained the views of
several industry groups and associations representing packers and
producers. We also interviewed several nationally recognized economic
experts knowledgeable about mandatory price reporting and related market
issues.

We conducted our review between February and November 2005 in accordance
with generally accepted government auditing standards.

Analysis of Transactions AMS Excluded From Reports Appendix II

Overall, from April 2005 through June 2005, we found approximately 8.8
percent of cattle transactions, 0.2 percent of hog transactions, and 2.7
percent of boxed beef transactions were eliminated. From May 2003 to
October 2003, a period of rapidly rising prices, we found that
approximately 22.8 percent of all cattle transactions were excluded from
AMS reports.1 Figure 1 shows that close to 95 percent of all excluded
dressed weight cattle transactions from negotiated sales were smaller
lots-groupings of cattle for sales purposes-of fewer than 25 cattle.2
However, as figure 2 shows, the proportion of negotiated live cattle
transactions that were eliminated consisted of lots that were relatively
larger than dressed cattle lots and more consistent in size; about 75
percent of lots were greater than the 0 to 25 lot category and over 10
percent were between 201 and 400 head of cattle. Information on the size
distribution of excluded lots is relevant because excluding large lots
could have a relatively greater impact on weighted average prices reported
by AMS than smaller lots. Also, the effects of excluding large lots could
be greater in daily reports when trade volume is light, and an
accumulation of excluded large lots could affect weekly and monthly
reports.

Figure 1: Percent of Negotiated Cattle Transactions Excluded by Size of
Lot, Dressed Basis, May through October, 2003

Figure 2: Percent of Negotiated Cattle Transactions Excluded by Size of
Lot, Live Basis, May through October, 2003

Market News reporters of hog trade eliminated significantly fewer
transactions than the cattle reporters early on in the livestock mandatory
reporting program. For hogs, from October 2003 to March 2004, we found
that approximately 0.1 percent of transactions were excluded, which was
less than 0.1 percent of all hogs. Figure 3 shows that, for negotiated
sales, while nearly 40 percent of excluded transactions were smaller lots
of 50 hogs or less, the largest category of slaughtered swine
excluded-over 35 percent-were somewhat larger lots, in the 151-200 head
lot category.

Figure 3: Percent of Hog Transactions Excluded by Size of Lot, Negotiated
Sales, October 2003 through March 2004

Effects of Eliminating Transactions

During a sample period of rapidly rising prices, our analysis of cattle
and hog livestock data shows that the elimination of transactions from
Market News reports narrowed price ranges while having a limited, but
frequently positive, effect on the average reported price.3 To illustrate
this process, figures 4 and 5 show the differences in the distributions of
cattle prices for dressed steers from May through July 2003 and how
reporters' exclusion of cattle transactions eliminated outlying prices and
narrowed the range of prices. During this same time period, reporters'
exclusions decreased the number of packer transactions from 4066 to 3334.
Excluding these transactions narrowed the associated price range-the
difference between the minimum and maximum price-from $117.95 to $16.50
per hundredweight.

Figure 4: Distribution of Cattle Prices Before Exclusions, Dressed Steers,
May through July, 20034

Price in dollars per hundredweight.

Note: The mean price or average price in the insert of the figure, unlike
the other averages in this appendix, is not a weighted mean, or adjusted
for head count.

Figure 5: Distribution of Cattle Prices After All Exclusions, Dressed
Steers, May through July, 2003

Price in dollars per hundredweight.

Note: The mean or average price in the insert of figure, unlike the other
averages in this appendix, is not a weighted mean, or adjusted for head
count.

Market News reporters' elimination of data for market reasons from reports
between May and October 2003 had the effect of narrowing price spreads or
ranges on a daily basis.5 For dressed steers, figure 6 shows the narrowing
of the range of prices over this period before and after all excluded
transactions, most of which were excluded for market reasons. As shown in
the figure, price ranges before any excluded transactions during this
period were from $2 to $20 per hundredweight while, after all market
exclusions were made, the range decreased to between $0 and $12 per
hundredweight. Market News reporters are instructed to exclude prices that
are $5 above or below the market to narrow the range of reported prices
and AMS record logs indicate that they do so. However, when prices are
rising or falling rapidly, this practice may exclude some transactions
that should reasonably be presented as reflecting the day-to-day
variations in the market. Also, since these are national daily reports,
price spreads tend to be larger since they encompass the full range of
prices for all regions.

Figure 6: Range of Cattle Prices Before and After Excluded Transactions,
Dressed Steers, May through October, 2003

During May to October 2003, a period of rapidly rising cattle prices, we
estimate that the effect of eliminating transactions for market reasons
was negligible about two-thirds of the time, while for the remaining third
the reported average prices were generally higher than they would have
been had these transactions not been eliminated. For live cattle sales,
figure 7 displays the differences between the average weighted daily
prices after AMS exclusions (as reported in Market News reports) and the
average weighted prices based on including the transactions that AMS had
excluded for market reasons for 35-65 percent choice steers from May
through October 2003. The average weighted prices published by AMS for
these dates were the same about 67 percent of the time, higher 31 percent
of the time, and lower 2 percent of the time over this period. This
suggests, and Market News record logs confirm, that during this period
when Market News reporters were excluding transactions, they were
predominantly excluding transactions for reasons of lower price rather
than high price. We found that over twice as many transactions were
excluded for low price as for high price during this period.

Figure 7: Difference in Weighted Average Prices With and Without
Transactions AMS Excluded for Market Reasons, Live Weight Cattle, May
through October, 2003

For 35 to 65 percent choice steers, dressed weight, figure 8 shows the
differences between the daily weighted average prices reported by AMS, and
the average prices that AMS would have reported if AMS reporters had not
eliminated transactions for market reasons. These differences display a
trend similar to the one we identified for live cattle prices. When we
compared our calculations of the weighted average prices with those AMS
reported, about 32 percent of the price differences were higher than those
AMS would have reported; about 67 percent were the same or about the same,
and 1 percent were lower. This result indicates that market reporters of
livestock were excluding a higher proportion of low prices during this
period. AMS reporters may have excluded low prices more frequently during
the period because prices were rising. What a reporter considered to be a
high price during one week may have appeared to be a much lower price by
the following week. Also, at the low end of the price ranges, transactions
may have been excluded because the prices represented low-quality
animals.6

Figure 8: Difference in Weighted Average Prices With and Without
Transactions Excluded for Market Reasons, Dressed Weight Cattle, May
through October, 2003

The effect of an excluded transaction on any particular day is determined
by how large that transaction is compared to the size and number of
transactions that took place on that day or that week, and how far it is
from the range of reported prices. While each transaction alone may be
considered a small lot, the total effect of a number of excluded
transactions for this reason can cumulatively have a large effect on the
weighted average price.

To determine if there was an overall statistical difference between our
replications of AMS prices and the prices we determined would have been
reported had reporters not eliminated transactions for market reasons, we
tested the two average weighted price series for both live and dressed
cattle. We found that for both live and dressed weight cattle, there was a
statistically significant difference in the weighted averages between
reported AMS prices and the prices that would have been reported if
exclusions had not been made for market reasons.7

Our analysis of data from AMS's daily hog reports from October 2003 to
March 2004 showed that, for the reports that we examined, reporters
frequently eliminated transactions that they believed to be errors that
would potentially widen price ranges.8 However, unlike cattle, there were
very few transactions eliminated from reports for market reasons. As a
result, for hogs, price ranges with and without exclusions by market news
reporters were more similar than for cattle. As illustrated in figure 9,
the difference between prices reported by AMS and prices that would have
been reported by Market News was notable on only 7 days for the National
Daily Direct Morning Hog Report from October 2003 through March 2004. A
similar analysis of the afternoon hog report shows the same pattern.

Figure 9: National Daily Direct Morning Hog Report: Differences of
Weighted Average Prices With and Without Transactions Excluded, October
2003 through March 2004

Additional Information Concerning AMS Warning Letters to 11 Packers
through September 2005 Appendix III

Source: GAO analysis of USDA warning letters and legal agreements with
packers.

Comments from the Department of Agriculture Appendix IV

GAO Contact and Staff Acknowledgments Appendix V

Robert A. Robinson, (202) 512-3841

In addition to the individual above Charles Adams, Assistant Director,
Aldo Davila, Barbara El Osta, Paige Gilbreath, Kirk Menard, Lynn Musser,
Karen O'Conor, Alison O'Neill, Vanessa Taylor and Amy Webbink made key
contributions to this report.

(360547)

www.gao.gov/cgi-bin/getrpt?GAO-06-202.

To view the full product, including the scope

and methodology, click on the link above.

For more information, contact Robert A. Robinson at (202) 512-3841 or
[email protected].

Highlights of GAO-06-202, a report to congressional requesters

December 2005

LIVESTOCK MARKET REPORTING

USDA Has Taken Some Steps to Ensure Quality, but Additional Efforts Are
Needed

Livestock producers, with gross income of $63 billion in 2004, depend on
USDA's daily, weekly, and monthly livestock market news reports. These
reports provide them and others in the industry with livestock and meat
prices and volumes, which are helpful as they negotiate sales of cattle,
hogs, lamb and meat products. Packers also use the average prices in these
reports as a basis for paying some producers with whom the packers have
contracts. In 1999, the Livestock Mandatory Reporting Act was passed to
substantially increase the volume of industry sales transactions covered
by USDA's market news reports and thereby encourage competition in the
industry. In the context of ongoing discussions about the renewal of this
act, GAO reviewed (1) USDA's efforts to ensure the quality of its
livestock market news reports and (2) the coordination between two USDA
agencies that are responsible for promoting competition in livestock
markets.

What GAO Recommends

GAO recommends, among other things, that USDA improve the transparency of
its market news reports, and its auditing of packers' transactions. GAO
provided a draft of this report to USDA for comment. USDA generally agreed
with the report and stated it has started to implement the report's
recommendations.

While the U.S. Department of Agriculture (USDA) took important actions to
produce quality livestock market news reports, GAO found that USDA could
improve the reports' transparency. Although packers with large plants must
report all of their livestock transactions to USDA, GAO found that USDA
market news reporters regularly excluded some transactions as they
prepared USDA's reports. For example, GAO's analysis showed that from
April through June 2005, USDA reporters excluded about 9 percent of the
cattle transactions that packers had reported. When USDA excluded
transactions, this sometimes changed the low, high, and average prices
that USDA would have otherwise reported. However, USDA has not informed
its readers of the extent of this practice. Moreover, USDA's instructions
for guiding its market news reporters as they prepared their reports
lacked clarity and precision, leading to inconsistency in their reporting
decisions.

In addition, GAO found the accuracy of USDA's livestock market news
reports is not fully assured. About 64 percent of 844 USDA audits of
packers-conducted over 36 months ending in April 2005-identified packers'
transactions that were inaccurately reported, unsupported by
documentation, or omitted from packers' reports. Moreover, some packers
have not promptly corrected problems. Since 2002, USDA has sent 11 packers
21 letters urging the packers to correct longstanding problems and warning
them of the consequences of delay. Twice USDA has levied $10,000 fines on
packers, but suspended the fines when these packers agreed to comply. As
of September 2005, USDA had continuing issues with 2 of the 11 packers.
USDA officials noted that packers' errors are relatively few compared to
the large volumes of data that packers report daily. However, USDA has not
(1) assessed the overall quality of packers' data, (2) used its audit
results to help focus future audit efforts, and (3) ensured that follow-up
promptly resolves problems.

Two USDA agencies have addressed competition in livestock markets-the
Agricultural Marketing Service (AMS) and the Grain Inspection, Packers and
Stockyards Administration (GIPSA). GAO found the coordination between
these agencies to be limited, primarily due to the legal authority within
which each operates. AMS has implemented the Livestock Mandatory Reporting
Act. That act did not provide authority for AMS to share individual packer
transaction data within USDA except for enforcement purposes. In two
investigations, AMS provided packers' data to GIPSA. On the other hand,
GIPSA enforces the Packers and Stockyards Act and is responsible for
addressing unfair and anti-competitive practices in the marketing of
livestock. Furthermore, GAO found that GIPSA monitors cattle and hog
markets by analyzing publicly available livestock market news reports-an
approach that has limitations because it lacks the company-specific
information that would be useful for detecting anti-competitive behavior.
*** End of document. ***