challenges opportunities for the -...

Coordinated by

Washington, DC | June 2009

CHALLENGES & OPPORTUNITIES

for the

21st CENTURY

A Review of the USDA-NASS Agricultural Prices Program:

Challenges &Opportunities

for the

21st Century

Coordinated by

Washington, DC | June 2009

NOTICE: The project that is the subject of this report was approved by the Council on

Food, Agricultural & Resource Economics (C-FARE) Board of Directors. The members of the

Review Panel responsible for the report were chosen for their special competences and with

regard to appropriate representation. This study was supported by Contract/Grant Nos.

08–OA–2090–038 and 09–OA–2090–001 between C-FARE and the U.S. Department of

Agriculture’s National Agricultural Statistics Service (USDA-NASS).

Additional copies of this report are available from C-FARE (www.cfare.org) or USDA-NASS

(www.nass.usda.gov). Printed in the United States of America. Copyright 2009 by the C-FARE

Board of Directors. All rights reserved.

Suggested citation: The Council on Food, Agricultural & Resource Economics (C-FARE). (2009).

A Review of the USDA-NASS Agricultural Prices Program: Challenges & Opportunities for the

21st Century. Washington, DC.

A R e v i e w o f t h e U S D A - N A S S A g r i c u l t u r a l P r i c e s P r o g r a m i

This Review reflects the efforts of many people beyondmembers of the Review Panel. The Council on Food,Agricultural & Resource Economics (C-FARE) is deeplyindebted to all who contributed. National AgriculturalStatistics Service (NASS) Administrator Cynthia Clark,Deputy Administrator Carol House and NASS seniorstaff provided publications, documents, information andconsultation essential to the Review Panel’s work. KevinBarnes, Chief of NASS’s Environmental, Economics andDemographics Branch, served as liaison to the ReviewPanel and deserves special thanks for his faithful andprompt responses to all requests for assistance andinformation. Several NASS staff members wereextremely helpful in ensuring that the Review Panel fullyunderstood the issues. Indeed, members of the ReviewPanel express great respect for the professionalism,integrity and dedication of NASS with regard to all itsstatistical efforts.

The Review Panel’s task was made easier by excellentassistance from the C-FARE office. C-FARE’s ExecutiveDirector, Tamara Wagester, arranged all the meetings,conference calls and financial matters, while maintaininga constant stream of communications and services tothe Review Panel. C-FARE intern, Alison Rozema,provided assistance. Marie E. Lee served as the generaleditor of the report. Colleen Clancy McGinley providedgraphic design and layout. NASS assumed responsibilityfor printing and distributing the report.

Four individuals chosen for their technical expertise andperspectives on agricultural prices reviewed the reportin draft form. The purpose of these independentreviews was to provide candid and critical comments tohelp make the report as sound as possible and to

ensure that it meets accepted standards for objectivity,evidence and responsiveness to the charge. We wish tothank Richard Just, University of Maryland; GeneNelson, Texas A&M University; Michael Sykuta,University of Missouri; and Robert Tortora, Gallup, Inc.,for providing timely and independent external reviews.Although these individuals provided constructivecomments and suggestions, responsibility for the finalcontent of this report rests entirely with the ReviewPanel.

The members of the Review Panel deserve specialcommendation for their outstanding work under tightdeadlines. Representing diverse interests andbackgrounds, the Review Panel came together quicklyto function as a team. They gave their time, energy andexpertise to make this an insightful and useful documentthat will assist NASS in improving the content,procedures and outputs of the Agricultural PricesProgram to the benefit of public policy and theAmerican public.

In closing, we wish to acknowledge the leadership andorganizational skills of John E. Lee Jr., who served asReview Director. Dr. Lee was diligent in guidingquestions to the right person and coordinating thedifferent responsibilities of the Review Panel. Hisknowledge and appreciation of U.S. agriculture and theimportance of accurate price data provided by NASSwere major factors in the success of this review.

Steve C. Turner, Chair Gail L. Cramer, Vice-Chair

USDA-NASS Agricultural Prices Program Review Committee

June 2009

Acknowledgements

J u n e 2 0 0 9 | The Counci l on Food, Agricultural & Resource Economics ii

A R e v i e w o f t h e U S D A - N A S S A g r i c u l t u r a l P r i c e s P r o g r a m iii

In April 2008, NASS asked C-FARE to assemble a panelof expert social scientists from academia, governmentand the private sector to conduct an “independent,comprehensive and objective review” of the NASSAgricultural Prices Program. The purpose of theReview was to identify the strengths and weaknesses ofthe Agricultural Prices Program and to recommendchanges to make the published statistics more accurateand useful. Implicit in that purpose were twoquestions:

1. Is NASS doing things right?

2. Is NASS doing the right things?

The first question relates to technical and proceduralissues involved in the collection, processing anddissemination of statistics on agricultural prices. Thesecond is much broader and relates to the rationaleand aim of providing price data, e.g., the uses and usersof the data, frequency and form of publication andtransparency to users of all aspects of the data sources,processes and products. To address those questions,the Review Panel examined the full range of activitiesrelated to agricultural prices in NASS, includingprogram objectives, data sources, processes andproducts.

In addition to the charge from NASS, the ReviewCommittee Chair charged the Review Panel as follows:

“Your responsibilities are both general and specific. Ingeneral you are to review all the input available to youcarefully, including both written and oral information;identify issues, questions and opportunities that you

perceive; ask questions and pursue informationnecessary to clarify issues and problems; assist inclassifying issues for purposes of categorizing andremanding them to appropriate sub-panels; and makesuggestions and recommendations useful to improvingthe quality and success of the Review.”1

The three components of the NASS Agricultural PricesProgram were the subject of this Review:

1. Prices received (for agricultural commodities);

2. Prices paid (for inputs to agricultural production);and

3. Price indexes (for prices received, prices paid andparity).

Sub-panels of the full Review Panel were organizedaround the preceding three components.

Each sub-panel prepared a report for its respectivecomponents of the Agricultural Prices Program with anevaluation and recommendations. The componentreports are Chapters 5, 6 and 7 of the report. Thatdivision of labor notwithstanding, all evaluations andrecommendations contained in this report representthe consensus of the entire Review Panel. Finally, in thecourse of its work, the Review Panel examined severalcommon themes and generated several recommenda-tions that either cut across the three component areasor pertained to more than one. These commonthemes and general recommendations are reported inChapter 4.

Charge to the Review Panel

1 Letter to Review Panel dated October 1, 2008.

J u n e 2 0 0 9 | The Counci l on Food, Agricultural & Resource Economics iv

The members of the Review Panel expressed theircollective intent to provide recommendations that areactionable, substantive and consistent with soundeconomic and statistical principles. It is their belief thatrecommendations in this report follow the guidelinessuggested by the National Research Council (NRC), asexpressed in that Agency’s Principles and Practices for aFederal Statistical Agency.2

For its part, NASS provided the Review Panel with dataand any information requested, but NASS did notintervene in the Review Panel’s deliberations. NASS alsogave the Review Panel the freedom to consider issuesnot specifically included in the charge, but, at the sametime, made it clear that NASS is constrained in its abilityto respond to recommendations on issues outside thescope of the Review. For example, NASS is not at libertyto dictate policies to other agencies whose data areimportant to the NASS Agricultural Prices Program.Likewise, NASS has to respond to legal mandates, evenin cases where the Review Panel felt the mandates nolonger served their original purposes. Finally, NASScommitted to publish this report without alteration orcomment and make it available to the public.

2 Martin, Margaret E., Straf, Miron L. and Citro, Constance F., editors, NationalResearch Council. Principles and Practices for a Federal Statistical Agency:Third Edition. 2005.

A R e v i e w o f t h e U S D A - N A S S A g r i c u l t u r a l P r i c e s P r o g r a m v

At-large Review PanelMember

William MotesInforma Economics, Inc.

Staff to Review Panel

John E. Lee Jr., Review DirectorMississippi State University (retired)

Tamara Wagester, Executive DirectorCouncil on Food, Agricultural &Resource Economics (C-FARE)

Alison Rozema, InternC-FARE

Review Committee and Review Panel

Sub-panel 1, Prices Received

Ismael Flores Cervantes, ChairWestat, Inc.

Brian Buhr, Co-chairUniversity of Minnesota

Richard Heifner, Co-chairUSDA-Economic Research Service(retired)

Corinne AlexanderPurdue University

Johnny BlairAbt Associates, Inc.

Richard S. SigmanWestat, Inc.

James ZavrelU.S. Department of Commerce, Bureauof Economic Analysis

Sub-panel 2, Prices Paid

William G. Tomek, ChairCornell University

Arlin BrannstromAmerican Society of Farm Managers and Rural Appraisers

Ronald S. FecsoU.S. Government Accountability Office

Jeffrey T. LaFranceWashington State University

Sub-panel 3, Price Indexes

Walter J. Armbruster, ChairFarm Foundation (President Emeritus)

Cathryn S. DippoU.S. Department of Labor, Bureau ofLabor Statistics (retired)

Robert ParkerU.S. Government Accountability Office(retired)

Michael K. WohlgenantNorth Carolina State University

Steve C. Turner, Chair

Mississippi State UniversityGail L. Cramer, Vice-chairLouisiana State University

J u n e 2 0 0 9 | The Counci l on Food, Agricultural & Resource Economics vi

A R e v i e w o f t h e U S D A - N A S S A g r i c u l t u r a l P r i c e s P r o g r a m vii

Executive Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .1

Part 1: Background and Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .11

1. Historical Perspective . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .11

2. Objectives, Uses and Users of NASS Price Statistics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .15

3. Current Issues and Challenges . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .19

Part 2: Reports and Recommendations of the Review Panel . . . . . . . . . . . . . . . . . . . . . . . . . . . . .23

4. Common Themes and General Recommendations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .23

5. Prices Received for Farm Commodities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .31

6. Prices Paid by Farmers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .35

7. Price Indexes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .41

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .49

Glossary of Terms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .51

Acronyms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .55

Appendices

A. Review Procedures and Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .57

B. Stakeholder Input to the Review . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .59

C. Biographies of the Review Panel . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .61

Contents

J u n e 2 0 0 9 | The Counci l on Food, Agricultural & Resource Economics viii

BAC K G R O U N D A N D OV E R V I E W

In April 2008, the U.S. Department of Agriculture’s(USDA)3 National Agricultural Statistics Service(NASS) asked the Council on Food, Agricultural &Resource Economics (C-FARE) to assemble a panel ofexpert social scientists from academia, governmentand the private sector to conduct an “independent,comprehensive and objective review” of the NASSAgricultural Prices Program. The purpose of theReview was to identify the strengths and weaknessesof the Agricultural Prices Program and to recommendchanges.

The collection and publication of statistics on agricul-tural prices have a long history in USDA. Collection ofprices received data began in 1866. Prices paid datawere first collected in 1911, as USDA and theCongress began to examine economic conditions inthe farm sector more closely. The collapse of agricul-tural commodity markets after World War I and theonset of the Great Depression in 1929 gave newimpetus to analyses of the well-being of farmers andto the generation of statistics to support thoseanalyses. At that time the concept of parity appeared.It established a relationship between prices receivedby farmers for the commodities they sold and pricespaid for the inputs they purchased. The parityconcept was used to define public policy objectivesuntil the 1980s, in some cases. The inclusion of theconcept of parity prices for farm commodities intothe Agricultural Adjustment Act (AAA) of 1938, apermanent act, effectively mandated USDA and the

The key findings ofthis Review are:� NASS needs a fresh vision

for the prices program.� Fundamental improvements

are needed in the conceptualbasis for prices and priceindexes.

� NASS must address theresponsibility that goes withheavy dependence on dataand indexes from otheragencies.

� To address these needs,NASS must commit to astronger program of future-oriented research to supportthe operations program.

� Increased transparency isessential to all aspects of theAgricultural Prices Program.

A R e v i e w o f t h e U S D A - N A S S A g r i c u l t u r a l P r i c e s P r o g r a m 1

Execut ive Summary

3 Refer to Acronyms, pp. 55.

predecessor agencies to NASS to collect and publishstatistics necessary to define parity indexes as well.Over the ensuing decades, coverage of commoditiessold and of inputs purchased has expanded, and indexbase periods and item weights have been updated peri-odically. Because of the modernization of agricultureand changes in farm policies, the need to calculate paritymeasures, while still mandated by law, is no longer theprimary reason for producing price statistics. Today,price statistics produced by NASS are heavily used inthe administration of Federal programs, the calculationof various economic indicators, analyses of commoditymarkets and in a host of other ways by public andprivate users. Despite the critical importance of pricestatistics, they have had to compete for limitedresources in recent years. It appears that price statisticshave received lower priority than other more marketsensitive commodity production statistics produced byNASS. It is in recognition of the need to reexamine andrevitalize the Agricultural Prices Program that NASSrequested this Review.

The three major components of the NASS AgriculturalPrices Program were the subject of this Review:

1. Prices received (for agricultural commodities);

2. Prices paid (for inputs to agricultural production);and

3. Price indexes (for prices received, prices paid, andparity).

Sub-panels of the larger Review Panel were organizedaround these three components. Each addressedobjectives and uses of the statistics, as well as statisticalissues such as data sources, sampling, collection,processing, analysis and index construction.Considerable attention was paid to transparency anddocumentation of all aspects of the Agricultural PricesProgram, including program purposes.

For its part, NASS provided the Review Panel with dataand information requested, but did not in any wayintervene in the Review Panel’s deliberations. NASS also

gave the Review Panel thefreedom to consider issues notspecifically included in the charge,but at the same time made clear

that the Agency is constrained in its ability to respond torecommendations on issues outside the scope of theReview. For example, NASS is not at liberty to dictatepolicies to other agencies whose data are important tothe NASS Agricultural Prices Program. Finally, NASScommitted to publish this report without alteration orcomment and make it available to the public.

It is the intent of the Review Panel that the recommen-dations in this report, and the text which explains andsupports those recommendations, be useful to NASS asit develops vision, purpose and content of theAgricultural Prices Program to meet the needs of the21st century. All the recommendations are alsointended to improve the quality of NASS price statistics.

Common Themes and GeneralRecommendations

Several common and important themes emerged fromthe Review Panel’s deliberations on needed improve-ments in the NASS Agricultural Prices Program. For themost part, the common themes cut across allcomponents of the Program. The themes pertain to:

1. transparency and documentation,

2. the critical importance of research,

3. the use of data from external sources,

4. NASS’s response to changes in agriculture anddemands of modern statistical systems,

5. the importance of specificity in defining attributesof items for which prices are collected, and

6. index construction (treated in Chapter 7).

RE C O M M E N DAT I O N 4–1

NASS should increase transparency and docu-mentation of all aspects of its AgriculturalPrices Program.

NASS’s data users should have easy access to under-standable, up-to-date information on data sources, datacollection methods, sampling methods, list complete-ness, processing and editing procedures, sources andmagnitude of errors and changes in the industry thataffect the collection and usefulness of the data. Thisapplies to statistics that NASS publishes which arederived from data from other agencies, as well as

J u n e 2 0 0 9 | The Counci l on Food, Agricultural & Resource Economics 2

statistics based on NASS surveys. It is equally importantthat the purposes and conceptual basis for pricestatistics be apparent to users. NASS should clearlydefine formulas for calculating price indexes and termsused in calculations and document justifications for theiruse.


NASS should substantially strengthen its coreresearch capacity.

NASS has a longstanding reputation as a supplier ofdependable, objective and accurate statistics to theFederal government and to the American public. Tocontinue its prominence, NASS should enhance itscapacity to conduct and gain access to research tosupport its operations programs. The Review Panelnoted that NASS’s research capability—once a sourceof public and professional confidence in the Agency’sstatistics—has been diminished over recent decades. Inthe report, the Review Panel identified the need formore research expertise in index construction,information systems and data-collection methodology.To the extent possible, NASS should justify withdocumented research all substantial decisions aboutprice collection procedures, use of data from externalsources versus internal survey data and index construc-tion methods. The documentation should includeestimates of effects of these decisions on the accuracyof the statistics produced. Documentation and researchare necessary to assure the use of up-to-date statisticalmethods and consistency across the Agency. It isimportant for Agency credibility, in that it allowsacademics and others outside the Agency to evaluateprocedures and revisions. Increased transparency couldalso generate some supportive external research, e.g.,by university faculty, if NASS establishes a workingprocedure for receiving and responding to input fromsuch external research.


NASS should accept full responsibility for howit uses, and explains the uses of, data fromother agencies for calculating prices and priceindexes.

Other agencies, principally the Agricultural MarketingService (AMS) and the Bureau of Labor Statistics (BLS),

are major sources of data used inthe calculation of NASS pricestatistics. The prices for mostlivestock and livestock productsand for most fruits, vegetable and specialty crops comefrom the AMS price reports. The statistical qualities ofthe AMS data are either unknown or not transparent.The total weight of items in the index of prices paidwhich come from AMS, BLS, ERS and DMRkynetec is48 percent for the base survey month of April and 72percent for other months. It was not apparent to theReview Panel that NASS had full access to, and madefull use of, information on statistical properties of theseexternal data, the nature and timing of changes and theappropriateness of uses being made of the data. TheReview Panel recommends that NASS should establishan internal team that works on methodological andother issues related to the uses of prices from otheragencies. Also, NASS should assume a leadership role incoordinating communications between NASS and otheragencies producing data used by NASS. Thesesuggestions imply that NASS should seek to betterunderstand the price collection procedures of AMS andindex construction concepts used in BLS and the impli-cations of changes in those agencies’ procedures.


NASS should undertake a comprehensive re-examination of its Agricultural Prices Programto develop a vision for a system of pricestatistics relevant to the emerging global foodand agricultural economy and consistent withbroader Federal and international statisticalsystems.

Changes in farm production and marketing methodspose challenges to estimating prices received and pricespaid. The emergence of new technologies and businesspractices in production and harvesting changes the mixand variety of inputs purchased and how they arepurchased. This increases the complexity of coverage,weighting and collection of prices paid data. Verticalcoordination and integration in the commoditymarketing system make it increasingly difficult todetermine farm level prices. Equally important, NASSshould rethink the role of the price statistics program in


the context of its linkage to national and internationalstatistical systems. With the introduction of the NorthAmerican Industry Classification System (NAICS), NASSshould consider its publication goals broadly withinNAICS–Sector 11: Agriculture, Forestry, Fishing andHunting. NASS should also consider producing priceindexes both conceptually and mathematicallycompatible to the BLS Producer Price Index (PPI), forexample, making it possible to combine indexes fromthe two agencies to produce a complete NAICS–Sector11 index. In short, the world in which agricultural pricestatistics are compiled and used has changed, and thatchange must motivate a comprehensive reassessment ofwhat the NASS price statistics program is all about.


NASS should address the issue of potentialbias in its price estimates caused by lack ofspecificity in defining attributes of items forwhich prices are collected.

NASS collects data on prices of items that are definedwith varying degrees of generality. This opens thepossibility of reported price changes coming from achange of attributes of the item or in the conditions ofthe transaction rather than from a change in the actuallevel of prices of items with precisely-defined attributes.Several recommendations in this report relate directlyor indirectly to this concern. (See Recommendations 5–1,5–3, 6–1 and 7–3.) The problem is illustrated in NASS’spractice of collecting expenditure and quantity data forcalculating prices of broadly-defined commodities,products and input items. In contrast, AMS surveyscollect actual reported prices for more narrowly-defined products, though not with probability sampling.NASS may have two conflicting objectives in thecollection of price data. One is to collect prices that,

when multiplied by quantity, givetotal revenue or expenditures.This helps measure farmincome, for example. But, if theobjective is to measure pricelevel changes, then prices shouldbe collected for commodities orinputs with specific attributesthat are held constant over time,while having a method of

dropping outdated items and adding new items, as BLSattempts to do with the Consumer Price Index (CPI).The Review Panel sees this as a critical issue for NASSto address, as it involves both program objectives andaccuracy of reported statistics.

In addition to the common themes, the Review Paneloffers three additional general recommendations forNASS’s consideration:


NASS should consider collecting andprocessing monthly prices received and pricespaid so that they can be published early in thefollowing month.

Both prices received and prices paid statistics are usedfor making agricultural projections and for administeringfarm programs. One of the most important uses ofNASS price data is for the monthly World AgriculturalSupply and Demand Estimates (WASDE) Reportprepared by the World Agricultural Outlook Board(WAOB). The WASDE Report is widely used and has amajor influence on commodity markets. It is usuallyreleased on the 10th, 11th or 12th of each month.Currently, its price predictions rely on the full-monthprices from two months earlier and the mid-monthprices for the previous month. The Review Panelexpressed concern that the mid-month price is subjectto point-estimate sampling problems and may not berepresentative of overall price patterns. Therecommended change would make full-month estimatesfrom the preceding month available for WASDE andother uses. Mid-month prices for field crops would nolonger be collected. Moreover, earlier availability ofmonthly average prices would be advantageous in theadministration of farm programs. For example, MilkIncome Loss Contract (MILC) payments could be madeearlier if NASS’s monthly average prices for feed wereavailable earlier. Any change in the schedule of reportingprices would have to be coordinated with WAOB andother users.



NASS should seek balance between qualityimprovement efforts and assuring consistencyof price statistics series over time.

Historical data on prices and price indexes are valuableto policymakers, researchers and the general public.Consistency in the price statistics series is important tomany users who follow price movements over time.Any changes in the methods used to collect data andcalculate price statistics should be made as often asnecessary to improve statistical quality. However, it isimportant that these changes be made deliberately andopenly. When estimation methods change for a series,reports should provide information for splicing the newestimates onto older series. Consistency is required sothe statistics are reliable ongoing indicators of pricerelationships. This recommendation is for transparency,not an argument against quality improvements. Anychanges in data sources, quality characteristics or char-acteristics of data obtained from other agencies shouldbe made transparent to users, along with informationon how to interpret and value the changes.

NASS must continue to seek balance between users’immediate needs and the longer-term goal of improvingthe quality of the price estimates. It became apparentduring the course of the Review Panel’s deliberationsthat maintaining a consistent series to administerexisting programs sometimes conflicts with thedevelopment of improved price estimates, which mightcontribute to better policies in the future. This canresult from a lack of strong research on what theoptimum price series should be. In cases where the dataare required for administering programs, but viewed asnot meeting best practices, NASS should develop a newseries to replace the deficient series and then bridge theold and the new series to minimize disruption to usersin the program agencies.


On its website and in itspublications, NASS shouldtreat “agricultural prices”as a major subject and notas a sub-category of“Economics.”

Identifying “agricultural prices”as a separate category wouldfacilitate user access toinformation about the prices data (especially online) andwould result in NASS combining information nowshown under several topics. Also, the resulting highervisibility would be consistent with the Review Panel’sview that agricultural prices are important and shouldbe supported with a new vision and clarity of purpose.

Prices Received for Farm Commodities

NASS prices received statistics are critically importantto the administration of many Federal farm programs.They are also required for the calculation of keyeconomic indicators for the farm sector (e.g., farmincome, commodity costs and returns and farm sectorproductivity), U.S. and world supply and demandestimates, important components of the NationalIncome Accounts and the calculation of the parity ratio.Statistics on prices received for livestock have fewerapparent uses in the administration of Federalprograms, but are important for all the other uses ofstatistics on prices received. Because prices receivedstatistics are only reported monthly, they are notintended for short-term marketing decisions. For thatpurpose, AMS Market News reports are more useful. Akey issue in the development of prices receivedstatistics is the heavy dependence on price datacollected by AMS.


NASS should review, update and improvecriteria for choosing between conductingprobability surveys and using price data fromAMS and other sources especially for thosecommodities not included in indexes.

The use of data from probability surveys and from othersources in the same agricultural prices tables and the


same price indexes raises questions about the compara-bility and accuracy of the two data sources. The surveyapproach used for field crops provides greaterassurance of complete coverage and allows measures ofprecision to be calculated. The use of price data fromAMS is less expensive, and the monthly estimatesappear less likely to be affected by monthly and annualchanges in the attribute mix within a commodity. NASSshould develop procedures for measuring the coverageand accuracy of monthly price estimates based on pricedata obtained from AMS and other sources. Suchprocedures should take into account completeness ofcoverage, bias and precision of estimates, users’ needsfor accuracy, the value of continuity in series, burdenson respondents and NASS’s costs. There may be waysto work more closely with AMS to make their pricedata collection methods more compatible to NASS’sneeds.


NASS should track and evaluate quantitativemeasures of prices received data collectionand processing operations to identify areas forimproving the processes and resultingstatistics.

The prices received program currently tracks theestimated coefficients of variation (CVs) at the nationallevel. Other quantitative measures that could be trackedinclude:

1. the proportion of reporters who fail to report inaccordance with NASS standards;

2. rates and patterns of item non-response and unitnon-response; and

3. edit failure rates for key items.

For each of these, the actual impact on estimates shouldbe assessed. Tracking such measures will provide datafor comparing performance over time. A periodicsummary and review of changes in these errorindicators will help identify points in the survey processin need of error reduction efforts.


NASS should review, update and improve thecriteria used to determine commodity, stateand attribute coverage.

The development and implementation of new coveragecriteria would serve to determine the overall scale ofthe price statistics program and which commodities andstates to include at the margin. Currently commoditiesand states to be included for each commodity aredetermined every three years so that at least 90percent of aggregate U.S. cash receipts are covered andeach included commodity has 90 percent coverage. The90 percent criterion is somewhat arbitrary. It should bereviewed from time to time and subjected to sensitivityanalysis to determine whether and how large deviationsfrom 90 percent would need to be in order to make ameaningful difference in application. In principle,commodities, states and attributes should be added ordropped at the margin depending on whether benefitsto users of the statistics exceed the costs to NASS andrespondents for generating the statistics.

Prices Paid by Farmers

Prices paid statistics are required for the calculation ofparity measures, but may be more important for thecalculation of farm income, farm costs and returns andother economic indicators, as well as for uses in theprivate sector. Data for producing prices paid statisticscome from numerous sources, including NASS surveys,AMS, BLS, USDA’s Economic Research Service (ERS)and private data companies. These data are largely forinputs used to produce traditional, economicallyimportant crops and livestock and livestock products.NASS surveys of prices paid are currently conductedannually. Until 2009, NASS surveyed prices for themonth of April; now NASS uses March prices. Theindexes of prices paid are reported monthly based ondata from BLS, AMS and internal NASS sources.



NASS should address the goals and principlesfor changing the mix of inputs and selectingthe attributes of items for which prices are tobe collected.

Profound changes in the agricultural sector requireNASS to be vigilant in determining what input pricesshould be collected. Large commercial farms arebecoming more specialized but in varied ways. Somepurchase inputs in bulk, wholesale and not fromtraditional local retailers. At the same time, there hasbeen an increase in organic farming and in production ofother specialty crops and livestock. These likely usevery different inputs and inputs from different sourcesthan do large commercial farms. Some inputs becomeless commonly used or become used in different forms.NASS needs documented criteria for guiding coverageof input items for which data are collected.

The quality and other attributes of some inputsconsumed in agricultural production change, sometimesdramatically, over time. For example, embedded tech-nologies make seeds for many crops far moreproductive today than in the past. A given size orhorsepower tractor may be very different and far moreproductive than a similar size tractor 30 years ago.These changes affect the meaning of price comparisonsover time. Criteria should be developed for pricingattributes and for incorporating attribute changes intoprice series. In any case, attributes of items for whichprices are collected should be carefully defined toreduce bias in estimates of price changes, i.e., todistinguish between price level changes and changes inattributes.


NASS should change to semi-annual or morefrequent surveys of prices paid.

Surveys in March or April to collect input prices may berelevant for major crops planted in the spring, but notnecessarily for winter wheat or livestock-related inputs.It is not clear that the combination of annual pricesurveys with BLS-based monthly adjustments canadequately capture the monthly variability in inputprices. A second survey period, perhaps August,September or October, would likely reduce the

adjustments needed in reportedmonthly prices paid statistics tobring them in line with surveydata.


NASS should develop a research program inprices paid statistics that addresses criticalproblems in survey design, documentationand identification of appropriate input itemsto cover.

Survey research is concerned not only with imprecisionarising from sample selection and estimation, but alsodue to various non-sampling errors such as coverage,specification, measurement, non-response and withprocessing errors. The Review Panel suggests a need forfurther information in each of these areas. Specificsuggestions are provided.

Price Indexes

The Index of Prices Received (1910) and the Index ofPrices Paid (1928) had their origins in attempts todevelop simple, aggregate measures of prices receivedand paid. Over the ensuing years, items (commoditiessold and inputs purchased) were added to both Indexesto provide coverage that was more complete. Also, theweights of items in both Indexes were periodicallyadjusted to reflect changes in commodities producedand inputs purchased. The parity index and relatedparity measures, based on prices paid and received,were developed in the 1930s as part of the effort tomeasure well-being of farmers. The years 1910–1914were the original base reference period for all agricul-tural indexes used for U.S. farm support programs.Additional base periods and reference periods wereestablished in the 1970s and updated in the 1990s, atwhich time 5-year moving average quantity weightswere introduced. The Review Panel found littleevidence of current official use of parity measures, sothey were given less attention.


NASS should have a well-defined conceptualbasis for its indexes.

Well-defined concepts underlying and justifying priceand index measures are useful to develop a strategy for


integrating changes in supply and demand into thosemeasures on an ongoing basis. NASS should define whattypes of change it wishes to reflect in what time frame.Decisions about concepts of change do not have to bethe same for both average prices and indexes, butconsistency is helpful. New index concepts need to beresearched and tested in parallel with existing indexesto permit bridging the old and new concepts tominimize disruption of official and other important usesof the indexes. Concern about disruption of ongoingofficial uses of existing indexes should not be a barrierto researching and testing improved indexes.


A high-level initiative should be undertaken toevaluate official and public interest in thecontinued calculation and frequency ofpublication of parity measures.

A program should be initiated to contact users of parityprices and related measures to determine the desiredfrequency of publication and scope of the paritymeasures. This question should be discussed withappropriate USDA officials and Congressional staff. Theresponses by users to past changes made in the calcula-tions of parity prices should be reviewed as a guide tohow much flexibility NASS has within its legal require-ments. The goal should be to minimize constraintsimposed by the required parity calculations on improve-ments in the more heavily used prices received andprices paid measures. Finally, NASS should reduce itsdependence on the requirement to calculate parityprices as justification of its Agricultural Prices Program.The recent submission to the Office of Managementand Budget (OMB) for approval of the AgriculturalPrices Program indicates that the collection of the pricedata is covered by U.S. Code Title 7, Section 2204 and

is not dependent on thelegislation related to the parityprices.


NASS should evaluate alternative indexformulations and revise the methodology usedin construction of its indexes.

The Review Panel carefully examined the calculationsunderlying the current construction of price indexes inNASS. It appears that the indexes now produced arenot true price indexes. The Review Panel offersalternative calculations and provides examples toillustrate the potential for bias if price changes are notcomputed correctly. NASS should revise itsmethodology as soon as possible.


NASS should consider improving its datacollection to include probability-based surveysfor its indexes.

In order to make an inference about a definedpopulation, probability sampling should be used at allstages. For price index estimation, an overall measureshould be estimated by weighting together sub-indexesfor the finest level of detail desired for analytical orpublication purposes. Probability-based mechanismsshould be defined for selecting among all businesseswhich might sell or buy any of the items and forselecting specific items to be priced over time within aclass. For each item stratum, the characteristics that arenecessary to distinguish among all the items belongingto the stratum should be enumerated. The level ofdetail and extent of the “checklist” must be sufficient toinsure that a current price for the exact same item atthe same outlet can be obtained over time. The long-term problem of reflecting change in the universe ofavailable items and outlets (e.g., the introduction of newtechnology or entirely new item classes) can beaddressed by defining a schedule of sample rotation.While switching to all probability-based surveys forindex components may be infeasible in the near-termfor budgetary reasons, NASS should seek everyopportunity to move in that direction.


CO N C LU S I O N

The Review Panel concluded that the NASS AgriculturalPrices Program produces statistics and economicindicators that are critically important for many officialand public uses. The program needs renewed attentionto its purposes—a new vision and a reexamination of allits data collections and processes. A new vision for theprogram should be accompanied by full documentationand transparency of all aspects of the program, includingthe program’s purposes and underlying conceptualfoundations. Decisions on improvements and newdirections should be based on sound research, whichwill require strengthening the core research programand linkage with the larger research community. Theserecommendations are consistent with NASS’s history ofcontinued improvement and reputation for professional-ism and integrity.

By tradition, and by virtue of authorities delegated to it,NASS is USDA’s flagship statistical agency—one of ahandful of key statistical agencies in the Federalgovernment. These agencies provide the informationbase for much of public policy and for decisions whichdrive the nation’s economy. NASS’s status bringsresponsibility for leadership. One of those leadershipresponsibilities, as seen by the Review Panel, is to dealwith the issues related to NASS’s heavy dependence onprice data from other agencies, especially AMS and BLS.At a minimum, NASS should institutionalize a processfor open communications among the agencies involvedthat is designed to bring transparency, documentationand understanding to all aspects of price data originatingin one agency and used by others. Ultimately, thisprocess should reveal opportunities for each agency toimprove the quality of data it provides to others, whileimproving overall efficiency of the collective statisticalsystem.

NASS’s response to these recommendations andchallenges will determine the future relevance andusefulness of the Agricultural Prices Program. TheReview Panel is confident that NASS has the will,competence and commitment to respond—to do theright things and to do them right. It is time to begin todevelop a program of agricultural price statistics thatmeets the needs and opportunities of the 21st century.


P A R T 1 | Background and Overv iew


The collecting and reporting of agricultural prices by theU.S. Department of Agriculture’s (USDA) NationalAgricultural Statistics Service (NASS) include pricesreceived by farmers for their commodities, prices paidby farmers for the production inputs they use and priceindexes. The prices are estimated by surveying buyersof farm products and sellers of farm inputs and fromdata provided by other government agencies andsources. The price indexes are calculated from pricesreceived and prices paid data as described later. Whilethe price statistics have a long history, they competewith difficulty for resources with the more publiclyvisible and market-sensitive crop and livestockproduction reports, also issued by NASS.

Estimates of prices received by farmers for theircommodities and prices paid for their production inputshave long been a staple of USDA’s statistical activities.USDA reports that it began collecting and reportingprices received by farmers as early as 1866, shortly afterthe Department was established. The early reportscovered December 1 prices received for ten crops. Thecollection of data to estimate January 1 values of sixspecies of livestock began in 1867. Prices received (as ofthe first of each month) were collected in 1908 for eightcrops. During the next two years, monthly pricesreceived for livestock and poultry and their productswere added. Policymakers and analysts became moreinterested in data on commodity prices as they watchedthe turbulence in the agricultural markets after WorldWar I. The depressed farm conditions in the 1920s and1930s increased their interest.

Throughout the 20th century, USDA’s estimates ofprices received for farm commodities evolved steadilywith adjustments in collection methods, commoditiescovered and frequency and timing of reporting. Thedetails of this evolution are chronicled in numerousreports and publications.4

Early in the 20th century, interest emerged in finding asimpler collective measure of prices received byfarmers. Work in USDA’s Bureau of AgriculturalEconomics (BAE)—the predecessor agency to NASSand the Economic Research Service (ERS)—led toconstruction of the first index of prices received.Beginning in 1910, the index was based on previousyear prices for ten crops. A second index was publishedin 1918 incorporating livestock prices. The third seriesof indexes was published in 1921 and was based onprices received for 31 crop and livestock products. Thebase period for this series was August 1910 to July1914, and the weights were based on Census ofAgriculture (COA) sales for 1909.

As with prices received, the 20th century brought forthmany changes in the index of prices received, such asthe base period, the commodities surveyed and theweights. Throughout the century, NASS and itspredecessor agencies tried to maintain continuity ofindexes by extrapolating backwards any changes inconcepts, commodity coverage and weighting periods.

Chapter 1Histor ica l Perspect ive

4 USDA-ERS. Major Statistical Series of the U.S. Department of Agriculture: HowThey Are Constructed and Used. Volume 1: Agricultural Prices, Expenditures,Farm Employment and Wages. ERS Agricultural Handbook No. 671, April1990.

At the time of this Review, NASS reported that itpublishes estimates of monthly prices received byfarmers for about 60 important crop and livestock itemsand annual or marketing year average prices for about60 additional items. Marketing year average prices areestimated for states where monthly estimates are notreleased because of limited marketing. Prices for fruitand vegetable crops for processing are estimated on amarketing year average basis because most productionis sold under contract, and prices are not final until afterthe crop is delivered. NASS publishes the pricesreceived by farmers and the price indexes each monthin its Agricultural Prices.

The history of the collection and publication of statisticson prices paid by farmers and the calculation of an indexof prices paid is similar to that for prices received forfarm commodities. In an effort to quantify changes inthe purchasing power of farm production, measures ofthe prices of farm inputs and family living expenseswere developed. USDA conducted an inquiry on pricespaid by farmers in 1911, based on prices of 86 farmproduction and family living items. These prices paiddata were collected annually until 1923. Then, quarterlysurveys were conducted through 1936. In 1937, thesurveys were semi-annual, quarterly or monthly,depending on the item. Collection of separate prices forfarm family living items was discontinued in 1976 in lightof evidence that price changes for family living expendi-tures followed the pattern of the Consumer Price Index(CPI). Currently, NASS collects prices paid on 297 inputitems.

Analysts first used prices paid data to show thepurchasing power of individual farm commodities. Forexample, they compared the number of bushels ofwheat required to purchase a binder or the number ofdozens of eggs required to buy a dress. Obviously, thisearly use of the prices paid data did not reflect farmers’purchasing power of farm products.

A composite measure of price changes, the index ofprices paid by farmers, was first published in 1928 byUSDA-BAE. The index measured changes in a selection

of prices of items purchased byfarmers. Just as for the index ofprices received, the initial baseperiod was significant for the

index of prices paid. In an effort to find a period whenboth prices received and prices paid were stable andthe agricultural economy was judged to be “healthy,”BAE undertook an extensive analysis of commodity andinput price trends from 1891 through 1926. The period1905–1915 exhibited relative stability. The most stableperiod, 1910–1914, was selected as the initial baseperiod for the index of prices paid. Later, theAgricultural Act of 1949 adopted the 1910–1914 periodas the original base reference period for all agriculturalindexes used for U.S. farm support programs. (Since1950, the index of prices paid and the index of pricesreceived were put on the same base: January1910–December 1914 = 100.)

Since the 1930s, revisions have been made periodicallyin weights and the number of items in the index ofprices paid. In 1995, the index was re-weighted andreconstructed. The current index of prices paid has abase period of 1990–1992 = 100, as well as the1910–1914 base. The weights are now assigned to alagged five-year moving average of 84 (of 297) items forwhich data are collected by NASS. Currently, approxi-mately 70 percent of the weights in the prices paidindex are assigned to indexes of selected input itemsand services obtained from the Bureau of LaborStatistics (BLS) and other sources outside NASS.

Another component of the NASS Agricultural PricesProgram—parity measures—evolved from the farmpolicy debates of the 1930s, when the nation and theagricultural sector were suffering through the GreatDepression. As noted previously, analysts and policy-makers examined the relationships between pricesreceived and prices paid to try to understand changes inthe economic well-being of farmers and farm families.Parity prices for farm products were first defined by theAgricultural Adjustment Act (AAA) of 1933. The Actdeclared it to be the policy of Congress to reestablishprices to farmers at a level which would give agriculturalcommodities a purchasing power with respect toarticles that farmers buy, equivalent to the purchasingpower of agricultural commodities in a base period. The1910–1914 period was chosen as the base because itwas considered a relatively normal period with generallystable price relationships.


For many years, the parity concept was cited as anindicator of the need for government intervention inagricultural markets, but its operational effectivenessdeclined rapidly after World War II. Rising farm produc-tivity and the liberalization of agricultural trade in the1960s and 1970s led to more flexible farm supportprograms. These programs reflected the modernizationand industrialization of American agriculture. The mostrecent farm acts do not mention parity. However, theparity requirement is still embedded in permanentlegislation, although it is widely considered as anachro-nistic. Still many provisions of recent farm acts aretemporary suspensions of permanent legislation,meaning that parity prices continue to have a potentialrole in the determination of agricultural support policy.This potential ensures a continuing interest by theCongress and farm leaders largely in reference tocontingencies, such as the expiration of temporary laws,or a basis for setting farm support prices in case oftrade embargoes and other food emergency conditions.

Mathematically, the parity index is the ratio of thegeneral level of prices farmers pay for the productiongoods and services they buy today compared with thegeneral level of such prices in the 1910–1914 baseperiod. When the index of prices received by farmersfor the products they sell is divided by the parity index(which is essentially the index of prices paid plus wagespaid to hired farm labor, interest on farm indebtednesssecured by farm real estate and taxes on farm realestate), the resulting parity ratio provides an indicationof the per unit purchasing power of farm commoditiesgenerally in relation to the purchasing power of farmproducts in the 1910–1914 base period. At any givendate, the parity price for any agricultural commodity isdetermined by multiplying the adjusted base price ofthat commodity by the parity index. (Parity prices arecomputed under the provisions of Title III, Subtitle A,Section 301(a) of the Agricultural Adjustment Act of1938 as amended by the Agricultural Acts of 1948,1949, 1954 and 1956.) Since parity calculations werebegun in the 1930s, numerous studies have focused onmodifying the parity concept and the calculationsunderlying it.5 It is important to note that, because of

the many changes in the agricul-tural sector since 1910–1914, theparity ratio is no longer ameaningful measure of farmincome, farmers’ purchasing power or farmers’ welfare.

Two other historical developments may be pertinent tothis Review: First, in recent decades NASS’s agriculturalprice statistics have received less public scrutiny thancommodity yield and production statistics. Also,additional responsibilities were transferred to NASS,including the Census of Agriculture. As a result, agricul-tural price statistics, by default, have received fewerresources for maintenance and research.

The second development is that other USDA agenciesare also charged with estimating prices and pricestatistics for a wide variety of program and policypurposes. For example, the Agricultural MarketingService (AMS), under the Agricultural Marketing Act of1946, reports daily prices for a wide range ofcommodities and markets, using definitions and surveytechniques that are quite different from those used byNASS. The Farm Service Agency (FSA) collects data onprices from a variety of sources to administer a numberof programs. By some counts, there are now more thana dozen sources of price data within USDA, withvarious combinations of sources used by agencies forspecific program purposes.

The tight funding for NASS prices collection andreporting and the existence of multiple price reportingefforts within USDA raise issues about duplication, inef-ficiency and the potential for misinterpretation andmisuse. Communication and collaboration amongagencies producing agricultural price statistics aresubjects for further examination, especially since thereare numerous data issues involved that cannot be solvedby NASS alone.


5 Tiegen, Lloyd D., Agricultural Parity: Historical Review and AlternativeCalculations. AER-571. June 1987. U.S. Department of Agriculture, EconomicResearch Service.

Chapter 2

Object ives , Uses and Users of NASS Price Stat ist ics


OB J E C T I V E S

An examination of the uses and users of NASS pricestatistics and the wording of the authorizing legislationsuggest that the objectives of the program are twofold:to provide essential data to meet the requirements ofFederal legislation; and to provide data essential tounderstanding the economic well-being of the agricul-tural sector and its contribution to the nationaleconomy. The implementation of numerousCongressionally-mandated programs now depends ondata provided under the first objective. The fulfillmentof the second objective provides a public record onfarm prices, the value of farm production, farm income,farm expenditures and the relative well-being of farmoperators. Such data enable analysts and members ofthe public to track the performance of the farmeconomy and quantify the contribution of the farmsector to the larger economy.

US E S A N D US E R S

It is difficult, of course, to identify all the uses and usersof any publicly available body of statistics. This isespecially the case for NASS prices received and pricespaid statistics, which have been published for manyyears. It would seem to be a simple task to identifylegally-mandated uses, but that is not always easy either,especially where those legal mandates trace tolegislation enacted long ago and kept on a standby basis.It is more difficult to track uses of NASS price statisticsto meet legislated policy objectives that presume theavailability of the data. Then there is the dependence onprice data to produce other estimates, reports andanalyses for which there may not be a specific legal

mandate, but on which policymakers and others havecome to depend.

Additionally, countless people and businesses worldwideuse NASS price statistics directly and indirectly forpurposes that can never be fully known. Some fairlycommon uses are noted in the following examples ofuses and users:

Legal Mandates. The Agricultural Act of 1949mandates the calculation of parity prices, indexes andratios. Prices received and prices paid data areeffectively mandated because both are required tocalculate parity prices. Even though the most recentFarm Act does not specifically mention parity, the Act isa temporary suspension of permanent legislation—theAgricultural Act of 1949. Moreover, as long as there arefarm price and income support programs, legislatorsand farm organizations, among others, will exertpressure on USDA to continue to calculate parityprices.

The Food, Conservation and Energy Act of 2008, themost recent Farm Act, contains several mandates thatrequire the use of NASS price statistics. Similarly,fulfilling the mandates of other legislation requires theuse of NASS price statistics.

� NASS weighted marketing year (season averagefarm) prices are required to determine levels ofPrice Counter-Cyclical Payments (PCCP).

� NASS monthly prices received are used todetermine loan deficiency payments for peanuts.

� FSA uses a variety of data sources to calculate dailyposted county prices (PCPs) used to determineloan deficiency payments for grains and soybeans.

FSA does not use NASS monthly state level pricesreceived directly in their PCP calculations, butnotes them as one of several checks on the reason-ableness of their daily county estimates.

� NASS season-average farm commodity prices areneeded to calculate crop revenue under the newSupplemental Revenue Assistance (SURE) program.

� National marketing year average (MYA) prices orseason-average commodity prices received forspecified crops—averaged for the previous twoyears—are required to calculate payments underthe new Average Crop Revenue Election (ACRE)program.

� NASS price statistics are essential to ERS’sdevelopment of Congressionally-mandated annualestimates of commodity costs and returns.Congress has also requested that ERS set monthlyestimates of milk cost of production, for which themonthly index of prices paid, published by NASS, isrequired.

� NASS monthly and season-average prices of non-price supported commodities (e.g., fruits, vegetableand livestock products) are often used as a basis forallocating so-called Section 32 funds. For example,in response to low prices, the Secretary ofAgriculture may announce that USDA is going topurchase a specific commodity with Section 32funds for donation to domestic food assistanceprograms.

� The Risk Management Agency (RMA), whichoversees the crop insurance programs, uses NASSprices received series for over 40 fruits, vegetablesand other minor/specialty crops for determining theprice election that is used in establishing the insuredproducer’s guarantee. For tobacco and cherries,NASS prices received data are used for determiningindemnities, as well as the settling of claims. Pricesreceived are also used to settle claims under thenew Actual Revenue History (ARH) pilot insuranceprogram, used currently for cherries. The ARHprogram may be expanded soon to include other

crops like California citrus, drypeas, sugar beets and lentils.

Other Government Uses and Users. Officials fromseveral USDA agencies testified to the Review Panelthat they depend heavily on NASS price statistics toconduct analyses of policy and program options.

� Price data are used heavily by USDA’s commoditymarket Situation and Outlook program,coordinated by the World Agricultural OutlookBoard (WAOB) to analyze market conditions andforecast commodity supply and demand, carry-over stocks, domestic and foreign utilization andtrade.

� The indexes of prices paid and prices received arekey components of the data base for USDA’s FarmIncome and Financial Indicator Accounts producedby ERS.

� ERS uses the NASS index of prices received andindex of prices paid in the development of USDA’sannual indexes of agricultural productivity andproductivity growth.

� USDA’s Commodity Credit Corporation (CCC)uses NASS monthly and season-average prices toestimate and forecast CCC outlays, which areincluded in forecasts of the Department’s andFederal expenditures.

� The index of prices received for beef is used byUSDA and the Bureau of Land Management (BLM)in the calculation of grazing fees on public lands.

� Prices received for corn, soybeans and hay areused to calculate the National Average Dairy FeedCost, a component in the formula used toadminister a monthly, direct payment to dairyproducers under the Monthly Income LossContract (MILC) program. Payments are based ondairy feed costs relative to Class 1 milk prices.

� The Congressional Budget Office (CBO) usesNASS commodity price data to estimate likelybudget outlays for various agricultural programs,actual and proposed.

� The Bureau of Economic Analysis (BEA) in the U.S.Department of Commerce makes multiple uses ofNASS price statistics.

• BEA’s National Income Accounts use theprice data to estimate agriculture’s contribu-tion to Gross Domestic Product (GDP) and


national income and to estimate the realconsumption of food on farms and CCCinventory change components of GDP.

• BEA’s regional accounts use agricultural pricedata to derive the annual regional estimatesof farm proprietors’ incomes, which, in turn,are used by other agencies to administerprograms that distribute over $214 billion inFederal funds to state and local areas. Thesefunds markedly affect economic well-being insome 400 rural counties.

• BEA is currently examining the possibility ofusing monthly or quarterly data on pricesreceived from NASS to estimate farmproprietors’ income on a quarterly basis.

Public and Private Uses and Users. NASS states,“Estimates of agricultural prices … are an importantpart of the nation’s economic data base. They are usedby industry management, economists, farmers, farmorganizations, legislators and government officials foranalysis ….”6

Members of the Review Panel, staff from USDA-NASSand industry representatives provided the followingexamples:

� Private commodity and farm advisory servicestrack monthly price reports to anticipate levels ofgovernment support for commodity producers andto provide production and marketing guidance tofarmers.

� Farm input suppliers and manufacturers trackUSDA prices as part of the process of anticipatingsales, stocking inventory and supplies and planningfurther production.

� Land Grant University Extension economists useNASS prices to develop commodity budgets, builddecision models for farmers, calculate value of agri-cultural production for their states and calculateagriculture’s contribution to the states’ economies.Numerous other purposes were mentioned bythese users.

� Prices received and pricespaid (along with futuresprices and other pricesinformation) are sometimesused by tenants and landlords in setting land rentalrates.

� NASS’s index of prices received and index of pricespaid are reportedly used to adjust prices paid toproducers under long-term contracts with foodprocessors.

� Farm and commodity organizations use NASS pricestatistics for tracking the economic conditionsfaced by their members, advising their membersand developing policy proposals and lobbyingpositions.

� Most of the NASS Field Offices have close workingrelationships with their state’s Department ofAgriculture. They supply those departments withstatistics, including state commodity and inputprices, to support the state’s monitoring ofeconomic conditions within their agriculturalsectors.

NASS AG R I C U LT U R A L PR I C E S

PR O G R A M IM PAC T S

A brief review of programs and their administrativerequirements and of price statistics used by public andprivate entities suggests those statistics guide andinfluence decisions involving billions of dollars. Thus, theimpact of the NASS Agricultural Prices Program issignificant as illustrated in the following examples:

� NASS estimates of marketing year average (MYA)prices received will be used to calculate supportpayments to farmers under the Food, Conservationand Energy Act of 2008, the current Farm Act.Thus, the accuracy of these estimates is veryimportant as small errors in the estimates ofcommodity prices can be costly for taxpayers orcommodity producers. For example, producers ofcertain commodities are eligible for incomeassistance under several provisions of the 2008Farm Act.

Under the Price Counter-Cyclical Payments(PCCP) provision, income payments are madewhen market prices of specific commodities are in


6 USDA-ERS. Major Statistical Series of the U.S. Department of Agriculture: HowThey Are Constructed and Used. Volume 1: Agricultural Prices, Expenditures,Farm Employment and Wages. ERS Agricultural Handbook No. 671, April1990.

a range determined by legislated formulas. Forsome basic commodities, the PCCP payment ismade on the amount of production determined bymultiplying that commodity’s base acreage timeshistoric yields times 85 percent.

At the market price ranges at which farmers areeligible for payments, a one cent ($0.01) error inthe marketing year average (MYA) price estimatecan make a difference of approximately $85 millionin government payments for corn, $23 million forwheat and over $101 million for upland cotton.

Considering the possibility of larger errors acrossnumerous crops, it is clear that hundreds of millions oftaxpayers’ dollars depend upon the accuracy of NASSprices received estimates for farm commodities underthis single provision of the current Farm Act.

� Similarly, small changes in prices received estimatesfor corn, soybeans and hay can drive changes inFederal direct payments to milk producers thatamount to millions of dollars under the MonthlyIncome Loss Contract (MILC). NASS pricesreceived estimates for these three dairy feedingredients are a key component in the formulaused to administer the MILC program as paymentsare based on dairy feed costs relative to Class 1milk prices.

� As noted previously, BEA of the U.S. Departmentof Commerce uses NASS price estimates tocalculate regional farm proprietors’ incomes, whichbecome part of the formulas used by otheragencies to distribute billions of Federal dollars tostates and local areas, including some 400 ruralcounties.

� A large national food company reports using theNASS index of prices paid to adjust the prices itpays farmers for potatoes to offset changes in costsof production. These adjustments alone—based onNASS prices—can amount to several million dollarsper year.

In short, the financial well-being of millions of individualsand businesses is affected in significant ways by themany decisions, large and small, which depend on theNASS Agricultural Prices Program. The billions of dollarsin impacts driven by these decisions contrast sharplywith the approximately $5 million dollars that NASSallocates annually to produce the prices estimates.7

This contrast begs the question of whether NASS couldimprove its Agricultural Prices Program within thatresource base. The Review Panel hopes that this reportstarts to answer that question.


7 Information emailed by Kevin Hinsman, NASS, April 16, 2009.


The Review Panel identified several issues andchallenges as part of its examination of the NASSAgricultural Prices Program. Some issues are technical,some policy-related and others related to the nature ofagriculture itself.

CH A N G E S I N T H E AG R I C U LT U R A L

EC O N O M Y

Until recent decades, there may have been a perceptionon the part of economists and statisticians that pricedata were relatively accurate and easy to compile in anagricultural economy regarded as a close approximationof Adam Smith’s world of perfect competition. Morerecently, however, the dramatic increase in contractualsales and vertical integration and the pervasive effects oftechnology and intellectual property on the concentra-tion of input industries (e.g., seed and pesticides) havebrought increases in both the importance and difficultyof compiling good price data.

� The changing organization and structure of thefood industry, especially vertical coordination andintegration, have blurred the lines between stagesof the food chain. This raises questions about theidentification and meaning of “farm-level” prices.For example, there are no “farm-level” prices forchickens; the first prices established are those forprocessed chicken products at the wholesale level.

� Similarly, prices received for crops are collectedfrom “first purchasers,” not from farmers.Consolidation (of grain elevator ownership, forexample), integration of once-local purchasers intohigher stages of the food chain and proliferation oftypes and ownership of buyers make it more

difficult to identify when and where a sale takesplace.

� Numerous and increasingly sophisticated marketingcontracts are used for selling farm products, whichmake it difficult for NASS to determine the datesand prices at which a transfer of ownership takesplace. The challenge is that there is a separationbetween pricing and delivery (ownership transfer),which makes it difficult to determine theappropriate observation point for the transaction.Producers use a variety of pricing mechanisms andset prices for various portions of their products atvarious times throughout the production cycle.This can affect the month to which transactions areassigned and distort the resulting statistics. Thechallenge to NASS is to find ways to minimizeerrors in making such assignments.

� In some transaction contexts, it is not alwaysobvious what the producer is being paid for andtherefore whether (or how much of) the receivedprice is for an agricultural product or for amanagerial service. The case of broilers is oneexample (and some other livestock would besimilarly involved), but the issue extends toidentity-preserved production and marketingcontracts more generally where producers are paidpremiums based on compliance with certainmanagement activities over and above the price forthe actual commodity. Some contracts specificallyseparate the pricing of the commodity from thepricing of the attribute premium; others do not.How are these different prices and attributesreflected or captured in NASS price data?

Chapter 3

Current Issues and Chal lenges

� U.S. agriculture is now integrated into the globaleconomy. Global supply and demand, trade policiesand exchange rates affect commodity and inputprices. Trade issues and policies influence thedemand for, and use of, price statistics. Closerglobal coordination of trade and food policies(under WTO, for example) increases the value ofconsistency in statistical conventions and methodsfor purposes of cross-country comparisons. Shiftsin global consumer demands and internationalflows of technology drive changes in the attributesof agricultural products and inputs.

These changes in the domestic and global agriculturaleconomy have implications for the nature of pricestatistics most useful to the various users. They alsopose challenges to the collection and utilization of pricedata to produce price statistics.

CH A N G E S I N DE M A N D F O R COV E R AG E

Changes in the use of price statistics affect expectationsof users about the commodities and input items forwhich price statistics are available and which areincluded in published indexes.

� Collection of data for prices received now extendsbeyond the original basic commodities. Yet, not allcrops (e.g., many specialty and minor crops) areincluded. What criteria should guide the selectionof commodities for inclusion in prices receivedstatistics beyond those commodities required tocompute parity or those needed for administrationof government support programs? How much ofthe production of those crops should be included?

� Not all products and services used by commodityproducers are included in estimates of prices paidfor various reasons such as uniqueness, infrequencyof use and the expense of finding data. Is this animportant issue? What criteria should guide NASS’sdecisions on which production items to include inprices paid and the index of prices paid?

CH A N G E S I N AT T R I B U T E S O F

CO M M O D I T I E S A N D IN P U T S

As uses of price statistics become more sophisticated,and as technology brings more rapid changes in the mixof product characteristics, more attention is focused onexactly what is being priced.

� Factors such as quality, utilization, seasonal varietiesand movement of old and new crops affect month-to-month price changes. Thus, a comparison ofmonth-to-month price changes may not reflectchanges in prices for identical products eachmonth. This can lead to bias and error of theestimates of prices and price indexes.

� Many of the prices reported by NASS representproducts with a range of specifications that areaggregated into a single price estimate. Differencesin variety, grade-weight class, protein level andorganic versus non-organic are examples. Howdifferent or unique should a product be to receiveseparate treatment? For example, should corngrown for ethanol under a fixed price or premiumprice contract be treated as a separate commodity?The point is that corn with a given set of attributesmay be priced differently from corn with anothermix of attributes. In that case, as an example, pricechanges reported for generic “corn” could resultfrom a change in supply and demand or a change inthe mix of attributes being priced, or both. Genericprices may be valid for estimating aggregate farmrevenue, but price indexes require prices forcommodities with well-defined attributes.

� Similarly, the mix and quantities of input items andservices that farmers buy and the qualities andattributes of those items change over time,sometimes rapidly. For example, how does onecompare the price of cotton seed today—with allthe technology now embedded in such seed—withthe price of cotton seed in earlier times? Moreover,the rapid changes in technology and the prolifera-tion of attributes (choices) present challenges toNASS in terms of what input items to report andhow to ensure comparisons of prices of like itemsover time.


� Many varieties of fertilizers, pesticides and feedmixes are customarily prepared and sold. NASSgroups these items by category for reporting pricespaid and for weighting in indexes. The prices ofitems within a category may not change in concert.For example, prices of nitrogen fertilizers may berising rapidly while prices of other fertilizers maybe rising more slowly or declining. As a result, theaggregated fertilizer component of prices paid maynot reflect current market developments for, say,farmers who use mostly nitrogen. The usefulnessof prices paid statistics and indexes is affected bythe fact that changes in component prices andweights do not always reflect the price movementsof some items within that category. What criteriashould guide NASS in determining how far to go indisaggregating groups or categories of prices paiditems?

SO U R C E S O F DATA

In most cases, more than one possible source of dataexists for prices received or prices paid. Some of thesesources are other government agencies, particularlyAMS and BLS. Some sources are private pricereporters. Efficiency considerations and budgetconstraints suggest the need to evaluate thesealternative sources as substitutes to, or additions to,collecting data by NASS surveys.

� NASS now uses probability surveys to estimateprices for selected field crops, defined generically.AMS uses non-probability surveys to collect pricesof well-defined products in selected markets. Howimportant is it for NASS price statistics to beunbiased estimates with known precision ofpopulation means? How difficult would it be toevaluate the trade-offs between the strengths andshortcomings of NASS and AMS price data tominimize bias in price estimates? What criteriashould guide the choice or mix of data sourcesused?

� The demand for more complete coverage ofcommodity and input prices and the growingdependence of NASS on other agencies for data toprovide that coverage challenge NASS’s role as theflagship statistical agency of USDA. Currently,NASS lacks full information about, and understand-

ing of, the characteristics ofprice data from AMS andother sources and indexesfrom BLS. NASS also is notalways aware of changes in data and indexes fromexternal sources, which can affect the quality ofNASS price statistics. It appears that NASS hasbeen a rather passive user of external data, havingexerted little influence on the qualities, characteris-tics and timing of data provided by other agencies.NASS may have to assume more responsibility forinter-agency communication and coordination to:

1. encourage other agencies to develop datamore consistent with NASS’s needs,

2. become more informed of qualities andchanges in price data collected and/orprovided by others (to improve anddocument NASS’s own published statistics),and

3. help develop and improve the vision andgoals for its own price statistics program.

The changes in the economic and statisticalenvironment in which NASS now manages its programsmake the issue of NASS leadership very important.

TH E CH A N G I N G DE M A N D F O R PR I C E

STAT I S T I C S

Agricultural price statistics were first collected to informfarmers and to provide those interested in agriculturewith some general indicators of economic conditions inwhat was once the nation’s largest economic sector.Later, in the 1930s, the legislated requirement for paritycalculations added to the importance of collecting andpublishing price statistics. There were few othersources of price data, and little, if any, formal orinformal connections of USDA statistical programs toother Federal statistical programs and economicindicators. In those earlier decades, there were fewinternational statistical regimes or standard statisticalconventions. Today, there are many sources of data onagricultural prices, even within USDA. The inter-dependence of statistical indicators generated by NASS,BLS, BEA and others increases the pressure forstatistical compatibility and more standardized statisticalconcepts and conventions among the domestic


statistical agencies. The same is increasingly true for theinternational statistical community, of which NASS is animportant member. In this changed and more complexmarket for price statistics, the Review Panel feels it isimportant to take a fresh look at the needs for and usesof price statistics and the attributes of those statisticsthat are important to users. As examples, it is importantto address price statistics issues such as:

Transparency. Many data users want and need moreinformation about all aspects of data collection,processing, analysis, degree and sources of bias anderror, and index content, weights and construction. Thisapplies to external data as well as that collected byNASS. Transparency requires a full commitment todocumentation and ease of access to that documenta-tion.

Compatibility and Comparability. It is important toview the NASS Agricultural Prices Program today as anintegral part of the body of Federal statistics. Thus, it isimportant to think about consistent, or at leastcompatible, conceptual bases for the various statisticalproducts, including indexes, that become componentsof statistical products produced by other agencies.International comparability is also important. Again,transparency and widespread understanding among thevarious statistical agencies are critical to progresstoward compatibility.

Timing. What time period should commodity pricescover? What are the arguments for monthly estimatesrather than some other time period? Are preliminarymid-month estimates of prices received useful and towhom? Could the timeliness of monthly price estimatesbe improved and costs reduced by eliminating the mid-month price estimates? What uses and which userswould be affected?

Education. Once price data, or any other statistics,become regularly and publicly available, USDA andNASS have little control over their uses. For example,there are reports of published prices received being

used to set land rental rates by landlords even thoughcommodity prices are not necessarily an indicator of theprofitability of land use. Should NASS be responsible foreducating the public so they will know how the statisticsare created and how to interpret them?

Legal Constraints. Does the “parity” concept, hencethe computation of parity prices and indexes, have anycurrent validity? How would de-emphasis of the legalmandate for parity computations relative to greateremphasis on the mainstream price statistics programimprove the flexibility of NASS to address programchanges to meet today’s needs? This is an issue overwhich NASS has no direct control and would requireguidance from higher policy levels.

In summary, rapid changes in today’s world result inmany challenges to managing a major agriculturalstatistics program. They also provide context for thisreview of the NASS Agricultural Prices Program. TheReview Panel has addressed several of the most criticalof these challenges in the following chapters.



Several common and important themes emerged fromthe Review Panel’s deliberations on needed improve-ments in the NASS Agricultural Prices Program. For themost part, the common themes cut across allcomponents of the Program. The themes pertain to:

1. the value of improved transparency and documen-tation,

2. the critical importance of research,

3. the use of data from external sources,

4. NASS’s responses to changes in agriculture anddemands of modern statistical systems,

5. the importance of specificity in defining attributesof items for which prices are collected, and

6. improvement of index number construction(treated in Chapter 7).

The Review Panel also offers additional recommenda-tions that are more generally applicable to theAgricultural Prices Program.


NASS should increase transparency and documen-tation of all aspects of its Agricultural PricesProgram.

NASS’s data users should have easy access to under-standable, up-to-date information on data sources, datacollection methods, sampling methods, list complete-ness, processing and editing procedures, levels andsources of errors and data interpretation. Suchinformation should be updated annually, or more

frequently, as needed. All data revisions should beclearly identified to enable users to correct any seriesthey are using. This applies to statistics that NASSpublishes which are derived from data from otheragencies, as well as statistics estimated by NASSsurveys. Increasing transparency and documentation forthe Agricultural Prices Program would meet NASSrequirements as stated in its Information QualityGuidelines.

It is equally important that the purposes and conceptualbasis for price statistics be apparent to users. Forexample, prices received and prices paid data, whilerequired for parity calculations and for administration ofsome programs, are essential to calculation of real valueadded by the farm sector in the national incomeaccounts. This value-added concept is one determinantof the coverage of farm expenditures, as well ascoverage of prices received for the entire range ofproducts of the farm sector. When (if) calculations ofprice measures to conform with legal mandates conflictwith other conceptual bases (e.g., national incomeaccounting) or sound statistical theory (e.g., indexconstruction), those conflicts and how NASS addressesthem should be readily transparent. A major benefit toNASS of documenting the conceptual basis for pricestatistics is that doing so provides the Agencyopportunity and incentive to clarify those conceptualissues for internal management purposes.

NASS provided several excellent manuals describingsampling, data collection and processing. However,many users of statistics are more interested in a better

P A R T 2 | Reports and Recommendat ions of the Rev iew Panel

Chapter 4

Common Themes and GeneralRecommendations

understanding of what the data represent than on howthey are collected. The Review Panel’s suggestions onimprovements in transparency range from including aguide to how NASS incorporates various sources ofprice data to providing a citation index for sources.

Statistical organizations document their procedures inorder to communicate what they do and how they doit. Part of this is documenting an agency’s statisticalstandards. Some agencies have manuals of statisticalstandards to facilitate a bigger and more consistentcross-surveys perspective. Such communications can beinternal, such as employee training or programmanagement, or external, such as for legal reasons,answering questions from statistics users and recruitingstaff or to provide advice to or obtain advice from otherstatistical organizations. Documentation that discussesthe software used to carry out statistical procedures canbe helpful for internal communications about statisticalprocedures, but it generally does not facilitate effectiveexternal communications.

The Review Panel recommends that NASS preparedocumentation of its statistical procedures on a regularbasis for purposes of effectively communicating withexternal stakeholders. The documentation should befree of reference to software familiar only to NASS orspecialized data users.

The information about how price data are collectedshould be readily available to users and include the toolsto find the information (e.g., searches within the NASSwebsite). Although the statistics in the hard copyreports include references and notes with additionalinformation, none is provided with the statisticsavailable online. These supplementary referencedocuments and similar publications should be readilyavailable for download on the NASS website. Some ofthese documents could be placed on the “UnderstandingAgricultural Estimates” webpage, but others that directlyexplain how these data are collected should be availableon the website with the reports and the data querypage. All of the statistics available in the reports shouldbe made available for download to a spreadsheet.

In some cases, members of the Review Panel felt thatNASS’s formulas for calculating price indexes were notclearly defined. NASS should define all terms used incalculations and provide brief justifications for their use.

Several additional suggestions to NASS that mayimprove documentation and transparency follow:

1. Seek public comment before implementingsignificant revisions;

2. Publicize significant revisions in press releases;

3. Mark statistics that have been revised since lastpublished;

4. Consider providing a single comprehensive docu-mentation report that conforms to NASS’s owntransparency and quality-of-information guidelineson a regular basis;8 and

5. Consider providing leadership across USDAagencies to update and re-publish on a regular basisthe bulletin, “Major Statistical Series of the U.S.Department of Agriculture.”9


NASS should substantially strengthen its coreresearch capacity.

NASS has a longstanding reputation as a supplier ofdependable, objective and accurate statistics to theFederal government and to the American public.Continued prominence requires that NASS enhance itscapacity to conduct and gain access to research thatsupports its operations programs.

In its deliberations, the Review Panel found severalinstances where additional research was needed toaddress problems in data collection, coverage andprocessing, and index construction. A key research needis to evaluate the tradeoffs between the strengths andweaknesses of NASS and AMS price data for use inNASS price indexes. The Review Panel noted thatNASS’s research capability—once a source of public and


8 USDA-NASS. About NASS: Information Quality Guidelines. Retrieved January 17,2009, fromhttp://www.nass.usda.gov/About_NASS/Information_Quality_Guidelines/index.asp

9 USDA-ERS. Major Statistical Series of the U.S. Department of Agriculture: HowThey Are Constructed and Used. Volume 1: Agricultural Prices, Expenditures,Farm Employment and Wages. ERS Agricultural Handbook No. 671, April 1990.

professional confidence in the Agency’s statistics—hasbeen diminished over recent decades.

To the extent feasible, NASS should justify withdocumented research all substantial decisions aboutprice collection procedures, use of external versusinternal survey data and index construction methods.The documentation should include estimates of effectson the accuracy of the statistics produced. Documen-tation and research are necessary to assure the use ofup-to-date statistical methods and consistency acrossthe Agency. It is also important for Agency credibility, inthat it allows academics and others outside the Agencyto evaluate procedures and revisions.

NASS should consider augmenting its internal researchcapacity through external contracts and other arrange-ments that increase access to statistical expertise inacademia and elsewhere. This should be in addition to,but not a substitute for, enhancing internal researchcapacity. Cooperative research agreements wouldprovide opportunity for greater transparency andinteraction between a strengthened NASS researchcapacity and the broader research community. Inaddition, NASS’s staff should be encouraged to publishtheir research and otherwise maintain their involvementin professional statistical societies. Doing so will enhancethe reputation of the Agency and improve its recruitingcontacts.

Nearly all of the information NASS provided to theReview Panel was about operating the AgriculturalPrices Program, not studying and improving it. At theoutset of their deliberations, the Review Panelrequested copies of NASS’s research reports about theAgricultural Prices Program. NASS responded with justone report, and it was dated 1983.10

The goal for NASS’s research should be to improve theefficiency and statistical legitimacy of future operations,as well as the quality and usefulness of statisticalproducts. Part of the research should also involve thedesign and development of new statistical programs ormethods.

Another aspect of researchinvolves monitoring existingoperations and searching for waysto improve them. Some examplesare traditional survey assessment studies (coveragestudies, non-response studies, recordkeeping studies)and meetings of focus groups of data users, conductedon a reoccurring basis. Other possibilities are follow-upsurveys with samples of respondents so that NASS canask them questions about the survey process and its“fit” with how data are available to them and thesystems they use or plan to use. Such samples could bepartly random and partly judgment samples.

Monitoring existing operations permits a statisticalorganization to determine if there have been anychanges in the survey environment. If there have beenchanges, then the knowledge gained can be used toguide changes in the organization’s survey practices.

Surveys of users by NASS researchers would helpensure the statistics are not becoming less useful overtime.


NASS should accept full responsibility for howit uses, and explains the uses of, data fromother agencies for calculating prices and priceindexes.

Other agencies, principally AMS and BLS, are majorsources of data used in the calculation of pricesreceived, prices paid and price indexes. For pricesreceived, NASS uses probability surveys to collect pricedata for selected field crops. The prices for mostlivestock and livestock products and for most fruits,vegetable and specialty crops come from the AMS pricereports. The statistical qualities of those AMS data areeither unknown or not transparent. The AMS price datawere designed to provide the market with timely priceinformation at major markets. In the index of pricespaid, approximately 70 percent (48 percent for the basemonth of April and 72 percent for other months) of theweight is placed on items that come from BLS, AMS,ERS and DMRkynetec. It was not apparent to theReview Panel that NASS had full access to, and madefull use of, information on statistical properties of theseexternal data, the nature and timing of changes and the


10 House, Carol, et al. June 1983. Alternative Sampling Frame Construction forthe Prices Paid by Farmers Survey: An Overview. U.S. Department ofAgriculture, National Agricultural Statistics Service. AGES830516.

appropriateness of uses being made of the data. Neitheris it clear to the Review Panel that the implications ofchanges in external data are made transparent to theusers of NASS price statistics.

In support of this recommendation, the Review Paneloffers several specific suggestions:

1. NASS should establish an internal team that workson methodological and other issues related to theuse of prices from external sources. Many of theissues related to the quality of external prices aresimilar to those of the survey-based prices, e.g.,breadth of coverage, appropriate specifications ofitems priced and others. Given the overlap intopics, the team working on external data mayneed to overlap with staff working on survey data.Users of NASS indexes should know how externalprices are processed and used to obtain thoseindexes. The team working on external pricesshould be responsible for providing this transparen-cy. This transparency may help NASS staff and datausers identify opportunities for improvement.

2. To fulfill its responsibility as USDA’s flagshipstatistical agency, NASS should assume a leadershiprole in coordinating communications betweenNASS and other agencies producing and using pricedata that flow to and from NASS. There could begreat benefit from having NASS develop andmanage an on-going communication process thatfacilitates discussion among staff of the variousagencies about what each agency can do toimprove the overall usefulness of data to allagencies. The communications could focus on suchtopics as improving transparency and documenta-tion, data improvements that cost little but whichimprove their usefulness to others, improvingcompatibility and comparability of data and indexesamong agencies and improving notifications aboutchanges in data collection and processes.

NASS’s role could be informal or formalized byauthority of the Secretary of Agriculture.Participating agencies should also include those

outside USDA that supply or useNASS statistics. Such a formalizedcommunication arrangementwould assist with the transfer of

knowledge, expedited notifications and consistencyin reporting rather than relying on informal institu-tional knowledge that can be lost as people retireor leave employment. The information basegenerated by this coordination will also be usefulfor improving transparency and documentation.(See Recommendation 4–1.)

3. NASS should explore with AMS possible changes inprocedures that would facilitate improvedmeasures of the reliability of price estimates thatNASS makes based on AMS data. Such changesmight include increased sharing of sampling listsand information about sampling methods, responserates and omitted categories. Closer collaborationand consultation between the two agencies mightlead to changes in collection methods that wouldresult in more reliable estimates for either or bothagencies. Following are examples:

a. Additional steps might be taken to assure thatthe lists of buyers contacted by AMS are ascomplete as possible. This might involvemore efforts to build lists or more sharing oflists between the agencies.

b. The methods used by AMS to select buyersto be contacted might be reviewed. In somecases, procedures to change buyer contactssystematically might help to reduce bias.Records of buyers who refuse to participateneed to be kept, and procedures should bedeveloped to adjust for such omissions.

c. AMS may be able to add marketing locationsand quality categories to achieve more nearlycomplete coverage. Where this is impractical,methods could be developed to adjust foromitted product categories. Ongoing collabo-ration between NASS and AMS could allowAMS to consider NASS’s needs as changesare made and as opportunities arise.

The Review Panel does not wish to imply that NASS canor should tell AMS how to collect data or what data tocollect. Rather, it is being recommended that NASSexplore the possibilities with AMS. Such explorationseems desirable even if improvements in methodscannot be assured. For those commodities where pricereporting is mandatory, accuracy is likely high, in part


because all or nearly all of the population is covered. Itwould nonetheless be useful to have estimates ofaccuracy or reliability for both AMS and NASS priceestimates for those commodities. Mandatory reportingapplies only to some of the livestock and none of thefruits and vegetables. Moreover, there are examples ofFederal agencies adding statistical requirements tomandated data collection to fulfill other legitimateneeds.11 Developing new measures of reliability andimproving the reliability of NASS estimates based onAMS data will be a difficult task for NASS. AMS cannotbe expected to fully meet accepted statistical standardsfor surveys while reporting daily prices at a relativelydetailed level. Yet, it seems possible that ways can befound to attach meaningful measures of reliability toestimates derived from their data.

The point is this: if NASS is going to be successful inrevitalizing its prices program, it will be essential toengage AMS in the process. It is the responsibility ofNASS to lead that engagement.


NASS should undertake a comprehensivereexamination of its Agricultural PricesProgram to develop a vision for a system ofprice statistics relevant to the emergingglobal food and agricultural economy andconsistent with broader Federal and interna-tional statistical systems.

In general, the current set of published statistics appearsto have been built around a requirement first set in theAgricultural Adjustment Act of 1938 to produce a parityindex based to the 1910–1914 time period. The originallegislation did not specify what type of index should beformulated, how it should be computed, what itemsshould be included or what sectors of the economyshould be included.

NASS should look at its mission in the area of pricestatistics in the context of the US/Canada/Mexico

statistical system, EUROSTAT andFAO. With the introduction of theNorth American IndustryClassification System (NAICS) andtaking on the responsibility for the U.S. Census ofAgriculture, NASS should consider its publication goalsbroadly within NAICS–Sector 11: Agriculture, Forestry,Fishing and Hunting. For example, NASS shouldevaluate any content gaps in the PPI. Whereappropriate, NASS should work with BLS for it toexpand its coverage, produce particular sub-indexes andincrease its sample size.

Where NASS-produced indexes are deemed moreappropriate, NASS should consider producing priceindexes both conceptually and mathematicallycomparable to the BLS Producer Price Index, making itpossible to combine indexes from the two agencies toproduce a complete NAICS–Sector 11 index.

Changes in farm production and marketing methodspose challenges to estimating prices received and pricespaid. The emergence of new technologies and businesspractices in production and harvesting changes the mixand variety of inputs purchased and how they arepurchased. For example, some survey forms requestreporting the prices of animal feed in bags of certainweights, whereas most animal feed today is sold in bulk.This increases the complexity of coverage, weightingand collection of prices paid data. Changes inconsumption patterns and the growth in importance ofnew commodities require a continuous review of thecoverage and weights in prices received as well ascoverage of new inputs in prices paid.

The fact that NASS has field offices in most statescertainly helps keep the Agency abreast of changes inagriculture. However, the demands of day-to-dayoperations can easily prevent adjusting NASSprocedures to take these changes into account. TheReview Panel recognizes that NASS is aware of thisissue and suggests that the Agency may want to developexplicit processes for periodically assessing sectorchanges and their implications. The process couldinvolve the use of knowledgeable external industry andacademic persons.


11 As an example, IRS adds questions to business tax returns to facilitate theCensus Bureau’s use of tax return information to develop its universe frameand for using tax return data in lieu of census reports. Also, the FederalReserve Board has added questions to its mandatory Call reports to aid theBEA in handling certain financial transactions for the National Accounts. Thereare numerous other examples involving BEA.


NASS should address the issue of potentialbias in its price estimates caused by lack ofspecificity in defining attributes of items forwhich prices are collected.

NASS collects data on prices of items that are definedwith varying degrees of generality. This opens thepossibility of reported price changes coming from achange of attributes of the item or in the conditions ofthe transaction rather than from a change in the actuallevel of prices of items with precisely-defined attributes.If, for example, the proportion of high-moisture wheatis higher this year than last, the lower reported price isa function of quality change, not a price change. Of

course, ceteris paribus, farm income is lower. But thehigh-moisture wheat price is not the relevant price forcomputing a price index. The prices used in the indexesshould, theoretically, be based on a set of well-defineditems in terms of attributes. This is true for both pricesreceived by farmers for their products and prices paidfor input items. Several recommendations in this reportrelate directly or indirectly to this concern. (SeeRecommendations 5–1, 5–3, 6–1 and 7–3.) The problemis illustrated in NASS’s practice of collecting expenditureand quantity data for calculating prices of broadly-defined commodities, products and input items. Incontrast, AMS surveys collect actual reported prices formore narrowly-defined products, though not withprobability sampling. (See box: “What is a Price?”)

WH AT I S A PR I C E?The estimation of a price would appear to be a simple task: survey sellers or buyers and ask what they received or paid. Inessence, that has been the basis for collecting prices paid and received data in NASS and in its predecessor agencies formore than a century. Statisticians have developed sophisticated probability surveys that produce data to which well-knownstatistical properties can be ascribed.

But, items being priced have properties which must be considered in collecting price data. Those properties havedimensions that vary, such as item attributes, geographic location, packaging, timing of sales and services associated withthe sale or item. The “appropriate” price to be collected depends on the intended use of the data. For example, aprobability survey that averages all prices for a given commodity across attributes, locations, times of transactions andembodied services will produce a price that, when multiplied by quantity sold, results in an approximate estimate of grossreceipts. But, if the objective is to produce an index of prices received or to contribute to a national index of pricechanges, it will not be clear whether the reported price changes were actually changes in price levels, attributes, location,services or some mix of these.

Which provides a more accurate estimate of a change in the price level for a given commodity: a probability survey of avery broadly defined product (corn) or a non-probability survey of sellers or buyers in key markets using a consistent,precise item definition (number two yellow corn with specified moisture content)? The question cannot be answeredwithout more information about the details of both types of surveys. The ideal approach, from the standpoint of statisticalaccuracy, is a probability survey with a precise definition of the item attribute being priced. But, as a practical matter, itcannot be stated with certainty that NASS survey data provide more accurate estimates of changes in price levels than doAMS Market News data. We don’t know whether the NASS data represent price changes or attribute changes. We knowthe AMS data are more likely to represent changes in price levels, but we don’t know whether the changes represent thepopulation.

This issue is further complicated by the fact that NASS probability survey data and AMS Market News data are mixed inthe calculation of price indexes. As an example, NASS average price for “all corn” is mixed with AMS data for the price ofRed Delicious apples of a specified grade for selected locations in the construction of a prices received index.

It is not clear to the Review Panel what the resulting index represents.

Development of an improved Agricultural Prices Program for NASS requires sorting out these issues. Doing so will requirecareful analysis and research to determine the optimum program within given budget constraints and to estimate likely

improvements at the margin from additional resources. In this report, the Review Panel urgesNASS to identify and prioritize the objectives of the Agricultural Prices Program as part of its newvision for the 21st century. In that process, it will be necessary to address the price definitionissues raised in this report.


NASS may have two conflicting objectives in thecollection of price data. One is to collect prices that,when multiplied by quantity, give total revenue orexpenditures. This helps measure farm income, forexample. But, if the objective is to measure price levelchanges, then prices should be collected forcommodities or inputs with specific attributes that areheld constant over time, while having a method ofdropping outdated items and adding new items, as BLSattempts to do with the CPI. The Review Panel sees thisas a critical issue for NASS to address, as it involvesboth program objectives and accuracy of reportedstatistics.

The Review Panel offers three additional general recom-mendations for NASS to consider:


NASS should consider collecting andprocessing monthly prices received and pricespaid so that they can be published early in thefollowing month.

Both prices received and prices paid statistics are usedfor making agricultural projections and for administeringfarm programs. One of the most important uses ofNASS price statistics is for the monthly WorldAgricultural Supply and Demand Estimates (WASDE)Report prepared by the World Agricultural OutlookBoard (WAOB). The WASDE Report is widely used andhas a major influence on commodity markets. It isusually released on the 10th, 11th or 12th of each month.Currently, its price predictions rely on the full-monthprices from two months earlier and the mid-monthprices for the previous month. Obviously, any change inscheduling of reporting prices would require coordina-tion between NASS and WAOB (as well as other users)to assure that a new schedule would work as intended.The Review Panel expressed concern that the mid-month price is subject to point estimate samplingproblems and may not be representative of overall pricepatterns. This change would make full-month estimatesfrom the preceding month available for WASDE andother uses. Mid-month prices for field crops would nolonger be collected. Moreover, earlier availability ofmonthly average prices would be advantageous in theadministration of farm programs. For example, Milk

Income Loss Contract (MILC)payments could be made earlier ifNASS’s monthly average pricesfor feed were available earlier.


NASS should seek balance between qualityimprovement efforts and assuring consistencyof price statistics series over time.

Historical statistics on prices and price indexes arevaluable to policymakers, researchers and the generalpublic. They provide a longer-term perspective oneconomic and market issues of concern to all users.

Continuity in the price statistics series is important tomany users who follow price movements over time.Any changes in the methods used to collect data andcalculate price statistics should be made as often asnecessary to improve statistical quality. However, it isimportant that these changes be made deliberately andopenly. When estimation methods change for a series,reports should provide information for splicing the newestimates onto older series. This may call for recalculat-ing and publishing earlier segments of the series usingthe new method.

Consistency is required so the data are reliable ongoingindicators of price relationships, including spreads. Anychanges in data sources, quality characteristics or char-acteristics of data obtained from other agencies shouldbe made transparent to users, along with informationon how to interpret and value the changes. Much of thedata for prices received for livestock is derived fromAMS and other sources that may change their samplingand collection procedures over time. Therefore, to theextent that NASS is using these data series, it isrecommended that NASS monitor the sampling,collection and reporting of the data sources. Likewise,NASS should communicate this information to the usersor at least notify them of these changes with referencesprovided for the appropriate originating administrativeunits.

As an example, NASS’s Agricultural Prices reports have apage on “Reliability of Prices Received Estimates.” Thispage references that prices received are based onUSDA-AMS reports. It would help users if NASSidentified specific reports and even the variables within


USDA’S MU LT I P L E PR I C E RE P O R T I N G PR O G R A M S

While this Review focused on the NASS Agricultural Prices Program, it was undertaken in the context of the responsibili-ties and authorities of USDA and, to some degree, of the Federal government. Basic authorities and resources for NASSprograms flow through the Secretary of Agriculture, as do those for all USDA agencies.

NASS is one of several USDA agencies responsible for estimating agricultural prices. NASS has established a number ofspecific reports and estimating procedures for that purpose. Other USDA agencies with overlapping responsibilitiesinclude, but are not limited to: AMS which primarily uses price survey-based estimates from selected buyers; ERS whichsupports a number of very large surveys of important characteristics of farms and farm families; FSA which develops dailyprices for thousands of locations and dozens of commodities nationwide and which relies both on its own price surveys aswell as those developed by other agencies and by private firms; and FAS which reports on prices, production and tradeglobally. The Office of the Secretary and the WAOB use all of these estimates in the preparation of their own officialestimates and in their other many economic monitoring responsibilities, as do many other Federal, state and privateofficials.

The Review Panel deliberated both the extent to which price estimating responsibilities are scattered throughout USDAagencies and the relative effectiveness and qualities of the various ways prices are estimated. The Review Panel recognizedthat while the price reporting programs of other USDA agencies arose at various times in response to legitimate needs,the result today is a fragmented system that could benefit from a fresh vision that should guide a comprehensive review.Thus, while the programs of other USDA agencies are outside the purview of this Review and beyond the authorities ofNASS, the Review Panel felt that NASS—as the flagship statistical agency in USDA and in its advisory role to theSecretary—could provide informal leadership for a comprehensive review of all price reporting and price statisticsoriginating in USDA agencies or used by USDA agencies. Such a review could examine the policy objectives of the various

price statistics, and how the entire USDA program might be improved in terms of rationale,objectives, division of responsibility and accountability, quality of information and service to alllegitimate users, including Federal policymakers. The Review Panel encourages NASS to considerhow progress might be made on this topic.


the cited reports. Then users would be able to identifyat once the current status of prices for those series inbetween reports.

NASS must continue to seek balance between users’immediate needs and the longer-term goal of improvingthe quality of the price estimates. It became apparentduring the course of the Review Panel’s deliberationsthat maintaining a consistent series to administerexisting programs sometimes conflicts with thedevelopment of improved price estimates, which mightcontribute to better policies in the future. This canresult from a lack of strong research on what theoptimum price series should be. In cases where thestatistics are required for administering programs, butviewed as not meeting best practices, NASS shoulddevelop a new series to replace the deficient series andthen bridge the old and the new series to minimize

disruption to users in the program agencies. (SeeRecommendations 4–1 and 4–2.)


On its website and in its publications, NASSshould treat “agricultural prices” as a majorsubject and not as a sub-category of“Economics.”

Identifying “agricultural prices” as a separate categorywould facilitate user access to information about theprices data (especially online) and would result in NASScombining information now shown under several topics.Also, the resulting higher visibility would be consistentwith the Review Panel’s view that agricultural prices areimportant and should be supported with a new visionand clarity of purpose.


BAC K G R O U N D

The collection and publication of statistics on pricesreceived by American farmers for their commoditieshave a long history, as documented in Chapter 1. Aswith other price statistics published by NASS, theprincipal purposes are:

1. to provide a public record that contributes tounderstanding economic conditions in agriculture,

2. to provide statistics needed to implementprograms for which the Congress has providedmandates and authorities,

3. to facilitate analysis of and decisions about policy,and

4. to provide for calculating parity measures.

A more complete summary of the uses of pricesreceived statistics is included in Chapter 2.

NASS reports monthly and annual average prices only;thus, it is not the intent that these statistics would beuseful to farmers and farm product marketers in makingshort-term selling and buying decisions.

However, NASS price statistics and the income andvalue of farm product sales statistics derived from theprices may be of help in making longer-term investmentdecisions in farming and farm product marketing. Forsuch purposes, the broadly based state average pricesreported by NASS supplement the more specific pricesreported by AMS.

The primary users of NASS prices received statistics arethose concerned with farmers’ incomes and theperformance of the farm economy: policymakers,

administrators of farm programs, other USDA agencies,researchers and the public. Federal expenditures forsome farm programs related to crops are especiallysensitive to NASS’s estimates of prices received.

In contrast to estimates of prices received for crops,there are relatively few direct uses of prices receivedfor livestock statistics. Much of the use of pricesreceived for livestock statistics is for research andanalysis and for calculation of other measures, such asfarm income and parity prices. Some governmentprograms such as the Milk Income Loss Contract(MILC) program and the Wool and Mohair program domake reference to NASS prices received for livestock,but those statistics are usually used for general analysisrather than for implementing specific policies orpayments. These uses call for price statistics that areunbiased, reasonably precise and broadly representativeof all segments of U.S. agriculture.


NASS should review, update and improvecriteria for choosing between conductingprobability surveys and using price data fromAMS and other sources especially for thosecommodities not included in indexes.

The use of data from probability surveys and from othersources in the same agricultural prices tables and thesame price indexes raises questions about the compara-bility and accuracy of the two data sources. The surveyapproach used for field crops provides greaterassurance of complete coverage and allows measures ofprecision to be calculated. The use of price data fromAMS is less expensive, and the monthly estimates

Chapter 5

Prices Received for Farm Commodit ies

appear less likely to be affected by monthly and annualchanges in the attribute mix within a commodity.Currently, the choice of data collection method islargely determined by commodity characteristics andthe resulting marketing practices. However, it isconceivable that some field crops should be switchedfrom surveys to AMS data, if NASS can obtain satisfac-tory quantity weights for calculating state monthlyaverages. On the other hand, NASS surveys may bejustified for some fruits, vegetables or livestock toobtain higher levels of accuracy or to obtain estimateswith probability properties.

NASS should develop procedures for measuring thecoverage and accuracy of monthly price estimates basedon price data obtained from AMS and other sources.Such procedures should take into account completenessof coverage, bias and precision of estimates, users’needs for accuracy, the value of continuity in series,burdens on respondents and NASS’s costs. Bias andimprecision are likely problems. There may be ways towork more closely with AMS to make their price datacollection methods a better fit to NASS’s needs. Ofcourse, the price data needs of AMS’s clientele aremuch different from the needs of NASS’s clientele.AMS’s clientele needs daily prices for many differentgrades and quality specifications. AMS’s reporters havetime to contact only a limited number of buyers eachday and may rely heavily on those buyers who have thebroadest grasp of the market. In contrast, NASS pricesserve the needs of policymakers and administrators forwhom broad coverage and accuracy are especiallyimportant. The cost savings achieved by combining thetwo price data collection efforts for index constructionare significant, but added investments should beconsidered in some cases to provide the accuracy thatNASS’s clientele needs.

Monthly and yearly changes in the product attribute mixwithin a commodity may affect the average pricesestimated from NASS surveys. For example, theproportion of corn that grades No. 2 may vary betweenmonths or wheat protein levels may differ from year toyear. An analysis of relationships between NASS’s

monthly prices and averages ofAMS’s prices for specific gradesand qualities should help

determine the magnitude of this effect. For example,research could determine the correlation betweenNASS’s monthly price for corn and AMS’s average pricefor No. 2 yellow corn, and how the difference betweenthe two varies over time.

Whenever NASS publishes reports based on dataprovided by AMS and other organizations, NASS shouldtake the lead in documenting and reporting to users thedata sources and the methods used to construct thepublished estimates. This calls for formalizing relation-ships with the agencies supplying the price data. (SeeRecommendations 4–1, 4–2, 4–3 and 4–5.)


NASS should track and evaluate quantitativemeasures of prices received data collectionand processing operations to identify areas forimproving the processes and resultingstatistics.

The prices received program publishes monthlyestimates and an annual report. It is very common forsurvey staff, who must meet a monthly publicationschedule, to be almost exclusively oriented to thepresent, not to the past or future. Because of thecompressed processing cycle, as soon as the report forone month is released, it is time to start working on thereport for the next month. This leaves little time toassess whether current performance is the same or isdifferent from past performance. If it is unrealistic toexpect the same staff to handle the assessment, thefunction of tracking performance may need to beassigned elsewhere, perhaps to the research staff.

The Review Panel recommends that NASS track aspecified set of quantitative measures of the pricesreceived data collection and processing operations. Theprices received program currently tracks the estimatedcoefficients of variation (CVs) at the national level.Other quantitative measures that could be trackedinclude:

1. the proportion of reporters who fail to report inaccordance with NASS standards,

2. rates and patterns of item non-response and unitnon-response, and

3. edit failure rates for key items.


For each of these, the actual impact on estimates willneed to be assessed. Unit non-response, for example,may have a serious effect on some estimates, but notothers.

By tracking such measures, it will be easier for theprices received program staff to make comparisonsbetween the present and the past, depending on theextent of changes in the measures over time.

An example of a possible area of process improvementfor the prices received survey is a study of question 16on the Grain, Oilseed and Pulse Profile Questionnaireused by NASS. This question asks the respondent whichsoftware or company the firm utilizes to maintain itsrecords. The operation-profile interviewer manualstates that “[t]he purpose of this question is to possiblyreduce reporting burden in the future. NASS may beable to accommodate firms reporting in the monthlyprices received survey if NASS can determine howrecords are maintained.”12 A data-collection researchmethodologist, who is familiar with the literature oneffects of recordkeeping practices on response burden,could analyze the responses from the 2008 monthlyreports, develop an experiment and then report on thefeasibility of using information about a respondent’srecordkeeping software or service company to reducerespondent burden.

NASS should implement a comprehensive survey qualityassessment, review and improvement system thataddresses both sampling and non-sampling sources oferror. Performance measures (e.g., non-coverage, frameerror, non-response rates, response error and othermeasures) should be summarized, along with descrip-tions of the survey procedures, in regular methodologyreports. Periodically (perhaps every few years) a higherlevel summary and review of changes in these errorindicators should be prepared in a Quality Profile. Thistracking of error sources will help identify points in thesurvey process in need of error reduction efforts.

Error reduction may suggest changes in design, training,operations (including technological changes) andmanagement. Such ongoing error tracking, review andcorrections should contribute to timely and focused

process changes to improveprices survey data quality.Research reports and other docu-mentation of ongoing errortracking will also add transparency to NASS operationsand product quality. (See Recommendations 4–1 and4–2.)


NASS should review, update and improve thecriteria used to determine commodity, stateand attribute coverage.

Legal requirements to calculate parity prices call forcoverage of most U.S. farm products, although there isno requirement for what specific commodities toinclude. On the other hand, NASS has a specified list ofcommodities for which prices are used in the adminis-tration of several major farm programs. Coverage issuesarise most frequently among the commodities notcovered by farm programs, particularly the minor fruitsand vegetables and specialty crops. Another significantconcern is the definition of “commodity” or“commodity group” because of the variation inattributes within products that may have once beenconsidered “generic.”

The development and implementation of new coveragecriteria would serve to determine the overall scale ofthe price statistics program and which commodities andstates to include at the margin. Currently, commoditiesand states to be included for each commodity aredetermined every three years so that at least 90percent of aggregate U.S. cash receipts are covered,and each included commodity has 90 percent coverage.The 90 percent criterion is somewhat arbitrary. Itshould be subjected to sensitivity analysis to determinewhether and how large deviations from 90 percentwould need to be in order to make a meaningfuldifference in application. In principle, commodities,states and attributes should be added or dropped at themargin depending on whether benefits to users of thestatistics exceed the costs to NASS and respondents forgenerating the statistics. This cost/benefit comparison isdifficult to perform in practice because benefits areespecially hard to measure. Measuring benefits involvesestimating users’ gains from having more or better


12 USDA-NASS. 2008 Prices Received Operation Profile: Interviewer’s Manual.May 2008.

information. The need for consistency in a series is acomplicating factor, which implies that changes shouldnot be made without strong evidence that benefits willbe increased.

It seems best to start by examining commodities thatare on the margin for coverage or for being droppedfrom coverage. The first step would be to identify thespecial characteristics of each commodity, such as highproduction growth rates, movement of productionbetween states, high price variability or proposedgovernment programs. Next, identify the key decisionsthat would be affected by improved price information,such as decisions about investment in production,processing and marketing facilities and aboutgovernment programs. Third, develop and applyprocedures for quantifying the benefits from improvingthese decisions through the use of improved priceinformation. This calls for assistance from persons withknowledge of and experience in evaluating informationsystems. At first, it may be possible to do little morethan rank commodities for coverage at the margin. Asmethods are developed further, the goal would be todevelop approximate dollar measures of benefits thatwould be useful in determining the overall scale of theprice reporting program within a cost-benefitframework.



BAC K G R O U N D

Prices paid by farmers are collected not only forpurchased inputs, but also for interest rates, taxes andwages. NASS conducts surveys to obtain prices formany farm inputs, but also relies on other sources fordata. Feeder cattle and feeder pig prices are obtainedfrom AMS while ERS provides interest rates and taxes.Custom rates are obtained from a private marketservice, DMRkynetec, and BLS provides “otherservices” prices, specifically electric utility prices. Farmsupplies and repairs, auto and truck, and buildingmaterial prices are also provided by BLS. And, ofcourse, BLS provides the CPI. Machinery and tractor,feeds, fuels, fertilizers and chemicals, and seed pricesare collected by NASS via survey instruments. NASSalso collects wage rate and rent data. In sum, pricespaid come from numerous sources and are largely forinputs used to produce traditional, economicallyimportant crops and livestock and livestock products.

NASS’s surveys of prices paid are currently conductedannually. Until 2009, NASS surveyed prices for themonth of April; now NASS uses March prices. Theindexes of prices paid are reported monthly based onBLS data. Price data from AMS are used for some inputsoriginating in the farm sector (e.g., feeder cattle andfeeder pigs), whereas NASS uses its own prices for milkcow replacements and poultry. NASS uses either theBLS CPI or PPI index depending on the particular sub-index (e.g., feed, fuels, etc.) that is being estimated. Themonthly indexes are adjusted—benchmarked—to theannual NASS prices each year. NASS reviews thecorrelation of the BLS adjustment factors with theNASS prices, but problems have occurred. A related

concern is providing transparency about revisions,which can occur because of the annual benchmarkadjustments and because BLS data may be revised orcorrected. (See Recommendation 4–3.)

Changing Structure and Prices Paid


NASS should address the goals and principles forchanging the mix of inputs and selecting theattributes of items for which prices are to becollected.

Profound changes in the agricultural sector areoccurring. On average, farms have become larger andmore productive, but they are also more specialized butin varied ways. Organic farms are a growing marketsegment, and their input (and output) prices are likelymuch different from those for conventional farms.Organic soybean meal used in dairy feed is perhapstwice the price of non-organic meal, but littleinformation exists about the prices of inputs used onorganic farms. At the same time, many conventionalinputs are not used by organic farmers. There willalways be a need for clarity about what is beingmeasured (and not measured) by NASS’s programs.

Large conventional farms often receive discounts fromlist prices. They may pre-purchase inputs and useforward, futures or options contracts to acquire inputs.

The horticulture sector is another example of increasingdiversity. Horticultural products have been a relativelysmall part of the income of the farm sector, but havebeen increasing. Grape (wine) production is an exampleof a growth area; another is ornamentals. Still another

Chapter 6

Prices Paid by Farmers

area that we know little about is horse farms. Whatattention should be given to all of this diversity?

Input suppliers and the items sold change as well. Seedproduction is increasingly sophisticated and concentrat-ed. The attributes of seeds can change dramatically;seed corn today is much different from the seed corn ofjust a few years ago. The same can be said formachinery producers and dealers. The quality attributesof tractors and other machinery are changing rapidly.

It is the Review Panel’s understanding that NASS isresponsive to requests for collecting prices on newinputs, and they have added items to surveys when it iseconomical to do so. These “new” prices are notnecessarily used in the index of prices paid, butaccommodate the interests of potential users of theirstatistics. According to NASS, they collect prices for 35seeds, but use just eight in the index of prices paid.Likewise, they collect prices on 108 chemicals, but use19 in the index. At present, NASS collects prices for 297inputs, 84 of which are used in the index of prices paid.

Decisions about what prices to collect, what prices touse in the index and how to accommodate the changingquality of inputs are a part of the broad issue of howbest to represent input prices in an increasingly complexfarm sector within the confines of limited resources.Attributes of items for which prices are collected mustbe carefully defined to reduce bias in estimates of pricechanges, i.e., to distinguish between price level changesand changes in attributes. Also, without carefully definedattributes, NASS data do not facilitate analysis of theprice premium for, say, glyphosate-ready soybean seedor the impact of non-competitive pricing of inputs dueto market concentration.

If, as we believe, the most important uses of prices paidrelate to economic intelligence about farms (and not tocomputing parity prices), what prices should becollected? What sectors and types of farms should becovered? More generally, what principles should guidethe addition and deletion of items for which price dataare collected?


NASS should change to semi-annual or morefrequent surveys of prices paid.

Prices for some inputs have become more volatile.Recent examples are the prices of fuels, chemicals andfertilizers, which in turn depend on the prices ofpetroleum or natural gas. Given the volatility of pricesand the diversity of the timing of the purchase of inputsby farmers, a concern is that collecting prices annually isnot sufficient to capture the complexity of the pricespaid by farmers.

How frequently should prices be collected? March orApril prices may be relevant for major crops planted inthe spring, but not necessarily for winter wheat orlivestock-related inputs. Moreover, it is not clearwhether the combination of annual price surveys withBLS-based monthly adjustments can adequately capturethe monthly variability in input prices. A second surveyperiod, perhaps August, September or October, wouldlikely reduce the adjustments needed in reportedmonthly prices paid statistics to bring them in line withsurvey data.

Survey Design and Documentation of Methods


NASS should develop a research program inprices paid statistics that addresses criticalproblems in survey design, documentation andidentification of appropriate input items tocover. (See Recommendations 4–1 and 4–2.)

Survey research is concerned not only with imprecisionarising from sample selection and estimation, but alsodue to various non-sampling errors such as coverage,specification, measurement, non-response and withprocessing errors. The Review Panel suggests a need forfurther information in each of these areas. The ReviewPanel specifically agrees with Statistical Policy DirectiveNo. 3: Compilation, Release and Evaluation of PrincipalFederal Economic Indicators,13 which states that any

13 Office of Information and Regulatory Affairs, U.S. Office of Management andBudget. Statistical Policy Directive No. 3: Compilation, Release, and Evaluationof Principal Federal Economic Indicators. Revised September 25, 1985.Retrieved on March 4, 2009, from http://www.whitehouse.gov/omb/assets/omb/inforeg/statpolicy/dir_3_fr_09251985.pdf


economic indicators program must meet high standardsof accuracy and reliability.

The following discussion focuses on prices collectedfrom firms selling inputs to farmers, but, as notedearlier, prices for some inputs are also obtained fromother agencies such as AMS. The need to address issuesrelated to price data obtained from other agencies iscovered in Recommendation 4–3, but recommendationsmade herein generally apply to all of the prices paiddata, whatever the source.

Coverage

Conceptually, the universe for prices paid consists of allthe purchases of goods and services used for farmproduction by farmers.14 To obtain information aboutthe value of the transaction, each transaction must beuniquely associated with an entity capable of reportingthe value. Thus, to gain reporting access to such trans-actions, NASS constructs a sampling frame of retailersand dealers who supply farmers and have knowledgeabout those sales. However, construction of such asampling frame, so that it approaches completeness,could be cost and time prohibitive. The sampling framethat NASS uses consists of firms or merchants known tosupply the goods or services. Further, the items selectedto be surveyed are a subset of the totality of inputsfarmers use (and those used in the index of prices paidare a subset of those collected). Thus, NASS reducesthe conceptual total universe to a target population ofsales from retail outlets or establishments wherefarmers purchase farm production inputs in 48 of the 50states. Each state develops a list of selling firms usingvarious sources. From seller lists, a sampling frame isthen constructed, and firms to be surveyed are selectedinto the sample. This current process of framedevelopment creates what is called frame incomplete-ness; that is, some units in the ideal universe of

transactions would not have anychance of being reported becausethe firm with knowledge toreport is not included in theframe.

The prices paid data are obtained from a survey sampleof approximately 8,500 businesses with the firmsselected from the state-developed lists by type of itemsold. The prices paid series currently consists of almost300 items that account for approximately 90 percent offarm expenditures for the most commonly usedproduction inputs. Each year, six prices paid surveys areconducted:

1. agricultural chemicals,

2. farm machinery,

3. feed,

4. fertilizer,

5. fuel, and

6. seeds.

The target population is all retail establishments wherefarmers purchase farm production inputs. As alreadynoted, the firms are selected randomly from lists ofsellers by the type of item sold. Firms are asked toreport the price of the most commonly purchased itemor of the item with the largest volume sold. Noadjustments are made for changes in the quality or char-acteristics of the inputs.

Prior to 1980, the prices paid surveys were non-probability surveys having chronic non-response andcoverage issues. NASS—in its widespread list framedevelopment efforts of the 1970s—purchased variousbusiness lists and created a list frame that was used forsampling beginning in 1980. Sampling was a clustersample of counties in each state, followed by a simplerandom sample of firms within each selected county.Problems remained because the list was believed to beincomplete. It included out-of-scope firms (no farmrelated sales) and the auxiliary information (e.g., size offirm—information that can be used to make samplingmore efficient) was poor or nonexistent. Even today,concerns exist about the auxiliary information becauseNASS’s list updating process uses a screening survey toupdate “control data.” For previous surveys, non-


14 For certain commodities or services, such as feeder pigs purchased, samplingframes to gain access to reporters are aimed at farmers who then supply theprice data. For example, when farmers in major hog-producing states aresurveyed on the number of hogs and pigs on hand, they are also asked aboutthe number, weight and prices of feeder pigs purchased during the previousmonth. For months having no survey, price estimates for feeder pigs arebased on sales reported by AMS and by selected auction markets. Stocker andfeeder cattle prices are averages from actual transactions at auctions andthrough dealers. Sales of cattle weighing less than 800 pounds are used toestimate the U.S. price of stocker and feeder cattle.

respondents to the mail-out screener were notfollowed up and the non-response rate for thisscreening operation was not transparent. NASS staffindicated that, pending OMB approval, the nextscreening survey is planned for 2010, and it is intendedthat there be follow-up to non-respondents.15

In the early 1980s, NASS conducted research onalternative sampling frames for prices paid surveys.16

The researchers warned that maintenance of the then-current frame would be a burden because establish-ments change so often that the frame would need to beessentially reconstructed each year to maintaincoverage. As an alternative design, the research investi-gated the possibility of adding questions about whereexpenditures were recorded for given commoditygroups at the end of the Farm Production andExpenditures Survey (FPES) (now ARMS). Theresearchers made a compelling argument that thismethod drastically improves coverage (from an averageof 50 percent incompleteness).17 Further, it adjusts tobuying patterns as well as providing other desirableproperties of probability sampling and provides costsavings. They found that estimates from the then-existing procedures did not compare very well toestimates made from their more technically groundedprocedures.

The early research revealed promising options towardimproving sampling frames for prices paid through theuse of probability surveys. Continued modification,development and testing of these options wererecommended. The approach was tried for severalyears and then discontinued.18 No further research onthe topic is evident.

NASS research should focus on acceptable coveragelevels (both overall and by subgroups, e.g., sizes of

farms) relative to the totalpopulation of potentialrespondents, based on statistical

principles related to possible bias and low precision inthe estimates. Researchable questions include: What is asatisfactory coverage rate? Are the estimates biased?How should probability surveys be used to help controlcoverage error?

Specification

Prices obtained from the surveys of firms should reflectprices for items actually purchased. The quality of pricedata is dependent upon the reporter’s understanding ofthe price and volume concepts specified by NASS andthe ability to quantify and report on those concepts. Inthis subsection we address two aspects of specification:

1. item comprehension, and

2. item selection.

Regarding item comprehension, merchants are asked toreport prices for items meeting certain specificationsthat were “most commonly bought by farmers” or forthe item that was the “volume seller.” Items mostcommonly sold in one community may not be ascommon in another, and the volume seller in one yearor community may not be the volume seller in the next.

General specifications are provided for items beingpriced, with the aim of comparability in terms ofeconomic use and function. The specifications areintended to prompt the reporter to report prices ofcomparable items on successive surveys. As an example,a respondent will report the price for a new machinerymodel that had the same functional use as the one itreplaced, even if styling or technology has changed. Still,the guidance provided to respondents generally consistsof five or six bullets about the item, each bullet usuallyhaving ten or fewer words. These instructions areplaced in a simple box near the top of the form. It is notevident that reporters understand the instructions oreven recognize the need to review them carefully.There have been considerable recent advances inunderstanding how form design impacts respondentnavigation and understanding of survey forms. Expertassistance in form design for ARMS has demonstratedthe potential for improvement of survey forms.

While changes to the input item mix appear to beinfrequent, NASS has a mechanism for solicitingcomments about needed changes. According to NASS’s

15 Per information from NASS Statistician, Jennifer Sissom, February 27, 2009.16 See Alternative Sampling Frame Construction for the Prices Paid by Farmers

Survey–An Overview, House, C. et al, AGES830516, June 1983.17 Researchers estimated incompleteness of the frame in use at the time to be

over 70 percent for many commodities. Also, House, Carol, et al. June 1983.18 Based on conversation with NASS Deputy Administrator, Carol House, on

February 4, 2009.


Information Collection Supporting Statement (OMB No.0535–003), NASS frequently consults with interestedparties within USDA (ERS, FAS, AMS, FSA) and othergovernment agencies (Commerce Department, BLS)and holds regular meetings around the country toobtain feedback. While attendees may focus on theusefulness of the data to them (e.g., current coverage,frequency of release), they may not appreciate thesurvey methodological issues involved in engineering thedata product. Thus, it is reasonable to ask if thefeedback methods and questions at the user meetingsare appropriate to obtain user input on methodologicalconcerns, and whether attendees have the appropriateknowledge to address such questions. Understandingthese issues requires considerable training as well asfamiliarity with the data. NASS needs a cadre of suchexpert users (e.g., survey methodologists andeconomists) of price statistics.

NASS attempts standardization of questionnaires andhas a “Questionnaire Repository System,” as brieflymentioned in section 4 of the 2008 Prices Paid Surveysdocument.19 State-to-state comparability20 is notnecessarily ensured by standardized questionnaires,21

nor is it necessarily best for paper and internet versionsof forms to be “kept identical” as implied in section 4 ofthe 2008 Prices Paid Surveys document. While NASS hasa very knowledgeable and experienced staff involved inconducting the price surveys, a survey methodologyprogram for prices is lacking.

Sample Selection and Estimation

The current sample design, as distinct from coverage,has the consequence that a small number of reports areobtained for each item in each state (often well underten reports). Price and volume variability of the itemcan result in bias problems, especially with the smallsamples. This sampling strategy is rationalized bystatements that the actual price level is secondary to

measuring price changes, and abiased price level can providegood measures of price changefor index use. Of course, this isdependent on stability of the “most common” or“volume seller” items. Yet, to have all prices at theproper level would require large increases in samplesizes, and the data collection costs would far exceedcurrent resources. The tradeoffs between cost and thevalue of sample size increases should be explored.

While the survey methods are somewhat complex(various levels of stratification), the summaryprocedures for prices are merely a simple average ofresponses, and they are not designed to provideestimates of sampling variability. NASS’s stated goal is toprovide three to five usable reports per survey item perstate. The Review Panel did not verify that sampling atthis low rate could produce estimates with desirablelevels of precision.

Data Review

NASS has a distributed workload process in which theField Offices (FOs) are responsible for collecting,reviewing and loading data on the data processingnetwork. FOs perform initial editing. Reports from thestates are merged into regional files and processedthrough a more rigorous consistency check. Data arethen summarized to regional and U.S. levels, based onsurvey indications, by statisticians in the Washington,DC headquarters. Editing software and processing arereasonable, but the small number of reports makesstatistical editing methods ineffective. Thus, no wayexists to determine the magnitude of sampling and non-sampling errors or to determine if the volume of sales isadequate to obtain reliable prices.

Measurement Error—Reliability of Data

The Field Offices collect data by mail, telephone,electronic data reporting and personal interview for allprices paid surveys. The Review Panel found noresearch that assesses whether respondents reportprices based on actual transactions or on list prices.That is, how do the respondents “retrieve” therequested information? Do the prices come fromrecords or from memory? Further complicating theretrieval process, little is known about what adjustments


19 USDA-NASS, 2008 Prices Paid Surveys, Washington, DC. January 2008.20 Standardization does not mean that there are no differences. Questionnaires

used for prices paid surveys are regionalized to accommodate varyingpurchasing patterns across the country while seasonal questionnaires reducethe number of items on a given survey.

21 We note that some types of standardization can be very important, and statesshould not have the freedom to implement changes without the appropriateup-front methodological support and going through an appropriate approvalprocess.

22 Retrieved January 17, 2009, fromhttp://www.nass.usda.gov/About_NASS/Information_Quality_Guidelines/index.asp

23 House, Carol, et al. June 1983. Alternative Sampling Frame Construction forthe Prices Paid by Farmers Survey: An Overview. U.S. Department ofAgriculture, National Agricultural Statistics Service. AGES830516.

24 Retrieved January 17, 2009, from http://www.nsf.gov/statistics/sestat/technical-info.cfm

25 Retrieved January 17, 2009, from http://www.ers.usda.gov/Briefing/ARMS/

are made by reporters for discounts, rebates, loyaltypayments, credit, delivery, sales tax and other problemsassociated with modern transactions and marketingpractices. These market practices can vary greatly bytype of inputs. For example, rebates and loyaltypayments dominate marketing practices for pesticideswhile they may play an insignificant role for fertilizer.The effects of these practices should be captured in theprice data.

Non-response

Response rates by the voluntary respondents to theprices paid survey in 2008 varied:

� farm machinery, 76%� retail seeds, 81%� fertilizer, agricultural chemicals, 81%� feed, 86%� fuels, 88%.

At this point, it is unclear what causes these responserates to differ or what direction, magnitude or signifi-cance of any bias in the price averages results from this.NASS indicated in their OMB submission that non-response bias analysis has not been conducted for theprices paid surveys. Given the ad hoc design of thesurvey, it would be difficult to assess the impact of non-response bias separately from other sources of error.

Redesigning the Survey

Given the problem of incompleteness of the samplingframe and the small state sample size, significant qualityimprovement would require a full redesign starting withthe frame. There are a number of technical subjectsthat require further analysis and call for thedevelopment of a methodology program.

Transparency and Documentation

(See Recommendation 4–1.) While the general topics oftransparency and documentation are covered inChapter 4, this section addresses additional concernsspecific to prices paid.

Retrieved from its website, NASS's Information QualityGuidelines state—

“NASS will make the methodsused to produce information as

transparent as possible. NASS internal guidelinescall for clear documentation of data and methodsused in producing estimates and forecasts such thatit can operate a truly national programimplemented in 46 State Statistical Offices.Implementation of those guidelines ensures thereproducibility of disseminated information.

NASS estimates and projections are not directlyreproducible by the public because all underlyingdata sets are confidential. However, reproducibilitycan be evaluated through periodic reviews byoutside panels of technical experts and throughdocumentation of methods, assumptions, datasources and related information.”22

While internal documentation exists for conducting theprices paid survey, there are few survey design andmethodology materials (covering methods, assumptions,data sources and related information) upon which toassess many aspects of the overall quality of prices paiddata. If there were prior outside reviews of the pricespaid survey, the Review Panel was not made aware ofthem.

NASS has an online listing of past research, but it isdifficult to search. Only one research paper related toprices paid survey methodology (covering aspects suchas survey design, sampling error and non-samplingerror) could be found.23

This lack of readily available information about thesurveys hinders any effort to assess adequacy of theprices paid methodology. An example of a detailed webrelease of technical information for a complex datacollection effort is found for the Scientists and EngineersStatistical Data System (SESTAT) surveys.24 The ERSwebsite for the ARMS surveys is another example ofgood depth of survey information.25 Similar informationfor the prices paid survey program would betteraddress NASS’s Information Quality Guidelines.



BAC K G R O U N D

The Index of Prices Received (1910) and the Index ofPrices Paid (1928) had their origins in attempts todevelop simple, aggregate measures of prices receivedand paid. The indexes were part of the effort in theformer USDA-BAE to provide policymakers withmeasures of the well-being of farmers. A modifiedLaspeyres Index was adopted as the conceptual basisfor the indexes. Over the ensuing years, items(commodities and inputs purchased) were added toboth indexes to provide coverage that was morecomplete. Also, over the years, the weights of items inboth indexes were adjusted to reflect changes incommodities produced and inputs purchased. TheAgricultural Act of 1949 confirmed 1910–1914 as theoriginal base reference period for all agricultural indexesused for U.S. farm support programs.

The concept of parity evolved from the examination ofthe relationship between prices received and prices paidto shed light on the economic status of farmers. Parityprices are computed under provisions of theAgricultural Adjustment Act of 1938, as amended by theAgricultural Acts of 1948, 1949, 1954 and 1956. Theparity price is the price needed for a unit of acommodity to have an equivalent purchasing power asin the base period (1910–1914).

Parity measures for any agricultural commodity at agiven date are determined by a series of legally-definedand somewhat complex calculations. But, in simpleterms, the relationship between the index of pricesreceived and the index of prices paid is used to calculate

a parity ratio. The parity ratio provides an indication ofthe per unit purchasing power of farm commoditiesgenerally in terms of goods and services currentlybought by farmers in relation to the purchasing powerof farm commodities in the 1910–1914 (or later) baseperiod. Thus a parity ratio less than 100 indicates thatthe average per unit purchasing power of all farmproducts is lower than in the base period.

An adjusted parity ratio is computed and publishedquarterly which incorporates payments from Federalfarm programs.

The parity ratio was intended as a measure of pricerelationships. It is no longer a meaningful measure offarm income, farmers’ total purchasing power orfarmers’ welfare. The well-being of farmers dependsupon a number of factors other than price relationships,such as changes in production efficiency and technology,quantities of farm products sold and supplementaryincome, including that from off-farm jobs and Federalprograms.

The Review Panel found no examples of parity prices orother parity measures being used currently in theadministration of farm commodity programs. However,they are still valued by producer interests for purposesof policy discussion. (See Chapter 1 for an historicalperspective of price indexes.)

In addition to being required for the calculation of parityprices, the indexes of prices received and prices paidhave many other official and public uses, which aresummarized in Chapter 2.

Chapter 7

Price Indexes

Recent Changes26

Federal regulations require that NASS publish parityprices, the indexes and relevant price data in theirmonthly publication, Agricultural Prices. Until 1994, boththe index of prices received and the index of pricespaid, which are used to compute parity prices, had notbeen revised since 1976. The fixed weights used inconstructing these indexes became outdated as the mixof commodities farmers produced and the input itemsthey purchased changed. Weights and base andreference periods were updated in 1994, and newprocedures were put in place for constructing the priceindexes.

Similar changes in the indexes of prices received andprices paid were adopted to maintain consistency intheir construction due to their use in the computationof parity prices. The prior fixed-weight indexes wereconverted to five-year moving average weight indexes.Sources for the weights for the major items are farmexpenditure data from USDA’s annual Farm FinanceSurvey (now ARMS), the Farm Costs and ReturnsSurvey and farm cash receipts as published by ERS.They are updated annually. For the more detailed items,the weights are updated less frequently because manyof these weights were derived from the 1990–1992detailed ARMS expenditure data, which are no longercollected because of reduced funding and respondentburden concerns. Three-year moving average weightswere considered in order to keep the weights ascurrent as possible, but these proved to be too volatilefor the index of prices received since major droughtslike those in 1983 and 1988 caused wide swings in theproduction and in the corresponding prices receivedand cash receipts for many commodities.

For the index of prices received, monthly adjustmentsto the weights have also been developed to reflect the

normal marketing patterns of commodities during theyear.

Price indexes and parity ratios are now calculated using1990–1992 as a reference period in addition to therequired 1910–1914 reference period. The 1990–1992indexes are calculated using 1990–1992 as the baseperiod and moving average weights. The 1910–1914indexes are calculated by linking the 1990–1992 indexesat year 1975 with indexes that use 1910–1914 baseprices and quantity weights. Between 1976 and 1996,the base price and weight periods were 1971–1973 andthe reference period was 1977. The 1990–1992 periodwas a time of relatively stable overall prices receivedand paid. The three-year reference and base periodwere selected since they provide reference period andbase period prices for comparison purposes that arecloser overall to historical price trends than a one-yearperiod provides.

The fixed-weight indexes used for the interval from1910–1914 to 1975 are modified Laspeyres Indexes. ALaspeyres Index is the ratio of a weighted average pricefor the current period divided by the weighted averageprice for a base period. It uses expenditure weightsfrom the base period. The indexes that use 1990–1992as a reference period are calculated as weightedaverages of price relatives, where a price relative is thecurrent price divided by the 1990–1992 price. Recentmoving average weights are used. The 1990–1992based indexes differ from the Laspeyres Index and othercommon index formulas. The effects of thesedifferences on actual results and the degree to whichthe formulas used meet the standards desired in priceindexes are not well known. Taking recently observedexpenditures into account is clearly desirable, but theideal price index generally uses both current and baseperiod weights.27

It is still the case that attributes of commodities andinputs priced are not highly specific. This means that theitems being priced over time are not strictlycomparable. Prices reported reflect not only price levelsbut also the specific attributes. Hence, price changes


27 Stone, Richard. Quantity and Price Indexes in National Accounts, Organizationfor Economic Cooperation and Development, Paris, 1956, Section VIII, pp.106-107.

26 This section draws heavily on: Milton, B., D. Kleweno and H. Vanderberry.Reweighting and Reconstructing USDA’s Indexes of Prices Received and Paidby Farmers. USDA, National Agricultural Statistics Service, Economic StatisticsBranch Report No. ESB-95-01. Washington, DC. 1995.

reported may reflect attribute changes as well as pricelevel changes or both. The Review Panel feels that thecurrent treatment of attribute specificity is likely asource of bias in NASS price reporting.

Crops coverage was expanded in the index of pricesreceived, largely for greenhouse/nursery products,fruits, vegetables and specialty crops. BLS data are beingsubstituted for much of the prices paid data used in theindexes, such as new autos and trucks, buildingmaterials and farm supplies. For instance, BLS sub-component indexes like hand tools, power tools andbuilding and construction materials are replacing pricespaid survey information for dozens of small input itemssuch as hammers, wrenches, electric saws and roughboards. This use of BLS data with broader itemcoverage follows the 1976 index revisions thatsubstituted the Consumer Price Index for UrbanConsumers (CPI-U) data for the USDA family livingsurveys that included food, clothing and housing items.

The 1910–1914 indexes of prices received and paidhave been updated and revised based upon the changesin the newly constructed 1990–1992 indexes. Linkingthe prior and revised 1910–1914 index was done at thebeginning of 1975, the point in time when the priorweights of 1971–1973 were most current for use inconstructing the indexes. From the beginning of 1975forward, the changes in the 1990–1992 indexes areused to link forward the changes in the revised1910–1914 indexes.

Price Program Objectives

Clarification and documentation of program objectives,a common theme of the Review, are also important forprice indexes.


NASS should have a well-defined conceptualbasis for its indexes.

A new vision for NASS price statistics should includeexamination and definition of the most appropriate anduseful index construct to meaningfully interpret andtrack changes in the agricultural economy. Change is anever present factor in statistical measurement, andNASS should consider the conceptual basis for indexproduction most meaningful to today’s index uses.

Although it cannot ignore itslegislative mandate to produceindexes for parity calculations,NASS should consider thebroader picture and determine what it wants for thefuture and then determine how best to incorporate itsancillary needs (e.g., parity) into its production system.Consider, for example, the BLS decision to produce aprobability-based Consumer Price Index for UrbanConsumers in the 1970s. Knowing that the traditionalwage-earner CPI was fast losing its relevance to anincreasingly service-oriented economy, BLS beganproducing the CPI-U, while subsuming the productionof the Consumer Price Index for Wage Earners (CPI-W)into its computation systems.

Well-defined concepts underlying and justifying priceand index measures are useful to developing a strategyfor integrating changes in supply and demand into thosemeasures on an ongoing basis. Decisions about conceptsof change do not have to be the same for both averageprices and indexes, but consistency is helpful. Concernabout disruption of ongoing official uses of existingindexes should not become a barrier to researching andtesting improved indexes.

Parity

The Review Panel struggled with the issue of parity.Since parity prices and indexes are not used in theadministration of current farm programs and are rarelymentioned in NASS’s meetings with its data users, theReview Panel gave little attention to parity calculationsor to the appropriateness of alternative parity formula-tions. NASS reports little response to past changes inthe calculation of parity prices. The Review Panelagreed that the parity concept is a wholly inadequatemeasure of the well-being of farmers and the farmsector. Nevertheless, permanent law continues torequire the monthly calculation of the parity index,parity ratio and parity prices.



A high-level initiative should be undertaken toevaluate official and public interest in thecontinued calculation and frequency ofpublication of parity measures.

A program should be initiated to directly contact usersof parity prices and related measures to determine thedesired frequency of publication and scope of the paritymeasures. This question should be discussed at a higherlevel with appropriate USDA officials and Congressionalstaff. The responses by users to past changes made inthe calculations of parity prices should be reviewed as aguide to how much flexibility NASS has within its legalrequirements. The goal should be to minimizeconstraints imposed by the required parity calculationson improvements in the more heavily used pricesreceived and prices paid measures. Finally, NASS shouldreduce its dependence on the requirement to calculateparity prices as justification of its Agricultural PricesProgram. The recent submission to the Office ofManagement and Budget for approval of the AgriculturalPrices Program indicates that the collection of the pricedata is covered by U.S. Code Title 7, Section 2204, andis not dependent on the legislation related to the parityprices.

Index Construction

The following two recommendations address thecritically important topic of how indexes areconstructed. The topic came up repeatedly during theReview Panel’s deliberations because it is seen as centralto the quality and statistical integrity of NASS’spublished indexes.


NASS should evaluate alternative indexformulations and revise the methodology usedin construction of its indexes.

NASS’s formula for calculating item indexes follows:

Ij = Subcomponent Indexi = Item j = SubcomponentC = Current W = WeightB = Base

Item weights sum to one (1.0).

Prices received item weights are calculated from thefive-year average cash receipts data. The weights forprices received are adjusted monthly based on1990–1992 average monthly marketings.

Prices paid item weights are generated from NASS/ERSfarm expenditure survey data (ARMS).

While the general form of the index used is correct, theproblem lies in the particular

prices selected for .

From the discussion and materials presented by NASS,these current and base prices are often taken to beaverage prices for a given month or year. For example,the corn price for a given month is the average of cornprices across the different locations where corn is sold.Average prices are computed as unit values, total valueof sales divided by total volume sold or, equivalently,prices at each location averaged across all locations byusing current quantity shares of sales. From theperspective of constructing index numbers, what shouldbe used is an index of prices across those locations, nota weighted average of prices across those locations: Cijin the current period divided by a weighted average ofprices across those locations in the base period Bij .


The following two examples show the potential errorswhen average prices are used instead of index numbers:

Example 1: Compare the Index of Average Prices,

The appended note at the end of this chapter includes aspreadsheet that shows two examples:

1. a shift in demand from locations #1 through 4 tolocation #5 such that total demand is the same,and

2. an increase in overall demand with shift in salesfrom locations #1 through 4 to location #5.

With respect to Example 1, in equilibrium, prices shouldnot change at each location because they are separatedby fixed per unit transfer costs. This example assumes$0.20 per unit sold difference from the previouslocation. Location #5 is assumed to be the baselocation, so it has the highest price. As can be seen, thecalculations show that the index of average pricesincreases approximately three percent while none ofthe other index numbers changes.

Thus, in Example 1, the index of average priceserroneously indicates that prices have increased when infact none of the prices for any of the individual locationshas increased!

The second example indicatesthat all prices rise, but with stillfixed differences due to the factthat arbitrage will eliminate anyprice differences to the point where prices differ exactlyby transfer costs. The total volume sold increases from1,000 to 1,200 and a higher proportion of sales goes tolocation #5 than to the other locations. In this case, allfour index numbers show an increase in price from thebase period. However, the index of average pricesoverstates the prices index compared to the Laspeyres,Paasche or Fisher Indexes. Indeed, the latter threeIndexes all indicate very close changes of approximatelynine percent, while the index of average prices indicatesa price increase of 11 percent.

These two examples indicate that the index of averageprices will overstate price changes. Unfortunately, wecannot say with any confidence what the direction ofbias will be. Other examples could be constructedshowing that the index of average prices will understateprice changes.

While the examples indicate the bias is not all that large,it needs to be stressed that the error indicated here isonly for comparison between the current period andthe base period. Each time the index is calculatedfurther errors may be present so that, over time, theerrors may accumulate and it will be unknown what themagnitude of the errors will be.

It also needs to be stressed that construction of theseindex numbers at the most elementary level requires nomore information than what is being used now. All thatis needed are the prices, quantities and expenditures ateach location at a point in time. That is what is nowused to compute average prices.

Three indexes (Laspeyres, Paasche and Fisher Indexes)have been used, and the examples indicate that all threeindexes yield similar results. While evaluating alternativeindex formulations, NASS should also review researchdone in BLS and elsewhere to determine the appropri-ateness of Unit Value Indexes.

The analysis does not indicate that average prices arewrong, only that use of average prices in computationof index numbers is inappropriate. Therefore, use of


average prices when computing such indexes should beminimized to the extent possible.28

Price indexes should reflect, as closely as possible,prices for markets cleared in particular time periods.Using moving average quantities can understate oroverstate price movements. The usual justification forweights in Laspeyres Index numbers is finding normaltime periods. Selecting weights based on the notion thatone wants to smooth out fluctuations due to weather,for example, is not justified. It is also unclear how thefive-year moving averages are computed, whetherNASS uses the current period or lagged periods andwhich ones? More clarity is needed in definitions andcalculations.

Use of five-year moving averages may not be a problemfor prices received for livestock, given that asset fixityand specialization within the industry will generally limitthe proportional shifts in aggregate quantities. One issuewould be the incorporation of state level indexes andsimple averages in the parity prices calculation. A uniqueissue for livestock is that the location and numbers oflivestock have shifted dramatically over relatively shortperiods, and while prices tend to be nationallydetermined, there can be significant differencesbetween regions.


NASS should consider improving its datacollection to include probability-based surveysfor its indexes.

In order to make an inference about the definedpopulation, probability sampling should be used at allstages. For prices index estimation, an overall measure,such as the CPI for all items, is usually estimated byweighting together sub-indexes for the finest level ofdetail desired for analytical or publication purposes.

Further, stratified sampling is more efficient than simplerandom sampling. For these reasons, the universe ofinterest is usually subdivided into all-encompassing andmutually-exclusive classes or strata that correspond tothe sub-indexes of interest. Probability-basedmechanisms for selecting among all businesses whichmight sell any of the items and for selecting specificitems to be priced over time within a class should bedefined. For each item stratum, the characteristics thatare necessary to distinguish among all the itemsbelonging to the stratum are enumerated. The level ofdetail and extent of the “checklist” should be sufficientto insure that a current price for the exact sameitem/outlet can be obtained over time.

Sometimes a price for a selected item/outlet becomestemporarily or permanently unavailable. In the first case,some form of item imputation can be used; in thesecond, a new item should be selected by a probability-based process.

In addition, the long-term problem of reflecting changein the universe of available items/outlets (e.g., the intro-duction of new technology or entirely new item classes)can be addressed by defining a schedule of samplerotation. For example, one-fifth of the stratum could beresampled each year insuring that no segment of thepopulation is represented by a sample more than fiveyears old.

While switching to all probability-based surveys forindex components may be infeasible in the near-termfor budgetary reasons, NASS should seek opportunitiesto move in that direction.


28 In principle, indexes should be constructed from observations at the lowestlevel of aggregation where prices and quantities can be matched across years.In some cases this is by geographical location, in some cases by grade, weightclass or variety, and, perhaps, in some cases by individual buyer. Someaveraging of prices before they are used to construct indexes generally isunavoidable for practical reasons.

AP P E N D E D NO T E

Examples to Support Recommendation 7–3

The following spreadsheet contains two exampleswhich demonstrate the points made in the discussion ofRecommendation 7–3. The first example illustrateswhat happens when there is a shift in the locations oftransactions, but the price at each location stays thesame.

The second example illustrates what happens whenthere is a shift in the locations of transactions whileprices rise.



Diewert, Erwin. Index Number Theory: Past Progress andFuture Challenges. Paper presented at SSHRCConference on Price Index Concepts and Measurement,Vancouver, Canada. July 3, 2004

House, Carol, et al. Alternative Sampling FrameConstruction for the Prices Paid by Farmers Survey: AnOverview. U.S. Department of Agriculture, NationalAgricultural Statistics Service. AGES830516. June 1983

Martin, Margaret E., Straf, Miron L. and Citro,Constance F., editors, National Research Council.Principles and Practices for a Federal Statistical Agency:Third Edition. 2005

Milton, B., Kleweno, D. and Vanderberry, H.Reweighting and Reconstructing USDA’s Indexes of PricesReceived and Paid by Farmers. U.S. Department ofAgriculture, National Agricultural Statistics Service,Economic Statistics Branch, Estimates Division. ESBReport ESB-95-01

NASS Information Collection Supporting Statement(OMB No. 0535-003). Issued November 17, 2008

National Science Foundation. Technical Information.Retrieved on February 16, 2009, fromhttp://www.nsf.gov/statistics/sestat/technicalinfo.cfm

Stone, Richard. Quantity and Price Indexes in NationalAccounts, Organization for Economic Cooperation andDevelopment, Paris, 1956, Section VIII, pp. 106-107

Tiegen, Lloyd D. Agricultural Parity: Historical Review andAlternative Calculations. AER-571. June 1987. U.S.Department of Agriculture, Economic Research Service

U.S. Congress. Agricultural Adjustment Act of 1938. TitleIII, Section 301, (a)(1)

U.S. Congress. Agricultural Adjustment Act of 1938. Title7, Chapter 35, Sub-Chapter 11, Part A, Sections 1301and 1303

U.S. Congress. Agricultural Act of 1949. Title 1, Sections408 and 409

U.S. Department of Agriculture, Agricultural MarketingService. Northeast Milk Order Prices. RetrievedDecember 6, 2008, fromhttp://www.fmmone.com/northeast_order_prices/NE_prices_main_new.htm#component

U.S. Department of Agriculture, Economic ResearchService. Data Sets: Commodity Costs and Returns.Retrieved December 6, 2008, fromhttp://www.ers.usda.gov/Data/CostsAndReturns

U.S. Department of Agriculture, Economic ResearchService. Farm Income Briefing Room. RetrievedDecember 6, 2008, fromhttp://www.ers.usda.gov/Briefing/FarmIncome

U.S. Department of Agriculture, Economic ResearchService. Major Statistical Series of the U.S. Department ofAgriculture: How They Are Constructed and Used. Volume1: Agricultural Prices, Expenditures, Farm Employmentand Wages. ERS Agricultural Handbook No. 671, April1990

U.S. Department of Agriculture, Farm Service Agency.Retrieved December 6, 2008, fromhttp://www.fsa.usda.gov/FSA/webapp?area=home&subject=prsu&topic=mpp-mi

References


U.S. Department of Agriculture, Farm Service Agency.Milk Income Loss Contract Handbook, Short Reference:11-LD, Washington, DC. Retrieved December 6, 2008,from http://www.fsa.usda.gov/Internet/FSA_File/11-ld.pdf

U.S. Department of Agriculture, National AgriculturalStatistics Service. 2008 Prices Paid Surveys, Washington,DC. January 2008

U.S. Department of Agriculture, National AgriculturalStatistics Service. 2008 Prices Received for Crops Survey,Washington, DC. May 2008

U.S. Department of Agriculture, National AgriculturalStatistics Service. 2008 Prices Received Operation Profile:Interviewer’s Manual. May 2008

U.S. Department of Agriculture, National AgriculturalStatistics Service. About NASS: Information QualityGuidelines. Retrieved January 17, 2009, fromhttp://www.nass.usda.gov/About_NASS/Information_Quality_Guidelines/index.asp

U.S. Department of Agriculture, National AgriculturalStatistics Service. Agricultural Prices. Various monthlyissues

U.S. Department of Agriculture, National AgriculturalStatistics Service. Estimation Manual, Chapter 10:Livestock Prices. December 2001

U.S. Department of Agriculture, National AgriculturalStatistics Service. Guide to NASS Surveys. RetrievedJanuary 15, 2009, fromhttp://www.nass.usda.gov/Surveys/Guide_to_NASS_Surveys/Paid_and_Paid_Indexes/index.asp

U.S. Department of Agriculture, National AgriculturalStatistics Service. Standards for Statistical Precision ofMajor Probability Surveys. March 1, 2004. Policy andStandards Memorandum No. 45

U.S. Department of Labor, Bureau of Labor Statistics.BLS Handbook of Methods, Chapter 14: Producer Prices.Updated June 2008

U.S. Department of Labor, Bureau of Labor Statistics.The Experimental CPI Using Geometric Means. (CPI-U-XG) Retrieved on February 10, 2009, fromhttp://data.bls.gov/cgi-bin/print.pl/cpi/cpigmrp.htm

U. S. Government Accountability Office. LivestockGrazing: Federal Expenditures and Receipts Vary,Depending on the Agency and the Purpose of the FeeCharge. GAO-05-869, September 2005

Office of Information and Regulatory Affairs, U.S. Officeof Management and Budget. Statistical Policy DirectiveNo. 3: Compilation, Release, and Evaluation of PrincipalFederal Economic Indicators. Revised September 25,1985. Retrieved on March 4, 2009, fromhttp://www.whitehouse.gov/omb/assets/omb/inforeg/statpolicy/dir_3_fr_09251985.pdf

Administrative data: Information that is derived frommarketings, inspections, acreage contracted and/orcertified, assessments, ginnings and other public orprivate sources that has a direct relation to acommodity and is used to supplement survey data toprepare or revise estimates of average prices in a regionin a given time period.

Average price: Total value divided by the number ofunits.

Base period: The time interval for which the weightsfor a price index are calculated.

Bias: The difference between the average (mean) valueof a statistic over all possible sample outcomes versusthe population value of the parameter it is estimating.

Coverage: The extent to which the sampling frameincludes all the units in the target population and nounits not in the population. Missing units can createunder-coverage biases while out-of-scope units createsampling inefficiencies.

Coverage adjustment: A mathematical procedure thatattempts to correct the potential bias resulting fromimperfect coverage by the sampling frame.

Data editing: Procedures used to attempt to correctsampling data entries deemed in error or to providedata for certain missing entries.

Enumeration: A complete count.

Estimation: Process by which sample data are used toapproximate the value of an unknown quantity (aparameter) of a population.

Expanded price: An average price that has beenadjusted to account for incompleteness of samplecoverage, or to obtain the total value of a commodity.

Frame: A list of units or groups of units in a populationthat can be used for sample selection. Ideally, the listshould collectively cover the whole population, and theentries of the list must not overlap, so that every unit ofthe population belongs to exactly one entry in theframe.

Index: A statistical device that serves to indicate a valueor quantity of a given statistic at a given time, usuallyrelative to the value of that statistic in a base period =100.

Parity: A concept defined by the U.S. Congress in theAgricultural Adjustment Act (AAA) of 1933, madepermanent in the AAA of 1938 and reconfirmed in theAgricultural Act of 1949. The concept was designed tomeasure the purchasing power of products sold byfarmers relative to that in the base period 1910–1914.

Price index: A price index (plural: price indices or priceindexes) is a weighted average of normalized pricerelatives for a given group of goods or services, in agiven location or region, during a given interval of time.Price indexes are designed to help compare how theseprices, taken as a whole, differ between time periods orgeographical locations.

Given a basket of goods in time period t, the totalmarket value of transactions for those goods is

where is the prevailing price inperiod t and the quantity sold in time period t.


Glossary of Terms

Across two time periods, t0 and t1, one might betempted to calculate the ratio,

to measure the effects of price changes. However, thisdoesn’t distinguish between changes in quantities tradedfrom changes in prices. For example, if all prices aredouble in period t1 relative to period t0 but quantitiestraded do not change, P=2. On the other hand, if allquantities are double in period t1 relative to period t0but prices do not change, P=2 as well. Several priceindexes have been constructed to address this issue.Four common formulas are the Divisia, Fisher’s Ideal,Laspeyres and Paasche Indexes. Definitions of eachfollow:

Divisia Price Index: The Divisia Price Index is calculatedby the formula,

The Divisia weights usually are defined in terms of thevalue shares in the initial period,

or as the average of the value shares across the twoperiods,

The former weights are the same as in the LaspeyresIndex (see definition of Laspeyres Index), while thelatter are a midpoint approximation in discrete time tothe continuous time definition of the Divisia Index as thetotal differential of the logarithmic change in value. In both cases, we have Therefore, the

Divisia Price Index is a geometric average of the relativeprices, equivalently, the natural logarithm of the DivisiaPrice Index is the value share weighted average oflogarithmic relative prices,

Fisher’s Ideal Price Index: Fisher’s Ideal Price Index iscalculated as the geometric mean of the Laspeyres, PL ,and Paasche, PP , Indexes (see definitions of LaspeyresIndex and Paasche Index),

The inequality implies that .This is one motivation for Fisher’s Index, since PFneither systematically overstates nor understates theimpact of prices changes.

Laspeyres Price Index: The Laspeyres Price Index iscomputed as

where

is the expenditure or value share of the ith good in theoriginal time period, t0, and . Thus, theLaspeyres Price Index is a weighted average of pricerelatives using the initial period’s value shares asweights.

A Laspeyres Index of 1 states that an agent in period t1can afford to purchase the same bundle that waspurchased in period t0 given that total expenditure doesnot change. Due to substitution among goods atdifferent prices,


where we take the previous period’s bundle,as given and is

the utility function if the agent is a consumer and theproduction function if the agent is a producer. Thisimplies that the Laspeyres Index systematicallyoverstates the effect of price changes.

Paasche Price Index: The Paasche Price Index iscomputed as

where

is the pseudo-expenditure or pseudo-value share of theith good for period t0 prices and period t1 quantities, and

Thus, the Paasche Price Index is apseudo-value share weighted average of price relativesusing the final period’s quantities times the initialperiod’s prices to calculate the pseudo-value shares asweights.

Due to substitution among goods at different prices,

where the final period’s bundle of quantities,

, are taken as fixed. The left-hand side isthe denominator in the Paasche Price Index, so thePaasche Price Index systematically understates the effectof price changes, and for any change in pricebetween t0 and t1.

Reference period: The timeinterval used to calculate theindex numbers used for a priceseries. For example, the parityindex uses a 1910–1914 reference period. The indexesfor those periods are equal to 100.

Probability survey: A survey in which every element inthe population has a known and non-zero probability ofbeing selected for the sample.

Sample: A subset of the population selected as thegroup from which data will be collected.

Section 32 Funds: A provision of Public Law 74–320(1935) which earmarks an appropriation the equivalentof 30 percent of annual custom receipts to USDA tosupport the farm sector in a variety of ways.

Survey: A set of procedures designed to obtain datafrom some or all members of a population.

Total survey error: A mathematical description of theaverage (mean) of the squared difference between astatistic and the parameter it is estimating. This willinclude components of variability and biases arising fromvarious non-sampling errors, such as under-coverage,imputation and non-response.

Under-coverage: Failure to include all units belongingto the target population in a sampling frame.

Weight: Multiplicative factors that are used to combineprice data on individual commodities to construct anaggregate index. The weights reflect the proportion oftotal expenditures associated with the individualcommodities in an index, which sum to 1.0.


AAA: Agricultural Adjustment Acts of 1933 and 1938.

AAEA: Agricultural and Applied Economics Association(formerly American Agricultural EconomicsAssociation).

ACRE: Average Crop Revenue Election, an optionalincome support program for farmers authorized by theFood, Conservation and Energy Act of 2008.

ARH: Actual Revenue History, a pilot program ofUSDA’s Risk Management Agency.

AMS: Agricultural Marketing Service, an agency of theU.S. Department of Agriculture.

ARMS: the Agriculture and Resources ManagementSurvey conducted by NASS for ERS.

BAE: Bureau of Agricultural Economics, an agency ofthe U.S. Department of Agriculture that is thepredecessor agency to NASS and ERS.

BEA: Bureau of Economic Analysis, U.S. Department ofCommerce.

BLM: Bureau of Land Management, U.S. Department ofInterior.

BLS: Bureau of Labor Statistics, U.S. Department ofLabor.

CBO: Congressional Budget Office.

CCC: the Commodity Credit Corporation, an agency ofthe U.S. Department of Agriculture that provides fundsfor farm support programs.

C-FARE: The Council on Food, Agricultural &Resource Economics.

COA: Census of Agriculture. Conducted every fiveyears, the COA is a widely used source of datadescribing the nation’s production agriculture sector.

CPI: Consumer Price Index, published by the Bureau ofLabor Statistics.

CPI-U: Consumer Price Index for Urban Consumers.

CPI-W: Consumer Price Index for Wage Earners.

ERS: Economic Research Service, an agency of the U.S.Department of Agriculture.

EUROSTAT: the statistical agency of the EuropeanCommission. EUROSTAT maintains comprehensivestatistics on the European Union and its member states.

FAO: Food and Agriculture Organization of the UnitedNations.

FAS: Foreign Agriculture Service, an agency of the U.S.Department of Agriculture.

FSA: Farm Service Agency, an agency of the U.S.Department of Agriculture that services farmers withoffices in most counties.

GDP: Gross Domestic Product, a measure of the totalvalue added by the domestic production of goods andservices.

GNP: Gross National Product, a measure of the totalvalue added from all sources.

IPA: Interagency Personnel Agreement.


Acronyms

MILC: Milk Income Loss Contract, a Federal pricesupport program for milk.

MYA: Marketing Year Average price.

NAICS: North American Industry Classification System.

NASS: National Agricultural Statistics Service, anagency of the U.S. Department of Agricultureresponsible for a wide array of agricultural surveys andstatistics.

NMY: National Marketing Year or season-averagecommodity prices received for specified crops.

OMB: Office of Management and Budget, ExecutiveOffice of the President.

PCCP: Price Counter-Cyclical Program, a provision ofthe Food, Conservation and Energy Act of 2008 thatprovides payments to producers under specified marketconditions.

PCE: Personal Consumption Expenditures, a measureof aggregate consumer expenditures, published by theBureau of Economic Analysis, U.S. Department ofCommerce.

PPI: The Producer Price Index program measures theaverage change over time in the selling prices receivedby domestic producers for their output. The pricesincluded in the PPI are from the first commercialtransaction for many products and some services.

RMA: Risk Management Agency, an agency of the U.S.Department of Agriculture.

SESTAT: Scientists and Engineers Statistical DataSystems.

SURE: The Supplemental Revenue Assistance programof the Food, Conservation and Energy Act of 2008.

USDA: United States Department of Agriculture.

WAOB: World Agricultural Outlook Board, an agencyof the U.S. Department of Agriculture that coordinatesagricultural commodity supply and utilization estimates,published in WASDE.

WASDE: World Agricultural Supply and DemandEstimates, a monthly USDA report, published byWAOB, that contains global agricultural commoditysupply and utilization estimates.



The procedures for this Review were designed toassure objectivity, independence, freedom from conflictof interest and the highest possible level of professionalcompetence. In general, the procedures were similar tothose used by the National Academies of Sciences(NAS).

During the winter and spring of 2008, NASS officialsmet with the C-FARE Board of Directors to frame thecontent and objectives of the proposed Review. Theyagreed on a statement of task, scope of review and abudget. The two parties signed a contract on June 11,2008. After that, C-FARE began the process ofassembling a Review Committee and a Review Panel ofexperts. The strength of C-FARE, relative to theReview, is its familiarity with, and access to, social-science expertise in academia, government and theprivate sector. Utilizing that strength, the board selecteda Chair and Vice-Chair of the Review Committee. TheChair, working with the C-FARE Board, selected aReview Director. The Review Director and Chair thenselected 16 experts for the Review Panel based on twocriteria: an appropriate range of expertise and a balanceof perspectives.

The exact range and mix of expertise required on theReview Panel flowed from the major issue areas NASSidentified for review. Each member of the Review Panelwas chosen for his or her ability to address one or moreof these major issue areas. The resulting mixrepresented a variety of statistical and economicsspecialties. Members of the Review Panel were alsoselected to provide a balance of experience andperspectives. Since the question of how price data are

used and by whom was cited by NASS as an importantquestion, several members, having the prerequisiteexpertise, were selected with that experience, interestand perspective in mind. Thus, the Review Panelreflected academic, government and business interests.In the end, the Review Panel represented a high level ofexpertise and a balance of disciplinary, gender, ethnicand user experience and interest. Once selected, eachmember served as an individual expert and not as arepresentative of any group, organization or interest.Prior to confirmation, each potential member of theReview Panel was screened for conflicts of interest.Each candidate was verified by NASS as having nocontractual, advisory or financial links (past or present)to NASS that would constitute a conflict of interest.USDA employees were ineligible to serve on theReview Panel. Once the entire Review Panel wasassembled, the full list of members was checked a finaltime by the C-FARE Executive Committee and NASSfor conflicts of interest.

Prior to the beginning of deliberations of the ReviewPanel, the Review Director and Chair gave the chargeto the Review Panel, laying out individual and groupresponsibilities and expectations. They also provided atimeline with major benchmarks to assure completionof a satisfactory review by the July 1, 2009 deadline.NASS provided pertinent background materialspertaining to the Agricultural Prices Program and theissues under review. Finally, C-FARE publicized theReview and invited input from users of price data andother stakeholders. (See Appendix B.) The ReviewPanel began deliberations in Washington, DC, on

Appendix A

Review Procedures and Methods

October 15–17, 2008. The first day was devoted topresentations by NASS staff followed by questions fromthe Review Panel. On the second and third days, severalhours were devoted to input from USDA-ERS, a majoruser of price data, and from a panel of USDA andCongressional policy experts, who explained how theyused price data and spoke of issues pertaining to thosedata. All these information-gathering sessions wereopen to the public.

To focus more intently on specific issues, the ReviewPanel divided itself into sub-panels with a chair for each.Provision was made for addressing issues that did not fitneatly into the boundaries of any single sub-panel. Bythe end of the first meeting of the Review Panel, thesub-panels had made a first cut at defining problems andissues around which they would deliberate and, possibly,make recommendations. Between the formal ReviewPanel meetings, sub-panel chairs and members workedboth independently and as teams via conference callsand email. All draft materials from all sub-panels wereshared by email with other sub-panel chairs, the ReviewDirector and Review Committee Chair. One purpose ofsharing was to identify overlaps and gaps in issuescoverage. Another was to identify issues of commoninterest. Sub-panel members also held occasionalconference calls.

To assure full access to information needed by theReview Panel, NASS assigned a liaison, who respondedquickly and completely to requests. NASS was notprovided access to draft materials arising out of sub-panel deliberations. However, the NASS liaison wasasked to review the preliminary and final manuscriptsfor any errors of fact. Prior to the second meeting ofthe Review Panel in Washington, DC, in February 2009,drafts of key parts of the final report were circulated toall members of the Review Panel. These included draftsof deliberations and recommendations by the sub-panels. At this second meeting, the Review Panelscrutinized, debated and refined these drafts. Shortlyafter the February meeting, sub-panel chairs forwardedrewritten drafts of their reports to the Review Director.

During March, the Review Director merged the sub-panels’ drafts into a complete and internally consistentreport. Several themes common to the overall Reviewwere identified. Recommendations related to thosecommon themes and other general recommendationswere consolidated into a separate chapter of the reportto eliminate duplication. The resulting comprehensivemanuscript was edited. The resulting edited draft wasreviewed again by the Review Panel. It was sent to fourexternal reviewers at the end of March 2009. At a finalmeeting of the Review Panel on April 23, 2009, theReview Panel considered external reviewer commentsand any final concerns about the report. The ReviewDirector reminded the members of the Review Panel oftheir right to file a minority report on any topic whereagreement could not be reached. The Review Panelreached consensus on all major issues, and there wereno minority reports.

When all external and internal comments wereaddressed, a layout specialist completed preparation ofthe manuscript in June. On May 14, 2009, NASS officialswere briefed on highlights of the forthcoming report. Asrequired by the contract, a camera-ready manuscriptwas provided to NASS prior to July 1, 2009. NASS willpublish the report without comment and release it tothe public.



As part of its assessment of the NASS Agricultural PricesProgram, the Review Panel solicited input from repre-sentatives within the public, private and academiccommunities. Efforts to contact price statistics stake-holders for public comment included email campaignsand individual phone calls to agricultural economists andother professionals in the agricultural community.Forums for public comment were held at the 2008Annual Meeting of the Agricultural and AppliedEconomics Association (AAEA) (formerly the AmericanAgricultural Economics Association) in Orlando, Florida,and the first meeting of the Review Panel on October16, 2008, in Washington, DC. Those who could notattend one of the open forums were invited to submitwritten statements.

Those who commented included representatives offarm and commodity organizations, governmentofficials, policy analysts, commodity market analysts anduniversity researchers. Their data interests varied, butall wanted continuation of prices received and pricespaid and the indexes for both. Interest in parity pricesand the parity index was mixed, with most supportcoming from commodity organizations. Most interestcentered on coverage issues (commodities and pricespaid items), accuracy of data and a clear understandingof what the data represent. Interest was also expressedin improved timeliness of data availability. Some specificexpressions of interest and concern included:

� Concerns about how contracting and verticalintegration in the food chain affect accuracy andweights of “farm-level” price data.

� Questions about what a “farm-level price” reallymeans given rapid changes in the structure ofagriculture and the food chain and the growingdominance of large farms and large buyers.

� The need to update and expand the list of fruitsand vegetables included in prices received. Somenew Federal risk management programs, especiallypilot revenue insurance programs, may requireprices received data on more specialty and minorcrops.

� Questions about the costs and usefulness ofcalculating parity prices and indexes. However,commodity organizations strongly support thecalculation and publication of parity prices.

� Strong interest on the part of government analystsand policy officials in maintaining and improvingprices received and prices paid statistics for a hostof official uses. (See Chapter 2.)

� Strong support for price data on the part of USDAand Congressional analysts for policy analysis andanalyses of commodity markets.

� Interest on the part of one commodity organizationfor more state-level reporting of monthly prices ofvarious types of a particular grain.

� The importance, as seen by commodity organiza-tions, of continuing to report monthly prices paidand prices paid indexes for the current list ofcategories and sub-categories of inputs, because ofincreased instability in input markets.

Appendix B

Stakeholder Input to the Review

Verbal presentations weremade by the following stakeholders:

William McBride, Senior Agricultural Economist, for Katherine Smith (Administrator, USDA-ERS)

Larry Salathe, Senior Analyst, Office of the Chief Economist,USDA

Craig Jagger, Chief Economist, Agricultural Committee, U.S. House of Representatives

Joy Harwood, Director, Economic Analysis Division, USDA-FSA

Written comments werereceived from the followingstakeholder organizations:

USDA-Economic ResearchService

USDA-Risk ManagementAgency

U.S. Department of Commerce,Bureau of Economic Analysis

USA Rice Federation

Participants in the forum onthe Review of the USDA-NASS Agricultural PricesProgram at the AAEA AnnualMeeting held in Orlando,Florida, July 2008:

Academic Community

John Brandt, North Carolina State University

Jerry Fletcher, West Virginia University

Gail Cramer, Louisiana State University

Steve Turner, Mississippi State University

John Nichols,Texas A&M University

Ron Plain, University of Missouri

Timothy Eagan, Iowa State University

Government Agency Community

Kevin Barnes, USDA-NASS

Lars Brink, Agri-Food Canada

Mark Denbaly, USDA-ERS

Carrie Litkowski, USDC-Bureau of Economic Analysis

Michael Steiner, USDA-ERS

Farm Organization Community

Bob Young,American Farm Bureau Federation



Corinne Alexander is Assistant Professor ofAgricultural Economics at Purdue University. She servesas an Extension Specialist in the area of grain marketing.Her goal is to assist farmers and agricultural businesseswith the marketing of their grain both in commoditymarkets and in specialty markets. Dr. Alexanderreceived her Ph.D. in Agricultural Economics fromUniversity of California, Davis, with fields in agriculturaleconomics and resource economics. Her researchgenerally focuses on interactions between members ofthe supply chain, with a particular interest in contractualrelationships. Her current research examines howfarmers’ production decisions interact with theirmarketing decisions. In two specific projects, she isevaluating the costs and benefits of an on-farm qualityassurance program developed at Purdue, and examininghow European regulations preventing the import ofnon-approved transgenic crops have affected Indianafarmers’ decisions to adopt Bt corn resistant to cornrootworm.

Walter J. Armbruster is President Emeritus of FarmFoundation. Throughout his early career in USDA andhis 30-year tenure with Farm Foundation, he has reliedon NASS prices and other statistics to develop anunderstanding of the economic and policy issues thatacademic, industry and government participants mustaddress to maintain a competitive U.S. agricultural andfood production and marketing system. He has servedas President of the American Agricultural EconomicsAssociation, the International Food and AgribusinessManagement Association, and the American AgriculturalLaw Association. He has served on numerous research,

education and policy advisory committees and iscurrently a member of the National AgriculturalResearch, Extension, Education and Economics AdvisoryBoard. Dr. Armbruster is a Fellow of the AmericanAgricultural Economics Association, the AmericanAssociation for the Advancement of Science and theInternational Food and Agribusiness ManagementAssociation and is a Distinguished Agricultural Alumnus,Purdue University. He received his B.S. and M.S.degrees from Purdue University and his Ph.D. inAgricultural Economics from Oregon State University.

Johnny Blair has served as Principal Scientist andSurvey Methodologist at Abt Associates, Inc., since2001. Previously he spent 30 years in academic surveyresearch centers, including serving as the ActingDirector of the Survey Research Center at theUniversity of Maryland at College Park and the Managerof Survey Operations for the Survey ResearchLaboratory at the University of Illinois at Urbana-Champaign. Mr. Blair has conducted research in anumber of areas, including sampling for rarepopulations, within-household respondent selectionmethods for random-digit dialing surveys, measurementerror in proxy reporting, data quality in converted-refusal interviews, and the design and analysis ofcognitive interviews for pre-testing.

Arlin Brannstrom is a financial expert, professionaleducator and researcher specializing in farm financialmanagement, accounting and farm business planning andanalysis. He has extensive experience in teaching andconducting workshops for professionals in agriculturalmanagement and finance. Mr. Brannstrom has been the

Appendix C

Biographies of the Review Panel

Academic Vice President for the American Society ofFarm Managers and Rural Appraisers since 2006. In1972, Mr. Brannstrom received a B.S. with Honors inAgricultural Economics; in 1973, he received an M.S. inAgricultural Economics and in 1991 an MBA in Financeall from the University of Wisconsin-Madison.

Brian Buhr is Head of the Applied EconomicsDepartment and also serves as the E. Fred Koller Chairin Management Information Systems at the University ofMinnesota. Professor Buhr works in the areas ofcommodity marketing with an emphasis in livestockmarkets. He has worked extensively with commoditymarketing groups on risk management, value addedmarketing and the economic impacts of policy andtechnology. Current research includes analysis of theeconomic value of traceability in agribusiness and theincentive mechanisms that can improve product qualityand conduct of market participants. Professor Buhr isalso conducting policy research on issues of animalwelfare, the impacts of biofuels on the livestock andmeat industry and the economic impacts of animaldiseases in livestock. Recent publications have appearedin the American Journal of Agricultural Economics, Journalof Agricultural and Resource Economics, AgriculturalFinance, Review of Agricultural Economics and the Journalof Food Distribution Research. Dr. Buhr received hisPh.D. from Iowa State University.

Ismael Flores Cervantes is a statistician at Westatwith 16 years of experience in sample design andselection, variance estimation and data analysis. He hasworked on several, large random digit dialing (RDD)surveys, such as the National Survey of America’sFamilies. He has been the lead statistician of theCalifornia Health Interview Surveys since 2003. Histopics of interest are weighting as sampling of rarepopulations. He has also published on weightingadjustments for non-telephone households in RDDsurveys, and he was a contributor to the Advances inTelephone Survey Methodology (Wiley Series in SurveyMethodology). He was a consultant for the World Bankon survey design issues and currently is a consultant tothe Mexico Department of Education in Mexico City on

the sample selection andweighting of education surveys.He has conducted workshops in

Variance Estimation, Sampling and Weighting inColombia and Mexico. He was an American StatisticalAssociation associate fellow at USDA-NASS from1991–1992. He received his M.S. in Statistics from theUniversity of Texas at Austin, 1991, and a B.A. inElectrical Engineering from the Universidad de lasAmericas in 1989.

Gail L. Cramer is currently Professor and Head ofthe Department of Agricultural Economics andAgribusiness at Louisiana State University. He was theL.C. Carter Chair Professor in the Department ofAgricultural Economics and Agribusiness at theUniversity of Arkansas from 1987 to 2000. He attainedhis Bachelor’s degree in 1963 from Washington StateUniversity, his master’s degree from Michigan StateUniversity in 1964, and his Ph.D. in AgriculturalEconomics from Oregon State University in 1967. From1967 to 1987, Gail Cramer was Assistant Professor,Associate Professor and Professor in the Department ofAgricultural Economics and Economics at Montana StateUniversity. He has received numerous teaching andresearch awards, both domestic and international. He isespecially noted for his research in wheat and ricemarketing. He has published more than 200 journalarticles and other publications in the general area ofgrain marketing. His rice research has taken himthroughout the world, and he has presented seminarson his research in more than a dozen countries. Hereceived the International Award from Gamma SigmaDelta in 2003 and was selected a Fellow of theInternational Agribusiness Management Association,IAMA, in June of 2003. In addition, he was selected forthe Lifetime Achievement Award by the SouthernAgricultural Economics Association, SAEA, in 2002.

Cathryn S. Dippo retired as associate commissionerof the Office of Survey Methods Research at BLS, U.S.Department of Labor. While at BLS, she chaired theFedStats R&D Working Group and the CurrentPopulation Survey Redesign. She started the NationalScience Foundation/American Statistical Association/BLSSenior Research Fellow Program in the mid-1980s andthe BLS Behavioral Science Research Center in the late1980s. An office holder and member of severalstatistical societies, she has published a number ofarticles and has served as a referee for various statistical


journals. She was a member of the Committee on SocialSecurity Representative Payees, National ResearchCouncil, which prepared the following book forpublication in 2007: Improving the Social SecurityRepresentative Payee Program: Serving Beneficiaries andMinimizing Misuse, and other National Academiescommittees. She received a Ph.D. in MathematicalStatistics from George Washington University.

Ronald S. Fecso, Chief Statistician, U.S. GovernmentAccountability Office. His role at GAO is to support andextend the work of GAO teams and contribute to theoverall quality of the information GAO supplies toCongress. Ron came to GAO from the National ScienceFoundation (NSF) where he served as Chief Statisticianfrom 1998 to 2006. At NSF, he was responsible for thetechnical supervision of and confidentiality issues relatedto the survey and statistical activities of NSF’s Divisionof Science Resources Statistics. He has contributed tothe development of OMB directives and guidelineswhile serving on the Federal Committee on StatisticalMethodology and the Confidentiality and Data AccessCommittee. Formerly, he was a Senior ResearchStatistician for NASS and a Study Director for theNational Research Council’s Committee on NationalStatistics (CNSTAT). He was elected fellow of theAmerican Statistical Association, served on ASA’s Boardof Directors for the 2006–2008 term, and is a past-president of the Washington Statistical Society. He hastaught sampling at USDA’s Graduate School and atGeorge Mason University. His degrees are inMathematics from Rider and Mathematical Statisticsfrom the University of Rochester.

Richard (Dick) Heifner is a retired agriculturaleconomist who has devoted his career to evaluatinggovernment agricultural programs and farmers’marketing strategies. His work is represented bynumerous journal articles, bulletins, and administrativereports produced over some 40 years. As a graduatestudent at Iowa State University, he examined effects ofprice support programs and soil bank programs. As anassistant/associate professor at Michigan StateUniversity, he concentrated on farmers’ marketingproblems, including grain storage strategies, processingplant location, and the potential for electronicmarketing. During his 29-year tenure at USDA, mostly

at ERS, Dick helped to evaluatethe potential for farmers to usefutures markets, futuresregulatory programs, dairyprograms, grade standards for fresh produce, proposalsto support farmers use of agricultural options contracts,crop insurance programs, and other Departmentprograms. He was among the first to show how futureshedging can be optimized using the tools of stockportfolio analysis. As senior staff economist at AMS, heled a major study on fruit and vegetable marketingorders. Dick grew up on a farm in Iowa. He holds a B.S.in Agricultural Education (1956) and a Ph.D. inAgricultural Economics (1963), both earned at IowaState University. He served six years in the U.S. Armyon active duty and in the reserve. He was a director ofthe American Agricultural Economics Association in1979–1982 and received USDA awards for superiorservice in 1975 and 1985.

Jeffrey T. LaFrance is a professor in the School ofEconomic Sciences at Washington State University. Healso has been a faculty member at Montana StateUniversity, the University of Washington, the Universityof Arizona, and the University of California, Berkeley.His research interests are broad, including agriculturalpolicy, microeconomic theory, the economics ofconsumer behavior and nutrition, the economics ofnatural resource use and the environment, dynamiceconomic systems, and econometrics. Honors andawards include Outstanding Ph.D. Dissertation, Qualityof Research Discovery, Outstanding Master’s Thesis,Best Journal Article and Distinguished Fellow from theAAEA; Outstanding Master’s Thesis, OutstandingPublished Research, Outstanding Professional Serviceand Distinguished Scholar from the WAEA; and a SeniorFulbright Research Scholarship from the U.S. Council onInternational Education. He has received grants fromthe USDA, EPA, DOD, NSF, the States of Arizona,California, Montana and Washington and other agencies.He has a Bachelor of Science in Economics (1977) andMaster of Science in Applied Economics (1979) fromMontana State University in Bozeman and a Ph.D. inAgricultural and Resource Economics (1983) from theUniversity of California, Berkeley.


John E. Lee Jr. is retired Professor and Head,Department of Agricultural Economics, Mississippi StateUniversity. Previously, he spent 32 years at USDA-ERSin Washington, DC, the last 13 of which he wasAdministrator of the Agency. His fields of interest andwriting are agricultural policy, credit and evolution ofthe domestic and global food system. He has authoredmore than one hundred professional publications,papers and presentations. He continues to be profes-sionally active with consultation and contract work. JohnLee received his undergraduate and M.S. degrees fromAuburn University and his Ph.D. in Economics fromHarvard University. He is a Fellow of the AmericanAgricultural Economics Association, past President ofthe AAEA Foundation and recipient of the SAEALifetime Achievement Award.

William Motes is the Chief Economist of InformaEconomics, Inc., where his primary responsibilitiesinclude client consulting, economic policy studies, inter-national commercial client consulting and internationaldevelopment projects. Formerly, he was a principalmember of Economic Perspectives, Inc., a Washingtonconsulting firm. He spent more than 25 years in Federalgovernment service, most recently as director of policyanalysis for the Secretary of Agriculture (1979–1981).He was associate director of USDA’s Budget andProgram Evaluation Office, was legislative assistant foragriculture for U.S. Senator Dick Clark and wasDirector of the Economic Development Division ofUSDA-ERS. He holds degrees in Agricultural Economicsfrom Kansas State University and a doctorate inAgricultural Economics from Iowa State University.

Robert Parker is a nationally and internationallyrecognized expert in economic accounting concepts andmethods and in U.S. economic statistics. Since retiringfrom Federal service in 2005, he has been consultingwith government, non-profit and business clients. Forthe Federal government, he served his last five years asChief Statistician of the Government AccountabilityOffice. Previously he spent 30 years at BEA where heserved as Chief Statistician, Associate Director forNational Economic Accounts and Chief of the National

Income and Wealth Division. AtBEA he was a major contributorto BEA’s switch in quantity and

price index number formulas from Laspeyres to Fisher’sIdeal. Prior to moving to BEA, he served nine years atthe Census Bureau with responsibility for the EnterpriseStatistics Program. Mr. Parker is an author and editor ofthe “Focus on Statistics” columns in Business Economics,the quarterly journal of the National Association forBusiness Economics (NABE). He currently serves asVice President of the Association of Public Data Usersand is a member of the Board of Directors of theCouncil of Professional Associations for FederalStatistics. He also represents the American EconomicsAssociation on the Census Bureau’s ProfessionalAdvisory Committee. Mr. Parker has served on severalstatistical review panels, authored numerous officialreports for U.S. and international organizations and hasheld offices in several leading statistical and businessorganizations. Mr. Parker was the recipient of the JuliusShiskin Award for 1999. He was elected Fellow of theAmerican Statistical Association in 1986.

Richard S. Sigman is a senior statistician at Westat.He has over 30 years of experience in statisticalmethods and research associated with Federal surveyprograms. Prior to joining Westat, Mr. Sigman was Chiefof the Statistical Methods and Sample Design Staffsupporting the Economic Statistical Methods andProgramming Division at the U.S. Census Bureau. Hespent 14 years at USDA, most recently as senior statisti-cian in the Quality Control Branch of the Food andNutrition Service and prior to that as mathematical stat-istician in the Research Division of NASS. His careerwith the Federal Government also includes work withthe Social Security Administration and the U.S.Department of Defense. Mr. Sigman earned a master’sdegree in Statistics at the University of Maryland and abachelor’s degree in Mathematics at Colorado StateUniversity.

William (Bill) G. Tomek is a professor emeritus,Department of Applied Economics and Management,Cornell University. His teaching and researchemphasized topics related to agricultural commoditymarkets and prices. He is the co-author of AgriculturalProduct Prices (Cornell University Press), now in itsfourth edition and of numerous journal articles andmonographs. Tomek has received awards from theAmerican Agricultural Economics Association (AAEA)


for an outstanding journal article and for two publica-tions of enduring quality. The Chicago Board of TradeEducational Research Foundation has honored hiscontributions to the literature on futures markets. Hewas the editor of the American Journal of AgriculturalEconomics, 1975–1977, and is a past-president andFellow of AAEA. A native of Nebraska, he holdsbachelor’s and master’s degrees from the University ofNebraska and a Ph.D. from the University of Minnesota.He was named a master alumnus of the University ofNebraska, and he has been honored as a distinguishedalumnus of the Department of Applied Economics,University of Minnesota.

Steven Cornell Turner is Head and Professor in theDepartment of Agricultural Economics at MississippiState University. Prior to August 2003, he was aProfessor and Undergraduate Coordinator in theDepartment of Agricultural and Applied Economics atthe University of Georgia. His research has concentrat-ed on marketing and finance with emphasis on thegreen and livestock sectors, along with futures andoption markets. Dr. Turner received the AmericanAgricultural Economics Association (AAEA)Distinguished Teaching Award in 1992 and the 1996Distinguished Professional Contribution in the Teachingof a Course Award from the Southern AgriculturalEconomics Association (SAEA). In 2001, Dr. Turner waselected President of the SAEA. In 2004, he was electedChairman of the Board for C-FARE and served in thatcapacity until 2007. He remains on the Board of C-FARE.

Michael K. Wohlgenant is William Neal ReynoldsDistinguished Professor of Agricultural and ResourceEconomics at North Carolina State University. Hisspecialty is development of economic models of agricul-tural marketing, policy, demand and price analysisproblems. Dr. Wohlgenant has had extensivecommodity experience, including work on applied priceand marketing problems for cotton, dairy, beef, pork,grapes, sugar, tobacco, wine and horticultural crops. Hisresearch has been published in professional journals,including the American Journal of Agricultural Economics,Review of Economics and Statistics, Australian Journal ofAgricultural and Resource Economics, Journal ofAgricultural and Resource Economics, and the European

Review of Agricultural Economics.His research is widely cited andrecognized by awards, with BestArticle awards in four journals.He is also a recipient of the Publication of EnduringQuality Award by the American Agricultural EconomicsAssociation. He has served as an economic consultantfor a number of organizations, including the ResearchTriangle Institute, the Food and Agriculture Organizationof the United Nations and for various law firms inexpert witness testimony (including expert witnesstestimony before the U.S. International TradeCommission). Dr. Wohlgenant has been called upon tobe a consultant to the North Carolina legislature oneconomic issues related to the N.C. swine industry,USDA-ERS and the U.S. Government AccountabilityOffice. He has received numerous awards for hisresearch and teaching, including a William NealReynolds Distinguished Professorship. Dr. Wohlgenantwas elected a Fellow of the American AgriculturalEconomics Association in 2001.

James (Jim) Zavrel is Chief of the Regional IncomeBranch of BEA, U.S. Department of Commerce, wherehe is responsible for the preparation and analysis ofstate and county estimates for all of the non-wagecomponents of personal income—including annualestimates of net farm proprietors’ income, as well as forthe production of quarterly state personal income.Recently, he has overseen efforts to develop advancedestimates of personal income for metropolitan areas aspart of the Bureau’s initiative to accelerate the releaseof its regional statistics. He has extensive experienceworking with agricultural sector economic accounts andhas served on numerous Federal committees andworking groups, including the Advisory Group oncontent for the Census of Agriculture and theAgricultural Sector Subcommittee of the EconomicClassification Policy Committee. Jim has a master’sdegree in Managerial Economics from George MasonUniversity, and a B.A. in Economics from the College ofWilliam and Mary in Virginia.


challenges opportunities for the -...

Documents