Choices 22(2) Full Issue · Oral Capps, Jr., Bruce A. McCarl (Coordinating Editor), ... Robert L. Thompson, University of Illinois ... promote new energy technologies,

CHOICESThe magazine of food, farm, and resource issues

2nd Quarter 2007 • 22(2) CHOICES i

A publication of theAmerican AgriculturalEconomics Association

2nd Quarter 2007 • 22(2)

©1999–2007 CHOICES. All rights reserved. Articles may be reproduced or electronically distributed as long as attribution to Choices and the AmericanAgricultural Economics Association is maintained. Choices subscriptions are free and can be obtained through http://www.choicesmagazine.org.

Table of Contents2nd Quarter 2007—Volume 22, Number 2

A Statement from the Editors and AAEA President. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75Washington Scene . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77 Coordinated by Joe L. Outlaw, Co-Editor, Choices

ArticlesTheme: Resources and the Environment

Water Quality and Agriculture(Catherine Kling, Guest Editor)

Overview: Agriculture and Water Quality in the Cornbelt: Overview of Issues and Approaches . . . . . . 79 Matthew J. Helmers, Thomas M. Isenhart, Catherine L. Kling (Guest Editor), Thomas B. Moorman, William W.

Simpkins, and Mark TomerA Tale of Three Watersheds: Nonpoint Source Pollution and Conservation Practices Across Iowa . . . . 87 Keith E. Schilling, Mark D. Tomer, Philip W. Gassman, Cathy L. Kling, Thomas M. Isenhart, Thomas B. Moorman,

William W. Simpkins, and Calvin F. WolterPrivatizing Ecosystem Services: Water Quality Effects from a Carbon Market . . . . . . . . . . . . . . . . . . . . . . . . 97 Silvia Secchi, Manoj Jha, Lyubov Kurkalova, Hongli Feng, Philip Gassman, and Catherine L. Kling Nitrate Reduction Approaches . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103

Christopher Burkart and Manoj K. Jha

Grab Bag

Organic Demand: A Profile of Consumers in the Fresh Produce Market . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109 John Stevens-Garmon, Chung L. Huang, and Biing-Hwan Lin Water Quality Credit Trading and Agriculture: Recognizing the Challenges and Policy Issues Ahead. 117 Charles Abdalla, Tatiana Borisova, Doug Parker, and Kristen Saacke BlunkFarm Growth, Consolidation, and Diversification: Washington Dairy Industry . . . . . . . . . . . . . . . . . . . . . . 125 Tristan D. Skolrud, Erik O’Donoghue, C. Richard Shumway, and Almuhanad MelhimFruit and Vegetables Go Back to School . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129 John L. Park, Benjamin L. Campbell, Andres Silva, and Rodolfo M. Nayga, Jr.

ii CHOICES 2nd Quarter 2007 • 22(2)


2nd Quarter 2007 • 22(2) CHOICES 75


2nd Quarter 2007 • 22(2)


A Statement from the Editors and AAEA PresidentWelcome to the twelfth and finalissue of our editorship of Choices(Q2 2007). As discussed in theAAEA President's statement justbelow, the Association intends tocontinue Choices and hopes to haveit back online by the end of 2007.Watch for announcements from theAAEA later this year.

Our term as editors expires withthis issue. We wish to thank thosewho have served on the editorialboard as well as those who haveserved as reviewers during our edito-rial term. Special thanks are due tothose individuals who served as guesteditors for specific issues. Choices hada fantastic run the last 3 years as anoutreach vehicle for the association.Thank you for your interest.

As our final offering, this issuecontains a theme on water quality inthe Cornbelt dealing with the prob-lem origin, and issues regarding con-servation programs, multiple serviceprovisions and tradeoffs and watertreatment options. This issue alsocontains articles on

• Organic Produce ConsumerCharacteristics

• Challenges in Water QualityCredit Trading in Agriculture

• Dairy Farm Growth, Consolida-tion, and Diversification

• Fruit and Vegetables in SchoolFood

Editorial Staff

EditorsOral Capps, Jr., Bruce A. McCarl (Coordinating Editor), Rodolfo M. Nayga, Jr., Joe L. Outlaw, John B. Penson, Jr., Texas A&M University

Associate EditorLinda Crenwelge, Texas A&M University

Editorial BoardRichard Adams, Oregon State UniversityWalt Armbruster, Farm FoundationJulie Caswell, University of MassachusettsRalph Christy, Cornell UniversityKeith Collins, Chief Economist, USDARoberta Cook, University of California-DavisAllen Featherstone, Kansas State UniversityAllan Gray, Purdue UniversityHal Harris, Clemson UniversityCraig Jagger, US House Committee on AgricultureCarol A. Jones, Economic Research Service-USDAMaureen Kilkenny, University of NevadaJoost Pennings, University of IllinoisLarry Sanders, Oklahoma State University Brent Sohngen, Ohio State UniversityRobert L. Thompson, University of IllinoisSteven Turner, Mississippi StateChoices is the outreach vehicle of the American Agricultural Economics Association (AAEA) and is designed to provide current coverage regarding economic implica-tions of food, farm, resource, or rural community issues directed toward a broad audience. Choices publishes thematic-oriented groupings of papers and individual papers. The broad themes we will repeatedly visit in Choices are agriculture and trade, resources and the environment, consumers and markets, and agribusiness and finance. Submitted manuscripts are subject to peer review for publication con-sideration.

Choices is published at the end of each quarter of the year by the American Agricul-tural Economics Association. Visit our web site at http://www.choicesmagazine.org.

Editorial CommunicationsPotential manuscripts, thematic proposals, and comments can be submitted through http://www.choicesmagazine.org/submissions.htm or directly emailed to the editors at [email protected]. Editorial communications can be sent to [email protected].

76 CHOICES 2nd Quarter 2007 • 22(2)

Again thank you for your readershipand participation.

Co-Editors: Oral Capps, Jr., RodolfoNayga, Jr., Bruce McCarl, Joe Out-law, and John B. Penson, Jr., andAssociate Editor, Linda Crenwelge

Department of Agricultural Econom-ics, Texas A&M University.

A Message from AAEA President Steve BuccolaCHOICES has been the AmericanAgricultural Economics Association’soutreach arm for 23 years, our princi-pal means of communicating withthose interested in food, farm, natu-ral resource, and rural communityissues but who are not necessarily

professional economists. CHOICESbrings economic research alive to apolicy audience, and internet hitsand downloads suggests it has beenincreasingly successful in doing so.Policy communication is essential tothe AAEA’s ethic because it is essen-tial to the mission to which many ofour members – and their employers –are committed.

CHOICES has always run at asubstantial financial loss. AAEAmembers have, in constant 2006 dol-lars, contributed $1,262,018 to thisjournal ($21 per member per year)since 1990 alone. Annual losses werereduced when CHOICES went elec-tronic but still have hovered around$50,000 and, because our member-ship has declined, is still $20 permember per year. The financial envi-ronment in which the AAEA oper-

ates requires that it further reducethese costs by providing CHOICESwith a new publishing and editorialstructure. Plans for doing so are wellunderway.

The AAEA is deeply grateful tothe editorial team of Bruce McCarl(Coordinating Editor), Oral Capps,Jr., Rodolfo Nayga, Jr., Joe Outlaw,and John B. Penson, Jr. at TexasA&M University, to Associate EditorLinda Crenwelge, and to the 17-member Editorial Board for raisingCHOICES to new standards of clar-ity, relevance, and accessibility. Theirwork provides a bridge to our nextformat. Stay tuned.

Steve Buccola

President, American AgriculturalEconomics Association




2nd Quarter 2007 • 22(2)


Washington Scene Coordinated by Joe L. Outlaw, Co-Editor, Choices

Congress is currently trying to wrap up loose ends on sev-eral key pieces of legislation before their summer recess,which encompasses most of August and the early part ofSeptember. The Farm Bill, a new Energy Bill, and immi-gration reform, among other issues such as appropriationsbills, have been on their agenda.

Farm BillThus far, the House and Senate Agricultural Committeesare each moving forward at their own pace. The House hasheld subcommittee hearings to debate almost all of thetitles of the Farm Bill. The Subcommittee on GeneralCommodities and Risk Management voted 18-0 to use the2002 Farm Bill as their bill for Title I. That doesn’t meanthere won’t be changes in full committee, but that at leastthe subcommittee intends for the basic structure of the2007 Farm Bill to look much like the 2002 Farm Bill. Fullcommittee markup in the House has been postponed, butwould still happen in time for the scheduled debate in theentire House of Representatives in late July.

Mr. Harkin, the Senate Agriculture Committee Chair,has been working through all of the requests from interestgroups to develop his proposal (generally referred to as theChairman’s mark). To date, there has not been any floortime scheduled to debate the farm bill in the Senate.Observers of the farm bill process over the past 20 yearswill say that the Senate tends to move through the processslower than the House of Representatives, so a slower paceisn’t unexpected.

When a bill is finally passed, what is it going to looklike? At this point, most farm policy observers would saythat either an extension of the 2002 Farm Bill or a slightlymodified 2002 Farm Bill are the two most likely outcomesfor Title I. Does this mean that there wouldn’t be anychanges? No, there will likely be modest changes in com-

modity programs, along with increased financial supportin many areas such as food, conservation, and energy pro-grams to name a few. However, the wholesale changes inthe commodity program are less likely than some wantbecause there isn’t enough money available to change thecommodity programs and provide a safety net that worksas well as the current one.

Doha RoundOnly a week after WTO Director-General Pascal Lamyreported modestly high hopes for gaining an agreementwithin the next six months, negotiators from the G4(European Union, United States, Brazil, and India) failedto move forward at their meeting in Potsdam. What next?Trade Promotion Authority (TPA) expires on June 30th..Based on comments from members of Congress, it isn’tclear whether a Doha Round agreement would have sur-vived the up or down vote with TPA. Without TPA, thechances of the Congress passing (without amendment)any Doha Round agreement seem highly unlikely.

Energy The Senate has been debating a new energy bill (HR 6)that was passed on June 22nd. In order to gain passage, acompromise was made on corporate average fuel economy(CAFE) standard increases – lowering the mandatedincrease to a fleetwide average of 35 miles per gallon. Thebill would mandate 36 billion gallons of renewable fuel by2022, increase efficiency standards for appliances and fed-eral buildings, promote new energy technologies, and pro-vide federal grants and loan guarantees for research intofuel-efficient vehicles.

The House has not cleared their version of the EnergyBill.





2nd Quarter 2007 • 22(2)


Theme Overview: Agriculture and Water Quality in the Cornbelt: Overview of Issues and ApproachesBy Matthew J. Helmers, Thomas M. Isenhart, Catherine L. Kling, Thomas B. Moorman, William W. Simpkins,

and Mark Tomer

JEL Classification Code: Q25

More than three decades have elapsed since the passageof the Federal Water Pollution Control Act with its statedgoal of zero discharge of pollutants into the nation’s water-ways. Yet, water quality remains poor in many locationsand considerable loading of pollutants continues. This isparticularly true for agricultural sources of water pollutionand is typified by the Upper Mississippi River Basin,where more than 1,200 water bodies appear on the currentU.S. Environmental Protection Agency (EPA) listing ofimpaired waterways. Additionally, nitrate export from thisregion has been implicated as a significant cause of thehypoxic zone in the Gulf of Mexico, which covered nearly20,000 km2 in 1999 and more than 17,000 km2 in 2006(http://www.epa.gov/gmpo/nutrient/hypoxia_pressrelease.html). Although a substantial body of evidence on theeffectiveness of agricultural conservation practices onwater quality continues to be developed, the net effect ofthese programs and practices at the watershed scale isunclear. Increasingly, studies are being focused on thewatershed (or landscape) scale and complex interactionsbetween agricultural practices and inputs, the types andconfiguration of conservation practices on the landscape,and the resulting downstream water quality. While lowcost methods to reduce agricultural non-point source pol-lution exist, large changes in water quality in agriculturalregions are likely to be costly and met with resistance. Thisis because to achieve large changes in water quality, majoralterations to land use or installation of expensive struc-

tural practices may be required, and the costs are bornedirectly by producers and landowners, or by the taxpayer.

Given the potentially large cost for significantimprovements in water quality, it is critical to developtools that can support cost-effective design of conservationpolicy and/or voluntary implementation of watershedplans focused on water quality. The following set ofthemed papers related to water quality and agriculture dis-cuss these issues, with a specific focus on using integratedwater quality and economic models to support better pub-lic policy and watershed-based solutions to these prob-lems. The article following this one describes detailedfield-scale data collected as part of a Conservation EffectsAssessment Project supported by CSREES and ARS. Inaddition to assessing the effects of current conservationactivities on water quality in these watersheds, data areused to calibrate a water quality model and are being inte-grated with economic cost information to study the opti-mal placement of additional conservation activities in thewatershed. That article discusses the historical evolution ofconservation activities in the three watersheds, the currentwater quality challenges in the watersheds, and the rolethat the integrated models can play in solving the prob-lems.

In the third paper of the series, Secchi et al. employ amore aggregate unit of analysis (scale) for calibrating awatershed model and a biophysical carbon sequestrationmodel and integrating them with economic data coveringthe entire state of Iowa. The focus of their analysis is on


the potential unanticipated environ-mental effects of developing marketsin ecosystem services that focus on asingle service, such as carbon seques-tration.

The final paper in the setaddresses a different water qualityissue: drinking water and nitrate lev-els. Specifically, the paper by Burkartand Jha considers whether it wouldbe cost-effective for farmers to reducenitrogen applications at the farmlevel, thereby reducing nitrate con-centrations in the water supplies forresidential consumers, rather thancontinue to treat the water in a deni-trification plant prior to use.

In the remainder of this themeoverview, we attempt to provide thecasual reader with adequate back-ground information on agriculturalwater quality problems, as well as theinstitutional framework withinwhich these water quality problemsin agriculture are currently managed.This includes a brief primer on thekey pollutants, their sources, and therange of conservation methods thatcan attenuate their effects. It is alsonecessary to understand the funda-mentals of the policy environment,which differs markedly fromapproaches taken in other industries.Specifically, voluntary actions are thefocus of state and federal agencyefforts under the requirements thatthey have to develop and implementTotal Maximum Daily Loads(TMDLs). We briefly describe theTMDL process and note the range offederal and state conservation pro-grams that provide funding for vol-untary conservation efforts.

Agriculture and Water Quality PrimerProduction of food and fiber haveinevitable impacts on land and waterresources. Conservation practices are

intended to reduce those impacts,ideally with as little effect on the pro-ductive and ecosystem service capaci-ties of the land. The critical questionsfor planning and implementation ofeffective conservation systems arethen: What water quality pollutantsare of primary concern and whattypes of conservation practices willprovide benefits for various environ-mental impacts? Here, briefly, weprovide generic answers to thesequestions that are most pertinent toagricultural watersheds in the CornBelt generally, and Iowa and theUpper Mississippi Basin, specifically.Through this discussion, we empha-size key differences among specificpollutants, in terms of the hydrologicpathways from field to stream, andthe types of conservation practicesthat can minimize their transport toreceiving waters. The primary pollut-ants of concern in the Corn Beltinclude nitrate-nitrogen, phospho-rous and sediment, and pathogens.

Nitrates-NitrogenNitrate-nitrogen (NO3-N) is a keypollutant of concern for its potentialwidespread impact on both publichealth and ecosystem function.Nitrate-nitrogen is readily leachedthrough soils to groundwater andenters surface water systems directlyby groundwater flow and through thesubsurface drainage systems (tiledrains), which were installed acrosslarge areas of poorly drained Mid-western soils beginning about 100years ago. These drainage systemshave allowed the Midwest to becomethe highly productive agriculturalarea that it is today, while short-cir-cuiting the much slower, naturalgroundwater pathway to the stream.Concentrations of nitrate-nitrogen indrainage and stream water oftenexceed 10 mg NO3-N /L, resulting

in losses exceeding 20 kg N/ha insome years (Tomer et al., 2003).Regional nitrogen budgets for theMississippi River Basin have impli-cated tile-drained regions of the Mid-west as disproportionately contribut-ing to N loads to the Gulf (Burkartand James, 1999). Nitrogen fertilizeris commonly applied to corn, at ratesvarying from 100 to 200 kg/ha. Theefficiency of N uptake by the cropvaries because of environmental con-ditions. Nitrogen losses are mostprevalent in early Spring when cropsare not present or are too small toeffectively immobilize the availablenitrate.

The problem of nitrate-nitrogenexport is not solely caused by N fer-tilizer management or any other sin-gle factor, but rather it is a combina-tion of soil management practicesand physical, chemical, and biologi-cal characteristics of the soil, alongwith temperature and precipitationpatterns (Dinnes et al., 2002). As aresult, reducing nitrate loss is morethan a matter of reducing N-fertilizerrates and improving timing of appli-cations (Jaynes et al., 2004). Effectivepractices to control N losses includediversified crop rotations thatincrease use of forages and improvednitrogen management (includingimproved timing and rates of applica-tion, and use of nitrification inhibi-tors). Improved engineering of agingdrainage infrastructure, and use ofwetlands, cover crops, and denitrifi-cation walls or subsurface drainagebioreactors are other alternatives thathave been shown effective. Becausenitrate in extensively tiled areas istransported to streams primarily insubsurface drainage water, any filter-ing ability of riparian buffers andedge of field filter strips will bebypassed.


Phosphorous and SedimentSurface runoff is the dominant mech-anism that transports phosphorus,sediment, and pesticides and bacte-ria from agricultural fields, asopposed to the subsurface pathwaysof nitrate. Ecological impacts of Pand sediment include eutrophicationand sedimentation of receivingwaters. Phosphorus losses from agri-cultural fields may be only a fractionof those observed for N (< 1 kg/ha.yris commonly reported), but suchlosses can have major implicationsfor the ecological integrity of lakesand streams. Phosphorus runoff fromagricultural fields is largely controlledby soil P concentrations and crop res-idue cover (Sharpley et al., 2002).Residue cover encourages infiltrationand discourages erosion. To improvephosphorus management at water-shed scales, the use of “P indices” arebeing implemented that identify soilerodibility, soil P concentrations, res-idue management practices, andproximity to streams, to rank fieldsfor runoff P losses. These indices canbe used to target conservation prac-tices to control P losses (Birr andMulla, 2001) via reduced tillage, lim-ited manure or fertilizer applica-tions, terraces, vegetated filter strips,and/or riparian buffers. These prac-tices are known to reduce erosion andphosphorus. Watershed responses tothese conservation practices may beless than initially expected becausestreambank erosion, rather than agri-cultural fields, can contribute signifi-cant amounts of sediment and phos-phorus to streams and rivers. Thesesources may result from past manage-ment activities.

Sediment and nutrient lossesfrom agriculture, therefore, can resultin a legacy of impacts within water-sheds, necessitating a long-term com-mitment to their amelioration. For

example, elevated nitrate concentra-tions in groundwater have beenshown to remain for decades (Rod-vang and Simpkins, 2001). Also,phosphorus accumulations in sedi-ment may have a legacy, providing along-term, internal loading source ofmineral P to the water column(Christophoridis and Fytianos, 2006)and may ultimately affect groundwa-ter P concentrations (Burkart et al.,2004).

Bacterial Pathogens and Livestock ConcernsLivestock is an important economiccomponent of U.S. agriculture,accounting for over 60% of agricul-tural sales. Production estimates for2005 include 72.6 million hogs, 10.9million beef cows, 3.1 million milkcows, 150 million egg layers, and 131million broilers for the 12-stateNorth Central Region. In the Mid-west, swine are increasingly producedin concentrated animal feeding oper-ations (CAFOs) making manuremanagement increasingly important,both as a source of nutrients for sub-sequent crops and as a potential envi-ronmental problem. CAFOs are alsoimportant in poultry and beef pro-duction. Potential water qualityissues arising from manure applica-tion are nitrate leaching and loss intile drainage networks, and loss ofphosphorus and pathogens in over-land runoff. Conservation practicesseek to prevent accumulation ofexcess nutrients (nutrient manage-ment plans), reduce and/or treat run-off from feed lots, and mitigate run-off from manured fields (buffers,filter strips). Several studies suggestthat increasing CAFO size offers cer-tain economic advantages in produc-tion, but increases the amounts ofmanure applied to land near theCAFO, which increases the risk of

loss of excess nutrients (Kellogg et al.,2000).

Bacterial pathogens that threatenwater quality include Escherichia coliO157:H7, Salmonella, Enterococ-cus, Listeria, and Campylobacter.Pathogenic protozoa includeCryptosporidium and Giardia.Although these microorganismscause disease in humans, they arecommonly carried in livestock with-out visible symptoms. Because of thedifficulty and cost involved in screen-ing water samples for these patho-gens, public health and water supplyauthorities have long relied uponindicator bacteria. In the past, fecalcoliforms tests filled this function,but two indicators are now beingpromoted by U.S. EPA, Escherichiacoli and Enterococcus. Quick andreliable tests for both of these micro-organisms are now available and thepresence of these bacteria has beencorrelated with the presence of dis-ease-causing microorganisms. Mea-sured E. coli densities in stream watercan be evaluated against EPA’s cur-rent standards, but the identificationof the E. coli sources is more complexand important to developing effectivewatershed management strategies.Microbial source tracking is anemerging technology that allows thesource animal to be determined.Potential sources in most watershedsinclude wildlife, farm animals, andhumans.

Heterogeneity of Conservation PracticesThere is a wide range of conservationpractices used on agricultural landintended to provide water qualitybenefits, including engineered struc-tures, edge-of-field practices, in-fieldnutrient and crop residue manage-ment practices, and land retirement.Government programs since the


1930s have promoted installation ofconservation practices on agriculturallands. Much of the early focus ofconservation practices was specifi-cally on soil conservation, where thegoal was to preserve the soil and tomaintain its productivity.

Structural practices that havebeen used for controlling soil loss andthe formation of gullies include ter-races, grassed waterways, sedimentbasins, and grade stabilization struc-tures. Terraces are used to decreasethe length of the hill-slope to reducerill erosion and the formation of gul-lies. Many early conservation prac-tices were intended, in part, for waterconveyance to improve trafficability,and thereby maximize agriculturalproduction. In addition to structuralpractices, there are a variety of in-field management practices such ascontour farming and strip croppingand tillage management, such as con-servation tillage and no-till. Also, insome areas marginal lands that arehighly susceptible to soil loss havebeen taken out of agricultural pro-duction and converted back to peren-nial vegetation.

Over the past thirty years, therehas been an increased concern relatedto the overall water quality impacts ofagriculture, including nutrient, pesti-cide, and pathogen loss from agricul-tural lands. Some conservation prac-tices have been installed with anintended purpose of reducing theexport of these contaminants. Two ofthese are buffer systems (riparian orgrassed) and the reintroduction ofwetlands back into the landscape. Inaddition, relative to nutrient losses,there has been an emphasis on appro-priate nutrient management practiceswithin agricultural fields to reducethe application of excess nutrients.

We have also learned that someagricultural practices have effects thatwere not intended. Subsurface drain-

age was used historically to enhanceproductivity of poorly drained lands,but these production benefits are off-set by the environmental impacts ofincreased export of nitrate-nitrogenfrom these drainage networks. Sur-face inlets to subsurface drain systemsalso create a direct conduit for surfacewater to enter streams effectivelybypassing riparian buffers or wet-lands. Much of the agricultural land-scape has been altered throughstream straightening channelization.Stream straightening and subsurfacedrainage have significantly altered thehydrology of the landscape, whichhas led to significant streambank sta-bility problems in many areas. So,while many of the conservation prac-tices mentioned above may reducesoil loss from agricultural fields, ifthey do not significantly reduce waterflow in the streams, the stream poweris not reduced. As a result, ratherthan carrying sediment from fields,the streams may erode sediment fromthe streambed and streambanks.

While there is a wide range ofpractices that can be used on agricul-tural lands for providing water qual-ity benefits, many times the locationswithin the watershed where practicesare implemented have not been spe-cifically targeted to achieve the great-est reduction of contaminants indownstream water bodies. This islikely the result of the voluntaryenrollment in federal conservationprograms combined with ineffectivetargeting technology. Recentadvances in remote sensing and geo-graphic information systems offer anopportunity for dramatic improve-ments in our ability to target conser-vation practice installation in largewatersheds. With the limited amountof resources available for conserva-tion practices, there will likely beincreased importance on targetingimplementation to those areas where

there may be the greatest benefitfrom a water quality perspective. Oneprogram that has used targeting withsome effectiveness is the Conserva-tion Reserve Program (CRP), whichtargets land choices based on an envi-ronmental benefits index. While theeffects of CRP on soil quality, carbonstorage, and wildlife have beenassessed, the aggregate effects at thewatershed scale are less understood.

Finally, it is important to under-stand that water quality monitoringin the United States is done by a vari-ety of state and federal agencies,including USGS and USEPA, andmany municipal and commercialwater supply entities, but the greatmajority of streams and rivers are notroutinely monitored. Thus, in manycases, the actual level of pollutants issimply unknown.

The Policy Environment: TMDLs and Voluntary ImplementationVoluntary cost-share and incentiveprograms sponsored by USDA andStates are large in geographic scaleand fiscal commitment (over $4.5billion was spent in 2005 by USDA-funded programs alone). These pro-grams generally provide varyingincentives to farmers for the installa-tion of structural or managementpractices described above. The crite-ria for participant eligibility varyfrom program to program, and con-servation compliance provisionsrequire that landowners who farm onhighly erodible land undertake someconservation activities in order to beeligible for other government incen-tives or subsidies. In addition to thelargest program, the CRP, there is acost-share program entitled the Envi-ronmental Quality Incentive Pro-gram, which provides cost share toproducers willing to install variousconservation structures or practices


on their farms. Notably, the 2002Farm Bill contained a new programthe Conservation Security Program awatershed-based initiative intendedto compensate farmers for adoptingconservation practices. Like the CRP,which covers the full cost of retiringland from production, the programwas intended to cover the full cost ofadopting conservation practices(rather than less than 100% of thecost as traditional cost share pro-grams do), but the focus of the Con-servation Security Program is on landthat stays in production. However,funding constraints have preventedthe program as it was initially envis-aged from being fully implemented.

Ironically, while there are largeconservation programs funded andadministered through USDA, theprimary law that addresses nonpointsource agricultural pollution loadingsis under the auspices of the U.S.Environmental Protection Agency(EPA) via the Clean Water Act.Rather than assign standards andrequire that sources implementchanges in production or invest inabatement technology to meet thosestandards, as has been the norm forair and water quality problems stem-ming from point sources, the TotalMaximum Daily Load (TMDL)approach was adopted. Under theTMDL framework, states are respon-sible for compiling lists of water bod-ies not meeting their designated uses,which are then reported as “impairedwaters.” The sources of impairmentvary across locations. For example,Iowa has 213 water bodies on the listand pathogens (bacteria) are the lead-ing cause of listing, accounting forabout 20% of the impaired waterbodies, with sediment/turbidityaccounting for about 10%. Nation-ally, it has been estimated that 40%of rivers and estuaries fail to meetrecreational water quality standards

because of microbial pollution(Smith & Perdek, 2004).

Note that water bodies are viewedas impaired only if they do not meettheir “designated use.” Thus, twowater bodies can be equally contami-nated with only one being listed asimpaired if their designated uses aredifferent (e.g., boatable vs. swimma-ble). This is part of what makes theTMDL rules so difficult to interpretand why a simple indication ofwhether a water body is listed or notis not necessarily a good indication ofits level of water quality.

Once a water body has been iden-tified as not meeting its designateduse, the state is required to identifythe sources of the impairment andthe “maximum allowable daily load”of pollutants that would eliminatethe impairment. Finally, states are tosuggest reductions for the variouspollutant sources that would allowthe watershed to reach the TMDL.Importantly, there is no regulatoryauthority by the states or EPA torequire that these reductions occur.Thus, the institutional environmentin which nonpoint source water qual-ity reductions may occur is funda-mentally voluntary.

In the TMDL process, modelingand monitoring can play importantroles in allocating pollutant loads tovarious sources, such as helping todetermine the relative contributionsof row crops, CAFOs, and urbansources to loads of nutrients and bac-teria observed in large watersheds.Two models, the Soil and WaterAssessment Tool and the Hydrologi-cal Simulation Program-FORTRANmodels are most often used to sup-port TMDL assessments (Benham etal., 2006). These models combineGIS-based spatial data of watershedphysical features with information oncropping systems, animal densities,fertilizer and pesticide use, and point

sources. For non-point source pollut-ants, conservation practices are a keyto developing mitigation strategiesthat allow watersheds to meetTMDL goals. Since TMDLs may bedesigned to mitigate multiple pollut-ants (e.g., nitrate and bacteria), com-binations of conservation practicesmay be necessary to achieve the nec-essary improvements in water quality.

Final RemarksThe purpose of this overview is tointroduce readers to the set of waterquality problems associated withrow-crop agriculture and livestockoperations in the Corn Belt andUpper Mississippi River Basin. Theproblems are complex, with a greatmany individual decentralized deci-sion makers contributing, both posi-tively and negatively, to their solu-tions. Adding to these complexproblems are the ever-changingdemands on agriculture to supplyfood, feed, fiber, and fuel. Thesedemands are leading to new ques-tions and concerns related to agricul-ture and may allow for some solu-tions that are economically viableand environmentally beneficial.Some concerns are related to poten-tial use of marginal lands for rowcrop agricultural production andincreasing continuous corn acreage tosupply the bioeconomy. At the sametime, the bioeconomy, particularly ifcellulose biofuels become feasible,may provide opportunities for morediversified cropping systems thathave environmental benefits. Associ-ated with some of these issues is theincreasing importance of agribusinessthrough decisions such as siting ofCAFOs and ethanol plants. Sitingdecisions should consider the poten-tial environmental impacts of thesefacilities both from a water qualityand water quantity perspective.


In the remaining three papers ofthis water quality theme, the authorsdescribe how data and models can beused to characterize the problems,model the underlying biophysicaland economic processes, and ulti-mately (hopefully) contribute tosolutions. Given the policy environ-ment described above one of volun-tary-based action and a myriad ofconservation programs with diversegoals and ever-present funding con-straints we believe that models ofwater quality processes carefully inte-grated with economic models areessential, both to assess existing pro-grams, and more importantly, todesign and implement cost-effectiveapproaches to meeting society’s waterquality goals. These modeling effortswill be difficult and will appropri-ately come under a great deal of scru-tiny.

The complexity of the ecologyand the social issues (including a hostof topics not addressed here such asinternational trade agreements, ruralcommunity viability, rural-urbanconflicts, etc.) indicate a need foradditional research that considers thebreadth of the systems involved atscales that are appropriate. For exam-ple, much of our current knowledgeof the efficacy of conservation prac-tices is based on field scale researchwhich cannot be simply “scaled-up”to understand the workings at water-shed levels. While current researchefforts are beginning at this morechallenging scale, definitive resultswill be, in many cases, many yearsoff.

Before we leave the reader to diveinto the three following papers, wenote a final thorny point concerningthe potential for significant “legacy”problems possibly hiding in ground-water supplies. Over many decades ofagricultural activity, we have addednutrients and other effluents to

groundwater systems that haveundoubtedly not yet emerged at thesurface. When and where such pol-lutants will appear is not clear, but ifconservation programs are designedonly with current pollutant contribu-tions in mind, our efforts may wellfall short due to these legacy sources.

For More InformationBenham, B.L., Baffaut, C., Zeckoski,

R.W., Mankin, K.R., Pachepsky, Y.A., Sadeghi, A.M., Brannan, K.M., Soupir, M.L., & Habersack, M.J. (2006). Modeling bacteria fate and transport in watersheds to support TMDLs. Transactions of the American Society of Agricultural and Biological Engineers, 49(4), 987-1002.

Birr, A.S., & Mulla, D.J. (2001). Evaluation of the phosphorus index in watersheds at the regional scale. Journal of Environmental Quality, 30(6), 2018-25.

Burkart, M.R., & James, D.E. (1999). Agricultural nitrogen contributions to hypoxia in the Gulf of Mexico. Journal of Environmental Quality, 28, 850-59.

Burkart, M.R., Simpkins, W.W., Morrow, A.J., & Gannon, J.M. (2004). Occurrence of total dissolved phosphorus in surficial aquifers and aquitards in Iowa. American Journal of Water Resources Association, 40(3), 827-34.

Christophoridis, C., & Fytianos, K. (2006). Conditions affecting the release of phosphorus from surface lake sediments. Journal of Environmental Quality, 35(4), 1181-92.

Dinnes, D.L., Karlen, D.L., Jaynes, D.B., Kaspar, T.C., Hatfield, J.L.,

Colvin, T.S., & Cambardella, C.A. (2002). Nitrogen management strategies to reduce nitrate leaching in tile-drained Midwestern soils. Agronomy Journal, 94, 153-71.

Follet, R.F., & Hatfield, J.L. (ed.). (2001). Nitrogen in the environment: Sources, problems and management. Amsterdam: Elsevier Science.

Jaynes, D.B., Dinnes, D.L., Meek, D.W., Karlen, D.L., Cambardella, C.A., & Colvin, T.S. (2004). Using the late spring nitrate test to reduce nitrate loss within a watershed. Journal of Environmental Quality, 33(2), 669-77.

Kellogg, R.L., Lander, C.H., Moffitt, D.C., & Gollehon, N. (2000). Manure nutrients relative to the capacity of cropland and pastureland to assimilate nutrients: Spatial and temporal trends for the United States. USDA Economic Research Service Publication No. nps00-0579.

Kross, B.C., Hallberg, G.R., Bruner, D.R., Cherryholmes, K., Johnson, J.K. (1993). The nitrate contamination of private well water in Iowa. American Journal of Public Health, 83(2), 270-72.

Rodvang, S.J., & Simpkins, W.W. (2001). Agricultural contaminants in Quaternary aquitards: A review of occurrence and fate in North America. Hydrogeology Journal, 9(1), 44-59.

Sharpley, A.N., Kleinman, P.J.A., McDowell, R.W., Gitau, M., & Bryant, R.B. (2002). Modeling phosphorus transport in agricultural watersheds: Processes and possibilities. Journal of Soil and Water Conservation, 57(6), 425-39.


Smith, J.E., & Perdek, J.M. (2004). Assessment and management of watershed microbial contaminants. Critical Reviews in Environmental Science and Technology, 34, 109-39.

Tomer, M.D., Meek, D.W., Jaynes, D.B., & Hatfield, J.L. (2003). Evaluation of nitrate nitrogen fluxes from a tile-drained watershed in Central Iowa.

Journal of Environmental Quality, 32(2), 642-653.

Matthew J. Helmers ([email protected]) is Assistant Professor,Department of Ag/Biosystem Engi-neering, Thomas M. Isenhart ([email protected])Assistant Professor,Department of Natural ResourceEcology and Management. CatherineL. Kling ([email protected]) Profes-sor, Department of Economics and

Center for Agricultural and RuralDevelopment, and William Simpkins([email protected]) is Professor,Department of Geology and Atmo-spheric Sciences, Iowa State Univer-sity, Ames, IA. Thomas B. Moorman([email protected]) isMicrobiologist, and Mark Tomer([email protected]) isResearch Soil Scientist, USDA-ARS,National Soil Tilth Laboratory,Ames, IA.





2nd Quarter 2007 • 22(2)


A Tale of Three Watersheds: Nonpoint Source Pollution and Conservation Practices across IowaBy Keith E. Schilling, Mark D. Tomer, Philip W. Gassman, Cathy L. Kling, Thomas M. Isenhart, Thomas B.

Moorman, William W. Simpkins, and Calvin F. Wolter


Many conservation practices and implementation pro-grams exist to address nonpoint source (NPS) pollutionlosses from agricultural landscapes (Helmers et al., thisissue). In order to select the most appropriate practices andprograms for reducing NPS pollution in a specific regionwhile maintaining economic return for the landowner, theinteracting processes of agricultural management andwatershed hydrology need to be understood across broadspatial scales. On a nationwide basis, it is easy to see howNPS pollution in one part of the country might be differ-ent than those in another region of the country. For exam-ple, cotton growers in the South, dairy farmers in theNortheast, cattle ranchers in the West, and grain farmersin the Midwest all face unique challenges based on differ-ences in climate, soil types, and cropping patterns. Eachregion relies on a different set of conservation practicesand programs to address NPS pollution. To be effective,conservation systems must be based on an understandingof specific management impacts on water quality prob-lems, and therefore be targeted to reduce, intercept, and/ortreat contaminants moving via surface or sub-surface path-ways from working agricultural lands.

Within agricultural regions, one might expect greaterhomogeneity in biophysical features and cropping prac-tices and be tempted to think that one size fits all; i.e., thatone set of conservation prescriptions can be used toaddress the negative impacts of agriculture on aquatic andterrestrial integrity. If this generalization could be madeanywhere, certainly a state such as Iowa, dominated by itsvast extent of corn and soybean fields, would be suited fora limited set of conservation prescriptions. However, as

described in this tale of three watersheds, conservationpractices must instead be tailored to individual landownerobjectives and local landscape conditions in order to opti-mize their effectiveness.

The research described in this paper was conducted aspart of USDA’s Conservation Effects Assessment Project(CEAP) and its Watershed Assessment Studies (Mausbachet al., 2004). The objectives of the project are to evaluatethe effects of agricultural conservation practices on waterquality, with a focus on understanding how the suite ofconservation practices, the timing of these activities, andthe spatial distribution of these practices throughout awatershed influence their effectiveness. An additionalcomponent of the project is to evaluate social and eco-nomic factors influencing implementation and mainte-nance of practices.

Watershed DescriptionsTo evaluate the effects of watershed conservation practiceson water quality, and to assess the spatial distribution ofthese practices, we are focusing on three watersheds in dis-tinct landscape regions of Iowa (Figure 1). By studyingthree watersheds with differing physical characteristics andpossessing a unique set of pollutants, practices and pro-grams, we can better assess the effectiveness of conserva-tion activities and land management decisions.

LandformsThe Sny Magill Creek, Squaw Creek, and the South Forkof the Iowa River (South Fork) watersheds are representa-tive of three distinct landform regions of Iowa (Prior,


1991). In Northeast Iowa, Sny Mag-ill Creek is a third-order stream inClayton County that drains 35.6 mi2

of the Paleozoic Plateau landformregion before discharging directlyinto the Mississippi River. The land-scape of this region is characterizedby narrow, gently sloping uplandsthat break into steep slopes withabundant outcrops of sandstone andlimestone. The characteristic lime-stone bedrock of the area gives rise tokarst features (sinkholes, caves, andsprings) that are found throughoutthe Sny Magill Creek watershed (Fig-

ure 2). Nearly 80% of annual stream-flow is ‘baseflow’ attributable toground water discharge from thesesubsurface sources. This results in“cold water” conditions suitable forhighly popular trout fisheries.

The 18.3 mi2 Squaw Creekwatershed is located in South-CentralIowa in Jasper County in the South-ern Iowa Drift Plain. The landscapeof this region is characterized bysteeply rolling hills and a well-devel-oped stream network that developedon a landscape composed of geologi-cally old (>500,000 years) glacial till

(poorly sorted mixture of gravel,sand, silt, and clay) overlain by geo-logically recent (17,000 to 31,000year old) windblown silt (loess).Because of the sloping hillsides andpoor infiltration capacity of the soils,rainfall is primarily directed tostreams via overland runoff, and only55% of the stream discharge is attrib-utable to baseflow originating asground water.

The largest of the study water-sheds is the South Fork of the IowaRiver, which covers 301 mi2 withinHardin and Hamilton counties inCentral Iowa. The landscape is repre-sentative of the Des Moines Lobe,the dominant landform region ofNorth-Central Iowa. The terrain isyoung (about 12,000 years since gla-cial retreat), and thus much of thelandscape is dominated by low reliefand poor surface drainage. Prior tosettlement by Europeans, the land-scape was a complex of wetlands, andthe stream network was poorly devel-oped due to the relatively younglandscape. The geology of the DesMoines Lobe region consists largelyof glacial till deposits in morainesand flat to rolling uplands, clay andpeat in depressional “prairie pothole”areas, and sand and gravel deposits infloodplains of rivers and streams. Soilwetness is a major concern for landmanagement and agricultural pro-duction. Hydric soils (indicative ofsoil saturation on at least a seasonalbasis) occupy about 54% of thewatershed, and artificial tile drainage(Figure 2) was installed in thesehighly productive and nutrient richsoils to lower the water table andallow crops to be grown. Thus, about70% of the stream flow in the SouthFork watershed originates from sub-surface drainage (Green et al., 2006),with most tile discharge occurringduring spring and early summer.

Figure 1. Location of the three watersheds (and controls) and EPA Ecore-gions in Iowa.

Figure 2. Subsurface hydrologic features in Sny Magill and South Fork water-sheds. In the Sny Magill watershed (left photo), springs and caves dischargenatural drainage from unknown areas of karst terrain. In the South Fork water-shed, once-prevalent wetlands were converted to cropland with tile drainage.This 36-inch clay pipe (right photo) has discharged drainage collected fromabout 4,500 acres of cropland for nearly 100 years (note monitoring lines).


Physical features of the threewatersheds have a great influence onthe timing and magnitude of therouting of water to streams. Water-sheds draining older landscapes havegreater slope and greater stream den-sity (number of streams per squaremile) than younger landscapes (Fig-ure 3). For example, the Sny Magillwatershed has twice the average slopeas Squaw Creek, which has morethan twice the slope of the SouthFork watershed. Slopes in Sny Magillare further accentuated because of

the bedrock terrain and its proximityto the Mississippi River. The SquawCreek and Sny Magill watersheds alsohave nearly three times more streamsper square mile than the South Fork.The well-dissected landscape of theSny Magill watershed shows a greaterstream density; thus, rainfall can bequickly routed as overland runoff tosinkholes or streams. In the SouthFork watershed, where natural drain-age is poor, excess rainfall would col-lect in potholes or other surfacedepressions if not for prevalence of

subsurface tile drainage, which hasaccelerated the routing of rainfallwater off the land. Watersheds drain-ing the Des Moines Lobe may yieldas much water as those draining frac-tured carbonate bedrock (Schillingand Wolter, 2005).

Relation of Land Use to the Landform RegionDifferences in land cover among thethree watersheds can be traced largelyto their watershed morphologies andthe suitability of land for intensive

Figure 3. Land use, conservation practices, and other characteristics of the three watersheds.


row crop agriculture. Row crops inthe Sny Magill watershed, primarilyfound on narrow upland divides andbottomlands, only comprise 26% ofthe land area (Figure 3). Grasses andforest are widespread in the Sny Mag-ill watershed, located on steep terrainthat is difficult to cultivate. In theSquaw Creek watershed, slopes arenot as severe as in the Sny Magillwatershed, and row crops are foundon 81% of the land area. Grasses aredistributed around the watershed onhighly erodible land, a practiceencouraged by conservation pro-grams. The till plain of the DesMoines Lobe, represented by theSouth Fork watershed, is also heavilyutilized for row crop production,which occupies 85% of the water-shed area.

Animal AgricultureIn the early 1900s, most small farmsin Iowa had livestock, often includ-ing cattle, swine, and chickens. As aresult of changes in farm policy andeconomies of scale, all three water-sheds have experienced shifts in ani-mal agriculture that are representa-tive of changes across the largerlandform regions (Figure 1). Histori-cally, the Sny Magill watershed hadsignificant numbers of dairy cattleutilizing available grasslands for for-age. While still a significant industry,dairy cattle within Sny Magill havedecreased greatly, with a resultingshift in some grassland acreage to rowcrop agriculture (soybean acreageespecially increased in Iowa as pas-ture and hayland decreased). Live-stock is comparatively absent in theSquaw Creek watershed except forseveral cow-calf operations. Nowhereis the concentration of livestock moreapparent than in the South Forkwatershed, where most swine andchickens are raised in confined ani-mal feeding operations (CAFOs).

There are nearly 100 CAFOs (mostlyswine) in the watershed. Based oninventories reported for permittedfacilities, hogs and chickens in theSouth Fork watershed number 1,654and 2,880 per square mile, respec-tively, which are densities consider-ably greater than the other twowatersheds combined (Figure 3). Allthe reported chickens are housed inone large egg-producing facility,while swine facilities are abundantacross the central part of the water-shed. We estimate that about a quar-ter of the watershed receives manureapplications annually, assuming thisis applied prior to corn at a rateequivalent to that crop’s uptake ofnitrogen (about 190 lb N/ac). Usu-ally these applications are done byinjection, and carried out in the fallwhen soils tend to be dry and mosteasily trafficked by manure tankersand applicators.

NPS Pollutants and the LandscapeBecause of their different watershedcharacteristics, land use, and live-stock histories, non-point pollutantsources and transport vary greatlyamong the three watersheds (Figure3). Pollutants of particular concern inIowa are sediment, nutrients (nitrateand phosphorus), and fecal bacteria(E. coli). In Iowa, nitrate concentra-tions in streams relate to the amountof row crops in a watershed (Schilling& Libra, 2000), and nitrate-N con-centrations are highest in the SouthFork and Squaw Creek watersheds,with median concentrations of 14.2and 9.5 mg/L, respectively. Tiledrains contribute greatly to nitratelosses in the South Fork watershed.In the mid-1990s, the USGS foundstream nitrate concentrations in theSouth Fork watershed to be among

the highest observed in the UnitedStates (Becher et al., 2001).

In contrast, nitrate concentra-tions are considerably lower in theSny Magill watershed, averaging 3.3mg/L over 10 years, a value thatwould be the envy of most otherregions of Iowa. The smaller concen-tration results from the differences inland use (Figure 3). Fecal bacteriacounts are also highest in the SouthFork watershed; however, multiplesources of bacteria are suspectedbecause patterns do not always followthe distribution of livestock. Yet,research suggests that these bacteriallosses in runoff are greatest when thatrunoff occurs within several weeks ofmanure application. Fecal bacteriaconcentrations in Squaw Creek arealso elevated, which may be surpris-ing, given the lower livestock densi-ties. However, cattle with directaccess to the streams, wildlife, andinadequate septic systems may allcontribute to fecal contamination ofIowa streams.

Sediment loss is also a major con-cern in these watersheds (Figures 3and 4). The greatest annual sedimentloss per unit area was associated withthe Squaw Creek watershed (0.69tons/ac per year), whereas meanannual sediment loss from Sny Mag-ill and South Fork watersheds aver-aged 0.26 and 0.28 tons/ac, respec-tively. Considering that row cropscover only 26%of the land area in theSny Magill watershed, actual soil lossper acre of cropland may be substan-tially greater. Long-term sedimentmonitoring data in the Sny Magilland Squaw Creek watersheds indi-cates that sediment transport is veryflashy in both watersheds, with muchof the annual sediment loss trans-ported by runoff from a few intenserainfall events. In Squaw Creek, onaverage, about 40% of the water-


shed’s annual sediment loss occurs onthe single day of greatest runoff.

Sediment losses from watershedsresult from overland flow across thelandscape, causing sheet, rill, andgully erosion, as well as substantialcontributions from streambanks. Inthe South Fork watershed, sedimentlosses are actually about three timeshigher than typically measured in theDes Moines Lobe region. In thelower third of the watershed, landsbecome more highly erodible in anarea of hilly moraines near the DesMoines Lobe’s edge, and the rivererodes its banks as it meanders acrossan alluvial valley (Figure 4). In someIowa watersheds, streambank ero-sion can contribute more than half ofthe annual sediment load exportedfrom a watershed.

Phosphorus is strongly adsorbedto sediment, which was reflected inSquaw Creek having both the great-est average sediment yields and great-est median concentrations of total P(0.14 mg/L) among the three water-sheds. Squaw Creek’s median P con-centration is more than twice whatEPA has proposed as the standard forMidwest streams. Concentrations ofP in the South Fork, by comparison,had a median of 0.07 mg/L during

three years of weekly-biweekly sam-pling. Recent groundwater samplingfrom 24 wells located throughout theSouth Fork watershed has shownmedian and maximum total P con-centrations of 0.030 and 0.340 mg/L, respectively. These groundwater Pconcentrations are found in similarmaterials and landscapes in Iowa(Burkart et al., 2004), and suggestthat groundwater can also be a P con-tributor to streams.

Tailoring Conservation Practices to Watersheds and NPS PollutantsConservation practices used on rowcrop fields in the three watershedsreflect respective watershed charac-teristics and land use histories. Afield-by-field assessment of conserva-tion practices was conducted in eachwatershed to assess the variety anddistribution of practices. An analysisof tillage practices, terraces, and con-tour farming shows the degree towhich land managers have used theseconservation practices to reducenutrient and sediment losses fromthe three watersheds. Of the threetillage practices assessed (conven-tional tillage, mulch till, and no till),

mulch till was most widely utilized inall watersheds. Mulch tillage (>30%residue cover) was used on 62 to91% of all row crop fields, whereasno till was used on 8 to 16% of thefields. Conventional tillage (<30%residue cover) was rarely used in SnyMagill and Squaw Creek, but wasused on 30% of cropland in SouthFork. Erosion losses from crop fieldsin South Fork are not a major con-cern in the flat, till plain portion ofthe watershed. In areas of the North-ern United States, with relatively flatterrain and poorly drained soils,many producers still view conven-tional tillage as the most viable prac-tice because the exposed soil iswarmed faster in spring, often allow-ing earlier seeding and emergence ofthe crop.

In the Sny Magill watershed, con-tour farming, terraces, and otherengineered structures are prevalentpractices for reducing sediment lossesfrom the steeper slopes in that water-shed. Although row crop fields occu-pied only 26% of the land area inSny Magill, most are terraced (77%)and/or farmed using contour plant-ing (92%). Other engineered conser-vation practices are also used exten-sively throughout the Sny Magill

Figure 4. Eroding streambanks and erodible soils are possible contributors to sediment loads. In the South Fork water-shed, both cut-bank meanders and erodible soils are mostly found in the lower (eastern) part of the watershed.


watershed, including a total of over150 sediment basins and grade stabi-lization structures. Terraces are not ascommon in Squaw Creek (23%), buthalf the farm fields are planted oncontours. Terraces and contour farm-ing are not common in the SouthFork watershed; fields with terracesoccupy less than 10% of the water-shed’s cropland.

A Tale of Three Watersheds - revisitedIn this tale of three Iowa watersheds,significant differences in NPS pollut-ants and practices emerged in a stateconsidered by many to be uniformlyagricultural. Much of the differencescan be attributable to their uniquelandform history that has beenexploited uniquely for intensive rowcrop and livestock development. Inthe South Fork watershed, extensivewetlands on recently glaciated tillplains were drained by settlers, andagricultural development then inten-sified during the past century. Theland is well suited for crop and live-stock production, but subsurface tiledrainage increases losses of nitrate,and the rapid routing of tile dis-charge, combined with surface run-off, may enhance movement of bac-teria, P, and sediment. In SquawCreek, with steeper slopes in rowcrops, conservation practices such asreduced tillage and contour farmingmethods are more prevalent. How-ever, losses of nutrients, sediment,and fecal bacteria remain as majorconcerns in the watershed, possiblybecause hydrologically sensitive areasare used for row crops or grazing.Row crop acreage in the Sny Magillwatershed constitute only about 25%of the land area and most row cropfields have conservation tillage orstructural practices such as terraces.However, the steep slopes and karst

drainage in the watershed make SnyMagill watershed perhaps the mostvulnerable among the three streamsevaluated.

Apparent in this tale of threeIowa watersheds is that, in order toprovide the greatest return on thepublic’s investment in conservation,it is imperative that practices be tai-lored to the most local of landscapeconditions and landowner objec-tives. Targeting is needed to placespecific conservation practices on theland to either reduce pollutant con-centrations or attenuate their trans-port. No single practice can beviewed as the answer in all cases, anda one-size-fits-all approach is likelydoomed to failure, or at least doomedto provide little return on the publicinvestment. Recent advances inassessment technologies and recordkeeping are only now beginning toallow us to understand the distribu-tion of practices on the land andtheir impacts on water quality. Sig-nificant challenges remain to developbetter assessment, monitoring, andmodeling techniques to capture theinherent differences among ourwatersheds in order to design conser-vation practices and programs pro-viding greater water quality benefitsfor lower cost. The challenges are notonly in assessing resource needsagainst the mosaic of land use andterrain that occur within watersheds,but also to then develop better policyand planning tools that can helpachieve watershed-scale conservationgoals through implementation at theindividual farm scale.

For More InformationBecher, K.D., Kalkhoff, S.J.,

Schnoebelen, D.J., Barnes, K.K., & Miller, V.E. (2001). Water-Quality Assessment of the Eastern Iowa Basins – Nitrogen,

Phosphorus, Suspended Sediment, and Organic Carbon in Surface Water, 1996-98. U.S. Geological Survey Water-Resources Investigations Report 01-4175. Iowa City, IA.

Burkart, M.R., Simpkins, W.W., Morrow, A.J., & Gannon, J.M. (2004). Occurrence of total dissolved phosphorus in surficial aquifers and aquitards in Iowa. Journal of the American Water Resources Association, 40(3), 827-834.

Fields, C.L., Liu, H., Langel, R.J., Seigley, L.S., Wilton, T.F., Nalley, G.M., Schueller, M.D., Birmingham, M.W., Wunder, G., Polton, V., Sterner, V., Tial, J., & Palas, E. (2005). Sny Magill Nonpoint Source Pollution Monitoring Project: Final Report, Iowa Department of Natural Resources, Geological Survey Bureau Technical Information Series 48, Iowa City, IA.

Green, C.H., Tomer, M.D., DiLuzio, M., & Arnold, J.G. (2006). Hydrologic evaluation of the Soil and Water Assessment Tool for a large tile-drained watershed in Iowa. Transactions American Society of Agricultural and Biological Engineers, 49(2), 413-422.

Helmers, M.J., Isenhart, T.M., Kling, C.L., Mooreman, T.B., Simpkins, W.W., & Tomer, M. (2007). Agriculture and water quality in the Cornbelt: Overview of issues and approaches. Choices 22(2), this issue.

Mausbach, M.J., & Dedrick, A.R. (2004). The length we go: Measuring environmental benefits of conservation practices. Journal of Soil and Water Conservation, 59(5), 96A-103A.


Prior, J.C. (1991). Landforms of Iowa. Iowa City, IA: University of Iowa Press, 153 pp.

Schilling, K.E., Hubbard, T., Luzier, J., & Spooner, J. (2006). Walnut Creek watershed restoration and water quality monitoring project: final report. Iowa Department of Natural Resources, Geological Survey Bureau Technical Information Series 49, Iowa City, IA. Available online: http://www.igsb.uiowa.edu/.

Schilling, K.E., & Libra, R.D. (2000). The relationship of nitrate concentrations in streams to row crop land use in Iowa.

Journal of Environmental Quality, 29(6), 1846-1851.

Schilling, K.E., & Wolter, C.F. (2005). Estimation of streamflow, baseflow and nitrate-nitrogen loads in Iowa using multiple linear regression models. Journal of American Water Resources Association, 41(6), 1333-1346.

Keith E. Schilling ([email protected]) is Research Geologistand Calvin F. Wolter ([email protected]) is Geologist, IowaGeological Survey – Iowa Depart-ment of Natural Resources, Ames, IA.Mark D. Tomer ([email protected]) is Research Soil Scientist

and Thomas B. Moorman ([email protected]) is Research Leader,USDA-ARS, National Soil TilthLaboratory, Ames, IA. Philip W. Gas-sman ([email protected]) isAssociate Scientist, Center for Agri-cultural and Rural Development,Cathy L. Kling ([email protected])is Professor, Department of Econom-ics and Center for Agricultural andRural Development, Thomas M.Isenhart ([email protected]) isAssistant Professor, Department ofNatural Resource Ecology and Man-agement, and William W. Simpkins([email protected]) is Professor,Department of Geology and Atmo-spheric Sciences, Iowa State Univer-sity, Ames, IA.

Watershed Highlight 1: Historical and Human Dimension: Squaw Creek and Walnut Creek Paired Watershed Study.

Because Squaw Creek represents typical agricultural land management in Southern Iowa, the watershed was selected to be the control basin for a large land use experiment occurring in the neighboring Walnut Creek watershed (Figure 5). In the Walnut Creek watershed, large tracts of row-cropped land are being recon-structed to native prairie at the Neal Smith National Wildlife Refuge by the U.S. Fish and Wildlife Service. Before restoration began, land cover in both water-sheds was about 70% row crop. From 1992 to 2005, nearly 220 acres of prairie was planted each year, so that by 2005, native prairie occupied 23.5% of Walnut Creek watershed. Surface water samples collected in the treatment (Walnut Creek) and control (Squaw Creek) watersheds from 1995 to 2005 documented the effects of prairie restoration on water quality (Schilling et al., 2006). Stream nitrate concentrations were found to have decreased 1.2 mg/L over the 10-year project period at the Walnut Creek outlet, with nitrate concen-trations decreasing up to 3.4 mg/L over the same time period in one monitored subbasin with substantial landuse conversion. Interestingly, land use in the con-trol basin of Squaw Creek did not remain static during the same 10-year monitoring period. Row-crop land area increased 9.2% in Squaw Creek as lands previously enrolled in the Conservation Reserve Program as grass-land were converted back to row crop production in the late 1990s. Stream nitrate concentrations increased 1.9 mg/L at the Squaw Creek outlet, with annual nitrate in one monitored subbasin increasing nearly 12 mg/L in 10 years where substantial acres were converted to row crops. These results attest to the sensitivity of water quality parameters to changes in watershed management that are, in aggregate, the result of indi-vidual landowner decisions.

Figure 5. Extent of Prairie plantings in the Neal Smith National Wildlife Refuge within the Walnut Creek watershed.


Three separate projects were carried out spanning the time period of 1988 to 1999 to improve water quality in the Sny Magill Creek watershed. The cumulative adoption percentages and total levels of key BMPs implemented during the 1990s through the Sny Magill Hydrologic Unit Area (HUA) and Sny Magill Creek Watershed projects are listed for selected years in Table 1. The cumulative adoption of terraces in the watershed is also shown in Figure 6 for 1991, 1995, and 2005. A paired watershed approach was used to assess Sny Magill Creek water quality improvements from 1992 to 2001 (Fields et al., 2005). Analysis of Sny Magill stream flow and water quality data collected during 1991-2001 was performed using a pre/post statistical model.

The statistical results indicated that discharge at the watershed outlet increased by 8% over the 10-year period; this could partly be due to routing of runoff water captured by terraces into surface inlet drains (that are often installed just upslope of a terrace) and to the stream. The statistical analysis also showed that the BMPs installed during the 1990s resulted in a 42% decrease in turbidity but only a 7% decrease in total suspended solids (TSS). The TSS results imply that stream bed and bank erosion continued to contribute significant sediment loads to Sny Magill Creek, even after BMP installation reduced sediment delivery from upland areas. The increase in discharge may have further magnified the in-channel sedi-ment contributions. Overall, the TSS results suggest that a long lag time may occur before the full impacts of the installed BMPs are realized.

The statistical analysis also revealed that an increase in nitrate concentra-tions of 15% was found at the SMCW outlet. This indicates greater N leach-ing, which is consistent with increased infiltration of rainfall that naturally results when conservation practices successfully decrease sur-face runoff. However, the nitrate con-centration level still only slightly exceeded 3 mg/L at the end of the 10-year time period, which is quite low compared with the concentra-tions measured in most other Iowa stream systems, including the South Fork and Squaw Creek watersheds.

Table 1. Cumulative percentages of total BMP adoption that was cost shared by year (expressed as a percentage of the total amount implemented as given in the bottom line).

Year Terrace Subsurface tile Sediment basin Grade stabilization Field border Contouring

1992 28 22 28 92 16 11

1995 65 65 79 93 99 53

1998 95 94 98 100 100 100

2001 100 100 100 100 100 100

Total Units 269,585 ft 160,345 ft 61 total 90 total 26,700 ft 1,907 ac

Figure 6. Cumulative additions of terraces to specific land tracts in the Sny MagillCreek Watershed.

Watershed Highlight 2: Long-term Implications of BMP Implementation in the Sny Magill Creek Watershed.


The Conservation Title of the 1985 Farm Bill (Food Security Act) included provisions to reduce soil erosion on highly erod-ible land (HEL) through conservation practices such as Conservation Reserve Program (CRP) plantings and reduced tillage. Land enrolled in CRP was planted to perennial, non-harvested vegetation for at least a ten-year period in exchange for annual rental payments. Soil survey data, including slope, soil texture and depth are used to identify HEL. Those producers farming on HEL-dominated fields were to employ reduced tillage practices to remain eligible for USDA commodity pro-grams; this was known as the conservation compliance provision of the 1985 Farm Bill.

A one-time inventory of conservation practices in the South Fork watershed was conducted during 2005. We compared the distribution of no-tillage management and CRP plantings with the distribution of HEL, which occupies 12% of the watershed (Figure 7). Very little (2.4%) of the watershed’s cropland had been enrolled into CRP by producers, partly because this is some of the most productive rain-fed agricultural land in the U.S. While CRP has also been used to install buffers along streams and around livestock facilities, there has been apparent success in targeting of CRP towards HEL. That is, the proportion of HEL in the watershed in CRP is 4.6%, as opposed to only 2.2% of non-HEL (Table 2). The same is true of no-tillage practices that are highly effective in controlling erosion: although relatively few producers in this water-shed have implemented no-tillage, largely due to concerns about planting delays during wet, cool spring conditions, a greater proportion of HEL (11.3%) is under no-tillage than is non-HEL (6.7%). There is little comparative data to evaluate

whether these practices are better targeted towards HEL in this watershed than in other areas. However, targeting success may also be indicated if conventional tillage prac-tices that increase soil susceptibility to ero-sion have shifted away from HEL as these conservation practices were implemented. This does not appear to be the case, as con-ventional tillage occupies nearly the same proportion of HEL and non-HEL cropland, to within 2%.

It is important to note that in 1985, when current policies where initiated, most of this watershed was probably tilled conven-tionally. This inventory offers only a snap-shot of conservation practices. Current conservation policies, which have had a goal of controlling soil erosion from the most sensitive soils for 20 years, have encouraged better management on the most vulnerable lands in the watershed. Yet, the least desirable tillage practices apparently have not preferentially shifted away from HEL. This raises questions about social-behavioral responses to conserva-tion policies, which are made by individual producers, yet in sum determine the impact of those policies in each watershed.

Table 2. Comparative distributions of CRP and no-tillage conservation prac-tices on highly erodible and non-highly erodible lands in the South Forkwatershed, along with conventional tillage practices.

Management HEL (12%) Non- HEL (88%) Total (100%)

Conservation Reserve Program 4.6% 2.2% 2.4

No-tillage 11.3% 6.7% 7.2

Conventional tillage 28.0% 30.0% 29.8

Figure 7. Distribution of key conservation practices for erosion control inthe South Fork watershed, compared to the distribution of Highly ErodibleLand.

Watershed Highlight 3: Evaluating Targeting of Conservation Practices in the South Fork Watershed.





2nd Quarter 2007 • 22(2)


Privatizing Ecosystem Services: Water Quality Effects from a Carbon MarketBy Silvia Secchi, Manoj Jha, Lyubov Kurkalova, Hongli Feng, Philip Gassman, and Catherine Kling


With the specter of a new farm bill on the horizon, con-siderable discussion is occurring concerning the possibleredirection of conservation programming and financing.Notably, interest in the increased use of incentive systemsand market-like instruments continues to expand. Onesource of this interest lies in the desire to shift some of theburden of providing ecosystem services, such as protectingstream and river channels from erosion, maintainingbiodiversity, and providing clean water and air, to privatesector pockets. For example, in the fall of 2006, USDAand EPA announced a joint partnership to supportexpanded water quality credit trading for nutrients in theChesapeake Bay watershed, allowing farmers to receivecompensation for water quality improvements. Carbonmarkets, such as the active program in the EuropeanUnion, are also being discussed as a possible model forexpanded market-like programs in agricultural conserva-tion policy.

While the potential cost effectiveness of providingenvironmental goods from incentive-based methodsappears to be broadly understood, there is an additionalattribute that is less broadly acknowledged: due to numer-ous inter-linkages in natural ecosystems, the developmentof a market that provides one ecosystem service may sig-nificantly change the level of provision of other ecosystemservices. Thus, by developing the institutional structure tosupport and encourage the provision of one ecosystem ser-vice, changes, either positive or negative, in other servicesmay result.

The example we consider here is the case in which acarbon market that would allow U.S. farmers to receivepayment for sequestering carbon when they retire landfrom production is implemented. This could occur if the

United States were to unilaterally implement such a mar-ket, or if at some future time the United States chose tosign on to a Kyoto-like accord, where carbon sinks wereallowed to generate credits that could be traded to meetmandatory carbon reduction requirements. Under such ascenario, land retirement decisions would be driven by theprices paid for carbon and the amount of carbon that aparticular parcel could sequester (we abstract from theimportant question of measuring the carbon storagepotential of each parcel see the excellent work of Mooneyet al., 2004).

Removing a parcel of land from production willchange the suite of environmental benefits associated withthe parcel. In many cases, these effects are likely to be pos-itive for example, taking land out of active agriculturalproduction and placing it in perennial cover or forestedlands will usually yield reduced erosion and nutrient run-off relative to row crop agriculture. Indeed, the findings ofour study are consistent with this outcome. However, ifconservation practices are already in place on a workingland field, water quality improvements from retiring theland might be small and, in fact, could be negative. Thelatter could occur if land retirement results in the plantingof land cover that, on the whole, is not as effective in cap-turing nutrients and sediments as a working land systemthat already has effective conservation practices and man-agement in place.

In this paper, we consider the possible water qualityconsequences of a carbon trading policy that allows farm-ers to receive carbon credits from retiring their land fromagricultural production. To do so, we make a number ofsimplifying assumptions about the structure of the carbonmarket and the choices farmers make in response to the


existence of that market; many ofthese assumptions may not, in fact,represent how an actual marketmight be implemented. Rather thanview the results of this analysis asdefinitive, we present the findings inthe spirit of raising awareness of thepotential environmental conse-quences that can occur when a singleenvironmental benefit or target (car-bon sequestration) forms the basis ofenvironmental policy, as would bethe case if carbon trading marketsthat allowed land retirement to yieldcarbon credits were functioning withhigh carbon prices.

A Bit about Our Data and ModelsTo develop our models, we drawheavily from the National ResourceInventory to provide data on the landuse, cropping history, and farmingpractices in the state of Iowa. Thisinventory is the most comprehensivedata set on land use in the UnitedStates, and we use data on the 14,472physical points in Iowa that representcropland. Conceptually, our data andmodels are based on individual pro-

ducer and farm-level behavior, andwe treat an NRI point as a producerwith a farm size equal to the numberof acres represented by the point (theexpansion factor provided by the sur-vey). Figure 1 illustrates the 35watersheds corresponding to theUnited States Geological Survey 8-digit Hydrologic Cataloging Unitsthat are largely contained in the stateand are modeled in this study. Tocompute the amount of carbonsequestered when a land unit isretired from cropland, we rely onestimates from the EnvironmentalPolicy Integrated Climate Model ver-sion 3060. When land is retired fromcrop production, we assume thatannual grasses are planted and main-tained on the land, and we run a 30-year simulation to predict the carbonsequestration level associated withthis change.

In addition, we also rely on esti-mates from a watershed-based modelto assess the conservation policies.Unlike carbon sequestration, thedegree to which land retirementimproves in-stream water qualitydepends on critical interactions

between land uses in different loca-tions within a watershed. Thus, forotherwise identical tracts of land,more water quality improvementmay occur from retiring a piece ofland from production that is locateddownstream from numerous othercropped points relative to one that isnot. The potential filtering effect isjust one example of the physical pro-cesses that need to be captured toassess the in-stream water qualityeffects of land retirement.

So that we can capture these landuse interactions within a watershedsetting, we employ the Soil andWater Assessment Tool, a biophysicalwater quality model, to estimatechanges in nitrogen, phosphorous,and sediment loads from retiring aparticular set of parcels from produc-tion within a watershed. To estimatethe in-stream water quality conse-quences of the increase in land setaside, we have calibrated the waterquality model for each of the water-sheds identified in Figure 1 to base-line levels (Jha et al., 2005; Gassmanet al., 2005). By running the modelat the set-aside levels “after” the pol-icy, we can compute the changes inwater quality attributable to theincrease in land set-aside. The water-sheds studied correspond to 13 out-lets, at which the in-stream waterquality is measured. The water qual-ity measures of interest are sediment,nitrogen, and phosphorus.

Water Quality Effects of a Carbon MarketTo demonstrate the possible conse-quences of a carbon market that paysfarmers for the sequestration of car-bon in agricultural soils on waterquality, we consider a simple sce-nario. Suppose that through an activecarbon market, the price of carbon issuch that about 10% of Iowa crop-

Figure 1. Study area and watershed delineations.


land is retired; suppose further thatthe cost of retiring all land within thestate is about the same. While rentalrates for farmland do vary across thestate, they vary relatively little withrespect to productivity (see Secchi &Babcock, forthcoming), and this sim-plifying assumption allows us tofocus on environmental outcomes ofthe scenario without overly compli-cating the analysis. Under theseassumptions, the cropland that willbe removed from production will bethe land that produces the highestcarbon sequestration benefits per acreas this land will earn the highestreturn from carbon sequestrationcredits. In consequence, the landremoved from production may ormay not represent significant por-tions of the watersheds under consid-eration.

Based on this scenario, the landretired would be focused in the cen-tral part of Iowa, in the ecoregionknown as the Des Moines Lobe, a flatarea, with very productive agricultureand particularly suited for carbonsequestration. Figure 2 illustrates thequantity and location of the carbonthat the carbon simulation modelpredicts would be sequestered acrossthe state under this scenario. Approx-imately 2 million acres of land isremoved from production under thisscenario with about 2.7 million tonsof carbon being sequestered annually.

Does this land retirement,induced by a private market that paysfor ecosystem services, yield otherenvironmental benefits to the region?To answer this question, we estimatethe in-stream water quality effects ofthis land retirement using the waterquality model and present the per-centage reductions in three commonindicators in Figures 3-5.

Figure 3 reports the estimated in-stream sediment reductions from theretirement of this set of land parcels.

We find that for the two largestwatersheds whose sediment dynamicsis influenced by the presence of largereservoirs, the Des Moines and IowaRiver watersheds, there is only a littleimprovement in sediment. In con-trast, there are larger reductions insediment in the South-Western partof the state, likely because the land

that is retired from production thereis more erodible. In general, however,sediment reductions in percentageterms are lower than the reductionsin nutrients, because land that hashigh carbon sequestration potentialalso has good productivity levels andis, therefore, more heavily fertilized.

Figure 2. Carbon sequestration from carbon trading.

Figure 3. In-stream sediment changes from carbon trading (percentagereduction).


The retirement of land generallyimproves the total N level as seen inFigure 4. The reason for theimproved N levels in many cases, asmentioned above, is that the landtaken out of production is largelyprime agricultural land: heavily fertil-ized and reliant on tiled drainage sys-

tems. Some of the highest reductions,a total of over 10,000 tons annualaverage, are in the Des Moines Riverwatershed, which comprises largeparts of the Des Moines Lobe andincludes some of the most productiveland in the state where most of theacres of land retirement are located.

Finally, Figure 5 reports theresults for the in-stream phosphorouslevels predicted to occur as a result ofthe carbon trading program. Like thesediment results, the Western water-sheds show the highest improve-ments. This is not surprising giventhe sediment results, since phospho-rous typically moves with sediment.

The development of more mar-ket-like programs to provide ecosys-tem services from agriculture is aconcept with expanding interest. Inthis paper, we have identified anadditional issue associated with thisstrategy, changes in other environ-mental goods of interest. In the caseanalyzed here, these changes were allpositive; thus, the market-based sys-tem yields positive gains for otherecosystem services. By recognizingthat a system that pays for carbonsequestration via land retirementpotentially has effects on other envi-ronmental services, and that the spa-tial distribution of different environ-mental services is likely to differ,policy makers can incorporate theseeffects in planning and implementingmarkets for ecosystem services.

For More InformationFeng, H., Kurkalova, L.A., Kling,

C.L., & Gassman, P.W. (2007). Economic and environmental co-benefits of carbon sequestration in agricultural soils: Retiring agricultural land in the Upper Mississippi River Basin. Climatic Change, 80(1-2), 91-107.

Gassman P.W., Secchi, S., Kling, C.L., Jha, M., Kurkalova, L., & Feng, H. (2005). An Analysis of the 2004 Iowa Diffuse Pollution Needs Assessment using SWAT. Proceedings of the SWAT 2005 3rd International Conference, 11-15 July, Zurich, Switzerland.

Figure 4. In-stream nitrogen changes from carbon trading (percentagereduction).

Figure 5. In-stream phosphorous changes from carbon trading (percent-age reduction).


Jha, M., Gassman, P.W., Secchi, S., & Arnold, J.G. (2005). Nonpoint source needs assessment for Iowa part I: Configuration, calibration and validation of SWAT. Watershed Management to Meet Water Quality Standards and Emerging TMDL (Total Maximum Daily Load), Proceedings of the Third Conference 5-9 March 2005 Atlanta, Georgia. American Society of Agricultural Engineers, St. Joseph, Michigan. pp. 511-521.

Mooney, S., Antle, J.M., Capalbo, S.M., & Paustian, K.H. (2004). Influence of project scale and

carbon variability on the costs of measuring soil carbon credits. Environmental Management, 33(S1), S252-S263.

Plantinga, A.J., & Wu, J. (2003). Co-benefits from carbon sequestration in forests: Evaluating reductions in agricultural externalities from an afforestation policy in Wisconsin. Land Economics, 79(1), 74-85.

Secchi, S., & Babcock, B.A. (2007). Impact of corn prices on land set-aside: The case of Iowa. Forthcoming CARD working paper.

Silvia Secchi ([email protected]) isAssociate Scientist, Manoj Jha([email protected]) is Assistant Sci-entist, Hongli Feng ([email protected]) is Associate Scientist, and PhilipGassman ([email protected]) isAssociate Scientist, Center for Agri-cultural and Rural Development,Iowa State University, Ames, IA.Lyubov Kurkalova ([email protected]) is Associate Professor, School ofBusiness and Economics, NorthCarolina A&T State University,Greensboro, NC. Catherine Kling([email protected]) is Professor,Department of Economics and Cen-ter for Agricultural and Rural Devel-opment, Iowa State University, Ames,IA.

mailto:[email protected]








2nd Quarter 2007 • 22(2)


Nitrate Reduction ApproachesBy Christopher Burkart and Manoj K. Jha


As noted in the overview to this set of papers, water qual-ity continues to be a growing concern. Nutrients appliedas commercial fertilizer and manure enter surface andground water, leading to several forms of water qualityimpairment. These impairments manifest themselves in anumber of ways. Excess phosphorus is responsible foralgae blooms, losses in water clarity, and even the presenceof toxic cyanobacteria in fresh water. Excess nitrogen isbelieved to be the limiting factor in low-oxygen dead zonesin several dozen locations around the globe. In somelocales, nitrate concentrations reach levels that are toxic toboth humans and aquatic animals. In the United States,local nitrate concentrations are largely uncontrolled. Theonly widely applied standard affects water used for humanconsumption. This is regulated by the Environmental Pro-tection Agency via National Primary Drinking Water Reg-ulations (EPA NPDWR). Similar requirements and guide-lines exist in Canada and Europe.

Several technologies can remove nitrates directly fromwater and are employed by municipal water works in orderto comply with drinking water standards during periods ofhigh nitrate concentrations in source water. These technol-ogies are costly to operate, suggesting an opportunity forcost savings via upland reductions of fertilizer application.This article explores possible tradeoffs in the context of anutrient-application-right trading scheme. Simulations ofboth water quality and economic effects in a test water-shed suggest that simple upland fertilizer reductions aremore costly than direct nitrate removal if the goal is com-pliance with drinking water standards. Other water qualitygoals merit consideration, but are difficult to model with-out objective standards and given the current nitrateremoval technology.

Watershed Background The area used for simulation is the Raccoon watershed,located in the state of Iowa in the United States. The Rac-coon River is the main stream for the watershed and drainsa large area containing an abundance of fertile soil. Thetotal area of the watershed is approximately 2.3 millionacres, 1.7 million of which are devoted to rotations of cornand soybean production. Nitrogen and phosphorus fertil-izer are applied at high levels on the corn crop and consti-tute the primary nonpoint nutrient pollutant source in thewatershed. Figure 1 shows a land-use map of the water-shed. The outlet of the watershed is near the capital city ofDes Moines, which along with other municipalities in thearea, uses the Raccoon River as a source of drinking water.The Des Moines Water Works is the supplier of drinkingwater and currently operates the world’s largest denitrifica-tion facility.

In-stream nitrate levels frequently exceed the maxi-mum allowed concentration of 10 milligrams per liter. Inthese instances, source water is run through the denitrifi-cation facility before being treated for use as drinkingwater. The facility uses an ion exchange process which pro-duces waste water with a high saline content in addition tothe nitrate removed. This waste water is currently dis-charged downstream at no cost to the facility. Downstreammunicipal water supplies are not adversely impacted bythis discharge, as they are able to meet their water needsfrom deeper ground water aquifers. For purposes ofNPDWR compliance this is not an issue, and the dis-charge is permitted by the EPA under the National Pollut-ant Discharge Elimination System.

The nitrate removal facility was constructed in 1990 ata cost of approximately $3 million. The scrubbers andmedia were the primary components of this large sunkcost, and would also be the bulk of the cost associated with


an expansion of the facility unlessanother removal technology wereemployed. Current processing vol-ume does not appear to requireexpansion in the near term, and therehas been no observed deterioration ofthe scrubber components. Operatingcosts of the facility are approximately$300 per million gallons of water,with a capacity of 10 million gallonsof water per day. In an average year,the facility runs approximately 50days.

Modeling ApproachWhile drinking water standards aregiven high importance due to theirdirect effects on human health, highnitrate levels cause other problems.However, control of ambient water

pollution in this watershed is stillbeing developed, and there are noexisting regulations outside of drink-ing water standards. Amelioratingproblems such as hypoxia and nitratetoxicity for aquatic animals wouldrequire both a lower threshold fornitrates and complete removal of thenitrate from the watershed. Meetingthe latter requirement with the tech-nology currently used for drinkingwater purposes is inappropriate as itreintroduces the nitrate to the envi-ronment. The analysis here proceedsin the framework of existing regula-tions and the technology currently inplace, but it is important to note thatthere are other impacts that meritconsideration: namely, the effects ofnitrate levels outside of drinkingwater considerations. Upland fertil-

izer reductions prevent nitrates fromentering waterways in the first place,and have positive effects beyond con-tributing to drinking water standardcompliance.

The goal of the modeling frame-work is to capture changes in waterquality generated by implementationof policy, as well as the associatedeconomic effects. This requires thecoupling of an economic model witha physical model. Nutrient applica-tion levels predicted by the economicmodel are used to supply land-useinputs to the physical model. Theoutput from the physical model inturn provides the water quality mea-sure of interest: nitrate concentra-tion over time. A hydrologic model isused to link the effects of upland fer-tilizer reductions to direct nitrateremoval at the outlet. The watershed-based Soil and Water AssessmentTool (SWAT) simulates the effects ofwatershed management on waterquality and water flow on a dailytime step. It is primarily used formodeling nonpoint source contribu-tions to nutrient and sediment loadswithin a watershed. The SWATimplementation employed uses datafrom the National Resources Inven-tory (NRI) to populate the watershedwith spatially detailed information. Apoint in the NRI effectively repre-sents a farm. Site-specific nutrientapplication data are generated by theeconomic model. The economicmodel predicts nitrogen fertilizerapplication rates based on prices ofcorn and fertilizer and a site-specificsoil characteristic. It also predictsyield, and thus returns to fertilizerapplication. Changes in nitrogen fer-tilizer prices, for example, via a tax onfertilizer or a cap on application, willcause a loss in returns for the farmer.This provides a measure of the costimposed by the policy. Data used toconstruct the model comes from a

Figure 1. Land use in the Raccoon watershed.


farm operator survey, the Agricul-tural Resource Management Survey,historical prices, and from a detailedsoil grid.

Policy Simulations Three scenarios are run through themodeling system described above.One is a baseline in which the eco-nomic model leaves prices and nitro-gen fertilizer applications unchanged,and the water quality model predictsthe associated nitrate concentrationsat the watershed outlet. The othertwo scenarios represent reductions infertilizer applications simulated bythe imposition of a nonpoint sourcetrading scheme. This scheme worksas follows: each farm is allocated fer-tilizer application permits for thetotal acreage it farms; for example, a100-acre farm might receive 12,000pounds worth of permits if the per-mit level is 120 pounds per acre. Afarm has three choices in using its

permits. One is to apply exactly thepermitted amount. Another is toapply less than permitted, and sellthe surplus permits to the thirdgroup, those who purchase permits inorder to apply at greater levels thaninitially permitted. Farmers maketheir choice of total applicationaccording to the model, taking intoaccount the prices they face, their soiltype, and the market price of a per-mit, which is determined by the dis-tribution of farmer types. The totalwatershed application is reduced aslong as the total permit allocation issmaller than the total amount origi-nally applied. For purposes of simula-tions, this is done at two levels of per-mit allocations. From a baselineaverage application rate of 135pounds per acre, one scenariorestricts the per-acre permit alloca-tion to approximately 120 poundsper acre and results in a simulated6% reduction in annual load ofnitrate at the watershed outlet. The

other restricts the allocation toapproximately 108 pounds per acreand results in an approximate 12%reduction in annual nitrate load.These reductions are the result of thetotal mass of nitrogen being appliedin the watershed being reduced.

Imposing the permit restrictionsbenefits those farmers who can sellexcess permits, but increases the costsof those who must purchase addi-tional permits. Since the totalamount of nitrogen application isbeing reduced, the net result is a lossfor farmers in the watershed as awhole. Loss or gain from the policyscenarios can be measured for indi-vidual farms and then aggregated tothe watershed level to gauge the costof the policy. Under the small reduc-tions, the total farm watershed loss isapproximately $161,000, and underthe larger reductions, losses areapproximately $700,000.

To compare the water qualitychanges resulting from the imple-

0.0

1.0

2.0

3.0

4.0

5.0

6.0

7.0

8.0

9.0

10.0

11.0

Jan Feb March April May June July Aug Sept Oct Nov Dec

NO 3

-N C

once

ntra

tion

(mg/

L)

Baseline (56 run-days)

Small reduction (51 run-days)

Large reduction (48 run-days)

Figure 2. Nitrate loads by monthly average.


mentation of these policies to opera-tion of the nitrate removal facility,the water-quality model is run on adaily time step and the nitrate con-centration for each day recorded. Thetrigger concentration for the nitrateremoval facility to run is 9mg/L (thelegal limit is 10mg/L). Under thebaseline scenario, that level wasexceeded 56 days of the year. Thesmall and large reduction scenariosreduced the number of run-days to51 and 48. Figure 2 shows a sum-mary of nitrate loads by monthlyaverage. Saving days of operation forthe nitrate removal facility impliescost savings and illustrates the short-ening in the number of run-daysrequired to maintain a safe level ofnitrate. The energy, labor, and rawmaterial costs of one run-day areapproximately $3,000. The lifetimeof the media used in the removal pro-cess is currently uncertain, making itdifficult to calculate the true cost ofoperation. The original media is stillin use and shows no sign of deteriora-tion after 14 years of use. As nitrateloads and water demand grow, theremay be a need for expansion in thefuture, involving significant capitalcosts and raising the cost of a day ofoperation. Such expansion may alsoinvolve a change in nitrate removaltechnology.

Trading nitrogen permitsbetween point and nonpoint sourcescan lower costs of reductions (Ran-dall & Taylor, 2000). This is usuallyconsidered in the context of a non-point source generating excess per-mits by purchasing upland reduc-tions. In that type of tradingarrangement, a trading ratio is estab-lished to equilibrate a pound ofupland reduction to a pound of pointsource discharge. Conceptually, thisapproach could work in reverse aswell: nonpoint sources could gener-ate permits for themselves by paying

for the removal system. While thesetrading opportunities are attractivepossibilities, a quick look at the dif-ference in costs in this case suggeststhat it would be much more efficientto simply run the nitrate removalfacility a few extra days rather thanimplement any restrictions on farmerapplication. Five run-days at $3,000per day is $15,000, far less than the$160,000 in losses that would beincurred by farmers. Eight run-daysof the nitrate removal facility are like-wise much less expensive than the$700,000 in losses associated withthe stricter cap-and-trade policy.

While the upland fertilizer reduc-tions examined here are more costlythan direct nitrate removal, this anal-ysis does not take into account otherpossibilities. There are also concernsbeyond drinking water standards,such as hypoxia and low-level nitratetoxicity (Camargo et al., 2005), thathave important impacts on ambientwater quality. Perhaps because drink-ing water issues pose the most imme-diate threat to human health, it is theonly form of existing pollution regu-lation that impacts this watershed. Asnew standards with broader impactsin mind are developed, such as TotalMaximum Daily Loads, this analysiscan be revisited, possibly with differ-ent conclusions. The upland reduc-tions have an effect on the ambientand downstream nitrate loads thatthe removal process does not andwould be more effective at meetingexpanded standards. Even if undermore comprehensive standardsupland reductions become more costeffective, there would be transactioncosts involved in any trading schemethat would also need to be consid-ered.

There are also combinations ofreduction strategies that could resultin superior reductions with similarcosts, even in the existing framework.

Coupling reductions with bufferstrips, grassed waterways, changes intillage, and application timing all cancontribute to reductions in nutrientloads to a watershed. In addition, amore complete comparison wouldrequire information on possible dete-rioration of the nitrate removalmedia and the associated replacementcosts, though these are at presentuncertain.

For More Information Atwood, J., McCarl, B., Chen, C.,

Edelman, B., Nayda, B., & Srinivasan, R. (2000). Assessing Regional Impacts of Change: Linking Economic and Environmental Models. Agricultural Systems, 63, 147-59.

Camargo, J., Alonso, A., & Salamanca, A. (2004). Nitrate toxicity to aquatic animals: A review with new data for freshwater invertebrates. Chemosphere, 58, 1255-67.

Jones, C. (2005). Nitrate and bacteria in the Raccoon River: Historical perspective and 2004 Summary, Des Moines Water Works. Available online: http://www.dmww.com/Laboratory/.

Nusser, S., & Goebel, J. The National Resources Inventory: a long-term multi-resource monitoring programme. Environmental and Ecological Statistics, 4(3), 181-204.

Randall, A., & Taylor, M. (2000). Incentive-based solutions to agricultural environmental problems: Recent developments in theory and practice. Journal of Agricultural and Applied Economics, 32, 221-34.

United States Environmental Protection Agency, National Primary Drinking Water Regulations (EPA NPDWR).


Current Drinking Water Rules. Available online: http://www.epa.gov/safewater/regs.html.

United States Geological Survey, National Water-Quality Assessment (USGS NAWQA). Nutrients in streams and rivers

across the Nation–1992-2001 National Water-Quality Assessment Program. Available online: http://water.usgs.gov/nawqa/.

Christopher Burkart ([email protected]) is Assistant Professor, Depart-ment of Marketing and Economics,University of West Florida, Pensa-cola, FL. Manoj K. Jha([email protected]) is Assistant Sci-entist, Center for Agricultural andRural Development, Iowa State Uni-versity, Ames, IA.






2nd Quarter 2007 • 22(2)


Organic Demand: A Profile of Consumers in the Fresh Produce Marketby John Stevens-Garmon, Chung L. Huang, and Biing-Hwan Lin


Demand for organic produce in the United States hasincreased steadily since the early 1990s. In 2000, for thefirst time, conventional supermarkets sold more organicfood than any other venue (Dimitri & Greene, 2002).According to the Organic Trade Association (OTA),organic food sales in the United States totaled $13.8 bil-lion in 2005, making up 2.5% of the retail food market.This is an increase from 1.9% in 2003 and from 0.8% in1997 (OTA, 2006). This increase coincides with theimplementation of national organic standards by theUSDA in October of 2002, which provided uniform label-ing for consumer recognition. Demand trends areexpected to continue as more conventional retailers takeup a larger portion of the organic market. Sales of organicfoods are estimated to rise to $23.8 billion by 2010 (NBJ,2004).

The phenomenal growth in organic sales in recentyears has brought additional farmland into organic agri-culture industry. Dimitri and Greene (2002) estimatedthat between 1997 and 2001, U.S. farmers and ranchersnearly doubled the acreage of certified organic land, total-ing to 2.3 million acres. With increasing production andsupply of organic produce and meats, organic food, onceconsidered a niche product, has become more availableand affordable for consumers in mainstream grocerystores. It is estimated that 46% of total organic food salesare now handled by the mass-market channel, whichincludes supermarkets, grocery stores, mass merchandisers,and club stores (OTA, 2006). A popular perception tendsto suggest that most organic consumers are white, wealthy,and have young children. However, the consumer base oforganic food appears to have become more diverse andcannot be easily pigeonholed as the market is growing with

increased availability and popularity. A study by the Hart-man Group (2002) found that half of the respondentswho purchased organic food frequently have an annualincome below $50,000, and that African Americans, AsianAmericans, and Hispanics purchase more organic productsthan Caucasians.

Our analysis used the Nielsen Homescan data from2001 and 2004 (Box 1) to determine the characteristics oforganic consumers, what they buy, how much they spend,and the price premiums they pay for organic produce.These two years give us a sample from before and after theimplementation of the National Organic Program’s(NOP) labeling standard. We focus on fresh producebecause produce represents the largest sector, at about39% of the organic market (OTA, 2006). One may specu-late that the growing popularity of organic consumptioncould be attributed at least partially to the implementationof NOP. However, it is not our intention to contribute tothe debate on the effect of NOP, mainly because Homes-can data are not suitable for examining such an issue. Wesimply present a cursory look at the data to examinewhether any notable changes have occurred after NOP bycomparing household purchases of fresh produce in 2001and 2004.

Who Buys Organic Produce?Of all demographic characteristics, race seems to be themost correlated with organic expenditures. In 2001, wefound that Asian Americans, compared to other ethnicgroups, spent the most food dollars to purchase organicproduce on a per capita basis. Though they bought com-parable amounts of fresh produce, Asian Americans, onaverage, spent more on organic produce than White, Afri-


can, or Hispanic Americans. By2004, Asian Americans’ expenditureson organics fell, while White, Afri-can, and Hispanic Americansincreased their spending on organicproduce (Figure 1). Further, AfricanAmericans have replaced AsianAmericans to become the ethnicgroup that spent the most on organicproduce. The proportion of AfricanAmericans who purchased organicproduce also increased from 34% in2001 to 37% in 2004, while the pro-portion of organic users among othergroups have remained relatively thesame. These findings are in generalagreement with the report that Asian,Hispanic, and African Americans arethe ethnic groups more likely to pur-chase organic foods than Whites(Hartman Group, 2002). Accordingto a more recent study by the Hart-man Group (2006), Asians and His-panics are motivated primarily byfamily concerns in buying organicproducts.

Organic expenditures vary byregion. We found that in 2004,households in the Western UnitedStates purchased more organic pro-duce than those residing in otherregions, spending on average about$4.90 per capita. This spending

amount represents an increase of19% over 2001, after adjusting forinflation. Households residing in thenortheastern and southern regionsalso registered an increase in averageper capita spending on organic pro-duce. The Central United Statesshowed the lowest average per capitaexpenditure in 2004, whichremained virtually unchanged from2001. In terms of proportion ofhouseholds that purchased freshorganic produce, the western regionalso showed the largest increase(almost 4%) of organic users from2001 to 2004. The West and Southappear to be the two fastest growingmarkets for organic produce in theUnited States.

According to Homescan data, theaverage per capita spending onorganic produce increased by 12% inreal terms between 2001 and 2004.As shown in Figure 2, this increase inspending is observed for all house-holds across various income groups.It is interesting to note that averageper capita spending on organic pro-duce exhibited a U-shape relation-ship with income for householdsearning less than $45,000 annually.Among households earning $45,000and more, organic spending appears

Box 1. The Sample Data and Description.

The Nielsen Homescan panel data include purchases of both random-weight and Uniform Product Code (UPC) food items. According to Nielsen, the panel consists of rep-resentative U.S. households that provide food purchase data for at-home consumption. For 2001 and 2004, more than 8,164 and 8,430 households, respectively, partici-pated in the Homescan panel. In general, panelists report their pur-chases weekly by scanning either the UPC or a designated code for random-weight (unpackaged) products of all their purchases from grocery stores or other retail outlets. For packaged or UPC-coded food products, organic pro-duce can be identified by the pres-ence of the USDA organic seal or with organic-claim codes created by Nielsen. For random-weight items, the descriptions of desig-nated codes can be used to iden-tify organic produce. Homescan panelists do not report prices they pay for each food; they report total quantity and spending for each food. In addition, the Homescan data include product characteris-tics and promotion information, as well as detailed socio-demo-graphic information of each house-hold. For our analysis, household spending on selected fresh pro-duce was calculated as average expenditures on per purchase record basis. Prices for organic and conventional produce were derived as unit values based on the household’s reported expenditures and quantities. Average house-hold expenditures on fresh pro-duce were expressed in terms of per capita to control for household size effect. Furthermore, all expen-ditures and prices were expressed on nominal current dollars for 2001 and 2004, respectively. The Con-sumer Price Index for food and beverages increased by 7.49% from 2001 to 2004, and this infla-tion rate is used to calculate changes in real terms.

Figure 1. Average per capita spending on fresh organic produce forhome consumption by race.

0

1

2

3

4

5

White Black Hispanic Asian Other

Race

Dolla

rs

2001 2004


Figure 2. Average per capita spending on organic produce for home con-sumption by income class.

to rise with income. These patternsbetween household income andorganic spending are observed forboth 2001 and 2004. It is somewhatsurprising to find that householdswith the lowest income level of lessthan $25,000 spent the most—morethan $4 per capita on organic pro-duce in 2001 and 2004. Further-more, households in the $35,000–$44,999 income bracket spent aboutas much on organic produce per cap-ita as those households earning over$100,000 annually ($3.94 versus$4.09 in 2004). For households withannual income at $25,000 or above,there appear little variations on aver-age per capita spending on organicproduce in 2001 and 2004. Overall,there is little consistent associationbetween per capita expenditures onorganic produce and householdincome. Studies suggest that lowerincome families choose to buyorganic when possible as a means ofpreventative medicine, and thus areat least as likely to purchase organicas other income groups (HartmanGroup, 2003; OTA, 2004).

The lack of a clear positive associ-ation between organic expenditureand income level may have promptedLaurie Demeritt, President of the

Hartman Group, to observe that“income is about the only thing thatdoesn’t skew at all by user and non-user. You get little skews in age, littleskews in geography, little skews ineducation, but there’s nothing at allfor income, so we don’t even look atthat any more” (Fromartz, 2006). Arecent survey conducted by the FoodMarketing Institute (2004) showedthat only 11% of organic shopperspolled bought organics at a natural-food supermarket, while 57% boughtat mainstream grocery stores and dis-count stores. The fact that main-stream grocery stores are replacingthe specialty food stores as the majoroutlets for organic foods couldexplain the seemingly fading relation-ship between organic expenditureand household income. It appearsthat income may no longer be a goodpredicator to profile organic consum-ers as the industry continues to growand evolve into maturity.

What Do Organic Consumers Buy and How Much Organic Premium Do They Pay?According to Homescan, tomatoes,potatoes, carrots, onions, lettuce,apples, oranges, bananas, grapes, and

strawberries were the top five vegeta-bles and fruits in terms of their sharesof fresh produce expenditures forhome consumption. Americanhouseholds spent more on organicproduce between 2001 and 2004 forall produce except oranges and let-tuce. Overall, average per capitaspending on these organic fruits andvegetables increased from $1.64 in2001 to $1.91 in 2004, an increase of8.5% in real terms. Tomatoes appearto be the most favored organic vege-table among American consumerswith average per capita spendingamounts 3–4 times those of otherorganic produce in both 2001 and2004. Per capita spending on organicapples and lettuce held distant sec-ond and third places in 2001, whilecarrots and apples were ranked sec-ond and third, respectively, in 2004.Strawberries and bananas registeredthe largest increases in organic expen-ditures by 45% and 33%, respec-tively.

Since organic agricultural pro-duction is typically more cost inten-sive than conventional agriculture,many organic farmers rely on the pre-miums that organic foods carry tocover their extra costs. High premi-ums usually indicate high demand,signaling to producers which marketsmay be expanded. As indicated previ-ously (Box 1), we calculated unit val-ues (spending over quantity pur-chased) to derive price premiums forselected fresh produce becauseHomescan panelists do not reportprices of organic and conventionalproduce. Thus, the organic premi-ums derived from unit values are notstrictly the same as would beobserved from the unit prices, ifavailable. Except for oranges andonions, average organic premiums forthe most valuable produce increasedfrom 2001 to 2004 (Figure 3). In2001, average organic premiums var-

0

1

2

3

4

5

<25,000

25K-34,999

35K-44,999

45K-59,999

60K-99,999

> 100K

Household Income

Dolla

rs

2001 2004


ied from 1% ($0.01/lb.) above theconventional produce for carrots to78% ($0.32/lb.) for potatoes. Incomparison, organic premiums var-ied from 9% ($0.08/lb.) for orangesto 78% ($0.36/lb.) for potatoes in2004. According to our calculations,organic potatoes carried a substan-tially higher price premium thanother organic produce in both 2001and 2004. This finding can be usefulto organic producers who are lookingfor new crops to improve their profitmargins. The changes in organic pre-miums between 2001 and 2004 wererelatively moderate among the mostvaluable produce, except for lettuceand carrots.

In terms of dollar amount, aver-age organic premiums that consum-ers paid in 2004 for apples, grapes,strawberries, tomatoes, and potatoeswere fairly uniform at about $0.35/lb. above their conventional counter-parts. There are substantial variationsamong individual fresh produce,most notably in carrots and lettuce,which registered the largest increasesin price premiums between 2001 and2004. Tomatoes and apples alsoshowed an increase in average pricepremiums by 52% and 75%, respec-tively. Overall, the average price pre-mium for the selected produceincreased from $0.19/lb. in 2001 to$0.29/lb. in 2004, which represents a42% increase in real terms.

Price plays an important role inconsumers’ purchase decisions. A sur-vey by Walnut Acres (2002) reportedthat 68% of consumers cited highprices as the main reason they didnot buy organic foods. However, tomany organic consumers, price couldbe of secondary consideration. Theyare willing to pay a price premiumbecause they value and demand cer-tain attributes from organic products.To them, the organic attributes arewell worth the price difference. The

fact that we find the organic premi-ums for most selected fresh produceincreased from 2001 to 2004 sug-gests that the demand for organicproduce remains strong, and con-sumers are willing to pay additionaldollars for the organic attribute.

Based on limited data on organicprices over the period 2000–04 at thefarmgate and wholesale levels, Ober-holtzer, Dimitri, and Greene (2005)show that prices for organic varietiesare comparatively more volatile thantheir conventional counterparts andorganic price premiums were higherat the wholesale level than at thefarmgate level. Of the three produce(broccoli, carrots, and mesclun mix)studied, they found that averageannual organic price premiums atwholesale, as a percent of conven-tional prices, increased for carrots(143% to 148%) and for broccoli(141% to 153%) between 2001 and2004, while the price premiumsdecreased for mesclun mix from 9%to 7%. It should be noted that theorganic premiums calculated fromthe Homescan data are not directly

comparable with those reported inOberholtzer, Dimitri, and Greene(2005). However, one would expectrelatively lower organic price premi-ums at the retail level than at thewholesale or farmgate level as organicfoods are becoming more competi-tive and increasingly marketedthrough mainstream supermarketsand discount club stores.

A Profile of Consumer by User GroupIn our analysis, we categorized eachhousehold into user or nonusergroup according to whether or notthe household purchased organicproduce. Then user households areclassified into one of three usergroups based on sample distributionof per capita spending on organicproduce. In 2004, the first quartile oforganic users with per capita spend-ing greater than $0 but less than$0.75 is defined as light users, thesecond and third quartiles are definedas medium users (between $0.76 and$3.65), and the fourth quartile is the

0

20

40

60

80

Apples

Banan

as

Grapes

Strawberr

ies

Oranges

Tomato

es

Potatoes

Lettu

ce

Onions

Carrots

Fresh ProduceO

rgan

ic P

rem

ium

, %

2001 2004

Figure 3. Average price premiums as % of conventional price for the top

10 most valuable organic produce.


heavy users (> $3.65). The nonusersaccount for 62.5% of the 2004 sam-ple, while light, medium, and heavyusers account for 9.6%, 18.6%, and9.4%, respectively. In comparison,the proportion of user groups in2001 are 62.9% (nonusers), 9.5%(light users), 18.4% (medium users),and 9.2% (heavy users). Overall, theresult shows that proportionally moreconsumers have become organicusers in 2004 than in 2001, with aslightly higher increase in bothmedium and heavy user groups. Percapita spending on organic produceby medium users increased from$1.45 in 2001 to $1.81 in 2004; anincrease of 16% in real terms. For thelight and heavy users, the growth inreal per capita organic spendingincreased by 10% from 2001 to2004.

With respect to market shares ofselected organic produce across usergroups, Figure 4 shows that lightusers spent the largest proportion oftheir organic expenditures on carrots,bananas, and tomatoes. The mediumusers purchased more tomatoes andcarrots relative to other kinds of freshproduce, while heavy users seemed to

expend a larger proportion of theirorganic budgets on tomatoes, apples,and grapes. Overall, organic toma-toes appear to be the favorite freshproduce among the organic users,accounting for more than 15% oflight and medium users’ organic pro-duce expenditure and more than10% for heavy users. It is interestingto note that organic vegetables appearto be the preferred organic produceof light users, while the heavy usersseem to have an affinity for organicfruits, especially apples and grapes.Heavy users buy proportionatelymore of each fruit than either thelight or medium users, except forbananas. On the other hand, theytend to buy less of each vegetablethan either the light or mediumusers, except for potatoes.

Comparing demographic infor-mation across user groups in 2004gives us further insights in terms ofhow organic expenditures are relatedto these characteristics. As shown inTable 1, heavy and medium usershave the largest proportions of thosewho have at least a bachelor’s degree,while a larger portion of nonusersand light users have either a high

school diploma or some college.Interestingly, households whoseheads have less than a high schooleducation account for 1.9% of heavyusers, the highest among all usergroups. With respect to age, heavyusers seem to comprise the largestproportion of the youngest house-holds (household head age < 30years), while the light users’ grouphas the largest proportion of house-hold head age between 30 and 49years old. Medium and heavy usersalso have the largest proportion ofolder households relative to nonusersand light users, with the age ofhousehold head 50 years and older.Most heavy users are found in theSouthern and Western United States,and the fewest are found in the cen-tral region. Medium users have thelargest proportion of Whites relativeto other user groups, while a rela-tively large proportion of Hispanicconsumers belong to the light users’group. In comparison, heavy usersare proportionally few amongWhites, with the reverse being truefor African, Asian, and other Ameri-cans.

SummaryWe used the Nielsen Homescan datafrom 2001 and 2004 to analyze con-sumer purchase patterns of freshorganic produce. Our analysis showsthat Asian and African Americanstend to purchase organic over con-ventional produce more than Whitesand Hispanics. Households residingin the western region spent more onorganic produce on a per capita basisthan those residing in other regions.Contrary to popular opinion, we donot find any consistent positive asso-ciation between household incomeand expenditures on organic pro-duce. Although certified organicacreage has increased rapidly in

0

5

10

15

20

Apples

Banan

as

Grapes

Strawberr

ies

Oranges

Tomato

es

Potatoes

Lettu

ce

Onions

Carrots

Fresh Produce

Org

anic

Exp

endi

ture

, % Light usersMedium usersHeavy users

Figure 4. Market shares of selected organic produce by user group, 2004.


boosting the production of organicfoods, our analysis suggests thatdemand appears to be growing fasterthan the supply so that organic pricepremiums for most selected freshproduce remained relatively high in2004, varying from 9% for orangesto 78% for potatoes. Among all freshproduce studied, organic potatoesappear to command the highest per-centage of price premiums in both2001 and 2004.

We classified all households intofour groups: nonusers, light users,medium users, and heavy users,according to their per capita expendi-tures on organic fresh produce. Theproportion of consumers buyingorganic produce increased between2001 and 2004, suggesting anincreasing organic penetration. Interms of demographic characteris-tics, medium and heavy users are rep-resented proportionately more byolder households with the age ofhousehold head 50 years and older.Heavy users also comprise the largestproportion of the youngest house-holds (household head age < 30years), while light users have the larg-est proportion of household head agebetween 30 and 49 years old. Inaddition, we find that light usersexpend a relatively larger share oftheir organic expenditures onbananas and carrots than both themedium and heavy users. Organicvegetables appear to be the preferredorganic produce of light users, whilethe heavy users seem to preferorganic fruits, especially apples andgrapes. For all organic users, organictomatoes are clearly the preferredchoice over other vegetables.

AcknowledgementsThe views expressed in this study arethose of the authors, and do not nec-

essarily reflect those of the UnitedStates Department of Agriculture.

For More InformationDimitri, C., & Greene, C. (2002).

Recent growth patterns in U.S. organic foods market.

Agricultural Information Bulletin No. 777, U.S. Department of

Agriculture, Economic Research Service.

Food Marketing Institute. (2004). U.S. organic food sales continue to grow. Food Institute Report, Arlington, VA, May 10.

Fromartz, S. (2006). Organic, Inc.: Natural foods and how they grew. New York: Harcourt Trade Publishers.

Table 1. Selected Household Characteristics by User Group, 2004.

Category Nonusers Light UsersMedium

Users Heavy Users Total

Educational Level (%)

Less than high school diploma 1.86 0.86 1.66 1.90

1.73

High school diploma 19.60 15.80 15.02 11.53 17.63

Some college 31.49 31.60 30.73 26.36 30.88

College degree 32.05 35.56 33.61 32.83 32.75

Post-college degree 15.00 16.17 18.98 27.38 17.01

Age of Household Head (%)

< 30 years 1.44 0.86 1.21 2.15 1.41

30–39 years 12.46 13.59 9.33 11.91 11.93

40–49 years 24.06 29.01 20.38 19.90 23.46

50–64 years 38.72 35.19 39.17 39.04 38.49

65 years and older 23.32 21.36 29.90 27.00 24.70

Region (%)

Northeast 20.41 22.59 25.24 23.45 21.80

Central 18.76 16.79 14.25 10.65 16.98

South 41.36 37.41 34.89 31.43 38.85

West 19.46 23.21 25.62 34.47 22.37

Race (%)

White 74.08 71.48 76.61 67.93 73.72

African 13.44 12.84 11.69 15.59 13.26

Hispanic 8.17 10.49 6.84 8.37 8.06

Asian 2.72 3.46 2.81 4.06 2.93

Other 1.60 1.72 2.04 4.06 1.92

Number of households 5,266 810 1,565 789 8,430

Note: User groups are classified based on sample distribution of average per capita spending on organic produce per purchase record. The first quartile is defined as light users with per capita spending greater than $0 but less than $0.75, the second and third quartile are defined as medium users (between $0.75 and $3.65), and the fourth quartile is the heavy users (> $3.65).


Hartman Group. (2002). Hartman organic research review: A compilation of national organic research conducted by the Hartman Group. Bellevue, WA.

Hartman Group. (2003). Organic consumer evolution 2003. Bellevue, WA.

Hartman Group. (2006). Organic2006: Consumer attitudes & behavior five years later & into the future. Bellevue, WA.

Nutrition Business Journal (NBJ). (2004). The NBJ/SPINS organic foods report 2004. Cleveland, OH: Penton Media, Inc.

Oberholtzer, L., Dimitri, C., & Greene, C. (2005). Price premiums hold on as U.S. organic produce market expands. VGS-308-01. Washington, DC: U.S. Department of Agriculture, Economic Research Service.

The Organic Trade Association (OTA). (May 2004). The Organic Trade Association’s 2004 manufacturer survey. Greenfield, MA.

The Organic Trade Association (OTA). (2006). The OTA 2006 manufacturer survey overview. Available online: http://www.ota.com/organic/mt.html.

Walnut Acres. (2002). Many Americans make eating organic food a top choice to protect health. Boulder, CO, April 9.

John Stevens-Garmon ([email protected]) is Graduate Research Assis-tant, and Chung L. Huang([email protected]) is Profes-sor, Department of Agricultural andApplied Economics, University ofGeorgia, Athens, GA. Biing-HwanLin ([email protected]) is SeniorEconomist, Economic Research Ser-vice of the USDA, Washington, DC.This research was supported byUSDA-ERS Cooperative AgreementNo. 43-3AEM-5-80043.





2nd Quarter 2007 • 22(2)


Water Quality Credit Trading and Agriculture: Recognizing the Challenges and Policy Issues Aheadby Charles Abdalla, Tatiana Borisova, Doug Parker, and Kristen Saacke Blunk


Economists have long championed market-basedapproaches over regulatory “command and control”approaches for addressing environmental problems.Recently, federal and state policymakers and some stake-holder groups have promoted market-based approaches fordealing with agricultural water pollution. At the federallevel, the U.S. Environmental Protection Agency (USEPA) in 2003 issued a trading policy that allowed indus-trial and municipal point sources (PS) to meet their dis-charge requirements through purchase of “credits” fromfarmers and ranchers who implemented conservation mea-sures that improved water quality. In October 2006, USEPA and the U.S. Department of Agriculture (USDA)reached an agreement to establish and promote waterquality credit trading. In January 2007, USDA SecretaryJohanns stated that in the upcoming farm bill, the admin-istration will view market-based solutions as an importanttool in federal environmental protection efforts aimed atagriculture (USDA, Natural Resources Conservation Ser-vice, 2007). At the state level, Idaho, Michigan, Ohio,Oregon, Pennsylvania, and Virginia have enacted laws orcreated regulatory programs to encourage water qualitytrading (see the summary of state efforts in ETN, 2007).

Given these activities, what can realistically beexpected from market-based programs, and specificallywater quality credit trading, in addressing the difficultissue of water pollution from agriculture? We concludethat policymakers are expecting far too much from tradingas a tool to address agricultural nonpoint source (NPS)water pollution. Currently, water quality credit trading inagriculture is in its infancy and significant implementation

challenges exist. Trading program design and implementa-tion must address complex physical, social, economic,legal and public policy issues. This will require moreexchange among economists, policymakers, farmers,municipalities, and other stakeholders than is currentlyoccurring. Only then will trading as a potential tool befully understood and appropriately implemented.

In a previous issue of Choices, King (2005) examinedissues encountered by water quality trading programs. Weobserve that changes in conditions affecting the supplyand demand sides of potential water quality credit tradingmarkets suggest the need to re-evaluate challenges thatconfront trading programs.

Non-Point Source Ag Pollution in the United States According to the 2000 National Water Quality Inventory,agricultural NPS pollution is the leading source of impair-ment to rivers and lakes, and a major contributor to degra-dation of estuaries (Figure 1). Pollutants from agriculturalcroplands and livestock operations include excess fertilizer,herbicides and insecticides, sediment, and bacteria.

Controlling agricultural runoff is a longstanding anddifficult problem. Agricultural NPS pollution loading isspread over large areas, and monitoring and measuring it istechnically difficult and expensive. Agricultural runoff ishighly variable due to the effects of weather variability,site-specific characteristics of the natural environment(e.g., soil type and land slope), and non-observable farmmanagement practices (such as timing and precision of fer-tilizer application). While the cumulative effect of agricul-tural runoff can be observed through ambient water qual-


ity monitoring, it is generallyimpossible to trace the pollutionback to specific farms. Existing com-puter models provide imperfect esti-mates of agricultural pollution loads.As a result, actual pollution amountsfrom a specific field or property arenot fully known to regulators orfarmers. Moreover, due to the vari-ability in pollution loading, produc-ers only partially control the runofffrom their fields (Horan & Shortle,2001; Braden & Segerson, 1993).

Accordingly, policymakers havelong avoided environmental regula-tory requirements for the agriculturalsector. For example, the federal CleanWater Act excludes all agriculturalsources (except for concentrated ani-mal feeding operations - CAFOs)from federal regulation. Also, impos-ing environmental regulation willlikely reduce agricultural producers’profits and may make U.S. agricul-ture less competitive than othernations with rules that are less strin-gent (Abdalla & Lawton, 2006).Instead of imposing environmentalregulation, policymakers have offeredincentive payment programs to

encourage farmers to voluntarilyadopt environmental protection mea-sures in the form of best managementpractices (BMPs) (NRCS, 2006).

This approach has failed to solvethe water quality problems caused byagricultural runoff. Limited federaland state budgets constrain expan-sion of incentive payment programsfor agricultural BMP implementa-tion, and the existing programs havenot always been cost-efficient (Bab-cock et al., 1995). At the same time,the policy of further reductions ofpoint source (PS) pollution loads isno longer feasible. Increases in urbanpopulation bring about increases inpollution loading from municipal PS(wastewater treatment plants), andthe necessary upgrades of industrialand municipal PS are costly.

Water quality credit trading pol-icy seems to offer an easy solution tothese problems. Economists havelong argued that allowing PS to pur-chase pollution reduction creditsfrom NPS will provide a low-costalternative to PS upgrades (Baumol& Oates, 1988; Pearce & Turner,1990; Faeth, 2000). Trading pro-

grams provide PS with flexibility inhow to achieve their pollution load-ing limits, which creates incentives todiscover cheaper and more efficientabatement methods. Credit salescould provide farmers with neededfinancial resources for BMP imple-mentation. Trading is also attractiveto policymakers and some stakehold-ers because it may provide privatefunds to supplement (or replace) fed-eral and state incentive programs(King & Kuch, 2003).

However, agricultural pollutionrunoff does not meet the economictextbook definition of a tradablecommodity. As a result, designing awater quality credit trading programposes a set of challenges. These arediscussed in the next section.

Realizing the Potential: What Does Economic Theory Suggest as Critical Elements of a Water Quality Trading Program?A water quality credit trading pro-gram is established to meet specificpollution reduction goals. Table 1summarizes elements that are neces-sary for inclusion or consideration inthe implementation of a program.

Critical elements of a water qualitytrading program. Even with thesecomponents in place, certainchallenges must be addressed for awater quality credit trading programto operate. Many of the challengesrelate to PS-NPS trades, where theregulated community (NPDES1

permit holders) meets theunregulated community (agricultureand other NPS). In a 2005 issue ofChoices, King examined the poten-tial supply and demand for waterquality credits. King states that, on

1. National Pollutant Discharge and Elimination System.

Figure 1. Leading sources of impairment of surveyed rivers and streams inthe United States. Source: USEPA, National Water Quality Inventory: 2000 Report, No. 841R02001, August 2002.

020

000

4000

060

000

8000

0

1000

00

1200

00

1400

00

Hydrologic modification

Habitat modification

Urban runoff

Forestry

Municipal point source

Resource extraction

Unknown

Agriculture

Impaired stream miles


the demand side (PS), few discharg-ers are interested in buying waterquality credits if the discharge restric-tions are weak or un-enforced. Morerecently, many states have set limitsfor PS that are either already bindingor are expected to become binding aspopulations grow. In some states,sources will be required to com-pletely offset all new pollution loads.Thus, we expect that the demand forcredits will change. In relation toNPS, King associates the lack of sup-ply with agricultural producers’desires to avoid environmental regu-lation. King argues that by participat-ing in trading programs, producersmake the implicit admission thatNPS pollution can be measured andcontrolled. As a result, some farmersare concerned that trading could leadto increased regulation. However, in

the past few years, a shift in perspec-tive has been occurring. The federaland many state governments havepassed regulations that require agri-cultural producers to implementpractices to better manage runofffrom their farms. Thus, producers arebeginning to view trading as a way tohold off the implementation offuture regulations. Despite thesechanges in conditions affecting boththe supply and demand of potentialwater quality credit trading markets,other significant challenges still con-front trading programs. Some of thekey challenges are discussed below.

Challenges to Water Quality Credit Trading Setting pollution caps. In order toensure that a water quality credit

trading program achieves publicwater quality goals, a maximumloading or “cap” for each pollutantmust be set for a watershed andenforced by the regulatory agency.While public water quality goals areoften linked to services a water bodyprovides (e.g. fish habitat), tradingrequires that a cap be defined forspecific pollutants. This presents achallenge for accurately estimatingthe amount of pollution reductionnecessary to achieve the public goals.In addition, many trading programsleave unregulated agricultural NPSout of the pollution cap, eliminatingthe link between public water qualitygoals and the program results (King& Kuch, 2003). Moreover, consistentenforcement of the cap is a necessarycondition for trading.Establishing allowable pollution lim-its (baselines). An unrestrictive capon PS can diminish or eliminate thedemand for credits. Conversely,setting a high baseline can reduce theNPS will or ability to produce anadequate supply of credits. Besidesaffecting the functionality of thecredit market, assigning baselinesraises the fairness issue since theparties with restrictive limits need toincur costs to achieve these limits.Baseline limits also raise questionsabout responsibility for pollutionclean-up and about property rights oflandowners. For example, manyagricultural BMPs are funded withpublic cost-share money. A debateexists about whether BMPs installedwith public funds are the property offarmers, and if so, whether thesecredits should be eligible for trades(Horan et al., 2004).

Theoretically, the agriculturalbaseline load should be linked topublic water quality goals. This guar-antees that the reductions beyond thebaseline (“credits”) reflect additionalenvironmental benefits produced by

Table 1. Critical elements of a water quality trading program.

Public water quality goals Set by federal, state, or local authorities based on public input and can be defined in terms of ecosystem function, fish population or public safety, or as surface water quality standards.

Pollution cap for a watershed

Limit on the total pollutant load from all sources to a water body. The justification for and size of the cap is based on the public’s water quality goals. Usually the cap is set for an annual load of specific pollutants.

Regulated baseline for point sources or nonpoint sources

Numeric level of pollutant load allowed at a particular point in time. If all polluters meet their regulated baseline, the pollutant cap for the watershed will be obtained.

Unregulated baseline for agricultural nonpoint sources

Minimum level of pollution abatement that an unregulated agricultural operation must achieve before it can participate in a trading program. Sometimes called the threshold.

Credits Units of goods (pollution reduction) to be traded in the water quality credit market. Credits are generated for every unit of pollution reduction beyond the baseline level.

Sellers (credit suppliers) Dischargers that reduce pollution below the baseline and generate credits for sale in the market. Credits can also be sold by intermediaries, if allowed by program rules.

Buyers (demanders for credits)

Dischargers with regulated baselines for whom pollution reduction is expensive. For these sources, it is less costly to buy pollution credits from other parties and use these credits to help achieve their baseline loads. Credits can also be purchased by intermediaries and third parties, if allowed by program rules.

Trading ratio Number of load reduction credits from one source that can be used to compensate excessive loads from another source.

Regulator Entity that determines the water quality goals, establishes caps for pollutants in a watershed, approves and administers the trading program, and monitors and enforces the rules.


the source and supplied to the waterquality credit market. In practice, thebaseline is often set in relation to thecurrent level of pollution, withoutregard to public water quality goals.In addition to jeopardizing publicwater quality goals, such baselinesmay create perverse incentives. Forexample, baselines may penalizethose who have already implementedBMPs and reward those who havenot by paying them for BMP imple-mentation through credit sales (King& Kuch, 2003).Complexities in establishing creditsand associated risks with agriculturalcredits. For NPS, pollution reductionfrom a BMP is difficult to accuratelypredict and monitor. Theeffectiveness of a BMP depends on itsage, implementation factors, howwell it has been maintained, and onsite-specific conditions. Scientificmodels are often used to estimateload reductions from BMPs.However, imperfections persist inmodels and estimated reductionsfrom a BMP likely differ from actualloadings. This complicates theprocess of credit verification andcreates uncertainty about themagnitude of water qualityimprovements from a trade (Ribaudoet al., 1999). Also, requirements toimprove credit verification processesand increase accuracy in pollutionreduction estimation cansignificantly increase costs associatedwith credit trading. Consequently,the number of willing credit sellersand buyers may be reduced (King &Kuch, 2003).

In addition to these measurementand verification complexities, theuncertain nature of agriculturalpollution reduction also implies thatcredit sellers (farmers) do not havecomplete control over the “goods”they sell (Shortle, 2007), while creditbuyers “face the risk of having the

quantity bought falling belowclaimed level” (McCarl, 2006). In themajority of trading programs,variability in NPS pollutionreduction is averaged and annualaverages are used. The risksassociated with agricultural creditsare addressed in existing programs byrequiring PS to purchase several NPSpollution reduction units tocompensate for one unit of their ownpollution increase (i.e., uncertaintytrading ratio). However, such tradingratios implicitly increase the price thePS needs to pay for NPS pollutionreductions. While the majority oftrading programs employ ratios ofgreater than one, it has also beenargued that trading ratios can beeither less than or greater than one,depending on the variability of theagricultural discharges (Horan, 2001;Horan et al., 1999; Horan & Shortle,2001).Transaction costs. Transaction costsare costs that must be incurred tocarry out a trade. Examples includethe degree of difficulty in finding abuyer or seller, verifying credits, andnegotiating and enforcing a trade.Trading will not occur if thetransaction costs exceed the benefitsof a potential trade (Stavins, 1995;Malik, 1992; Krutilla, 1999). Waterquality credit trading programs thatinvolve agricultural NPS arecharacterized by higher transactioncosts than programs involving PSonly. The transaction costs of findinga trading partner are higher becauseNPS are widely distributed across awatershed, and each source cangenerate only small numbers ofcredits in comparison with the largerdemand of PS credit buyers(Woodward, 2006).

In addition to the costs of findinga trading partner, the measurement,verification, and enforcement of agri-cultural NPS pollution reduction can

be costly because of the nature ofNPS pollution runoff (Woodward,2006). Also, for all water qualitycredit trading programs, negotiatinga trade can be difficult because of thenovelty of the markets (Woodward,2003). Unlike other environmentalmarkets (e.g., wetland banking orSO2 emissions trading programs),rules for water quality credit tradingare not yet clearly defined and varyacross programs. Many programs arecomplex, which increases the transac-tion costs for reaching agreementsbetween potential credit trading part-ners. Examples of unclear and com-plicated rules include the credit certi-fication process, credit resale, creditlife span, monitoring and mainte-nance, liability, sale approval prices,and pricing. For example, in theirsurvey of farmers and agency staff toassess perceptions of policies for NPScontrol, McCann and Easter (1999)found transaction costs for waterquality credit trading to be problem-atic. The survey revealed that thetrading program’s administrativecosts were perceived to be the fifthmost expensive among the policiesconsidered. Woodward (2003) sug-gested that transaction costs associ-ated with NPS trades may decrease asprogram participants become morefamiliar with the rules and othertrading partners, and become moreconfident in credit estimation andapproval procedures.Enforcing contracts and liabilityissues. For the benefits of trading tobe realized, there must be amechanism to ensure that agreementsarrived at are met. For example, in aPS-NPS trade, the potential buyers(PS) are liable for achieving pollutionreductions as mandated by theirNPDES permit limits. In contrast,the only document binding thepotential sellers (NPS) is the privatecontract with the buyer. Most


existing trading programs hold thebuyer responsible for monitoring theseller and enforcing the tradeagreement. However, because thecredit buyer and seller are more likelyto focus on the credit price (asopposed to credit quality, i.e.,delivering actual pollution loadreductions), the regulator may bearmore responsibility for verifyingcredits, and enforcing agreements(King & Kuch, 2003). By holdingonly the credit buyer liable forachieving pollution reductions, theregulator reduces the buyer’s (PS)willingness to engage in anagreement. Suggested approaches toalleviate liability issues include theuse of a mediator that can monitorand enforce the trading contract andplace purchased credits in an“insurance pool” to guarantee thatthe NPDES limits are met even ifone of the sellers fails to deliver thecredits. The latter approach is usedby the Pennsylvania and Ohioprograms.Leakage. Implementation of a trad-ing program in one watershed orregion can potentially lead to coun-tervailing actions in areas outside thewatershed. Stephenson et al. (2005)define leakage as an occurrence inwhich “a trade results in a netincrease in loads.” For example, it isproposed that some of the agricul-tural credits certified by Pennsylva-nia’s trading program be generated bytransporting manure/poultry litter tonutrient-deficient regions outside ofthe Chesapeake Bay watershed inPennsylvania. If information con-cerning nutrient availability in soilsin the receiving watershed is not wellknown, manure/litter importationcan lead to increases in water pollu-tion. Stephenson also provided anexample about a farmer who installs ariparian buffer as a BMP, and gener-ates and sells credits. However, to

compensate for reductions in pro-ductive land due to buffer implemen-tation, the farmer expands the pro-ductive acreage in a different place,increasing the nutrient and sedimentloads in that vicinity.Scale of the trading program. Manyexisting water quality credit tradingprograms have been developed forrelatively small watersheds. How-ever, in a larger watershed, moreopportunities exist to find a tradingpartner with significantly differentpollution abatement costs. Thus,greater reductions in costs of meetingpublic water quality goals can be real-ized. Also, in a large watershed, alarge number of buyers and sellerscan help ensure that the market par-ticipants do not exploit the marketpower or distort efficient trading(Woodward, 2003; Hahn, 1989:King & Kuch, 2003). Currently, theregulatory driver for developing trad-ing programs for larger regions isoften lacking. Water quality stan-dards or Total Maximum Daily

Loads (TMDLs), which are consid-ered a driving force for trading pro-grams, are usually set for small water-sheds. The Chesapeake Bay Region isan exception.

Sizing up the Evidence A variety of trading and other mar-ket-like programs have been createdover the past 20 years. Failure toaddress the challenges identifiedabove is a reason why many of theseprograms have been short-lived orhave not resulted in much tradingactivity (Breetz et al., 2004). Thereare also examples of successful waterquality credit trading programs. TheLong Island Sound (CT) and Tar-Pamlico Basin (NC) trading pro-grams both experienced relativelylong lives and resulted in docu-mented pollution load and costreductions. The programs resulted inreallocations of pollution caps amongPS and did not address the challengesposed by NPS runoff.

The genesis of water quality credit trading in the Chesapeake Bay Region. Years of nutrient and sediment related pollution have caused significant impairments in the Chesapeake Bay. In 2000, the renewed Chesapeake Bay Agreement established ambitious targets for member states to signifi-cantly reduce their portion of nutrient pollution by 2010. These regional goals have had a ripple effect on each state in the watershed.

By 2003, Maryland instituted tributary strategies that placed caps on nutrients entering the state’s waters. These caps were the impetus for Pennsylvania to seek new means for reducing its nutrient loads that flow across state lines. In 2006, Pennsylvania adopted its nutrient trading policy that enables PS-NPS trades. Pennsylvania is also attempting to ratchet down municipal sewage treatment plants’ nutrient limits.

Virginia recently established a PS-PS trading program with the intent to eventually include NPS in trades.

Maryland and West Virginia are also in varying stages of developing state trading programs.

It is much too soon to judge how successful water quality credit trading will be in meeting the collective reductions necessary for improv-ing and restoring the Chesapeake Bay.


Alternatively, trading programs inthe Miami Watershed (Ohio), SouthNation River Basin (Ontario, Can-ada), Cherry Creek (Colorado), BeetSugar Cooperative and Rahr MaltingPollutant Offsets (Minnesota), andRed Cedar River (Wisconsin) allinvolve both PS and agricultural NPS(Breetz et al., 2004). Some of thechallenges associated with agricul-tural runoff have been addressed inthese programs by creating an inter-mediary between credit sellers (NPS)and credit buyers (PS). Such interme-diaries (referred to as aggregators,credit banks, or brokers) can reducethe transaction costs of finding trad-ing partners, and credit verificationand monitoring. Intermediaries mayalso potentially bear some liability fordelivering pollution reductions speci-fied in trading agreements. In exist-ing programs, such intermediaries area joint venture of the regulatoryauthorities and public and privateentities. The funds used to purchasecredits from agricultural NPS havebeen drawn from both PS and fed-eral- and state programs. In otherwords, the programs are essentiallyhybrids between market-based trad-ing programs and government-man-aged tax-and-subsidy schemes. Therules for selecting NPS projects togenerate credits and for selectingprices that PS must pay differ amongthe programs, making some of themmore like market-based trading andothers more like government-directed offsets (Stephenson et al.,2005).

Conclusions

Unresolved Public Policy Questions Water quality credit trading has beenperceived by many as an alternativeto command and control regulations.Yet, water quality regulation is a nec-essary driver for trades to occur.

Thus, trading alone cannot solve thechallenges posed by largely uncon-trolled agricultural pollution. Anumber of important public policyquestions have been raised as discus-sions of trading have occurred. Forthe most part these questions remainunanswered or various interestgroups and governmental agencieshave answered them differently.Among the questions are:

Do the political will and resourcesexist? Do federal and state decision-makers have the political will andresources to enforce regulatory capson PS and NPS? This is a critical stepbecause the value of a water qualitycredit is dependent on the enforce-ment of this cap. Without anenforced cap, there is nothing ofvalue for potential market partici-pants to trade (King & Kuch, 2003).Will government define the right topollute and the right to clean water?This question focuses on a keyunderlying issue in trading programdesign. In terms of trading, “you can’tsell what you don’t own”. The answerto this question determines who paysand who benefits from trading. Asnoted, an example of the unresolvednature of this question is the debateover farmers’ ownership rights topublicly funded BMPs. Opinionsvary over assignment of propertyrights to private parties from publiclyfunded projects, raising questionsabout the water quality benefits fromsuch “double-dipping.” Allocatingand enforcing property rights is afundamental role of government. Aregovernments willing to reconcileproperty rights questions of thisnature?Will trading programs be accepted asequitable? Is it fair when one cate-gory of polluters – PS – have regula-tory effluent limits placed on themwhile NPS are required to meet only

program-specific baseline require-ments? King and Kuch (2003) sug-gest that PS dischargers believe thatthere is an inequitable allocation ofpollution rights to NPS dischargers.Will market approaches for environ-mental goods or services be accepted?Some stakeholder groups opposetrading because they believe that it isinappropriate to put a price on natu-ral and environmental resources andtrade them in a market (Goodin,1994; Hahn, 1989; Hahn & Hester,1989; King & Kuch, 2003; Wood-ward, 2003). Some are suspect ofmarket-based approaches and per-ceive trading as excluding environ-mental interests. Also, the terminol-ogy associated with water qualitycredit trading is not well understoodby environmental, farming, anddevelopment communities. This mayprevent them from effective partici-pation in trading program develop-ment.

Where to from Here?Recently, federal and state policy-makers as well as some stakeholdergroups have promoted market-basedapproaches to address the agriculturalwater pollution. We described someof these policy activities and raise thequestion: what can be expected frommarket-based programs, and specifi-cally trading, in addressing the long-standing and difficult issue of waterpollution from agriculture?

After assessing these challenges,we conclude that federal and statepolicymakers are expecting too muchfrom trading as a tool to address NPSwater pollution from agriculture.Since King’s 2005, Choices article,institutions have made progress creat-ing a more supportive trading envi-ronment. Nevertheless, the physicaland regulatory context for agricul-tural NPS pollution does not match


the conditions that economic theorysuggests are needed for widespreadtrading to occur. The majority ofagricultural sources do not face anenforceable cap. Thus, it is difficultto ascertain whether PS-NPS tradeswill create new water qualityimprovements. King and Kuch(2003) state that PS-NPS trades can-not achieve water quality standards inwatersheds where the NPS discharg-ers are responsible for the bulk ofnutrient discharges or where verylarge reductions in nutrient loadingare necessary. The lack of docu-mented success in water qualitycredit trading adds credence to theidea that there is a mismatch of the-ory and practice.

We believe that water qualitytrading in agriculture should con-tinue to be explored and be field-ver-ified for its use as a tool to reduce thecosts associated with pollution reduc-tions. However, in the interim, poli-cymakers must reduce their expecta-tions and reliance on market-basedsolutions until there is more evidencethat validates that these programs canhelp meet pollution reduction goals.Policymakers must recognize thatwater quality credit trading in agri-culture is still in its infancy and thatthe challenges identified are not yetwell understood.

Ongoing trading efforts shouldbe regarded as experiments. Increasedattention must be paid to designingfuture experiments to better under-stand the physical, social, economic,legal, and policy considerations. Thiswill require greater exchange amongeconomists, physical scientists, poli-cymakers, farmers, municipalities,community members, and otherstakeholders. As a more thoroughknowledge of trading as a tool toaddress agricultural water qualityproblems is gained, its potential for

use to help meet public water qualitygoals increases.

For More InformationAbdalla, C.W., & Lawton, J.L.

(2006). Environmental issues in animal agriculture. Choices 3rd

Quarter, 21, (3). Available online: http://www.choicesmagazine.org.

Babcock, B.A., Lakshminarayan, P.G., & Wu, J. (February 1995). Renewing CRP: Results from a study of alternative targeting criteria. CARD Briefing Paper 95-BP 6, Center for Agricultural and Rural Development, Iowa State University.

Baumol, W.J., & Oates, W.E. (1988). The Theory of Environmental Practice. Cambridge: Cambridge University Press, 2nd ed.

Braden, J.B., & Segerson, K. (1993). Information problems in the design of NPS Pollution Policy. In C.S. Russell and J.F. Shogren (eds.), Theory Modeling and Experience in the Management of NPS Pollution. Boston: Kluwer Academic Publishers.

Breetz, H.L., Fisher-Vanden, K., Garzon, L., Jacobs, H., Kroetz, K., & Terry, R. (2004). Water quality trading and offset initiatives in the U.S.: A comprehensive survey. A report for the U.S. Environmental Protection Agency. Hanover, NH: Dartmouth College Rockefeller Center. Available online: http://www.dartmouth.edu/~kfv/waterqualitytradingdatabase.pdf (Accessed February 26, 2007).

Environmental Trading Network (ETN). (2007). Water Quality Trading – State Programs and Rules. Available online: http://www.envtn.org/wqt/

stateprograms_page.html#usprograms (Accessed February 26, 2007).

Faeth, P. (2000). Fertile ground: Nutrient trading’s potential to cost effectively improve water quality. Report by World Resources Institute, Washington, DC.

Goodin, R.E. (1994). Selling environmental indulgences. Kyklos, 47, 573-96.

Hahn R.W. (1989). Economic prescriptions for environmental problems: How the patient followed the doctor’s orders. Journal of Economic Perspectives, 3(2), 95-114.

Hahn R.W., & Hester, G.L. (1989). Marketable permits: Lessons for theory and practice. Ecology Law Quarterly, 16, 361–406.

Horan, R.D. (2001). Differences in social and public risk perceptions and conflicting impacts on point/nonpoint trading ratios. American Journal of Agricultural Economics, 83(4), 934–41. doi:10.1111/0002-9092.00220.

Horan, R.D., Shortle, J.S., & Abler, D.G. (April 2004). The coordination and design of point-nonpoint trading programs and agri-environmental policies. Agricultural and Resource Economics Review, 33(1), 61-78.

Horan, R., Shortle, J., Abler, D., Ribaudo, M. (1999). Second-best markets for point/nonpoint pollution trading. Paper presented at the European Association of Environmental and Resource Economists annual meeting, Oslo, Norway.

Horan R.D., & Shortle, J.S. (2001). Environmental instruments for agriculture. In Shortle, J.S. and D.G. Abler (eds.), 224 pp. Environmental Policies for


Agricultural Pollution Control, New York: CABI Publishing.

King, D.M. (2005). Crunch time for water quality trading. Choices, 20(1), 71-75.

King, D.M., & Kuch, P.J. (2003). Will nutrient credit trading ever work? An assessment of supply and demand problems and institutional obstacles. Environmental Law Reporter News and Analysis, 33(5), 10352-19368.

Krutilla, K. (1999). Environmental policy and transactions costs. In J.C.J.M. van den Bergh (ed.), pp. 249-64, Handbook of Environmental and Resource Economics. Glos, UK: Edward Edgar Publishing Limited.

Malik, A.S. (1992). Enforcement costs and the choice of policy instruments for controlling pollution. Economic Inquiry, 30, 714-21.

McCann, L.M., & Easter, K.W. (1999). Differences between farmer and agency attitudes regarding policies to reduce phosphorus pollution in the Minnesota River Basin. Review of Agricultural Economics, 12, 189-207.

McCarl, B. (March 2006). Measurement and quantity uncertainty. Presentation at a Workshop on Environmental Credits Generated Through Land-Use Changes: Challenges and Approaches, March 8-9, 2006. Baltimore, Maryland. Available online: http://www.envtn.org/LBcreditsworkshop/Uncertainty_Intro.pdf (Accessed February 27, 2007).

Natural Resources Conservation Service (NRCS) (2006). NRCS Conservation Programs. Available online: http://

www.nrcs.usda.gov/PROGRAMS/ (Accessed April 26, 2007).

Pearce, D.W., & Turner, R.K. (1990). Economics of Natural Resources and the Environment. Baltimore, MD: Johns Hopkins University Press.

Ribaudo, M.O., Horan, R.D., & Smith, M.E. (1999). Economics of water quality protection from nonpoint sources. Washington, DC: USDA ERS Ag. Econ. Rep No. 782.

Shortle, J.S. (January 2007). Research opportunities and challenges. Presentation at Water Quality Credit Trading Symposium, USDA CSREES National Water Conference, January 29, 2007, Savannah, Georgia. Available online: http://www.mawaterquality.org/themes/trading_symposium_savannah.htm (Accessed February 27, 2007).

Stavins, R.N. (1995). Transaction costs and tradeable permits. Journal of Environmental Economics and Management, 29, 133-48.

Stavins, R.N., & Whitehead, B.W. (November 1996). The next generation of market-based environmental policies. Discussion Paper 97-10, RFF, 48 pp. Available online: http://rff.org/rff/Documents/RFF-DP-97-10.pdf.

Stephenson, K., Shabman, L., & Boyd, J. (2005). Taxonomy of trading programs: Concepts and applications to TMDLs, In Tamim Younos (ed.), pp. 253-285, TMDLs: Approaches and challenges. Tulsa, OK: Penn Press.

U.S. Department of Agriculture (USDA) Natural Resources Conservation Service. (January 2007). Agricultural Secretary Mike Johanns’ remarks to the American Water Resources

Association Meeting, January 23, 2007, Arlington, VA. Available online: http://www.usda.gov/wps/portal/!ut/p/_s.7_0_A/7_0_1RD?printable=true&contentidonly=true&contentid=2007/01/0011.xml.

U.S. Environmental Protection Agency (USEPA). (August 2002). National Water Quality Inventory: 2000 Report, No. 841R02001.

Woodward, R.T. (March 2006). The challenge of transaction costs. Presentation at a Workshop on Environmental Credits Generated Through Land-Use Changes: Challenges and Approaches, March 8-9, 2006. Baltimore, Maryland. Available online: http://www.envtn.org/LBcreditsworkshop/Transaction_Costs_Intro.pdf (Accessed February 27, 2007).

Woodward, R.T. (2003). Lessons about effluent trading from a single trade. Review of Agricultural Economics, 25(1), 235–45.

Charles Abdalla, ([email protected]) is Associate Professor of Agricul-tural and Environmental Economics,Department of Agricultural Econom-ics and Rural Sociology, PennsylvaniaState University, University Park, PA.Tatiana Borisova ([email protected]) is Research Assis-tant Professor, Division of ResourceManagement, West Virginia Univer-sity, Morgantown, WV. Doug Parker([email protected]) is AssociateProfessor, Department of Agriculturaland Resource Economics, Universityof Maryland, College Park, MD.Kristen Saacke Blunk ([email protected]) is Extension Associate, Agri-cultural Environmental Policy,Department of Agricultural Econom-ics and Rural Sociology, PennsylvaniaState University, University Park, PA.




2nd Quarter 2007 • 22(2)


Farm Growth, Consolidation, and Diversification: Washington Dairy IndustryBy Tristan D. Skolrud, Erik O'Donoghue, C. Richard Shumway, and Almuhanad Melhim


The shrinking number of farms in the United States iswell-documented. Between 1974 and 2002, the totalnumber of farms in the United States declined by 21%.While this represented a large drop in the overall numberof farms, the number of farms with milk cows declinedmuch more dramatically, falling by 79% during thisperiod (USDA/NASS, 2002). With four times as manymilk cows per farm in 2002 than in 1974, it is obviousthat the dairy industry has become much more concen-trated. Further, the entire decline in number of farms withmilk cows occurred in size categories with fewer than 500cows. The number of farms with 500-999 milk cows grewby 36% and the number with 1,000 or more milk cowsmore than doubled. Changes in the State of Washingtongenerally followed those of the Nation.

The growth in the number of the largest-sized farmscreates the most intrigue for economists and policymakersalike. As one of the last bastions of nearly perfectly com-petitive production, does this growth in farm size hint at amajor change in the historically competitive nature of agri-cultural commodity supplies? For example, in 2002, morethan 30% of milk sales came from just 1.5% of dairyfarms. This situation warrants careful attention sinceadverse environmental effects often accompany increasesin farm sizes, particularly for confined animal operations.

While we know that significant changes are occurringin farm size, no one has yet identified which farms aregrowing or shrinking in size. Nor has anyone documentedthe extent of commodity diversification on farms of differ-ent size. Which farms grow? Do farms in the larger sizecategories actually grow the most rapidly? Or do medium-sized farms combine with other farms of comparable sizeto create new large organizations? Do farms in different

size categories increase or decrease their levels of diversifi-cation over time?

To answer these questions, we examined longitudinaldata from the Census of Agriculture in 1992, 1997, and2002 for dairy farms in Washington. This is an importantindustry in both the state and Nation. In the UnitedStates, dairy products rank second among all agriculturalcommodities in value of production (USDA/NASS,2006a). Washington ranks 10th in the nation in milk pro-duction and first in milk production per cow, while thevalue of milk production in the state also ranks it secondin importance among all agricultural commodities(USDA/NASS, 2006b). The state’s dairy industry is highlyconcentrated, but geographically divided. More than halfthe milk cows are located in two counties; Whatcom onthe west of the Cascades and Yakima on the east. Thedemographics are changing with rapid movement of cowsto the east side of the Cascades. Cow numbers in YakimaCounty grew by more than 30% between 1997 and 2002,while those in Whatcom County declined.

Sample Selection and Information CollectedFor our analysis, we included all farms for which theowner checked farming as his/her main occupation andfor which at least 50% of all agricultural income (notincluding government payments) came from the sale ofmilk and dairy products. As a result, 781 farms areincluded in our sample, representing 65% of all Washing-ton dairies in the 1992 census.1 We ranked the farms fromlowest to highest in terms of agricultural sales and thendivided them into 10 equally sized cohorts. In otherwords, each cohort had the same number of farms in 1992with the smallest 10% of dairy farms in the state in


Cohort 1 and the largest 10% inCohort 10. The approximate rangeof sales for each cohort is reported inTable 1. Where possible, we trackedindividual farms in each cohort overthe next two censuses. We also cre-ated new cohorts for entrants in1997 and 2002, for a total of 12cohorts.

We recorded each farm’s tenurestatus, total agricultural sales (exclu-sive of government payments), andmilk and dairy product sales in eachcensus year that it appeared. Basedon this information, we calculatedthe number of farms in production,the number that entered and exited,farm size distributional statistics(mean, median, standard deviation,skewness, kurtosis, and range ofsales), and the percent of cohortfarms in each of four diversificationcategories. The percent of total farmsales (exclusive of government pay-ments) derived from milk and dairyproduct sales determined the diversi-fication categories: (1) 90% or more,(2) 75 - 89.9%, (3) 50 - 74.9%, and(4) less than 50%.

Farm GrowthMean growth rates of 1992 dairyfarms that remained in productionvaried considerably both amongcohorts and between censuses. Afteradjusting for inflation between thecensuses, the dairy farms grew at anaverage compound rate of 1.6% peryear between the 1992 and 1997 cen-suses and 1.1% per year between the1997 and 2002 censuses, averaging1.4% between 1992 and 2002.

Figure 1 shows the annual growthrates we computed for each cohortfor the 5- and 10-year periods. Theaverage size of the smallest cohort ofdairy farms decreased over the 10-year period, while the average size offarms in the three largest cohortsincreased substantially and steadilyover time. Farms in the intermediatesize ranges generally grew slowly andmore erratically. Overall trends sug-gest that, as farm size increased, sodid the corresponding growth rate.

Distribution of Farms within CohortsFarms were close to being uniformlydistributed within most cohorts in1992. Only in the largest cohort wasthe distribution of farms appreciably

skewed. In this cohort, the majorityof farms lay in the lower part of therange and only a small number ofmuch larger farms resided in theupper end of the range. In successivecensuses, as farms tended to grow insize, the surviving farms in all cohortsbecame positively skewed, similar tothe largest cohort in 1992. This find-ing implies that a small number offarms in every cohort grew muchmore rapidly than others.

These results suggest that averagecohort sales were particularly influ-enced by a small number of farmsthat grew rapidly within each cohort.In fact, in each of the five smallestcohorts, a majority of the survivingfarms were smaller in each successivecensus than in 1992. Therefore, ifused improperly, average farm sizecan result in very misleading conclu-sions.

Farm Size and DiversificationBecause of the criteria used to selectfarms to include in the sample, nodairies in 1992 were in the mostdiversified sales class (with less than50% of agricultural sales from milkand dairy products). As apparentfrom Figure 2, the smallest three

1. The remaining 35% of dairy farms consisted of retired and residential/lifestyle farmers.

Table 1. 1992 agricultural sales.a

Cohort Range

1 < 95,000

2 95,000 – 155,000

3 155,000 – 215,000

4 215,000 – 270,000

5 270,000 – 325,000

6 325,000 – 405,000

7 405,000 – 505,000

8 505,000 – 685,000

9 685,000 – 1,085,000

10 > 1,085,000

a Because of data confidentiality conditions, these ranges are only approximate.

Figure 1. Annual growth rates.

-10.00-8.00-6.00-4.00-2.000.002.004.006.008.00

Perc

ent

1 2 3 4 5 6 7 8 9 10

Cohort

1992-19971997-20021992-2002


cohorts were the most diversified andall larger cohorts were more special-ized.

In successive censuses, everycohort became more diversified. Forexample, the percent of farms thatreceived 90% or more of their agri-cultural sales from milk and dairyproducts declined from 35% in 1992to 27% in 2002 in Cohort 1 andfrom 78% in 1992 to 67% in 2002in Cohort 10. Much more dramaticwas the shift of farms to the mostdiversified sales class. By 2002, nearly75% of farms in Cohort 1 receivedless than half of their agriculturalsales from milk and dairy products,while none did in 1992.

Across cohorts, diversificationfollowed roughly the same pattern in1997 and 2002 as in 1992. Thesmallest cohorts were the most diver-sified and specialization increasedwith farm size (see Figures 2-4). Wetested this graphical evidence byexamining the correlation betweenfarm size and level of diversification.Confirming our results, we foundstatistical evidence that as farm sizeincreased, farms tended towardgreater specialization. This tendencybecame stronger over time.

While the diversification trendsbetween 1997 and 2002 followedthose between 1992 and 1997, somecaution should be exercised wheninterpreting the most recent statistics.Milk and dairy product sales do notinclude cull dairy cow or other cattlesales, and milk prices were lower in2002 than in 1992 or 1997. Conse-quently, it is possible that part of theapparent increase in diversification in2002 was due to a higher than nor-mal culling rate induced by the lowermilk price.

A further caution should be madeabout the diversification levels. Wemeasure farm size by value of agricul-tural sales (exclusive of government

Figure 2. Farm diversification, 1992.



0%

20%

40%

60%

80%

100%

Perc

ent o

f Far

ms

1 2 3 4 5 6 7 8 9 10

Cohort

Less than 50% 50-74.9%

75-89.9% 90-100%

0%

20%

40%

60%

80%

100%

Perc

ent o

f Far

ms

1 2 3 4 5 6 7 8 9 10

Cohort

50-74.9% 75-89.9%90-100%

Percent of Farm Sales from Milk & Agricultural Products

Percent of Farm Sales from Milk &

Agricultural Products

0%

20%

40%

60%

80%

100%

Perc

ent o

f Far

ms

1 2 3 4 5 6 7 8 9 10

CohortLess than 50% 50-75%

75-90% 90-100%

Percent of Farm Sales from Milk &

Agricultural Products


payments), and our sample wasselected to include only those farmsfor which milk and dairy productsales accounted for at least 50% ofagricultural sales. Consequently, themost diversified farms with milkcows did not enter our initial sample.If they had been included, the evi-dence of diversification within thedairy industry would be even greater.

Farm Entry and ExitBetween each pair of censuses, morethan twice as many dairy farms exitedthe industry in Washington as newfarms entered. Smaller dairy farmstended to exit at higher rates than didlarger farms. In Cohorts 1-7, an aver-age of 3.5 farms exited for each farmthat entered between 1992 and 2002.In contrast, an average of just overone farm exited for every farm thatentered in Cohorts 8-10, implying avery low net exit rate. Further, thelargest farms (Cohort 10) had fewerexits than entrants, which resulted inpositive growth in the number oflargest dairy farms.

Farms of all sizes entered the mar-ketplace. However, their distributionand behavior differed widely fromincumbent farms. While their meansize was much larger than the incum-bents, falling between the means ofthe two largest incumbent cohorts,their growth rates tended to be muchslower than the growth rates of thelargest incumbents. They averagedless than 1% growth per year. Theyalso entered the marketplace with ahigher average level of diversificationthan any of the large incumbent farmcohorts in the initial sample and con-tinued to diversify at a much morerapid rate.

What Does All This Information Mean?This analysis of longitudinal agricul-tural census data for the Washingtondairy industry has produced impor-tant insights about the relationshipbetween initial farm size and bothsubsequent growth rates and the ten-dency to diversify. The largest groupof cohorts is growing the fastest, sug-gesting that, despite earlier evidencethat economies of scale were largelyexhausted by 750-cow farms (e.g.,Matulich, 1978), dairy farms in thestate are not yet converging toward asize that minimizes average costwithin the current size range. How-ever, the fact that it was Cohort 8rather than Cohort 10 that grew atthe fastest rate does suggest thateconomies of scale may be diminish-ing for the very largest farms.

Additionally, we found that thelarger the farm, the greater the ten-dency to specialize. In other words,larger dairy farms derived more oftheir revenues from milk and dairyproduct sales, while smaller farmsturned to a more diverse range ofoutputs to generate their agriculturalrevenue streams. The only exceptionsapplied to new entrants. While theiraverage size was very large at entry,they were much more diversifiedthan large incumbent farms and grewmuch more slowly. However, theaverage level of diversification in allcohorts has increased over the 10-year period examined. This finding isparticularly surprising for an agricul-tural commodity that has been one ofthe last bastions of the single-productfarm.

Tristan D. Skolrud ([email protected]) is Undergraduate Honors Stu-dent, School of Economic Sciences,Washington State University, Pull-man, WA. Erik O'Donoghue

(eo’[email protected]) is Agri-cultural Economist, EconomicResearch Service, U.S. Department ofAgriculture, Washington, DC. C.Richard Shumway ([email protected]) is Professor, School of EconomicSciences, Washington State Univer-sity, Pullman, WA. Almuhanad Mel-him ([email protected]) is Gradu-ate Research Assistant, School ofEconomic Sciences, Washington StateUniversity, Pullman, WA.

AcknowledgementsThe views expressed are those of theauthors and do not necessarily corre-spond to the views or policies of ERSor the U.S. Department of Agricul-ture.

For More InformationMatulich, S.C. (1978). Efficiencies

in large-scale dairying: Incentives for Future Structural Change. American Journal of Agricultural Economics, 60(4), 642-67.

United States Department of Agriculture National Agricultural Statistical Service (USDA/NASS). (2002). Census of Agriculture. Available online: http://www.nass.usda.gov/Census_of_Agriculture.

United States Department of Agriculture National Agricultural Statistical Service (USDA/NASS). (2006a). Agricultural Statistics. Available online: http://www.nass.usda.gov/Publications/Ag_Statistics/agr06/.

United States Department of Agriculture National Agricultural Statistical Service (USDA/NASS). (2006b). Washington 2006 Annual Bulletin. Available online: http://www.nass.usda.gov/Statistics_by_State/Washington.




2nd Quarter 2007 • 22(2)


Fruit and Vegetables Go Back to Schoolby John L. Park, Benjamin L. Campbell, Andres Silva, and Rodolfo M. Nayga, Jr.

JEL Classification Code: I38, Q18

Perhaps one of the most alarming trends plaguing ourmodern food system is the seemingly rampant increase inthe prevalence of obesity across the United States. TheDepartment of Health and Human Services reports thatone in three adults is obese, and two out of three are con-sidered overweight or obese. Even more alarming is thetrend among children, where obesity rates have nearly tri-pled since 1980 (NCHS-CDC, 2006). Policy makersacross the country have responded with efforts to drivefoods of minimal nutritional value out of our schools andreplace them with whole grains and fresh fruit and vegeta-bles (Schmid, 2007; Zhang, 2007).

The resulting policies and programs may representopportunities for marketers and producers of fresh fruitand vegetables to reach a growing market segment withinour schools. However, it is not enough to simply providean appealing product to students. Instead, successful mar-keters will appeal to the needs, perceptions, and prefer-ences of those responsible for wholesale purchasing (Park,2001). They need insight into the mentality of the schoolfoodservice director. The effectiveness of these programs toimprove dietary quality and presumably health is currentlybeing debated. Externalities such as the influence of schoolfoodservice buying habits and constraints may impact theeffectiveness of these programs to achieve their statedobjectives.

The Road to ObesityTo understand our present situation, let’s step back andlook at how we got to this point of a national health crisis.We believe that one major influence on our current pre-dicament is the change that has occurred in our lifestyles.Think back fifty years ago—families generally consisted oftwo parents, and subsisted on one income. Family mealswere prepared at home and enjoyed around the dinner

table. The newspaper was a major avenue for the flow ofinformation, and businesses competed with the guy acrossthe street.

Fast forward to the present—the composition of thefamily unit has changed, as well as the economic condi-tions in which it operates. Today, meals of convenience arethe norm, and businesses conduct operations on a globalscale. Information is transmitted as quickly as ideas aredeveloped. The widespread use of cell phones, text mes-saging, and the internet have compounded the amount ofinformation available to an individual at any given pointin time. Consequently, the modern consumer expectsinstant satisfaction and greatly values added services andconveniences. Not surprisingly, the food industry hasshifted toward providing indulgent, value-added foodproducts that are highly convenient (see Capps and Park,2003, for further discussion of food marketing channels).When you put this together with the facts that U.S. con-sumers generally have less discretionary time, more discre-tionary income, and lead sedentary lifestyles, you get a rec-ipe for obesity.

In a continual effort to provide consumers with prod-ucts they want, food marketers are watching these trendsclosely. Some recent new product trends emphasize the useof wholegrain ingredients, while others offer portion con-trol like Nabisco’s “100 Calorie Packs.” Even so, marketerscontinue to struggle to increase per capita consumption offresh fruit and vegetables, despite continued reports on theassociated health benefits (Wang & McKay, 2006). How-ever, the public outcry over the poor state of school food-service offerings may signal an opportunity for increasedsales of fresh fruit and vegetables. In support of this, thegovernment offers programs intended to improve thedietary intakes of school children while simultaneouslysupporting agricultural producers.


Back to SchoolMost (if not all) school districts havea foodservice director that is incharge of purchasing food for the stu-dents within the district. Althoughtheir primary concern is providinglunch, many schools also offer break-fast and snacks. The foodservicedirector will combine funds availablefrom state and local government aswell as federal programs. In general,he/she can purchase products from

whatever source he/she chooses; how-ever, participation in certain govern-ment programs requires purchasingspecific products through specificsources of distribution.

A variety of programs are avail-able to help foodservice directorsprocure food for their schools. Suchprograms include the NationalSchool Lunch Program (NSLP) andthe National School Breakfast Pro-gram (NSBP) among others. TheNSLP and NSBP differ from some

food aid programs in that they areavailable, at a slightly higher cost, tochildren who may not qualify forpoverty-based assistance. The spend-ing of these program funds are typi-cally administered by a state depart-ment of agriculture.

The National School Lunch Pro-gram (NSLP) is the major govern-ment program that foodservice direc-tors use to purchase their lunchfoods. The NSLP provides nutrition-ally balanced low-cost, or sometimesfree lunches to millions of childreneach school day. Since the inceptionin 1946, daily student participationin NSLP has grown from 7.1 millionto 29.6 million in 2005, withapproximately 100,000 schools par-ticipating. With regards to the NSBP,daily participation has grown from1.8 million children in 1975 to 9.3million children in 2005, withapproximately 83,000 schools partic-ipating. Based on the large numberof students using the NSLP andNSBP daily, their influence on nutri-tion, both in consumption and inestablishing life-long behaviors,could be considerable.

There are also other programsthat exist to encourage the consump-tion of specific food products inschool programs. The Fresh Fruitand Vegetable Program, instituted bythe USDA, reimburses schools fortheir purchases of fresh fruit and veg-etables outside of those purchased aspart of the NSLP. Initially availablein 8 states, the program has beenexpanded, but funds are limited. Forexample, in Texas this program wasmade available to only 24 of the7,203 schools that are eligible to par-ticipate in NSLP.

The methods school districts useto implement these programs gobeyond putting nutritional foods onthe menu. Some schools make theseproducts available on demand,

Figure 1. Percent of obese adults.Source: National Center for Health Statistics.

Figure 1. Percent of obese children.Source: National Center for Health Statistics.

0%

10%

20%

30%

1991 1995 1998 1999 2000 2001 2003

0%

5%

10%

15%

20%

1963-70 1971-74 1976-80 1988-94 1999-2000

6 to 11 years old12 to 19 years old


throughout the day. Finally, manystates have initiated Farm-to-Schoolprograms in conjunction with federalprograms. These programs help tokeep federal funds within the stateeconomy by allowing schools to buyproduce from local growers at subsi-dized prices, sometimes only payingthe cost of delivery (TDA, 2006).

Program EffectivenessAs part of the Centers for DiseaseControl and Prevention, theNational Center for Health Statisticscollects data through various meth-ods in an effort to document thehealth status of the U.S. population.The information they gather is alsoan important part of research effortsto evaluate health policies and pro-grams. However, quality of health is acomplex issue. It can be measured inmany different ways and is impactedby many different factors. For thatreason, there is an abundance ofresearch examining the effectivenessof these programs to provide onlyselected groups of nutrients at anyone time.

Currently, we are examining datafrom the National Health and Nutri-tion Examination Survey(NHANES) to see if the NSLP andNSBP actually improve the con-sumption of fresh fruit and vegeta-bles among school age children.Since obesity is rising and a largenumber of students eat at least onemeal (lunch) and perhaps two meals(lunch and breakfast) at school eachday, measuring the effectiveness ofthe NSLP and NSBP is extremelyimportant in order to determine ifthe current guidelines are having aneffect on healthy eating habits, par-ticularly related to the consumptionof fruits and vegetables.

Some preliminary results suggestthat student participation in only the

NSLP has a positive impact on freshfruit and vegetable consumption.However, student participation in theNSBP has a negative impact on freshfruit and vegetable consumption(Campbell et al., 2007). Reasonsbehind these results are being investi-gated, but we need to remember thatthese results are influenced by thechoices available to the students. Forexample, in the course of our researchwe were able to interview many dif-ferent school foodservice directors.On one occasion, we ran across a ref-erence to what foodservice personnelcalled “Hot Cheetos and cheese” thatwas sold to the students a la carte.The product involved taking a singleserving bag of Flamin’ Hot Cheetos(a popular brand of spicy extrudedcorn snack from Frito-Lay), pouringa scoop of melted nacho cheese overthe contents, and putting a fork in it.This cheesy treat was a favoriteamong the students and provided theschool district with sizeable revenue.

Although the product was admit-tedly unhealthy, the income that itgenerated gave the school districtgreater freedom and flexibility inoperations. Any profit from the saleof a la carte items of this nature goesback to the district office, in essence

increasing its budget. The rareopportunity of an actual profit centerin a school foodservice program is atemptation that can completelyundermine nutritional objectives.This illustrates how the factors sur-rounding the implementation offoodservice programs can confoundthe ability of national programs toachieve their stated goals. Further, wefound it interesting that smallernearby school districts also admittedto selling the Flamin’ Hot Cheetosand cheese mixture, but stopped thatpractice due to nutritional concerns.From discussions with these foodser-vice directors, it was evident thattheir action might be due in part to agreater sense of accountability to par-ents and increased parental involve-ment with school administrators.

As a final note, researchers needto be aware of the Cheetos effect.Seemingly, conflicting results sur-rounding federal program initiativesmay not be entirely due to the pro-gram, but also due to the conditionsof its implementation. National sur-veys sometimes have difficulty inaccounting for quality differencesamong the experiences of theirrespondents. In order to be moreeffective, policy makers and food


marketers alike must be aware of thebehavior of channel intermediarieslike school administrators, in addi-tion to the constraints they face.

For More InformationCampbell, B.L., Silva, A., Nayga,

R.M., Jr., & Park, J.L. (2007). Do the national school lunch and school breakfast programs improve children’s fresh fruit and vegetable consumption? A matching technique analysis. Working Paper, Department of Agricultural Economics, Texas A&M University.

Capps, O., Jr., & Park, J.L. (2003). Food retailing and food service. In Koontz, S. (ed.) Veterinary Clinics Food Animal Practice, 19, 445-61.

National Center of Health Statistics-Center for Disease Control (NCHS-CDC). (2006). Preva-

lence of overweight among chil-dren and adolescents: United States, 1999-2002. Revised Feb. 8, 2005. Available online: http://www.cdc.gov/nchs/products/pubs/pubd/hestats/overwght99.htm (Accessed April 19, 2006).

Park, J.L. (2001). Supermarket prod-uct selection uncovered: Manu-facturer promotions and the channel intermediary. Interna-tional Food and Agribusiness Review, 4, 119-31.

Schmid, R.E. (2007). Standards urged for school snacks. Associ-ated Press. Available online: http://news.yahoo.com/s/ap/20070425/ap_on_he_me/diet_school_food_10.

Texas Department of Agriculture (TDA). (2005). Texas Selected for Fresh Fruit and Vegetable Pro-gram. News Release dated November 15, 2005. Austin, TX.

Wang, S., & McKay, B. (2006). More reasons to eat your veggies; Government revamps effort to boost consumption of produce as understanding of benefits grows. Wall Street Journal, (Eastern Edition). New York, NY: July 25, 2006, p. D.1.

Zhang, J. (2007). Proposal would tighten school-food standards. Wall Street Journal, (Eastern Edi-tion). New York, NY: April 26, 2007, p. D.6.

John L. Park ([email protected]) isAssociate Professor, Benjamin L.Campbell ([email protected]) and Andres Silva ([email protected]) are Graduate Assis-tants, and Rodolfo M. Nayga, Jr.([email protected]) is Professor,respectively, Department of Agricul-tural Economics, Texas A&M Uni-versity, College Station, TX.