SPIA TAC CGIAR

Prabhu L. Pingali

With an annotated bibliography

Matthew P. Feldmann

Milestones in ImpactAssessment Research in theCGIAR, 1970-1999

Prabhu L. Pingali*

With an annotated bibliography of impact

assessment studies conducted in the CGIAR, 1970-

1999, prepared by Matthew P. Feldmann*

Standing Panel on Impact Assessment (SPIA),Technical Advisory Committee (TAC) of theConsultative Group on InternationalAgricultural Research (CGIAR)

* Prabhu Pingali is Director of the Economics Program at the International Maize andWheat Improvement Center (CIMMYT), Mexico. He presented an earlier version of thispaper at the Impact Assessment Workshop organized by the Standing Panel on ImpactAssessment (SPIA) of the Technical Advisory Committee (TAC) of the ConsultativeGroup on International Agricultural Research (CGIAR), 3-5 May 2000. When this workwas underway, Matthew P. Feldmann was a CIMMYT Economics Program ResearchAssistant. The views expressed in this paper and bibliography do not necessarilyreflect policies of SPIA, TAC, the CGIAR, or CIMMYT.

The Consultative Group on International Agricultural Research (CGIAR) (www.cgiar.org) isan international association of 58 public and private sector members that supports 16international Future Harvest centers (www.futureharvest.org) that conduct research withfarmers, scientists, and policymakers to alleviate poverty and increase food security whileprotecting natural resources. The Standing Panel on Impact Assessment (SPIA) of theCGIAR Technical Advisory Committee (TAC) provides information on the impact of pastCGIAR outputs with respect to CGIAR goals, complements the ex post assessment activitiesof the Future Harvest centers, and provides feedback to strategic planning and prioritysetting in the CGIAR. CIMMYT (www.cimmyt.org) , a Future Harvest Center, is a non-profit, scientific research and training organization that works with agricultural researchinstitutions worldwide to improve the profitability, productivity, and sustainability ofmaize- and wheat-based cropping systems for poor farmers.

The opinions expressed in this publication are the sole responsibility of the authors. Thedesignations employed in the presentation of materials in this publication do not imply theexpression of any opinion whatsoever on the part of the CGIAR, SPIA, CIMMYT, orcontributory organizations concerning the legal status of any country, territory, city, or area,or of its authorities, or concerning the delimitation of its frontiers or boundaries. Fair use ofthis material is encouraged by the publishers. Proper citation is requested.

Correct citation: Pingali, P.L. 2001. Milestones in Impact Assessment Research in the CGIAR,1970-1999. With an Annotated Bibliography of Impact Assessment Studies Conducted in theCGIAR, 1970-1999, Prepared by Matthew P. Feldmann. Mexico, D.F.: Standing Panel on ImpactAssessment, Technical Advisory Committee of the Consultative Group on InternationalAgricultural Research.

Abstract: This report identifies “milestones” in impact assessment research by theinternational research centers of the Consultative Group on International AgriculturalResearch (CGIAR). Milestones are theoretical or methodological research contributions thatidentify and analyze new areas of impact assessment research and that have been publishedin refereed publications. The report covers the development of impact assessment researchin the CGIAR, charting the progression from assessments of the impact of germplasmadoption and crop management research in the 1970s and 1980s, to formal rate of return andbenefit distribution studies in the 1980s, research on spillover and intersectoral impacts inthe 1980s and 1990s, and gender and environmental impact assessment research as the 1990sprogressed. The role of the CGIAR scientists in this evolution and broadening of scope isdocumented, as well as the CGIAR’s contribution to (and benefits from) impact assessmentresearch in universities and non-CGIAR research centers The report also highlights the roleof impact assessment research in producing a picture of solid accomplishments of CGIARresearch with respect to poverty, food security, and the environment. The report concludeswith comments on future needs for impact assessment research within the CGIAR. Anannotated bibliography of CGIAR impact assessment studies follows the report and servesas a useful guide to this significant area of CGIAR research.

ISBN: 970-648-076-5

AGROVOC descriptors: CGIAR; Agricultural development; Economic development;Research institutions; Research projects; Diffusion of research; Technology transfer;Innovation adoption; Environmental impact assessment; International cooperation

AGRIS category codes: E14 Development Economics and PoliciesA50 Agricultural Research

Dewey decimal classification: 338.16

Printed in Mexico by the International Maize and Wheat Improvement Center (CIMMYT)for the Consultative Group on International Agricultural Research (CGIAR).

ii

Acknowledgments

Hans Gregersen, the Chairman of the Standing Panel on ImpactAssessment (SPIA) of the Technical Advisory Committee (TAC), andGuido Gryseels, the Deputy Executive Secretary of TAC, wereinstrumental in helping me conceptualize and write this paper. Theyprovided substantive comments on drafts of the paper and providedfinancial support for its publication. The first draft of this paper waswritten during a one-week personal retreat at Stanford University inMarch, 2000. My long discussions with Wally Falcon at that time, andhis subsequent comments on various versions of the paper, helpedme enormously. Comments, suggestions, and help from AlexMcCalla, Robert Herdt, Timothy Reeves, Michael Morris, and severalmembers of the CIMMYT Economics Program are gratefullyacknowledged. Comments and criticisms from the participants of theSPIA workshop on impact assessment, held in May 2000 in Rome,helped clarify my thinking and provided useful leads to referencesthat I had previously overlooked. I am also grateful to CIMMYTcolleagues Kelly Cassaday, Head of Information Services; SatwantKaur, CIMMYT Economics Program Editor; and Miguell Mellado,Head of Production, for editing and publishing this review.

iv

Dedication

This paper is dedicated to Vernon Ruttan, who was the firsteconomist to work for a CGIAR Center (the International RiceResearch Institute, 1963-1965). Economists—past and present—inthe CGIAR owe him our gratitude for leading the way.

Foreword

It is quite widely accepted that the CGIARhas had, over its lifetime, a significant,sustainable impact on poor people byhelping to develop the technologies andagricultural management tools that havepermitted increased food security anddramatic lowering of the cost of producingthe major food crops of the world. This, inturn, has benefited both poor producers andconsumers. A recent SPIA study hasdocumented and confirmed this conclusionat the CGIAR-wide level (Evenson andGollin 2001).

What is not so well known is the significantcontributions that CGIAR scientists havemade in advancing the theory and methodsof impact assessment which have been usedto identify and estimate the people-relatedimpacts of agricultural research. For thefirst time, this “other” role of the CGIARhas been systematically documentedthrough the accompanying study of“milestones” in CGIAR impact assessmentresearch. The author, Dr. Prabhu Pingali,Director of the Economics Program atCIMMYT, has been intimately involved inthis work over the past decades and is anappropriate person to document CGIARcontributions. This paper was originallyprepared for a May 2000 workshopsponsored by SPIA, “The Future of ImpactAssessment in the CGIAR: Needs,Constraints, and Options” (TAC/SPIA,forthcoming).

Dr. Pingali carries out the study in anobjective fashion by identifying“milestones” in impact assessment researchconducted within the CGIAR centers.Milestones are defined as researchcontributions that identify and analyze newareas of impact assessment research,whether they be theoretical or

methodological. A key characteristic ofmilestones is that they are quickly followedby other, similar studies within and outsidethe CGIAR that verify the findings of themilestone research. In all cases, themilestone research is published in refereedpublications. Given these definingcharacteristics of “milestone” research, it ispossible to develop an objective picture ofthe importance of CGIAR contributions inthe evolving world of impact assessmentresearch.

The study traces the evolution of thisresearch in the CGIAR, concluding thatthere is a logical evolution from therelatively narrow focus in the 1970s and1980s on assessment of impacts ofgermplasm adoption and crop managementresearch, to formal rate of return and benefitdistribution studies starting in the 1980s.The next major broadening in the 1980s wasto work on spillover and intersectoralimpacts. Finally, in the 1990s the activity hasbroadened further into gender andenvironmental impact assessment research.

The role of the CGIAR scientists in thisevolution and broadening of scope has beenimportant. In addition to contributionsmade by CGIAR scientists while workingwith the CGIAR, it should be noted thatmany of these researchers have moved onto become leaders in impact assessmentresearch in universities and research centerswhere they still build on and benefit fromtheir early impact assessment research inthe CGIAR. Thus the indirect contributionsof the CGIAR are widespread. At the sametime, as Dr. Pingali emphasizes, the CGIARhas benefited greatly from the work ofothers and the interactions that have been

v

possible with scientists in both developingand developed countries.

In tracing the milestones in CGIAR impactassessment research, Dr. Pingali paints abroad yet in-depth picture of how activitieswithin the CGIAR System have contributedto this growing and dynamic field. At thesame time, as he points out, the milestoneresearch and use of its methodologicalresults in more routine impact assessmentsof CGIAR research have produced a pictureof solid accomplishments and impacts ofCGIAR research with respect to the maingoals of the System, which are to have

agricultural research results applied insuch a way that there are sustainableimpacts on poverty, food insecurity, andthe environmental resources on which allagricultural production depends. Heprovides some insights into the roadahead and the gaps that remain to befilled by future generations of CGIARimpact assessment researchers.

SPIA commends Dr. Pingali for this study,one that itself will be a milestone insynthesis of past activity in the CGIAR inthe area of impact assessment research.

Hans GregersenChair, Standing Panel on ImpactAssessmentTechnical Advisory CommitteeCGIAR

vi

1

IntroductionThe Consultative Group on InternationalAgricultural Research (CGIAR) probablyhas had greater impact on agriculturalproduction, productivity, and thelivelihoods of the rural poor in the worldthan any other agricultural researchorganization. This paper does not intend todocument the substantial impacts of theCGIAR but rather to dwell on the stock andstate of impact assessment researchconducted by CGIAR scientists, especiallyeconomists. What is not so widely realizedis that over the past three decadescontributions by CGIAR economists andother scientists to the science of impactassessment have in many cases beengroundbreaking, and they have createdmilestones in the ever-increasing body oftechnical literature on impact assessmenttheory and methods. In numerous instances,the CGIAR was the forerunner of asubstantial body of academic research onparticular themes related to impactassessment.

For the purposes of this paper, a milestoneis defined as a research contribution thatidentified a new area or theme of impactassessment research. It can also includemethodological contributions. In all cases,the seminal contributions (milestones) werequickly followed by other studies, bothfrom within and outside the CGIAR, whichverified the findings or applied themethodology developed in the study. In allcases, the milestone contributions werepublished in refereed journal articles, books,or official, peer-reviewed publications of a

CGIAR Center (not working papers and othernon-refereed documents).

Before proceeding with this review, it isimportant to be explicit about its uniquefeatures and shortcomings. This reviewdocuments and highlights the importantmilestones in impact assessment research thatcan be attributed to current and past CGIAReconomists and social scientists. It coversonly research on the impacts of technologyand management practices; it does not coverthe contributions of CGIAR scientists toimpact assessment research related to policyresearch and policy advice. Owing to thelimitations of my own knowledge, the focushas been mainly on milestones related to cropresearch in the CGIAR. Activities related tosuch themes as livestock, trees, and capacitystrengthening have not received adequateattention. At the same time, it should bepointed out that impact assessment researchrelated to these themes is also much less, andthere are few CGIAR milestones in theseareas that can be discussed. The need forsubstantial gap filling, in areas covered andstudies cited, is recognized and allsuggestions will be appreciated.

Categories of ImpactAssessment ResearchFor the purposes of this paper, CGIAR impactassessment research is divided into sixcategories:

1. Adoption and farm-level production/productivity/profitability impact ofmodern varieties;

Milestones in Impact AssessmentResearch in the CGIAR, 1970-1999Prabhu L. Pingali

2

2. Adoption and impact of technicalchange in crop management andimproved input-use efficiencies;

3. Distribution, equity, and food securityimpacts, including poverty alleviation;

4. Environmental, ecological, and humanhealth impacts;

5. Intersectoral linkages: agriculture as anengine of growth; and

6. Impacts on the research systemthrough spillover benefits, training,and networks.

Most of the early (1970s and 1980s) impactassessment work in the CGIAR wasconcentrated in the first two categories.Interest in—and significant contributionsto—the remaining categories of researchcame about only in the late 1980s and1990s (Figure 1). The broadening agendaof impact research in the CGIAR,especially in the 1990s, reflects to a largeextent the broadening research agenda ofthe CGIAR itself and also the changingpublic perceptions about the impacts ofagricultural modernization andtechnological change. The impactsresearch agenda of the 1990s also reflects adesire among CGIAR economists to takeon more challenging and difficult-to-measure impacts, such as theenvironmental and ecological impacts ofagricultural intensification andmodernization.

Adoption and Farm-levelProduction/Productivity/Profitability Impact ofModern VarietiesThe extent of adoption of modern varieties(MVs) is now well established, at least for themajor cereal crops. CGIAR contributions tothe content of varieties released by national

agricultural research systems (NARSs) are wellknown on a global basis for rice, wheat, andmaize and on at least a regional basis for otherCGIAR mandate crops.

Since the early 1970s, substantial work hasbeen done on the extent of adoption andfarm-level impact of modern, high-yieldingvarieties of rice, wheat, and maize. Similarwork for other crops, such as cassava,sorghum, millets, and potatoes, followed inthe 1980s and beyond. The pioneeringstudies in assessing the global extent of MVadoption and CGIAR impact wereconducted on rice and wheat by Dalrymplein the early to mid-1970s (Dalrymple 1977,1978, 1986). Dalrymple (1978) provided aquantitative verification of the commonperception that the Green Revolution wasunderway for rice and wheat (Figure 2): “By1976/77 approximately 30 million hectaresof wheat and 25 million hectares of rice inthe developing world was planted to highyielding varieties.”

Dalrymple’s work triggered several in-depth adoption studies, particularly by theInternational Rice Research Institute (IRRI)and International Maize and Wheat

Figure 1. The expanding agenda of impactassessment research.

1970 1980 1990 2000

Germplasm adoption

Crop management research

Returns to research

Equity consequences

Spillover effects

Sectoral linkages

Gender

Environment/health

3

Improvement Center (CIMMYT). Herdt andCapule (1983) and Byerlee and Moya (1993)provided detailed assessments of MVadoption for rice and wheat, respectively, inwhich global, regional, and nationaladoption figures were differentiated by typeof production environment (favorable andunfavorable). Byerlee’s work in particularconcentrated on identifying the CIMMYTcontent of wheat varieties released aroundthe world. Other adoption and impactstudies that traced CGIAR content in cropreleases included Walker and Ryan (1991)for sorghum and millet, Evenson and David(1993) for rice, López-Pereira and Morris(1994) for maize, and Walker and Crissman(1996) for potatoes. The recent effort by SPIAto evaluate the impacts of cropimprovement research builds on a longhistory of CGIAR efforts and expertise inthis area (see Evenson and Gollin 2001, whoprovide detailed adoption and impactinformation on all CGIAR crops).

The CGIAR Centers have generated substantialempirical evidence on the biophysical andsocioeconomic factors that influence theprofitability of adopting improved germplasm.IRRI’s “consequences of modern rice

technologies” studies of the mid-1970s (seeIRRI 1978) and CIMMYT’s “studies in theadoption of new agricultural technology”(see Winklemann 1976) were the absoluteforerunners in adoption case studies. Theseearly efforts identified the impact on farm-level production and income of switchingfrom traditional to modern varieties. TheIRRI and CIMMYT studies were done in astandardized case study format acrossseveral countries. At CIMMYT, for example,case studies initially started in Mexico andquickly expanded across Asia, Africa, andLatin America. By the early 1980s, abouttwo dozen detailed case studies fromaround the developing world haddocumented the farm-level consequences ofadopting modern rice, wheat, and maizegermplasm. In addition to documenting thepositive production and income benefits ofMVs, these early adoption studies helpedallay fears that the Green Revolution had abias toward large-scale farmers and anadverse effect on poor farmers and landlesslaborers.

Aside from documenting the consequences ofadopting MVs, CGIAR researchers have madesubstantial contributions to the literature on thefactors constraining the adoption of moderntechnologies. In his study of rice technology,Herdt (1979) documented the biophysicaland socioeconomic factors that explainedthe “yield gap” in farmers’ fields (the yieldgap is the difference between the yield thatis economically achievable and the yieldthat is actually obtained). A follow-uppublication by Herdt and Mandac (1981)identified the technical and economicconstraints to exploiting yields and profitsfrom the adoption of MVs. Herdt’s work onconstraints to adoption and technology useinspired a whole literature on technical andeconomic efficiency. CIMMYT’s adoptioncase studies for Latin America, Asia, andAfrica also provided detailed informationon constraints. Village-level studies (VLS)by the International Crops ResearchInstitute for the Semi-Arid Tropics(ICRISAT) provided similar information for

Figure 2. Green Revolution in Asia: Thefirst decade.

Rice

Wheat

30

25

20

15

10

5

0

Area on which MVs adopted(million ha)

1965/661966/67

1967/681968/69

1969/701970/71

1971/721972/73

1973/741974/75

1975/761976/77

Source: Dalrymple (1978).

4

the semi-arid tropics of India and, later, forWest Africa. The VLS dataset that enteredthe public domain in the mid-1980sspurred a series of doctoral theses (112 atlast count), mainly from universities in theUnited States, whose authors attemptedempirical assessments of a large gamut offarmer behavior in the context of technicalchange.

Other important contributions emergingfrom case studies of adoption andconstraints were detailed assessments ofrisk attitudes, studies of the variability offarm production, and credit studies.Binswanger’s study of risk, done atICRISAT, was a seminal contribution to theunderstanding of farmers’ attitudes to riskand how those attitudes constraintechnology adoption (Binswanger 1980a).Binswanger’s work also made significantcontributions to the theoretical literatureon risk. Anderson and Hazell’s (1989)volume on variability in grain yieldsprovided an important synthesis ofevidence on production variability inagricultural systems that had recentlyswitched to MVs. The volume coveredmost CGIAR crops and all continentswhere the CGIAR Centers worked. Thecontention that MVs may be more risky andtherefore less attractive to farmers does notseem to have held up in practice. Stochasticdominance tests of the distribution ofreturns from improved and traditionalvarieties typically show new varieties to bedominant. The following studies providedcrop-specific results for sorghum andmillets (Walker 1989; Witcombe 1989) andrice (Flinn and Garrity 1989; Coffman andHargrove 1989). More recent studies atCIMMYT (CIMMYT 1991) for wheat and atICRISAT for millet (Adesina 1988; Shapiro1990) have reported reduced coefficients ofvariation for yields over time (Table 1).

Studies estimating the rates of return to CGIARcommodity research investments have consistentlyshown the investment to be extremely profitable.The returns to investments in high-yieldingmodern germplasm have been measured ingreat detail by several CGIAR Centers overthe last few decades. These studies foundhigh returns to the CGIAR strategy ofgermplasm improvement. The very firststudies that calculated the returns to researchinvestment were conducted at IRRI for riceresearch investments in the Philippines(Flores-Moya et al. 1978) and at theInternational Center for Tropical Agriculture(CIAT) in Colombia (Scobie and Posada 1977,1978). More detailed evidence on the highrates of return to public-sector investments inagricultural research was provided by theInternational Service for NationalAgricultural Research (ISNAR) (Echeverría1990) and the International Food PolicyResearch Institute (IFPRI) (Pardey et al. 1992).Table 2 presents a synthesis of early rate ofreturn studies. We ought to recognize,however, that these high rates of return arepartly biased by the fact that in general onlysuccess stories are incorporated in rate ofreturn studies. For a detailed synthesis of thenumerous studies conducted across crops andcountries, see Evenson (2001) and Alston et al.(2000). Alston et al. concluded from a reviewof 289 studies that there was no evidence tosupport the view that the rate of return toagricultural research and development hasdeclined over time.

Table 1. Expected income and risk of traditionaland improved millet technologies in Niger

Author Location Traditional Improved

Adesina (1988) Maradi 156 195(52) (42)

Shapiro (1990) Libure 446 628(52) (56)

Kouka 301 409(49) (39)

Note: Income figures are in US$ (CFA 298/US$ 1). Thenumber in parentheses is the coefficient of variation ofincome (%).

5

Table 2. Returns to CGIAR research: first-generation studies

ResultsCountry (rate of

Study Year (region or institute) Commodity Period return) a

Evenson and Flores 1978 Asia Rice 1950-65 32-39%

Asia (national) Rice 1966-75 73-78%

Asia (international) Rice 1966-75 74-102%

Flores et al. 1978 Philippines Rice 1966-75 46-71%

Scobie and Posada 1978 Colombia Rice 1957-64 79-96%

Martínez and Sain 1983 Panama (IDIAP-Caisan) Maize (on-farm research) 1979-82 188-332%

Nagy 1983 Pakistan Maize 1967-81 19%

Pakistan Wheat 1967-81 58%

Ambrosi and Cruz 1984 Brazil (EMBRAPA-CNPT) Wheat 1974-90 59-74%

Monares 1984 Rwanda Potato seed 1978-85 40%

Salmon 1984 Indonesia Rice 1965-77 133%

Muchnik 1985 Latin America Rice 1968-90 17-44%

Unnevehr 1986 Southeast Asia Rice quality 1983-84 61%

Librero and Perez 1987 Philippines Maize 1956-83 27-48%

Norton et al. 1987 Peru (INIPA) Aggregate 1981-2000 17-38%

Peru (INIPA) Rice 1981-2000 17-44%

Peru (INIPA) Maize 1981-2000 10-31%

Peru (INIPA) Wheat 1981-2000 18-36%

Peru (INIPA) Potatoes 1981-2000 22-42%

Peru (INIPA) Beans 1981-2000 14-24%

Echeverría et al. 1989 Uruguay Rice 1965-85 52%

Evenson 1988 Paraguay Crops 1988 75-90%

Norgaard 1988 Africa Cassava 1977-2003 Benefit-

(biological control) cost ratio

of 149:1

Evenson and da Cruz 1989 South America (PROCISUR) Wheat 1979-88 110%

South America (PROCISUR) Soybeans 1979-88 179%

South America (PROCISUR) Maize 1979-88 191%

Schwartz et al. 1989 Senegal (CRSP) Cowpeas 1981-87 63%

Source: Echeverría (1990).a Depending on the study, these are average or marginal rates of return. More than one value means a range of

returns depending upon different assumptions or different periods of analysis. Results are rounded. Results ofconducting sensitivity tests on various parameters of the models are not presented in this table.

6

countries on the impact of small-farmmechanization (IRRI 1986). More recentstudies at IRRI on the consequences ofmechanization have concentrated onharvest and post-harvest operations,including small mills used predominantlyby female household members. Animportant contribution of themechanization studies was the finding thatfarm machinery is also adoptedsequentially. Machinery is first adopted forpower-intensive operations such as tillageand transport, even in labor-abundantsocieties. Control-intensive operations suchas weeding and harvesting, which requireknowledge and judgment, are mechanizedonly as wages rise (Herdt 1983; Binswanger1984). Consequently, most studies on theimpacts of tractor use (for land preparation)found minimal labor displacement even inlabor-abundant societies (see Pingali et. al.1987 for a detailed review of the farmmechanization literature).

Fertilizer UseGiven the importance of fertilizer forrealizing the yield potential of MVs, theCGIAR has given surprisingly littleattention to assessing the adoption andimpact of fertilizer. David’s (1976) study offactors that determine fertilizer use at thefarm level and David and Barker’s (1978)assessment of fertilizer responsiveness arestill the best pieces of work in this area.David and Otsuka (1994) provide morerecent information on the determinants offarm-level fertilizer use for rice. Despite thehistory of high fertilizer subsidies indeveloping countries, relatively littleresearch has been done on the economicincentives and technological options forincreasing fertilizer-use efficiency. Few expost impact studies have examinedalternatives to chemical fertilizer, includingorganic fertilizers (e.g., Azolla, Sesbania),crop rotations, and improved fallows. Thelimited CGIAR effort to develop andpromote these technologies and their

Adoption and Impact ofTechnical Change in CropManagement and ImprovedInput-use EfficienciesSavings in production costs have come aboutfrom technical change in crop management andincreased input-use efficiencies. Once MVshave been adopted, the next set oftechnologies to make a significantdifference in reducing production costsincludes machinery, land managementpractices (often in association withherbicide use), fertilizer use, integrated pestmanagement, and (most recently) improvedwater management practices. Althoughmany Green Revolution technologies weredeveloped and extended in package form(e.g., new plant varieties plusrecommended fertilizer, pesticide, andherbicide rates, along with water controlmeasures), many components of thesetechnologies were taken up in a piecemeal,often stepwise manner (Byerlee and Hessede Polanco 1986). The sequence of adoptionis determined by factor scarcities and thepotential cost savings achieved. Herdt(1987) provided a detailed assessment of thesequential adoption of crop managementtechnologies for rice in the Philippines.Traxler and Byerlee (1992) provided similarevidence on the sequential adoption of cropmanagement technologies for wheat inSonora, northwestern Mexico.

Machinery Adoption and UseThe early adoption of tractors in the high-potential agricultural areas where the GreenRevolution first occurred, particularly inAsia, prompted a series of in-depth studiesof the impacts of farm mechanization.Binswanger’s study for ICRISAT on the useof tractors in South Asia was a pioneeringeffort in this area (Binswanger 1978). In theearly to mid-1980s, IRRI conducted detailedfarm-level studies across several Asian

7

limited adoption probably explain the smallnumber of adoption and impact studiesconducted on non-chemical fertilizers by thecenters. There are two significant exceptions:the work of McIntire et al. (1992) at theInternational Livestock Center for Africa(ILCA)1 on crop-livestock integration and ofFranzel et al. (1999) at the InternationalCentre for Research in Agroforestry (ICRAF)on improved fallows. McIntire et al.conducted an Africa-wide survey on theextent of crop-livestock integration and itsimpact on soil fertility management. Franzelet al. have been documenting the extent ofadoption of improved fallow systems acrossEast Africa and the consequences forsustaining soil fertility in smallholderproduction systems.

Integrated Pest ManagementCompared to fertilizer use, integrated pestmanagement (IPM) has received far greaterattention from the CGIAR’s commodityresearch centers. See Waibel (2000) for adetailed review and assessment of IPMresearch across the CGIAR Centers. One ofthe earliest studies on insecticide use was astudy on rice (Herdt et al. 1984) conductedas part of a yield constraints study in severalcountries of Asia between 1973 and 1979.Herdt et al. concluded that the expectedreturns to rice production are lower forfarmers who apply insecticides on aprophylactic basis than for farmers whoapply no insecticides at all. This result wasvalidated in on-farm trials by Litsinger(1989) and Waibel (1986).

Three very important studies related to IPMhave been conducted in recent years: a studyby the International Institute of TropicalAgriculture (IITA) on biological control ofcassava mealybug in Africa; an IRRI study ofIPM in Asian rice production; and a studyby the International Potato Center (CIP) on

IPM practices in Andean potatoproduction. IITA’s work on cassavamealybug control (Norgaard 1988) was thevery first study that attempted to estimatethe ex post returns to a biological controlprogram. The benefit-cost ratio turned outto be 149:1, triggering an enormous interestin the cassava mealybug program (thenunder evaluation) and in biological controlprograms in general. Although Waibel(2000) has cautioned that the returns tocrisis research, such as the cassavamealybug program, are generally higherthan returns to non-crisis research, such asgermplasm improvement, it would be safeto conclude that investments made inmealybug control were extremelyworthwhile. IRRI’s research on the impactof IPM documented the declining levels ofinsecticide used on rice and validatedearlier findings through farm survey datathat natural control (a zero-pesticidestrategy) was the most profitable option forfarmers, especially when health costs weretaken into account (Rola and Pingali 1993;Pingali et al. 1994; Pingali and Roger 1995).The CIP study followed a design similar tothat of the IRRI study, but it found apositive productivity benefit to insecticideuse and hence a more crucial productionversus health trade-off for insecticideapplications in Andean potatoes (Crissmanet al. 1998).

Although host-plant resistance has alwaysbeen the cornerstone of any IPM strategy,little work has been done to assess theimpact of host-plant resistance oncontrolling pest losses and reducing yieldvariability. The exception is the recent workon returns to breeding for rust resistance inwheat at CIMMYT (Smale et al. 1998).Smale et al. not only determined theeconomic returns to breeding for durablerust resistance but provided a methodologyfor evaluating the returns to investments in

1 Now the International Livestock Research Institute (ILRI).

8

such “maintenance breeding” activities.2

Maintenance breeding is crucial to the long-term relevance and viability of any breedingprogram, yet it remains quite invisible indonor portfolios and CGIAR prioritysetting.

Distribution, Equity, andFood Security ImpactsWidespread adoption of modern seed-fertilizertechnology led to a significant shift in the foodsupply function, contributing to a fall in realfood prices. Some of the very early work onthe impact of the Green Revolution wasdirected towards understanding its effectson market prices and food security. Theprimary effect of agricultural research onthe non-farm poor, as well as on the ruralpoor who are net purchasers of food, isthrough lower food prices:

The effect of agricultural research onimproving the purchasing power of thepoor—both by raising their incomes andby lowering the prices of staple foodproducts—is probably the major sourceof nutritional gains associated withagricultural research. Only the poor gohungry. Because a relatively highproportion of any income gains made bythe poor is spent on food, the incomeeffects of research-induced supply shiftscan have major nutritional implications,particularly if those shifts result fromtechnologies aimed at the poorestproducers.

Alston et al. (1995:85)

Early efforts to document the impact oftechnological change and the consequentincrease in food supplies on food prices andincome distribution were made by Hayamiand Herdt (1977) at IRRI, Pinstrup-

Andersen (1976) and Scobie and Posada(1978) at CIAT, and Binswanger (1980b) atICRISAT. Pinstrup-Andersen arguedstrongly that the primary nutritionalimpact for the poor came through theincreased food supplies generated throughtechnological change.

The profitability of modern farming systemshas been maintained despite falling food prices(in real terms), owing to a steady decline in thecost per ton of production. The point thatproducers have continued to benefit fromtechnological change despite falling outputprices has not been emphasizedadequately in the literature on researchimpacts, although empirical evidence doesshow quite clearly that the cost per ton ofproduction has fallen significantly. Most ofthe empirical data on changes in unitproduction costs is on rice. The very firststudy comparing unit production costs ofmodern and traditional varieties was doneby IRRI (1972) for several Asian locationsand found significantly lower costs per tonof production of modern compared totraditional varieties. More recent studieshave documented trends in the cost perton of rice production over time (seePingali et al. 1997:43 for several Asianlocations and Hossain 1998:331 forBangladesh). Consistently across severalAsian countries, unit production costshave tended to decline over time, and overthe same period, production costs havegenerally tended to be lower than outputprices. Empirical evidence for the long-term decline in unit costs of productionalso exists for wheat; see Sidhu andByerlee (1992) for evidence from theIndian Punjab and numerous publicationsfor evidence from the Yaqui Valley ofnorthwestern Mexico, the starting point forthe Green Revolution in wheat.

2 In crop improvement research, “maintenance breeding” can be defined as any breeding activity thatseeks to preserve yield potential and yield stability by maintaining a crop’s genetic resistance toevolving biotic stresses such as diseases and pests.

9

The impact and benefits of technological changehave varied by ecological domain, socioeconomicfactor, and gender. Many studies haveaddressed the differential impact oftechnological change in favorable andunfavorable production environments.David and Otsuka (1994) conducted a studyon the differential impact of technologicalchange across rice environments in Asia.They found that although the favorable,high-potential environments gained themost in terms of productivity growth, theless favorable environments benefited aswell through technology spillovers andthrough labor migration to more productiveenvironments. According to David andOtsuka, wage equalization across favorableand unfavorable environments was one ofthe primary means of redistributing thegains of technological change. Renkow(1993) found similar results for wheatgrown in high- and low-potentialenvironments in Pakistan. Byerlee andMoya (1993), in their global assessment ofthe adoption of wheat MVs, found that overtime the adoption of MVs in unfavorableenvironments caught up to levels ofadoption in more favorable environments,particularly when germplasm developedfor high-potential environments was furtheradapted to the more marginalenvironments.

Income distribution effects across thevarious socioeconomic groups within arural community have received someattention in the CGIAR impact literature. Ina detailed study of North Arcot District ofTamil Nadu, India, Hazell and Ramaswamy(1991) estimated the distribution of benefitsof technological change across landlesslaborers, tenant farmers, and small andlarge landowners. David and Otsuka’s(1994) study paid particular attention toeffects on landless labor and tenant farmers.Their results validated the findings ofIRRI’s (1978) consequences study, whichfound that the benefits were shared acrossthe various groups. The early criticism that

the Green Revolution had benefited onlylarge-scale farmers was negated by thefindings of all of these studies.

In recent years, the differential impactliterature has focused on identifying thedistribution of benefits between men andwomen farmers and male- and female-headed households. Paris from IRRI, andQuisumbing et al. from IFPRI, haveconducted several studies on the subject.The general finding across crops andcontinents is that women farmers andfemale-headed households have gainedproportionally less than their malecounterparts. Paris (1998) has contendedthat it is not gender alone that determineswhether an individual benefits fromtechnological change, however, but ratherthe initial social and economic status of theindividual. Women from land-owninghouseholds who have some control over theland benefited substantially fromtechnological change relative to womenfrom poor, landless households.Quisumbing et al. (1995) concluded from aten-country study that among the very poorthe economic welfare of male- and female-headed households differed very little.Differences emerged only where cultural orinstitutional factors prevented equalparticipation in the labor force, as inBangladesh. For excellent recent reviews ofthe literature, see Doss (1999) on Africanmaize farming systems and Paris (1998) onrice in Asia.

Environmental, Ecological,and Human Health ImpactsThe environmental, ecological, and humanhealth impacts of modern technology havereceived limited attention from the CGIARCenters. There are a few significantexceptions to this generalization, includingthe research on genetic diversity conductedby CIMMYT, on pesticides and health byIRRI and CIP, and on soil erosion in the

10

hillsides of Central America by IFPRI, CIAT,and CIMMYT. A good deal ofenvironmental impact assessment is beingdone at the CGIAR Centers now, but resultshave not been published and therefore arenot included in this review. The works citedabove were all groundbreaking, in terms ofthe problems they tackled as well as theirmethodological contributions.

From the very early days of the GreenRevolution, the CGIAR has been accused ofreducing crop genetic diversity insubsistence farming systems through thewidespread promotion of MVs. Smale wasone of the first economists in the CGIAR toinvestigate how intensification and use ofMVs affected genetic diversity; she assessedthe current stock of diversity in farmers’fields as well as farmers’ incentives forconserving diversity (see Smale 1998 forcase studies of wheat, maize, and rice). Asubstantial joint research effort byCIMMYT, IFPRI, and the International PlantGenetic Resources Institute (IPGRI)continues to document changes in geneticdiversity in intensive wheat productionsystems in China, India, and Pakistan.

The impact of intensive and injudiciouspesticide use was another issue that doggedthe commodity centers of the CGIAR forseveral decades. While the CGIARresponded by promoting IPM and saferpesticide use, studies that directlyaddressed the impacts of pesticides werenot conducted until the late 1980s. IRRI’sstudy of the impact of pesticides in riceproduction on human health was apioneering effort that created renewedinterest in IPM research (Rola and Pingali1993; Antle and Pingali 1994; Pingali et al.1994; Pingali and Roger 1995). The CIPstudy on pesticides in potato productionwas done soon afterward (Crissman et al.1998). Both the IRRI and CIP studies,conducted in areas of intensive pesticideuse, found that adverse health effects couldbe attributed directly to pesticide use. Thesimilarity in health findings among rice

farmers in the Philippines and potatofarmers in Ecuador was striking, but asnoted previously, the IRRI study found thatreturns to pesticide use were negative whenhealth effects were explicitly accounted for,whereas the CIP study found continuedpositive returns even when health costswere considered.

Inter-sectoral Linkages:Agriculture as an Engine ofGrowthStudies by CGIAR economists have providedempirical support to the proposition that growthin the agricultural sector has economy-wideeffects. One of the earliest studies showingthe linkages between the agricultural andnon-agricultural sectors was done at thevillage level by Hayami while at IRRI(Hayami et al. 1978). Hayami provides thebest micro-level illustration of the impacts ofrapid growth in rice production on land andlabor markets and the non-agriculturalsector. More recent assessments on theimpacts of productivity growth on land andlabor markets have been done by Pinstrup-Andersen and Hazell (1985) and by Davidand Otsuka (1994). Pinstrup-Andersen andHazell argued that landless laborers did notadequately share in the benefits of the GreenRevolution because of depressed wage ratesattributable to migrants from other regions.David and Otsuka, on the other hand, foundthat migrants shared in the benefits of theGreen Revolution through increasedemployment opportunities and wageincome. The latter study also documentedthat rising productivity caused land pricesto rise in the high-potential environmentswhere the Green Revolution took off.

An IFPRI study by Hazell and Ramaswamy(1991) showed the development ofbackward and forward linkages fromincreased agricultural productivity growthin India. The term “backward linkages”

11

refers to the demand for inputs used in anew production activity, whereas “forwardlinkages” refers to new processingindustries stimulated by the availability ofraw materials provided by the newproduction activity. Delgado et al. (1998),also at IFPRI, found similar evidence ofgrowth being stimulated in the non-agricultural sector by growth in agriculturalproductivity. The work by researchers atIFPRI has provided strong empiricalsupport for the proposition that agriculturedoes indeed act as an engine of overalleconomic growth (Hazell and Haggblade1993; Fan et al. 1998)

Impacts on the ResearchSystem through Spillovers,Training, and NetworksResearch on the impacts of the CGIAR onnational public-sector research, technologydevelopment, and human resource capacityneeds to be enhanced substantially.

Research SpilloversCGIAR economists have made animportant contribution to increasedresearch efficiency through their work ontechnological spillovers. It often takes along time for knowledge to be developedthrough research and then adopted.Typically, ten years pass from the initiationof a research project to the dissemination ofresearch results. By borrowing researchresults (e.g., plant lines or varieties) fromother countries, a country can shorten itsresearch time and contribute to increasedreturns to research investments (Alston etal. 1995).

Several attempts have been made toidentify and incorporate such spillovers inresearch priority setting at the level of

individual countries. Brennan (1986), whileat CIMMYT, measured the benefits toAustralian wheat breeding programs ofaccess to CIMMYT breeding materials.Pardey et al. (1996) measured the benefitsto US wheat and rice production fromgermplasm developed at CIMMYT andIRRI (Table 3). Pardey and Wood (1994)explicitly accounted for cross-bordertechnology spillovers in agriculturalresearch priority setting in Latin America.Morris et al. (1994) conducted a similarassessment for wheat research spilloversfrom India to Nepal. Maredia and Byerlee(1999) quantified spillover benefits forwheat across agroecological boundaries—in other words, they measured thetransferability of wheat varietiesdeveloped for one productionenvironment (e.g., an irrigatedenvironment) to another (e.g., a rainfedenvironment).

Training and NetworksThe CGIAR has made immensecontributions to strengthening researchcapacity in the national research systemsof developing countries, primarily bybuilding human capacity (through trainingprograms) and by improving the exchangeof information and technology (throughnetworks). Surprisingly little has beendone to measure the impact of CGIARinvestments in training and networking. Arecent CIMMYT study of the RegionalMaize Program in Central America (amaize research network) found high

Table 3. US benefits and costs (in millions of1993 US$) from CGIAR wheat and rice research

Benefits and costs Wheat Rice

Present value of benefits 13, 653 1,042Present value of costs 71 63Benefit-to-cost ratio 190:1 17:1

Source: Pardey et al. (1996).

12

returns to participation in the network,especially for small countries that couldnot afford a critical mass of crop researchand development specialists (Gómez1999). Substantial work is needed on theeconomic and social returns to networkparticipation, since this mode of linkingresearchers in national programs and theCGIAR is expected to continue into theforeseeable future. Measuring the impactsof networks is very difficult, however,because of problems in clearly identifyinginputs and outputs and attributing themto the participants in the network.Training is another area in desperate needof a critical impact assessment. MostCGIAR Centers have counted numbers ofcourse participants and assembledanecdotal information on training, but todate no detailed impact assessment hasbeen done. Perhaps the recent SPIA effortto evaluate the impact of CGIAR trainingwill fill this large gap.

Concluding RemarksOver the past three decades, CGIAReconomists have been actively involved inassessing the adoption and impact of MVsand other technologies developed by theCGIAR Centers. Impact assessment at theCGIAR has been recognized for itssubstantive and methodologicalcontributions by the economics professionas well as by the donor community thatinvests in the CGIAR Centers. The scopeof impacts work done at the centers hasexpanded from a narrow effort to measurethe adoption of MVs to researchquantifying a wide array of impacts onproduction, productivity, equity, humanhealth, and the environment. Numeroushigh-quality publications in internationaljournals, as well numerous awardsreceived by CGIAR researchers, attest tothe high quality of the research conductedat the centers. Even so, several concerns

remain to be addressed to enhance andsustain the work that has been done, andalso to increase the impact of impactassessment research.

While the scope of impact assessmentresearch in the CGIAR has expanded overtime, there continues to be a need (andpressure) to further broaden the agenda ofimpacts research. Evidence that moderntechnology has contributed to povertyalleviation is particularly sought after bythe donor community. Similarly, studies onthe impacts of agricultural modernizationon the sustainable management ofagricultural ecosystems in general, andbiodiversity in particular, are also indemand. CGIAR economists have alsobegun measuring the impacts of policyresearch and advocacy, training and humancapacity strengthening, and networks fortechnology generation and exchange. Theimpact assessment community facesenormous methodological challenges as itgears up to measuring impacts in theseareas, but they are also the areas wherefuture milestones can be achieved.

Looking at the record of CGIAR impactassessment research over the past threedecades, it is clear that not all centers havebeen equally engaged in the work. A fewcenters have had an outstanding record,and many are trying to catch up. Whatfactors distinguish the few successfulcenters from the rest?

It is likely that the commodity mandate ofsome centers has made a big difference:centers that achieved technologicalbreakthroughs with enormous impacts alsoconducted outstanding impact assessmentresearch. Equally important was the valuethat center leadership placed on buildingand sustaining economics researchcapacity. It is perhaps not coincidental thatcenters with excellent impact assessmentcredentials also supported an independenteconomics group with a critical mass of

13

well-trained staff. The experience of the lastthree decades has shown that there are noshortcuts to sustainable, high-qualityperformance in impact assessment research.

To maintain professional credibility andensure high-quality research, it is importantthat the methods used in impact assessmentare the best available and that the work isdone in as transparent a manner as possible.The most effective means for centers tosustain the reputations of their impactassessment groups over the long term hasbeen to constantly subject research results topeer review. The long-term publicity valueof high-quality impact assessment researchcannot be underestimated and ought not besacrificed for short-term gains.

Although CGIAR economists have a goodrecord in impact assessment research, theirability to communicate their results to awider audience has been rather poor.Publication in high-quality journals is anessential means of maintaining professionalcredibility, but the findings of impactresearch may have their true impact on anaudience that does not necessarily obtain itsinformation from professional economicsjournals, such as taxpayers, developmentassistance agencies, and philanthropicorganizations. Multiple and targeted

outputs derived from a set of results wouldgo a long way in helping to achieve thegoals of certain audiences while enhancingthe mission of the research institution.

It is also very important to recognize thatlong-term credibility in impact assessmentresearch can be maintained only when theinputs and contributions of all partners areappropriately acknowledged. Many CGIARtechnologies are joint products that requiresignificant input from national programpartners and increasingly from the privatesector. It is sometimes unclear whether theCGIAR can claim exclusive or even partialcredit for certain impacts. The nature of thecontribution ought to be specified clearly, beit discovery, development, or facilitation.

Finally, there is a fine line between impactassessment research and public awarenessactivities. As noted above, impact resultscertainly ought to reach a wider audience,but it must be remembered that the primaryfunction of impacts research is to providefeedback to researchers and to theinstitution about what is going right andwhat is going wrong with research anddevelopment activities. The opportunity forintrospection and mid-course correction isoften lost when the balance is tipped tooheavily—or prematurely—towards thepublicity value of the results.

14

References

Adesina, A.A. 1988. Farmer Behaviorand New AgriculturalTechnologies in the RainfedAgriculture of Southern Niger: AStochastic ProgrammingAnalysis. Ph.D thesis, PurdueUniversity, West Lafayette,Indiana.

Alston, J.M., M.C. Marra, P.G. Pardey,and T.J. Wyatt. 2000. Researchreturns redux: A meta-analysis ofthe returns to agricultural R&D.Australian Journal of Agriculturaland Resource Economics 44(2): 185-215.

Alston, J.M., G.W. Norton, and P.G.Pardey. 1995. Science underScarcity: Principles and Practices forAgricultural Research Evaluationand Priority Setting. Ithaca:Cornell University.

Anderson, J.R., and P.B.R. Hazell.1989. Variability in Grain Yields:Implications for AgriculturalResearch and Policy in DevelopingCountries. Baltimore: JohnsHopkins.

Antle, J.M., and P.L. Pingali. 1994.Pesticides, productivity, andfarmer health: A Philippine casestudy. American Journal ofAgricultural Economics 76(3): 418–430.

Binswanger, H.P. 1978. The Economicsof Tractors in South Asia.Patancheru: International CropsResearch Institute for the Semi-Arid Tropics (ICRISAT) andAgricultural DevelopmentCouncil.

Binswanger, H. P. 1980a. Attitudestoward risk: Experimentalmeasurement in rural India.American Journal of AgriculturalEconomics 62 (3): 395-407.

Binswanger, H.P. 1980b. Incomedistribution effects of technicalchange: Some analytical issues.South East Asian Economic Review1(3): 179-218.

Binswanger, H.P. 1984. AgriculturalMechanization: A ComparativeHistorical Perspective. StaffWorking Paper No. 673.Washington, D.C.: World Bank.

Brennan, J.P. 1986. Impact of WheatVarieties from CIMMYT onAustralian Wheat Production.Agricultural Economics BulletinNo. 5. Sydney: Department ofAgriculture New South Wales.

David, C.C. 1976. Fertilizer demand inthe Asian rice economy. FoodResearch Institute Studies XV(1).

David, C.C., and R. Barker. 1978.Modern rice varieties andfertilizer consumption. InInternational Rice ResearchInstitute (ed.), EconomicConsequences of the New RiceTechnology. Los Baños:International Rice ResearchInstitute (IRRI).

David, C., and K. Otsuka (eds.) 1994.Modern Rice Technology and IncomeDistribution in Asia. Boulder:Lynne Rienner.

Delgado, L.C., J. Hopkins, and V.A.Kelly. 1998. Agricultural GrowthLinkages in Sub-Saharan Africa.IFPRI Research Report No. 107.Washington, D.C.: InternationalFood Policy Research Institute(IFPRI).

Doss, C.R. 1999. Twenty-Five Years ofResearch on Women Farmers inAfrica: Lessons and Implications forAgricultural Research Institutions;with an Annotated Bibliography.CIMMYT Economics ProgramPaper No. 99-02. Mexico, D.F.:International Maize and WheatImprovement Center (CIMMYT).

Echeverría, R.G. 1990. Assessing theimpact of agricultural research. InR.G. Echeverría (ed.), Methods forDiagnosing Research SystemConstraints and Assessing the Impactof Agricultural Research. Vol. 2. TheHague: International Service forNational Agricultural Research(ISNAR).

Echeverría, R.G., G. Ferreira, and M.Dabezies. 1989. Returns toInvestment in the Generation andTransfer of Rice Technology inUruguay. ISNAR Working PaperNo. 30. The Hague: InternationalService for National AgriculturalResearch (ISNAR).

Evenson, R.E. 2001. Economic impactsof agricultural research andextension. In Vol. 1 of B. Gardnerand G. Rausser (eds.), Handbook ofAgricultural Economics.Amsterdam: Elsevier Science.

Evenson, R.E., and D. Gollin (Eds.).2001. Crop Genetic Improvement andAgricultural Development.Wallingford: CABI (forthcoming).

Evenson, R., and C. David. 1993.Adjustment and Technology: TheCase of Rice. Development CentreStudies. Paris: Organisation forEconomic Cooperation andDevelopment (OECD).

Byerlee, D., and E. Hesse de Polanco.1986. Farmers’ stepwise adoptionof technological packages:Evidence from the MexicanAltiplano. American Journal ofAgricultural Economics 68(3): 519-527.

Byerlee, D., and M. Morris. 1993.Research for marginalenvironments: Are weunderinvested? Food Policy 18(5):381-393.

Byerlee, D., and P. Moya. 1993. Impactsof International Wheat BreedingResearch in the Developing World,1966-1990. Mexico, D.F.:International Maize and WheatImprovement Center (CIMMYT).

CIMMYT (International Maize andWheat Improvement Center). 1991.World Wheat Facts and Trends 1990-91: Wheat and Barley Production inthe Rainfed Marginal Environments ofthe Developing World. Mexico, D.F.:International Maize and WheatImprovement Center (CIMMYT).

Coffman, W.R., and T.R. Hargrove.1989. Modern rice varieties as apossible factor in productionvariability. In J.R. Anderson andP.B.R. Hazell (eds.), Variability inGrain Yields: Implications forAgricultural Research and Policy inDeveloping Countries. Baltimore:Johns Hopkins.

Crissman, C.C., J.M. Antle, and S.M.Capalbo (eds.). 1998. Economic,Environmental, and Health Tradeoffsin Agriculture: Pesticides and theSustainability of Andean PotatoProduction. Boston: KluwerAcademic Publishers.

Dalrymple, D.G. 1977. Evaluating theimpact of international research onwheat and rice production in thedeveloping nations. In T.M. Ardat,D.G. Dalrymple, and V.W. Ruttan(eds.), Resource Allocation andProductivity in National andInternational Agricultural Research.Minneapolis: University ofMinnesota.

Dalrymple, D.G. 1978. Development andSpread of High Yielding Varieties ofWheat and Rice in the less DevelopedNations. USDA ForeignAgricultural Economics Report No.95. Washington, D.C.: UnitedStates Department of Agriculture(USDA).

Dalrymple, D.G. 1986. Development andSpread of High Yielding WheatVarieties in Developing Countries.Washington, D.C.: United StatesAgency for InternationalDevelopment (USAID).

15

Fan, S., P. Hazell, and S. Thorat. 1998.Government Spending, Growth, andPoverty: An Analysis ofInterlinkages in Rural India. EPTDDiscussion Paper No. 33.Washington, D.C.: InternationalFood Policy Research Institute(IFPRI).

Flinn, J.C., and D.P. Garrity. 1989.Yield stability and modern ricetechnology. In J.R. Anderson andP.B.R. Hazell (eds.), Variability inGrain Yields: Implications forAgricultural Research and Policy inDeveloping Countries. Baltimore:Johns Hopkins University.

Flores-Moya, P., R.E. Evenson, and Y.Hayami. 1978. Social returns torice research in the Philippines:Domestic benefits and foreignspillover. Economic Developmentand Cultural Change 26(3): 591-607.

Franzel, S., D. Phiri, and F.K. Wesiga.1999. Assessing the adoption ofimproved fallows in EasternZambia. In S. Franzel and S.Scherr (eds.), Assessing theAdoption Potential of AgroforestryPractices in Africa. Nairobi:International Centre for Researchin Agroforestry (ICRAF).

Gómez, M.I. 1999. Economic Benefitsof Research Cooperation: TheCase of the Regional MaizeProgram for Central America andthe Caribbean. Ph.D. thesis,University of Illinois at Urbana-Champaign.

Hayami, Y., and R.W. Herdt. 1977.Market price effects oftechnological change on incomedistribution in semi-subsistenceagriculture. American Journal ofAgricultural Economics 59(2): 245-256.

Hayami, Y., with M. Kikuchi, P.F.Moya, L.M. Bambo, and E.B.Marciano. 1978. Anatomy of aPeasant Economy: A Rice Village inthe Philippines. Los Baños:International Rice ResearchInstitute (IRRI).

Hazell, P., and S. Haggblade. 1993.Farm-nonfarm growth linkagesand the welfare of the poor. In M.Lipton and J. van de Gaag (eds.),Including the Poor. Washington,D.C.: The World Bank.

Hazell, P.B.R., and. C. Ramaswamy.1991. Green RevolutionReconsidered: The Impact of theHigh-Yielding Rice Varieties inSouth India. Baltimore: JohnsHopkins.

Herdt, R.W. 1979. An overview of theconstraints project results. In FarmLevel Constraints to High RiceYields in Asia: 1974-1977. LosBaños: International RiceResearch Institute (IRRI).

Herdt, R.W. 1983. Mechanization ofrice production in developingAsian countries: Perspectives,evidence and issues. InConsequences of Small FarmMechanization. Los Baños:International Rice ResearchInstitute (IRRI).

Herdt, R.W. 1987. A retrospectiveview of technological and otherchanges in Philippine ricefarming, 1965–1982. EconomicDevelopment and Cultural Change35(2): 329–49.

Herdt, R.W., and C. Capule. 1983.Adoption, Spread, and ProductionImpact of Modern Rice Varieties inAsia. Los Baños: InternationalRice Research Institute (IRRI).

Herdt, R.W., L. Castillo, and S.Jayasuriya. 1984. The economicsof insect control in thePhilippines. In Judicious andEfficient Use of Insecticides on Rice.International Rice ResearchInstitute (IRRI).

Herdt, R.W., and A.M. Mandac. 1981.Modern technology andeconomic efficiency of Philippinerice farmers. EconomicDevelopment and Cultural Change29(2): 375–99.

Hossain, M. 1998. Rice research,technological progress, and theimpact on the rural economy: TheBangladesh case. In P.L. Pingaliand M. Hossain (eds.), Impact ofRice Research. Proceedings of theInternational Conference on theImpact of Rice Research, 3–5 Jun1996, Bangkok, Thailand.Bangkok and Los Baños: ThailandDevelopment Research Institute(TDRI) and International RiceResearch Institute (IRRI).

IRRI (International Rice ResearchInstitute). 1972. IRRI AnnualReport. Los Banos: IRRI.

IRRI (International Rice ResearchInstitute). 1978. EconomicConsequences of the New RiceTechnology. Los Banos: IRRI.

IRRI (International Rice ResearchInstitute). 1986. Small FarmEquipment for DevelopingCountries. Manila: IRRI.

Litsinger, J.A. 1989. Secondgeneration insect pest problemson high yielding rices. TropicalPest Management 35(3): 235–42.

López-Pereira, M.A., and M.L.Morris. 1994. Impacts ofInternational Maize BreedingResearch in the Developing World,1966–1990. Mexico, D.F.:International Maize and WheatImprovement Center (CIMMYT).

Maredia, M.K., and D. Byerlee (eds.).1999. The Global WheatImprovement System: Prospects forEnhancing Efficiency in the Presenceof Spillovers. CIMMYT ResearchReport No. 5. Mexico, D.F.:International Maize and WheatImprovement Center (CIMMYT).

McIntire, J., D. Bourzat, and P.Pingali. 1992. Crop-LivestockInteraction in Sub-Saharan Africa.Washington, D.C.: The WorldBank.

Morris, M.L., H.J. Dubin, and T.Pokhrel. 1994. Returns to wheatbreeding research in Nepal.Agricultural Economics 10(3): 269-282

Norgaard, R.B. 1988. The biologicalcontrol of the cassava mealybugin Africa. American Journal ofAgricultural Economics 70: 366-371.

Pardey, P.G., J.M. Alston, J.E.Christian, and S. Fan. 1996.Hidden Harvest: U.S. Benefits fromInternational Research Aid. FoodPolicy Report. Washington, D.C.:International Food PolicyResearch Institute (IFPRI).

Pardey, P.G., R.K. Lindner, E.Abdurachman, S. Wood, S. Fan,W.M. Eveleens, B. Zhang, andJ.M. Alston. 1992. The EconomicReturns to Indonesian Rice andSoybean Research. Jakarta and theHague: Agency for AgriculturalResearch and Development(AARD) and International Servicefor National AgriculturalResearch (ISNAR). Unpublishedreport.

Paris, T.R. 1998. The impact oftechnologies on women in Asianrice farming. In P.L. Pingali andM. Hossain (eds.), Impact of RiceResearch. Proceedings of theInternational Conference on theImpact of Rice Research, 3–5 Jun1996, Bangkok, Thailand.Bangkok and Los Baños:Thailand Development ResearchInstitute (TDRI) and InternationalRice Research Institute (IRRI).

16

Pingali, P.L., Y. Bigot, and H.P.Binswanger. 1987.Agriculturalmechanization and the evolutionof farming systems in Sub-Saharan Africa. Baltimore: JohnsHopkins.

Pingali, P.L., M. Hossain, and R.V.Gerpacio. 1997. Asian Rice Bowls:The Returning Crisis?Wallingford: CAB International.

Pingali, P.L., C.B. Marquez, and F.G.Palis. 1994. Pesticides andPhilippine rice farmer health: Amedical and economic analysis.American Journal of AgriculturalEconomics 76(3): 587–92.

Pingali, P., and P. Roger (eds.). 1995.Impact of Pesticides on the RiceEnvironment and Human Health.Boston: Kluwer AcademicPublishers.

Pinstrup-Anderson, P., and P.B.R.Hazell. (1985). The impact of theGreen Revolution and prospectsfor the future. Food ReviewInternational 1(1): 1-25.

Pinstrup-Andersen, P., N. Ruiz deLondoño, and E. Hoover. 1976.The impact of increasing foodsupply on human nutrition:Implications for commoditypriorities in agricultural research.American Journal of AgriculturalEconomics 58(2): 131-42.

Quisumbing, A.R., L. Haddad, andC. Peña. 1995. Gender and Poverty:New Evidence From 10 DevelopingCountries. Food Consumptionand Nutrition Division (FCND)Discussion Paper No. 9.Washington, D.C.: InternationalFood Policy Research Institute(IFPRI).

Renkow, M. 1993. Differentialtechnology adoption and incomedistribution in Pakistan:Implications for researchresource allocation. AmericanJournal of Agricultural Economics75(1): 33-43.

Rola, A., and P. Pingali. 1993.Pesticides, Rice Productivity, andFarmers’ Health. Manila andWashington, D.C.: InternationalRice Research Institute (IRRI)andWorld Resources Institute (WRI).

Scobie, G.M., and R.T. Posada. 1977.The Impact of High-Yielding RiceVarieties in Latin America—withSpecial Emphasis on Colombia.Series JE-01. Cali: CentroInternacional de AgriculturaTropical (CIAT).

Scobie, G.M., and R.T. Posada. 1978.The impact of technical change onincome distribution: The case ofrice in Colombia. American Journalof Agricultural Economics 60(1): 85-92.

Shapiro, B.I. 1990. Potentialconstraints, policy, and newtechnologies in the NiameyRegion of Niger. Ph.D. thesis,Purdue University, WestLafayette, Indiana.

Sidhu, D.S., and D. Byerlee. 1992.Technical Change and WheatProductivity in the Indian Punjab inthe Post-Green Revolution Period.CIMMYT Economics WorkingPaper 92-02. Mexico, D.F.:International Maize and WheatImprovement Center (CIMMYT).

Smale, M. (ed.). 1998. Farmers, GeneBanks, and Crop Breeding: EconomicAnalyses of Diversity in Wheat,Maize, and Rice. Boston: KluwerAcademic Publishers.

Smale, M., R.P. Singh, K. Sayre, P.Pingali, S. Rajaram, and H.J.Dubin. 1998. Estimating theeconomic impact of breedingnonspecific resistance to leaf rustin modern bread wheats. PlantDisease 82(9): 1055–61.

TAC/SPIA (Technical AdvisoryCommittee/Standing Panel onImpact Assessment). The Future ofImpact Assessment in the CGIAR:Needs, Constraints, and Options.Washington, DC: TAC/SPIA(forthcoming).

Traxler, G., and D. Byerlee. 1992.Economic returns to cropmanagement research in a post-green revolution setting. AmericanJournal of Agricultural Economics74(3): 573-582.

Unnevehr, L.J. 1986. Consumerdemand for rice grain quality andreturns to research for qualityimprovement in Southeast Asia.American Journal of AgriculturalEconomics 68(3): 634-641.

Waibel, H. 1986. The Economics ofIntegrated Pest Control in IrrigatedRice: A Case Study from thePhilippines. Berlin: Springer-Verlag.

Waibel, H. 2000. Impact of IntegratedPest Management Research in theCGIAR. Rome: ImpactAssessment and EvaluationGroup, Technical AdvisoryCommittee of the CGIAR.

Walker, T.S. 1989. High-yieldingvarieties and variability insorghum and pearl milletproduction in India. In J.R.Anderson and P.B.R. Hazell(eds.), Variability in Grain Yields:Implications for AgriculturalResearch and Policy in DevelopingCountries. Baltimore: JohnsHopkins.

Walker, T., and C.C. Crissman. 1996.Case Studies of the Economic Impactof CIP-related Technologies. Lima:International Potato Center (CIP).

Walker, T., and J. Ryan. 1991. Villageand Household Economies in India’sSemi-arid Tropics. Baltimore: JohnsHopkins.

Winkelmann, D. 1976. The Adoption ofNew Maize Technology in PlanPuebla, Mexico. Mexico, D.F.:International Maize and WheatImprovement Center (CIMMYT).

Witcombe, J.R. 1989. Variability in theyield of pearl millet varieties andhybrids in India and Pakistan. InJ.R. Anderson and P.B.R. Hazell(eds.), Variability in Grain Yields:Implications for AgriculturalResearch and Policy in DevelopingCountries. Baltimore: JohnsHopkins.

17

Antle, J.M., and P.L. Pingali. 1994.Pesticides, productivity, and farmer health:A Philippine case study. American Journalof Agricultural Economics 76(3): 418–30.

Rice farmers from two regions in thePhilippines were interviewed to obtain datafor an analysis on human health andproductivity tradeoffs from limitedpesticide use. Through the models utilizedin this paper, it is determined that“pesticide use has a negative effect onfarmer health, [and] that farmer health has apositive effect on productivity.” Afterexamining the health of the sample farmers,Antle and Pingali used regression analysisto estimate the cost of health treatments andthe opportunity cost of lost labor frompesticide-related illnesses. These resultswere then compared with the change inproductivity from possible taxes oninsecticides and herbicides which wouldaffect incentives and ultimately reducepesticide use. Antle and Pingali concludethat there are two probable solutions to theproblems of pesticide use on human health;either restrict pesticide use through policymeasures such as taxes or, alternatively, finddifferent pest management methods thatinvolve the use of safer chemicals or nochemicals at all.

Binswanger, H.P., and J.G. Ryan. 1977.Efficiency and equity issues in ex anteallocation of research resources. IndianJournal of Agricultural Economics 32(3):217–31.

This paper discusses the informationalcomponents necessary to make effectivedecisions regarding research resourceallocation. Two principal topics, efficiencyand equity, are thoroughly examined.Related to efficiency, Binswanger and Ryanpresent topics including returns to scale, theadvantages of private- and public-sectorresearch, and the effect of factor shortage onresearch endeavors. In discussing equity,the authors compare distribution oftechnological gains between producers andconsumers, laborers and labor owners, anddiffering regions. It is mentioned thatcertain technologies are site-specific andthus disproportionately benefit farmers orlaborers of particular regions.

Brennan, J.P. 1984. Measuring thecontribution of new varieties to increasingwheat yields. Review of Marketing andAgricultural Economics 52(3): 175–95.

In determining the consequences of varietalchange on production, Brennan explainsand evaluates three measures: (1) the indexof varietal newness, (2) the proportion ofarea sown to recent varieties, and (3) theindex of varietal improvement. Brennan

Impact Assessment StudiesConducted in the CGIAR, 1970-1999:An Annotated BibliographyMatthew P. Feldman

Note: This bibliography covers many, although not all, of the studies cited in this publication.Readers interested in more detailed information on the studies that do not appear in this bibliographyare encouraged to consult the authors of those studies or the original publications themselves.

18

describes each measure by explaining itsfunction, including the endogenous andexogenous variables, its advantages, and itslimitations. An empirical example isevaluated with data from Mitchell Shire insouthern New South Wales between 1950 and1981. The author concludes that one’s choiceof measure is critical since there is not a highstatistical correlation, in general, between theresults of each index.

Brennan, J.P. 1986. Impact of WheatVarieties from CIMMYT on Australian WheatProduction. Agricultural Economics BulletinNo. 5. Sydney: Department of Agriculture,New South Wales.

Benefits from CIMMYT-based varieties(CBVs) have been recognized around theworld in developed and developingcountries. In this analysis, Brennan evaluatesthe benefits and impact of CBVs of wheatthat are commercially cultivated in Australia.He assesses the advantages of CBVs inincreasing yields, quality, disease resistance,and their adoption by region, variety, type,class, and grade. Additionally, he conductsan extensive quantitative analysis of theestimated benefits of CBVs by Australianstates between 1974 and 1983. Brennanemphasizes the high adoption rates ofCIMMYT-based wheat varieties throughoutAustralia and concludes that the benefits ofCBVs in Australia have been profound sincethe 1960s.

Byerlee, D., and P. Moya. 1993. Impacts ofInternational Wheat Breeding Research in theDeveloping World, 1966–90. Mexico, D.F.:International Maize and Wheat ImprovementCenter (CIMMYT).

Since the Green Revolution, the area plantedto modern varieties (MVs) has steadilyincreased throughout the developing world.Byerlee and Moya conducted an extensiveanalysis of data from 38 developing countriesthat included data on the area planted to

modern wheat varieties, percentage ofadopted varieties from CIMMYTgermplasm, and the benefits of CIMMYTresearch during the post-Green Revolutionperiod of 1977–90. The authors estimate thatof the 600 varieties released in thedeveloping world between 1966 and 1990,75% came directly from CIMMYT crosses orwere the result of crosses with CIMMYTparents by national agricultural researchsystems (NARSs). The new varietiesembodied many benefits, including yieldgains and disease resistance, which helpedCIMMYT to obtain a 54% internal rate ofreturn on its investment. In conclusion, theauthors emphasize that additional researchwill need to focus on improving grainquality and preserving genetic diversity.

Byerlee, D., and E. Hesse de Polanco.1986. Farmers’ stepwise adoption oftechnological packages: Evidence from theMexican altiplano. American Journal ofAgricultural Economics 68(3): 519–27.

The adoption of inputs, either in a stepwisemanner or as a package, is dependent uponthe relative profitability of each input andnot solely on yield increases. Byerlee andHesse evaluate the adoption of improvedbarley varieties, fertilizer, and herbicides intwo regions in Mexico between 1975 and1980—a dry region with an annual rainfallof 450–550 mm and a wet region withannual rainfall of 600–700 mm. Their resultsemphasize that the three inputs studied arenot perfect complements in production,even though interaction between inputsdoes exist, and thus are frequently adoptedin a stepwise fashion to minimize risk.

Byerlee, D., and M. Morris. 1993.Research for marginal environments: Are weunderinvested? Food Policy 18(5): 381–93.

As remnants of the Green Revolution forwheat are seen today through relativelyhigh yields in the developing world,

19

research has begun to focus on the disparitybetween the benefits of research forirrigated lands and “marginal”environments plagued by poor watercontrol. By simply recognizing the lowadoption rates of modern varieties and lowyields in marginal environments, peopleoften believe that more investment isneeded in this area. As Byerlee and Morrisexplain, the focus on marginal regions mustcorrespond with the value of the crop fromthat locale. Background information,including possible reasons for why moreresources should be placed on research formarginal areas and reasons for lowadoption of modern varieties in marginalenvironments, is presented along with amodel to determine over andunderinvestment. By using a congruencymodel, the authors find that CIMMYT’swheat breeding program and India’sresearch on wheat were not marked byunderinvestment in research for marginalregions.

Crissman, C.C., J.M. Antle, and S.M.Capalbo (eds.). 1998. Economic,Environmental, and Health Tradeoffs inAgriculture: Pesticides and the Sustainabilityof Andean Potato Production. Boston:Kluwer Academic Publishers.

In this book, the authors use data from theCarchi Province of Ecuador to quantify theimpact of pesticide use on potatoproduction, on the environment, and onhuman health. The book addresses thedemographics of the study region, themethods of assessing environmental andhuman health impairments, the effect ofgovernment policies and exchange rates onpesticide use, and simulation modelsincorporating the environment, humanhealth, and economic tradeoffs. The authorsdetermine that a ban on pesticide use is nota likely solution; a more probable solutionto the problem would be achieved by

improving farmer safety measures andapplication strategies, as well as restrictingpesticide use. The models in this bookprovide a quantitative framework toillustrate the tradeoff between productivity,farmer health, and environmental factors.

Dalrymple, D.G. 1977. Evaluating theimpact of international research on wheatand rice production in the developingnations. In T.M. Arndt, D.G. Dalrymple,and V.W. Ruttan (eds.), ResourceAllocation and Productivity in National andInternational Agricultural Research.Minneapolis: University of Minnesota.

This early study of the impact of rice andwheat research in the developing worldestimates changes in yield and area as wellas the role of high-yielding varieties (HYVs)on these changes. It gives backgroundinformation on the International Maize andWheat Improvement Center (CIMMYT) andthe International Rice Research Institute(IRRI), two initial centers of theConsultative Group on InternationalAgricultural Research (CGIAR). Topicsinclude the direct and indirect effects ofHYVs, the gap between yields in researchplots and farmers’ fields, changes in areassown to HYVs and crop yield, as well as theeffects of innovative technologies on riceand wheat production in developingcountries. Two methods are used todetermine the impact of new technologies:production function and index numberanalyses. The implications and limitationsof these methods are described prior toestimations of the value of increased outputin the early 1970s, principally in Asia. Theanalysis concludes by emphasizing theimportance of impact research indemonstrating the benefits of agriculturalresearch.

20

Dalrymple, D.G. 1978. Development andSpread of High-Yielding Varieties of Wheatand Rice in the Less Developed Nations.USDA Foreign Agricultural EconomicsReport No. 95. Washington, D.C.: UnitedStates Department of Agriculture (USDA).

During the 12-year period of the GreenRevolution (1965/66–1976/77), the areaplanted to high-yielding varieties (HYVs) ofrice and wheat in developing countriesgreatly increased. This paper presentsinformation on the origins of semidwarfwheat and rice varieties and thedevelopment of many HYVs by theInternational Maize and WheatImprovement Center (CIMMYT) and theInternational Rice Research Institute (IRRI).A vast majority of the data regarding areaplanted to HYVs is a reflection of the areaplanted to varieties supplied directly fromCIMMYT or IRRI or having CIMMYT orIRRI parents. In the case of high-yieldingwheat, data on area planted to HYVs inspecific countries over all (or portions) ofthe 12-year period are presented for Asia,the Near East, Africa, and Latin America. Inthe case of high-yielding rice, data on areaplanted are available for Asia, theCommunist nations of Asia (Lao PDR,People’s Republic of China, and Vietnam),the Near East, Africa, and Latin America. Ina more aggregate form, the results showthat about 19.7 million hectares of HYVwheat and 24.2 million hectares of HYV ricewere planted in developing countries inAsia (excluding Communist nations andTaiwan) during 1976-77, making Asia thepredominant adopter of HYVs compared toother regions in the developing world. Thetotal area of HYVs of wheat and rice indeveloping countries (excludingCommunist nations, Taiwan, Israel, andSouth Africa) in 1976-77 was 29.4 millionhectares and 25.3 million hectares,respectively.

Dalrymple, D.G. 1986. Development andSpread of High-Yielding Rice Varieties inDeveloping Countries. Washington, D.C.:United States Agency for InternationalDevelopment (USAID).

This publication provides estimates of thearea planted to high-yielding rice varieties(HYRVs) from the mid-1960s to the mid-1980s in four geographical regions—Southand East Asia, the Near East, Africa, andLatin America, and in 67 developingcountries. Dalrymple presents specific time-series data for a few countries in South andEast Asia, where HYRVs are widelyadopted and data are more often available,and for a few countries in Latin America.Country details related to area planted inselected years and the choice of ricevarieties are presented for other countries inthe study. The aggregate area planted toHYRVs in developing countries in the fourregions, excluding communist nations andTaiwan, was about 39.2 million hectares in1982/83. When communist nations wereincluded (excluding North Korea), thefigure rose to 72.6 million hectares.

David, C.C. 1976. Fertilizer demand in theAsian rice economy. Food Research InstituteStudies XV(1).

Using aggregate data from 11 Asian rice-growing countries and two farm surveys(one from the Philippines, Indonesia,Thailand, Malaysia, India, and Pakistan,and the other from Laguna, Philippines),David presents a highly quantitativeanalysis of fertilizer demand functions.Determinants such as rice yield, fertilizeruse, fertilizer-rice price ratio, and priceelasticities are evaluated. The author studiesthe dissemination of modern varieties, thequality of the environment and irrigationsystems, productivity of fertilizers, and thefertilizer-rice price ratio in explainingvarying consumption of fertilizer acrosscountries and within regions.

21

David, C.C., and R. Barker. 1978. Modernrice varieties and fertilizer consumption. InInternational Rice Research Institute (ed.),Economic Consequences of the New RiceTechnology. Los Baños: International RiceResearch Institute (IRRI).

As land resources become more scarce andpopulations continue to grow, the ability ofmodern rice varieties (MVs) to respond tofertilizer application has become increasinglyimportant. In this quantitative analysis offactors contributing to varying fertilizerconsumption, the authors use three datasets:one aggregate set from 12 Asian countries,one farmer survey from six Asian rice-growing countries, and one farmer surveyfrom Laguna, Philippines. Before presentingthe results of their analysis, the authorscompare MV and traditional variety (TV)responsiveness to fertilizer applications.Then the datasets are used to estimatefertilizer demand functions for Asian rice-growing countries. By using two demandmodels, the authors derive the relativecontributions of the fertilizer-rice price ratio,percentage of area planted to MVs, quality ofirrigation, output value, and quality of theenvironment in explaining the gap betweenaverage and high levels of fertilizerconsumption. The Asian aggregate data andLaguna farmer data show that MV adoptionis a principal factor in explaining the gapbetween average and high levels of fertilizeruse. The fertilizer-rice price ratio alsoexplains a large portion of the differences infertilizer use in all three datasets.

David, C.C., V.G. Cordova, and K. Otsuka.1994. Technological change, land reform, andincome distribution in the Philippines. In C.C.David and K. Otsuka (eds.), Modern RiceTechnology and Income Distribution in Asia.Boulder: Lynne Rienner.

Through a highly quantitative analysis ofincome distribution in the Philippinesbetween favorable and unfavorable croppingenvironments, the authors evaluate the

factors that influence a distinct incomedifferential. In part because of a focus onimproving varieties for irrigated andfavorable rainfed regions, the productivitygap between regions increases. Regressionanalysis is utilized to determine the effect offarm size, land tenure, ownership of capital,and technological factors on income. Non-rice income proved to be the most importantform of income leading to the incomedifference between favorable andunfavorable environments. In conclusion, theauthors recommended that emphasis begiven to increasing human capital andimproving research on rice for unfavorableenvironments to decrease the inequality.

Doss, C.R. 1999. Twenty-Five Years ofResearch on Women Farmers in Africa:Lessons and Implications for AgriculturalResearch Institutions; with an AnnotatedBibliography. CIMMYT Economics ProgramPaper No. 99-02. Mexico, D.F.: InternationalMaize and Wheat Improvement Center(CIMMYT).

Even though broad, definitivegeneralizations cannot be made on whyenhanced agricultural innovations are notquickly adopted by female African farmers,this review of a wide array of literature helpsto clarify a few potential possibilities.Important questions are raised in this studyand potential answers are given for specificregions, but few generalizations for Africa asa whole are presented. Doss emphasizes thatlack of access to labor, land, credit, fertilizer,extension, and mechanization are somepotential reasons for the lack of use of newmaize technologies. Specific cases fromregions of Africa help illuminate the diversityin gender issues. The author cautions that thegender situation is continually changing andthus it is important to conduct before andafter research on technology adoption. Thereview is supplemented by a comprehensiveannotated bibliography for over 175references on gender issues and Africanagriculture.

22

Echeverría, R.G. 1990. Assessing theimpact of agricultural research. In R.G.Echeverría (ed.), Methods for DiagnosingResearch System Constraints and Assessingthe Impact of Agricultural Research. Vol. II:Assessing the Impact of AgriculturalResearch. The Hague: International Servicefor National Agricultural Research (ISNAR).

Echeverría reviews more than 100 studiesestimating rates of return to various cropsthroughout the developed and developingworld. He emphasizes that impactassessment has changed over the last fewdecades as ex post and ex ante research havestarted to move beyond general cost-benefitstudies of innovative technology. Currentresearch has begun to focus on spillovereffects, distribution of benefits and income,government intervention in agriculture, andthe effects of agriculture policy. Echeverríaemphasizes that high rates of return foragricultural research signifyunderinvestment in this area. Following thisintroduction, the first half of the bookfocuses on issues such as the consequencesof omitting private-sector research,measurement of changes in consumer-producer surplus resulting from qualityimprovement, and other topics. The secondhalf of the book contains more site-specificcase studies addressing research andextension in Peru, and expected returns topasture improvement in Latin America.

Echeverría, R.G., G. Ferreira, and M.Dabezies. 1991. Returns to Investment in theGeneration and Transfer of Rice Technologyin Uruguay: The Case of Rice. ISNAR StaffNotes 89-50(s). The Hague: InternationalService for National Agricultural Research(ISNAR).

In this study of rice production in Uruguaybetween 1965 and 1985, the authors give ahistorical account of rice in the countrybefore estimating returns from technologicalchange. Even though there are manycomponents that must be accounted for in

technological change, this study includespublic and private extension and privateresearch. This analysis gives backgroundinformation including the area, production,and yield of rice for Uruguay from 1931 to1988, rice exports from 1973 to 1986,countries importing rice from Uruguay, andrice prices between 1981 and 1987. Inaddition, it is determined that the rate ofreturn to research and extension investmentbetween 1965 and 1985 in Uruguay was 52%,while the benefit-cost ratio was 5.5. Theauthors also determine that producerscaptured a large portion of the benefits fromresearch and extension. In conclusion, specialemphasis is placed on the concept of spillins,as countries such as Uruguay must be able toinvestigate, study, and introduce geneticmaterials and management techniques fromother countries to realize high rates of return.

Evenson, R., and C. David. 1993.Adjustment and Technology: The Case ofRice. Development Centre Studies. Paris,France: Organisation for EconomicCooperation and Development (OECD).

During a time of market liberalization,privatization, and structural reform in manycountries recovering from the financial crisesof the 1980s, Evenson and David emphasizethat “for inherently public good activitiessuch as rice research the invisible hand of themarket is not relevant.” They believe thatprivatization of rice research cannot take theplace of public research. In addition topresenting information on past riceconsumption and production, the use ofgenetic resources, distribution of benefitsfrom research, and investment issues, theauthors evaluate the effects of“mismanagement” crises from the 1980s onrice research. Following the crises, manygovernments moved in support of marketeconomies in their efforts to bring aboutstructural reform in the public system. Fromthe available data, it is believed that the“effectiveness” of rice research did notchange during the 1980s, but investment in

23

research declined in several countries whichhave not seen an equal increase during thereform era.

Fan, S., P. Hazell, and S. Thorat. 1998.Government Spending, Growth, and Poverty:An Analysis of Interlinkages in Rural India.EPTD Discussion Paper No. 33. Washington,D.C.: International Food Policy ResearchInstitute (IFPRI).

Since the mid-1960s, the number of peopleliving below the poverty line in India hasdropped dramatically. For this trend tocontinue, it is important to evaluate themarginal impact of various governmentexpenditures geared toward achieving areduction in poverty. The authors use asimultaneous equation model to determinewhich factors can provide the greatest impacton rural poverty reduction and agriculturalproductivity growth. Governmentexpenditures studied include the impact ofrural roads, agricultural research anddevelopment, irrigation, rural electrification,and education. Through estimates ofelasticities and numbers of people whowould be brought above the poverty line bymarginal expenditures, it is determined thatadditional government expenditure on ruralroads would provide the greatest impact onrural poverty. On the other hand, additionalexpenditure on agriculture research anddevelopment would provide the greatestimpact on growth in agriculturalproductivity.

Flinn, J.C., and D.P. Garrity. 1989. Yieldstability and modern rice technology. In J.R.Anderson and P.B.R. Hazell (eds.),Variability in Grain Yields: Implications forAgricultural Research and Policy inDeveloping Countries. Baltimore: JohnsHopkins.

This paper looks at the effects of modern ricevarieties and related technologies onproduction stability in Asia. Factors such as

irrigation, cropping intensity, area, andpesticide and fertilizer use are analyzed inrelation to their effect on productionstability. A small case study of Luzon andMindanao, Philippines, between 1974 and1978 is used to illustrate some comparisonsof the variability of production based onwater control and varieties. In addition, theauthors explain the importance ofprotecting diversity and developingeffective crop and soil managementstrategies.

Franzel, S., D. Phiri, and F. Kwesiga.1999. Assessing the adoption potential ofimproved fallows in Eastern Zambia. In S.Franzel and S. Scherr (eds.), Trees andFarmers: Assessing the Adoption Potentialof Agroforestry Practices in Africa.

A five-year improved fallow and maizerotation plan has rapidly become a popularmethod of cropping in regions of easternZambia. The improved fallow techniqueinvolves planting trees for two years tohelp rejuvenate the soil before plantingthree consecutive years of maize. Franzel,Phiri, and Kwesiga attempt to evaluate theadvantages and disadvantages of thissystem compared to continuous plantingsof unfertilized or fertilized maize. Theyfind that over the five-year period, higherreturns to land and labor accrued to theimproved fallow technique over theunfertilized maize pattern, while thereturns to labor were higher for theimproved fallow method than thecontinuous fertilized maize technique.Although under certain conditions thebenefit difference between improvedfallow and fertilized maize is not stark, itmust be emphasized that the relationshipin benefits between these two methodsdepends greatly on the prices of maize andfertilizer. Over the five-year cycle, fertilizedmaize has the highest overall yield.

24

Flores-Moya, P., R.E. Evenson, and Y.Hayami. 1978. Social returns to rice researchin the Philippines: Domestic benefits andforeign spillover. Economic Developmentand Cultural Change 26(3): 591–607.

In this paper the authors estimate thebenefit/cost ratio, internal rates of return,supply and demand elasticities, andspillover effects of rice research in thePhilippines. A review of past rice researchin the Philippines precedes a description ofthe methodology used to calculate socialreturns and the actual calculations. Factorssuch as price elasticities, research costs atinternational and national research facilitiesin the Philippines, and the shift in the riceproduction function owing to technologicalinnovation are quantified. The estimationsection shows that in a closed economy andin the tropical world as a whole, consumersreap the rewards of research, whereasproducers fare worse than prior to theresearch. It is estimated that on averageevery dollar invested in research in thePhilippines translates into a benefit of US$ 4to the Philippines. Additionally, thebenefit/cost ratio is much higher for thetropical world, which illustrates the impactof spillover effects.

Gómez, M.I. 1999. Economic Benefits ofResearch Cooperation: The Case of theRegional Maize Program for Central Americaand the Caribbean. Ph.D. thesis, Universityof Illinois, Urbana-Champaign, Illinois.

Over the past decade, research on impactassessment has begun to include spillovereffects. When these effects are not taken intoaccount, estimates of returns to investmentand thus financial resource allocation canoften be skewed. This award-winning thesisdescribes and analyzes the sources of thevast benefits of the cooperative breedingprogram of the Programa Regional de Maíz(PRM, or Regional Maize Program) inCentral America and the Caribbean. Theauthor identifies and quantifies spillovers

within the region and conducts cost-benefitanalyses on public breeding research.Gómez estimates that close to two-thirds ofthe maize research impact in the study areacomes from spillins from outside the region.For the majority of countries in the region, itis more beneficial to invest in the PRM thanin national agricultural research. Investmentin the PRM opens the door to receivinglarge spillin benefits from the collaborativeinternational effort, which can capitalize onspecialization and returns to scale.

Hayami, Y., and R.W. Herdt. 1977.Market price effects of technological changeon income distribution in semisubsistenceagriculture. American Journal of AgriculturalEconomics 59(2): 245–56.

Hayami and Herdt emphasize the need fortechnological innovations to shift thesupply function at a faster rate than theshift in the demand function. Through astudy of rice variety innovations in thePhilippines, the authors quantitativelyanalyze the income distribution betweensectors and within the rural sector using amodel they develop. Estimated factorsinclude changes in consumer surplus,producer income, and income distributionbetween large- and small-scale farmers forboth fixed and variable consumption of thissemisubsistence crop. In concluding, theauthors state that more equality can bereached if innovations shift the supplyfunction faster than the demand function,allowing prices to fall.

Hazell, P.B.R., and C. Ramaswamy. 1991.The Green Revolution Reconsidered: TheImpact of High-Yielding Rice Varieties inSouth India. Baltimore: Johns Hopkins.

In this series of papers on the effects of theGreen Revolution in rice in North ArcotDistrict of South India, Hazell andRamaswamy show that the rural poorreceived both direct and indirect benefits

25

from the technological innovations. Initiallythe new technology was adoptedprincipally by large-scale farmers, but astime passed smaller farms began adoptinghigh-yielding varieties at similar rates aslarge-scale farmers. It is estimated that newtechnology increased rice output, loweredpoverty, improved diets, and positivelyaffected the non-farm economy, withoutleading to broad changes in the number oflandless people in adoption regions.Between 1973-74 and 1983-84 “small paddyfarmers and landless laborers gained thelargest proportional increases in familyincome.” The authors point out that theresults of these papers are quite differentfrom those of previous studies, whichevaluated the new technology too early inthe adoption process. Since many studieswere conducted too early, some concludedthat the new technology was not scale-neutral, had fostered inequality, and hadresulted in disproportionate gains by large-scale farmers.

Herdt, R.W. 1987. A retrospective view oftechnological and other changes in Philippinerice farming, 1965–1982. EconomicDevelopment and Cultural Change 35(2):329–49.

In response to reports stating thatinnovative technology was principallyutilized by large-scale farmers, Herdt usescase studies from central Luzon andLaguna, Philippines, over a 15-year periodto show that adoption of modern ricevarieties was relatively independent ofscale. In Laguna there was no significance inadoption based on farm size. In the Luzonsample, on the other hand, small-scalefarmers’ adoption of the new technologywas slower in the beginning, but at the endof the study period there was no differencein adoption rates between small- and large-scale farmers. Data indicate increases inproduction from 1966 to 1982, labor use perhectare, double cropping, the use of powertillers, and the use of fertilizer and

irrigation. In addition, many farmersswitched from sharecropping toleaseholding. Even though there were largeincreases in yields, farmers’ real incomes didnot change dramatically because of largeincreases in production costs. Thus, duringthe study period, real rice prices weredepressed and consumers were the principalbeneficiaries of technological change.

Herdt, R.W., and C. Capule. 1983.Adoption, Spread, and Production Impact ofModern Rice Varieties in Asia. Los Baños:International Rice Research Institute (IRRI).

In this report, the authors discuss adoptionof modern rice varieties (MVs) in South andSoutheast Asia, increased production since1965, and factors correlated with adoption.The authors describe adoption, key players,IRRI’s role in the changes, percentage of areaplanted with MVs, and land categories in 11countries: Bangladesh, Burma, India,Indonesia, the Republic of Korea, Malaysia,Nepal, Pakistan, the Philippines, Sri Lanka,and Thailand. Modern rice varieties wereadopted most rapidly in the Philippines,with 89% of irrigated rice land and 77% ofrainfed rice area planted to MVs in 1979-80.In addition, the authors discuss previousregression analyses of variables related toadoption of MVs and fertilizer, includingage, schooling, technological knowledge,social class, and family size, and factors thatinfluence variable costs, including farm size,tenure, wealth, fertilizer, and creditavailability.

Herdt, R.W., and A.M. Mandac. 1981.Modern technology and economic efficiencyof Philippine rice farmers. EconomicDevelopment and Cultural Change 29(2):375–99.

With data from Nueva Ecija, Philippines,between 1974 and 1977, Herdt and Mandacattempt to determine factors which lead tothe yield gap between potential and realized

26

on-farm rice yields. Even though semidwarfvarieties have proven to have relativelyhigh yield capability on experiment stationscompared with traditional varieties, thesevast benefits have not been entirely realizedon the farm. The authors determine thatthree factors–profit-maximization attitudes,allocative inefficiency, and technicalinefficiency–contribute to the yield gap andthen quantify the relative effect of eachfactor on the yield discrepancy. Afterestimating production functions for eachfarm and utilizing regression analysis, theyconclude that farm size, information, andnumber of days off the farm are threeprincipal contributors to the economicinefficiency. Herdt and Mandac find thatfarm size is negatively related withefficiency (i.e., larger farms are lessefficient), whereas information and days offthe farm are positively related to efficiency.

Herdt, R.W., and T.H. Wickham. 1978.Exploring the gap between potential andactual rice yields: The Philippine case. InInternational Rice Research Institute (ed.),Economic Consequences of the New RiceTechnology. Los Baños: International RiceResearch Institute (IRRI).

Even though modern rice varieties (MVs)have been exalted as having potential yieldsof between 6 and 10 t/ha, these loftyexpectations have not been realized infarmers’ fields: a 1969-70 study in thePhilippines found yields of irrigated MVs tobe 2.1 t/ha. This paper attempts to discoverreasons for the gap in rice yields onexperiment stations and in farmers’ fields.The authors begin with an analysis of datafrom an IRRI experiment on rice varietyIR20 and estimate the “maximum attainablenational average yield” for the Philippinesis 4.1 t/ha, taking into account climatic,water, and environmental conditions. Theprincipal explanation for the yield gap isthe motivation of farmers to maximizeprofits more often than yields. The authors

explain that farmers must take factors suchas risk and diminishing returns intoconsideration. Of eight input use scenarios,three result in the maximum yielddelivering the maximum profit, while fivedemonstrate that yields below themaximum can result in larger profits. Inconclusion, Herdt and Wickham state thatwater control, solar radiation, season, riskand other economic determinants, yearlyvariations, and insect, disease, and weeddamage are the principal factors accountingfor the vast difference between actual andpotential yields of modern rice varieties.Thus efforts to diminish the yield gap mustfocus on these research areas.

Hossain, M. 1998. Rice research,technological progress, and the impact onthe rural economy: The Bangladesh case. InP.L. Pingali and M. Hossain (eds.), Impactof Rice Research. Proceedings of theInternational Conference on the Impact ofRice Research, 3–5 Jun 1996, Bangkok,Thailand. Bangkok and Los Baños: ThailandDevelopment Research Institute (TDRI) andInternational Rice Research Institute (IRRI).

Hossain chronicles agricultural researchefforts in Bangladesh and their progresstowards reducing poverty within thecountry. He presents data on researchinvestment, education of researchers,characteristics of modern varietiesdeveloped for the country, anddissemination of improved varieties. Acomparison of unit costs of rice productionbetween traditional and modern varietiesfrom 1973 to 1993 is presented, whichfacilitates benefit-cost estimates of researchinvestment in rice. Estimates show thatwhen evaluating cost reductions in riceproduction the benefit-cost ratio ofinvestment in production and technologytransfer is 16.6:1, but when one includes thesavings from a decline in food imports, theratio more than doubles to 36:1. Even

27

though it is frequently argued that benefitsof modern varieties accruedisproportionately to large-scale farmers,this study shows benefits to the poorersegments of the population from reducedrice prices and a reduction in poverty withinthe country.

Islam, Y., and J.L. Garrett. 1997. IFPRI andthe Abolition of the Wheat Flour RationShops in Pakistan: A Case-Study onPolicymaking and the Use and Impact ofResearch. Impact Assessment DiscussionPaper No. 1. Outreach Division.Washington, D.C.: International Food PolicyResearch Institute (IFPRI).

Through research and dissemination ofresults to government policymakers, IFPRIwas able to provide information whicheventually led, in part, to the abolishment ofPakistan’s wheat ration-shop system. Theauthors explain IFPRI’s five-stage researchprogram and its application to a case studyof the ration-shop system in Pakistan. Thefive stages are: (1) client consultation andresearch program design, (2) researchprogram implementation, (3) thecommunication of research results,(4) policymaking, and (5) impactassessment. The case study emphasizes thatthrough close collaboration with key actorsinside and outside the government, IFPRIwas able to tailor the research topolicymakers’ interests and provide thisinformation on a continual, timely basisinstead of through a final report, which wasnot published until 1988. The subsequentabolition of this ration-shop system wasdue, in part, to a wasteful and corruptsystem which did not have a strong impacton the poor population.

Litsinger, J.A. 1989. Second generationinsect pest problems on high yielding rices.Tropical Pest Management 35(3): 235–42.

Recent advances in irrigation, modernvariety development, and increasedcropping intensity have put pressure on riceyields as farmers now must battle pests thatno longer have a natural break in theirdevelopment cycles. With increasedcropping intensity there are no long fallowperiods in which the majority of the pestsdie. The government recommended virulentinsecticides to kill pests, but the insecticidesalso killed natural predators. Because ofthese insecticides and year-round cropping,pests have become more difficult to control.Litsinger recommends that communities settimetables for the crop season to providesome fallow time and develop integratedpest management (IPM) strategies.

López-Pereira, M.A., and M.L. Morris.1994. Impacts of International MaizeBreeding Research in the Developing World,1966–1990. Mexico, D.F.: InternationalMaize and Wheat Improvement Center(CIMMYT).

In this analysis of the adoption of improvedmaize varieties and hybrids in thedeveloping world between 1966 and 1990,López-Pereira and Morris use data from 45countries to estimate the area planted toimproved varieties and CIMMYT’scontribution to the development of thosevarieties. The authors estimate that in 1990,24.6 million hectares of land in developingcountries, or 43% of the area planted tomaize, was planted to improved maize.Since 57% of the area in developingcountries planted to maize was planted tounimproved varieties, there is opportunityfor further dissemination of improvedvarieties and hybrids. Of the 24.6 millionhectares of improved maize, 13.5 millionhectares were planted to improved maizecontaining CIMMYT germplasm. Between

28

1966 and 1990, 53% of the 842 maizevarieties and hybrids released by publicbreeding programs in developing countriescontained CIMMYT germplasm. In additionto the aggregate analysis, the authors alsopresent a breakdown of countries in sub-Saharan Africa, Asia, North Africa, andLatin America, describing releases and areaunder local and improved varieties.

Maredia, M.K., and D. Byerlee (eds.).1999. The Global Wheat ImprovementSystem: Prospects for Enhancing Efficiencyin the Presence of Spillovers. CIMMYTResearch Report No. 5. Mexico, D.F.:International Maize and Wheat ImprovementCenter (CIMMYT).

This study of the efficiency of theinternational wheat improvement systemattempts to evaluate the effects of spilloversin developing countries. One principal topicaddressed is the decision to breed broadlyadapted varieties for vast arrays ofenvironments versus breeding site-specificvarieties. By including the effects ofresearch spillovers between programs, bothnational and international, the authorsattempt to eliminate a bias in previous cost-benefit analyses. After evaluating the effectsof spillover benefits, the authors conductedan econometric analysis which found that28 of the 69 research programs studied indeveloping countries invested too much inwheat improvement research. Many wheatimprovement programs in developingcountries are too large for their mandateareas and would benefit from investing inprograms to screen varieties developedelsewhere. Economic topics, includingeconomies of scale and scope, are addressedas they apply to the comparative advantageof various research programs. Case studiesfrom India and Australia are presented toanalyze actual spillover effects and returnsto investment.

McIntire, J., D. Bourzat, and P. Pingali.1992. Crop-Livestock Interaction in Sub-Saharan Africa. Washington, D.C.: TheWorld Bank.

In this comprehensive analysis of crop-livestock interactions in sub-Saharan Africa,the authors investigate the potential benefitsand limitations of integrating these twoagricultural components. To provide a well-rounded study, evidence is obtained fromprevious research on farming systems aswell as recent surveys from 33 Africanlocales. After describing the four climaticregions—humid, subhumid, semiarid, andhighlands—the authors provide an overviewof crop-livestock interactions, the potentialfor competition for scare resources, andpossible complementary effects. Issues suchas animal traction, soil fertility, crop residue,and animal production are integrated intothe analysis prior to an overall evaluation ofpolicy recommendations. In conclusion, theauthors single out three obstacles toincreasing productive crop-livestockinteractions: (1) previous trade andgovernmental policies working againstAfrican agriculture, (2) scarcity ofinformation on profitable technology, and (3)a lack of farming methods that are bothprofitable and relevant to the regions ofAfrica. It is determined that agriculturalgrowth will not dramatically change even ifconstraints to crop-livestock integration arealleviated; rather, exogenous technicalchange should be implemented to helpexpedite growth.

Morris, M.L., H.J. Dubin, and T. Pokhrel.1994. Returns to wheat breeding research inNepal. Agricultural Economics 10(3): 269–82.

Because Nepalese research institutes havebeen established to test and spreadimproved, imported wheat varietiesthroughout the country, and because ofinteraction between Nepalese and Indianfarmers, Nepal has achieved high internal

29

rates of return to wheat breeding researchduring the Green Revolution (1965-90). Theauthors estimate ex post returns to researchinvestment during the Green Revolutionand provide an ex ante estimate of futurereturns to wheat breeding research (1990-present). They find that the internal rate ofreturn was 84% during the GreenRevolution and predict it to be 49% in thefuture. In conclusion, they acknowledge thatit may not be feasible for smaller countriesto conduct extensive breeding research, butas the case of Nepal indicates, it is possibleto capitalize on spillover effects frominternational centers and other countries toachieve a relatively high rate of return.

Moya, T.B., W.C. de la Viña, and S.I.Bhuiyan. 1994. Potential of on-farm reservoiruse for increasing productivity of rainfed riceareas: The Philippine case. In S.I. Bhuiyan(ed.), On-farm Reservoir Systems forRainfed Ricelands. Manila: International RiceResearch Institute (IRRI).

In this study of on-farm reservoirs (OFR)in Central Luzon, Philippines, theauthors determine that constructing anOFR provides good returns on theinvestment and can be beneficial forfarms independent of scale. The paperuses a case study from the 1985 wetseason to evaluate the benefits of an OFR.They find that the average rice yielddifferential between farms with an OFRand rainfed farms is 0.5 t/ha. Thisdifference is compounded when oneconsiders that farmers using an OFR canplant, on average, 40% of their farm areaduring the dry season and obtain averageyields of 2.3 t/ha. Additionally, manyfarmers use the OFRs to raise fish forhome consumption and sale. Whencertain maintenance and life-spancharacteristics of an OFR are accountedfor, the authors estimate the benefit-costratio of OFRs to be 5.1.

Norgaard, R.B. 1988. The biologicalcontrol of cassava mealybug in Africa.American Journal of Agricultural Economics70(2): 366–71.

Cassava has been a crucial food source inAfrica since its introduction 300 years ago. Amealybug was introduced to the continentduring the 1970s and caused large croplosses during in the 1980s. Due to the abilityof international and national programs tocollaborate and capitalize on thebiodiversity of South America, a wasp,Epidinocarsis lopezi, was found to be aparasite of the cassava mealybug. Becauseof the low-cost methods of finding,reproducing, and disseminating thisparasite, the benefit-cost ratio has beenestimated to be 149:1. Empirical evidenceshows that measures to control peststhrough biological means can lead torelatively large benefit-cost ratios.

Palanisami, K., and J.C. Flinn. 1989.Impact of varying water supply on input useand yields of tank-irrigated rice. AgriculturalWater Management 15(4): 347–59.

Palanisami and Flinn use a simultaneous-equation model to estimate the impact ofwater supply on rice yields. With data fromsouthern India, the authors estimate theimpact of five inputs—tank water, wellwater, nitrogen, labor, and cropmanagement—on rice yields during yearscharacterized by a water surplus and thosemarked by a deficit. The model developedin the study allows the authors tonumerically explain the factors which causethe difference between low- and high-yielding rice farms. During water-deficityears, the calculated yield difference was3.74 t/ha, caused principally by a lack ofwater (60%) and inadequate cropmanagement (34%). During water-surplusyears, the calculated yield difference was1.64 t/ha, caused principally by poor cropmanagement (57%).

30

Pardey, P.G., J.M. Alston, J.E. Christian,and S. Fan. 1996. Hidden Harvest: U.S.Benefits from International Research Aid.Food Policy Report. Washington, D.C.:International Food Policy Research Institute(IFPRI).

Even though the principal focus of the 16international agricultural research centers ofthe Consultative Group on InternationalAgricultural Research (CGIAR) is toalleviate hunger in developing countriesthrough increased food production, benefitsalso accrue to developed countries. Thisreport presents results of a benefit-costanalysis of the US government’s researchinvestment in two CGIAR centers, theInternational Maize and WheatImprovement Center (CIMMYT) and theInternational Rice Research Institute (IRRI).These centers conduct research on rice andwheat, important crops for the US. Theauthors estimate that the benefit-cost ratiosof US government financial investment inCIMMYT and IRRI were 190:1 and 17:1,respectively.

Paris, T.R. 1998. The impact of technologieson women in Asian rice farming. In P.L.Pingali and M. Hossain (eds.), Impact ofRice Research. Proceedings of theInternational Conference on the Impact ofRice Research, 3–5 Jun 1996, Bangkok,Thailand. Bangkok and Los Baños: ThailandDevelopment Research Institute (TDRI) andInternational Rice Research Institute (IRRI).

As farming has become more efficientthrough mechanization, discussion hasheightened regarding the benefits anddisadvantages of new technology on womenfarmers. Paris emphasizes that to study thepositive and negative consequences oftechnology, researchers must pay carefulattention to differences in socioeconomicclass among women farmers. Such groupsinclude female-headed houses, females inlandowning households, and landless

female laborers. Through an extensiveliterature review it is determined that eachclass of female farmers is affecteddifferently by new technology. This analysislooks at various forms of technology,including modern variety adoption,mechanical reapers, direct vs. hand seeding,rice mills, and herbicides to determine howdifferent classes of female farmers areaffected. As long as wages are high, landlessfemale laborers will experience numerousnegative consequences of technologybecause their efforts are replaced by moreefficient technological innovations. On theother hand, women from landowninghouseholds experience both positive andnegative consequences, in which the netresult of new technology is still notdefinitive.

Perrin, R.K., D.L. Winkelmann, E.R.Moscardi, and J.R. Anderson. 1976. FromAgronomic Data to FarmerRecommendations: An Economics TrainingManual. CIMMYT Information Bulletin 27.Mexico, D.F.: International Maize andWheat Improvement Center (CIMMYT).

This economics text provides a frameworkby which agronomists can learn to useeconomic criteria to evaluate and makefarming recommendations. The authorspresent a step-by-step process of finding asimilar set of plots for trials, estimatingcosts of capital investments, determiningbenefits and costs along with theirvariability, and ultimately calculating theexpected rate of return. The authorsemphasize that factors such as risk and lackof access to capital must be considered inmaking recommendations to farmers. If thereturn to the additional investment in atechnology is not at least 40%, then itshould not be recommended to farmers.Special emphasis is placed on evaluatingbenefits and costs of potentialrecommendations at the margin andconsidering opportunity costs. In

31

concluding, the authors present twoexamples which help to apply theprocedures of evaluating a potentialrecommendation.

Pingali, P.L., and R.V. Gerpacio. 1997.Living with reduced insecticide use fortropical rice in Asia. Food Policy 22(2):107–18.

During the years following the GreenRevolution, insecticide use dramaticallyincreased in Asia. Pingali and Gerpacioestimate that insecticide expenditures inAsia increased from US$ 347 million toalmost US$ 1.1 billion between 1980 and1990. This increase has led researchers tostudy the returns to high insecticide use anda “zero-insecticide strategy” as well aspractices that lie between these twoextremes. A study initiated in 1993 in thePhilippines showed that areas practicing azero-insecticide strategy had the highesteconomic payoff; separate studies havebegun to corroborate this finding in otherregions of Asia. In determining whichmethod of insect control to use, one mustweigh the risks of possible pest infestationsfrom a lack of insecticide against the vastarray of health, environmental, andecological risks of insecticide use. Theauthors believe that government subsidyprograms for insecticides should beremoved and replaced by taxes.

Pingali, P.L., C.B. Marquez, and F.G.Palis. 1994. Pesticides and Philippine ricefarmer health: A medical and economicanalysis. American Journal of AgriculturalEconomics 76(3): 587–92.

In a study of 152 rice farmers in thePhilippines, Pingali, Marquez, and Palisevaluate the health effects of pesticide use.Regressions are used to determine thefactors—including farmer traits such asalcohol and tobacco consumption, pesticide

use, and age—that affect health. Theauthors estimate the probabilities of havingeye, respiratory, gastrointestinal, dermal,and neurological problems under severalinput decisions. Estimates of health costsrange from 1,084 Philippine pesos for noinsecticide applications to the opposingextreme of 2,792 pesos for roughly sixinsecticide applications. When the healthimpacts of pesticide use are taken intoaccount, it is more profitable for farmers notto apply insecticide. As a result, the authorsrecommend that there should be regulationsto reduce pesticide use and correspondinglyreduce the harmful effects on farmer health.

Pinstrup-Andersen, P., N. Ruiz deLondoño, and E. Hoover. 1976. The impactof increasing food supply on human nutrition:Implications for commodity priorities inagricultural research and policy. AmericanJournal of Agricultural Economics 58(2):131–42.

In this paper the authors develop aprocedure to evaluate the relative prioritythat should be given to research on specificcommodities to achieve increased calorieand protein intake. They use data gatheredfrom an urban population in Cali,Colombia, between 1969 and 1970. Theauthors divide the sample into five incomecategories and then analyze changes innutritional intake based on a supplyincrease. They provide estimates of calorieand protein intake, percentage of the supplyincrease consumed by nutrient-deficientcategories, priority rankings for agriculturalcommodity research based on calorie andprotein objectives, and price elasticities ofselected food products for each incomecategory. The authors point out that maizeis an extremely important means to increasecalorie and protein intake because itspriority level is in the top five for the testedassumptions of supply elasticity and cost.

32

Quisumbing, A.R., L. Haddad, and C.Peña. 1995. Gender and Poverty: NewEvidence From 10 Developing Countries.Food Consumption and Nutrition Division(FCND) Discussion Paper No. 9.Washington, D.C.: International Food PolicyResearch Institute (IFPRI).

Quisumbing, Haddad, and Peña analyzethe relationship between gender andpoverty by using the stochastic dominancemethod. The 10 countries studied areBotswana, Côte d’Ivoire, Ethiopia, Ghana,Madagascar, Rwanda, Bangladesh,Indonesia, Nepal, and Honduras. Afterexplaining the method used to measuregender differences, the authors discuss thedifficulties in classifying households intomale- and female-headed households. Abrief description of stochastic dominanceprecedes the analysis. The authors findonly a few cases in which female-headedhouseholds are worse off than male-headed households, including cases fromrural Ghana and Bangladesh.

Renkow, M. 1993. Differential technologyadoption and income distribution inPakistan: Implications for research resourceallocation. American Journal of AgriculturalEconomics 75(1): 33–43.

Renkow utilizes a multimarket model totest income distribution undertechnological innovation in Pakistan. Thepaper begins with an overview of wheat inPakistan and the multimarket model. Forthe analysis of income distribution,Renkow divides the sample populationinto small and large irrigated farms,landless farmers, small and large rainfedfarms, and poor and rich urban dwellers.The model is used to identify the principalrecipients of technological innovationunder various conditions. Renkowconcludes that when wheat prices are

controlled by the government, net-producing households, which include large,rainfed farms, and small and large irrigatedfarms, are the primary recipients ofeconomic gains. When the market controlsprices, net-consuming households, whichinclude small, rainfed farms, landlessfarmers, and the urban population, stand tobenefit the most from technologicalinnovations.

Rola, A., and P. Pingali. 1993. Pesticides,rice productivity, and health impacts in thePhilippines. In P. Faeth (ed.), AgriculturalPolicy and Sustainability: Case Studies fromIndia, Chile, the Philippines, and the UnitedStates. Washington, D.C.: World ResourcesInstitute.

In this study, Rola and Pingali estimate thenet benefits of four insect control practicesin the Philippines when taking farmerhealth into consideration. The four pestmanagement practices evaluated are:(1) complete protection, (2) economicthreshold, (3) natural control, and(4) farmers’ practice. Initially, the benefits ofeach treatment are estimated under severalclimatic and market conditions. Afterexplaining health consequences of pesticideuse and comparing the number of cases ofcertain illnesses in exposed and unexposedvillages, the authors calculate the expectedutility of each practice, taking health costsinto consideration. When health costs areaccounted for, the management practiceswith the greatest to lowest net benefits arethe natural practice, farmers’ practice,economic threshold, and complete control.The authors recommend that governmentsrestrict the use of the most hazardouspesticides, thereby creating an incentive touse safer chemicals.

33

Sain, G.E., and H.J. Barreto. 1996. Theadoption of soil conservation technology inEl Salvador: Linking productivity andconservation. Journal of Soil and WaterConservation 51(4): 313–21.

Combining productivity-enhancing andsoil-conserving techniques into onerecommendation proved successful inGuaymango, El Salvador. Even though Sainand Barreto give numerous possible reasonswhy farmers in other parts of El Salvadordid not adopt the recommended method ofsoil conservation when it was proposed 15–20 years ago, they dedicate the majority ofthe paper to emphasizing reasons why alarge proportion of Guaymango farmers didadopt them. The authors explain how theformat of the program, which dividedfarmers into groups and supplied creditonly to groups whose members followedthe entire recommendation, allowed for thepackage recommendation to be widelyadopted. Empirical evidence shows that theconservation component of the packageincreased costs in the short run, but thisincrease was offset by the strong benefits ofthe productivity component.

Scobie, G.M. 1979. The demand foragricultural research: A Colombianillustration. American Journal of AgriculturalEconomics 61(3): 540–45.

Scobie uses the case of rice in Colombia toexplain seemingly contradictory efforts ofgovernment investment in agriculturalresearch and national policy to protect thecountry’s manufacturing industry. Thesepolicies economically hurt the Colombianrice industry by artificially altering theexchange rate and affecting returns toinvestment in several sectors. Scobie bringspolitical and economic issues together indiscussing funding for research and thedistribution of resulting benefits. Between1957 and 1974, rice output increasedfivefold, thus lowering the domestic priceand benefiting consumers. Between 1968

and 1974, rice exports were minimalbecause of an overvalued exchange rate. In1975, rice prices fell enough to makeexporting a viable option. This change willbe evident in the distribution of benefits, asproducers and foreign consumers reap morerewards from Colombia’s agriculturalresearch.

Scobie, G.M., and R.T. Posada. 1978. Theimpact of technical change on incomedistribution: The case of rice in Colombia.American Journal of Agricultural Economics.60(1): 85–92.

This study is one of the earliest endeavors toquantify the distribution of benefits ofsemidwarf rice varieties in Colombia. Aftera crippling virus affected rice production in1957, the Colombian Ministry ofAgriculture, the Centro Internacional deAgricultura Tropical (CIAT, theInternational Center for TropicalAgriculture), and the International RiceResearch Institute (IRRI) partnered todevelop a semidwarf variety resistant to thevirus. Instead of studying the rate of returnor efficiency of the new varieties, theauthors empirically studied the distributionof benefits or equity to producers andconsumers as well as to income levels. Thispivotal research concluded that householdsin lower income brackets received themajority of the benefits of the research.

Sidhu, D.S., and D. Byerlee. 1992.Technical Change and Wheat Productivity inthe Indian Punjab in the Post-GreenRevolution Period. CIMMYT EconomicsWorking Paper 92-02. Mexico, D.F.:International Maize and Wheat ImprovementCenter (CIMMYT).

Since the Indian Punjab accounts forroughly one-quarter of Indian wheatproduction, it is important to analyze past,present, and future possibilities of

34

productivity growth. During the GreenRevolution, wheat production increased at arate of 10% per year in the Punjab. Today,though, the increase is 4.8%. What haschanged over the past two decades? Is therea possibility of future productivity growth inthe double digits? In this paper, Sidhu andByerlee document trends in input use, realinput and output prices, and productivity.They find that while real wheat prices havefallen, so have total production costs. Inaddition, there has been a shift from usingorganic manure and animal labor tochemical fertilizers and tractors. Inconclusion, the authors explain that eventhough farmers captured the majority of theeconomic gains during the GreenRevolution, today consumers are receiving alarge share of the gains from reduced outputprices.

Smale, M. (ed.). 1998. Farmers, GeneBanks, and Crop Breeding: EconomicAnalyses of Diversity in Wheat, Maize, andRice. Boston: Kluwer Academic Publishers.

By presenting the results of initial economicinvestigations of genetic diversity in thethree major food crops (wheat, maize, andrice) in developing countries, this volumefurthers the understanding of the economiccontext in which crop breeders make use ofgenetic resources and their diversity. Itprovides an annotated catalog of the toolsused to measure and value genetic diversity.The book also explores fundamentalquestions related to the value and efficiencyof conserving seed ex situ, in gene banks.Several chapters analyze farmers’ objectivesand incentives for conserving crop geneticresources in centers of crop diversity (insitu), offering procedures for monitoring,predicting, and developing potentialmechanisms to encourage the conservationof crop genetic resources in farmers’ fields.In addition, the book exploresmethodological issues that are important forstudying the economics of crop species

diversity in farmers’ fields, and it examineshow diversity is mediated by policies andinstitutions.

Smale, M., R.P. Singh, K. Sayre, P.Pingali, S. Rajaram, and H.J. Dubin. 1998.Estimating the economic impact of breedingnonspecific resistance to leaf rust in modernbread wheats. Plant Disease 82(9): 1055–61.

Using data from the Yaqui Valley innorthwestern Mexico from 1970 to 1990, theauthors estimate the economic benefit andinternal rate of return on the investment inbreeding for nonspecific resistance to leafrust in modern bread wheats. Afterexplaining the importance of maintenancebreeding, the authors quantify yield lossesthat would have occurred if the varietieshad not been bred for nonspecific leaf rustresistance. In conducting the analysis,emphasis is placed on the benefits ofnonspecific resistance over specific leaf rustresistance. Using costs of the InternationalMaize and Wheat Improvement Center’s(CIMMYT) pathology program, andbenefits to the Yaqui Valley between 1970and 1990 of US$ 17 million, the authorsestimate rates of return under severalassumptions. The internal rate of return oninvestment is 13% if a ten-year research lagis present and almost 40% if a five-year lagis present.

Traxler, G., and D. Byerlee. 1992.Economic returns to crop managementresearch in a post-Green Revolution setting.American Journal of Agricultural Economics74(3): 573–82.

Despite numerous endeavors to conductstudies of the impact of agriculturalresearch, crop management research (CMR)has not received adequate attention. Traxlerand Byerlee emphasize the need to evaluatean entire CMR portfolio instead of

35

individual projects, because the evaluationof returns to individual projects can inflatethe benefits. Whereas the benefits ofbreeding and actual manufacturing ofinputs can be quantified relatively easily, theoutputs of public sector research on cropmanagement consist largely of informationand knowledge, and their impact is muchmore difficult to evaluate. From a case studyof the Yaqui Valley in northwestern Mexicofrom 1977 to 1988, the authors find that onlytwo of nine practices developed by the CMRprogram were adopted by farmers based onrecommendations of the local researchagency. Thus, despite individual projectreturns of 50% and 100%, the return on theentire CMR portfolio in this study was 16%.

Traxler, G., and D. Byerlee. 1993. A joint-product analysis of the adoption of moderncereal varieties in developing countries.American Journal of Agricultural Economics75(4): 981–89.

Although grain yield is the principalcriterion for evaluating the success of cerealvarieties, Traxler and Byerlee evaluate thebenefits of joint-product profit maximizingin cereal production. Modern varieties(MVs) are semidwarf and therefore havesmall straw yields. In joint-product profitmaximization, the producer evaluates therelative prices of grain and straw indetermining the payoff matrix to MVs andtraditional varieties (TVs). This paperpresents a model for evaluating joint-product profit maximization under severalclimatic conditions, during different periods(the pre-Green Revolution, GreenRevolution, and post-Green Revolutioneras), and under varying relative input andoutput prices. The authors conclude thatMV adoption is slowest in areas with lowrainfall where the grain yield differential isnot as dramatic between TVs and MVs andthe price of fodder is relatively higher.

Unnevehr, L.J. 1986. Consumer demandfor rice grain quality and returns to researchfor quality improvement in Southeast Asia.American Journal of Agricultural Economics68(3): 634–41.

As the benefits of the Green Revolution areseen through large supplies of rice,researchers are recognizing the increasingneed to invest in quality-improvementresearch. Unnevehr suggests that previousresearch on physical-quality improvementwill now be followed by research onchemical-quality improvement. Theunderinvestment in quality-improvementresearch can be compensated for by aconcerted, collaborative effort betweeninternational and national researchprograms. The author uses the “consumergoods characteristics model” developed byLadd and Suvannunt to estimate implicitprices of specific rice qualities. Rice marketsin three countries, Thailand, Indonesia, andthe Philippines, are studied to quantifyregional estimates of returns to research.Unnevehr concludes that the demand formilling quality is relatively constantbetween countries, whereas the demand forchemical characteristics can varydramatically.

von Braun, J., and P.J.R. Webb. 1989.The impact of new crop technology on theagricultural division of labor in a WestAfrican setting. Economic Development andCultural Change 37(3): 513–34.

Von Braun and Webb analyze the effect of acentralized-pump irrigation system for1,500 hectares in central Gambia on genderroles in agriculture. With theimplementation of this new technology in1983, rice, once a “woman’s crop,” becameprimarily a “male-controlled crop.” Thisshift principally occurred because as riceyields increase, rice is no longer grown onindividual plots by women, but rather oncommunal lands under the control of men.

36

It is this division of labor—betweenindividual and communal plots—andtechnological change that affect genderroles in rice production.

Walker, T.S. 1989. High-yielding varietiesand variability in sorghum and pearl milletproduction in India. In J.R. Anderson andP.B.R. Hazell (eds.), Variability in GrainYields: Implications for Agricultural Researchand Policy in Developing Countries.Baltimore: Johns Hopkins.

Sources of production variability havebecome critical areas of study as researcherstry to ascertain why variability around theworld has been rising. Walker uses datafrom 48 sorghum and 40 pearl milletproducing regions in India for two 12-yearperiods, one before the Green Revolutionand one after the Green Revolution, todetermine factors contributing toproduction variability. Initially, it isexplained that the vast majority ofvariability is derived from the covariancebetween regions. Walker uses a weightedleast squares regression to determine theeffects of high-yielding variety adoption,irrigated area, and variation in rainfallcovariance over the two time periods.Higher adoption of both sorghum and pearlmillet has increased yield covariancebetween regions, a higher rainfallcovariance has contributed to higherproduction covariance, and irrigationchanges have contributed to highercovariance for sorghum and lower yieldcovariance for pearl millet.

Walker, T.S., and C.C. Crissman. 1996.Case Studies of the Economic Impact ofCIP-Related Technologies. Lima, Peru:International Potato Center (CIP).

Recent emphasis has been placed ondemonstrating practical applications of newtechnologies. This collection of nine case

studies describes the impact of researchfrom the International Potato Center (CIP).Walker and Crissman have put together acollection of three case studies on varietal,integrated pest management (IPM), andseed system technology. The studies presentcost-benefit analyses of adoption in Easternand Central Africa, China, Peru, Tunisia, theDominican Republic, Vietnam, and India.Topics include CIP-24 and Canchán-INIAAvarieties, potato tuber moth, sweetpotatoweevil, and the Andean potato weevil inIPM research, and various seed projects.

Winkelmann, D. 1976. The Adoption ofNew Maize Technology in Plan Puebla,Mexico. Mexico, D.F.: International Maizeand Wheat Improvement Center (CIMMYT).

This study describes the results of a plan todevelop, test, and disseminaterecommendations to increase maize yieldsin Mexico. Winkelmann discusses theobjectives and organization of the initiative,as well as factors limiting adoption of therecommendations. In general, therecommendation of increased nitrogen,phosphorus, and planting densitydramatically increased yields in the testarea. The point of contention, though, wasthe different ways of measuring adoptionrates. Even though adoption of the entirerecommendation may be relatively low, thepercentage of farmers who implementedcomponents of the new technology is muchhigher. Winkelmann emphasizes that small-scale farmers will modify their farmingpractices in anticipation of higher profits,but only to the extent to which they canbear the additional risk. Since implementingthe entire recommendation requiresnumerous increases in inputs, to reducerisks many farmers settle for using a level ofinputs intermediate between their originalmethod and the recommendation. Theauthor explains that the rate of return toinvestment in “intermediate” input researchwill be higher than intensive input research.

37

Thus collaborative efforts of research andextension institutions in making anddisseminating recommendations have beenmuch more beneficial for farmers thanestimates of the adoption of the entirerecommendation would illustrate.

Witcombe, J.R. 1989. Variability in theyield of pearl millet varieties and hybrids inIndia and Pakistan. In J.R. Anderson andP.B.R. Hazell (eds.), Variability in GrainYields: Implications for Agricultural Researchand Policy in Developing Countries.Baltimore: Johns Hopkins.

Witcombe uses data on varieties andhybrids in India and Pakistan from theInternational Crops Research Institute forthe Semi-Arid Tropics (ICRISAT) todetermine the comparative stability andyields of varieties and hybrids. Individual

data are compared with populationaggregates for low-, average-, and high-yielding environments through a regressionanalysis. It is determined that hybrids havehigher yields, but varieties are more stable.Mean variance and stochastic dominanceanalyses are conducted to determine thetradeoff between stability and variance inchoosing suitable varieties or hybrids. Meanvariance analysis revealed that for allenvironments studied the variety or hybridto be chosen was the highest yielding of theoptions. The stochastic dominance methodshowed hybrids to be better than varietiesfor low-yielding environments and overmost yield values.

ISBN: 970-648-076-5

SPIA TAC CGIAR

INTERNATIONAL MAIZE AND WHEAT IMPROVEMENT CENTER

WWW.CIMMYT.ORG