success in high value horticultural export markets for the small farmers: the case of mahagrapes in...

17
Success in High Value Horticultural Export Markets for the Small Farmers: The Case of Mahagrapes in India DEVESH ROY International Food Policy Research Institute (IFPRI), Washington, DC, USA and AMIT THORAT * Jawaharlal Nehru University, New Delhi, India Summary. In spite of being the second largest horticultural producer in the world, India is a failed exporter mainly because of the inability to meet the food safety standards. Hence, successes in horticultural exports are rare. Here, we study one unique success story, Mahagrapes, a marketing partner to farmer cooperatives attributing its success to a combination of collective action and pub- lic private partnerships. Our results indicate that Mahagrapes farmers earn significantly higher in- come vis-a ` -vis their outside marketing option and smallholders face no bias in selection. Together with the farmer’s ability to consistently meet standards, this implies that the model can be scaled up. Ó 2008 Elsevier Ltd. All rights reserved. Key words — standards, food safety, exports 1. INTRODUCTION Since 1971, global agricultural exports have grown at 3.0% per year in real terms, on aver- age, while agricultural production has grown at 0.7% per year. As a result, the share of agri- cultural production that is exported has dou- bled, rising from 19% in 1971 to 40% in 2003. The composition of agricultural exports has changed as well. The grain exports have fallen sharply as a proportion of the total value of agricultural trade, declining from 15% in the 1960s to less than 8% in the 1990s. At the same time, exports of higher-value agricultural prod- ucts such as fruits and vegetables, meat, dairy products, eggs, and fish and seafood have grown from 29% to 42% of the total over the same period. Concomitantly, the share of pro- cessed products in agricultural exports has in- creased from 41% in the 1980s to 51% in the 1990s. The growth in the share of agricultural exports that are processed can be seen in almost all commodity categories (FAO, 2005). The exceptions to this pattern are fruits and vegeta- bles: both primary and processed fruits and vegetables have increased their share of world agricultural exports over this period, but growth in fresh fruit and vegetable exports has been greater. This reflects both consumer willingness to pay for fresh produce and improvements in transportation and logistics * Our heartfelt gratitude goes to the Mahagrapes and independent farmers who shared the information dili- gently on a wide range of questions. Thanks also to the people at Mahagrapes especially Mrs. Bhushana Karan- dikar who accompanied us on our field visits and made it an enriching experience for us. We also acknowledge the contribution of the participants at the Plate to Plough conference in New Delhi in September 2006 and seminar participants at the Markets Trade and Institutions Di- vision at the International Food Policy Research Insti- tute. Final revision accepted: September 6, 2007. World Development Vol. xx, No. x, pp. xxx–xxx, 2008 Ó 2008 Elsevier Ltd. All rights reserved 0305-750X/$ - see front matter doi:10.1016/j.worlddev.2007.09.009 www.elsevier.com/locate/worlddev 1 Please cite this article in press as: Roy, D., & Thorat, A., Success in High Value Horticultural Export ..., World Development (2008), doi:10.1016/j.worlddev.2007.09.009 ARTICLE IN PRESS

Upload: independent

Post on 21-Nov-2023

0 views

Category:

Documents


0 download

TRANSCRIPT

World Development Vol. xx, No. x, pp. xxx–xxx, 2008� 2008 Elsevier Ltd. All rights reserved

0305-750X/$ - see front matter

ARTICLE IN PRESS

doi:10.1016/j.worlddev.2007.09.009www.elsevier.com/locate/worlddev

Success in High Value Horticultural Export

Markets for the Small Farmers: The Case

of Mahagrapes in India

DEVESH ROYInternational Food Policy Research Institute (IFPRI), Washington, DC, USA

and

AMIT THORAT *

Jawaharlal Nehru University, New Delhi, India

Please..., W

Summary. — In spite of being the second largest horticultural producer in the world, India is afailed exporter mainly because of the inability to meet the food safety standards. Hence, successesin horticultural exports are rare. Here, we study one unique success story, Mahagrapes, a marketingpartner to farmer cooperatives attributing its success to a combination of collective action and pub-lic private partnerships. Our results indicate that Mahagrapes farmers earn significantly higher in-come vis-a-vis their outside marketing option and smallholders face no bias in selection. Togetherwith the farmer’s ability to consistently meet standards, this implies that the model can be scaledup.� 2008 Elsevier Ltd. All rights reserved.

Key words — standards, food safety, exports

* Our heartfelt gratitude goes to the Mahagrapes and

independent farmers who shared the information dili-

gently on a wide range of questions. Thanks also to the

people at Mahagrapes especially Mrs. Bhushana Karan-

dikar who accompanied us on our field visits and made it

an enriching experience for us. We also acknowledge the

contribution of the participants at the Plate to Plough

conference in New Delhi in September 2006 and seminar

participants at the Markets Trade and Institutions Di-

vision at the International Food Policy Research Insti-

tute. Final revision accepted: September 6, 2007.

1. INTRODUCTION

Since 1971, global agricultural exports havegrown at 3.0% per year in real terms, on aver-age, while agricultural production has grownat 0.7% per year. As a result, the share of agri-cultural production that is exported has dou-bled, rising from 19% in 1971 to 40% in 2003.The composition of agricultural exports haschanged as well. The grain exports have fallensharply as a proportion of the total value ofagricultural trade, declining from 15% in the1960s to less than 8% in the 1990s. At the sametime, exports of higher-value agricultural prod-ucts such as fruits and vegetables, meat, dairyproducts, eggs, and fish and seafood havegrown from 29% to 42% of the total over thesame period. Concomitantly, the share of pro-cessed products in agricultural exports has in-creased from 41% in the 1980s to 51% in the1990s. The growth in the share of agriculturalexports that are processed can be seen in almost

1

cite this article in press as: Roy, D., & Thoratorld Development (2008), doi:10.1016/j.worldd

all commodity categories (FAO, 2005). Theexceptions to this pattern are fruits and vegeta-bles: both primary and processed fruits andvegetables have increased their share of worldagricultural exports over this period, butgrowth in fresh fruit and vegetable exportshas been greater. This reflects both consumerwillingness to pay for fresh produce andimprovements in transportation and logistics

, A., Success in High Value Horticultural Exportev.2007.09.009

2 WORLD DEVELOPMENT

ARTICLE IN PRESS

that has made it possible to transport perish-ables over long distances.

The growth in fresh fruits exports offers a win-dow of opportunity to those developing coun-tries that have a low processing capacity buthave high labor and suitable agro-climaticendowments. During 1971–2002, the largestnet exports from developing countries werefruits and vegetables (FAO, 2005). At the sametime, there exists differentiation across develop-ing countries in their ability to benefit from thismarket opportunity mainly because of the dif-ferences in capacity for meeting internationalfood safety standards (Athukorala & Jay-asuriya, 2003). The ability to meet high foodsafety and quality standards is lowest in small-holder agriculture because of the scale econo-mies that standards lead to. India is one suchcountry dominated by smallholders (with 82%of landholdings less than 2 ha in 2001) but thepolicy debate on whether or not smallholderscan access high value export markets encom-passes the whole gamut of developing countries.

India is one of the leading producers of fruitsand vegetables in the world. In 2005, India con-tributed to about 9% of world fruit productionand was the third biggest producer of fruits afterChina and Europe, producing nearly 50 milliontons of produce (FAO, 2005). Similarly, Indiaalso contributes to about 14% of world vegetableproduction. In 2005, it was the second largestvegetable producer after China, at around 35million tons. 1 At the same time India is a rela-tively small horticultural exporter. Accordingto the 2005 World Bank report on Indian horti-culture, India is a competitive producer of horti-cultural products on farm but owing to off-farmdisadvantages, she fails to compete abroad. In-dia’s share in world horticultural trade was just2.3% in 2004, a figure disproportionate to itsshare in global production. 2 Diversifying intohorticulture is often seen as an opportunity foruplifting incomes for the small farmers. Addi-tionally, if rich country markets are accessibleto farmers, it would result in a significant in-crease in farm incomes.

What are the constraints that make India anunsuccessful exporter of horticultural products?The most important reason is the inability ofthe smallholder dominated production systemsto meet the food safety and quality require-ments of the rich country markets. The small-holder-based agriculture in India imposeslimits on the number of farmers capable toadopt more sophisticated farm practices andundertake the necessary investments (e.g., land

Please cite this article in press as: Roy, D., & Thorat..., World Development (2008), doi:10.1016/j.worldd

improvements, obtaining necessary certifica-tions, and cold storage) to meet more stringentfood quality and safety requirements (Umali-Deininger & Sur, 2006).

The western food standards such as Euro-Re-tailer Produce Working Group Good Agricul-tural Practices (EurepGAP) have often turnedout to be too stringent for Indian producersto remain competitive given their cost of com-pliance. The inability to meet these stringentstandards results from several factors, likeunderdeveloped physical infrastructure, suchas limited supply of power and irrigation, lim-ited number and narrower spread of pre cool-ers, cold storages and refrigerated transportvehicles and accredited certification agencies.Most importantly, these standards require ahigh degree of information on requirements re-lated to inputs, packaging, etc., that is costly toobtain and process.

Also, the standards keep changing and theproducers have to be abreast with that informa-tion. Acquiring information from buyers, get-ting access to post harvest infrastructure orgetting certified for meeting the standards areall subject to scale economies. 3 The small farm-ers typically face scale disadvantages and arethus unable to meet the rich country standards. 4

Often, the product attributes are imperfectlyobserved especially if the goods have the prop-erty of credence. In such situations, the signifi-cance of reputation goes up. Without perfectobservability, expected quality from an individ-ual seller is determined by the average. SinceIndian horticultural producers are largely per-ceived to be inept at meeting the standards,negotiating and initiating marketing contractsin developed country markets can be difficultfor the individual producers owing to the aver-age reputation effect.

India’s reputation in complying with SPSstandards in richer markets is dismal. ConsiderIndia’s exports to the United States. DuringMay 1999–April 2000, the total number ofdetentions by the US originating from all (52)countries was 9875, of which 860 shipments orig-inated from India. This was the highest numberof shipments rejected by United States Food andDrug Administration (USFDA) that originatedfrom a single country (Mehta & George, 2003).Recently, during February 2006–January 2007,the number of detentions was 16,818. Indiaagain faced the highest number of detentionswith a total of 2,085 detentions (FDA, 2007).

What are the institutional solutions that canbe proposed to overcome these problems? To

, A., Success in High Value Horticultural Exportev.2007.09.009

SUCCESS IN HIGH VALUE HORTICULTURAL EXPORT MARKETS 3

ARTICLE IN PRESS

overcome scale diseconomies for the small pro-ducers or to exploit the scale economies, themost common solution that is proposed is thecreation of producer organizations. 5 Europeis a good example of this, with substantial shareof national demand for fruits and vegetablesbeing supplied by the farmer cooperatives.Dairy in many countries is dominated by coop-eratives including in India. In general, fruitsand vegetables have a much smaller share ofmarkets with the cooperatives in many coun-tries. In the United States, in 1997 where theshare of cooperatives in dairy was 83%, in fruitsand vegetables it was merely 19%. In 1997,across developed countries in Europe, the high-est market share of cooperatives in fruits andvegetables was in Belgium (greater than 70%)(Bekkum & Dijk, 1997).

Producer cooperatives can exploit economiesof scale by collectivization of output of thefarmers. Economies of scale in productionare, however, different from economies of scalein marketing that are dominant in high valueexport markets. The requirements of marketingin terms of information procurement and dis-semination, the role of reputation and brandingand the importance of negotiations (in initiat-ing and maintaining contracts) lead to differentkinds of scale economies.

Here, we study the case of Mahagrapes, a mar-keting partner to a group of producer coopera-tives in Maharashtra state of India. Amarketing cooperative was first described byHelemberger and Hoos (1962) as an organiza-tion that has the objective of maximizing theprice paid to the members of the marketed prod-uct. Mahagrapes is not a marketing cooperativebut a marketing partner to the cooperatives. Itnot only negotiates better prices for its membersbut also provides technical assistance, inputsand information to the farmers to enable themin meeting the stringent requirements of the ex-port markets. The role of Mahagrapes is broadlytwofold. First, as a facilitator it provides market-ing expertise (in both inputs and output markets)which is lacking among the producer coopera-tives. In providing information, negotiating con-tracts and supplying certification, Mahagrapesovercomes the diseconomies of scale of smallfarmers through collectivization. As a providerof technical assistance and inputs, in its secondrole Mahagrapes again overcomes the scale dis-economies through bulk buying or through in-house production.

As a success story of smallholders in the highstandard high return export markets, the issue

Please cite this article in press as: Roy, D., & Thorat..., World Development (2008), doi:10.1016/j.worldd

of interest to the development practitioners iswhether this kind of arrangement can be scaledup or replicated. This is the issue we address inthis paper. As discussed below, policy initiativesof various types (public and public private) havebeen involved in the success of Mahagrapes.However, in spite of the political and institu-tional support, for the model to be scaled up,three conditions must be satisfied: (i) consistentability of the smallholders to meet the standardsrequired in the developed country markets, (ii)absence of any bias in the selection/participa-tion of the farmers in terms of commonly foundattributes in the population (such as small landsize), and (iii) conditional on participation andcompliance with the standards, there should besignificant gains to the farmers.

Several papers have looked at the issue ofparticipation of small farmers in export ori-ented agriculture. Here, we approach the issueof selection more broadly. Linking with largefarmers is likely to involve lower transactioncosts but catering to high standard markets im-plies that different characteristics of the farmerscould be important such as education and expe-rience in farming. Thus, we posit the selection/participation question as one of negative versuspositive selection based on several characteris-tics. We find that it is not the land size but hu-man capital attributes of the farmers thatsignificantly determine participation.

Conditional on participation, do the farmerslinked with Mahagrapes earn significantly high-er profits relative to independent farmers aftercontrolling for farmer characteristics? The netreturns to the farmers has to take into accountthe incremental costs of standardizing the prod-uct for western markets with the use of im-proved inputs in production (such as the useof bio-pesticides and fertilizers) and in market-ing (the use of better packaging materials). Ourresults indicate that after controlling for bothobserved and unobserved characteristics ofthe farmers, the net returns to farmers per acreof cultivated land are significantly higher.

Though there exists a large body of literaturethat estimates the gains to the farmers fromparticipation in high value markets, most stud-ies do not control for the potential selectionbias that could affect the results (Warning &Key, 2002 being an exception). Several studiessimply compare the gross margins per hectare.von Braun and Immink (1994) found that ex-port horticultural production in Guatemalagenerated gross margins per hectare 15 timesas large as maize production and gross margins

, A., Success in High Value Horticultural Exportev.2007.09.009

4 WORLD DEVELOPMENT

ARTICLE IN PRESS

per labor day twice as large. McCulloch andOta (2002) show that horticultural exportgrowers have higher incomes after controllingfor farm size, education, irrigation, and otherfactors.

None of these studies, however, takes into ac-count the fact that those who participate in highvalue markets (for exports or otherwise) makerelatively higher gains simply because they havebetter skills or face better conditions some ofwhich are unobserved and hence not capturedin the data. The problem of non-random selec-tion creates an upward/downward bias in theestimates of the benefits of linking up. We con-trol for the endogeneity bias using instrumentalvariables technique. The result that Mahagra-pes farmers earn significantly higher profits isrobust to controlling for selection bias.

We further find that the impact of member-ship in Mahagrapes varies between small andlarge farmers. The impact on small farmers inmodern marketing channels like supermarketsor in export horticulture has been extensivelystudied (e.g., Birthal, Joshi, & Gulati, 2005;Faiguenbaum, Berdegue, & Reardon, 2002;Farina, 2002; Ghezan, Mateos, & Viteri, 2002;Reardon & Berdegue, 2002). This strand of lit-erature, however, has largely been descriptive.We find that controlling for other farm charac-teristics, the marginal impact on the smallfarmers from participation is higher.

The paper is divided into the following sec-tions. Section 2 gives background informationon Mahagrapes including details on its forma-tion and the organizational and institutionalstructure. Section 3 highlights the features ofMahagrapes especially with regard to the en-abling role it plays in farmers’ ability to meetthe requirements of the export markets. Section4 discusses the data and the descriptive statis-tics from the primary survey. Section 5 presentsthe regression results. Section 6 addresses theissue of selection/participation. Section 7 con-cludes.

2. BACKGROUND INFORMATION ONMAHAGRAPES AND ITS ROLE AS AFACILITATOR TO SMALL FARMERS

Mahagrapes came into existence in 1991.Several government agencies supported itsestablishment including some federal agencies(like National Cooperative Development Cor-poration (NCDC), National HorticulturalBoard (NHB), Agricultural and Processed

Please cite this article in press as: Roy, D., & Thorat..., World Development (2008), doi:10.1016/j.worldd

Food Products Export Development Authority(APEDA)). Alongside there were state agenciesthat helped in establishing Mahagrapes (e.g.,Department of cooperation, Government ofMaharashtra and Maharashtra State Agricul-ture Marketing Board (MSAMB)).

The creation of Mahagrapes is unique in theIndian context as it is the first to make use of aspecial provision (following an amendment)under the cooperative laws of the state. Thisamendment in 1984, allowed cooperatives toassociate with other agencies (including mar-keting partners). Such an association of cooper-atives with other agencies was forbidden priorto the amendment. Mahagrapes is registeredas a marketing partner to the producer cooper-atives under the clause following this amend-ment in the cooperative law of the state.

The organization of Mahagrapes is as fol-lows. At the apex are two executive partners.This is followed by an executive council com-prising seven elected cooperative heads, fol-lowed by a board of directors composed ofthe heads of the 16 cooperatives. The executivecouncil and the board of directors are responsi-ble for transmitting the views of the farmers.All decisions finalized by the executive partnersare taken after achieving a consensus of theexecutive council. The members of the execu-tive council in turn being the representative offarmer cooperatives, voice the opinion of thefarmers. The executive council, however, alsohas a discretionary power to make emergencydecisions including financial decisions up tothe tune of 40 million rupees.

Mahagrapes is a for profit organization. Theprice which each farmer gets is decided basedon the quality of the grape that can be ascer-tained due to traceability of the produce. Sowhen the price is decided for grapes from a par-ticular farmer, Mahagrapes deducts a facilita-tion fee per unit of output and passes on therest to the cooperatives. The cooperative thenkeeps a part of the price for itself and passeson the rest to the farmer. The amount whichthe cooperative keeps varies from one coopera-tive to another depending upon their specificrequirements.

The primary source of funding of Mahagrapesis membership equity. The government does notprovide any explicit assistance to Mahagrapes inits functioning. The only government assistancethat Mahagrapes receives is the general assis-tance provided to all horticultural and processedfood exporters in the country in the form oftransport assistance (from APEDA).

, A., Success in High Value Horticultural Exportev.2007.09.009

SUCCESS IN HIGH VALUE HORTICULTURAL EXPORT MARKETS 5

ARTICLE IN PRESS

Since Mahagrapes works as a facilitator, animportant source of current financing forMahagrapes is the facilitation charge. Addi-tionally, the transfer of inputs by Mahagrapesto the cooperatives is a commercial transactionthough the inputs are sold at little or no profit.Thus, the inputs transacted between the coop-eratives and Mahagrapes do not involve anysubsidy.

In the beginning, Mahagrapes had 29 grapegrowing farmer cooperatives as its members.This number came down immediately withinthree years and currently there are 16 farmercooperatives. The initial time periods werecharacterized by difficulty for Mahagrapesresulting from high rates of consignment rejec-tions in the European markets. In the secondyear after exports started in 1991, a large con-signment was rejected and there were losses tothe tune of 20 million rupees. As a result manycooperatives left Mahagrapes to concentrate onthe domestic market or to look for alternatemarketing arrangements thus bringing a reduc-tion in the number of cooperatives that contin-ued to link up with Mahagrapes.

Though not as acute as in initial time periods,Mahagrapes has had to deal with the issues ofconsignment rejection on an annual basis fora substantial period of time. 6, 7 As the fruitquality and SPS measures as well the methodsused to ascertain these fruit characteristicschange year to year, Mahagrapes has to keepabreast with them and amend their own pro-duction, processing, storing and testing meth-ods regularly. For instance, consignmentrejection occurred from the Netherlands mar-ket, as the method employed in India for testingchemical residues was different. 8

Other product attributes which could be thebasis for rejection are: berry size, fruit color,bunch weight, blemish, bag weight (mini-mum–maximum), stem color, berry shrived,split berry, SO2 damage, waste berry, pest dam-age, shatter berry, chill damage, temperature,residue, taints and odor, packing quality andaverage check weight. Over the years, Mahag-rapes has learnt how to minimize these poten-tial problems and rejection rates have gonedown substantially. 9 In the initial periodswhen the consignment rejections occurred, thegovernment stepped in with soft loans to thecooperatives to help set up pre-cooling and coldstorage infrastructure in the cooperatives.

The institutional arrangement linking smallfarmers to export markets through Mahagrapescan be broadly summarized as a mix of collec-

Please cite this article in press as: Roy, D., & Thorat..., World Development (2008), doi:10.1016/j.worldd

tive action and public private partnerships.Conceptually, the role of collective action ariseswherever there are economies of scale in pro-duction or marketing. This includes the roleof farmer groups in being better able to ensuretraceability. In these chains, the costs for theestablishment of traceability are lower for firmsand farms with collective action than without.Similarly, collective action has a rationale ifagents in the supply chain have different com-parative advantages. Thus, a producer group(with comparative advantage in production)could benefit from collective action with agentsthat have expertise in marketing.

Additionally, public private partnerships(PPPs) might be required to play a complemen-tary role in linking small farmers with HVAmarkets. The traditional public supply func-tions can be inadequate to meet the needs ofthe HVA supply chains from the perspectiveof the smallholders. Traditional public sectoractivities such as extension, research and devel-opment, and price and marketing policies havebeen largely commodity-based and hence maynot provide the support smallholders requirein a HVA supply chain.

Mahagrapes is a good example of public–pri-vate partnership. 10 Ownership of Mahagrapeslies solely in the hands of the farmers; as theyhave collectively contributed their share in thefixed and operating costs of Mahagrapes andthey also handle the governance of the firm.However, as discussed above, in the initial peri-ods (including in the initiative to establish), therole of public institutions such as MSAMB wasessential. MSAMB deputed and paid the sala-ries of the first governing officers of Mahagra-pes for three years who were brought in fromother state departments. MSAMB also pro-vided for consultancy services from experts onagri-marketing, packaging, technical servicessuch as refrigeration and cooling. All the liai-soning with institutions such as the CentralFood Technology Research Institute (CFTRI)was done by Mahagrapes. The National Coop-erative Development Commission (NCDC)gave loans to the societies for pre-cooling andpack-houses. Loans from the private sector thathas a higher discount rate would not be forth-coming (as success in export markets withsmallholders usually takes time).

With assistance from a spectrum of govern-ment bodies, it is important that the governmentassumed the role of a mere facilitator. In con-trast to the system of other cooperatives in India(e.g., in dairy and sugar), the government was

, A., Success in High Value Horticultural Exportev.2007.09.009

6 WORLD DEVELOPMENT

ARTICLE IN PRESS

not assigned any direct role in the decision mak-ing of cooperatives or in Mahagrapes. In manydairy cooperatives, for example, elite captureand rent seeking by the top management has be-come quite common in India. Similarly, Banner-jee et al. (2001) present evidence for rent seekingand elite capture by the management of sugarcooperatives in Maharashtra.

What has Mahagrapes done to facilitatefarmers in meeting the export standards? As amarketing partner of the producer cooperativesfacilitating exports, Mahagrapes has enabledthe farmers in several ways: Kleinwechter andGrethe (2005) while analyzing the adoption ofEurepGAP standards by mango exporters inPeru categorize the compliance process intothree stages viz. the information stage followedby a decision stage and finally an implementa-tion stage. Mahagrapes has been active in allthe three stages of the compliance process.

In course of facilitation Mahagrapes solvesthe diseconomies of scale problem in marketingin several ways. In the input markets and in theoutput markets, procurement of informationand negotiation of contracts involve economiesof scale. Mahagrapes undertakes this activityon behalf of the farmers. Mahagrapes also pur-chases inputs in bulk (like bio-fertilizers fromAustralia) and to the extent that it collectivizesthe input demand of all the members, it benefitsfrom economies of scale more than what coop-eratives would enjoy separately.

Related to compliance, Kleinwechter andGrethe (2005) point out that certain knowledgeof the standard is necessary to make a decision.Creating a knowledge base about standardsand about export markets in general involveslarge fixed costs (e.g., foreign travel) and infor-mation once procured can be disseminated atlow marginal cost. Small farmers would findit prohibitive to incur such fixed costs and thusbenefit from Mahagrapes procuring this infor-mation on their behalf.

Once the information on the market require-ments is available with Mahagrapes, it dissem-inates this information to the farmers free ofcost. This follows the long tradition of holdingquality and product safety related workshopsand field demonstrations which the grape grow-ers association in Maharashtra had been con-ducting from before. Farmers and grapehandlers/sorters (primarily women) are contin-uously informed about and trained in the latestgrape growing and handling methods and pro-cesses, as well as the latest weather and climateupdates.

Please cite this article in press as: Roy, D., & Thorat..., World Development (2008), doi:10.1016/j.worldd

Mahagrapes continuously updates the list ofbanned and approved pesticides and fertilizers,which keep changing from year to year acrossmarkets. Similarly, the changes in the permissi-ble levels of chemical residues are also providedby them regularly. All this information is pub-lished in the form of a yearly hand book in thenative language and distributed free of chargeamong the farmers.

Once the information on the standards isavailable, action is needed relating to the deci-sion on implementation. What are the stepsthat need to be taken, if there are restrictionson inputs, and how and where from can theybe procured are some of the decisions thatemerge regarding implementation of the stan-dards in the next stage. The Mahagrapes deci-sion to produce organic pesticides itself wouldfall under its actions in the decision stage.

In the implementation stage, Mahagrapesprovides materials and technical help alongwith infrastructural support to facilitate theimplementation of the standards. Mahagrapes,for instance, provides the farmers with packag-ing material which comply with internationalnorms. Plastic bags and palettes are importedfrom Spain. Special SO2 sheets (sulfur di oxidesheets) are imported from China which are usedto cover the grapes before being sealed in corru-gated cardboard boxes. This releases the SO2

gas right after 15 days when the consignmentof grapes arrives at the destination, so as to re-strict the spread of bruising or any other dam-age sustained by the grapes in transit.

Regular and constant monitoring of thegrape plant by the farmers themselves is facili-tated by the scientists from the National Re-search Centre (NRC) in Pune. This ensuresthat the plant remain healthy throughout theyear and not just in the fruiting season. Bio-fer-tilizer and bio-pesticide are developed and pro-duced by Mahagrapes and provided to memberfarmers cheap, that helps them meet the Eurep-GAP requirements. 11

Acquiring a EurepGAP certificate individu-ally is costly for the small and medium farmers.However, Mahagrapes has managed to providethe entire cooperative societies with certifica-tion. Thus, each society gets certified as Eurep-GAP compliant along with the member farmerswho now have to pay just Rs. 1,200 per annum(approximately $28). In our sample, for thesmallest exporter also, this expenditure is lessthan 2% of the annual export revenues. Notethat the certification is marketing relationshipbased and holds for cooperatives/farmers only

, A., Success in High Value Horticultural Exportev.2007.09.009

SUCCESS IN HIGH VALUE HORTICULTURAL EXPORT MARKETS 7

ARTICLE IN PRESS

as long as they are linked to Mahagrapes astheir marketing partner.

With the infrastructure and information dis-semination system set up that has been criticalin enabling the farmers to meet the high stan-dards, the issue remains with regard to the real-ized gains from accessing the export markets.Since rejection rates have gone down substan-tially and have stabilized, risks have been re-duced significantly. Yet, it is possible thatcompliance with standards erodes margins ofprofitability substantially enough, so as to real-ize only insignificant gains. In order to addressthe issue of potential for scaling up, we nowanalyze the data from the primary surveyinvolving Mahagrapes and independent farm-ers and subject it to empirical investigationfor the issue of selection and significance ofthe realized gains.

Table 1. Access to infrastructure at the village level (%)

Independent Mahagrapes

Average distance ofvillage to urban area(km)*

13.9 16.1

Does the village havea cooperative creditsociety (%)

96.9 98.9

Does the village havea regional rural bank(%)**

80.4 89

Does the village havea land line phonefacility within adistance of 5 km (%)

100 100

Does the village havea post office (%)

94 93

Does the village haveaccess to a bus stopwithin the distance of5 km (%)

100 100

Does the village haveaccess to a coldstorage facility (%)

61 60

Does the village havean agricultural mandiin the vicinity (within5 kms)

76 76

Source: Primary survey; *, 10% level of significance; **,5% level of significance; ***, 1% level of significance fortest in difference of means between Mahagrapes andindependent farmers.

3. DATA AND DESCRIPTIVE STATISTICS

The data were collected from a primary fieldsurvey (involving questionnaire based-inter-views) in the districts of Pune, Nasik and San-gli, comprising two types of grape farmers;those who are members of the Mahagrapesand those who are independent, selling mostlyin the local markets. Out of a total of 183 sur-veyed farmers, the number of farmers associ-ated with Mahagrapes is 88 and ofindependent farmers is 95. The survey wasundertaken in 2005. The collected informationpertains to the year 2004. The Mahagrapesfarmers were randomly chosen from a list offarmers provided by the cooperative while theindependent farmers were randomly selectedfrom the list of registered grape farmers in thesame area (registered with grape growers asso-ciation that has 22,000 members) but not asso-ciated with Mahagrapes.

The data were based on the recall memory ofthe households while some of the informationwas cross-checked with the cooperative/Mahagrapes for validation. The survey col-lected detailed information on household char-acteristics, infrastructural environment of thefarmers, details of the inputs and outputs in/from grape production and details on transac-tion costs. 12

The survey comprised five parts. In the firstpart, information on the access of the farmersto village level infrastructure was collected (togauge their physical access to markets andconnectivity through phones, transport, etc.).

Please cite this article in press as: Roy, D., & Thorat..., World Development (2008), doi:10.1016/j.worldd

Table 1 compares the availability of infrastruc-ture across the Mahagrapes and the indepen-dent farmers. The Mahagrapes farmers on anaverage are marginally farther away from thecity centers (with the difference in means beingsignificant) and all the farmers in the samplehave access to bus facilities in the vicinity. Interms of access to communication infrastruc-ture, though there is near parity between thetwo types of farmers, the variable is an imper-fect measure of actual access. Private access totelephone has a much greater impact on connec-tivity than the presence of a telephone facility ina locality. A simple test for difference in meansindicates that there is not much to distinguishthe two groups of farmers in terms of their ac-cess to infrastructure (except in their access toregional rural banks).

The second type of information in the surveyrelates to the characteristics of the farmersincluding education, experience in farming andcurrent and past occupation. Clearly, the

, A., Success in High Value Horticultural Exportev.2007.09.009

8 WORLD DEVELOPMENT

ARTICLE IN PRESS

Mahagrapes farmers have substantially higherexperience in farming. Moreover, relative tothe independent farmers, the average levels ofeducation are also higher for the Mahagrapesfarmers especially in the category graduate edu-cation and above. Labor economists define skillas a combination of education and experience(Borjas, 2003). Thus, the Mahagrapes farmers,on an average are higher skilled with the differ-ence in means being statistically significant. Inlight of the stringent standards requirements inthe western markets, this kind of positive selec-tion is quite likely. Note that (from Table 2), skilland not the land size is the distinguishing charac-teristic between the two groups of farmers.

In terms of occupational characteristics (e.g.,primacy of grape farming as an occupation),the two groups of farmers are nearly identical.An interesting observation from Table 2 is thata substantial proportion of grape farmers in the

Table 2. Individual characteristics of M

Experience in grape farming* (number of years)*

Average age (years)

Levels of educationa,**

– Illiterate (%)– Schooling (%)– Graduate (%)– Post graduation (%)– Specialization (%)

Proportion of farmers whose main occupation is grape f

Note: There is only one illiterate farmer in the sample andSource: Author’s calculations based on primary survey.a Schooling refers to a high school degree, graduate refer

masters degree.* denotes significance at 10% level.** denotes significance at 5% level.***denotes significance at 1% level.

Table 3. Annual input charges per

Average area under grape cultivation (acres)

Variable inputs

Chemical fertilizers (Rs./ton)Bio-fertilizers (Rs./ton)Pesticides (Rs./ton)Bio-pesticides (Rs./ton)IrrigationElectricity (Rs./ton)water charges (Rs./ton)

Source: Author’s calculations based on primary survey.a Data missing.

Please cite this article in press as: Roy, D., & Thorat..., World Development (2008), doi:10.1016/j.worldd

sample (both Mahagrapes and independent)have prior non-farming experience. Several ofthe grape farmers have invested capital accu-mulated elsewhere to get into commercial grapeproduction.

The third part of survey collected informationabout the inputs, outputs, prices and revenuesfor the farmers over a one-year period. Table 3presents information about the levels of inputuse for the Mahagrapes and the independentfarmers. More detailed information on the inputcosts of the farmers is provided in the Appendix.The details about inputs are particularly rele-vant as standards alter the mix of inputsthrough compliance requirements. Costs ofcompliance ultimately determine the relativeprofitability of production for the western mar-kets vis-a-vis domestic markets. For example,the grapes produced for the European marketshave restrictions on their pesticide residues

ahagrapes and independent farmers

Independent Mahagrapes

6 1744 45

0 1.281.5 58.516.3 30.51.1 6.11.1 3.7

arming (%) 51.5 47.3

he is a member of Mahagrapes.

s to a bachelors degree and post graduation refers to a

ton: annual averages (Rs./ton)

Independent Mahagrapes

5.56 5.48

2350.93 2861.631893.53 1793.193225.74 4023.66767.42 603.88

–a –a

867.54 963.561035.82 2094.56

, A., Success in High Value Horticultural Exportev.2007.09.009

SUCCESS IN HIGH VALUE HORTICULTURAL EXPORT MARKETS 9

ARTICLE IN PRESS

and farmers are restricted to using bio-fertilizersrather than chemical fertilizers.

Constrained to using better quality and differ-entiated inputs from the producers for the localmarkets, the role of cooperatives/Mahagrapesin providing inputs cheaper becomes veryimportant particularly for the small farmers.Thus, for requirements on specific factors, theinput costs of Mahagrapes farmers are statisti-cally not different (on an average) from the inde-pendent farmers (Table 3). The specific inputsthat are imported are purchased by Mahagrapesand sold to the farmers at little or no profit.Mahagrapes imports an expensive hormonefrom Australia on behalf of the farmers anddue to bulk buying charges a lower price fromthe members. Additionally, Mahagrapes pro-duces its own brand of specialized inputs suchas bio-fertilizer/organic manure (MahaTiko)and bio-pesticide (MahaVert), both of whichare provided cheaper to members. 13

Mahagrapes farmers have higher yields andalso earn higher revenues per acre of grape land(Table 4). The difference in means between thetwo groups of farmers is highly significant inthis case. There are several possible reasonsfor this. First, the average land size for theMahagrapes farmer under grapes is slightlyhigher by 0.08 acres. 14 Also, Mahagrapesfarmers get better quality inputs includingextension services at cheaper rates. Though allthe outputs produced by the Mahagrapes farm-ers are intended for the export markets, a partof the output gets rejected for quality reasons.The rejected output is then sold in the openmarket. Since, the rejected output is of betterquality than the average stock in local markets,it fetches a higher price. In the export markets,the gross price is three times higher than in lo-cal markets. However, selling to the exportmarkets requires availing of marketing servicesand use of cold chain facilities for which farm-

Table 4. Annual output a

Output sold in open market (kg) annuallyOutput sold in export markets (kg annually)***

Revenue (Rs./acre of grape land)***

Revenue (Rs./kg) (total = open + contract market)***

Gross price in the export market (Rs./kg)Net price in the open market (Rs./kg)Price open (Rs./kg)***

Source: Author’s calculations based on primary survey; *, 11% level of significance for test in difference of means betw

Please cite this article in press as: Roy, D., & Thorat..., World Development (2008), doi:10.1016/j.worldd

ers pay a fee. Net of these charges the prevail-ing price in 2004 was about twice as high.

Table 5 shows that though the revenues forthe Mahagrapes farmers are moderately higher,net of costs of compliance, the difference be-comes much smaller. However, the returnscould differ within the group of Mahagrapesfarmers and we are particularly interested inthe differential impact among the farmers basedon size. Importantly, those independent farm-ers (there exist very few such farmers and theyare large farmers) who sell some of their out-puts to export markets have to purchase inputsat higher prices (even if they purchase fromMahagrapes).

4. GREATER EARNINGS FOR THEMAHAGRAPES FARMERS

Are the Mahagrapes farmers significantlybetter off (in terms of operating profits) relativeto the non-members? The comparator is thefarmer who is outside Mahagrapes but sharessimilar characteristics. If the comparator wereother similar farmers who exported but werenot part of Mahagrapes, then the comparisonwould represent the contribution of Mahagra-pes in enhancing the gains from exporting.Comparing across similar farmers, those whoare linked with Mahagrapes and others whoare independent (and selling mostly domesti-cally) captures the difference between sellingdomestically and in export markets albeit withthe premise that small farmers are able to ex-port only when linked with Mahagrapes. 15

Summary comparisons of profit (as in Sec-tion 3) among the two groups does not controlfor the fact that the two groups might not sharesimilar characteristics that matter for profit-ability. If being an independent or a Mahagra-pes farmer were random, average returns would

nd revenues (averages)

Independent Mahagrapes

20,452 19,582765.96 7417.5880,667 144,474

15 23– 32.56– 19

14.89 16.14

0% level of significance; **, 5% level of significance; ***,een Mahagrapes and independent farmers.

, A., Success in High Value Horticultural Exportev.2007.09.009

Table 5. Operating profits for Mahagrapes and independent farmersa

Independent Mahagrapes

Revenues 206282.9 248235.3Operating profits (per kg)*** 9.39 14.98Operating profits/acre grape land # (without the specific costs for the exportmarket)***,b

52,633 119,370

Operating profits/acre grape land (discounted for costs specific for export markets)*** 26,633 59,586a Note: The profits are based on variable costs only, fixed costs are not included due to data constraints.b Note that a part of cost of compliance is included here in terms of the type of pesticides and fertilizers used that aredictated by the standards. The parts that are not part of production are included separately such as cleaning, grading,packing and certification.*** denotes significance at 1% level.

10 WORLD DEVELOPMENT

ARTICLE IN PRESS

be a good measure to demonstrate greater prof-its from joining Mahagrapes.

It is possible that Mahagrapes and indepen-dent farmers earn different profits precisely be-cause of a process of screening by Mahagrapesor self selection by the farmers. For example,Mahagrapes might select farmers based on cer-tain attributes such as land sizes or skills (edu-cation and experience). If the farmers picked upby Mahagrapes on an average are higher skilledor larger farmers than the average of the inde-pendent ones, this will imply a case of positiveselection. Similarly, if the pool of farmers se-lected (or chose to be selected) is inferior tothe pool of independent farmers then there isnegative selection. The positive or negativeselection can occur in terms of both observablesand unobservables, for example, farmer’s edu-cation level and skill, respectively.

The profit regression that we are interested inestimating is given as

Pi ¼ aþ d � di þ c � Xi þ ei; ð1Þwhere Pi denotes the profit of the ith grapefarmer, di is a dummy variable that assumes va-lue equal to 1 if the farmer is a Mahagrapesfarmer and zero otherwise and Xi is a vectorof control variables relating to the characteris-tics of the ith grape farmer. The matrix Xshould ideally include all those variables thatdetermine whether a particular farmer is amember of Mahagrapes or not. Further, if thevariables in the X matrix are uncorrelated withthe error term (e), then the ordinary leastsquares estimates of (1) will be consistent(Ramaswami, Birthal, & Joshi, 2006). In thatcase d, the OLS estimate from the regressionsin 1 can be treated as the true level effect onprofits from membership of Mahagrapes.

The specification in Eqn. (1) if estimatedusing OLS can suffer from omitted variablesbias. Unobserved ability or characteristics like

Please cite this article in press as: Roy, D., & Thorat..., World Development (2008), doi:10.1016/j.worldd

reputation that cannot be controlled for in theregression will lead to a selection bias. If thefarmers, for example, who are the members ofMahagrapes have greater unobserved abilitythen the estimated coefficient d captures the ef-fect of this omitted variable and not the effectof membership per se. Similarly, it is possiblethat those who become the members of Mahag-rapes have poor prospect outside owing to low-er ability or knowledge and thus the coefficientd is downward biased. In both cases the dummydi will be correlated with the error term ei.

In order to correct for the selection bias,standard instrumental variable technique canbe used. The ideal instrument is such that it iscorrelated with di, the participation variableand not correlated with variable Pi. Here, weappeal to the role of transaction costs in deter-mining the benefits of coordination. If transac-tion costs are higher then the incentive for thefarmers to become the members of a coordina-tion arrangement such as Mahagrapes is likelyto be stronger. This follows from the fact thatfarmers have a greater motivation to save onother costs if the transaction costs on the inputsnot provided by the cooperative or Mahagrapesare higher.

Consider the following equation that deter-mines participation in Mahagrapes:

di ¼ bþ hZi þ ui; ð2Þwhere Zi vector comprises variables that deter-mine membership of Mahagrapes. For the esti-mated d i to satisfy the exclusion criterion of aninstrument, there must be one variable in Zi

that is not in Xi. The exclusion variable (theinstrument) is the transaction costs on the in-puts not provided by the Mahagrapes or thecooperatives.

Imbens and Angrist (1994) show that with adummy endogenous variable, instrumental vari-ables methods estimate causal effects for those

, A., Success in High Value Horticultural Exportev.2007.09.009

SUCCESS IN HIGH VALUE HORTICULTURAL EXPORT MARKETS 11

ARTICLE IN PRESS

whose behavior would be changed by the instru-ment if it were assigned in a randomized trial.That is, the effect is estimated for subjects whowill take the treatment if assigned to the treat-ment group, but otherwise not take the treat-ment. This parameter is known as the LocalAverage Treatment Effect (LATE). If everyonein the population has the same response to aparticular intervention or treatment, the distinc-tion between LATE and other parameters doesnot matter. But with ‘‘heterogeneous treatmenteffects,’’ the parameter identified by instrumen-tal variables may differ from the average effectof interest. Since there are farmer level heteroge-neity, the instrumental variables measure theLATE and not the average effect necessarily.

As Angrist and Krueger (1991) have arguedusing probit or logit to generate first-stage pre-dicted values is not necessary and may even dosome harm. In two-stage least squares, consis-tency of the second-stage estimates does not de-pend on getting the first-stage functional formright (Kelejian, 1971). So using a linear regres-sion for the first-stage estimates generates consis-tent second-stage estimates even with a dummyendogenous variable (see Table A.3 in theAppendix for the first-stage regressions). More-over, using a nonlinear first-stage to generate fit-ted values that are plugged directly into thesecond-stage equation does not generate consis-tent estimates unless the nonlinear model hap-pens to be exactly right, a result which makesthe dangers of misspecification high. Thus, theIV regressions here use the first-stage linear mod-el to obtain predicted values for membership.

The results presented in Table 6 show thatcontrolling for both observed and unobserved

Table 6. Profit equation—dependent varia

Explanatory variables Coefficients (tOLS estima

Membership (the level effect) 27.6 (2.53)Distance from the city center �0.093 (�0Schooling 2.28 (0.30Age �0.28 (�0.Land �2.39 (�4.5Experience �1.76 (�0.Experience squared 0.060 (1.1Total land* membership �1.5 (1.2Grapeland* membership 2.5 (2.28)Number of observations 132

Instrument—transaction cost per unit of output on commo* denotes significance at 10% level.** denotes significance at 5% level*** denotes significance at 1% level.

Please cite this article in press as: Roy, D., & Thorat..., World Development (2008), doi:10.1016/j.worldd

characteristics, Mahagrapes farmers earn sig-nificantly higher profits than their independentcounterparts most of whom sell to local mar-kets where quality and safety requirements arelow but so are the returns. Since the IV estimateof the coefficient on membership is higher, itimplies that OLS estimates are downwardlybiased. Thus, the farmers who participate inMahagrapes and export are likely to havepoorer prospects as independent farmers thuspointing to negative selection. One caveat forthese profit results is important. Since theseare operating profits, they do not include someof the fixed costs that are important for exportmarkets and would thus further lower the prof-its for Mahagrapes farmers. For example,EurepGAP compliance requires that there beadequate sanitation facilities in places wheregrapes are sorted and pruned.

5. SELECTION/PARTICIPATION INMAHAGRAPES

Given the significantly higher net returns tothe Mahagrapes farmers, the question is whythere still are farmers who are not associatedwith Mahagrapes? These farmers are producingtheir outputs in the same area, facing similarproduction conditions and having access tosimilar public infrastructure.

The probit analysis of participation in Mahag-rapes below captures the role of different factorsin explaining the sorting of the farmers intoMahagrapes and independent farmers. Amongthe determinants of participation, the variablethat we are most interested in is the land size.

ble profit in thousands per acre of land

values)tes

Coefficients (t values) instrumentalvariable estimates

** 96.36 (2.53)*

.18) �0.58 (�1.10)) 11.68 (1.59)59) �0.67 (�1.55)4)*** �2.56 (�2.65)**

87) �2.95 (�1.15)0) 0.08 (1.10)) �3.51 (�1.29)** 0.22 (0.09)

132

n inputs.

, A., Success in High Value Horticultural Exportev.2007.09.009

12 WORLD DEVELOPMENT

ARTICLE IN PRESS

The participation equation does not supportthe hypothesis of significant influence of landsize in participation. Neither the total landholdings of the farmer nor the land of the farm-er under grapes is a significant determinant ofwhether a farmer belongs to Mahagrapes.Since, the land size data correspond to currenttime period, the insignificance of land sizes iseven more informative. Most farmers have in-creased their land sizes since their joining ofMahagrapes (Table 7).

Thus, at the time of selection, the true unob-served coefficient on land sizes is likely to besmaller. Indeed, given that farmers need capac-ity to meet the requirements of the productstandards, other attributes like age of the farm-er (proxy for experience) and his education le-vel are influential in determining participation.Most importantly the transaction costs relatedto the purchase of inputs not provided by thecooperative or Mahagrapes significantly affectsparticipation. This is the variable that we use asinstrument in profit regressions above. The listof variables included in the transaction costs isgiven in the Appendix (Table A.2).

To summarize, the results on participation donot support any evidence for a systematic biasagainst the smallholder. Neither the aggregateland size nor the land area under grapes deter-mines participation significantly. Conditionalon participation, all farmers gain. As regressionresults show, the level effect on profits per unitof land does not differ significantly by landsize (the coefficients on the interaction termsin Table 7 are insignificant).

Table 7. Probit estimation: factors determiningparticipation/selection in Mahagrapes

Explanatory variables Coefficients (t values)

Distance from the citycenter

0.021 (1.90)**

Schooling 0.38 (2.08)**

Age 0.01 (1.64)*

Land 0.007 (0.61)Experience 0.05 (1.18)Experience squared �0.0008 (�0.75)Land under grapes 0.008 (0.48)Transaction costs forfactors not provided bycooperative or Mahagrapes

0.0002 (1.92)**

Number of observations 132* Significance at 10% level.** Significance at 5% level.

Please cite this article in press as: Roy, D., & Thorat..., World Development (2008), doi:10.1016/j.worldd

(a) The issue of small farmers

Figures 1 and 2 show the distribution offarmers in the sample by land size (in acres).The distribution of the independent farmers iscentered in the middle while that of Mahagra-pes farmers is skewed towards smaller landsizes.

Ceteris paribus, small farmers are less likelyto have access to foreign markets owing toscale diseconomies that they face. Large farm-ers have more resources and better access totechnology and information. In the sample,42% of the small farmers (less than 2 ha ofcultivated grape land) belong to Mahagrapes,while rest are independent. The summary sta-tistics with regard to the land sizes for thegrape farmers in the sample are given inTable 8.

Within the class of small farmers, the operat-ing profit per kg of output sold in the market issubstantially higher for the Mahagrapes farm-ers. The operating profit per acre of small farmis also substantially higher for Mahagrapesfarmers. Though the numbers are not presentedin Table 8 but the difference is substantial forthe class of medium farmers as well. The differ-ence in profits for both the classes of farmerseither per unit of land or per unit of output isstatistically significant.

6. CONCLUSION

Overall, Mahagrapes has emerged as a suc-cessful example of the cooperative model inthe horticulture sector in India. It providesfarmers with a platform for collective bargain-ing especially in negotiation with foreign buy-ers. Most importantly, Mahagrapes epitomizesthe role of scale economies in information pro-curement and processing regarding the exportmarkets. A group of lead farmers procuresinformation and because of greater skill canprocess and disseminate it to smaller and lesserskilled farmers. Also, the scale economies inbranding (creating and preserving brand value)involves fixed costs, which get averaged outover greater outputs in an organization suchas Mahagrapes.

Our results indicate that conditional on par-ticipation, the Mahagrapes farmers earn signif-icantly higher net profits. The result thatfarmers earn significantly higher profits is ro-bust to controlling for endogeneity. The resultson factors determining participation provide no

, A., Success in High Value Horticultural Exportev.2007.09.009

Frequeny distribution of Maha Grapes farmers

0

2

4

6

8

10

12

14

16

5 7 9 11 13 15 17 18 20 25 30 40land size

Nu

mb

er o

f fa

rmer

s

Figure 1. Distribution of Mahagrapes farmers in the sample by size (land size in acres).

Table 8. Summary sample comparison among small farmers

Mahagrapes Non-Mahagrapes

Percentage of medium sized farmers in terms of total land area (4.9–12.3 acres) 43% 49%Percentage of medium sized farmers in terms of total grape land area (4.9–12.3acres)

41% 34%

Percentage of farmers who are small in terms of total land area (<4.9 acres) 15% 19%Percentage of farmers who are small in terms of grape area cultivated (<4.9acres)

42% 57%

Average total land among the small farmers (in acres) 11.8 9.6Average grape area for small grape farmers (in acres) 2.5 2.9Average grape area for small farmers (in acres)* 4.6 3.9Average operating profit per acre of small farm (in thousand rupees perannum)**

135 78

Average profit (in rupees) per kg of output among small farmers** 14.5 10.3Authors’ calculations

* denotes significance at 10% level.** denotes significance at 5% level.

Frequency distribution of Independent farmers

0

2

4

6

8

10

12

14

1 3 5 7 9 11 13 15 17 19 21 23 25 27Land size

No

of

farm

ers

Figure 2. Distribution of independent farmers in the sample by size (land size in acres).

SUCCESS IN HIGH VALUE HORTICULTURAL EXPORT MARKETS 13

ARTICLE IN PRESS

evidence for a systematic bias against the smallfarmers being selected.

Given the higher net returns the question iswhy are there farmers who are not associatedwith Mahagrapes? First, there are reasons pos-sible for which Mahagrapes could restrict itsmembership. Most importantly, demand for

Please cite this article in press as: Roy, D., & Thorat..., World Development (2008), doi:10.1016/j.worldd

new members by the cooperative is a deriveddemand and Mahagrapes is demand con-strained. Mahagrapes supplies its output tofew retail outlets in the European markets.Each year, the buyers specify the amount ofgrape they will buy. Procurement from farmersis done according to the demand.

, A., Success in High Value Horticultural Exportev.2007.09.009

14 WORLD DEVELOPMENT

ARTICLE IN PRESS

The more important reasons are the strin-gent quality norms that farmers have to com-ply with. These include standards for size,shape and color of grapes. There are strictstandards also for permissible levels of pesti-cides and other chemicals. A farmer whowants to be a part of the Mahagrapes shouldhave the capacity to bear the costs of comply-ing with the standards. Further, foreign buyersreject the consignments if it fails to meet anyof the quality criteria. In such a case, it isimportant that the farmer has the capacity tobear this risk. Both these things imply thatnot all farmers can contract with the firmand gain from this association. However, in

Please cite this article in press as: Roy, D., & Thorat..., World Development (2008), doi:10.1016/j.worldd

terms of common attributes in the population,we find no bias against the small farmers inparticipation.

However, the case of Mahagrapes is impor-tant from the point of view of success in highstandard high return markets for small farm-ers. Our results indicate that net of costs ofcompliance, the small farmers do earn signif-icantly higher profits compared to their out-side option. Since there is no bias inselection based on land sizes and conditionalon participation the small farms can complywith standards and earn significant profits;without demand constraints, the model canbe scaled up.

NOTES

1. Calculations based on FAO Data, 2005.

2. FAO Stats, 2006.

3. Consider, for example, the cost of obtaining infor-mation on standards. It often requires communicationand possibly travel. The information once procuredapplies to the whole output.

4. India’s average farm size has decreased from 2.2 hain 1950 to 1.4 ha in 1995/96 (FAO Statistics Division).Smallholders, defined to be those operating less than2 ha, operate about 36% of the aggregate land area.

5. Alternatively, institutional arrangements such ascontract farming have been suggested as a means ofstrengthening the linkage between markets and produc-ers which at the same time improves access of thefarmers to inputs such as credit, information andimproved varieties of the crops. In India, contractfarming though quite prevalent is still not institutional-ized in most places. The state run cooperatives have beenin existence for decades and have been pioneers insuccessfully consolidating farmers and helping them tomarket their produce collectively. Sugar cooperatives area good example of this. Over time however, bureaucraticred tape-ism and rent seeking (Bannerjee, Mookherjee,Munshi, & Ray, 2001) has shifted the decision makingand economic powers away from the small farmermajority into the hands of the governing few.

6. In the early years of exports, consignment rejectionat times was as high as 80%. At present the Mahagrapesclaims a negligible rate of rejection (less than 1% after2001, less than 10% after 1995).

7. In the initial years, consignment rejections occurredbecause there was imperfect understanding of marketrequirements—from seed quality, to harvest and postharvest care to packaging practices. The average rejec-tion rate prior to 1995 was 50% (Naik, 2006).

8. In Indian laboratories residues were tested usingGC/MS method (gas chromatography/mass spectrome-try) while in EU laboratories they were tested by LC/MSmethod (liquid chromatography/time-of-flight massspectrometry).

9. EU Regulation EC No. 1148/2001, introduced onJuly 1, 2002, requires all fresh produce arriving in theEurope to undergo an ISO 9000 style inspection to verifyconformity with marketing standards. This inspection iscarried out at the port or airport before release intocirculation. The only exceptions are for consignmentswith certificates of compliance from countries whosesystems of export quality inspection have been approvedby the Europe. Such consignments with both a phyto-sanitary and a quality conformity certificate are releasedwhen customs procedures are complete. Exporters withapproved certification systems obviously have a com-petitive edge. Some countries, such as Jordan, haveestablished a pesticide residue testing laboratory and aheavy metals testing laboratory, and issue conformitydocuments on consignments to the Europe (Kane,Sallah, Alex, & Meer, 2002).

10. Land distribution pattern of Maharashtra showsthat 70% of the holdings are less than equal to 2 ha and20% are between 2 and 4 ha. The remaining 10% aremedium, large and very large farmers. In the threedistricts surveyed 64% of land holding are less then 2 ha.It would be safe to categorize all the farmers with less

, A., Success in High Value Horticultural Exportev.2007.09.009

SUCCESS IN HIGH VALUE HORTICULTURAL EXPORT MARKETS 15

ARTICLE IN PRESS

than 2 ha of land as small farmers (normally definedmarginal farmers being included in the definition ofsmall farmers).

11. These inputs are also sold to non-members but athigher prices implying cross-subsidization.

12. Of the 183 farmers chosen, 47 are from Nashik, 26are from Pune and 111 are from Sangli districts.

13. Since Mahagrapes farmers also sell domesticallyand to export markets in the Gulf and in Sri Lanka, theyneed not necessarily use only specialized inputs.

Please cite this article in press as: Roy, D., & Thorat..., World Development (2008), doi:10.1016/j.worldd

14. In the three surveyed districts, Pune, Nasik andSangli, the average sizes of the landholdings are 5.5, 6.2and 4.3 acres, respectively (Agricultural Census, Gov-ernment of India 2001). Thus, by the average size of theirgrape farms, the Mahagrapes and the independentfarmers are near the area average. However, comparedto the land sizes in countries that compete withMahagrapes in the same market (e.g., Chile and Aus-tralia), the average size of landholdings is several timessmaller.

15. There are no independent small farmers in the areawho export and the few independent exporters whoexist, they are all large farmers.

REFERENCES

Angrist, J. D., & Krueger, A. B. (1991). Instrumentalvariables and the search for identification: Fromsupply and demand to natural experiments. Journalof Economic Perspectives, Summer 1992.

Athukorala, P., & Jayasuriya, S. K. (2003). Food safetyissues, trade and WTO rules: A developingcountry perspective. World Economy, 26(9),141–162.

Bannerjee, A., Mookherjee, D., Munshi, K., & Ray, D.(2001). Inequality, control rights, and rent seeking:Sugar cooperatives in Maharashtra. Journal ofPolitical Economy, 109, 138–190.

Birthal, P. S., Joshi, P. K., & Gulati, A. (2005). Verticalcoordination in high-value food commodities: Impli-cations for smallholders. MTID Discussion paper.Washington DC: International Food Policy Re-search Institute.

Borjas, G. J. (2003). The labor demand curve isdownward sloping: Reexamining the impact ofimmigration on the labor market. Quarterly Journalof Economics, 118(4), 1335–1374.

Faiguenbaum, S., Berdegue, J., & Reardon, T. (2002).The rapid rise of supermarkets in Chile: Effects ondairy, vegetable, and beef chains. Development PolicyReview, 20(4), 459–471.

Farina, E. M. M. Q. (2002). Consolidation, multina-tionalization, and competition in Brazil: Impacts onhorticulture and dairy products systems. Develop-ment Policy Review, 20(4), 441–457.

Food and Agricultural Organization (FAOSTAT) 2005.Food and Drug Administration (FDA). (2007). Import

Detention Reports. Available at: http://www.fda.-gov/ora/oasis/ora_oasis_ref.html.

Ghezan, G., Mateos, M., & Viteri, L. (2002). Impacts ofsupermarkets and fast-food chains on horticulturesupply chains in Argentina. Development PolicyReview, 20(4), 389–408.

Helemberger, P., & Hoos, S. (1962). Cooperative enter-prise and organization theory. Journal of FarmEconomics, 44, 275–290.

Imbens, G. W., & Angrist, J. D. (1994). Identificationand estimation of local average treatment effects.Econometrica, 62(2), 467–475.

Kane, S., Sallah, T., Alex, G., Meer, K.v.d. (2002).Agricultural Investment Note. World Bank.

Kelejian, H. (1971). Two stage least squares andeconometric models linear in the parameters butnon linear in the endogenous variables. Journal of theAmerican Statistical Association, 66, 373–374.

Kleinwechter, U., & Grethe, H. (2005). The significanceof food quality and safety standards in developingcountries – A case study for the EurepGAP standardin the Mango Export Sector in Piura, Peru. Poster atthe Deutscher Tropentag 2005, Hohenheim.

McCulloch, N., & Ota, M. (2002). Export horticultureand poverty in Kenya’. IDS Working Paper. 174.Brighton, Institute of Development Studies.

Mehta, R., & George, J. (2003). Processed food productsexports from India: An exploration with SPS regime.Research report of Australian National University,University of Melbourne, Research and InformationSystem (India) and Thammasat University (Thai-land).

Naik, G. (2006). Bridging the Knowledge Gap inCompetitive Agriculture: Grapes in India. In Chan-dra Vandana (Ed.), Technology adoption and exports:How some developing countries got it right?. Wash-ington DC: World Bank.

Ramaswami B., Birthal, P.S., & Joshi, P.K., 2006.Efficiency and distribution in contract farming: Thecase of Indian poultry growers, MTID DiscussionPapers 91, International Food Policy ResearchInstitute, Washington DC, USA.

Reardon, T., & Berdegue, J. (2002). The rapid rise ofsupermarkets in Latin America: Challenges andopportunities for development. Development PolicyReview, 20(4), 371–388.

Umali-Deininger, D., & Sur, M. (2006). Food safety in aglobalizing world: Opportunities and challenges forIndia. Paper presented at the International Associ-ation of Agricultural Economists Conference, GoldCoast Australia, August.

van Bekkum, O. F., & Dijk, G. (Eds.) (1997). Agricul-tural co-operatives in the European Union, trends andissues on the eve of 21st century. Holland: Royal vanGorcum.

, A., Success in High Value Horticultural Exportev.2007.09.009

16 WORLD DEVELOPMENT

ARTICLE IN PRESS

von Braun, J., & Immink, M. D. C. (1994). Nontradi-tional vegetable crops and food security amongsmallholder farmers in Guatemala. In J. von Braun,& E. Kennedy (Eds.), Agricultural commercialization,economic development, and nutrition. Baltimore:Johns Hopkins University Press.

Table A.1. Detai

Annual input charges for output: annual averages (Rs./t

Fixed inputs

Green houses (Rs./ton)Tube wells (Rs./ton)Tractors (Rs./ton)Nursery raising (Rs./ton)Seed (Rs./ton)

Variable inputs

Chemical fertilizers (Rs./ton)Bio-fertilizers (Rs./ton)Pesticides (Rs./ton)Bio-pesticides (Rs./ton)IrrigationElectricity (Rs./ton)Water charges (Rs./ton)

a Data missing.

Table A.2. List of transactio

Items in transaction costs (in Rs.)

Cost of communication — telephone, etc.Travel costsLoading costsUnloading costsCommission to agentAverage transaction cost

Cost of communication — telephone, etc.Travel costsLoading costsUnloading costsCommission to agentAverage transaction cost

Cost of communication — telephone, etc.Travel costsLoading costsUnloading costsCommission to agentAverage transaction cost

APPEND

See Tables A.1–A.3.

Please cite this article in press as: Roy, D., & Thorat..., World Development (2008), doi:10.1016/j.worldd

Warning, M., & Key, N. (2002). The social performanceand distributional consequences of contract farming:An equilibrium analysis of the Arachide de BoucheProgram in Senegal. World Development, 30(2),255–263.

led input costs

on) Independent Mahagrapes

–a –a

32.96 8.0714.75 19.16982.05 1016.761127.93 1446.63

2350.93 2861.631893.53 1793.193225.74 4023.66767.42 603.88

–a –a

867.54 963.561035.82 2094.56

n costs for selected inputs

Inputs (figures below show the sample average)

Tubewell48.02115.0233.9

––

65.64

Pesticides31.70121.2832.92

––

61.96

Manures6785168––

106.6

IX A

, A., Success in High Value Horticultural Exportev.2007.09.009

Table A.3. First-stage regressions for IV profit equation

Dependent variable (contract = 1, independent = 0)

Grape land size 0.002 (0.53)Distance to urban center 0.007 (1.96)**

Farmer’s education level 0.13 (2.28)**

Age of the farmer 0.006 (1.61)Experience of the farmer in grape farming 0.01 (1.18)Experience of the farmer in grape farming squared 0.09 (1.55)Transaction cost 0.0009***

Constant �0.31 (�1.55)

SUCCESS IN HIGH VALUE HORTICULTURAL EXPORT MARKETS 17

ARTICLE IN PRESS

Please cite this article in press as: Roy, D., & Thorat..., World Development (2008), doi:10.1016/j.worldd

Available online at www.sciencedirect.com

, A., Success in High Value Horticultural Exportev.2007.09.009