Community Based

Community-Based Approaches in Philippine Agricultural Projects: Fad, Facts, and Framework

FINAL REPORT* Roehlano M. Briones Postdoctoral Fellow (Economics) WorldFish Center Jln. Batu Maung, Batu Maung 11960 Bayan Lepas Penang, Malaysia

Submitted to the Philippine Institute for Development Studies , 2003

*

The comments of Cristina David are gratefully acknowledged. The author alone is responsible for the contents of this paper.

1. Introduction Community-based approaches (CBAs) are now widely accepted as an essential element of development at the grassroots. Development discussions are often fashionably laced with the rhetoric of “participation” and “empowerment”. Such rhetoric however risks obscuring a critical assessment of what makes CBAs work, or even whether they work as well as believed. This paper contributes to such an assessment by surveying the literature on CBAs in agricultural development. Its main problem is to identify, at least conceptually, the contribution of CBAs to the efficient implementation of development interventions. This survey consolidates the various models of community-based participation and organization into a coherent framework. Some of the evidence for these models are also reviewed. The analysis will form the basis of drawing broad implications for public policy. This study adopts an economic approach, which is most appropriate for the analysis of efficiency. Even measures aimed at equity and sustainability – two other major aims of development policy – need to be evaluated in terms of cost-effectiveness relative to alternative measures of equity and sustainability. Part of the efficiency analysis of this paper is to delineate the roles between public and private sector, to avoid wastage of public funds in activities that can be done as well or better by the private sector. The economic approach is of course only one of several viewpoints by which to evaluate CBAs. Where appropriate perspectives from other social sciences (e.g. anthropology and sociology) would be taken into account. In terms of normative analysis, the other social sciences would posit objectives other than efficiency; Michener (1998) for example idealizes the “transformative” attitude towards beneficiary participation (i.e. participation as seen as an end in itself, rather than a means to an end). This paper takes an instrumentalist attitude, not because it is normatively “better,” but only to maintain analytical focus. In terms of positive analysis, economics emphasizes methodological individualism. Social phenomena are reduced to individuals and their interactions; individual motivation is identified with rational self-interest, while group interactions result from different individuals simultaneously trying to realize their own interests. This is formalized by the mathematical construct of the utility function, which each individual attempts to maximize subject to constraints; stable patterns of behavior across individuals are described by an equilibrium in the system of interactions. Insistence on methodological individualism is an analytical simplification; it should be recognized that such an insistence may undervalue some important facets of organizations, such as the formation of norms, power relations, and similar phenomena stressed by the other social sciences. These factors shall be referred to occasionally throughout the paper. Definitions of CBAs abound. For this paper the essential characterization of a CBA is beneficiary participation, collectively organized at the community level, in designing and implementing development interventions. This definition identifies three elements, namely participation, collective action, and organization. This study will provide a

1

detailed treatment of the elements of participation, collective action, and organization. The rest of the paper is organized as follows: CBAs in Philippine agriculture and policy are discussed in Section 2. The basic argument behind beneficiary participation in development projects is presented in Section 3. Section 4 deals with the problem of collective action. Collective action in agribusiness functions (processing and trade) require deeper analysis of organizations, which is most formally embodied in the agricultural cooperative to be discussed in Section 5. Section 6 discusses the policy implications. Section 7 concludes. 2. CBAs in Philippine Agriculture1 Community-based arrangements are not a new development in Philippine agriculture. In irrigation for example, communities have evolved traditional systems in handling the management of communal facilities. Attention has been drawn to the ingenious zangjera in Ilocos Norte, a communal arrangement covering small brush dams spanning a few villages. The zangjera exhibits features that have sustained collective efficiency in water use, despite some inefficiency in intra-farm utilization (Cruz and Cruz, 1984). The cooperative movement also began making inroads into the Philippines from the late 19th century. By 1952, support for cooperatives reached the legislative level in the enactment of the Agricultural Credit and Cooperative Financing Administration Law. The law organized cooperatives who were provided collateral-free loans, mostly in restive areas in Central Luzon. Unfortunately the program was unsuccessful as 72% of loans went unrepaid; numerous state-sponsored cooperatives folded up, often as a result of incompetence and corruption (Sibal, n.d.). The Martial Law Period The thrust towards agricultural modernization gathered momentum in the Martial Law period. With it came a renewed emphasis on community-based approaches. Considerable investment was made in irrigation systems, which were initially characterized by a command-and-control mechanism. Later the National Irrigation Administration (NIA) began turning over fee collection and management to user associations. The NIA also organized Irrigators Associations (IAs) to manage communal irrigation facilities (with a coverage area of 1,000 ha. or smaller.) Based on its internal guidelines, at least 80% of farmers in the coverage area should be members of the IA prior to the construction of a facility. An IA is administered by its own board of representatives, elected by majority vote of members. Ideally the IA should be involved in in all phases of a project (planning, construction, and maintenance) as well as in shouldering project cost. In addition to an engineer, who oversees the technical side of irrigation works, the NIA provides an Irrigation Development Officer who works full-time at community organizing, as well as coordinating between farmers and technical staff.

1

This study distinguishes agriculture from the natural resource sector, in which CBAs figure even more prominently. The latter though is outside the scope of the paper.

2

Measures were also instituted to further strengthen agricultural cooperatives, as well farmer associations known as the Samahang Nayon (SN). The Samahang Nayon was organized to facilitate land reform, and train farmers in cooperation and business activities; after adequate preparation, these SNs were to be transformed into full-fledged cooperatives with business functions. SNs and farmer cooperatives were given key roles in the Masagana 99, a direct credit program for increasing yields and achieving agricultural self-sufficiency. The Post-Martial Law Period The return of democracy ushered in a new Constitution, which mandates the promotion non-governmental, community-based, and sectoral organizations as state policy. On this basis, Congress enacted several laws that recognized CBAs in development. Foremost of these is the RA 6838, or the Cooperative Code, and its companion RA 6839 which formed the Cooperative Development Authority (CDA) under the Office of the President. The Code defines cooperatives as follows (Section 3): a duly registered association of persons, with a common bond of interest, who have voluntarily joined together to achieve a lawful common social or economic end, making equitable contributions to the capital required and accepting a fair share of the risks and benefits of the undertaking in accordance with universally accepted cooperative principles.

These principles include (Section 4): Democratic control. Members of primary cooperatives shall have equal voting rights on a one-member-one-vote basis. Cooperatives shall be administered by persons who are democratically appointed. Limited Interest in Capital. Share capital shall receive a strictly limited rate of interest. Division of Net Surplus. Net surplus arising out of the operations of a cooperative belongs to its members and shall be equitably distributed for cooperative development common services, indivisible reserve fund, and for limited interest on capital and/or patronage refund in the manner . Cooperative Education. All cooperatives shall make provision for the education of their members, officers and employees and of the general public based on the principles of cooperation. Cooperatives are classified into the following types under the Cooperatives Code: Credit cooperatives function as savings and loan associations. Consumers cooperatives specialize in procurement and retailing of commodities, at least in part for the benefit of members. Producers cooperatives are organized for agricultural or industrial production and may run farms or factories. Marketing cooperatives engage in the supply of production inputs and marketing services for members. Service cooperatives engage in various household and community services, including health, insurance, housing,

3

transport, labor, electricity, and communication. Finally, multi-purpose cooperatives combine two (2) or more of the business activities of various kinds of cooperatives. Cooperatives, like corporations, are limited liability entities (Section 12). Unlike corporations though, the Code explicitly exempts cooperatives from taxes on transactions with members (Section 62). Other tax provisions include: Small and medium size cooperatives (with net savings of below 10 million pesos) are exempted from all taxes, including custom duties on imports of machinery (where said equipment is locally unavailable). All new cooperatives (ten years or below) are exempted from income and sales taxes. Cooperatives are also accorded a number of privileges, including: •

The preferential right to supply government institutions and agencies with agricultural commodities produced by their own members;

•

Preferential treatment in the allocation of cargo in commercial shipping vessels;

•

Preferential rights in the management of public markets and lease of public market facilities;

•

For credit cooperatives, entitlement to loans, credit liens, rediscounting, and other eligible papers with government financial institutions;

•

Exemption from pre-qualification bidding requirements in transactions with the Philippine Government and its corporations.

In addition to the Cooperative Code, the other relevant landmark legislations are the Comprehensive Agrarian Reform Law, the Local Government Code, the Magna Carta for Small Farmers, and the Agriculture Fisheries Modernization Act (Castillo et. al., 2003). Cooperatives are recognized as a crucial link in the government service delivery at the local level. Funds from agricultural modernization are allocated to farmer organizations and cooperatives. Cooperatives are provided special treatment in the promotion of marketing activities (such as fertilizer distribution), and are the main conduit for the conduct of livelihood projects. In the area of agricultural credit support, the post-Martial Law period saw a shift away from direct credit programs owing to low repayment rates afflicting these schemes. The new generation of credit programs place more emphasis on group-based lending and savings mobilization (Esguerra, 1995). Credit delivery was increasingly shifted to government financial institutions. A major source of agricultural finance (particularly for developing agrarian reform areas) is the Land Bank of the Philippines (LBP), which courses its small farmer lending through cooperatives. Despite all this legislative support, Castillo et. al. (2003) note that the current legal framework differs from the previous ones in the relative absence of explicit funding

4

allocation support for cooperatives. Rather, under the principle of subsidiarity invoked by the Cooperative Code, the cooperative sector itself is given the primary responsibility for its own progress. Limited funding for agricultural development (in particular the meager appropriation of the CDA) has ruled out the aggressive state promotion of cooperatives. Nevertheless the cooperative sector remains sizable; around 2.8 million persons are organized in nearly 60,000 cooperatives all over the country (Table 1). In Philippine agriculture, the most common types of cooperatives have provide marketing and credit services. Only 57% however are operating; the rest, while registered, are non-operating or are effectively dissolved. Of the registered cooperatives, by far the greatest proportion is found in agriculture, mostly in the form of multipurpose cooperatives. Multi-purpose nonagricultural cooperatives are second, while credit cooperatives are a distant third.

Table 1. Selected statistics on cooperatives in the Philippines, 2001

Number of registered cooperatives Total membership (As of December 31, 2000) Paid-up share capital of cooperatives (pesos) Number of cooperatives, by operational status a. Operating Coops b. Non- Operating Coops c. Dissolved and cancelled Number of cooperatives by type Multi-Purpose (Agricultural) Multi-Purpose (Non-Agricultural) Credit Service Consumers Producer Marketing Other cooperatives

59,765 2,838,913 3,720,677,553

% of total 100.0 100.0

33,989 21,479 4,297

56.9 35.9 7.2

33,012 17,641 3,934 1,425 1,208 1,066 767 712

55.2 29.5 6.6 2.4 2.0 1.8 1.3 1.2

SOURCE: CDA

3. Participation in Development: Basic arguments Evidence on participation While anecdotal accounts of the benefits of participation abound, solid evidence is only recently accumulating. Alkire et. al. (2001) enumerates several cases in which participatory/community-based approaches have promoted the efficiency and effectiveness of rural development programs. These include:

5

•

A study of 1,875 households in rural communities in six countries (Benin, Bolivia, Honduras, Indonesia, Pakistan, Uganda) suggests that water system sustainability is greater when communities make key investment decisions and when they pay part of the investment costs, as they can assert their preferences in the asset design (Sara and Katz, 1997).

•

A study for Bangladesh suggests that Parent Teacher Associations might be in the best position to determine which children should receive tuition or school lunch subsidies (Ravallion, 1999).

•

Research on community-based irrigation systems in Asia usually find that systems built and operated by users (often with little external assistance) are more productive compared to government-built systems (Lam, 1998).

•

Studies on education and communities in the Philippines (Jimenez and Paqueo, 1996; Jimenez and Sawada, 1998) find that community-managed primary schools appear to supply similar levels of enrollment and quality with lower costs. Community management may likewise lead to better educational outcomes.

•

A case study of water supply projects in five Indian states finds that the most successful projects were those in which decision-making authority was decentralized in such matters as the location of water standpoints and latrines, and where organizations to monitor and maintain these facilities were developed (Manikutty, 1998).

In addition, Korten and Siy (1989) have found that participation motivates sustainability and cost-recovery in irriagation systems. Ostrom, E. and R. Gardner (1998), citing Svendson (1992) note that where farmers have been allowed to effectively organize, greater diversity of rules have emerged from the annual bargaining process. Such localized rule formation promotes efficiency of water use. Admittedly, correlations between beneficiary participation and project success are prone to selection bias, i.e. the factors accounting for project success may be the same factors that make participation more likely. For example, in communities were average ability and educational attainment is higher, development interventions are more likely to be effective; simultaneously, individuals are more likely to welcome community action. Some studies deal with this source of bias. Isham et. al. (1995) show that participation has improved service delivery in water supply projects, and that causation runs from participation to service, rather than the other way around. Grootaert (1999) as well as Grootaert and Narayan (2000) relate generic community participation directly to household outcomes by applying multivariate regression to cross-section data. With measures of household welfare, assets, savings, and credit access as their dependent variables, they include as determinants indicators of community involvement, such as membership in a peasant organization. In Indonesia and Bolivia, greater community participation leads to larger household assets, more savings, better credit access, and

6

higher per capita spending overall. The causation runs from participation to household outcomes rather than the reverse. Hodinott et. al. (2001) apply multivariate regression to link participation with the impact of public works programs in South Africa. They found that participation leads to the following: a lower ratio of project to local wages; increases in labor intensity; lower costs of creating employment; and lower costs of transferring funds to the poor. To account for the possibility that participation is also an effect of underlying success factors, the study controls for selectivity bias. There is no evidence that community participation increases cost overruns or the ratio of training to employment. Rationale for participation Hodinott et. al. (2001) also present a model that explains why participation should have salutary effects documented in the foregoing. Three entities are posited, namely the financier, the provider, and the community (i.e. the beneficiaries). In the traditional topdown arrangement, these functions are sharply delimited; beneficiaries are passive recipients of externally provided and funded services. This arrangement though may encounter problems when the flow of benefits is sensitive to local conditions and is dependent on information that only beneficiaries are likely to have. For example, irrigation systems best designed when detailed information on water flows and soil conditions are available. Engineers and external planners often lack such detailed and site-specific knowledge. Furthermore, when external monitoring is costly, financiers and providers may undertake opportunistic behavior that reduces welfare impact on beneficiaries. For example, providers may impose their own preferences for public prestige on the design and implementation of the project; the financier can also distort project design towards their own ends owing to their control of project funds. In other words, one of the parties can “holdup” the entire project and the economic surplus it could generate by threatening withdrawal. Such problems can be addressed by permitting overlap of functions. In the model, participation is interpreted as a form of role-sharing, in which the community also acts as a provider and sometimes a financier. This arrangement is a more effective means of tapping the information endowment of communities. Furthermore, it is more likely to align project design and implementation towards a more efficient realization of beneficiaries’ welfare. The foregoing model is highly simplified, in that communities appear as monolithic entities. They act as if they were a single agent, utilizing their information endowment, in designing the project, and in monitoring or undertaking project operations. However motivating and coordinating individual beneficiaries in a group endeavor is far from a trivial matter. The problem of collective action, is discussed in the following section.

7

4. The problem of collective action In the standard model of general equilibrium, marginal benefits and costs of economic decisions are all internal. In this special case collective action is superfluous, i.e. the competitive price system mediates all transactions needed to achieve allocative efficiency. However there are numerous instances in which internalization of all marginal benefits is infeasible. In the case of public goods, reliance on the price system leads to undersupply, as beneficiaries would attempt to free ride on others’ contributions. Collective action can occur at the state level; public good provision by the state is in fact the standard assumption in the public finance literature. However, especially for local public goods, provision by the state may not be the most appropriate. The whole idea behind CBAs is the superiority of community level cooperation over state level collective action. To explain community level cooperation, game theoretic models of strategic interaction between individuals becomes a relevant analytical framework. Game theoretic analysis In game theory, individuals are “players” who select one out of a set of possible strategies, with each combination of individual strategies yielding a set of individual payoffs. A cooperative solution, which corresponds to collective action, is that combination of strategies that yield the largest group pay-off. The modeling thrust therefore is to ascertain whether a cooperative solution is an equilibrium. The “Prisoner’s Dilemma” describes a non-cooperative equilibrium. The dilemma arises when a player who does not cooperate, or defects, can realize a pay-off larger than that under cooperation, as long as other players do not defect. However if other players defect, the maximal group pay-off is forfeited. The Prisoner’s Dilemma is a common way to describe the breakdown of cooperation. The standard answer to this dilemma is the folk theorem (Fudenberg and Maskin, 1986). According to this theorem, cooperative solutions are equilibria for games that are repeatedly played, as long as the following conditions are satisfied: players who defect can be identified and penalized; defection is not too rewarding; and the future is sufficiently important to each player. The basic idea is that players who permanently cooperate obtain a net gain from a continued relationship with other players; this imposes an opportunity cost on defection. It is noteworthy though that in games that seem to closely approximate real world situations, a cooperative solution is just one of many equilibrium outcomes. A noncooperative solution may be an equilibrium as well; other factors must be introduced into the game structure to sharpen the model’s predictions. Exogenous factors such as group norms and group trust may tilt equilibrium towards cooperation. Norms lay down a set of expectations regarding individual behavior shared by group members. Trust provides assurance that other players will cooperate in a predictable manner, subject to past known behavior. The stock of trust specific to a network of cultivated relationships

8

has been referred to as “social capital” and may be crucial to the solution of the collective action problem. Aggarwal (2000) presents a model that captures the importance of group norms. Consider a group of N farmers, e an N-dimensional vector of inputs from by the farmers (which can include effort). The cost to each farmer is given by C (ei ) , while the benefit accrues to the group and is given by f (e) ; suppose each individual receives an equal share in this group benefit. Then the individual’s maximization problem is given by

max

f (e) − C (ei ) N

for individual i, i = 1, 2, …, N

(1)

On the other hand, the Pareto optimum (which maximizes group net benefit) is obtained from the solution of a different problem: max f (e) − ∑ i =1 C (ei ) . N

(2)

Clearly the solution to (2) diverges from the solution to (1); it is easy to show that the solution to (1) implies underprovision of effort relative to the solution of (2).2 To avoid such a solution, one may conceive of a “peer pressure function” Pi specific to individual i. This function is also dependent on the effort levels of the members. Hence (2) is modified as max

f (e) − C (ei ) − Pi (e) N

for i = 1, 2, …, N

(2’)

Each member must take into account not just the benefits of free-riding, but also the sanctions of peer pressure in deciding upon individual effort level. The higher the extent of peer pressure, the smaller the difference between actual contributions and the group norm. Based on the foregoing, Aggarwal identifies several factors contributing to the likelihood of group cooperation, such as: a smaller group size, kinship ties among members, interaction between members in a variety of social settings, a history of past cooperation among members, availability of information about benefits and costs of collective activity, and high group regard for norms. Evidence on collective action In addition to group norms, Molinas (1998) cites other factors as relevant for collective action, namely: gender, outside intervention, and heterogeneity of players. With respect to gender, cooperation is usually deemed more likely the more actively women participate in group. Outside agents meanwhile are expected to promote cooperation if 2

The first order conditions for (1) and (2) respectively are: f i ' (e1 ) = NC '(ei1 ) ; f i ' (e 2 ) = C '(ei2 ) . Then

e1 < e 2 if C " > 0, f " < 0 .

9

they provide training and education in cooperation, as well as assist in identifying and penalizing defectors. On the other hand, they may impede cooperation by exacerbating internal conflict, or create dependency relations that undermine organizational autonomy. As for heterogeneity, the literature contains mixed expectations. In the first place, too much homogeneity may inhibit the rise of “organizational entrepreneurs” who would invest time and effort in forming a functioning group; on the other hand, homogeneity may improve awareness about mutual dependency, a factor identified by Ostrom and Gardner (1993) as an element of successful local cooperation. Molinas’ empirical analysis uses data from 104 peasant organizations in Paraguay. He constructs an index measure of organizational performance based on: conduct of poverty reducing activities (i.e. joint commercialization, rotating loan funds, local public good provision); attendance of members; leaders’ evaluation; members’ satisfaction; and the degree to which organizational experience was emulated. He related this index to proxy measures of collective action determinants, namely: gender – an index of degree of women’s involvement in the organization outside help – measured by the number of training courses conducted by higher level peasant organizations and undergone by lower level organizations social capital – an aggregated subjective index of frequency of informal interaction relative size – measured by the fraction of member households in total households of a community age of the organization initiative – a dummy variable for whether an organization is a community initiative tenure – measure of the stability of leadership, i.e. number of years of tenure of current group of officials welfare – proxy of well-being of the community, uses the median size of landholding adjusted by the average land price within a community heterogeneity - measured by the Gini ratio of landholding at the community level He finds that the significant monotonic determinants of collective performance are: social capital, gender, and community welfare. Only the last has a negative sign, i.e. organizations from poorer communities tend to be more successful. Meanwhile, outside help appears to be in an inverted-U relationship to collective performance. That is, at an initial range these factors promote collective performance, but eventually a saturation point is reached, from which further increases in outside help adversely affect collective performance. Interestingly, Molinas finds that heterogeneity (proxied by the Gini ratio of landholding) is also in an inverse-U relation to collective performance. The foregoing discusses the collective action problem in general. The ensuing discussion sets the problem in the context of two major areas of collective action in agriculture, that of irrigation and credit delivery.

10

Collective action in communal irrigation The problem associated with managing and maintaining an irrigation system is wellknown; water delivery is too costly to meter and price individually, hence some collective mechanism must be enforced to prevent excess consumption by some users (especially those upstream) as well as to maintain the facility (to reverse the effects of normal wear, to prevent weed growth, to repair dams and canals in the event of flood, etc.) Monitoring and maintenance generates group benefits, opening up opportunities for free riding among members. In many developing countries, centralized administration has given way to management by user associations. A prominent example of successful cooperation between parastatal and user associations is that of Taiwan; Lam (1996) attributes this to the parastatal’s decision to set up working stations to act as intermediaries with the user associations. In practice, the working stations acted as the de facto authority at the field, allowing timely decisions to be taken in response to contingencies. Delegation to the working station entails softening hierarchy lines, hence the field staff were compelled to develop good will relations with the user associations in order to maintain a high performance level. Experience in other countries, especially in Southeast Asia, has been far more uneven. Reliable empirical regularities relating success or failure are hard to identify. For Philippine irrigation, the most extensive empirical study has been conducted by Fujita et. al. (2001). Their study focuses on 46 IAs in national irrigation systems found in Region 4. First they measure cooperation using a composite index obtained by principal components analysis. The index incorporates the following: • • • •

collective work for cleaning canals and laterals; coordination in rice cropping schedules; practice of water rotation; and organized monitoring on cropping schedule and/or water rotation.

They then run a regression analysis to identify factors significantly associated with cooperation. It is found that collective action is difficult to organize where: • • • • • •

water supply is uniformly abundant; water supply is greatly different between upper and lower streams in lateral size of association is large population density is low the ratio of non-farm households is high; farmers have a background of rainfed farming and little history of irrigated farming

These findings are consistent with the institutional economics of collective action, namely, that collective action is easiest in: small groups; where benefits from organization is high; where density of interaction is high; where exit options are constrained; and where social capital has been built up by community experience. 11

Collective action in credit delivery Another area in which group cooperation shows promise is in credit. The benefits of group action can be explained in relation to the market failures that prevent efficient credit delivery. A credit transaction hinges on a commitment made by a borrower; the quality of this commitment is subject to uncertainty owing to weak enforcement and asymmetric information. Repayment problems fall under two major categories, namely moral hazard and adverse selection. Moral hazard arises when ex post a borrower opts to default, or neglects to ensure the profitability of the funded project. Adverse selection meanwhile refers to the increase in average risk in the pool of potential borrowers as interest rates rise to take into account the average default risk. Under heterogeneity of default risk, charging the average default risk may lead to the exit of low risk borrowers and therefore higher average default risk (the “lemons” problem). At the extreme, interest rates may escalate to levels that leave only sure defaulters inside the market. This eliminates the credit market entirely. Institutional adaptations and practices have evolved in response to these problems. Lenders can screen their borrowers in advance, monitor their borrowers’ behavior, and verify default claims. They can ration credit by limiting loan sizes. Dynamic incentives can be offered by allowing loan sizes to expand over time, while threatening to cut off a credit line in the event of a default. In the formal sector, a common practice is to require collateral. Joint liability lending, which introduces the collective element in finance, is another means to deal with the moral hazard and adverse selection problem, Consider a bank that lends to the rural poor. Instead of lending to individuals, they may lend to a group, which then channels the funds to the members. The group as a unit assumes responsibility for repaying the entire loan. The arrangement clearly shifts transaction costs from the lender to the group members. There is however good reason to suppose that total transaction cost may fall3 as the joint liability arrangement provides and incentive for members to exercise peer monitoring. Typically, group members reside in the same community, hence information concerning individual risk type and economic behavior is readily available. Moreover, peer disapproval can be a strong deterrent against naked opportunism. Joint liability with contract enforceability at the peer level solves the moral hazard problem (Stiglitz, 1990). Recent models have tried to explain how group lending addresses not just moral hazard, but also adverse selection. Unlike peer monitoring models, in which organized groups are given, adverse selection models are able to account for group formation. Moreover, the peer selection model is simpler than that of peer monitoring, as the joint liability scheme need not require a technology for monitoring and enforcing contracts at the intragroup level. Ghatak (2000), considers the following set-up: suppose there are two risk types 3

“Economies of scale in lending” is often cited as a reason for organizing borrowers into groups; however, as lending must still occur within a group, ultimately scale economies must be attributed to better loan processing and monitoring technologies within a group.

12

(safe and risky), and borrowers know each other’s risk types though this is unknown to the lender. Peer group formation involves, at equilibrium, a sorting of groups (of pairs or partners) into those composed of only safe borrowers, and those composed of only risky borrowers. Ghatak shows that, compared to the without-group scenario, group formation can induce safe borrowers to re-enter. De Aghion and Gollier (2000) offer an even more general model in which borrowers need not be aware of each others’ risk types. Suppose there are only two types of project – a safe one, with zero probability of zero returns, and a risky one, with positive probability p of zero returns (loss) and a probability of 1 - p of a fixed yield. However, confirming that indeed a loss has occurred requires a positive verification cost. Each borrower borrows a fixed amount for one type of project. Suppose banks insist that loans only be extended to pairs, each pair being jointly liable for a loan. This arrangement deals with the adverse selection problem in two ways: First, it decreases verification cost, as loss claims are transferred from individuals to pairs; second, joint liability implies an ex post subsidy from a safe to a risky project in the event of loss, hence functions as ex ante collateral. Interest rate increases can therefore be moderated (as long as returns from safe projects are not large enough to repay the loss of a risky project). Joint liability lending is a scheme to channel funds from external sources. Meanwhile savings and loan associations (e.g. credit unions and cooperatives) provide intermediation services for funds internal to the community. Of course, membership in such associations must be sizable, curtailing reliance on small group lending (unless such a strategy is itself adopted by the association). Barham and Boucher (1996) have pointed out that for such savings and loan associations, the incentive for net depositors to monitor net borrowers is even stronger than in the case of pure external finance. Moreover, the obligation to repay is reinforced by the imperative to maintain a pool of funds which all association members find valuable for its insurance-like protection. The microfinance movement, fueled by popular accounts of successful microcredit schemes, is currently sweeping poverty alleviation programs worldwide. The movement was pioneered by the Grameen Bank in Bangladesh, which introduced very small loan sizes, joint liability lending, frequent collection, mandatory savings, preference for female borrowers, and values training. Repayment rates are reportedly much higher, while cost recovery more feasible, compared to traditional direct credit systems. However, evidence for or against group-based lending is far from clear-cut. Huppi and Feder (1990) examined the performance of such schemes and identified several factors that seem correlated to repayment rates, namely: small group size; homogeneity of the borrowing groups; previous experience in borrowing; and finally, specialization of the group in credit (i.e. the group itself was not engaged in other joint commercial activities, as is common in multipurpose agricultural cooperatives). Interestingly, for the Philippines (as well as India) the cost of borrowing was not lower for group-based lending, contrary to the trend in other developing countries.

13

The finding on group size is echoed by Ghatak and Guinnane (1999). They note that in the Grameen program, the group size is typically set at five, a figure that was reached by trial and error. On the negative side, experiences in Nepal have shown that mutual trust is meager in larger groups (20 or more). Moreover, the origin of the group matters; groups formed by program officials tend fall under the delinquent category. Morduch (1999) cites a survey of 72 microfinance programs worldwide with professed commitment to financial sustainability; contrary to impression, financial performance of programs lending to individuals are similar to those lending to solidarity groups. This finding certainly questions the necessity of group-based lending, though of course does not detract from its value in numerous settings. Morduch’s survey also found some interesting facts about sustainability. The covered programs are generally characterized as operationally sustainable (with a few exceptions). However, incorporating the opportunity cost of capital, the programs typically remain subsidized (by around 17% on average). Moreover, the subsidy tends to be equally dispersed across beneficiaries and not concentrated upon the poorest clients. The limited bias towards the poorest is similarly found in Guatemalan credit unions, which have provided measurable relief from binding credit constraints. However, the poorest individuals typically did not benefit from increased access to credit (Barham and Boucher, 1996). Morduch also finds that estimates of net welfare impacts are scanty. In general his review finds that empirical regularities regarding group-based lending (in terms of success factors and impacts) have yet to be developed. Much of the evidence offered remains anecdotal. Likewise, Ghatak and Guinnane (1999) concede that joint liability lending is correlated with improved repayment, but most successful programs have other features that could account for this correlation. No study apportions reasons for success. Needless to say, research isolating these contributory factors would be of great interest and value to policy. Aside from credit support to agriculture, CBAs have been recommended for other agribusiness activities such as processing and trade. Collective action in these areas is closely associated with the most formal and commercialized form of community organization, which is the cooperative. This form is treated in the next section.

5. The cooperative as a community-based approach to agribusiness organization The fundamental difference between the cooperative and the standard firm is the ownership structure (Sykuta and Cook, 2001): while the former may be regarded as a patron-owned firm (POF), the latter may be referred to as an investor-owned firm (IOF). In the IOF, decisions are made by investors, towards the joint maximization of the residual claim. As a background to the POF, the discussion will deal with the IOF structure in some detail.

14

Theory of the Investor-Owned Firm In the elementary theory of the IOF, the firm appears as a monolithic entity with the sole objective of maximizing profit. Accordingly, input and output choice is determined at the equality of marginal revenue and marginal cost; competitive equilibrium implies equality of price and marginal cost. Allocation can be inferred solely from the production function and prices; the only function of ownership structure is to inform the distribution of the surplus. Early models of the IOF pinpointed the locus of monolithic choice in the abstract ownerproprietor, or entrepreneur. In a seminal paper, Coase (1937) identified transaction coast as the key consideration in identifies the boundaries of the firm (i.e. which activities to undertake internally, and which to be purchased as input or service from the market). However, inasmuch as the allocative function can itself be discharged by a manager hired from the labor market, it was not clear why control needed to be associated with residual claim or asset ownership. Models which recognize the fact that IOF is really a set contracts between individuals (rather than a monolithic maximizing entity) are needed to understand its ownership structure. Alchian and Demsetz (1972) used team theory to arrive at an agency explanation for the residual claim. Consider a production team whose technology is subject to economies of scale and scope. This makes the marginal product of each team member’s effort difficult to measure; hence, each member has the incentive to shirk. Suppose though that one team member can specialize in metering marginal product, thus aligning effort with rewards. This monitoring task itself is vulnerable to shirking on the part of the monitor; to remedy this, the monitor may have to be paid a residual claim. The residual claim meanwhile is more credible if supported by actual control over decision-making; in a private property system, the owner is given the right to decide whom to give access to a resource. This confers decision-making powers on the asset owners. Hence when asset ownership is held by the residual claimant, who is also responsible for monitoring, then all incentives are properly aligned to address the inefficiency of free riding. In practice, while the ultimate monitors are the firm’s owners, daily monitoring tasks must be delegated to managers, which adds another layer to the principal-agent relation. This management layer commands the most attention, particularly in the corporate governance literature. A common theme is that the equities market itself is device to ensure productive use of an IOF’s assets, by ensuring flexibility of ownership, and providing a means to value the firm. Shareholders incapable of proper selection and monitoring of managers can be bought out; at the extreme, management can be revamped as a consequence of a “hostile takeover” (Shliefer and Vishney, 1996). Free riding behavior is not the only problem dealt with by investor ownership. Another form of opportunistic behavior is the “holdup” problem, which arises under conditions of asset specificity (Klein et. al., 1978). In many production processes, assets become specific to a combination or pool if the value of the pool exceeds the sum of individual asset values. Under divided ownership, any party can bargain for a greater share in the

15

surplus on the threat of withdrawal from the combination. Transaction cost theory predicts that when holdup is a serious problem, assets will tend to fall under common ownership (Williamson, 1985). Early models of the Patron-Owned Firm Early models of the POF parallel those of the IOF in adopting a monolith assumption. these models, the major preoccupation was to identify the objectives of the cooperative, as well as allocative implications of each objective (e.g. Coterill, 1987). Statements of the objective are stated in terms of maximizing one of the following: • • • •

Producer surplus; Patronage refund, or consumer surplus (Helmberger and Hoose, 1962). Combination of producer surplus and patronage refund (Enke, 1945) Dividend (Ward, 1958).

Allocative outcomes are easily determined from these objectives. Maximization of producer surplus entails the standard MR = MC condition; maximization of consumer surplus (subject to a break-even constraint) implies equality of price with average cost; and maximization of producer and consumer surplus entails the P = MC outcome. Meanwhile if the maximand is the dividend, some algebra is required to work out the implications: let the production function therefore be Q ( M ) , where Q is output and M is the only variable input, measured by the number of members . Let Q increase with M at a decreasing rate. Suppose output price is fixed at P, and the wage w is fixed at w. There is a positive fixed cost F. The profit-maximizing firm would therefore set value of marginal product equal to wage. In contrast, the dividend maximizing firm would select M so as to equalize the value of marginal product with the average revenue (net of fixed cost): P ⋅ Q '( M ) = ( P ⋅ Q − F ) M . Interestingly, strict concavity of the production function implies a supply curve that is downward sloping.4 This conjecture, while interesting, remains largely unsubstantiated by evidence (Bonin et. al., 1993). As these models posit potential departures from profit maximization, they tend to suggest that cooperatives perform no better than and may even be worse than IOFs in terms of efficiency. However, studies on the technology and performance of cooperatives fail to produce evidence for this view (Sexton and Iskow, 1993). Hind (1994) also compares agricultural cooperatives in UK with IOF in agribusiness: on the basis of profitability, capital gearing, liquidity, sales/working capital ratios, and creditor ratios, there were no significant differences between the two type of firms.

Cooperatives and market structure Total differentiation yields (dM dP ) ⋅ P ⋅ Q "( M ) ⋅ M + ( Q '( M ) ⋅ M − Q ) = 0 . Since the term in parenthesis is negative, then dM dP must also be negative.

4

16

The implications of the various cooperative objectives are sharpened by examining the implications of market structure. Under short run perfect competition, profit maximization is equivalent to maximizing the sum of producer and consumer surplus; however consumer surplus maximization implies inefficient oversupply of the output, whereas dividend maximization entails undersupply. Under long run perfect competition though, all four objectives enumerated above lead to the efficient equilibrium at which price equals long run average cost. It is in the case of imperfect competition that the contribution of the cooperative to efficiency can be clearly noted (Normark, 1996). In the case of a monopoly, a cooperative pursuing total surplus maximization would clearly be more efficient than a profit-maximizing IOF. Cooperatives in direct competion with at least one large IOF generates an mixed oligopoly or oligopsony, a setting which greatly complicates the analysis. Sexton and Sexton (1987) examine the effect of potential entry of a supplier cooperative on a market; they find that the effect differs markedly from that of potential entry by an IOF. Potential cooperatives therefore exert a powerful role in beneficially regulating free market performance. Tennbakk (1995) shows that total welfare is higher in a mixed duopoly with involving cooperative compared to a pure duopoly consisting only of IOFs. An implication of the imperfect competition argument is that, in the long run, erosion of imperfect competition would render cooperatives superfluous for efficient resource allocation. However, even in developed economies, instances of imperfect competition in agricultural markets abound. In the US market, food processing and marketing has in fact exhibited a sharp trend towards consolidation (Sexton, 2000). In many agricultural markets, raw products are costly to transport, and processors are sparsely distributed. The persistence of market power would then support the long term contribution the cooperative to agricultural efficiency. Cooperatives and the new institutional economics So far the foregoing models all apply the monolithic assumption in analyzing cooperatives. The “new institutional economics” enriches the analysis by examining the contractual arrangements among the individuals comprising the cooperative. The “incomplete contracts” approach of Hart and Moore (1990) contains an in-depth theoretical analysis of the IOF, which they extend to the case of the POF. As in the transaction cost model, asset specificity is posited; in addition, they allow for investments that can further increase the surplus from an asset combination, although such prior investments are by assumption noncontractible (e.g. human capital investments). The structure of ownership affects the incentive to provide these investments; the theory predicts that the observed ownership structure would conform to the optimal incentive design. The solution concept for ex post bargaining over surplus (“division of spoils”) is the Shapely value.5 The ex ante investment decisions are determined as a Nash equilibrium of investment strategies. 5

This solution concept roughly corresponds to the average of the marginal contribution of each agent to the value of the asset combination.

17

The case of the POF can be explained as follows: Consider one asset, a set of S individuals, and actions or investments taken by each individual to increase total benefits available to the entire set. (Assume that no action yields any benefit without the asset). We can form subsets or groups from S; suppose there exists a subset G such that combinations containing less than or equal to half of the members of G generate no marginal benefits from investing over all individuals. Then G constitutes a key group. Their theory predicts that such a key group would own the asset. Hence, in the case of a supplier cooperative, a key group would consist of the customers; in the case of a marketing cooperative, a key group would consist of the suppliers. The model however does not precisely identify instances in which a group of customers or suppliers might emerge as a key group. Moreover it is not easy to find examples of noncontractible prior investments for agricultural cooperatives, undermining the relevance of the Hart and Moore model. Sykuta and Cook (2001) posit the holdup problem as an explanation for the processing cooperative. They point out that, once agricultural output has been delivered, the division of the surplus from postharvest processing and marketing transforms into a zero sum game. The prospect of holdup introduces ex ante distrust between contracting parties. Vertical integration, with production and processing under common ownership, may solve the holdup problem. However, vertical integration may not be feasible due to diseconomies of scale in production, which may spill over into diseconomies of scope. An alternative structure would be the processing cooperative, in which small independent producers jointly own a downstream processor. Decision-making by majority vote in cooperatives is examined by Albaek and Schultz (1997) as well as Hendrikse (1998). The former discusses, in the context of agricultural cooperatives, whether majority voting is inferior to voting rules which favor large farmers (i.e. “one cow, one vote”. ) In a variety of cases the inferiority of majority voting is cannot clearly be demonstrated; indeed in the case of a large cooperative which finances investment out of retained earnings, the majority rule yields efficient decisions. The latter paper by Hendrikse meanwhile focuses on the transaction cost of the decisionmaking process in a cooperative. His model identifies a tradeoff between higher transaction cost of cooperative and better “quality” of decision-making. More novel approaches to account for the cooperative is that nonlinear pricing and signaling. Vercammen, Fulton, and Hyde (1996) posit a purchasing cooperative as a means to practice nonlinear pricing through a scheme of membership fees and rebates. The efficient pricing rule, which is constrained by asymmetric information and membership heterogeneity, entails a higher average price for higher delivery volumes. The pricing rule can be modified to reflect equitable distribution of benefits across members. Another price discrimination scheme is found in Hart and Moore (1998). The supplier cooperative is modeled as a mechanism for charging different prices for insiders

18

(members) and outsiders (nonmembers). Consider a group with M members, attempting to provide each with one unit of a good. Fixed cost of provision is F > 0, but marginal cost of provision is zero. The good can be purchased outside the group at a competitive price p* ; by assumption, F M < p* , i.e. the average cost price is below the competitive price (the group earns some rent). Members pay a nonnegative transaction cost for purchasing this good from outside. Members value the good heterogeneously, and this valuation is private knowledge. An IOF which acts as a supplier must choose between a competitive strategy and a monopolistic strategy. The former entails setting a price p* for all consumers, while the latter involves selling only to insiders at a price higher than p* , which is possible when transaction cost is strictly positive. The monopolistic strategy is inefficient due to the excessive exclusion of consumers. Alternatively the producer can be organized as a supplier cooperative, which is stylized as a nonprofit entity adhering to the break-even condition. Then price charged to outsiders and insiders can be differentiated. Insiders are charged the average cost price, while excess capacity can be eliminated by sales to outsiders at the competitive price. Due to the low inside price, the POF ‘s inefficiency results from excessive inclusion of consumers. On the balance it is not clear whether inefficient exclusion (a possibility for the IOF) is a greater distortion than inefficient exclusion (a certainty for the POF). Two factors would favor the superiority of the POF: first is a greater departure from perfect competition; second is a greater similarity in members’ benefits, i.e. a more homogenous membership profile. Meanwhile the signaling explanation is taken up by Albaek and Schultz (1998). Consider a farmers' dairy cooperative that competes against a profit-maximizing firm. The cooperative maximizes the total profit of itself and its suppliers. Members of the cooperative individually decide how much to supply; this turns out to be a commitment device for credibly gaining a large market share. This device can enable it to realize higher profit per farmer than would the profit maximizing firm. Signaling could also explain adherence to a nonprofit status. Glaeser and Schliefer (2000) present a model of the nonprofit firm, whose activities have an important quality dimension that are essentially unobserved. A for-profit status is a signal that the firm faces high-powered incentives for quantity provision, and relatively weaker incentives for quality provision. A nonprofit status implies low-power incentives for quantity provision, thus signaling better quality provision. Hence, a nonprofit firm may gain a competitive advantage when customers, employees, financiers, or donors feel protected by its nonprofit status. This explanation could also account for the emphasis on members’ education in much of the cooperative literature. Suppose donors or government seeks to finance quality training of farmers in entrepreneurship, cooperation, community action, and civic consciousness. 19

Then nonprofit cooperatives may be more reliable suppliers of high quality training, compared to for-profit entities such as agribusiness firms, and therefore command higher amounts of financing.

6. Policy Implications The arguments in favor of CBAs developed over the last three section seem to support a policies for supporting and even subsidizing CBAs. There are however also good reasons for exercising caution in extending such subsidies, mainly due to governance breakdown, cooperation breakdown, and participation breakdown, to which CBAs are often prone. Governance breakdown While the cooperative can resolve some organization problems (as described in Section 5), it faces drawbacks of its own. Cook (1995) suggests the headings of portfolio, control, and influence cost as categories of these problems. The portfolio problem arises when a cooperative’s selection of assets is inconsistent with the interests or risk attitudes of any given member; unlike for an IOF, the investor cannot easily withdraw her individual share of the asset pool of the cooperative. Meanwhile the control problem is a general difficulty inherent in agency relations, manifesting itself most sharply in the monitoring of managers; for cooperatives however the control problem is exacerbated by the absence of external pressure applied by the equities market. Finally, influence costs arise when members or member coalitions expend effort and resources to sway organizational decisions towards their own interest, possibly to the detriment of other members. The Hart and Moore (1998) paper also analyzes governance problems associated with majority rule in a cooperative. As usual this rule faces a median voter problem. For the price discriminating cooperative, consider the following: two types goods can be sold, conditional on investment outlay; the quality of the alternatives as well as the investment requirements differ. [S]uppose that, whatever the investment decision, a majority of members have small payoff (net of cost), but they are slightly better off if the investment does not go ahead. And suppose the minority consists of members who stand to gain a great deal if the investment does go ahead. In such circumstances, the efficient decision is clearly to invest, but this will be blocked by the majority. Moreover, any attempt by the minority to bribe the majority into changing their minds is thwarted by free-riding on the part of individual members of the minority: the asymmetry of information (combined with large numbers) means that no individual can be forced to contribute to the bribe (p. 8).

Banerjee (2001) argues for the rent-seeking model of the cooperative, in which influence cost is a dominant factor in cooperative decisions. Within cooperatives, the elite minority group typically has an edge in mobilizing collective action, as it faces smaller transaction cost and a less serious free rider problem. This model is applied ot the case of sugar cooperatives in Maharashtra in 1971-1993. The elite is extracts rent by depressing the purchase price and diverting the retained earnings of the cooperative. The study confirms

20

the prediction that in areas characterized by heterogenous asset holdings (as proxied by land size), prices would tend to be lower and processing capacity more limited. Cooperation breakdown Cooperation between community members can break down for a variety of reasons. In the area of credit, for instance, joint liability lending may simply shift part of the moral hazard problem onto co-borrowers; opportunistic members may free-ride on the group’s access to credit while reneging on their obligations. Close community ties could work against efficiency as opportunists bank on peer leniency. Corrupt or weak group leaders can divert funds, or coddle irresponsible members. Third, the community-based finance would tend have a highly concentrated loan portfolio due to similarities in economic activities of the members. This would therefore limit risk diversification, render portfolio returns vulnerable to covariate shock, and invite liquidity problems as cash inflows and outflows of members tend to be synchronized (Huppi and Feder, 1993). In the Philippine experience, cooperative repayment performance has been far from exemplary. The LBP for one faces a mounting challenge to maintain financial soundness while sustaining its cooperative finance thrust, due to the repayment problems encountered in the latter (Briones, 2001). On a positive note, state support can make a difference in preventing cooperative breakdown. In irrigation for instance, the Fujita et. al. (2001) study includes two policy variables in the regression. The first is the provision of special incentives by the local NIA; the second is the presence of dedicated and proficient local NIA staff. The incentives take the form of subsidies and resource sharing in maintenance activity. To detect the presence of excellent local staff, the interview method relied on a unanimous identification by IA leaders and farmers (agreement was observed for around 15% of the respondent IAs). These policy variables are found to be statistically significant factors for promoting cooperation. Participation breakdown Empirical and theoretical analysis of participation presupposes that instances of participation can be identified. This turns out to be a nontrivial exercise; considerable anthropological and sociological analysis has gone into verifying the authenticity of claimed participatory processes (Castillo, 1978). A critical source of participation breakdown is the maneuverings of the local elite. Abraham and Platteu (forthcoming) enumerate various reasons for frustration of participatory processes due to intervention by the local elite: 1. In traditional communities, the villagers and their leaders may feel threatened by th potential disruption of existing socioeconomic hierarchies due to development projects. Participatory approaches also conflict with traditional patterns of authority and status. Hence the traditional elites may resist its introduction, or redistributive pressures deleterious to project efficiency.

21

2. Meanwhile in societies where markets are well entrenched and elites dominate both economically and politically, project benefits become prone to elite capture. For example, cost-sharing arrangement are common to many projects, which the elite can easily shoulder to project design and beneficiary selection towards their own preferences. Organizational efforts incur transaction cost which are charged to community members; often it is only the members of the elite that can bear these costs. This jibes with Michener’s (1998) observation that “participation” in practice entails the active work of only a few community residents, typically in leadership positions. 3. Local elite can assume the role of community benefactors, partly due to their extensive networks with government officials and funding agencies. In so doing they are able to deflect participatory processes away from their original objectives. NGOs are organized for the most cynical purposes reasons, e.g. fund raising, diversion, and employment of downsized bureaucrats. For these reasons, Abrahams and Platteau argue against a rapid rollout program based on massive donor financing of a participation-based strategy. Authentic participatory processes require considerable investments of time and effort, which should be factored in when designing community projects. Implications for policy Theory and evidence in favor of public support for CBAs is ambiguous. The strongest case can be made for CBAs in the provision of local public goods, such as irrigation; the case becomes progressively weaker in the areas of financial intermediation and especially in agribusiness activities. For these gray areas therefore, the best recommendation appears to be a targeted strategy of promoting CBAs. The credit review suggests an adherence to microfinance principles, including group lending; in the absence of evidence to the contrary, small groups should be preferred over large groups, which implies deemphasizing cooperatives as conduits of finance. Support for trade and processing cooperatives should be selective of underdeveloped areas, where markets are fragmented and perfect competition unlikely to be approximated. Needless to say, any financial support should comply with minimum standards, such as studies of feasibility, allowance for risk (especially risk that is covariate across farmers), etc. The most significant policy stance towards CBAs, and cooperatives in particular, are the list of tax exemption privileges. These privileges are non-targeted and non-selective, and may therefore seem to be a candidate for phasing out. One must however be cautious of making these adjustments, in view of probable strong resistance from farmer organizations and their lobbyists. Priorities in term of subsidy phaseout should also be kept in mind: the financial burden of fiscal incentives for export-oriented industries probably dwarf by an order of magnitude the implicit subsidy to the cooperative sector; hence, the any review of subsidy policy should take up this sector last.

22

6. Concluding Remarks This survey began by noting the need for a rigorous framework of analyzing CBAs, as well as a compilation of available facts about the contribution of CBAs to community development. The review finds that the shift from centralized participatory and community-based strategies, far from being a mere fad, has a solid conceptual and empirical basis. An obvious efficiency argument for CBAs is the provision of local public goods. However even for commercial activities, CBAs may be beneficial in the presence of market failures associated with imperfect competition, asymmetric information, and imperfect contract enforcement. Joint liability lending and the cooperative form may arise in response to these failures. A solid factual basis is being laid for the development role of CBAs. The evidence though seems weakest for agribusiness activities such as trade and processing, as it cannot be clearly proven that a CBA counterpart of a private sector firm performs as efficiently. The evidence grows progressively stronger for activities in which private sector supply faces serious gaps (such as rural finance), and is most convincing for the case of local public goods such as irrigation. These provide the groundwork for building a policy framework regarding the promotion of CBAs. While this study cannot prescribe the specifics of that framework, minimal elements can be spelled out as follows: First, public support should be sensitive to the every-present possibility of breakdowns in cooperation, participation, and governance. Engaging community members in participatory processes should therefore not be taken for granted. Second, support should not be extended in a blanket manner, but should be selective and targeted to cases in which neither market nor command-and-control mechanisms are deemed workable. Precisely what and where these cases are, should be the object of further research in this important and exciting field of institutional analysis.

References Abraham, A. and Platteu, J. (forthcoming). “Participatory development: where culture creeps in.” Culture and Public Action: A Cross-Disciplinary Dialogue on Development Policy. V. Rao and M. Walton, eds. Stanford University Press. Adato, M., and L. Haddad (2002). “Targeting poverty through community-based public works programs: experience from South Africa.” Journal of Development Studies 38(3):1-36. Aggarwal, R. (2000). “Possibilities and limitations to cooperation in small groups: the case of group-owned wells in Southern India.” World Development 28(8):148197.

23

Aghion, B. and C. Gollier (2000). “Peer group formation in an adverse selection model.” Economic Journal 110:632-43. Albaek, S. and C. Shultz (1997). “One cow, one vote?” Scandinavian Journal of Economics 99(4):597-615. Albaek, S. and C. Shultz (1998). “On the relative advantage of cooperatives.” Economics Letters 59(3):397-401. Alchian, A., and H. Demsetz (1972). “Production, information costs, and economic organization.” American Economic Review 62(5):777-95. Alkire, S., A. Bebbington, T. Esmail, E. Ostrom, M. Polski, A. Ryan, J. Van Domelen, W. Wakeman, and P. Dongier. “Community-Driven Development.” World Bank: mimeo. Banerjee, A. et. al. (2001). “Inequality, control rights, and rent seeking: sugar cooperatives in Maharashtra.” Journal of Political Economy 109(1):138-90. Barham, P. and S. Boucher (1996). “Credit constraints, credit unions, and small-scale producers in Guatemala.” World Development 24(5):793-806. Bebbington, A. (1997). “New States, new NGOs? Crises and Transitions among Rural Development NGOs in the Andean Region.” World Development 25(11): 17551765. Bierschenk, T., J.P.O. de Sardan, and J.P. Chauveau, eds. (2000). Courtiers en développement _Les villages africains en quête de projets, Paris : Karthala. Bonin, J., D. Jones, and L. Putterman (1993). “Theoretical and empirical studies of producer cooperatives: will the twain ever meet?” Journal of Economic Literature 31(3):1290-320. Bolton, P. and D. Scharfstein (1998). “Corporate finance, the theory of the firm, and organizations.” Journal of Economic Perspectives 12(4):93-114. Briones, R. (2001). “How will farmers’ borrowing respond to increases in the crop insurance premium? The credit demand of small rice farmers in the Philippines.” Report submitted to the Agricultural Credit Policy Council. Castillo, G. (1978). How Participatory is Participatory Development? Castillo, E., J. Baskiñas, W. Medina, A. Albano, A. Peria, and A. Manilla (2003). “Cooperativism in agriculture: the case of top four cooperatives in Region IV, Philippines.” Philippine APEC Study Center Network Discussion Paper No. 2003-01.

24

Castillo, M. and M. Casuga (1999). “Small farmer credit delivery after the 1980s reforms.” Proceedings of the ACPC Rural Finance Symposium 18 October 1999. Conning, J., and M. Kevane (1999). “Community Based Targeting Mechanisms for Social Safety Nets”. Mimeo. Department of Economics, Santa Clara University, Cal., USA. Coase, R. (1937). “The nature of the firm.” Economica 4:386-405. Cotteril, R. (1987). “Agricultural cooperatives: a unified theory of pricing, finance, and investment.” Cooperative Theory: New Approaches. J. Royer, ed. Cooperative management Division, Agricultural Cooperative Series Report 18. United States Department of Agriculture. Cook, M. (1995). “The future of U.S. agricultural cooperatives: a neo-institutional approach.” American Journal of Agricultural Economics 77(5):1153-59. Cruz, W., and M. Cruz (1984). “Nonseparable externality, collective action, and irrigation institutions.” Philippine Economic Journal 23(2-3):235-248. De Aghion, B., and C. Gollier (2000). “Peer group formation in an adverse selection model.” Economic Journal 110: 632-643. Dorward, A. (2001). “The effects of transaction costs, power and risk on contractual arrangements: a conceptual framework for quantitative analysis.” Journal of Agricultural Economics 52(2):59-74. Enke, Stephen (1945) "Consumer Coöperatives and Economic Efficiency." American Economic Review 35:148–55. Fudenberg, D. and E. Maskin. (1986). “The folk theorem in repeated games with discounting or with incomplete information.” Econometrica 54:533-554. Fujita, M., Y. Hayami, and M. Kikuchi (1999). “The Conditions of Collective Action for Local Commons Management : The Case of Irrigation in the Philippines.” Aoyama-Gakuin University, Tokyo, mimeo. Glaeser, E. and A. Shleifer (2001). “Not for profit entrepreneurs.” Journal of Public Economics 81(1):99-116. Ghatak, M. (2000). “Screening by the company you keep: joint liability lending and the peer selection effect.” Economic Journal 110:601-631. Ghatak, M. and T. Guinnane (1999). “The economics of lending with joint liability: theory and practice.” Journal of Development Economics 60(1):1-34.

25

Grootaert, Christiaan and Deepa Narayan (2000). “Local Institutions, Poverty and Household Welfare in Bolivia.” Local Level Institutions Working Paper 9. Social Development Department, World Bank, Washington, D.C. Grootaert, Christiaan (1999b). “Social Capital, Household Welfare and Poverty in Indonesia. Local Level Institutions Working Paper 6. Social Development Department, World Bank, Washington, D.C. Grossman, S. and O. Hart (1986). “The costs and benefits of ownership: a theory of vertical and lateral integration.” Journal of Political Economy 94(4):691-719. Hart, O. and J. Moore (1990). “Property rights and the nature of the firm.” Journal of Political Economy 98(6):1119-1158. Hart, O. and J. Moore (1998). “Cooperatives and outside ownership.” Working Paper 6421. National Bureau of Economic Research: Cambridge, MA. Helmberger, Peter, and Sidney Hoos (1962). "Cooperative Enterprise and Organization Theory." Journal of Farm Economics 44:275–90. Hendrikse, G. (1998). “Screening, competition, and the choice of the cooperative as an organizational form.” Journal of Agricultural Economics 49(2):202-17. Hendrikse, G., and CP. Veerman (2001). “Marketing cooperatives and financial structure: a transaction costs economics analysis.” Agricultural Economics (26)3 : 205-216. Hind, A. (1994). “Cooperatives-Under Performers by Nature? An Exploratory Analysis of Cooperative and Non-cooperative Companies in the Agri-business Sector.” Journal of Agricultural Economics 45(2): 213-19. Hoddinott, J., M. Adato, T. Besley, and L. Haddad (2001). “Participation and poverty reduction: issues, theory, and new evidence from South Africa.” Food and Consumption Nutrition Division Discussion Paper 98. International Food Policy Research Institute: Washington, D.C. Hossain, M. (1988). “Credit for the alleviation of rural poverty: the Grameen Bank in Bangladesh.” Research Report 65, International Food Policy Research Institute. Huppi, M. and G. Feder (1990). “The role of groups and credit cooperatives in rural lending.” World Bank Research Observer 5(2):187-204. Isham, J., and S. Kahkonen. (1999). “What determines the effectiveness of communitybased water projects? Evidence from Central Java, Indonesia, on demand responsiveness, service rules and social capital.” World Bank, Washington, D.C.

26

Isham, J., D. Narayan, and L. Pritchett (1995). “Does participation improve performance? Establishing causality with subjective data.” World Bank Economic Review 9 (2):175-200. Jimenez, E. and V. Paqueo (1996). “Do Local Contributions Affect the Efficiency of Public Primary Schools?” Economics of Education Review 15(4): 377-86. Jimenez, E. and Y. Sawada (1998). “Do Community-Managed Schools Work? An Evaluation of El Salvador’s Educo Program.” Impact Evaluation of Education Reforms Paper 8. Development Research Group, World Bank, Washington, D.C. King, E., and B. Ozler (1998). “What’s decentralization got to do with learning? The case of Nicaragua’s school autonomy reform.” Paper presented at the Northeast Universities Development Economics Consortium Annual Conference, Yale University, New Haven, Conn., U.S.A. Klein, B., R. Crawford, A. Alchian (1978). “Vertical integration, appropriable rents, and the competitive contracting process.” Journal of Law and Economics 21:297-326. Krahnen, J., and R. Schmidt (1995). “On the theory of credit cooperatives: Equity and onlending in a multi-tier system – A concept paper.” Enterprise and Cooperative Development Department Poverty-oriented Banking Working Paper No. 11. Geneva: International Labour Office. Korten, F. and R. Siy , eds. (1989). Transforming a Bureacracy: The Experience of the National Irrigation Administration. Kumarian Press: West Hartford, CN. Lam, W. (1996). “Institutional design of public agencies and coproduction: a study of irrigation associations in Taiwan.” World Development 24(6):1039-54. Lam, W. (1998). Governing Irrigation Systems in Nepal. Institutions, Infrastructure and Collective Action. Oakland, CA: ICS Press. Lamberte, M. (2001). “Expanding banking services to micro, small, and medium enterprises and poor households in the Philippines.” Discussion Paper No. 200124. Philippine Institute for Development Studies. Manikutty, S. (1998). “Community participation: Lessons from experiences in five water and sanitation projects in India.” Development Policy Review 16(3): 373-404. Meyer, C. (1995). “Opportunism and NGOs - Entrepreneurship and Green North-South Transfers.” World Development, 23(8):1277-1289. Michener, V. (1998). “The participatory approach: contradiction and co-option in Burkina Faso.” World Development 26(12): 2105-18.

27

Molinas, J. (1998). “The impact of inequality, gender, external assistance and social capital on local-level cooperation.” World Development 26(3):413-431. Morduch, J. (1999). “The microfinance promise.” Journal of Economic Literature 37(4):1569-614. Normark, P. (1996). “A role for cooperatives in the market economy.” Annals of Public and Cooperative Economics 67(3):429-39. Ostrom, E. and R. Gardner (1993). “Coping with asymmetries in the commons: selfgoverning irrigation systems can work.” Journal of Economic Perspectives 7(4):93-112. Ostergaard, Geoffrey, and A. H. Halsey (1965). Power in Cooperatives. Oxford, England: Basil Blackwell and Mott. Raby, N. (1991). “Participatory management in large irrigation systems: issues for consideration.” World Development 19(12):1767-76. Ravallion, M. (1999). “Is More Targeting Consistent with Less Spending?” International Tax and Public Finance 6(3): 411-19. Sara, J. , and T. Katz (1997). “Making Rural Water Sustainable: Report on the Impact of Project Rules.” United Nations Development Programme and World Bank Water and Sanitation Program, Washington, D.C. Schroeder, T. (1992). “Economies of scale and scope for agricultural supply and marketing cooperatives.” Review of Agricultural Economics 14(1):93-103. Sexton, R. (1986). “The formation of cooperatives: a game-theoretic approach with implications for cooperative finance, decision-making, and stability.” American Journal of Agricultural Economics 68(2):214-25. Sexton, R. (2000). “Industrialization and Consolidation in the US Food Sector: Implications for Competition and Welfare.” Waugh Lecture, American Agricultural Economics Association Annual Meeting, Tampa, Florida. Sexton, R., and T. Sexton (1987). “Cooperatives as Entrants.” Rand Journal of Economics 18(4):581-95. Sexton, R. (1990). “Competition in Agricultural Markets and the Role of Cooperatives: Spatial Analysis." American Journal of Agricultural Economics 72:709–20. Sexton, R., and J. Iskow (1993). “The competitive role of cooperatives in market-riented economies: a policy analysis.”Csaki,-Csaba; Kislev,-Yoav, eds. Agricultural

28

Cooperatives in Transition.. Boulder and Oxford: Westview Press in cooperation with Balaban, 1993, pages 55-83. Shliefer, A., and R. Vishny. (1996). “A Survey of Corporate Governance.” National Bureau of Economic Research Working Paper Series No. 5554. Sibal, J. (nd). “A Century of the Philippine Cooperative Movement.” www.wis.edu/info/abroad/sibal.html. Stiglitz, J. (1990). “Peer monitoring and credit markets.” World Bank Economic Review 4(3):351-66. Svendson, P. (1992). “Assessing effects of policy change on Philippine irrigation performance.” Working Papers on Irrigation Performance No. 2. International Food Policy Research Institute. Sykuta, M. and M. Cook (2001). “New institutional economics approach to contracts and cooperatives.” American Journal of Agricultural Economics 83(5):1273-1279. Tennbakk, B. (1995). “Marketing cooperatives in mixed duopolies.” Journal of Agricultural Economics 46(1):33-45. Vedeld, T. (1998). “Village politics: heterogeneity, leadership, and collective action.” Journal of Development Studies 36(5):105-34. Vercammen, J., M. Fulton, and C. Hyde. (1996). “Nonlinear pricing schemes for agricultural cooperatives.” American Journal of Agricultural Economics 78(3):572-84. Williamson, O. (1985). The Economic Institutions of Capitalism. New York: Free Press. Ward, B. (1958). “The firm in Illyria: market socialism.” American Economic Review 48: 566-589. Zeller, M., G. Schrieder, J. von Bran, and F. Heidhues (1997). “Rural finance for food security for the poor: implications for research and policy.” Food Policy Review No. 4. International Food Policy Research Institute.

29