Dayton-Johnson (2001) constructs a game-theoretic model of social cohesion, which he differentiates from “community”: The latter is based on shared values and shared interpretation of social reality, a stronger condition than social cohesion. These concepts must be viewed in the present context as exogenous in the first instance. As Agrawal (1999:103) notes, “The aspect of community that stands for shared understanding is precisely what external interventions can do very little about. States, NGOs [nongovernmental organizations], bureaucratic authorities, aid agencies, and policy makers cannot directly create community-as-shared-understanding.”


Recent empirical research on producer cooperatives in developing countries can be interpreted as part of the same research agenda. See the recent studies by Banerjee et al. (2001) and Seabright (1997). Similarly, empirical studies of people’s propensity to join voluntary organizations in Paraguay (Molinas, 1998) and rural Tanzania (La Ferrara, 1999) demonstrate a negative effect of economic inequality.


A much earlier quantitative study of irrigation systems, not considered in this chapter, was carried out by de los Reyes (1980).


The estimated coefficient on the Gini variable is negative and significant, while the estimated coefficient on the square of the Gini is positive and significant.


The estimated coefficient on the square of the Gini term was positive and significant in two of the three models in Dayton-Johnson (2000b), suggesting a strongly positive effect of inequality on maintenance. Nevertheless, inequality was significantly related to proportional water allocation; that rule, in turn, was associated with lower levels of maintenance. The full effect, direct plus indirect, of inequality was negative.


Bardhan (2000) includes a variable indicating whether an ayacut is located at the tail end of a larger system; an entire village of tail-enders, however, does not behave differently from other villages, all else equal.


Baker’s argument is more nuanced than that stated here: He claims that the effect of exit options is mediated not only by differentiation, but by reliance on the water source. Where reliance is high and differentiation is low, management regimes can withstand increased exit options. Nevertheless, our reading of his argument is that where differentiation is high, regardless of the level of reliance, the stress on the institutions of governance is critical.


In the Mexican study, landholding inequality is essentially not endogenous because the distribution of landholding was frozen by the agrarian reform. Otherwise, it would be more difficult to determine the indirect effect of inequality on performance via the choice of rules.


To recapitulate, landholding inequality is associated with proportional cost sharing in the Indian study, and with proportional water allocation in the Mexican study. Proportional cost sharing is, in turn, associated with better performance in India, while proportional water allocation is associated with poorer maintenance in Mexico.


This is the principal spillover exploited in the model of Dayton-Johnson and Bardhan (in press).


These problems are considered in Young (this volume:Chapter 8).


As an example of this kind of research beyond the realm of unequal irrigators, Agrawal and Goyal (1999) analyze the question of group size based on data from 28 forest councils from the Indian Himalaya. Their appealing result is that there is a U-shaped relationship between group size and effective monitoring, rather than the classic monotonic result.


Aggarwal, R., and T. Narayan 1999 Does Inequality Lead to Greater Efficiency in The Use of Local Commons? The Role of Sunk Costs and Strategic Investments. Unpublished paper, Department of Agricultural and Resource Economics, University of Maryland, College Park.

The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement