12.3 The twin issues of group size and homogeneity
Back to contents - Previous file - Next file
Small is beautiful
One of the conditions for successful collective action most tirelessly and unanimously emphasized in the empirical literature is that user groups must have a small size. In Chapters 4, 5, and 6, we have offered strong theoretical reasons why small size may be such an important determinant of co-operative success and these reasons are strikingly similar to those advanced by social scientists. When groups are small, members tend to have frequent and highly personalized relationships and they have therefore a strong incentive to consider the more indirect and long-term consequences of their choices instead of paying exclusive attention to immediate costs and benefits. Moreover, the close, faceto-face nature of these relationships guarantees that people are well informed about each other's actions and preferences. Identity feelings are also likely to be strong with the result that emotions easily come into play to sustain co-operative behaviour. This effect, however, may not be as decisive as is too often implied. As a matter of fact, personal antagonisms and rivalries are also very much pervasive in village life, and they are probably less disruptive when diluted over a large group than when concentrated in a small one. Lastly, a point on which economic theory has laid much emphasis is that the smaller the group, the less diluted are the incentives to behave in a socially efficient way since the externality is correspondingly reduced.
The important question to ask is how small is small supposed to be. An opinion that is frequently encountered is that user groups ought to have a size much smaller than that of a village community. For example, in attempting to understand the causes of widespread failure with village woodlots experiments, an already-cited report has laid stress on the fact that 'tine community was not effective as a social unit in tree-growing programs', except in China and the Republic of Korea where community forestry has been supported authoritatively by the government (Gregersen et al., 1989: 133; Cernea, 1989: 33, 62). A similar conclusion is reached by Wilson and Thompson in their account of the breakdown in ejido productivity on extensive, livestock-herding areas in Mexico. According to them, indeed, pastoral management at community level (the level of the ejido) has proved a failure, presumably owing to the excessive size of the groups concerned. In a revealing manner, indeed, this failure of group management has led in a significant number of cases to the formation of grazing coalitions within smaller groups 'where cooperation is assured and benefits are enjoyed under very severe ecological conditions' (Wilson and Thompson, 1993: 300). The most effective of these (smaller) grazing associations are based upon the extended family. They practice intensive grazing management with short grazing periods and hoof action contributing to a more sustainable and productive range resource. Outside these short periods, multiple families combine their livestock into one large herd (200-600 animals) and move the entire grazing operation to range camps situated in remote areas less subject to population pressure (ibid. 310-12).
In the field of irrigation, to take yet another example, empirical evidence tends to show that successful irrigation systems usually operate in relatively small communities (Tang, 1992). Some sociologists consider that water-users' associations should correspond to hydrologically defined outlet units such as blocks of field neighbours under canal irrigation. Coward thus writes that 'for purposes of irrigation organization the critical unit is the "irrigation community", composed of field neighbours, and not the village community, composed of residential neighbours' (Coward, 1980: 208, quoted from Wade, 1988a: 214). There is a good ground, therefore, for subdividing an irrigation system into relatively separable units so that the actual number of farmers whose actions directly affect one another is kept quite small and it is relatively easy for each farmer to monitor other farmers (Weissing and Ostrom, 1991: 244-5). Wade has expressed strong dissent with this view arguing that in the villages he studied in Andra Pradesh the only corporate water organization is based on the village rather than the outlet, and the effectiveness of this pattern of collective action has been amply demonstrated by experience. Why it is so is explained by him in terms of features of social organization and in terms of economies of organizational scale achieved by combining water and grazing (Wade, 1988a: 213-14). Essentially, Wade's position—to which we shall soon revert—is that existing social ties and authority structure assume greater importance than strictly ecologically defined parameters in determining the optimal size of user groups (in this case, irrigation groups).
Other examples confirm Wade's intuition that small size is not a necessary condition of success in collective action. Thus, in Panama, the Kuna Indians have succeeded in systematically administering their own natural resources on a large scale, and this success included the setting up, during the early 1980s, of a wildland park and botanical sanctuary in an area covering 600 sq. km. (Utting, 1994: 251-2). The strong authority structure underlying this Indian society and the sharing of common norms by all its members probably go a long way towards explaining their remarkable performance. Another interesting illustration comes from the Gambian German Forestry Project (GGFP) in the Foni Brefet district. There, indeed, we are told that village leaders refused to divide a forest customarily shared among several villages on the ground that a clear delineation of the forest space could easily ignite intervillage conflicts. As a result of this opposition, it was decided to create a multivillage association that would 'enforce measures to exclude nonresident woodcutters from the use of the entire forest and to set up mechanisms to encourage controlled commercial woodcutting by resident villagers' (Freudenberger and Mathieu, 1993: 17). In this case, a tradition of good neighbourliness and, presumably, the need to unite against a common outside threat (represented by commercial woodcutters) have helped foster co-operation within a large group. The last point can also be derived from the experience of the Sehlabathebe Grazing Association in Lesotho. Indeed, if this large association covering eleven villages has been quite successful in enforcing exclusive use by local residents of pasture lands over which they have been granted exclusive grazing rights by the State, it has been far less effective in enforcing internal management rules among these local rightholders (Swallow and Bromley, 1994b: 7, 18). The lesson from these examples seems to be that there is some sense in saying that large groups are made more like small groups when their members share common norms possibly enforced by a well-recognized authority, or when they are confronted by a common challenge arising from without and they need to protect themselves against outsider encroachments. Note that the proposition according to which users are much more prone to take action to enforce a CPR boundary against outsiders than to enforce rules among themselves has been a central conclusion from Chapter 10. Clearly, as explained below, this discussion raises the issue of group homogeneity.
What bears emphasis, indeed, is that proponents of the view that small groups are more likely to be successful in collective actions do not base their argument entirely on considerations related to the closeness and frequency of relationships among members of such groups nor on the problem of incentive dilution with large groups. Another justification they often resort to is that small groups are more homogeneous than large groups. For example, the basic reason why Gregersen et al. think the village com munity is an ineffective social unit for tree-growing programmes is that, in their own words, 'the interests of community members often differ to such an extent that unified action is impossible'. The most important source of failure with the community approach to resource management lies in the large size and internal stratification of communities as social units. A small group is likely to be less diverse and less subject to internal strife (see, however, the aforementioned proviso regarding this effect); it is also better able to enforce rules about equal contributions by its members through peer pressures. Therefore, as in the case of a water-user association formed around a small branch of an irrigation system, a small group may succeed in operating a woodlot without the conflicts that surround community woodlots (Gregersen et al., 1989: 133, 135). A similar position is adopted by Lawry: 'where interests are heterogeneous and views toward appropriate resource-use standards vary, sufficiently strong support for enforcement of many kinds of rules will not emerge' (Lawry, 1989b: 7).
A few empirical studies have tried to test systematically the relationship between group homogeneity and success in collective action. In an aforementioned research about the determinants of effective CPR management in the dry regions of India, Jodha notes that the decline of CPR areas is less in villages with lower socio-economic differentiation, that is, in villages where access and benefits from CPRs are relatively equitable (Jodha, 1992: 41). Most authors concerned with village-level collective action for CPR management actually reach similar conclusions. This is especially so with field studies concerned with local water management and conflict resolution in water use. Thus, from a study of twenty-three community irrigation systems in different countries, Tang has reached the conclusion that a high degree of rule conformance and good maintenance tend to be associated with a low variance of the average annual family income among irrigators (Tang, 1991, 1992). That co-operation to form water-user organizations is easier to come by when the group involved is relatively egalitarian (as measured, for example, by the variation in farm size among the member farmers) is also evident from empirical studies conducted in the Indian states of Gujarat and Tamil Nadu (Jayaraman, 1981; Easter and Palanisami, 1986).
Unfortunately, the oft-heard argument about the comparative effectiveness of (small) homogeneous groups is not tight enough to be fully satisfactory. Too often, heterogeneity is blamed as a matter of principle without enough effort being devoted to spelling out the precise conditions under which it undermines collective action. In the following, we make an attempt to clarify the issue by considering various possible sources of heterogeneity in a group and the way they can possibly bear upon collective action capabilities. This will enable us to show that, contrary to what the above view implies, but in accordance to some of our theoretical findings in Chapter 5, there is no systematic link between group homogeneity and success in collective action.