MS5

Rule enforcement and local-level forest management

Clark Gibson ¹, John Williams ² and Elinor Ostrom ³

Abstract

Current studies of community-level resource management focus on the factors that impede or help individuals to overcome their collective action problems. Dozens of factors have been identified in this literature. While important, such work is limited by the lack of theory that connects these factors with outcomes, and the near total lack of hypothesis testing beyond the case level. We argue that, despite the possible differences between individuals or the characteristics of the resource they use, the regular monitoring and sanctioning of rules is a necessary condition for successful resource management. We offer a preliminary test of this hypothesis by pairing rule monitoring and sanctioning against other factors considered important in the literature, and by using a sample size of 178 user groups from the International Forestry Resources and Institution's (IFRI) research programme. We find support for the hypothesis that rule enforcement - i.e. regular monitoring and sanctioning - tends to dominate the other factors with regard to the probability that a forest is in good condition. While supportive of local-level management in general, our evidence questions efforts that attempt to set up such schemes based on ascriptive factors alone.

The foundations of successful resource management at the local level

Scholarship in the mid-1980s challenged the previous conventional wisdom that only by external enforcement could "the commons" be well managed (Hardin, 1968).⁴ Since that time, anthropologists, political scientists, economists and game theoreticians have examined the factors that appear to result in the successful management of natural resources. While these factors range from a group's dependence on a resource to their religious practices, they can be placed into three broad categories: characteristics of the resource, characteristics of the group and various facets of the external environment (markets, government agencies, etc.).

Central to these studies is the argument that local users can and have constructed rules to use their natural resources sustainably (see Ostrom, 1990). Essentially, most analysts theorize that various factors from the three categories above affect individuals' expected costs and benefits and these in turn, influence their decision to construct rules to manage their resource.

While successful at identifying more than 30 independent factors that they believe influence locals' decision-making, scholars have been less successful at explaining and testing the significance of the factors that they argue contribute to good outcomes. One of the principal failings is the lack of robust tests for their assertions, especially the small sample sizes used by those doing field studies (Agrawal, 2001; see also Gibson, McKean and Ostrom, 2000). We seek to contribute to this debate by exploring the importance of one factor - the enforcement of rules at the local level - and by testing our argument using a relatively large number of cases.

The primacy of rule enforcement

One of the most widely agreed upon factors leading to successful outcomes is rule enforcement (also called monitoring and sanctioning [Ostrom, 1990]). It is easy to see why. Regardless of what caused individuals in a group to create rules - their high level of social capital, their small size, their dependence on a resource, etc. - ultimately the rules must be enforced for the group's agreement to work. In many cases, well-known factors may have led individuals to construct a solution to their commons dilemma; in other cases, contextual idiosyncrasies may assume the most important role in creating rules. However, in all cases of success we would argue that there must be one or more individuals who monitor the rules and sanction the rule-breakers. This is true even in the unlikely case in which every individual in the group shares a preference for the same type of management scheme. Successful outcomes on the landscape still require enforcement: individuals can change their preferences, or outsiders can threaten the quality or quantity of a resource (Gibson and Koontz, 1998; Gibson and Becker, 2000). Given the theoretical primacy of rule enforcement, we argue that it is a necessary factor to effectively solve a commons dilemma.

Work employing game theory and computer simulations provides compelling evidence that not only is rule enforcement necessary, but it is also evolutionarily stable. Kameda, Takezawa and Hastie (2003) use evolutionary game theory and computer simulation to argue that rule enforcement actually can be rational for individuals. Boyd et al. (2003) also find enforcement evolutionarily stable, even in large populations.

Methods and data

We test our theory using the data produced by the International Forestry Resources and Institution's (IFRI) research programme. The IFRI research programme is a multilevel, multicountry, over-time study of forests and the institutions that govern, manage and use them. IFRI collaborators use the same ten research protocols to collect data about community-level rules, as well as socio-economic, demographic variables, and physical factors that affect human incentives and behaviour, and the impact of this behaviour on local forest ecologies. It is one of the first research programmes that combine systematic forest mensuration techniques with data about local institutions and socio-economic and demographic variables. Currently, the IFRI programme has 15 collaborating centres who have studied in 97 sites, which include 178 user groups and 220 forests (see Ostrom and Wertime in Gibson, McKean and Ostrom, 2000).

Even with these impressive data we faced significant impediments to creating a structural model for our data and theory. The first is missing data. As a "living" dataset with dozens of contributors, a number of potential cases cannot be used without further input from collaborators. Second, because there are few return visits, a structural model is haunted by endogeneity: are forest conditions causing rules or are the rules causing forest condition? Only when IFRI collaborators return to their sites for a second visit, can a time series be constructed. At this point we are forced to rely mainly on simple chi-square tests to evaluate our theory; we also present a regression equation that is limited to a very small sample size.

For our preliminary tests, we have chosen three explanatory variables: the regularity with which individuals in a user group monitor or sanction others' rule conformance (which we call rule enforcement), the group's social capital, the group's dependence on forest resources, and whether the group is a formal organization or not. Rule enforcement is measured by a scale about whether the user group undertakes monitoring and sanctioning efforts at certain time intervals: never, occasionally, seasonally, or year-round. We measure social capital by combining a number of variables regarding the cooperative activities that individuals from a user group undertake in the forest. We calculate dependence by adding the percentage of needs that individuals in a user group claim are met by a forest for their food, biomass, timber and firewood. The measure of user group's formal organization derives from whether or not they have meetings and officials. Full explanations of the construction of these variables and the definition of the user group are available in the unedited version of the paper.

We also have two dependent variables in this study, both measuring the condition of the forest based on an assessment by the user group. One of these variables is based on a user group's assessment of the condition of the forest at the time the IFRI study was done, the other on a more dynamic assessment about how the condition of the forest has changed over time. We generate six three-way tables, and given that the monitoring variable is our focus variable, it is always in one of the tables. We then use chi-square tests of these relationships. (We report only the results for the tables that used dynamic forest condition, given that the results from both the static and dynamic measures of forest condition are largely the same.)

The general idea is to use expectations from theory to predict the pattern in these three-way tables. It is important to understand that we are looking for patterns in the tables - as exhibited in both Tables 1-3 and Figures 1-3 - and not statistically driven causal relationships. In short, we try to see if the patterns that our theory predicts appear in the tables' patterns.

Hypotheses

We posit one general hypothesis: that higher levels of rule enforcement lead to better forest conditions. We delineate the effect of this general hypothesis on the relationships between social capital, rule enforcement and forest condition below.

Hypothesis A. If social capital is low and rule enforcement is sporadic, we expect forest condition to be poor.

Hypothesis B. If social capital is low and rule enforcement is frequent, we expect forest condition to be good.

Hypothesis C. If social capital is high and rule enforcement is sporadic, we expect forest condition to be poor.

Hypothesis D. If social capital is high and rule enforcement is frequent, we expect forest condition to be good.

We thus theorize that consistent rule enforcement is necessary for good outcomes in forest management. Whether a community has a high level of social capital or not, it will still need to monitor and sanction whether its members are actually conforming to their own agreements about how to manage a resource. Without such enforcement, agreements may become meaningless within a short time.

We argue that the same pattern holds in the relationships between enforcement and formal organization of user groups, and between enforcement and a group's dependence on forest resources: where rule enforcement is frequent, we expect better forests; where it is sporadic or non-existent, worse forests. Regardless of a group's level of formal organization or of a group's dependence on a forest's resources, the frequency of rule enforcement should be more important in determining forest conditions.

Results

The explanatory and dependent variables produce three tables for analysis. To provide an easy visual examination of the results, we have produced figures that illustrate the results based on chi-square tests of significance for each subtable. Information about the exact cell entries, as well as information about scales measuring each of the variables, can be obtained from the authors.

Figure 1 provides the dynamic forest condition by regularity of enforcement and level of social capital. To keep cell values large, we have dichotomized each, and this is the case in each figure. When paired with rule enforcement, we find that social capital does not appear to matter to dynamic forest condition. Frequent rule enforcement correlates with good forest conditions, and sporadic rule enforcement correlates with poor forest conditions. These results are significant at the .01 level. Thus, there is significant empirical support for all four hypotheses that we proposed for the relationship between rules and social capital. Further, no outcome contrary to our hypothesis is statistically significant, e.g. there is no significance to the path in which social capital is low, the frequency of rule enforcement is low, and the forest condition is high.

In Figure 2 we examine how formal group organization might influence forest conditions, when paired with rule enforcement. Our hypothesis about the importance of enforcement finds significant empirical support only in the case of no formal organization and sporadic rule enforcement. When there is no formal organization and only sporadic rule enforcement, there is a statistically significant result of poorer forest conditions.

There is also some, albeit weaker evidence found in this figure as well, in the sense that no significant outcomes are contrary to our hypotheses.

Figure 3 presents the results for the relationship between forest dependence and the regularity of rule enforcement. Here we see the same strong support for our hypotheses as in Figure 1: all four hypotheses that correspond to the importance of rule enforcement are statistically significant. Regardless of the level of dependence that a group may have on a forest's resources, the level of rule enforcement is associated with forest conditions. And, as in Figure 1, no relationship contrary to our hypotheses is statistically significant.

We also made a preliminary attempt to employ a structural model to test our hypotheses. This logistic regression neither solves the endogeneity nor the missing values problems, but does constitute a first run at what we hope to refine in the future. We use the same variables as in the chi-square tests. The results are at least suggestive that our hypotheses are on the right track (see Table 4). The overall model is almost significant at .126, and the only independent variable that comes close to significance is rule enforcement (.132).

Discussion

In this paper we try to deduce what factors associated with successful, local resource management are necessary in all cases. We argue that rule enforcement must be present for successful outcomes on the landscape. We tested this hypothesis using a relatively large amount of data collected from 12 countries using the same research protocols as the IFRI research programme. Employing simple chi-square tests, the data support our hypotheses.

The study begs several important questions. The first is of course the "so what?" question. Any seasoned policy practitioner knows that rule monitoring and enforcement are necessary to help a programme reach its goals, so it may be easy for said practitioner to roll their eyes at this "simple" theory. However, what is important is that this study finds that rule enforcement dominates all other factors. This should send a message to many: rather than focus on other factors per se, this study indicates that we should examine their impact on rule enforcement. That is, although studies claim that things like "social homogeneity" lead to better local-level forest management, it might be that this factor's real effect is through rule enforcement, and not some irreplaceable ascriptive characteristic of the group. This is good news for policy-makers, because it may allow the construction of effective policy, rather than a search for some demographic or cultural similarity for management partnerships.

Another question is "what role for government versus local rule creation and enforcement?" This study also provides some evidence that supports the argument that "locals can do it." In our study, we did not discriminate official user groups like forestry agencies from local user groups like a women's fodder harvesting group. In fact, only about one-fifth of the cases represented official agencies. Thus our evidence shows that local groups can be effective - evidence drawn from 13 countries and dozens of user groups of various kinds. Given that official forestry agencies never have enough resources to enforce their rules at all times, it appears that some type of joint arrangement seems one of the likely paths ahead (although the results of programmes such as India's Joint Forest Management and Nepal's Leasehold Forestry are unclear).

This study did not answer the central questions regarding the nature of the mix of local/official power and enforcement. While decentralization and devolution of rights have been popular recently, there are few systematic studies of the effects. Indeed, in the environmental realm there are theoretical reasons to think that wholesale devolution of power over natural resources is not a good idea, given their large externalities (in contradiction to the common property literature). Local-level groups generally have a more difficult time with larger tracts of forest and commercial invasions (factors directly related to rule enforcement!). And, barring effective federations, which require enormous resources to construct, locals must ally with more powerful units. This may be a government agency, but it also may be an informal and corrupt alliance with commercial interests.

The study also did not forward any evidence about the effect of property rights. Such cases range from forests located on private property where the owners are part of the user group, to government forests with user groups illegally harvesting products. There is evidence that enforcement on the ground is more important than de jure property rights (Gibson, Lehoucq and Williams 2002), but this awaits a systematic test that can hold other factors constant.

While this study makes clear that local-level management can lead to good outcomes on the forested landscape, future studies must examine the relative importance of the panoply of factors touted by scholars to be essential. Ultimately, the goal of such work should be reliable evidence about the structure of successful management schemes under different conditions.

Table 1. Social capital, rule enforcement and forest condition

Table 1a. Chi-square tests

Table 2. Formal organization, rule enforcement and forest condition

Table 2a. Chi-square tests

Table 3. Forest dependence, rule enforcement and forest condition

Table 3a. Chi-square tests

Table 4. Logit regression results

References

Agrawal, Arun. 2001. Common property institutions and sustainable governance of resources. World Development, 29 (10): 1649-1672.

Axelrod, Robert & Hamilton, William D. 1981. The evolution of cooperation. Science, 211(4489) (March):1390-96.

Baland, Jean-Marie & Platteau, Jean-Philippe. 1996. Halting degradation of natural resources: Is there a role for rural communities? Oxford, Clarendon Press.

Berkes, Fikret. ed. 1989. Common property resources: Ecology and community-based sustainable development. London, Belhaven Press.

Boyd, Robert, Gintis, Herbert, Bowles Samuel & Richerson, Peter J. 2003. The evolution of altruistic punishment. Proc. of the National Acad. of Science, 100 (March): 3531-3535.

Clutton-Brock, Tim, & Parker, Geof A. 1995. Punishment in animal societies. Nature, 373: 209-216.

Gibson, Clark & Koontz, Tom. 1998. When community is not enough: Communities and forests in southern Indiana. Human Ecology, 26: 621-647.

Gibson, Clark & Becker, C. Dustin. 2000. The lack of institutional supply: Why a strong local community in western Ecuador fails to protect its forest. In Clark Gibson, Margaret McKean & Elinor Ostrom, eds. People and forests: Communities, institutions, and the governance of forests: pp.135-162.

Gibson, Clark, McKean, Margaret & Ostrom, Elinor. eds. 2000. People and forests: Communities, institutions, and the governance of forests. Cambridge, MA, MIT Press.

Gibson, Clark, Lehoucq, Fabrice and Williams, John. 2002. Does privatization matter to resource management? Property rights and forests in Guatemala. Social Science Quarterly, 83(1): 206-225.

Hardin, Garrett. 1968. The tragedy of the commons. Science, 162 (Dec.): 1243-48.

Kameda, Tatsuya, Takezawa, Masanori & Hastie, Reid. 2003. The logic of social sharing: An evolutionary game analysis of adaptive norm development. Personality and Social Psychology Rev., 7(1): 2-19.

McCay, Bonnie J. & Acheson, James M. 1987. The question of the commons: The culture and ecology of communal resources. Tucson, University of Arizona Press.

Ostrom, Elinor. 1990. Governing the commons: The evolution of institutions for collective action. New York, Cambridge University Press.

Ostrom, Elinor, Walker, James & Gardner, Roy. 1994. Rules, games, and common-pool resources. Ann Arbor, University of Michigan Press.

Sugden, Richard. 1986. The economics of rights, co-operation, and welfare. Oxford, Basil Blackwell.

Wade, Robert. 1994. Village republics: Economic conditions for collective action in South India. San Francisco, CA, ICS Press.

¹ Department of Political Science, University of California, San Diego, 9500 Gilman Drive, 0521 La Jolla, California 92093 - 0521, USA. [email protected]

² Department of Political Science, University of California, Riverside, USA.

³ Center for the Study of Institutions, Population, and Environmental Change, Workshop in Political Theory and Policy Analysis, Indiana University, USA.

⁴ In this paper, we use the word "commons" interchangeably with common-pool resources. We use both words to refer to natural resources, although theoretically commons exist in all aspects of human society.