The Ecology of Games Framework: Some Responses to Critics
A major branch of my research is devoted to studying complex institutional systems, which I argue are the defining feature of real-world environmental governance and public policy more generally. Along with my colleagues (especially John Scholz and Ramiro Berardo) and students, we have updated the “ecology of games” idea originally developed by sociologist Norton Long in 1958 to describe the many different types of political actors and institutions operating in local political contexts. Our ecology of games framework (EGF) synthesizes a number of existing theoretical concepts, with a strong basis in the work of Elinor Ostrom and new institutional economics, network analysis, and complex adaptive systems. We have collected a large amount of data on water and climate governance in several different countries, submitted and published a number of papers, and given a wide variety of talks.
This scientific dialog and process has identified several persistent criticisms of the EGF, which we debate with reviewers, but usually do not have the space to fully explore in a publication. Furthermore, since our work is breaking new ground in a number of ways (at least I think so), we do not have strong empirical or theoretical answers to all of the questions yet. That means a blog is a perfect place to have dialog of ideas and speculations, without having to hew to the standard of strong scientific evidence. After all, where would science be without speculative curiosity and tentative hypotheses?
Why use the biological metaphor of “ecology”?
Various reviewers have warned of the (non-specific) dangers of introducing a biological metaphor into a social science discussion. While I largely chalk this up to close-minded reviewers trying to perpetuate disciplinary boundaries, the more constructive answer is that the concept of ecology embraces the idea of interdependence among various parts of a system. Resources flow between system components, and actions occurring in one part of the system may positively or negatively influence other parts. Understanding the structure and function of an ecological system requires analyzing how all of the parts relate to each other. Complex institutional systems feature these exact same type of phenomena, hence the term ecology is a useful metaphor.
Why don’t you use the term “policy forum” or “policy venue” instead of “policy institution”?
The EGF uses the term policy institution to describe the rules governing collective decision-making regarding some policy issue. Critics say this is just another name for a policy venue or forum, and in one sense they are correct—policy institutions constitute a social space where people interact. There is a long history of the term “venue” in the policy sciences, for example many papers talk about “venue shopping”. My problem with the terms “venue” and “forum” is that they do not go far enough in connoting what happens in these social spaces. The term “institution” specifically refers to the rules governing collective decision-making, including which actors are allowed and expected to interact. Forums and venues are just “black box” spaces for social interaction, while the term institution relates to how the social process unfolds.
What is a “policy game”?
In classical game theory, a game consists of actors, choice sets, and payoffs. Game theory is strategic because the payoffs for one actor depend on the choices of the other actors; game theory entails interdependence. In the EGF, a game occurs when actors make collective-decisions in the context of a policy institution, and the decisions apply to some underlying policy issue. I usually think about the policy issue as involving some basic collective-action problem like a common-pool resource or public good. The collective-decision making process entails payoffs for different actors, which are at least partially a function of how those collective-decisions shape how the use of the environmental resources under consideration. But a policy game only occurs when there is a constellation of actors, rules, and issues (one reason network analysis is a useful tool). Since I love to play soccer, I think a soccer analogy works well here---a soccer game occurs when the players (policy actors) arrive at the field and play according the rules (policy institution) of soccer, kicking around a ball (the policy issue). If nobody is on the field, the rules still exist, but there is no game in progress.
Isn’t this just “polycentricism” in new bottles?
The concept of polycentrism is no doubt a crucial part of the EGF. We are definitely studying polycentric governance arrangements. The difference is that polycentrism is only a theoretical concept, it is not a stand-alone theoretical framework that drives hypotheses about the structure and function of complex institutional systems. Polycentricism itself does not identify variables and the relationships among them. The original concept was largely normative—polycentric governance arrangements were considered more efficient in the delivery of public goods. The EGF does not make this normative assumption, but rather considers the fact that a large number of different types of governance arrangements might have better or worse fits to the set of collective action problems at hand. In addition, the EGF attempts to identify hypotheses about the structure and function of complex institutional systems—e.g.; collaborative institutions will be central within the system (see various papers for more hypotheses). Hence, the EG framework is really a theory of polycentric governance, not just another name for polycentric governance.
Why do the perceptions of actors matter?
A lot of our empirical work relies on surveys of policy actors, which measure their perceptions of the effectiveness of different institutions, availability of scientific knowledge, levels of cooperation among actors, and severity of various environmental problems among other things. A common criticism of this approach that institutions perceived to be effective are not always in reality effective if you could measure environmental outcomes, or that actors who think scientific knowledge is adequate may in fact not have very high levels of scientific knowledge if measured through some purported objective test. Of course this is descriptively true—we do not always have the capacity to directly measure some of the “real” variables being asked about on a survey.
However, granting the validity of this criticism does not mean perceptions are irrelevant or unimportant. In fact, one could argue that perceptions are even more important than reality because perceptions are the proximate drivers of decisions. Even if some policy actor is dead wrong about the level of some environmental parameter like water quality (for example), it is their beliefs about water quality that drive their decisions. The general public opinion debate about vaccines and climate change should make this painfully obvious.
Furthermore, the very important theoretical work of institutional economist Masahiko Aoki clarifies the link between subjective beliefs systems (perceptions) and strategic interaction. Even if all actors are operating from different subjective belief systems, they experience regular outcomes from repeated interaction, which allows them to develop strategic responses to other actors even if the other actors perceive the game differently. When the behaviors stabilize into an equilibrium, Aoki argues an institution has been established. For me, Aoki delivers an extremely strong theoretical justification for studying perceptions in the context of strategic interactions like those that occur in the EGF.
Obviously, a crucial aspect of the future research agenda is exploring the relationship between perceptions of various sorts and measurements of more objective outcomes like environmental quality. Are measurements of environmental quality correlated with perceptions of institutional effectiveness? My guess is that the correlations between “real” outcomes and perceptions varies across individuals, and over time. Different institutional arrangements will increase the correlation between environmental outcomes and perceptions, for example the integration of scientific information into policy decisions. Systems that feature higher correlations between environmental conditions and perceptions are likely to be more effective overall, because the proximate drives of decisions (perceptions) are more response to the factors that ultimately determine payoffs experienced by actors and society.
Can the EGF make policy recommendations?
This criticism is one of the more difficult ones for me to think about. The first thing to say is that yes we definitely do want to make strong policy recommendations at some point. We would like to make recommendations about how to structure complex institutional systems in ways that support cooperation, improve environmental problems, and increase system resilience. We would also like to develop strategies that can help actors navigate complex institutional systems to achieve their policy goals. This is especially poignant because I believe complex institutional systems are truly the reality of public policy, and that analysis of individual programs or single policy issues in isolation will always be incomplete (in grandiose moments I think of it like the difference between partial and general equilibrium analysis in economics). When I show policy-makers pictures of some of the network diagrams we use to describe complex institutional systems, they almost always agree that is the world they are facing. Then they ask me for advice on how to improve it…and I don’t really have good answers at this point.
The reason I don’t really have good answers is that we are so early in the scientific process of developing theory and empirical evidence about these complex institutional systems. We are beginning to get some idea of structural properties (at least in a few contexts….), and some insights into function at least as perceived by actors. But as mentioned earlier, we don’t really have a bunch of long-term measurements that allow us to relate structure, function and perceptions to actual environmental or other policy outcomes in many different research sites. Without a better evidence base and more refined and precise theories, I feel like our policy recommendations are premature and risky.
That said, I suppose some concentrated thought and dialog could produce some initial ideas; I’ve been contemplating about how to seriously engage in that exercise. I think returning to some observational and qualitative fieldwork would be an important aspect. Spend a lot of time interviewing policy actors, explaining to them how we’re seeing the complex institutional system, and asking them what they have seen work, and how they navigate it. Some of my original thinking about the EGF came from qualitative dialog with water policy stakeholders; now that we have done some empirical work, that dialog is better informed. I do jealously admire the work of Erik Hans-Klein and the crew at Erasmus University in Rotterdam, Netherlands. They have been studying the same issues for years using a similar framework, and have developed long lists of qualitative policy recommendations. They are really good at extracting some clear ideas from a lot of qualitative information. I like to think the EGF complements their work with a potentially simpler theoretical framework and quantitative analysis.
Well, this certainly doesn’t exhaust the list of criticisms but I can’t think of any other sore thumbs at the moment. Plus this blog is getting long. I will write another one or add to this one, as additional recurrent themes emerge.