Social Mechanisms?

by David K. Levine, Andrea Mattozzi and Salvatore Modica on October 15, 2019

Academics of all types are as puzzled by these events as they are convinced it proves the correctness of their pre-existing views. Political economists write subtle mathematical articles examining how fake news can drive out the real - while useful, we suspect that George Orwell would not be amused. Simultaneously the hidden truth behind political lies remains unseen. Meanwhile we economists make an industry of proposing carbon and wealth tax schemes that are doomed to fail.

Conventional wisdom in economics is that markets work well while political systems are subverted by special interest lobbying. The most extreme expression of this point of view is found in the Chicago school, but the range of disagreement is small: while the Chicago school argues that the benefits of government action are undone by regulatory capture, the MIT school argues that moderate regulation and anti-trust rules can be useful in the occasional circumstance of market failure. In evaluating this near consensus there are two crucial questions that ought to be addressed:

1. Is it true that markets work well? Both lobbying groups and cartels face a public goods problem. Groups that are effective at lobbying are small in relative terms but large in absolute terms. If these groups are able to overcome the public goods problem in order to lobby why can they not also overcome the public goods problem of cartel formation? If they can, markets will be just as subverted as political systems. More specifically - why do some large groups such as farmers successfully subvert the political system but fail to subvert markets while others such as trade unions successfully subvert both?

2. Is it true that political systems are subverted by special interest lobbying? Both lobbying and elections are political contests. In the conventional view small groups win at lobbying but fail in elections. Why are special interests successful in one type of contest but not the other? Yet some small groups such as teachers unions are effective at winning elections, while other small groups such as those arguing for minority rights often lack political influence.

Persuasive and systematic answers to these questions are important for a variety of reasons: not least that one of the driving forces behind populism is the subversion of democratic systems by special interests. Without understanding the disease it is difficult to propose a cure.

The purpose of our research is to explore a theory that provides coherent answers to these basic questions. Our conclusions are not always the obvious ones - for example, we find that making lobbying more difficult will increase the influence of special interests, while making voting easier can reduce voter turnout. Our theory enables us to examine the issue of why social and cultural norms at times persist after they become dysfunctional while at other times change with blinding speed. We explain why the taxes most favored by economists - taxes on wealth and Pigouvian taxes - face nearly insurmountable political obstacles. By laying down a firm foundation for thinking about political economy we uncover sources of our current malaise.

Our theory is well grounded in empirical research and in existing economic theory. It is not a behavioral theory although it explains many of the anomalies that motivated research in that area. The fact is that after several decades of research in behavioral economics no coherent theory explaining a broad range of facts has emerged. Moreover, we reject the strongly individualistic basis of both conventional and behavioral economic theory, adopting instead the collectivist point of view used both in political economy and in the theory of the firm.

Our theory is a theory of group behavior which takes into account individual incentive compatibility. It is a theory that takes as its point of departure that whenever group members are asked to contribute to a public good such as lobbying, voting, or by limiting output to exploit market power, that group faces a free rider problem: each member would prefer that other members bear the cost of the public good. It is a theory grounded in research showing that groups are effective at disciplining members through social means including exclusion and ostracism - in a word, through peer pressure. It is a theory grounded in empirical research showing that the social norms that emerge endogenously in self-organizing groups are functional and effective in overcoming free riding problems. It is a theory that recognizes that groups need to monitor compliance with social norms. Above all it is a theory that recognizes that groups respond to incentives in different ways than do individuals. It recognizes, for example, that distribution is not neutral so which group gets the proceeds of a tax is crucial to understanding the political consequences of tax policy.

Specifically, our theory views a group of individuals with relatively common interests as facing a mechanism design problem. The design problem is the choice of a social norm in the form of a target contribution to public good production together with a punishment for failing to comply with that norm. With imperfect information about the compliance of individual members how optimally should the group discipline members and what does this mean for the optimal production of public goods?

This theory of social mechanisms draws on existing mechanism design theory. The theory is individualistic in the sense that group members do not act out of altruism or concern for group welfare but pursue their own interest. It is collectivist - as is most mechanism design theory - in the sense that the group is assumed to be able to agree on an optimal mechanism that satisfies individual incentive constraints. It is behavioral in the sense that it recognizes the importance of social interaction and that exclusion is a powerful punishment that can be used to overcome free riding.

Our flavor of mechanism design theory has several features that have not played a prominent role in earlier work on mechanism design.

1. It emphasizes the use of punishments rather than rewards. This recognizes the fact that political and social groups unlike business firms do not generate much by way of revenue that can be used for rewards. It leads to an emphasis not on implementation of the first best, but rather on the costs of implementing a second best.

2. Our monitoring technology is endogenous. Standard cartel and mechanism design theory assume that there is a given noisy signal of individual output. We assume that the signal depends upon the targets established by the group - once the target is established the monitoring technology is specialized to the detection of violations. Not only is this a more sensible assumption but it greatly simplifies analysis of the mechanism design problem.

3. We focus on mechanisms that coexist and compete with each other - as is the case, for example, when two political parties each design a mechanism to employ in an electoral contest. Standard mechanism design theory focuses on stand-along mechanisms to solve public goods problems.

4. We explicitly account for adjustment costs. While is is costly for individuals to change plans in response to circumstances it is especially costly for groups. Moreover adjustment costs for groups differ from adjustment costs for individuals: groups have options not available to individuals including both disbanding and engaging in political activity.

Our perspective on economic research is that theory directs our thinking about data - what questions to ask, what data to collect, and how to analyze it. Strong theoretical and conceptual underpinnings we believe are essential to that thinking, so our theory is directed at foundational issues. We understand as well that analysis and policy cannot always wait for the most satisfactory theoretical underpinnings and a great deal of research in political economy has been conducted using different - and in our view not fully satisfactory - underpinnings. This does not mean that we reject this research - especially not as much of it has substantial empirical validation. A new theory if it is to be useful has to have substantial backwards compatibility - it must not unexplain that which has already been explained - it must agree with valid insights that already exist. While our primary focus is on what is new and not on what has been done by others, we also examine of existing models and explanations and how and when our theory is compatible with them.

The theory of social mechanisms is a new one founded in research we began about a decade ago. It raises new questions and proposes new answers to old questions. As it is a new theory it has not been subject to the scrutiny applied to existing theories. The goal of our research is to lay out the theory in a coherent way along with relevant facts and to argue that it has enough potential to deserve careful empirical and theoretical scrutiny.