Why do societies choose inefficient policies and institutions? An extension of the traditional Coase Theorem to the political sphere suggests that this should not be the case. Previous theoretical research has identified one of its key assumptions as the culprit. This post will explore the Political Coase Theorem, illustrate the likely reason why it often does not hold, and summarize recent experimental evidence on its validity.

The Coase Theorem (Coase 1960) characterizes  the economic allocation of resources in the presence of externalities. The theorem states that if trade in an externality is possible and there are sufficiently low transaction costs, bargaining will lead to a Pareto-efficient outcome regardless of the initial allocation of property rights and the bargaining power of the parties involved, since these factors only affect the division of the surplus.

This concept has also been extended to the political sphere. The Political Coase Theorem (hereinafter PCT) states that, in the absence of transaction costs, political players should agree to implement efficient policies regardless of the distribution of political power among them, which should only affect how gains are distributed across players.

The crucial implication of the PCT is that policy and institutional differences across countries should not be a major determinant of differences in economic outcomes, since societies choose, at least approximately, the best policies and institutions given their prevailing conditions. Yet, there is a great deal of evidence that this is not the case. Inefficient policies are pervasive, a fact that suggests that the application of the Coase Theorem cannot readily be extended to the political sphere. This somewhat puzzling situation can be explained by the fact that one key assumption of the Coase Theorem probably does not hold in politics: the assumption that agents can commit to their agreements. Without this assumption, the theorem breaks down as a result of time inconsistency, and only inefficient outcomes emerge.

In a standard economic transaction, this may not be a serious problem because agents can always rely on binding agreements enforced by a third party (e.g., a contract enforced by the courts). Politics, however, involves bargaining among powerful players, who, because they are so powerful, can renege on their agreements. As a consequence, powerful players face a commitment problem which may restrict the scope of the agreements that they can reach with other players.

A historical example may help to illustrate the crucial role of commitment problems in politics. North and Weingast (1989) studied the Glorious Revolution in Britain as an example of an institutional change that solved a crucial commitment problem. Before the revolution, Parliament was not willing to introduce new taxes because the king had considerable discretionary power over expenditures. Nor could the king borrow money because he could not credibly commit to repay his debts. As a consequence, the Crown used very inefficient sources of revenue, such as the sale of monopoly licenses. The Glorious Revolution ushered in the era of parliamentary supremacy. The king could no longer dissolve Parliament, Parliament significantly increased its power to control and monitor expenditures, royal prerogative powers were reduced and subordinated to common law, and judges stopped serving at the king’s pleasure. The successful dethroning of Charles I and James II posed a credible threat to future monarchs; however, the Crown continued to play an important role as a counterbalance to the power of Parliament. As a consequence, after the Glorious Revolution, the government gained access to more efficient sources of revenue.

The importance of the PCT cannot be overemphasized. On the substantive front, there is hardly a more relevant issue in the social sciences than the identification of the sources of inefficient policies. From a theoretical perspective, most formal political economy models now simply assume that the PCT does not apply. This is the end result of a shift in the literature on institutions and institutional change away from a tacit acceptance of the PCT.

However, only Acemoglu (2003) presents a formal political economy model in which social conflict and limited commitment are the key factors that undermine the validity of the PCT. His model is an infinitely repeated taxation game between a ruler and a citizen. Under no commitment, the unique sub-game perfect Nash equilibrium of the stage game leads to a very inefficient outcome. The citizen does not work hard because he knows that the ruler will tax all his income, and the ruler does not have any way to credibly commit herself to refraining from taxing the citizen’s income. When the ruler can credibly commit to limit taxation or the citizen can credibly commit to pay a transfer if the ruler resigns, the efficient outcome is restored. Note, however, that the latter case requires an institutional change (the ruler must resign). Even when neither the ruler nor the citizen is able to commit, repeated interactions open the door to limited commitment and, hence, to better social outcomes. In other words, repeated interactions make some promises credible.

In our paper "The Political Coase Theorem: Experimental Evidence", we employ a randomized laboratory experiment to explore how commitment problems affect the PCT.

We begin by adapting the model developed by Acemoglu (2003) to serve as a simplified version suitable for testing in a laboratory environment. The adaptation retains three main features of Acemoglu’s model. First, agents’ ability to commit is limited. Second, commitment opportunities could be asymmetrically distributed among players. Third, exploiting commitment opportunities could require an institutional change. The adaptation is simpler because we consider a two-period game in which promises are only partially binding. Specifically, there is some probability that a player must keep her promises and some probability that promises are not binding at all. By changing these probabilities, we induce different levels of commitment. Our simple two-period game with partially binding promises is able to capture the key insights on commitment problems as they relate to the PCT. In particular, we can handle situations in which nobody can commit; only one player has the ability to partially commit; and only one player has the ability to commit, but an institutional change is required.

Participants were first randomly assigned to one of two roles - Player 1 (the ruler) or Player 2 (the citizen) - and then to four different treatments. Each treatment differs only in terms of commitment opportunities: in Treatment 1, neither player can credibly commit (promises are not binding at all); in Treatment 2, Player 1 has a slight commitment opportunity, while Player 2 has none; in Treatment 3, Player 1 has significant commitment opportunities, while Player 2 has none; and, last, in Treatment 4, Player 1 has no commitment opportunities, while Player 2 has full credibility (her promises are binding). The first three treatments offer increasing commitment opportunities to Player 1 (the ruler), who initially holds the power to tax. The last treatment offers more commitment opportunities to Player 2 (the citizen), who initially does not have the power to tax, which means that the players must re-engineer the distribution of their relative power (an institutional change) in order to exploit commitment opportunities.

Overall, the results of this experiment support the hypothesis that commitment issues matter and that the existence of more commitment opportunities leads, on average, to better social outcomes. Indeed, we find that this is valid even when a reallocation of political power is required in order to take advantage of new commitment opportunities. However, we also find that, at low levels of commitment, there is more cooperation than would, strictly speaking, be predicted by our parameterized model, while the opposite is true at high levels of commitment. Furthermore, only large improvements in commitment opportunities have a significant effect on the social surplus, while small changes do not. These results are consistent with the previous literature.

Our results have interesting implications for political economy. The bad news is that the experiment suggests that small changes in commitment opportunities might not improve social outcomes. This result could be important for a longstanding debate in political economy about the relative virtues of gradual and radical reforms. The good news is that subjects seem to understand that sometimes, in order to take advantage of new commitment opportunities, political power must be reallocated to the agents who can credibly commit. This result could also be important to understand the political economy of reforms. Indeed, some historical reforms can be explained as the outcome of a redistribution of political power between parties with different abilities to commit.

References:

Acemoglu, Daron, 2003. "Why Not a Political Coase Theorem? Social Conflict, Commitment, and Politics", Journal of Comparative Economics, vol. 31, pages 620-652.

Coase, Ronald, 1960. “The Problem of Social Cost”, Journal of Law and Economics, vol. 3, pages 1-44.

Galiani, Sebastian, Gustavo Torrens and Maria Lucia Yanguas, 2014. "The Political Coase Theorem: Experimental Evidence,", Journal of Economic Behavior & Organization, Elsevier, vol. 103(C), pages 17-38. 

North, Douglass C., and Barry R. Weingast, 1898. "Constitutions and Commitment: The Evolution of Institutions Governing Public Choice in Seventeenth-Century England", The Journal of Economic History, vol. 4, pages 803-832.