Wittman argues that democratic failure is a myth because political institutions reduce information failures, transaction costs and other problems that could impede efficient outcomes. That is true enough, to some degree. Political competitors can gain some advantage by catering for voter ignorance. Coalitions that imply much dead weight loss will, ceteris paribus, tend to be outcompeted by more efficient ones that have more pie to divide. Candidates and parties, by building up reputations, can to some extent solve problems of time inconsistency and agent shirking. After all, there is competition both within and between parties.
However, any argument that democratic systems are to some degree competitive is trivially true otherwise one could scarcely call them democratic. The relevant question is by how much democratic institutions will tend towards efficient outcomes.
By most standards, Wittman does not provide an adequate theory of constrained political efficiency to start answering this question. By and large, he contents himself with presenting mere judgments and assertions. Thus, we are told (pp. 73, 108, 186, 174) that if academic analysts can detect asymmetric information and principal-agent problems, then it must follow that political principals can, too – hence those problems are ‘exaggerated.’ Elections, in turn, are deemed to be transaction cost-reducing ways to exercise political takeovers since ‘the time period between elections is relatively short (legislators are not elected for life), there are no super-majority requirements for being elected, and the opposition participates in the legislature’ (p. 23).
Such assertions are especially problematic for a theory that adamantly dismisses existing approaches in political science in favour of the view that ‘democratic markets do indeed have the qualities typically associated with efficient markets’ (p. 3). Regarding the question of degree, there is little doubt that the information failures and the costs of transacting are particularly pervasive in politics. Avinash Dixit’s MEP helps to clarify this in a number of ways. For instance, it is not inconsequential that politicians often allocate income on the basis of explicitly political considerations, such as prospective votes.
Take an industry that is affected by a negative shock. Politicians may want to compensate those workers who are affected. The economically efficient way to do so would be a lump-sum grant equalling the capitalized value of the workers’ income losses. Rather than distorting subsequent decisions by workers, such a grant would allow them to move at once to their most productive alternative jobs. Lack of credibility complicates matters, however. Workers cannot credibly promise to reward the politicians with votes over all elections that span their prospective loss. So the latter will offer compensation as partial payments over many periods, rather than up front. But politicians themselves cannot credibly promise to keep on compensating far into the future.
Workers know that they will keep on receiving compensation only if they retain enough political clout. This may lock them in the uncompetitive industry, and prevent them from moving to more scattered, albeit more productive, occupations. In equilibrium, time inconsistency results in inherently inefficient political redistribution.
Building on economic agency theory, Dixit systematically analyses the incentive structure in political institutions. Fairly consistently, the incentives for political agents to perform according to their principals’ objectives turn out to be substantially less powerful than in equivalent market contexts. This is due to many structural characteristics of political agencies, including the lack of transparency and of yardstick competition. The political agency contract itself is ill-defined. It differs from economic contracts ‘in several ways, all of which make them more complex and harder to define’ (p. 48). Electoral campaign promises are vague, multidimensional, and not legally enforceable. One might add that voters do not make the repeated, separate choices over many different issues that actors make in those economic markets that approximate efficient competition. Typically, their preferences over many different issues have to be bundled together, every so many years, in one single vote, for one out of only a few credibly competing parties. Assuming that voter control (not mere political takeover) matters in democracy, elections come out as a less than powerful weapon for voters to sanction their agents (Przeworski et al., 1999).
Political agencies often perform ‘multiple tasks,’ sometimes along many dimensions of input (effort) and output (performance), which are difficult for principals to observe. Moreover, they may have to deal with ‘multiple principals’ with partly conflicting interests. For instance, a bureaucratic agency may be simultaneously answerable to the executive and the legislative branch of government. Dixit (pp. 157–71) models such an incentive scheme with n principals. In equilibrium, it has only about (1/n)th of the power of a second-best scheme with a unified principal.
MEP adds a more complete conceptual framework to earlier transaction-cost analyses by Douglass North, James Q. Wilson, Oliver Williamson and others. But it reaches virtually the same conclusions. They are twofold. First, like Williamson before him (and in NSDS), Dixit notes that in political or bureaucratic institutions where property rights are poorly defined or costly to enforce, apparently inefficient arrangements may not actually be ‘remediable.’ That is, there may be no feasible (as opposed to hypothetical) alternative that is superior in all respects.
On its own, this argument might be seen as compatible with the Panglossian efficiency conclusions. But second, these political arrangements also incorporate incentive schemes that typically are significantly less powerful than those found in economic markets. Hence, far from being exaggerated, the information problems inherent in democracies constitute serious theoretical grounds for doubting Wittman’s assertion that political markets approximate the degree of competition in economic markets. Accordingly, the presumption of constrained efficiency in politics should be so much weaker.