Structure-induced equilibrium and legislative choice

Conclusion

Professor Tullock has raised a central question in the confrontation between abstract models of PMR and majority rule as practiced in real institutions. We believe the decision making stability of real-world legislatures lies in the way these legislatures institutionalize majority rule. Logrolling, vote trading, coalition formation, and bargaining are red herrings in this argument. Rather, it is the restrictions on such legislative exchange that promote structure-induced equilibrium. Put differently, institutional arrangements place constraints on the completeness of the majority rule relation by restricting social comparisons.

The framework developed here shows that an assumption implicit in the discussions of many majority rule theorists fails to hold. In part, the implicit rationale for focusing upon PMR was that results proved for this rule were presumed to hold forany institution based on PMR. In one sense this remains true, namely, that the majority rule win sets,W(x), are everywhere non-empty. In another sense, however, it is not true that all properties of institutions based upon majority rule are inherited from PMR. The theory outlined above shows that stability may not be as elusive as theorists of PMR have concluded.

The concept of equilibrium developed in the last section incorporates the major features of prominent choice institutions as well as capturing the special cases in the literature cited in Section II. We now turn to a brief discussion of future work. We address the question that remains, in our opinion, the salient one in the study of institutions and their effect on policy choice, namely, understanding the factors governing the choice of one institutional arrangement over another.

Throughout this paper, we have distinguished agreements that transform the rules from agreements (or vote-trades) that take place within a given set of rules. In principle, anything attainable under the former could also be attained under the latter if there were some form of mechanism to enforce vote-trades as contracts. Under such a rubric, complex legislative agreements in the form of contingent contracts achieve the desired result without resorting to the institutionalization of a rule. In practice, however, there are several problems with vote-trading agreements as contracts. First, the cost of writing these contracts is often quite high due to the number of potential contingencies for which provision must be made. Second, and more important, PMR lacks an enforcement mechanism. Individual parties to contracts in market settings have recourse to the courts. This provides protection beyond the assurance of good faith and brand names. No comparable institution exists within the legislature to supplement the natural though imperfect brand name phenomenon (i.e., that of ‘keeping one's word’ to preserve and enhance credibility for future trades).

While the legislature could create a court or committee to monitor contracts and enforce agreements, alternatively, it could simply impose a rule binding upon everyone which insured the outcome sought. Of the two alternative institutions, the latter probably economizes on transaction costs, particularly for those situations that recur with some frequency. With a rule, a new contract need not be negotiated each time between new sets of players. Moreover, a contingency clause might easily be appended to a rule to cover cases where there is widespread agreement that it is inappropriate. For example, in the Congress a special majority may vote to suspend the rules (note that if only a simple majority were required, then this would be no different from PMR).¹ In sum, logrolling solutions to the problem of forging agreements are unworkable because they lack enforcement mechanisms. Logrolling, then, cannot constitute an answer to the question, ‘Why so much stability?’²

If institutional rules are to constitute an answer to Tullock's stability question, then we must confront the manner in which those rules are chosen. There are very few theories about the choice of rules — exceptions include Buchanan and Tullock (1962), Buchanan (1975, 1979), and contributions in the property rights literature. Even in the absence of a theory, we may still worry that constitutional choice processes (the choice of rules) are vulnerable to the same instabilities found in PMR. We term this the ‘Riker Objection’ since this issue was recently posed by Riker (1980). If institutional constraints create equilibrium — that is, if transformations of a PMR institution into a non-PMR institution create a situation of equilibrium from one without an equilibrium — then preferences over outcomes lead naturally to aninduced set of preferences over institutional arrangements. In this sense, an individual prefers one institution over another if he prefers the equilibrium policy state of one over the equilibrium (or unpredictability) of the other. In the case of multiple equilibria, an individual prefers the institution that yields the highest expected utility given a probability distribution over equilibrium states (Plott, 1972).

As long as preferences for policy states differ, then preferences over institutions with differing equilibrium states (distribution of equilibria) should also differ. The Riker Objection suggests that a simple extension of McKelvey's Chaos Theorem predicts endless cycles here so long as PMR governs the choice over institutions. In this sense, the existence of institutions and their stability must remain, like policy choices under PMR, tenuous — what Riker calls ‘unstable constants.’ Nevertheless, empirically we observe institutions persisting for long periods; in light of the Riker Objection, Tullock's question applies at this level as well.

We may make several observations that imply an attenuation of endless cycling at the institutional-choice level. First, typically, non-PMR rules govern the choice of new rules. Second, it is risky to attempt to change the status quo contrary to the interests of those currently in control. Since failure may lead to the imposition of sanctions, expected gains must be weighed against the certainty of these sanctions. While this does not rule out changes, it will reduce the number of attempts. This is surely the conclusion to be drawn from a reading of the history of the U.S. Congress. The comparison between choice in this setting and the McKelvey world, then, is not parallel since proposals are costless to make in the latter but not in the former. Finally, there often exists a well-defined status quo alternative. In the case of the social contract, the status quo is the Hobbesian state of nature. For the case of the U.S. Constitutional Convention, it was the Articles of Confederation (Riker, 1979). In these and similar settings, even though there may be no formal rule that the status quo must literally be voted last; this restriction nevertheless may hold de facto. Consequently, the constitutional outcome is either the status quo ante or an alteration that cannot be vetoed, i.e., an element in the ‘win set’ of the status quo. With these qualifications in mind, the effect of the Riker Objection is mitigated. Even at the constitutional level, then, restrictions on the ability of individuals to make proposals may induce equilibrium.