Within the rational choice tradition there are two now-standard ways to think about institutions.2 The Wrst takes institutions as exogenous constraints, or as an exogenously given game form. The economic historian Douglass North, for example, thinks of them as ‘‘the rules of the game in a society or, more formal- ly, . . . the humanly devised constraints that shape human interaction’’ (North 1990, 3). An institution is a script that names the actors, their respective behavioral repertoires (or strategies), the sequence in which the actors choose from them, the information they possess when they make their selections, and the outcome resulting from the combination of actor choices. Once we add actor evaluations of outcomes to this mix—actor preferences—we transform the game form into a game.
To give an ancient example of a game form from Downs (1957), the actors are n voters and two candidates. The candidates each select a policy position represented by a point on the unit interval, [0,1]. They either do this simultaneously, or choose in a particular sequence but the candidate choosing second does not know the Wrst candidate’s choice in advance of his own choice. (While candidates do not know the choices of other candidates, they do know voter preferences as deWned below.) Voters then vote for one candidate, the other candidate, or abstain. The candidate with the most votes is elected. If each candidate obtains the same number of votes (including none if all voters abstain), then a random device determines which of them is elected. This is a game form, an exogenously provided script that gives the various ways the strategic interaction can develop. If (i) candidates prefer winning to tying to losing, and (ii) each voter i has single-peaked preferences on [0,1] symmetric about his or her most preferred policy, then we have characterized actor preferences and now have a game. The well-known Median Voter Theorem applies: The candidate who locates closest to the most-preferred policy of the median voter wins the election. In game-theoretic language, the Nash equilibrium of this game is for both candidates to locate at the median ideal point and one of them to be randomly chosen as the winner.3,4 Shepsle (1979) called this a structure-induced
equilibrium of the institutional game.
The second interpretation of institutions is deeper and subtler. It does not take institutions as given exogenously. Instead of external provision, the rules of the game in this view are provided by the players themselves; they are simply the ways in which the players want to play. A group of children, for example, might take the oYcial rules of baseball as a starting point to govern their interactions, but then adapt them to speciWc circumstances or tastes. A ball rolling into the creek that borders the Weld, as I recall from my childhood, allows the baserunner to advance only one additional base. On any particular day, however, the kid who brought the bat and ball might insist on a variation to that rule more to his liking—say, a ball in the creek is an automatic home run—and be in a position to induce the others to accept his preference. In this view of institutions, there is nothing exogenous about the rules of the game, and certainly nothing magical. They do not compel observance, but rather reXect the willingness of (nearly) everyone to engage with one another according to particular patterns and procedures (nearly all the time). The institutional arrangements are, in this view, focal (Schelling 1960) and may induce coordination around them. Calvert (1995), one of the intellectual architects of this perspective (see also Schotter 1981), puts it well:
here is, strictly speaking, no separate animal that we can identify as an institution. There is only rational behavior, conditioned on expectations about the behavior and reactions of others. When these expectations about others’ behavior take on a particularly clear and concrete form across individuals, when they apply to situations that recur over a long period of time, and especially when they involve highly variegated and speciWc expectations about the diVerent roles of diVerent actors in determining what actions others should take, we often collect these expectations and strategies under the heading institution .. . (Calvert 1995, 73–4).
Institutions are simply equilibrium ways of doing things. If a decisive player wants to play according to diVerent rules—like the kid who threatens to take his bat and ball home if the rules are not adjusted to his liking—then the rules are not in equilibrium and the ‘‘institution’’ is fragile.
We come to think of institutions (in the ordinary language sense) as scripts that constrain behavior—the Wrst interpretation above—because in many political contexts ‘‘highly variegated and speciWc expectations about the diVerent roles of diVerent actors’’ are involved, and decisive individuals or coalitions are not pre- pared to change the way business is conducted. Calvert’s point, however, is that this does not mean decisive actors are never inclined to push for change. Early in the last decade, for example, a newly elected Labour government in Great Britain, to the surprise of many, transformed the Bank of England from one of the most dependent central banks in the developed world into a much more independent agency. A revision of the Rules of the US Senate—particularly Rule 22 to make it easier to end Wlibusters—has been contemplated on many occasions (Binder and Smith 1996). Twice in the last century there were major changes in the rules to make cloture Wrst possible, and then easier. The Republican majority in the US Senate of the 109th Congress (2005–7) has raised this issue again in the context of the conWrmation of judges and justices.5
There is a third interpretation of institutions (indeed, there are many others) that is decidedly not rational choice in nature; it bears describing brieXy in order to contrast it with the two interpretations just given. I associate it with Sait (1938) and his legacy in various forms is found in the work of modern historical institution- alists. For Sait, institutions are magical. He describes them with the wide-eyed wonderment of someone examining a coral reef for the Wrst time.6 They just form, and re-form, according to complex, essentially unknowable forces. Law, slavery, feudalism, language, property rights—these are the ‘‘ediWces’’ Sait considers institutions. His emphasis diVers from that of the institutions-as-constraint and institutions-as-equilibrium schools of thought described above. Institutions for him are macrosociological practices deWned, and altered, by historical
contingency. There is microanalysis neither of the patterns of behavior they induce and sustain nor of the human attempts to alter institutional properties. There is for him no architect of Roman Law, for example. An institution is an accretion, changing ever so slowly and never by identiWable human agency. Perhaps we need a diVerent name for one of these.
