Tax reform is a good opening case for exploring the interactions among institutions, communication, and policy processes because it discourages ideological claims that competitive markets necessarily produce felicitous results. The U.S. case was notable for the open competition among parties, interests, and the president for definition of a policy situation that was covered extensively (and in some economic detail) by the news media. Yet when the reform had been signed into law, few of the players, whether elites or publics, ended up happy with the outcome. Although some experts pronounced the results impressive, both as a feat of economic redistribution and as an accomplishment of bipartisan politics, few of the leaders responsible concluded that their political capital had been well spent. Perhaps more disturbing, the majority of the public regarded the tax law as an example of the interest group struggles and party politics that had corrupted Washington. T h e source of the unhappiness, we think, was a generalized inability of any faction to control the rhetorical definition of the policymaking situation, resulting in diminished control over the course of the policy process and its outcome, and decidedly little control over how the process and the outcomes were communicated to the public. Such a competitive idea market may have discouraged political domination by particular factions but without the fruits of Madisonian consensus in the bargain.

This American experience with tax reform illustrates the capacity of poorly integrated or coordinated political institutions to reduce a good idea to hollow rhetoric in the minds of many citizens. Perhaps the greatest irony of this process is that government actually implemented the sort of public preferences observed by Page and Shapiro (1992) in their studies of opinion and policy, yet neither the proceedings nor the results were communicated in ways that led citizens to confer legitimacy on either their government or their representatives. The comparison case of Sweden suggests that different institutional arrangements, particularly involving parties and interest representation, permit different strategies for managing public debates with implications for both the coherence of resulting policies and levels of public support.