The agenda control models of Niskanen (1971), Romer and Rosenthal (1978), Denzau, Mackay and Weaver (1979), and others should be quite familiar to most of the readers of this journal. The essence of these models is the monopolization of the power to place proposals before public approval. These agenda monopolists are invariably bureau chiefs with a vested interest to budget maximize subject to voter approval constraints. Therefore, hapless voters find themselves confronted with various 'take it or leave it' choices and are induced to approve excessive levels of public services.
The most severe case of agenda 'extortion' results in the Niskanen model where voters approve public service levels twice the optimal quantities preferred by the decisive voter.1 Other papers have shown how the ability of the agenda monopoly to extract 'bureaucratic rents' depends upon the reversion level (Romer and Rosenthal, 1978), spending/tax limits (Denzau, Mackay and Weaver, 1979) and the budgetary mix (Buchanan, 1963). These models, however, implicitly or explicitly preclude the possibility of voter mobility.
How might migration effect agenda control behavior? What happens in a world when people can vote 'with their feet' in addition to the ballot? The answer, of course, depends upon many parameters of the model including migration cost, the distribution of voter preferences, the number of communities in the relevant market area, and the competitiveness/collusiveness between the various agenda controllers. 2
Intuitively, voter mobility might alter the nonmigratory results of Niskanen (1971) and Romer and Rosenthal (1978) in a number of ways. First, agenda controllers may not wish to extract the maximum possible * The author wishes to gratefully acknowledge the helpful comments and assistance from his dissertation committee including