A RATIONAL EXPECTATIONS MODEL OF LOCAL PAYROLL TAX REGIMES

A rational expectations model of straight-time and over-time labor demand is developed in order to determine whether expectations concerning a proposal to increase the payroll tax rate influenced firms' contemporaneous employment decisions. The model is estimated with monthly firm data for the City of Toledo, Ohio aggregated into ten major industry categories. The data series cover the period 1974 through 1981.

An empirical investigation of the Granger-causal structure of the model and the appropriateness of the cross-equation parameter restriction implied by the rational expectations hypothesis is carried out. A theoretical specification of the way in which expectations about future tax payments enter the employment decision rule was derived from a model of the payroll tax process. This specification is empirically tested by simulating two versions of the model over a nine-month period following the announcement of the proposal to increase the tax rate. The first version of the model is based on the assumption that firms' expectations about future tax payments are unaltered by the announcement, resulting in a continued use of historically established employment decision rules. The second version of the model includes a theoretical specification of changes in expectations about future tax payments after the announcement. The mean-squared error of prediction for the second version of the model during this nine-month period is lower in five of the fifteen employment decision rules estimated and significantly lower in two of these fifteen cases. This empirical test of the particular expectations specification developed here provides only weak evidence in support of the specification. A separate test for the existence of an announcement effect on employment decisions prior to the actual implementation of the payroll tax rate increase is carried out. This test is separated from the rational expectations hypothesis in two ways. First, an unconstrained version of the model is employed which does not impose the cross-equation restrictions embodied in the expectations model. Second, this test is free from any particular specification of the payroll tax process since it uses a general transition function which allows the parameters in the employment rules to vary over the transition period between announcement of the tax hike proposal and the actual implementation of the higher tax rate. . . . (Author's abstract exceeds stipulated maximum length. Discontinued here with permission of school.) UMI