Assessing Empirical Approaches for Analyzing Taxes and Labor Supply

Recent surveys on the labor-supply responses of men document a divergence in the estimates of substitution and income effects obtained using various estimation approaches. Generally, studies accounting for nonlinear tax schedules in a static setting via a piecewise-linear approach produce estimates that typically imply higher substitution and lower income responses than are suggested by empirical work applying other approaches. This paper demonstrates that maximum likelihood estimation of a consumer-choice problem with nonlinear budget sets implicitly relies on the satisfaction of inequality constraints that translate into behaviorally meaningful restrictions. These constraints arise not as a consequence of economic theory, but instead as a requirement to create a properly defined statistical model. In the analysis of piecewise-linear budget sets, the implicit constraints required by maximum likelihood in estimation amount to imposition of Slutsky conditions at all wage-income combinations associated with kink points. In the analysis of differentiable budget sets, the tacit constraints invoked by maximum likelihood also involve inequality restrictions on Slutsky terms. The empirical work presented in this study supports the contention that these implicit constraints play a major role in explaining the discrepancies in estimates found in the literature on men's labor supply.