The Labor Supply Effects of Government Transfers in Cash and in Kind

While there is a large body of literature which addresses the labor supply effects of cash transfer programs, almost all of these studies have ignored important features of our welfare system. Over the past several decades we have been increasingly relying on in-kind transfers to redistribute income-relatively little effort has been devoted to understanding the labor supply effects of these programs. Also, many studies have analyzed a single welfare program in isolation when many individuals are multiple program participants. To the extent that programs interact with one another, it is important to consider the group of transfers that an individual may receive, not one program in isolation.

To address these issues, we describe the effects of the AFDC, Food Stamp and Public Housing Programs on the individual's budget space, describe and implement a method to estimate a structural model that allows us to address the effects of these programs; and use the estimation results in policy simulations to analyze alternative programs. We specify a model that assumes that individuals maximize a displaced CES utility function by choosing housing, leisure and other goods consumption subject to a budget constraint that accounts for the three transfer programs. In order to estimate the indifference map parameters, we begin by adding stochastic structure by assuming that there are variations in tastes and stigma associated with welfare program participation. We then estimate wages for nonworkers and housing consumption for public housing recipients since these quantities are not directly observed. Once these quantities are obtained, we estimate the distribution of indifference map parameters with a method of simulated moments estimator due to the complicated error structure.

The estimates that we obtain suggest that we may have seriously misspecified the budget constraint. Our model predicts that individuals have very weak preferences for housing and leisure relative to other goods. We believe that the misspecification results from the fact that we have not adequately captured the rationing of public housing. We conclude by outlining a method for simulating the effects of alternative programs that takes into account the estimated distribution of indifference map parameters.