9511557 Jakubson Daughters of mothers who were welfare recipients are, other things equal, more likely to be welfare recipients themselves. In order for policy to address the issue of welfare transmission, policy-makers need to know more than the fact that there is an intergenerational link in welfare dependency. The nature of the transmission mechanism has important implications for which types of policy interventions will be effective. This project will partition the intergenerational correlation into components attributable to unobserved heterogeneity, associated with the transmission of preferences and/or constraints (which would be present in the absence of parental welfare participation), and state dependence, or scarring (which is directly attributable to parental welfare participation). Data from the Panel Study of Income Dynamics (PSID) will be used to investigate these questions. Two different statistical models are developed and tested using the PSID. One is based on a sequential choice approach in which the welfare participation decision is conditional on the fertility decision. The other allows for concurrent choice among fertility and welfare status. Within each model two approaches to distinguish the heterogeneity and state dependence components are used. One approach allows for both family- and individual- specific heterogeneity components and relies on a timing assumption for the decomposition. In the other, the two components can be identified without a timing assumption, but at the cost of allowing for only a family-specific effect.
