Content area
Abstract
This thesis explores the ramifications of the joint federal and state structure of social insurance in the United States on the level and distribution of welfare spending. I focus on three effects of this fiscal federalism: the income effect of federal grants, the price effect of federal subsidies, and the spillover effect of state spending.
The first chapter analyzes the effects of federally-mandated program changes on state budgets. I estimate that the entire state burden of mandated spending is borne by decreases in other public welfare spending. These reductions are due in part to the substitutability of programs in the voter utility function and in part to a flypaper-like "stickiness" of spending by budget category. States with greater racial differences and states with less generous neighbors reduce other public welfare spending by even more, alleviating the burden the medical expansions impose on their taxpayers.
The second chapter estimates state sensitivity to the federal matching rate for the program Aid to Families with Dependent Children. Using legislative changes between 1948 and 1963, I find that states are much more sensitive to the match rate than previous estimates indicated, with elasticities of AFDC spending of between $-$.5 and $-$1.0. The reduction in AFDC spending caused by an increase in the required state share is driven in part by reductions in the benefit amount per recipient and in part by reductions in the number of recipients. Higher prices drive an increase in recipiency requirements, and lead states to impose caps on benefits given out by localities.
The third chapter estimates the degree to which state spending is influenced by the spending of neighboring states. I use variation in medical spending to abstract from unobserved correlation between neighbors and estimate the responsiveness of state spending to neighbors' spending to be between 37 cents to 88 cents on the dollar. These results are similar to those of past studies, but I obtain more consistent results different measures of neighborliness, including not only those based on geography, income, and percent of the population that is black, but also interstate mobility and population-weighted proximity.
