The effects of education policy on college entry and household savings.

This dissertation focuses on the consequences of state college admission policies and federal financial aid. During the 1990s, several states eliminated their public colleges' affirmative action admissions. The first chapter examines the efficacy of new programs that grant admission for a fixed percentage of each high school's graduating class. Using data from the National Education Longitudinal Study, I find that preferences given to minority applicants are large and that minority enrollment in top-tier institutions would fall substantially. Furthermore, I find that there are not sufficient numbers of minorities in the top-x percent of their high school for the expected recovery from an x-percent program to be very large. Using a top-ten percent criterion, underrepresented minorities would regain only one-fifth of the share of acceptances that was lost with the elimination of affirmative action. The second chapter estimates students' application responses to changes in affirmative action admissions. A censored negative binomial model is used to predict students' applications to various types of colleges. Data are taken from a ten-percent national sample of SAT test takers in the years 1996 to 2000. The effect of the changes in state policy on student applications are simulated and compared to the actual application responses. The simulation predicts a strong decrease in minority applications and a small increase in white and Asian-American applications to top-tier public colleges as a result of changes in the probability of acceptance and changes in the colleges' racial compositions. The third chapter reevaluates the savings responses to the implicit tax on assets imbedded in the allocation of federal college financial aid. The prior literature on this topic uses strong and unrealistic assumptions leading to overestimates of the savings responses. Using data from the Survey of Income and Program Participation, a marginal tax on asset accumulation is estimated using more realistic assumptions. I find that the implicit asset tax has a small and often insignificant effect on various types of assets. In general, the parameters predict less than a three-percent reduction in asset holdings, which is far lower than prior estimates.Ph.D.EconomicsEducationEducation financeLabor economicsSocial SciencesUniversity of Michigan, Horace H. Rackham School of Graduate Studieshttp://deepblue.lib.umich.edu/bitstream/2027.42/132455/2/3058003.pd