| Thomas Dee
University of Maryland
Resource allocation and public school quality
FINAL REPORT: The policy determinants of educational attainment: An instrumental variables approach.
This dissertation consists of three essays that examine the policy determinants of educational attainment. In the first essay, I address the relationship between state alcohol policies, teen drinking and subsequent schooling decisions. I show that changes in state minimum legal drinking ages (MLDA) provide a source of exogenous variation in teen drinking that can be used to identify the impact of teen drinking on attainment. However, because no one data set contains within-state variation in both teen drinking and attainment, traditional instrumental variables (IV) procedures cannot condition on important, unobserved state attributes. Therefore, I employ a new estimation technique that generates IV estimates using the moments from two samples. Using data from the 1977-92 Monitoring the Future (MTF) surveys, I show that teens who faced an MLDA of 18 were substantially more likely to drink than teens who faced a higher drinking age. Using data from over 1.3 million respondents in the 1960-69 birth cohorts of the 1990 Public-Use Microdata Sample (PUMS), I show that teen exposure to an MLDA of 18 had small and statistically insignificant effects on high school completion, college entrance and college completion. Then, using matched cohorts from the MTF and PUMS data sets, I report two-sample instrumental variables (TSIV) estimates of the effect of teen drinking on educational attainment. These TSIV estimates are smaller than the corresponding ordinary least squares (OLS) estimates and are statistically insignificant, indicating that the strong correlation between teen drinking and measures of educational attainment does not represent causal effects. In the second essay, I use a unique data set with high school completion rates from the unified school districts in 18 states to present new evidence on the effect of competition from private schools on achievement in public schools. Using the population concentration of Catholics as an instrument, I find that such competition has a positive and statistically significant effect on public school quality. In the third essay, I address the influential but largely unsubstantiated assumption that school districts spend too few resources "in the classroom." The purposeful allocation of resources suggests that traditional educational production functions which relate resource levels to student achievement are misspecified. I estimate the effects of non-instructional and instructional resources on student achievement using two very different classes of instrumental variables. These results indicate that school districts spend too few of available resources on instruction. However, they also demonstrate that money spent on instruction in public schools is highly effective in promoting the level of student achievement when conditioned on the decision to spend outside the classroom.
Back to Funded Dissertation Grants Page |