The Effect of High School Exit Exams on Graduation, Employment and Wages (Job Market Paper), with Kevin Lang [pdf]
Over the last three decades, a number of states have required students to pass an exit exam in order to earn a high school diploma. In this paper, we investigate whether such exit exams have an effect on high school graduation, GED receipts, employment outcomes and the wage distribution. We construct a state/graduation year cohort dataset using the Current Population Survey data and information on the timing and difficulty of exit exams in different states. Using this dataset we analyze within-state variation in outcomes overtime. Overall, we find relatively modest effects of high school exit exams. We do not find consistent effects on graduation rates for exit exams that assess academic skills taught below the high school level; however, we find that more challenging standards-based exams reduce graduation rates by about one percentage point. For minorities, the adverse effects of exit exams on graduation rates are larger but less precise. Although students have multiple retesting opportunities, we find that exam-takers are more likely to drop out of high school before these opportunities are exhausted. We also find that higher dropout rates are not fully offset by an increase in GED diploma receipts. Our analysis of labor market outcomes shows that exit exams increase employment rates; however, exit exams do not change the distribution of wages.
Rank-Based Admissions and Academic Performance: Evidence from the Texas Top 10 Percent Law [pdf]
In 1998, state universities in Texas began using high school class rank as the sole factor in university admissions. This policy was implemented to increase enrollment of minority and economically disadvantaged students, but it generated criticism that such beneficiaries of rank-based admissions lack the academic preparation necessary to perform well in college. I test this claim by analyzing academic performance of rank-eligible students who attended UT Austin before and after the law. To account for grade inflation I use a difference-in-differences framework with students not eligible for rank-based admissions as controls. The difference-in-differences estimates may be overstated, however, because academic quality of the control group may have increased after the law. I use propensity score matching methods to correct for this. I also correct for the confounding effects of the GPA ceiling on the difference-in-differences estimates. Both the baseline and the adjusted estimates suggest that mean college GPA of rank-admitted students declined after the law.
The Maimonidesí Rule Revisited: Effect of California Class Size Reduction Program on Test Scores, with Daniele Paserman
In this paper we analyze the institutional and academic consequences of the California class size reduction program. In 1996 California introduced class-size caps of a maximum of 20 students per class. Policies that institute such caps have been used in quasi-experimental regression discontinuity (RD) studies because class-size caps create exogenous variation in class size as a function of school enrollment. However, in the case of California we find that financial and capacity constraints forced schools to adjust their enrollment levels in order to avoid creating classes with very few students. Moreover, this behavior was strongly associated with schoolís overall enrollment and studentsí socioeconomic characteristics. Such premeditated manipulation of enrollment levels violates the assumptions necessary for the implementation of an RD design. Carelessly applying an RD framework to the California policy would indicate that smaller classes are associated with lower test scores. Therefore, researchers should be aware that class-size cap policies may not always be suitable for an RD approach and that using RD design in such cases could generate severely biased results.
Asset Rundown after Retirement: The Importance of Rate of Return Shocks, with Phil Doctor, Eric French, Economic Perspectives, Federal Reserve Bank of Chicago, 2007, Vol. 31, 2nd Quarter (download here)
Right before the End: Asset Decumulation at the End of Life, with Mariacristina De Nardi, Eric French, John Bailey Jones, Economic Perspectives, Federal Reserve Bank of Chicago, 2006, Vol. 30, January (download here)