Research

• Misperception and Income Response to Means-Tested Programs: Evidence from the College Financial Aid Implicit Tax 

(Job Market Paper 2023-24)

Means testing of college financial aid creates large implicit tax rates that affect millions of middle income families each year. These implicit tax rates can exceed 30pp, with middle income families earning between $40k and $140k facing the highest rates. I present the first estimates of the elasticity of parent income with respect to these taxes. I use Free Application for Federal Student Aid (FAFSA) records covering the universe of aid applicants in California from 2010-2021 and a series of difference-in-differences designs that exploit year-over-year changes in a family’s effective tax rate. I estimate an elasticity of taxable income (ETI) for middle income families of 0.10. Responses are larger among families with a high share of flexible non-labor income (ETI=0.47), high assets (ETI=0.36), or higher income ($140k to $240k; ETI=0.28). The ETI is a sufficient statistic for the efficiency cost of a tax under the null that all individuals correctly understand the tax. However, I show based on an online survey that I conducted that many families misperceive the financial aid tax schedule. I show theoretically that when individuals misperceive a tax, the efficiency cost of the tax is affected by two channels: a bias channel measuring the average degree of misperception; and a variance channel measuring heterogeneity in misperception. The survey indicates that parents are not biased on average, but that their perceived tax rates are highly variable. Accounting for misperception, I estimate that means testing in college aid produces an efficiency cost equal to 2.3% of total aid among middle income families. Because of the substantial heterogeneity in perceived tax rates, I estimate that misperception increases the efficiency cost of means testing college aid by $18.8 million per year among middle income families in California alone.

• Achieving Race-based Equity through Race-blind Policies: Evidence from a Local Preference in College Admissions 

With Johnny Huynh

We study the enrollment and equity effects of a unique college admissions policy: a preference in admissions for students applying from local high schools. In the mid-2000s, 18 California State University (CSU) campuses were mandated to prioritize applicants from local high schools; however, only nine campuses offered a meaningful local preference in practice, which we call “adherent” campuses. We estimate the effects of exposure to a local admissions preference using a difference-in-differences design that interacts an indicator for being local to an adherent campus with an indicator for being pre or post policy implementation. Our results show that the policy induced students to enroll at their local campuses, without evidence of crowd-out from other public four-year colleges in California. Effects are only found for students from high schools with a high share of underrepresented minority (URM) students. As a result, the formally race-blind local preference policy nearly eliminates the pre-existing gap in enrollment at California public four-year colleges between students from high and low URM share high schools.