Working Papers

Welfare Added? Optimal Teacher Assignment with Value-Added Measures

with Julian Betts, Nathan Mather, and Michael Ricks

Under Review

Abstract: We study how teacher “value added” should inform optimal teacher-assignment policy. Our welfare-theoretic framework illustrates (1) how theoretically optimal assignments leverage variation in teachers’ impacts both across student types and across different outcomes, and (2) how empirically optimal assignments trade off improved targeting from estimating richer student heterogeneity against a growing winner's curse. In practice, optimal assignments use limited student types (only lagged achievement) and multiple outcomes (not just math). Even after correcting for policy overfitting, assignments raise average present-value earnings by $3,000 and increase lower-achieving students’ earnings by 66–142% more than benchmark policies using homogeneous effects, single-subject heterogeneity, or teacher deselection.

Presentations: NBER Public Economics Spring Meeting; NBER Economics of Education Spring Meeting; American Economic Association Annual Meeting; All California Labor Economics Conference

Second-Best Reassignments Outperform other Proposed Policies
Gains to below-median versus above-median students under alternative teacher-assignment policies

Broader Horizons: The Long-Run Impacts of Exposure to New Places

with Michael Ricks

Previously circulated with pilot results (~900 responses), updated with scale results (~18,000 responses)

Abstract: We study how volunteering in different cultural environments shapes individuals’ social attitudes and actions using variation from the location assignments of volunteer missionaries for The Church of Jesus Christ of Latter-day Saints. Administering an original survey to former volunteers, we find noticeable changes in attitudes about underrepresented minorities, political out-partisans, and women in the workforce. We find that volunteering in places with high Black or Latino populations increases positive sentiment towards these group and increase real-life behaviors like residing in more diverse zip codes and supporting social justice causes. These effects persist for decades after missionary service. Preliminary results also suggest that exposure to locations with more liberal politics seems to reduce the gap in sentiment towards Republicans and Democrats and the gap in related actions like voting and donations. Although we don’t find large effects on attitudes about working women, women who volunteer in more gender-equitable places may have more children.

Presentations: American Economics Association Annual Meeting; Southern Economic Association Annual Meeting; UCSD/UCLA/CIFAR Conference on Economic Development, Political Economy, and Culture; Western Economic Association Graduate Student Workshop; Rising Scholars Conference

Mission Assignment is Conditionally Independent
Change in mission assignment probability by application information: language, international, demographic, and leadership variables

(a) Application Information

Change in mission assignment probability by unobserved baseline characteristics: parental demographics and childhood zipcode characteristics

(b) “Unobserved” Baseline Characteristics

Work in Progress

Meeting Every Qualification: Job Ad Content and Applicant Demographics

with Amanda Bonheur

Pilot Complete, Scale Experiment Funded by the Russell Sage Foundation (Presidential Grant)

Abstract: Despite years of policy and revised corporate practice intended to correct inequality in the hiring process, application gaps persist for women and individuals from underrepresented racial minority groups. This study explores whether it is possible to narrow this application gap and promote diversity in the applicant pool by modifying the language around qualification requirements in job ads. We do so using a large-scale, “reverse audit study” field experiment where we randomize the content of job ads and observe job seeker behavior. Specifically, we established a non-profit recruiting firm to act as an intermediary in the job search process. This firm reposts real job ads and collects information from job seekers interested in applying. We randomize whether we encourage people to apply even if they don’t meet all of the listed qualifications and whether we inform them that companies routinely hire individuals who do not have all qualifications. In two waves of a small, pre-registered pilot, embedding this language in the job ad raised the share of job seekers who continued with the application by nearly 6 percentage points and pulled highly but not perfectly qualified women into the applicant pool, with no change in the qualifications of male applicants. Encouragement alone mainly raised men’s confidence, whereas adding information about hiring practices sharply increased women’s stated likelihood of applying and closed the gender gap in how strictly job seekers expected the employer to evaluate them. These effects appear to operate by changing perceptions of the specific employer rather than of the job market more broadly.

Presentations: W.E. Upjohn Institute Invited Seminar

What are the Pecuniary Returns to Foreign Language Acquisition?

with Michael Ricks and Alison Doxey

W.E. Upjohn Institute Early Career Research Award

The Impact of Generative AI on Labor-Market Matching

with David Arnold

The Long-Run Impact of Prejudice: Evidence from WWI Casualties

with Joe Price, Dan Rees, and Kerwin Charles