Research

Work in Progress:

Uniform Rate-Setting and the Provision of Quality: Evidence from Financial Advisors 

Second-Year Paper

Why do Registered Investment Advisor firms frequently hire employees who have histories of financial misconduct? Using new data on rate schedules, I show that major firms charge nearly uniform rates across wealth levels, creating substantial differences in revenue per client. I document that misconduct is less prevalent in branches located in high-wealth zip codes than middle-wealth ones, where per-client revenues are lower. A dynamic model of firm quality choice rationalizes these patterns: firms optimally substitute toward lower-cost, lower-quality advisors in less profitable markets. I estimate the model and use it to explore counterfactual competition scenarios. Absent competitive pressure, the share of employees with clean records in the average market would drop from 85% to 79%.

The Cost of Early Access (with Olga Aristova)

Using new hand-collected data, we document a previously unstudied feature of entry-level hiring in the financial services profession: at least 50 major firms operate a parallel hiring market eight months before the standard cycle, designed to recruit candidates from underrepresented demographic groups. We offer an economic explanation for how diversity-conscious recruiting can alter not only who firms hire but also when they hire. Because demographic characteristics are observable ex-ante, competition for these candidates may make it individually optimal for firms to make early offers instead of waiting until all information has been revealed. Exploiting the staggered adoption of early recruitment programs, we find that they did not change the demographic composition of intern cohorts, but did increase the share of interns from universities without established on-campus recruiting pipelines by 16.5 percentage points. These interns became 13 percentage points less likely to return to the firm after their internship and 23 percentage points less likely to remain two years later.