Research

Publications

• Understanding the Racial Employment Gap: The Role of Sectoral Shifts with Carter Bryson, Labour Economics, 2023. [Pre-print version.] [+] Abstract

  Employment outcomes of Black men worsened significantly relative to White men during the second half of the 20th century. We explore the role of broad sectoral shifts in labor demand over this period in explaining this trend. We first quantify changes in local employment and population in response to local labor demand shifts for both groups of workers. We then combine our estimates with a stylized model that incorporates frictional local labor markets and imperfect mobility across markets. Our framework enables us to aggregate local responses while accounting for geographic mobility and regional employment composition. We find that sectoral reallocation can explain around one-fifth of the total exacerbation in the employment-to-population ratio differential between Black and White men over 1970–2010. Out-migration from harder-hit markets, while large, does not mitigate the impact of negative labor demand shifts. We also find that most of the predicted change in the employment differential is due to differential response rather than differential exposure to shifts across two groups.

Working Papers

• Duration Dependence and Heterogeneity: Learning from Early Notice of Layoff [Online Appendix] [arXiv Link] [Slides] (Under Review) [+] Abstract

  This paper presents a novel approach to distinguish the impact of duration-dependent forces and adverse selection on the exit rate from unemployment by leveraging variation in the length of layoff notices. I formulate a Mixed Hazard model in discrete time and specify the conditions under which variation in notice length enables the identification of structural duration dependence while allowing for arbitrary heterogeneity across workers. Utilizing data from the Displaced Worker Supplement (DWS), I employ the Generalized Method of Moments (GMM) to estimate the model. According to the estimates, the decline in the exit rate over the first 48 weeks of unemployment is largely due to the worsening composition of surviving jobseekers. Furthermore, I find that an individual’s likelihood of exiting unemployment decreases initially, then increases until unemployment benefits run out, and remains steady thereafter. These findings are consistent with a standard search model where returns to search decline early in the spell.

• Firm Organization and Spillovers with Esther Lee [+] Abstract

  This paper studies whether the organizational decisions of new entrants in a market are influenced by the hierarchical structure of their incumbent peers. Using matched employer-employee data from Brazil, we classify establishments into one to four-layer entities and examine how a new entrant’s decision to add an organizational layer varies with the average number of layers of other establishments in their industry and location. To address the potential endogeneity of peers’ layers, we construct an instrument based on layers of other establishments in peers’ firms that operate in different markets. We find that new entrants are twice as likely to add a layer within five years if their average peer has one more layer at the time of entry. Our results suggest that organizational structure spillovers can provide a new source of agglomeration advantages. We also find that the influence of peers is stronger in industries that are more similar. Additionally, we show that new entrants with high-layer peers hire more workers from within the market in the newly created layers, indicating personnel exchanges as a mechanism for organizational spillovers.

Works in Progress

• Discrete Time Mixed Proportional Hazard Model (Draft coming soon!) [+] Abstract

  This paper presents a constructive proof of identification for the discrete-time variant of the Mixed Proportional Hazard (MPH) model. Leveraging the identification result, a root-N consistent non-parametric estimator is proposed based on the Generalized Method of Moments (GMM). The estimator is applied to firm-exit data, demonstrating its convenience and effectiveness when dealing with event-to-time discrete data.