Research

Working Papers

Same Shock, Separate Channel: House Prices and Firm Performance in the Great Recession [pdf]

Combining confidential business-level microdata with housing and banking data, I document large and persistent effects of local house prices on employment at small businesses, and particularly young businesses, during the Great Recession. I show that the effect on entry is important for explaining the disproportionate effect on young businesses, while young firm exit is also disproportionately affected. I then explore the channels through which house prices affect business outcomes. I use local bank balance sheet information to show both young and old firms are sensitive to local credit shocks, with some evidence of a larger effect on young businesses. I then use survey data to show that reliance on either personal assets or home equity is associated with increased sensitivity to house prices. I develop a macroeconomic model that is consistent with these findings where house prices work through two channels: a bank credit supply channel and a housing collateral channel.

Job Tasks, Worker Skills, and Productivity [CES working paper 25-63]

with Cindy Cunningham, Matthew Dey, Lucia Foster, Cheryl Grim, John Haltiwanger, Rachel Nesbit, Sabrina Wulff Pabilonia, Jay Stewart, Cody Tuttle, and Zoltan Wolf

Productivity dispersion within an industry is an important characteristic of the business environment, potentially reflecting factors such as market structure, production technologies, and reallocation frictions. The retail trade sector saw significant changes between 1987 and 2017, and dispersion statistics can help characterize how it evolved over this period. In this paper, we shed light on this transformation by developing public-use Dispersion Statistics on Productivity (DiSP) data for the retail sector for 1987 through 2017. We find that from 1987 through 2017, dispersion increased between retail stores at the bottom and middle of the productivity distribution. However, when we weight stores by employment dispersion, the middle of the distribution is lower initially and decreases over time. These patterns are consistent with a retail landscape featuring more and more activity taking place in chain stores with similar productivity. Firm-based dispersion measures exhibit a similar pattern. Further investigation reveals that there is substantial heterogeneity in dispersion levels across industries.

National Chains and and Trends in Retail Productivity Dispersion [CES working paper 25-64]

with Dominic Smith, Michael D. Giandrea, Cheryl Grim, Jay Stewart, and Zoltan Wolf

Productivity dispersion within an industry is an important characteristic of the business environment, potentially reflecting factors such as market structure, production technologies, and reallocation frictions. The retail trade sector saw significant changes between 1987 and 2017, and dispersion statistics can help characterize how it evolved over this period. In this paper, we shed light on this transformation by developing public-use Dispersion Statistics on Productivity (DiSP) data for the retail sector for 1987 through 2017. We find that from 1987 through 2017, dispersion increased between retail stores at the bottom and middle of the productivity distribution. However, when we weight stores by employment dispersion, the middle of the distribution is lower initially and decreases over time. These patterns are consistent with a retail landscape featuring more and more activity taking place in chain stores with similar productivity. Firm-based dispersion measures exhibit a similar pattern. Further investigation reveals that there is substantial heterogeneity in dispersion levels across industries.

Collaborative Micro-productivity Project: Establishment-Level Productivity Dataset, 1972-2020 [CES working paper 23-65]

with Cheryl Grim, Rachel L. Nesbit, Cody Tuttle, and Zoltan Wolf

We describe the process for building the Collaborative Micro-productivity Project (CMP) microdata and calculating establishment-level productivity numbers. The documentation is for version 7 and the data cover the years 1972-2020. These data have been used in numerous research papers and are used to create the experimental public-use data product Dispersion Statistics on Productivity (DiSP).

Publications

Opening the Black Box: Task and Skill Mix and Productivity Dispersion

with Cindy Cunningham, Matthew Dey, Lucia Foster, Cheryl Grim, John C. Haltiwanger, Rachel L. Nesbit, Sabrina Pabilonia, Jay Stewart, Cody Tuttle and Zoltan Wolf. In: Technology, Productivity, and Economic Growth, 2025.

Financial Markets and Misallocation: The Long and Short of Leverage and Productivity Dispersion [pdf]     Macroeconomic Dynamics, vol. 29, e94, April 2025.

Macro and Micro Dynamics of Productivity: From Devilish Details to Insights [pdf]                                                           with Lucia Foster, Cheryl Grim, John Haltiwanger, and Zoltan Wolf. American Economic Journal: Macroeconomics, vol. 13, no. 3, pp. 142-172, July 2021.

Work in Progress

Allocating Misallocation: Decomposing Measures of Aggregate Allocative Efficiency [slides]             with John Haltiwanger and Zoltan Wolf

We explore sources of measured misallocation using establishment data from U.S. manufacturing industries. We decompose standard productivity dispersion statistics into contributions by revenue margins over costs and plants' input-mix, and establish a link between these components and measured allocative efficiency. The results indicate the components contribute similarly to misallocation. We investigate the following mechanisms that can potentially generate these patterns: heterogeneous production elasticities, changing responsiveness and adjustment costs, and changes in market dynamics. Finally, we show rising misallocation is pervasive, and yet a few industries account for over half of the aggregate decline.  

Do the Means Matter for the Ends? An Evaluation of Optimal Unemployment Assistance under Search Frictions preliminary work with Neil White

We document the prevalence of recall hiring in the wake of recessions, particularly during the recovery after the COVID-19 pandemic. Decompositions by industry and firm size provide suggestive evidence that the Paycheck Protection Program (PPP) may have influenced recall employment. Using a modified DMP model with a temporary layoff state, we explore the effects of different unemployment assistance schemes on aggregates and welfare when workers and firms cannot perfectly bargain over flow payments while on temporary layoff. We focus on whether funds are better channeled through firms, thereby maintaining incentives to remain "in touch" while not working, or given directly to workers. We show the potential for policy to replicate (approximately efficient) outcomes from a benchmark model where bargaining in the layoff state is possible. We use this framework to evaluate policies in response to standard business cycles and to the COVID pandemic in particular. We ask whether the structure of PPP and other unemployment assistance policies that encourage "mothballing" are well-designed to achieve the explicit goals of policymakers.