Firm-Level Labor-Shortage Exposure

Jarrad Harford,Qiyang He,Buhui Qiu

DOI: https://doi.org/10.2139/ssrn.4410126

2023-01-01

SSRN Electronic Journal

Abstract:This paper presents a novel approach to measure labor-shortage exposure at the firm level by leveraging finance-specialized machine learning techniques. Specifically, we use FinBERT to extract information from earnings conference call transcripts to develop a reliable measure of labor-shortage exposure. We demonstrate the validity of our measure by showing that states with higher levels of labor-shortage exposure experience lower future unemployment rates but higher wage growth, while firms with higher labor-shortage exposure have greater future staff expenses. Furthermore, we exploit the 2017 U.S. immigration policy reform as a natural experiment to demonstrate that labor-intensive firms increase their labor-shortage exposure following a shock that reduces foreign labor supplies. We find that firms with labor-shortage exposures experience significantly lower cumulative abnormal returns within three days following their earnings conference calls. Moreover, firm-level labor-shortage exposure negatively predicts cross-sectional stock returns and operating performance over the next year. Firms respond to labor shortage by substituting labor with capital and R&D investments, and by producing more process patents to improve their production processes. Finally, we find that firms with high labor-shortage exposure prior to the COVID-19 pandemic perform significantly worse than other firms during the crisis. These results provide important implications for firms and policymakers seeking to address labor shortages and their impact on firm performance.