Corporate Decision-Making under Economic Policy Uncertainty: Evidence from U.S. Public Firms around the 2020 Recession

This paper studies the association between U.S. economic policy uncertainty (EPU) and corporate investment and cash holdings using a large panel of publicly listed firms constructed from SEC EDGAR XBRL filings. The preferred sample contains 2,532 firms and 20,539 firm-year observations over 2011–2024, with particular attention to the period surrounding the February 2020 NBER recession peak. In the preferred specification with firm fixed effects and no year fixed effects, a one-unit increase in lagged annual EPU is associated with a decline in capital expenditure scaled by assets (coefficient = −0.000046, firm-clustered SE = 0.000005, conventional p < 0.001) and an increase in the cash-to-assets ratio (coefficient = 0.000212, SE = 0.000023, conventional p < 0.001). Both associations retain the same signs when leverage is added on the smaller subsample of 1,716 firms and 8,786 firm-year observations. Because the EPU index is strongly persistent, the effective number of independent time-series observations is limited, so conventional confidence intervals understate the true uncertainty around these estimates and reported statistical significance should not be read as settled. The sample contains only one NBER recession peak, so the results are best read as evidence on firm behavior around the 2020 downturn rather than a general pre-recession pattern. The main contribution is a transparent EDGAR-based design that confronts the collinearity problem that arises when a national time-series uncertainty index is paired with year fixed effects.