Labor unionization and real earnings management: evidence from labor elections

Abstract

By exploiting the local randomness in close-call labor elections, the authors find a negative impact of labor unionization at a firm on its real earnings management (REM). The finding suggests a managerial pressure effect of increased labor power. In a local regression discontinuity (RD) analysis, firms that narrowly pass the 50% threshold show a significant decrease in REM, relative to their peers that narrowly fail. This effect is stronger for firms headquartered in right-to-work states and when managers have less pressure to manage earnings. Evidence from a global parametric RD analysis and a multivariate OLS test using industry-level unionization measures confirms the external validity of results in local RD analysis. Overall, the research sheds new light on the economic consequence of labor unionization on employers' accounting decisions.