Janus-Faced Consequences of Ending Agency Fees: Causal Evidence from the Janus v. AFSCME Shock

Abstract

This paper estimates the effects of the June 27, 2018 Janus v. AFSCME Supreme Court decision on union free-riding and wages of state and local government employees. Using Current Population Survey data from 2002 to 2025, I leverage the Janus shock in a two-way fixed effects difference-in-differences design that compares 23 treated jurisdictions to 22 control states with pre-2002 public-sector right-to-work (RTW) laws. Free-riding was largely unchanged for the first five years but increased by 3.05 percentage points (p < 0.05) by FY2024–2026. Wages fell 2.97% (p < 0.01), corresponding to $13.9 billion in annual lost earnings. These findings are robust to synthetic difference-in-differences, triple-differences, and Callaway-Sant’Anna and Borusyak-Jaravel-Spiess staggered estimators that incorporate four legislative RTW adopters. Three mechanism tests rule out a financial channel in which government employers respond to the realized loss of union capacity. First, the wage decline began immediately, reaching 2.07% within the first post-treatment year; second, it failed to scale with predicted union revenue losses; third, its magnitude was statistically indistinguishable for covered and non-covered workers. Instead, my findings point to a forward-looking threat channel: government employers compressed wages in anticipation of the union political-leverage decline Janus implied. This paper provides the first causal test of the wage prediction central to the political-leverage theory of public-sector unions. It also yields the first cleanly identified estimates of public-sector RTW effects, since Janus was judicially imposed rather than legislatively adopted.