Abstract
We study how much CEO replacements contribute to firm growth and show that traditional fixed effect estimates substantially overstate CEO influence and create spurious trends due to small-sample bias. We introduce a placebo-controlled, difference-in-differences method using non-changing firms to correct bias in second moments. Simulations confirm the method is robust to persistent shocks, short tenures, and unbalanced panels. Naive estimates on 60,000 Hungarian CEO transitions attribute 40–60% of revenue variance to CEO changes; our debiased method finds 20–30%. The corrected estimates show immediate, persistent effects without false pre-trends.
Please cite as
Koren, Miklós, Krisztina Orbán, and Álmos Telegdy. 2025. "CEO Replacements and Firm Growth: Correcting for Small-Sample Bias in Fixed-Effect Estimates"
