Bias-corrected estimation of panel vector autoregressions

•Least-squares estimation of the fixed-effect panel VAR is asymptotically biased.•Least-squares estimation of the impulse-response function is asymptotically biased.•Simple bias corrections are derived for the panel fixed-effect VAR and the IRF.

We derive a bias-corrected least-squares estimator for panel vector autoregressions with fixed effects. The estimator is straightforward to implement and is asymptotically unbiased under asymptotics where the number of time series observations and the number of cross-sectional observations grow at the same rate. This makes the estimator particularly well suited for most macroeconomic data sets.