Inference on time-invariant variables using panel data: A pretest estimator

For static panel data models that include endogenous time-invariant variables correlated with individual effects, exogenous averages over time of time-varying can be internal instruments. To pretest their exogeneity, we first estimate a random effects model includes all (Mundlak, 1978; Krishnakumar, 2006). Internal instruments are then selected if parameter is statistically different from zero Hausman and Taylor, 1981). Finally, Hausman-Taylor (1981) using these We evaluate the biases currently used alternative estimators in Monte-Carlo simulation: repeated between, ordinary least squares, two-stage restricted Oaxaca-Geisler estimator, fixed effect vector decomposition, (restricted generalized squares).