Two questions:
1) I was told that the Mundlak version of the Hausman test is low power, could anyone lend a hand to what can be inferred from this? Is it that its result (whether to use FE or RE is unreliable)?
2) Similarly if the Hausman suggests use of FE but I use RE anyway due to it better estimating my research goals, what is the inference econometrically from this on the estimation? Does it mean using a RE model is a less efficient but still consistent estimator?
1) I was told that the Mundlak version of the Hausman test is low power, could anyone lend a hand to what can be inferred from this? Is it that its result (whether to use FE or RE is unreliable)?
2) Similarly if the Hausman suggests use of FE but I use RE anyway due to it better estimating my research goals, what is the inference econometrically from this on the estimation? Does it mean using a RE model is a less efficient but still consistent estimator?