For our empirical approach, we use fixed and random effects models. Hausman
(1978) specification tests indicate that in some cases we cannot reject the hypothesis that
the coefficients from the fixed effects and random effects models are different. In other
cases, this hypothesis is rejected.