For a panel data linear regression model with both individual and time effects, empirical studies select the two-way random-effects (TWRE) estimator if the Hausman test based on the contrast between the two-way fixed-effects (TWFE) estimator and the TWRE estimator is not rejected. Alternatively, they select the TWFE estimator in cases where this Hausman test rejects the null hypothesis. Not all the regressors may be correlated with these individual and time effects. The one-way Hausman-Taylor model has been generalized to the two-way error component model and allow some but not all regressors to be correlated with these individual and time effects. This paper proposes a pretest estimator for this two-way error component panel data regression model based on two Hausman tests. The first Hausman test is based upon the contrast between the TWFE and the TWRE estimators. The second Hausman test is based on the contrast between the two-way Hausman and Taylor (TWHT) estimator and the TWFE estimator. The Monte Carlo results show that this pretest estimator is always second best in MSE performance compared to the efficient estimator, whether the model is random-effects, fixed-effects or Hausman and Taylor. This paper generalizes the one-way pretest estimator to the two-way error component model.

This paper derives a feasible GLS estimator for a two-way error component model with serial correlation on both the time effects as well as the remainder disturbances. This estimator is based on two methods, one proposed by De Porres and Krishnaku mar(2013) for deriving the spectral decomposition of a general error component structure and the other based on an inversion trick for the variance-covariance matrix of this model suggested by Skoglund and Karlsson (2001). While the last paper used maximum likelihood methods under the normality assumption, we use method of moments estimators following Baltagi and Li (1991) for the one-way error component model with serially correlated remainder disturbances and its extension by Brou et al. (2011) for the two-way model with serially correlated time effects as well as remainder disturbances. Monte Carlo simulations are performed to compare the performance of these two estimators as well as a bias correction version based on Nobach (2023). Our results find that the method based on the (Skoglund and Karlsson 2001) inverse that is bias corrected a la (Nobach 2023) performs the best in root mean square error (RMSE) as well as mean absolute percentage error (MAPE) and is recommended.

This paper proposes a Bayesian approach to estimate a factor augmented productivity equation. We exploit the panel dimension of our data and distinguish individual-specific and time-specific factors. On the basis of 14 technology and infrastructure indicators from 37 countries over a 10-year period (1998 to 2007), we construct summary indicators of these two components and estimate their e ect on the growth and the international diff erences in GDP per capita.