Moment-based tests for random effects in the two-way error component model with unbalanced panels

This paper extends Wu and Li (2014)'s moment-based tests for random effects to the case with unbalanced panel data. Based on the difference of variance estimators of the idiosyncratic errors at different robust levels, two statistics are constructed to test for the existence of individual and time effects, respectively. Some variants of the two statistics and joint tests for both the two effects are also discussed. Their asymptotic properties are obtained under some mild conditions. It is shown that the tests retain the desired properties observed in the balanced panel data case. Monte Carlo simulation experiments and a real data analysis are carried out for illustration.