4. Empirical results and discussion
Table 3 presents a summary of the estimation results: the full set of our sample countries, the group of “big” countries and the group of “small and medium” ones. The econometric estimation followed was the panel random effects (RE) for the full sample, the “small and medium” countries and the “big” countries group. We report two modes for each set (the full sample, the group of “small and medium” counties, and the group of “big”); the Mode 1 without the interaction term and the Mode 2 with the interaction term. We note that we have also lagged our right hand side variables LHOST and LEXP (and as a result their product, i.e. the INTER variable, as well) one period to control for the possible endogeneity of current values of the explanatory variables (see also Bloom and Van Reenen, 2002). As a result, this change does not noticeably affect the previous results reported in the main body of the paper.