Note that the population share of the economically active age groups now becomes more clearly insignificant. This suggests that its initial statistical significance might be entirely due to its correlation with the urbanization rate as pointed out above and its correlation with the average household size (partial correlation coefficient of )). In other words, it would appear that the urbanization rate and average household size are the demographic factors that really matter. Interestingly, the coefficient size of the energy intensity variable rises once average household size is controlled for. Table 2 reports results for a similar set of estimations, but with SO2 emissions as the dependent variable. In column I.
