At the global level, the relationship between population and environ- mental impact is not easy to test in a way that leads to reliable and non- spurious estimates. This is because of the lack of data on environmental degradation covering a sufficiently large number of countries over a suffi- ciently large period of time. It is for this reason that studies have tended to focus on CO2 emissions and energy use, for which cross-country and time- series data are available. The time dimension is necessary in order to avoid the problem of one-period cross-sectional regressions, which are likely to lead to spurious results if population size or growth.
