Munich Personal RePEc Archive

Population, Affluence, and Environmental Impact Across Development: Evidence from Panel Cointegration Modeling

Liddle, Brantley (2013): Population, Affluence, and Environmental Impact Across Development: Evidence from Panel Cointegration Modeling. Published in: Environmental Modelling & Software , Vol. 40, (2013): pp. 255-266.


Download (330kB) | Preview


This paper analyzes urban population’s and affluence’s (GDP per capita’s) influence on environmental impact in developed and developing countries by taking as its starting point the STIRPAT framework. In addition to considering environmental impacts particularly influenced by population and affluence (carbon emissions from transport and residential electricity consumption), the paper determines whether and, if so, how those environmental impact relationships vary across development levels by analyzing panels consisting of poor, middle, and rich countries. The development-based panels approach is an improvement on the GDP per capita polynomial model used in the Environmental Kuznets Curve and other literatures for several reasons: (i) it allows one to determine whether the elasticity of all variables considered varies according to development; (ii) it is arguably a more accurate description of the development process; (iii) it avoids potentially spurious regressions involving nonlinear transformations of nonstationary variables (GDP per capita squared); and (iv) unlike the polynomial model, it allows for the possibility that elasticities are significantly different across development levels but still positive—precisely the relationship expected for the environmental impacts considered here. Whether or not the elasticity for affluence was greater than that for population was a function of both the choice of dependent variable and the makeup of the panel (all countries, poor, middle, or rich). Furthermore, the estimated elasticities varied, in a nonlinear fashion, according to the development process: U-shaped, inverted U-shaped, and monotonic patterns were revealed, again, depending on the dependent variable.

MPRA is a RePEc service hosted by
the Munich University Library in Germany.