Consumption-driven environmental impact and age-structure change in OECD countries: a co-integration STIRPAT analysis

Brant Liddle, Victoria University, Melbourne

This paper examines two types of environmental impact for which population has a substantial demonstrated impact or influence, i.e., carbon emissions from transport and residential electricity consumption. It takes as its starting point the stochastic version of the IPAT model or the STIRPAT framework. The paper advances that framework in two important ways: (1) it disaggregates population into four key age groups: 20-34, 35-49, 50-69, and 70 and older; and (2) it employs advanced time-series-based techniques like panel unit root tests, panel cointegration, and panel Fully Modified OLS (FMOLS). Population has been shown to impact the environment in considerably different ways across age-groups, i.e., young adults (20-34) typically behave in more environmental intense ways than older ones (35-59). Also, the variables typically analyzed in STIRPAT studies are stock (population) or stock-related variables (GDP, emissions, and energy consumption), and thus, are likely nonstationary—i.e., their means change over time. Such data sets should be tested for panel-unit roots and panel-cointegration. Furthermore, the IPAT variables (e.g., population, affluence) are highly interrelated, for example, affluence is believed to affect population—through human capital’s influence on birth rates and higher income’s ability to lower death rates; meanwhile, population has been shown to impact affluence—when the size of the working-age population increases faster than the size of the dependent-age population. Our use of FMOLS to estimate the long-run elasticities corrects for this endogeneity as well as serial correlation and nonstationarity. Initial results indicate that population has a greater impact than affluence on carbon emissions from transit and residential electricity consumption, that young adults exert a grater influence on the environment than older adults, and that that influence is more pronounced in the transport than the residential sector.