Abstract
We focus on three environmental impacts particularly influenced by population age-structure—carbon emissions from transport and residential energy and electricity consumption—as well as aggregate carbon emissions for a panel of developed countries, and take as our starting point the Stochastic Impacts by Regression on Population, Affluence, and Technology (STIRPAT) framework. Among our contributions is to further disaggregate population into three particularly key age groups: 20–34, 35–49, and 50–64, and by doing so demonstrate that population’s environmental impact differs considerably across age groups, with the older age groups (ones typically associated with larger households) actually exerting a negative influence. Furthermore, those age-specific population influences are different (in absolute and relative terms) for the different environmental impacts we analyze. Also, we find that urbanization, in developed countries, best measures access to a country’s power grid, and thus, is positively associated with energy consumption in the residential sector. Finally, we suggest some modeling and methodological improvements to the STIRPAT framework.
Similar content being viewed by others
Notes
Perhaps the level of aggregation encountered in the literature is so popular because it is the level of aggregation in the widely used World Bank data set. The UN does publish (with free access) population data in 5-year age grou**s, but the data is only available at 5-year intervals (beginning in 1950), and that data set requires considerably more manipulation by the analyst to compile in a form suitable for regressions (http://esa.un.org/unpp).
The first published population–environment study to consider households as the unit of analysis we know of was MacKellar et al. (1995). However, household size can be a difficult variable to collect for an empirical panel analysis; thus, few other macro-level, cross-country studies have followed MacKellar et al.’s lead—two exceptions are Cole and Neumayer (2004) and Liddle (2004).
This number is estimated because the last household size category supplied in the data is “seven or more” members, i.e., the number of households with exactly eight, nine, etc., members is not explicitly known from the data.
The working or retired designation is merely to distinguish between two household types that do not include children. The data set used does not otherwise allow for disaggregation’s by employment status.
This point was made by an anonymous reviewer.
Unit root tests are used to determine stationary, and were originally developed for time-series, but have been expanded to cover panels.
Two or more non-stationary variables are said to be cointegrated if some linear combination of them is stationary. The finding of cointegration among economic variables is interpreted as evidence of a long-run, equilibrium relationship. Like for unit roots, cointegration tests were originally designed for time-series but have been expanded to cover TSCS data sets.
This variation is now available as an option on most statistical programs too.
Those countries are: Australia, Austria, Belgium, Canada, Denmark, Finland, France, Greece, Ireland, Italy, Japan, Netherlands, Portugal, Spain, Sweden, United Kingdom, and United States.
Of course, as a secondary energy source, electricity’s ultimate greenhouse gas impact depends on the extent to which fossil fuels are used to generate it.
Initially, we planned to use the population levels of these age groups, but were deterred due to the multi-colinearity problems such variables created. The size (but not the shares) of population cohorts is highly correlated.
This variable is constructed as follows: industrial energy consumption (from the International Energy Agency—IEA) is divided by industrial output. Industrial output is derived by scaling the IEA’s industrial production index, which is indexed to year-2000, by 2000s GDP from industry—itself calculated by multiplying total GDP (from Penn World Table) by industry’s share of value added (from the World Bank). A few missing observations in the IEA’s industrial production index are filled from a similar index produced by the International Monetary Fund.
The non-fossil fuel sources considered are: geothermal, nuclear, hydro, and solar/wind.
The panels used in the carbon dioxide from transport estimations are missing two observations as described in Table 3.
When including a lagged dependent variable, the Durbin–Watson test for serial correlation is no longer accurate. The recommended test is a Lagrange multiplier (LM) test that involves regressing the residuals from an OLS estimation on the first lag of those residuals as well as all the independent variables (including the lagged dependent variable) used in that OLS estimation. One then performs a LM test on the null hypothesis that the coefficient on the lagged residual term is zero—a rejection of that null is evidence of first-order serial correlation.
In addition, variance inflation factors were calculated, and all were found to be below 3. However, it is nearly impossible for regressions comprised of IPAT variables to avoid completely multicollinearity (mutual association among variables) based on the theories developed to explain how those variables interact. For example, affluence (or GDP per capita) is believed to affect population—through human capital’s influence on birth rates (e.g., Becker et al. 1990) 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 (e.g., Bloom and Williamson 1998); and human capital and technology have been recognized as drivers of economic growth (affluence) since Solow (1956). The above theories suggest that the best way to perform STIRPAT regressions may be to treat the variables as part of a cointegrated system; however, such analysis requires annual observations, and is beyond the scope of this article.
In addition, tests on the redundancy of the fixed effects were strongly rejected for our models, as were Hausman tests on the consistency and efficiency of a random effects alternative to fixed effects.
When economic or demographic variables are non-stationary, they are typically I(1), i.e., if differencing is applied once they become stationary. Orders of integration greater than I(2) are very rare among economic/demographic variables.
As pointed out by an anonymous reviewer, national, aggregate carbon emissions are calculated from national, aggregate energy consumption; thus, for countries with carbon intensive energy sources, aggregate carbon emissions and aggregate energy intensity run the risk of being highly correlated with construction. It can be seen from the correlation matrix (Table 4) that our industrial energy intensity variable does not suffer from this problem.
References
Ang, B. W., & Zhang, F. Q. (2000). A survey of index decomposition analysis in energy and environmental studies. Energy, 25, 1149–1176.
Badoe, D. A., & Miller, E. J. (2000). Transportation—land-use interaction: empirical findings in North America, and their implications for modeling. Transportation Research Part D, 5, 235–263.
Baltagi, B. (Ed.). (2000). Advances in econometrics, Vol. 15: Nonstationary panels, panel cointegration, and dynamic panels. Amsterdam: JAI Press.
Beck, N. (2008). Time-series–cross-section methods. In J. Box-Steffensmeier, H. Brady, & D. Collier (Eds.), Oxford handbook of political methodology. New York: Oxford University Press.
Beck, N., & Katz, J. (1995). What to do (and not to do) with time-series cross-section data. American Political Science Review, 89(3), 634–645.
Beck, N., & Katz, J. (1996). Nuisance vs. substance: Specifying and estimating time-series-cross-section models. Political Analysis, 6(1), 1–34.
Becker, G. S., Murphy, K. M., & Tamura, R. (1990). Human capital, fertility, and economic growth. Journal of Political Economy, 98(5), S12–S37.
Bento, A., Cropper, M, Mobarak, A., & Vinha, K. (2003). The impact of urban spatial structure on travel demand in the United States. World Bank Policy Research Working Paper no. 3007.
Bloom, D. E., & Williamson, J. G. (1998). Demographic transitions and economic miracles in emerging Asia. The World Bank Economic Review, 12(3), 419–455.
Cole, M., & Neumayer, E. (2004). Examining the impact of demographic factors on air pollution. Population and Environment, 26(1), 5–21.
Dietz, T., & Rosa, E. (1997). Effects of population and affluence on CO2 emissions. Proceedings of the National Academy of Sciences of the USA, 94, 175–179.
Ehrlich, P., & Holdren, J. (1971). The impact of population growth. Science, 171, 1212–1217.
Fan, Y., Liu, L.-C., Wu, G., & Wei, Y.-M. (2006). Analyzing impact factors of CO2 emissions using the STIRPAT model. Environmental Impact Assessment Review, 26, 377–395.
Heston, A., Summers, R., & Aten, B. (2006). Penn World Table Version 6.2, Center for International Comparisons, University of Pennsylvania.
Kenworthy, J., Laube, F., & Newman, P. (1999). An international sourcebook of automobile dependence in cities, 1960–1990. Boulder: University of Colorado Press.
Lee, C., & Chang, C. (2008). Energy consumption and economic growth in Asian economies: A more comprehensive analysis using panel data. Resource and Energy Economics, 30, 50–65.
Lee, C., Chang, C., & Chen, P. (2008). Energy-income causality in OECD countries revisited: The key role of capital stock. Energy Economics, 30, 2359–2373.
Liddle, B. (2004). Demographic dynamics and per capita environmental impact: Using panel regressions and household decompositions to examine population and transport. Population and Environment, 26(1), 23–39.
MacKellar, F. L., Lutz, W., Prinz, C., & Goujon, A. (1995). Population, households, and CO2 emissions. Population and Development Review, 21(4), 849–865.
Maddala, G. S., & Kim, I.-M. (2000). Unit roots, cointegration, and structural change. Cambridge, UK: Cambridge University Press.
Marland, G., Boden, T. A., & Andres, R. J. (2007). Global, regional, and national CO2 emissions. In G. Marland, T. A. Boden, & R. J. Andres (Eds.), Trends: A compendium of data on global change. Oak Ridge, TN: Carbon Dioxide Information Analysis Center, Oak Ridge National Laboratory, U.S. Department of Energy.
Martinez-Zarzoso, I., Benochea-Morancho, A., & Morales-Lage, R. (2007). The impact of population on CO2 emissions: evidence from European countries. Environmental and Resource Economics, 38, 497–512.
Narayan, P., & Smyth, R. (2008). Energy consumption and real GDP in G7 countries: New evidence from panel cointegration with structural breaks. Energy Economics, 30, 2331–2341.
O’Neill, B. C., & Chen, B. S. (2002). Demographic determinants of household energy use in the United States. Population and Development Review, 28, 53–88.
Pedroni, P. (1999). Critical values for cointegration tests in heterogeneous panels with multiple regressors. Oxford Bulletin of Economics and Statistics, 61, 653–670.
Prskawetz, A., Leiwen, J., & O’Neill, B. (2004). Demographic composition and projections of car use in Austria. Vienna Yearbook of Population Research, 2004, 247–326.
Rosa, E., York, R., & Dietz, T. (2004). Tracking the anthropogenic drivers of ecological impacts. Ambio, 33(8), 509–512.
Roy, N. (2002). Is adoptive estimation useful for panel models with heteroskedasticity in the individual specific error component? Some Monte Carlo evidence. Econometric Reviews, 21, 189–203.
Schipper, L., Unander, F., Murtishaw, S., & Ting, M. (2001). Indicators of energy use and carbon emissions: Explaining the energy economy link. Annual Review of Energy and Environment, 26, 49–81.
Shi, A. (2003). The impact of population pressure on global carbon dioxide emissions, 1975–1996: Evidence from pooled cross-country data. Ecological Economics, 44, 29–42.
Solow, R. (1956). A contribution to the theory of economic growth. Quarterly Journal of Economics, 70, 65–94.
White, H. (1980). A heteroskedasticity-consistent covariance matrix and a direct test for heteroskedasticity. Econometrica, 48, 817–838.
York, R. (2007). Demographic trends and energy consumption in European Union Nations, 1960–2025. Social Science Research, 36, 855–872.
York, R., Rosa, E., & Dietz, T. (2003a). STIRPAT, IPAT, and ImPACT: analytic tools for unpacking the driving forces of environmental impacts. Ecological Economics, 46, 351–365.
York, R., Rosa, E., & Dietz, T. (2003b). A rift in modernity? Assessing the anthropogenic sources of global climate change with the STIRPAT model. International Journal of Sociology and Social Policy, 23(10), 31–51.
Acknowledgement
Comments from anonymous reviewers and the editors of this special issue helped to improve the final version.
Author information
Authors and Affiliations
Corresponding author
Electronic supplementary material
Below is the link to the electronic supplementary material.
Rights and permissions
About this article
Cite this article
Liddle, B., Lung, S. Age-structure, urbanization, and climate change in developed countries: revisiting STIRPAT for disaggregated population and consumption-related environmental impacts. Popul Environ 31, 317–343 (2010). https://doi.org/10.1007/s11111-010-0101-5
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11111-010-0101-5