手法を比較

	IPAT Decomposition	STIRPAT Model
分野	Environmental Sociology	Environmental Sociology
系統	Process / pipeline	Regression model
提唱年	1971	1997
提唱者	Paul R. Ehrlich & John P. Holdren (IPAT); Yoichi Kaya (Kaya identity)	Thomas Dietz & Eugene A. Rosa; Richard York
種類	Multiplicative accounting identity and decomposition of environmental impact	Log-linear stochastic regression model of environmental impact drivers
原典	Ehrlich, P. R., & Holdren, J. P. (1971). Impact of Population Growth. Science, 171(3977), 1212-1217.	Dietz, T., & Rosa, E. A. (1997). Effects of population and affluence on CO2 emissions. Proceedings of the National Academy of Sciences, 94(1), 175-179.
別名	IPAT Identity, Ehrlich-Holdren Identity, Kaya Identity Decomposition, Impact Equation	Stochastic IPAT, STIRPAT Regression, Stochastic Impacts by Regression on Population Affluence and Technology, Dietz-Rosa Impact Model
関連	4	4
概要	IPAT decomposition expresses environmental impact as the product of three factors, population, affluence, and technology, providing a simple accounting framework for attributing degradation to its proximate human drivers. The identity was crystallized in the debate between Paul Ehrlich, John Holdren, and Barry Commoner around 1971, with Ehrlich and Holdren's Science article on the impact of population growth a foundational statement. In the equation, affluence is output per person and technology is impact per unit of output, so the three factors multiply back exactly to total impact, making IPAT a definitional identity rather than an empirical claim. Its best-known specialization, the Kaya identity, decomposes carbon emissions into population, GDP per capita, energy intensity of output, and carbon intensity of energy, and underpins much emissions-scenario work. By taking growth rates, IPAT also yields a clean additive decomposition that apportions the change in impact among its drivers. Because the identity assumes each factor contributes proportionally, it was the stimulus for the stochastic STIRPAT model, in which Dietz and Rosa relaxed that assumption to test the drivers statistically.	The STIRPAT model, short for Stochastic Impacts by Regression on Population, Affluence, and Technology, is a statistical reformulation of the IPAT identity that allows the drivers of environmental impact to be estimated and tested rather than merely asserted. Thomas Dietz and Eugene Rosa introduced it in 1997 to study national carbon dioxide emissions, recasting the deterministic accounting identity impact equals population times affluence times technology as a multiplicative stochastic model with an error term. Taking logarithms turns this into a linear regression whose coefficients are elasticities, the percentage change in impact associated with a one-percent change in each driver. This lets researchers ask whether impact rises strictly in proportion to population, as the original identity assumes, or whether there are increasing or decreasing returns to scale. Richard York, Rosa, and Dietz formalized and extended the approach in 2003, showing how additional drivers, quadratic terms, and panel structure can be incorporated within the same framework. STIRPAT has become the dominant quantitative tool in environmental sociology for analyzing the anthropogenic forces behind emissions, energy use, and ecological footprints.