نابرابری درآمد و محیط زیست: جانبداری کل در منحنی کوزنتس محیطی

The environmental Kuznets curve assumes an inverted U-shaped relation between environmental damage and per capita income. Recently it has been argued in the literature that in addition to income levels, the inequality in the distribution of power and income is (positively) related to environmental degradation. We provide an additional argument, based on simple aggregation, for including a measure of income dispersion in empirical analyses. When the relationship between environmental damage and household income is concave (e.g. resembles an environmental Kuznets curve), then income inequality is negatively related to total environmental damage. Results from an empirical analysis of cross-national variation indicate that the aggregation effect can run counter to and outweigh the political economy effect for some environmental indicators.

According to the environmental Kuznets curve (EKC) hypothesis, environmental damage first increases, but after a ‘turning point’ declines with per capita income. One reason for this inverted U-shaped relation is that the so-called ‘scale effect’, capturing the simple intuition that more output, ceteris paribus, results in more adverse effects for the environment, is (partly) offset by the ‘composition effect’ (due to changes in the underlying structure of the economy) and the ‘technique effect’ (referring to possible changes in the methods of production). The extent to which these factors operate depend on the incentives faced by economic actors and policy makers. As incomes increase, both the demand for improvements in environmental quality and the resources available for environmentally friendly investments will rise. Another, more simple, reason is put forward by Andreoni and Levinson (1998) who propose an explanation derived directly from the technological link between consumption and abatement of the undesirable byproduct (pollution). For empirical support for some damage indicators, see e.g. Shafik and Bandyopadhyay, 1992, Cropper and Griffiths, 1994, Selden and Song, 1994, Grossman and Krueger, 1995 and Panayotou, 1995 and studies in the special issue of Ecological Economics (vol. 25, 1998) on this topic. For critical discussions of progress on the EKC concept, see Stern et al. (1996) and Stern (1998). One approach to progressing research in this field is to add distribution issues. Boyce (1994) hypothesises that the inequality of power and income may be an important determinant of environmental degradation, which is examined empirically by Torras and Boyce (1998). More specifically, it is argued that greater equality of incomes results in lower levels of environmental degradation. This conclusion is challenged by Scruggs (1998) on both theoretical and empirical grounds. Boyce (1994) and Torras and Boyce (1998) take a public choice approach to explain the importance of income distribution. We call this the political economy argument. For example, redistributing income will affect society's demand for environmental quality (Magnani, 2000), and thus affect the ‘induced policy response’ (Grossman and Krueger, 1995). Also, an equitable distribution may contribute to social harmony that is conducive for the long-term perspective that investments in environmental quality require. In addition to this political economy argument, a simple aggregation argument can be used to motivate the importance of income inequality in explaining environmental degradation. The aggregation argument is based on two observations: (i) for some environmental problems, a non-linear (e.g. an inverted U-shaped) relation between income and degradation can be found at the micro (household) level; and (ii) aggregating over households with unequal incomes subject to such a non-linear relation will result in biased estimates of both the level and turning point of the aggregate relation (EKC?) when a measure of income dispersion is not included. The paper therefore makes two contributions. First, it adduces an additional reason for including measures of income inequality in EKC-type analyses of the relation between income and national level environmental degradation. Second, it demonstrates the importance of examining household level relations between income and environmental pressure, which can affect the sign and magnitude of the overall inequality effect.

In this paper we have demonstrated the importance of income distribution as an explanatory variable in environmental degradation caused by households. Due to non-linearities in the relation between environmental pressure and income at the micro level, estimates of aggregates should include a measure of income dispersion to avoid biased estimates of EKC levels and turning points. Boyce (1994) also proposes to add a measure of income inequality to regression models, but his argument is based on considerations pertaining to the political economy rather than simple aggregation algebra. Interestingly, when the relation between household income and environmental pressure describes a micro EKC, the predictions of Boyce run counter to our predictions. Specifically, while Boyce predicts that redistributing income for rich to poor households enhances environmental quality, we predict that the reverse should hold. The empirical results indicate that the aggregation argument may dominate the political economy argument, as is the case for nutrient balances and CO2-emissions. For particulate matter and SO2 concentrations, on the other hand, the net effect does not differ significantly from zero. For deforestation, the political economy effect appears to dominate. When the degradation-income relationship is convex, as is the case with lack of access to sanitation and safe water, the two effects operate in the same direction. The theoretical and empirical results thus indicate a potential trade-off between environmental conservation and income equality, both of which can be considered legitimate government objectives. Redistributing income is desirable in its own right, but may, at least in the short and medium term, contribute to loss of environmental quality if the household level relation between income and environmental pressure is concave and the aggregation effect is sufficiently large.