Matthew A. Cole

Department of Economics

University of Birmingham

A large number of studies have claimed to find evidence of an inverted U-shaped relationship between pollution and per capita income, generally referred to as an environmental Kuznets curve (EKC) (see for example Shafik 1994, Grossman and Krueger 1995, Cole et al. 1997). This relationship is potentially of great significance since it implies that countries can ‘grow out’ of environmental problems.

However, several recent studies have criticised the EKC relationship and the methodology used to estimate it on a number of grounds (see for example Arrow et al. 1995, Ekins 1997, Stern 1998, Stern and Common 2001). In the light of these criticisms, this paper assesses the strength of the Environmental Kuznets Curve (EKC). Specifically, answers are sought to the following related questions: (1) How robust is the EKC relationship?; (2) To what extent can the EKC relationship be explained by changing trade patterns as opposed to growth-induced pollution abatement?

With regard to question (1), the alleged weaknesses with the EKC are assessed and the sensitivity of EKC results is examined, with particular attention paid to the choice of functional form, to the inclusion of additional covariates and to the income range of the countries in the sample. Issues of stationarity and simultaneity between income and emissions are also considered. Turning to question (2), it is argued that previous studies that have incorporated trade into the EKC have done so in an inappropriate manner by assuming a unique relationship between trade and emissions. Instead, following Antweiler et al. (2001), it is here argued that the impact of trade liberalisation on the environment will differ from country to country depending on whether or not they have a comparative advantage in pollution intensive production. In turn, it is argued that this depends on a country's relative factor endowments and/or its relative environmental regulations. EKCs are therefore estimated in a manner that allows the impact of trade liberalisation on pollution to depend on these country characteristics.

The results indicate that, firstly, the inverted U relationship between per capita income and emissions is reasonably robust. Estimated turning points for each pollutant, for instance, show little variation across the variety of estimations that are undertaken. Secondly, the impact of trade on emissions appears to be small, particularly relative to the impact of income. There is little evidence to suggest that changing trade patterns are responsible for the shape of the EKC.

JEL Classification: O10, O13, F18, Q25