of the findings associated with the coefficients presented in this table; suffice it to say that the N-shaped functional
form continues to fit the data well even after the inclusion of these covariates in the equation.
Taken together, the findings presented here suggest that a cubic, N-shaped functional form better
describes the relationship between GDP and CO 2 emissions than the quadratic, downward-U functional form that
researchers regularly celebrate in the literature. If so, there is no reason to be sanguine about the possibility that
aggregate carbon dioxide emissions will be diminishing anytime soon. Researchers who support the idea of an
EKC for carbon dioxide sometimes suggest that it may be possible to “grow out” of problems associated with
global climate change. The findings presented here not only indicate that most of the world’s population resides
at points along the curve where increases in per capita income are positively associated with increases in CO 2
emissions, they also call into question the claim that per capita carbon dioxide emissions begin to diminish once
a certain income threshold has been met. Skepticism regarding the claim that economic growth will eventually
bring about a reduction in CO 2 emissions in turn implies that scholars and members of the public alike need to
focus much more attention on policy instruments if reductions in aggregate carbon dioxide emissions are to be
achieved. Since the early 1990s, governments have implemented a number of policies aimed at mitigating climate
change, such as carbon taxes and emission trading schemes. Because researchers have largely failed to incorporate
into their analyses direct measures of environmental policy, however, we know very little about the effectiveness
of various efforts on the part of governments to reduce carbon dioxide emissions. It is toward policy-oriented
research of this sort that scholarship needs to turn.
Table 1. Estimates from a fixed-effects model, 1980-2004 (lag term included, standard errors in
parentheses). Dependent variable: per capita CO 2 emissions in metric tons.
Model 3Per capita GDP (thousands USD)
.1283***(.0278)Per capita GDP^2
-.0055***(.0012)Per capita GDP^3
.0001***(.0000)Lagged per capita CO 2
.6470***(.0705)Constant
-.3916***(.0981)Number of observations
2223- p < .05 ** p< .01 *** p < .001