Is Trade Good or Bad for the Environment?

Summary of working paper 9021
Featured in print Digest

Openness, measured as the ratio of trade to income, appears to reduce air pollution.

Opponents of globalization claim that international trade harms the environment. They believe that, becausethat in open economies a "race to the bottom" in environmental standards will result from governments' fears that enhanced environmental regulation will hurt their international competitiveness, the result is a "race to the bottom" in environmental standards. In Is Trade Good or Bad for the Environment? Sorting out the Causality (NBER Working Paper No. 9021), NBER Research Associates Jeffrey Frankel and Andrew Rose examine the environmental effects of openness to trade in a statistical cross-section of countries in 1995. They find that the impact of trade on at least three kinds of air pollution appears to be, if anything, beneficial, not adverse, for a given level of income. Openness, measured as the ratio of trade to income, appears to reduce air pollution. The level of statistical significance is high for Sulfur Dioxide (SO2), and moderate for Particulate Matter and Nitrogen Oxides (NOx).

Correlation need not prove causation. The observed correlation between trade and pollution could arise in other ways. It is possible that countries that are more democratic tend to be both more open to trade and more responsive to environmental concerns. Also, higher levels of income can interact with trade and the environment in all sorts of ways. This paper tries to disentangle the causality between trade and the environment by first testing for the effect of openness on the environment while controlling for income. Then the authors focus on exogenous variation in trade attributable to geography (for example distance from major trading partners), and on variation in income per capita attributable to standard growth determinants (for example population, investment, and education).

How could trade be good for the environment? Trade allows countries to attain more of what they want, including environmental protection (the authors call this proposition the gains-from-trade hypothesis). Trade might lead to international pressures to increase environmental standards, or to beneficial technological and managerial innovations. Multinational corporations tend to bring clean state-of-the-art production techniques from higher-standard countries of origin to host countries where such standards are not yet known. Furthermore, trade economists believe that openness to trade encourages continual innovation both in technology and in management practice; such innovation likely will be applied to environmental concerns as well as to pure economic goals. In other words, Frankel and Rose suggest, environmental improvement may well accompany globalization.

Even if openness to trade does not raise air pollution worldwide, it may give rise to "pollution havens": that is, some countries that specialize in dirtier production and export their products to others who specialize in cleaner production. In this way, the geographical distribution of pollution might change, even if the average level did not. In one version of the pollution haven hypothesis, poorer countries are predicted to have a "comparative advantage" in pollution. But Frankel and Rose test the proposition that the combination of being poor and open makes for higher levels of pollution, and they find no evidence of it at all. Similarly they are able to reject the versions of the pollution haven hypothesis that say that low-density countries or capital-intensive countries have a comparative advantage in pollution.

The authors also document for thethree measures of air pollution the "Environmental Kuznets Curve" for the three measures of air pollution. This widely tested relationship says that growth harms the environment at low levels of income, but helps at high levels. At higher levels of income per capita, growth stimulates the public's demand for improving environmental quality, which in democratic societies is brought about through environmental regulation. Frankel and Rose estimate that SO2 pollution, for example, peaks at income levels of about $5,770 per capita, and thereafter starts to decline. All of this squares with economic theories that suggest that growth yields air and water pollution when industrialization is being introduced, but eventually results in reduction of reduced pollution as countries become prosperous enough to afford cleaning up their environments. In other words, production technology inevitably pollutes, but the rising income that results from this same production technology just as inevitably increases the demand for environmental quality.

A final finding is also familiar from studies of trade and income: globalization is good for growth. The authors find that every .01 increase in the ratio of trade to GDP raises income by 0.4 percent over the following 20 years. The effects of trade that operate via growth -- worsening pollution at first, and then reducing pollution later -- may be larger than the effects of trade that operate independently of growth.

In sum, Frankel and Rose find that after an initial adverse effect in the relationship between growth and environmental damage at low levels of income, a pattern emerges showing that growth eventually has a beneficial effect on air pollution. Still, the researchers caution that the results are less consistently positive in regard to broader measures of environmental quality. Some environmental problems such as emission of greenhouse gases are truly global and not local, they point out. A "free rider" problem prevents national governments from translating the demand for environmental improvement into reality, even when they collectively have the economic means to do so. Thus the authors are not surprised to find statistically that trade and growth do not seem to have beneficial effects on emissions of these gases. In such cases, say Frankel and Rose, international cooperation and not just local regulation is needed.