On the contrary, a strict environmental policy leads to a reduction in CO2 emissions in traded goods, both on the part of the exporter and the importer. To quote this article: ASSOGBAVI Koutchogna Kokou Edem and DEES Stéphane (2021), Environmental policies and CO2 emissions embodied in international trade, Bordeaux Economics Working Papers, BxWP2021-07. Emissions resulting from trade (about 8 billion tons in 2015) are very large and represent a quarter of total global emissions (about 32 billion tons). Developed countries are generally net importers of CO2 emissions, while emerging or commodity-producing countries are instead net exporters (Cezar and Polge, 2020).
First, we use data on CO2 emissions embodied in trade flows produced by the OECD (Yamano and Guilhoto, 2020) in a large set of countries over a more recent period. On the contrary, a tightening of environmental policy leads to a reduction in CO2 emissions embodied in trade. Therefore, the apparent disparities in CO2 emissions embodied in trade do not stem from changes in environmental policy, but from common motives of specialization.
Using this theoretical model of trade flows, let us now see how to model emissions in trade. This explains the double counting problems associated with emissions contained in intermediate trade flows, e.g.
Gravity model variables
We calculate the correlations between the CO2 contained in the exports from country i to country j and the de jure and de facto indicators in country i, in country j and with the difference between the two. We also distinguish among the pairs between exports from emerging economies and those originating from advanced economies. These pairwise cross-sectional correlations first show a negative relationship between the export of CO2 goods and the environmental policy of the exporting country.
In other words, stricter environmental policy is associated with greater exports of CO2-intensive goods, illustrating the potential 'carbon leakage' of environmental policy through international trade. These correlations are even stronger if we focus only on exports from emerging economies to developed economies. Finally, the stricter the policy of the importer compared to the exporter (meaning lower EP Sijt and CCHijt), the higher the correlation with the CO2 included in trade, again illustrating the possible phenomenon of 'carbon leakage'.
Similarity in the production structure
Empirical approach
Benchmark results
Overall, we consider 4 different models from the simplest, where we only look at GDP per per capita, trade costs and environmental policy indicators to the most comprehensive that includes all variables. Before looking at the environmental policy variables, a first point worth mentioning is related to the signs and meaning of GDP per capita. inhabitant. Contrary to gravity equations used to describe trade flows, our specification shows that the CO2 content of exports is inversely proportional to the level of development of the exporter.
The elasticity of CO2 embodied in exports to environmental policy variables is therefore key to proving this question. The stricter a country is on environmental policy (for both exporters and importers), the less CO2-intensive trade flows are. Adding energy-related variables shows that CO2 embodied in trade is also positively dependent on the exporter's energy consumption and its electricity production.
Looking at the results in more detail, we note that the magnitude of the coefficients of the policy variables is larger in absolute terms for the exporter policy indicator. In Table 3, we replace our policy variables with a single one, taking the difference in the degree of stringency between the exporting and the importing country, giving an indication of the relative level of stringency in environmental policy between each country. Such an index of policy dissimilarity appears with a negative sign, meaning that a country whose environmental policy is stricter than that of its partner will export goods with lower CO2 content.
Considering these results from the importer's side, it means that environmental policy measures have a relatively smaller impact on the CO2 intensity of imported goods. Taken together, our results show that although we find no evidence of carbon sequestration (the coefficient of EP Sij is negative in Table 3), the good countries from an environmental policy perspective import the most from the least good. From a policy perspective, since differences in environmental regulation between countries tend to support trade in carbon-intensive goods, this result points to the importance of international and interstate coordination in environmental policy to limit trade-induced CO2 emissions.
To go deeper into the analysis of the role of de jure policy in the CO2 content of trade, we divide the policy indicators into market and non-market instruments.
Results using de facto environmental indicators
The results found with the aggregate index are verified when it is broken down into market components and non-market components. Interestingly, the sensitivity of CO2 contained in exports tends to be lower for market-related measures (-0.05 for exporters and -0.03 for importers) compared to non-market ones (around -0.06). Non-market measures (regulatory measures or support for low-emission technology) therefore contribute more to avoiding trade-related carbon leakage compared to market measures (i.e. carbon tax).
Moreover, when using bilateral measures, although the coefficient on EP SijtN M is significantly negative (Model 5), the coefficient on EP SijtN M is not significant, meaning that the difference between exporters and importers disappears for non-market measures. Comparing the size of the coefficient is not easy, as the EPS is an indicator that ranges from 0 to 6, while CCH is measured as a percentage. However, if we multiply the CCH-related coefficients by a factor of 100/6, we find values that are similar (i.e. around −0.2 for the country i indicator and −0.1 for the country j indicator).
These results therefore confirm not only the sign but also the magnitude of the sensitivity of CO2 embodied in trade to environmental indicators. In the case of CCH, since the indicator covers more countries than VPA, this result is also satisfactory since doubling the size of the sample does not change the results. Since it includes many more emerging and commodity exporting countries, this sample also seems even more relevant to determine the presence of carbon leakage behavior.
We also consider another set of estimates that replaces environmental indicators with bilateral ones calculated as for EPS as the difference between a country's index and its partner. In terms of de jure indicators, this result suggests that large heterogeneity in environmental performance is detrimental to trade-induced reductions in CO2 emissions.
Robustness exercises
At the same time, this does not dramatically change the coefficient values of the environmental variables, making our main results robust to the choice of trade cost indicators. The signs are correct, but the impact of domestic environmental policy on CO2 emissions in exports is not significant. 3.1, we account for the role of structural differences by including an interaction variable between our environmental policy indicators for the importing countries and this bilateral similarity index.
At maximum (theoretical) level of inequality (when Sij = 2), the coefficient associated with the interaction term should be half (in absolute value and opposite sign) of the importer's policy indicator coefficient to point to any evidence of carbon leakage via trade with the most disparate countries. As seen above, the maximum value of Sij is around 1.20 and most of the distribution varies between 0.2 and 0.5, so the interaction term coefficient must be at least twice as large (in absolute value) as the corresponding coefficient to cancel the negative impact of policy stringency on CO2 emissions embodied in trade. In the extreme case where a country only has to trade with countries that are structurally different (i.e. the few cases at the maximum value of our inequality index distribution), this result would only indicate a cancellation of the impact of environmental performance on reducing the import of CO2-intensive goods.
These results show that limiting our sample to such country pairs yields no evidence of the effects of environmental policy on reducing CO2 emissions embodied in trade. However, the value of the CCHi coefficient is much lower now (from 0.09 to 0.06 in Model 1), which means that the impact of environmental performance on exports of CO2-intensive goods is smaller in the case of developing economies. When a country imports from partners that are less similar, the effects of environmental performance on the CO2 intensity of goods are lower.
On the contrary: the higher the environmental performance of a country, the lower the CO2 emissions that trade entails. Our empirical evidence also shows that CO2 emissions associated with trade are more likely to be explained by the usual determinants of international trade, such as shipping costs or revenues from both exporting and importing trading partners. From a policy perspective, since differences in environmental regulations between countries tend to support trade in carbon-intensive goods, this result points to the importance of international coordination and harmonization between countries in environmental policies to curb CO2 emissions reflected in trade. .
Kyoto and Carbon Leakage: An Empirical Analysis of the Carbon Content of Bilateral Trade. The proper panel econometric specification of the gravity equation: A three-way model with bilateral interaction effects. Contribution of Working Group III to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change;.
CO2 emissions incorporated into international trade and domestic final demand using the OECD inter-country input-output database. World Development Indicators (WDI), World Bank Energy use World Development Indicators (WDI), World Bank Climate Change Yale Center for Environmental Law & Policy CO2 emissions incorporated in in-.
LAREFI
