قوانین و مقررات تجارت آلوده و پنج مؤلفه آن

Based on two extensions, this paper proposes a re-appraisal of the concept of the pollution terms of trade (PTT) introduced by Antweiler (1996). First, detailed data allows capturing the effect of differences in emission intensities across countries and over time. Second, relying on Johnson and Noguera (2012), the revised PTT index controls for trade in intermediate goods and is based on value-added rather than gross output figures. Applied to a database for SO2 emission intensities for 62 developed and developing countries over the 1990–2000 period, it turns out that the first extension has a larger empirical importance than the second one. The global pattern is one in which the major rich economies exhibit a PTT index below one (higher pollution intensity in imports than in exports). Trade imbalances tend to exacerbate this asymmetry, allowing rich economies to further offshore their pollution through trade.

By disconnecting production from consumption sources, international trade leads to a worldwide distribution of polluting emissions which does not reflect final demand. A common suspicion is that rich countries, with higher environmental standards, tend to offshore their pollution to poor countries, according to a “pollution-haven” effect (e.g. Levinson and Taylor, 2008). These concerns, along with a growing pressure to curb down global emissions, have led to a flurry of studies analyzing the emission-content of trade (e.g. Wiedmann, 2009). However, it is fair to say that most of these studies have been national or regional in scope, limited to one year, and that available evidence at the world-wide level is still scant. This lack of world-wide evidence is linked with important data requirements and limitations. First, input–output matrices are needed in order to capture the additional emissions generated by the derived demand for inputs (e.g. Levinson, 2010). Second, data on imported input requirements by trade partner are necessary to attribute intermediate imports to their final destination (Johnson and Noguera, 2012). Third, reliable trade and production data must be made available and compatible at a reasonable degree of disaggregation to identify the influence of the most polluting sectors. Fourth, country (and year)-specific emission coefficients are necessary to control for the fact that the emission content of a given amount of output varies across countries and over time, because of differences in both technologies and input–output relationships. Taking the best out of available data for a specific pollutant, which is sulfur dioxide (SO2), the first objective of this paper is to provide evidence of the pollution content of trade at the world-wide level, illustrating in particular the importance of capturing differences in (total) emission coefficients between countries and years and taking properly into account trade in intermediate goods. SO2 is a particularly useful pollutant to investigate because it is primarily anthropogenic, primarily industry-driven (rather than generated by transportation or household activity), and primarily a local (rather than a trans-boundary or global) pollutant.1 In previous work Grether et al. (2009) have decomposed world-wide SO2 manufacturing emissions over 1990–2000 into the well-known scale, technique and composition effects. It has been shown that despite the considerable increase of manufacturing activity by 10% (scale effect) total emissions have fallen by roughly 10% thanks to the adoption of cleaner techniques (technique effect) and a small shift towards cleaner industries. When using output as the scale measure, there is however an important shift towards “dirty” countries. These results confirm the importance of analyzing in more detail the SO2 pollution content of trade. The second objective of this paper is to propose a reconsideration of the concept of the pollution terms of trade (PTT) introduced more than fifteen years ago by Antweiler (1996). Most of the recent literature has produced results in terms of environmental trade balances, normally captured by the difference between import-embodied and export-embodied emissions, or balance of emissions embodied in trade (BEET) according to Muradian et al. (2002). As noted by Straumann (2003), this measure is sensitive to trade imbalances, which may disappear or be reversed over time. By simply taking the ratio between the average pollution content per dollar of exports and the average pollution content per dollar of imports, the PTT index abstracts from this source of bias and appears more appropriate as a long run structural indicator. The original application reported by Antweiler (1996), based on a large range of pollutants (CO2, SO2, NO2, lead, particulate matter, volatile organic compounds), came to the rather paradoxical conclusion that rich countries tended to exhibit a larger PTT index than poor ones. As already discussed by Antweiler himself, a possible reason for this result could come from data limitation. Indeed, he had to rely on US input–output adjusted emission intensities, and apply them universally, as if there were no technological differences across countries. Relying on his own words the original PTT was only capturing the “trade-composition” part of the PTT variation, not the “technological” part. Since then, although the calculation of input–output based embodied emissions has been burgeoning, there has been, to our knowledge, no systematic attempt to reconsider the issue of PTT estimates at the world-wide level. This is all the more regrettable that recent contributions point towards the importance of trade in intermediate goods in shaping the factor content or the value-added content of trade (e.g. Johnson and Noguera, 2012 and Trefler and Zhu, 2010). In the presence of intermediate trade, the relationship between demand and polluting emissions becomes more complex. This is so because imports may correspond to inputs which are used to produce other goods which are further exported to another destination country. This affects the measurement of import or export-embodied emissions (and thus also PTT calculations) in a non-trivial way. The new methodology developed recently makes it possible to control for these effects, by uncovering the implicit trade flows that relate the original producer (and the corresponding emissions) to the final consumer (in the destination country that does manage to offshore pollution). This paper proposes to revisit PTT calculations by exploiting newly available data and recent methodological advances in the analysis of intermediate trade. The approach is directly borrowed from Antweiler (1996), the basic difference being that we include time and country variations into the analysis, as well as trade in intermediate goods, which allows for an original decomposition of the PTT index into five components: a between-sector, a between-country, a technique, an intermediate trade and a value added effect. The first effect corresponds to the “trade-composition” index measured by Antweiler (1996), the second (third) effect captures the influence of different emission intensities across countries (over time), the fourth effect reflects the impact of using implicit (i.e. final demand driven) rather than reported trade flows, and the fifth the impact of considering value added rather than output trade flows. Regarding empirics, the sample period is 1990–2000 with a good coverage (63 developed and developing economies), and a particular care given to capturing technological heterogeneity. Trade‐ and country-specific input–output tables are taken from the Trade Production and Protection database of the World Bank (Nicita and Olarreaga, 2007), while country‐ and time-specific polluting manufacturing emission intensities come from the recent database elaborated by Grether et al. (2009). We also impose the consistency between trade and input–output data, control for reexports and apply the proportionality assumption to spread intermediate input requirements across trade partners (see Johnson and Noguera, 2012). As it turns out, the new empirical evidence reverses the paradoxical pattern observed by Antweiler (1996), and confirms the importance of including the newly computed correction terms. The next section reviews the empirical evidence regarding pollution content of trade calculations. Section 3 outlines the theoretical derivation of the PTT index when trade in intermediate goods is taken into account and makes the link between the environmental trade balance, the trade ratio, and the PTT index. Section 4 shortly describes the data, Section 5 reports and discusses the main results and the last section concludes.

In an increasingly integrated but highly heterogeneous world, international trade leads to a spatial distribution of pollution across countries that might benefit some countries and hurt others in a non-trivial way. Starting with trade imbalances, even if embodied emissions per dollar were identical for all goods, a country running a trade deficit would be able to shift emissions abroad as those emissions contained in its imports are larger than those included in its exports. But trade imbalances are normally temporary, and even if they persist, their effect on net imported emissions from abroad may be secondary when compared with the other main source of unbalanced trade in emission, which is the relative pollution content of exports, better known as the pollution terms of trade (PTT) of a country. The more a country specializes in clean activities, the lower its PTT index, and the more it manages to offshore pollution abroad. This paper discusses the issues surrounding the appropriate measurement of the PTT index and looks for empirical regularities based on SO2 emission data for a large sample of countries. Two major caveats are raised regarding measurement. First, one should control for differences in emission intensities, over time, between sectors and across countries, by relying on richer databases. This allows decomposing changes in the PTT index into three initial basic effects: the between sector, between country and technique effect. Second, as intermediate goods are traded as well as final goods, an appropriate calculation of trade-embodied emissions should be based on a world input–output table rather than separate national ones. This presents the additional advantage of attributing each production activity to the final demand it is aimed to satisfy, sorting out the complexities of internationally integrated production chains. Moreover, it avoids double counting in trade flows by relying on the value-added content of trade rather than output flows. This leads to two further correction terms adjusting the PTT index: an intermediate trade effect and a value-added effect. Relying on a rich database for SO2 emissions intensities, and controlling for input–output relationships and intermediate trade flows, our analysis suggests that of the two caveats raised above, the first one has the largest empirical importance. Amongst the five effects that compose the PTT index, the differences of emission intensities between countries and over time turn out to be the strongest determinants. Incidentally, this allows to overcome the original paradox reported by Antweiler (1996), as the adjusted PTT index turns out to be negatively associated with GDP per capita (and significantly so in a regression controlling for fixed effects, with an elasticity of − 0.65). The intermediate trade and value-added effects exhibit a weaker influence. This may appear relatively surprising as these effects imply substantial modifications of trade flows, but one should keep in mind that the PTT index is a relative measure, and the fact that the intermediate trade and value-added effects do not change much PTT estimates simply means that the corresponding absolute changes in trade flows have not a strong correlation with emission intensities. Overall, and combining PTT values with trade imbalances, the general pattern in terms of pollution offshoring is one under which most large environmental winners are large rich countries (except for India) and where large environmental losers are large emerging economies (except for Spain and Australia). Gains and losses are fairly concentrated, representing more than 15% of world trade embodied emissions for the two countries located at the extreme of the distribution (USA, the largest winner, and China, the largest loser). It should be clearly stated that the zero-sum game indicator of net environmental gain used in this paper (and many others) is only an accounting measure of trade-embodied emissions, which lies far away from a proper general equilibrium analysis of the impact of trade on the environment (see Antweiler et al., 2001 for a thorough analysis of the SO2 case). However, it deserves interest for at least two reasons. First, when the impact of pollution is mainly local, it is indeed a measure of the environmental damage which is transferred abroad through trade. Second, even when the impact of pollution is regional (SO2) or global (CO2), it is a measure of the additional burden that domestic consumption imposes on the community of trading partners, a critical dimension to design appropriate international environmental agreements. Two final caveats are in order. First, even if we tried to take the best out of available data, caution is required in analyzing results given the adjustment procedures that had to be followed, in particular to estimate emission intensities, input–output coefficients or implicit trade flows. More empirical efforts based on better quality data are certainly needed in the future, in particular for local pollutants and for those two effects of major revealed importance i.e. the between-country and the technique effect. Second, given the greening of production technologies, one may argue that the severity of the problem may vanish over time. However this technique effect has to be balanced with the scale effect arising from the continuous increase in trade flows. In the case of SO2 during the nineties, although emission coefficients decreased in most countries, exports and imports flows boomed to such an extent that the total amount of emissions embodied in world trade did increase. Whether these trends will continue or be reversed in the future is still an open question.