In the Kyoto Protocol the absence of Green House Gases (GHGs) commitments of developing countries (non-Annex 1) and the more flexible terms of implementation which are allowed to countries shifting toward a market economy (transition economies) naturally lead to the absence or to less constraining national measures and policies of reduction of the GHGs emissions which, in turn, may determine a comparative advantage in the production of the highest energy/carbon intensive commodities for these countries. These arguments are valid also considering the future implementation of the European Emission Allowance Trading Scheme (EATS). Thus, developing countries may become a haven for the production of not environmental-friendly commodities;, in this case, the so-called Pollution Haven Hypothesis, stating that due to freer international trade the comparative advantage may change the economic structure and consequently the trade patterns of the countries linked by trade relationships, could occur. This would lead to the increase of the transfers of energy and carbon embodied in traded commodities from developing countries and transition economies toward Kyoto or EATS constrained countries. The aim of this paper is to verify if for Italy, as a Kyoto and EATS complying country, evidence of a change in the trade patterns, occurred on the basis of the Pollution Haven Hypothesis, does exist and to estimate the magnitude of the under-estimation of the carbon actually emitted: the carbon leakage. The Input-Output model has been used to calculate the intensities of energy consumption and the related Green House Gases emission, for each Italian economic sector. (c) 2004 Elsevier Ltd. All rights reserved.