WHICH do you value: production or consumption? The preference has long defined economic questions ranging from tax policies to development. Now it matters in national carbon accounting too. If you look at production, you count the carbon that comes from a territory’s smokestacks, exhaust pipes and forest fires. With consumption you tot up the carbon emitted when providing the goods and services bought there. Looked at that way, international trade is a carbon-dioxide pipeline moving responsibility for its effects on the climate from place to place.

Just over a quarter of all industrially emitted carbon moves about in this way, up from a fifth in 1990. The net flow is from the developing to the developed world. So in terms of consumption the rich countries are still the world’s biggest emitters (see chart). New research shows that the flow’s volume is rapidly increasing.

Glen Peters at the Centre for International Climate and Environmental Research, Oslo, with colleagues elsewhere, has looked at the carbon content of international trade since 1990—the benchmark year for measuring emissions under the UN’s Kyoto protocol. The annual growth of CO2 emissions from exported products was 4.3%—faster than GDP or carbon emissions in general, slower than world trade. But it was 17% for trade between developed countries (those expected to meet the Kyoto emissions cuts) and the rest of the world, rising from 400m tonnes in 1990 to 1.6 billion in 2008.

On a production basis, many of the rich countries (but not America, which has not ratified Kyoto) have cut their emissions—by 6% in 1990-2008 in the case of the European Union. But the EU’s imports of embodied carbon from developing countries rose a lot more than its local emissions fell. Overall, the rich world’s increase in “carbon imports” is six times bigger than cuts in the developed countries’ own industrial emissions. The lion’s share of this carbon comes, predictably enough, from China; 18% of the global increase in emissions since 1990 is embodied in Chinese exports.

Mr Peters and his colleagues see no evidence so far that carbon-control policies, weak as they are, are shifting production to less regulated countries. Carbon follows trade patterns set by other factors; it does not shape them. Sterner carbon restrictions, though, might provoke rich-world industrialists to press for tariffs on carbon-intensive imports with which they cannot compete. A more fruitful approach might be to see the trend in terms of the need for greener investment outside the rich world. Spreading low-carbon technologies there matters as much or more than decarbonising developed countries.