A British academic group has published a major report on geoengineering technologies that could be used to mitigate climate change. These technologies come in two flavors: carbon dioxide removal and solar radiation management (reflecting the sun's energy back into outer space). The bottom line is that many of these technologies have potential, but also pose significant risks and are likely to be extremely costly. For example, capturing carbon dioxide from ambient air would be the most precise and accurate method of addressing climate change, since it would remove the source of the problem; but the know technologies are not cost-effective. Stratospheric aerosols (sending particles into the air to reflect solar energy back into space, as occurs after volcanic eruptions) are cheaper but could have potentially catastrophic effects on other aspects of the climate, such as the ozone layer.

It is not hard to anticipate that, despite these problems, geoengineering will become the rallying cry for skeptics of a climate treaty. Indeed, the British scientists clearly anticipated this reaction, and prepared for it with a rather gloomy title for their press release: "Stop emitting CO2 or geoengineering could be our only hope"—an otherwise puzzling start to what might seem like a pretty optimistic take on an approach to the climate problem which seemed out of the realm of possibility not many years ago. The first sentence of the press release drives the message home: "The future of the Earth could rest on potentially dangerous and unproven geoengineering technologies unless emissions of carbon dioxide can be greatly reduced, the latest Royal Society report has found." Yet the conclusion of the report is that we should invest heavily in researching these technologies.

But the authors of the report are right to worry about climate-treaty skepticism. The right way to approach the climate change problem is first to force people and business to internalize the costs of their emissions. This can be done only through a climate treaty that imposes a tax or a (more-or-less) equivalent quantity restriction with tradable permits. If geoengineering can remove carbon from the atmosphere at an economical cost, then taxes can be reduced or quantities limits raised as necessary. But it is unrealistic to expect that geoengineering will simply eliminate the problem; much more likely, it will mitigate the problem at the margins, so that over the long term greenhouse gas emissions will not need to be lowered to zero, but they will still need to be curtailed, and only a climate treaty can accomplish that goal.

The most worrying implication of the report is that it injects an enormous dose of uncertainty into an already extraordinarily complex problem. It is not hard to imagine countries like China and India deciding that they would rather gamble with geoengineering than risk the short-term economic costs that would be imposed by any meaningful climate treaty. Indeed, they might reasonably expect to free ride on geoengineering initiatives undertaken by the United States and Europe, which would benefit everyone in the world. Ideally, international cooperation would ensure a fair distribution of costs. But the United States and Europe probably cannot credibly tell China that they will not fund geoengineering projects unless China participates, whereas they can credibly tell China that they will not cut emissions unless China does as well. The strategic settings are different. With respect to cutting emissions, the United States and Europe cannot solve the climate problem without Chinese (and Indian and Russian and Brazilian…) participation because of their enormous contributions to the climate problem (China is the leading emitter of greenhouse gases) and because of the "leakage" problem—industry will migrate from regulated countries to unregulated countries. With respect to geoengineering, it is at least possible that the United States and Europe can solve this problem, or at least make great inroads. If so, and if China and other countries respond rationally by free riding, then geoengineering may end up costing the United States a lot more than a reasonable climate treaty would.