It is with the less-developed countries that we have to be most careful about superimposing the climates of the future on the economies and societies of today. As it was in our own country during this century, the trend in developing countries is to be less dependent on agriculture and less vulnerable to climate in transportation and other activities and health. If per capita income growth in the next 40 years compares with the 40 years just past, vulnerability to climate change should diminish, and the resources available for adaptation should be greater. I say this not to minimize concern about climate change, but to anticipate the question of whether developing countries should make sacrifices in their development to minimize the emission of gases that may change climate to their disadvantage. Their best defense against climate change may be their own continued development.

This is a point worth emphasizing. Some environmentalists argue that developing countries should sacrifice some of their hopes for economic development in the interest of slowing the climate change that may prove disastrous. But the advice contains a contradiction. Any disaster to developing countries from climate change will be a disaster to their economic development.
...
At this point, I appear to have reached the conclusion that the developed world has no self-interest in expensively curtailing carbon consumption and that the developing cannot afford to incur economic penalties to slow the greenhouse effect. There is a mismatch between those who may be vulnerable to climate change and those who can afford to do anything about it.
...
A strong argument for trying seriously to slow climate change is that the developing countries are vulnerable and we care. Developed countries are currently providing $50 billion per year of assistance to the developing world; we would be talking about expending or forgoing perhaps 4-8 times that much to slow emissions and slow climate change. Whether people in the developed democracies could be mobilized to contribute so much to benefit, half a century from now, the people in the countries we now call developing I do not know, but I believe that if the developed countries were prepared to invest, say, $200 billion per year in greenhouse-gas abatement, explicitly for the benefit of developing countries 50 years or more from now, the developing countries would clamor to receive the resources immediately in support of their continued development. There would undoubtedly be abatement opportunities so cheap that they could compete with direct aid to developing countries, but it would be hard to make the case that the countries we now perceive as vulnerable would be better off 50 or 75 years from now if 10 or 20 trillions of dollars had been invested in carbon abatement rather than in their economic development.
...
The third argument for spending heavily to slow climate change is that the conclusions I reported earlier may be quite wrong. I said that the climate models predict that climates will change slowly and not much; the models do not produce discontinuities, surprises, catastrophes. What is known about weather and climate constitutes an equilibrium system...The possibility has to be considered that if global temperature increases, not by the median estimate of three degrees Celsius for a doubling of carbon in the atmosphere, but by four or five degrees and continues to rise beyond the doubling because carbon fuels are still in use worldwide, some atmospheric or oceanic circulatory systems may switch to alternative equilibria, producing regional changes that are both sudden and extreme.
...
Insurance against catastrophes is thus an argument for doing something expensive about greenhouse emissions. But to pay a couple percent of GNP as insurance premium, one would hope to know more about the risk to be averted. I believe research to improve climate predictions should be concentrated on the extreme possibilities, not on modest improvements to median projections.

I said that current estimates suggest that it might cost a couple percent of GNP to postpone the doubling of carbon in the atmosphere by several decades. Is 2 percent a big number or a small one?

That depends on your perspective and on what the comparison is. In recent years 100 billion dollars per year in budgets or taxes has been a politically unmanageable magnitude in the United States. On the other hand, subtracting 2 percent from GNP in perpetuity lowers the GNP curve by not much more than the thickness of a line drawn with a number-two pencil, or to formulate it as I did earlier, it postpones the GNP of 2050 until 2051. I say this not to belittle the loss of 10 trillion dollars from the American GNP over the next 60 years, but only to point out that the insurance premium, if we choose to pay it, will not send us to the poorhouse. The proper question is whether, if we were prepared to spend 2 percent of our GNP in the interest of protecting against damage due to climate change, we might find better use for the money.

I have mentioned one use: directly investing to improve the economies of the poorer countries. Another would be direct investment in preserving species, ecosystems, or wilderness areas.