1. Stern never says what discount rate he is actually using. I find
this bizarre, to say the least. One account from the FT estimates he uses a
figure between 2 and 3 percent.
2. The correct rate depends on a society’s rate of investment. If
government regulates, taxes, or otherwise pulls resources from the
private sector, we need to estimate how much of these resources would have gone
into investment and how much would have gone into consumption. Stern
never does this.
3. The resources that would have gone into investment should be
discounted by the (risk-adusted) rate of return on investment. This will be much higher than two or three percent, although of course it does not apply to the entire gross upfront cost.
4. The resources that would have gone into consumption are harder to discount, especially if we are comparing those resources across
the generations, and if the change in question is "large"
rather than "small." I tend to favor a very low or zero discount rate in these settings, if only because there is no pure time preference across the
generations. (Before you are born, you are not sitting around impatiently, waiting, unless of course you are a character in Maeterlinck’s The Blue Bird.) In any case this is predominantly an ethical question, and no correct answer follows directly from examining marginal analysis and market prices.
If the change is "small" for the affected people, in the precise sense of not much affecting their marginal utility of wealth, we should discount by the market rate of interest, adjusted for risk, taxes, transactions costs, etc. I don’t find Stern very clear on such matters.
5. Since the Chinese save and invest more than do Americans, the
correct social discount rate should be higher for China than America.
If the Chinese are earning ten percent a year on savings of fifty
percent, that gives a rough discount rate of about five percent. Don’t
tell me that US and A is saving zero percent a year (no way), but we are saving less than the Chinese.
All other things equal, that means we should invest more to stop global warming than should the Chinese, and that is not even considering our higher income.
6. Stern argues that if the environment is worsening, this might
justify a negative discount rate for some environmental amenities.
This is theoretically possible, but with substitutability between
environmental and non-environmental goods, it is unlikely.
7. Cost-benefit analysis works best for small changes which can be
evaluated at market prices. I don’t think it tells us much about
evaluating the costs of, say, one hundred million poor "climate-change"
refugees. Under my ethical views, which refer to a notion of property rights, the true costs of those refugees are
higher than market prices/incomes will indicate.
8. Lower discount rates don’t always make global warming costs more important. Say the rate of discount is zero. This implies that one-time adjustment costs fade into insignificance, compared to ongoing gains from economic growth.
9. For this entire exercise, the results are very very sensitive to
the choice of discount rate. Some of this requires ethics, not just economics. Stern
notes this clearly, but the relevant caveats don’t seem to find their
way into his final presentation of the estimates.
The bottom line: Stern avoids many of the common mistakes in
this area. He stresses that a multiplicity of discount rates is required. But his treatment of discount rates is far from transparent and it is in some regards incorrect. That said, the "mistakes" slant the analysis in both directions, rather than confirming any prior that global warming is a significant economic problem.
The other bottom line: I do understand, and accept, the case for doing something. But I don’t yet see how this report adds to that case. Maybe I’ll read on.
Here is a critique from The Economist, and here. Here is Cass Sunstein on the study.
It also seems the report
relies on excessively high estimates of econoimc growth. Here is one critique of the science. Here is a detailed Bjorn Lomborg critique. New Economist blog points to these supporting materials.