Nordhaus on the Stern Review

Sometimes on this blog I'll say half-baked things, get a few comments, think I'm a real dimwit and wish I never started blogging. And then sometimes I feel OK about it. This is one of those rare latter times. From Mankiw's Blog:

Yale economist Bill Nordhaus, an expert on climate change, reads the Stern Review of the topic and does not like what he finds. The Review's apocalyptic conclusions are, according to Nordhaus, severely overstated because of its assumption of a near-zero discount rate (0.1 percent per year) and log utility (so marginal utility does not decline much as technological progress causes consumption to rise). Here is an excerpt from Nordhaus:

The Review proposes using a social discount rate that is essentially zero. Combined with other assumptions, this magnifies enormously impacts in the distant future and rationalizes deep cuts in emissions, and indeed in all consumption, today. If we were to substitute more conventional discount rates used in other global-warming analyses, by governments, by consumers, or by businesses, the Review’s dramatic results would disappear....

Suppose that scientists discover that a wrinkle in the climatic system will cause damages equal to 0.01 percent of output starting in 2200 and continuing at that rate thereafter. How large a one-time investment would be justified today to remove the wrinkle starting after two centuries? The answer is that a payment of 15 percent of world consumption today (approximately $7 trillion) would pass the Review’s cost-benefit test. This seems completely absurd. The bizarre result arises because the value of the future consumption stream is so high with near-zero discounting that we would trade off a large fraction of today’s income to increase a far-future income stream by a very tiny fraction.

Thanks to Tyler Cowen for the pointer.

Thanks to Greg Mankiw via Tyler Cowen for the pointer.

Here is Nordhaus' full review of the Review.

 

Comments

[Joe Otten]

The discounting question is discussed here, point 8:

http://www.hm-treasury.gov.uk/independent_reviews/stern_review_economics_climate_change/sternreview_faq.cfm

It still isn't that clear in my mind. However, I can see that if we are examining the risks of a substantial loss of life and livelihood due to climate change, then assuming those risks don't exist - assuming ongoing growth - for the purposes of discounting seems wrong.

[Charles Young]

When talking about Nordhaus’ review of Stern, it’s best to start, as he does, by saying that “The Stern Review’s results are fundamentally correct in sign if not in size”. No commentator as far as I am aware has made the case that the costs of following Stern’s policy recommendations outweigh the benefits. Yet some reports of these discussions have wrongly inferred that there is controversy among serious students about whether they should be implemented.

Nordhaus is right to say that it is the discount rate that makes Stern an outlier, and also right to distinguish between the social time preference rate and other forms of discounting.

It is very important to notice that at no point in the discussion does Nordhaus ever make the case for a larger discount rate than the one used by Stern, which, as he rightly says, is based on the idea that ethics requires intergenerational neutrality.

If one had to judge between a group of people at different locations in space, the default must surely be that they all have equal value. If instead one were to propose that the welfare of those situated further North should be prioritised, one would expect this proposition to be followed by some justification. Equally, if we propose that a specific location in time gives priority (so that our grandparents’ welfare is to be judged much more important than our children’s), one would expect to find some argument justifying this claim.

Nordhaus advances no such argument. At most he advances some alternative (but less intuitively appealing) assumptions, though he does not make any case in favour of any of them.

The only case he advances against near-zero discounting is Arrow’s argument that it could be used to require massive sacrifices from one generation to offset minimal gains from many future generations. This is a familiar problem with utilitarian arguments, and not specific to discounting. If I were to prove that the total welfare of everybody today would be improved by robbing those living in latitude 20’06” and distributing the spoils to others, that would not make a case for violating the rights of that group of people.

This is the discussion, in embryo:

Non-discounter: This generation’s welfare counts no more and no less than that of other generations.

Discounter: If that were so, then this generation should make massive sacrifices.

N-d: You misunderstand my assumption. I am not saying that this generation has a uniquely under-privileged position, only that it does not have a privileged one. Its rights matter as much as any.

The “reductio ad absurdum” argument can be used even more cogently against the discounters who, as I’ve pointed out before, are obliged to concede that an extra glass of wine for Alexander the Great matters more than all today’s capital stock.

In judging between these two arguments, common sense is the best guide. If you think that what is at issue in the discussion of climate change is that there is a risk that our generation may be uniquely disadvantaged, and is being called upon to make massive sacrifices in order to produce negligible benefits for future generations, then we are in the Arrow-Nordhaus scenario and you should be questioning whether near-zero discount rates are being misused to put your generation in a uniquely disadvantaged position, compared to past and future ones.

If, by contrast, we think that our generation is a historically privileged one that is being called on to make a fairly small sacrifice of, say, 1-2% of its income, in order to avoid costs of 3-20% to future generations, then we are in the Cowen-Parfit (or Alexander) situation in which misuse of positive discount rates is more of a danger. (I think 3% is Nordhaus’ current estimate, excluding any distributional considerations, which would be bound to inflate it).

By the way, since Mankiw acknowledges Cowen as the source, let's hear what Cowen has to say: "I agree with Stern that the discount rate should be zero or non-zero for alternative consumption streams through time" (which is all that Stern is considering, since he takes the propensity to invest as a given).

[Richard Tol]

Charles:

All very true. But, you cannot use one discount rate for climate change and another discount rate for other problems. A pure rate of time preference of 0.1% per year (as Stern favours on the flimsy ground that the probability of humanity going extinct is 10%per century) would imply a massive increase of investment in education, research, pensions, and so on. Not just in emission reduction.

Besides, a democratically elected government should be careful in deviating from the will of the people. Every study shows that most people have high discount rates. No amount of public goods and risk pooling could defend a social discount rate as low as Stern's.

Finally, if you insist on using a low discount rate, then you better be careful in your modelling of the far future. Stern essentially assume that the only things that will be different in the future are the global mean temperature, the size of the population, and the size of the economy. This is less than careful.

[Richard Tol]

Charles:

All very true. But, you cannot use one discount rate for climate change and another discount rate for other problems. A pure rate of time preference of 0.1% per year (as Stern favours on the flimsy ground that the probability of humanity going extinct is 10%per century) would imply a massive increase of investment in education, research, pensions, and so on. Not just in emission reduction.

Besides, a democratically elected government should be careful in deviating from the will of the people. Every study shows that most people have high discount rates. No amount of public goods and risk pooling could defend a social discount rate as low as Stern's.

Finally, if you insist on using a low discount rate, then you better be careful in your modelling of the far future. Stern essentially assume that the only things that will be different in the future are the global mean temperature, the size of the population, and the size of the economy. This is less than careful.

[J.S.]

One thing that I think is missing from this is that CBA and discounting are simply an efficiency criteria- we can use low or zero discount rates to come up with scenarios where sacrifice today are worth it with respect to benefits for people in the future and still oppose it. Efficiency is not the final say in the matter. We may want to look at distributional issues both with respect to who sacrifices today and who benefits tomorrow. Also, we can look at the fundamental nature of the issue in question- if we're talking about investment in new ice cream that doesn't melt for people in the future to enjoy we can reasonably say that this does not merit sacrifice today based on the non-essential nature of the good in question, whereas species extinction, severe weather, and ecological destruction can rightly be viewed in a different light. Let's not miss the forest for the trees.

J.S.

[John Whitehead]

The discussion on this post and others has really clarified my thinking on the economics of climate change. In short, benefit-cost analysis is a tool for identifying efficient government policies and climate change is a matter of intergenerational equity. We ought to estimate the benefits and costs of climate change to get a handle on the potential problem but any societal decision about what to do ought to be based on equity. If we want to discount at a 0% rate and call it an equity discount rate then OK. But, to say that the costs of climate change will be between 5% and 20% "now and forever" as in the Stern Report is bogus.

[J.S.]

John- I essentially agree- CBA starts to lose focus when the time-horizon is so long as to become an essentially intergenerational issue way off in the future- such as 100-200 years out- at this point it really is equity considerations that are dominant.

J.S.

[John Whitehead]

Richard

Thanks for agreeing that what I said is all very true, and let me repay the compliment. I quite agree with you that the same time preference rate should be used for all decisions. As to the alleged implication that this generation should drastically cut its consumption and invest massively, I tried to deal with that above.

John

I'm not that inclined to argue about whether "5-20% now and forever” is misleading. The important thing about the Stern review is whether or not the policy recommendations get implemented. Can I draw you out on whether or not you are convinced that they should be?

For some reason I can no longer post on this site, and it doesn’t seem to be the site’s fault. So unless some kind person is willing to volunteer to post e-mails on my behalf (I don’t like to go on troubling John, who’s kindly posted this and my previous one), I’ll have to sign off.

Charles Young

[Danø]

Charles, fix your cookie preferences in your browser as a troubleshooting first step [and if you use IE, stop and use Firefox].

I agree with Charles that

I'm not that inclined to argue about whether "5-20% now and forever” is misleading. The important thing about the Stern review is whether or not the policy recommendations get implemented.

as the number is a value expression.

It is fine to disagree on values (what would be left to talk about??).

It is not fine to quibble over whether or not to play a fiddle that is slightly out of tune while Rome rapidly oxidizes.

Best,

D

[John Whitehead]

Charles,

At some point in this debate I've said that I think some non-Bush-style-voluntary climate change policy should be implemented. Here's what I like:

1. tradeable pollution permits on nontransportation energy use and/or a carbon tax
2. an increase in the gas tax

The magnitudes of these economic incentives should be somewhere below that which might avoid the present value of an undiscounted 5% loss in global GDP. In other words, I don't think the impacts will be enormous.

The cap in #1 and the tax in #2 can be adjusted as new information rolls in.