Reviewing the Stern Review

by John Q on November 21, 2006

The release of the Stern review on the economics of climate change has had a huge impact on the climate change debate in Australia. There had already been signs of movement, but the government was still adamant in rejecting both the Kyoto protocol and any form of emissions trading. And, although the offical position did not dispute the science of climate change, many of the government’s supporters in the media, and even some ministers, were pushing the denialist line.

That was only a few weeks ago. Now the Australian government has endorsed emissions trading (in principle at least) and is calling for a ‘new Kyoto’. Ratifying the old Kyoto is still a step too far for a government that has never disagreed with George Bush on anything, but it’s hard to see how long this position can last.

Given the impact of the Stern review, it’s important to see if it stands up to scrutiny, and I’ve done a series of posts on parts of the report at my blog. My main conclusions:

(i) Stern’s estimates of the cost of stabilising CO2 levels (1 per cent of GDP by 2050) are optimistic, but in the right ballpark
(ii) Stern’s treatment of discounting is correct (More to come on this, I hope)
(iii) Stern underestimates the costs of Business As Usual, particularly in relation to environmental damage
(iv) Headline reporting of Stern overstates the risks of worst-case outcomes in the long tail, but critics are wrong to suggest that low-probability extreme outcomes should be ignored.

Overall, my conclusion is that the Stern review gets the basic economics and the policy recommendation right, even if the presentation is inevitably political.

{ 1 trackback }

Reviewing the Stern Review « Carbonara: CET in Europe
11.21.06 at 1:32 pm



kharris 11.21.06 at 9:21 am

From far, far away, my knowledge of Australian news is slim, and perhaps based on sensationalism in the headlines. But didn’t I hear recently that the government thinks vast areas of Australian farm and pasture land may have to be abandoned for lack of water? Hasn’t the government urged farmers to consider pulling up stakes and moving farther from urban areas, to wetter lands? Strikes me as odd that a government looking the consequences of climate change right in the face has been refusing to take steps to slow the change.


William Goodwin 11.21.06 at 10:18 am

John, it sounds like you may say more about this in a future post (and I read your cursory discussion of the issue on your blog), but I can’t see the justification for the minuscule discount rate if a sizeable chunk of the future damage we’re talking about is economic. If you use a zero discount rate, ensuring that people in the distant future are kept whole in their consumption requires major cuts in our consumption today. But we know to a high degree of certainty that those people are going to be significantly richer than we are, which means that declines in their future consumption (more accurately, declines in their projected consumption) will be significantly less meaningful, in a marginal sense, than declines in consumption today. I may be completely confused about this, but what’s the justification for using a zero (or 0.1%) discount rate and then going on to measure damage, etc., in dollar terms?


Tom T. 11.21.06 at 11:27 am

I may be asking the same question as in #2, but I’m also confused by the implications of a zero discount rate. I understand that if civilization ends, then future consumption drops to zero, but isn’t a zero discount rate assuming a 100% probability of civilization ending?

In perhaps more conventional terms, isn’t a zero discount rate implying a zero opportunity cost? Money spent on climate-change change abatement today will not be available for spending on health care or poverty relief tomorrow. Or more directly, money spent on carbon sequestration technique A today will not be available when carbon sequestration technique B gets invented in ten years. All of these may be valid social choices, but why shouldn’t they be discounted to reflect these trade-offs?


William Goodwin 11.21.06 at 11:58 am

Just to be clear, I’m not disagreeing with the idea of using a zero discount rate for the purposes of intertemporal utility comparisons (that is, we should assume the welfare of future generations is as important as our own – with a slight discount included because there’s a slim chance they might not exist). I’m disagreeing with the idea that when a significant part of that utility is quantified in economic terms, the fact that distant generations will be significantly richer than we are today is irrelevant. (I think this idea is implicit in Stern’s use of a zero discount rate.)


Adamsmithee 11.21.06 at 12:44 pm

I’m also fine with using a low (not zero, but low) discount rate, but I’m not sure it actually makes a difference to the argument on the relative importance of abating climate change.

If you are worried about future generations as much as you are about this generation, you can spend money today on abating climate change and (at a very low discount rate) the costs to you are considerably outweighed by the benefits to future generations. Or you could put the money in the stock market today, deed the stocks and earnings to future generations and (at a very low discount rate) the costs to you are *hugely* outweighed by the benefits to them. Driving a wedge between the social discount rate and the market discount rate makes *any* investment that will pay off for our kids more attractive, not just climate-change-abating investments. And a lot of the other investments that are good for our kids are likely to have even bigger payoffs than climate change abatement… reducing the spread of AIDS and malaria, putting in sewage systems, and so on and so forth. Changing the discount rate does very little to answer the Copenhagen Consensus critque, if you will…


lemuel pitkin 11.21.06 at 1:07 pm

Can you spell out what you mean by “the cost of stabilising CO2 levels”?

Looking at the linked post, you seem to mean that an increase in the price of fossil fuels sufficient to reduce CO2 emissions to a stable level would work out to 1% of total output. I assume you have some reason for thinking this is the right approach even though any actual reduction in fossil fuel use would involve many mechanisms other than an increase in their price?

Bigger question is what this means in terms of more familiar notions of cost. Does it mean that consumption will be percentage point lower, or that GDP growth will be one point lower per year?

It seems to me that to assume the sort of tradeoff implied by “cost” requires unrealistic assumptions, for instance a full-employment economy where increased investment implies decreased consumptionand vice versa. In the real world, a major investment in e.g. alternative energy needn’t involve any decrease in consumption at all.

Second, what’s the relationship between a one-off “cost” like e.g. a major increase in the price of fossil fuels and a decrease in the rate of economic growth?

Finally — and this ties in with the bigger questions — elacticities really depend on the time-scale, don’t they? Measured elasticities will be over periods of months or years. Presumably the elasticity over the decades-long period we’re talking about here would be greater, as people could adjust living arrangements (e.g. commuting distance) to higher fuel prices. Which would make the “cost” by your approach lower, no?


Maynard Handley 11.21.06 at 3:20 pm

“Overall, my conclusion is that the Stern review gets the basic economics and the policy recommendation right”

And could you tell us what those policy recommmendations are. I’ve read a dozen posts on the science the report gives (nothing I haven’t heard before) and arguing about the economics (not my field, so I’ve no opinion on appropriate discount rates) but I’ve seen nothing that actually tells me what these recommendations are (and more importantly) what relationship they have to the real world.

What I mean is, it’s a fine thing to say “we need to implement a carbon tax that adds 1 euro/l to the price of gasoline, and need to start building eighty new nuclear plants”, but that’s nothing people haven’t been saying for ten years; what’s of interest is what’s going to make this year different in terms of moving forward compared to those previous ten years?


John Quiggin 11.21.06 at 3:25 pm

Some quick responses:

kharris, the severity of the drought has played a major role in solidifying public opinion and bringing about the collapse of the government’s earlier position

William and Tom T, I obviously need to explain myself better. Discounting in Stern arises for precisely the reasons you state.

Adam Smithee, the relevant market discount rate here is the real rate of returns on government (or high-grade corporate) bonds, not the return to equity, which incorporates a large premium for systematic risk. The real bond rate is typically close to the rate of economic growth, and this is consistent with Stern’s approach.

Lemuel, standard economic analysis says the effects will be on levels, not growth rates. As you say, elasticities increase and costs decline the longer the time a price increase is in place. That’s an important reason why short-term observations of shocks to energy markets overstate the implied cost of reducing consumption.


leederick 11.21.06 at 4:38 pm

Reading Chapter 2 of Stern really made me feel the whole argument was very shaky. The entire argument is constructed on top of a huge philosophical edifice. Unless you buy into all this, which looks to me like some sort of intergenerational utilitarianism, then it’s not going to convince you.

I’m not sure there really is an obligation to avoid reducing the welfare of future generations. For example: I’m not sure I have a right to be pissed that WWI went and damaged the economy of Europe, thus preventing me from being richer than I would be if it hadn’t been fought.


Chris Bertram 11.21.06 at 4:42 pm

leederick: unless you are very very old, you probably wouldn’t exist if WW1 hadn’t taken place.


William Goodwin 11.21.06 at 5:08 pm

John, given your comment, it may be that my confusion is just based on a misreading — or a missed reading — of the report. What discount rate (or is it discount factor? Stern distinguishes between them) does Stern actually use? I couldn’t find anything in the report that explicitly officially states it. But there’s a suggestion in the Appendix to Chapter 2 that the rate is somewhere between 3-4% (which seems reasonable, if a bit low). Is that the rate he’s using?


leederick 11.21.06 at 5:23 pm

leederick: unless you are very very old, you probably wouldn’t exist if WW1 hadn’t taken place.

That’s kind of my point. In 2100 there may well be quite a few people who wouldn’t exist if spending 1% of GDP on something other than stabilising CO2 levels hadn’t taken place. If that’s the case, do I really have to worry about them pissing on my grave for not doing what Sir Nicholas recommended?


John Quiggin 11.21.06 at 5:44 pm

The theoretical analysis means that the discount rate depends on the rate of growth of income and the elasticity of intertemporal substitution. These parameters vary from case to case, but the implied rate is 2-4 per cent in most cases. This is broadly in line with the market-determined real bond rate (currently US 10-year Treasuries are returning 1-2 per cent real, depending on which inflation measure you use).


William Goodwin 11.21.06 at 6:04 pm

Yes, I was confused. Thank God for the blogosphere. I would have been walking around wrongheaded for months.


Michael Sullivan 11.21.06 at 6:33 pm

cb on comment 11: Huh?

Why wouldn’t various people exist if WW1 hadn’t taken place? I would have supposed that a large number of people might exist who don’t now if WW1 hadn’t taken place. Or do you mean if instead of a WW, we had just let germany and others overrun the land? Color me obtuse, but I don’t get the comment.


Michael Sullivan 11.21.06 at 6:41 pm

jq: What is your reasoning behind ignoring the returns on equity when considering a discount rate? Yes, there is a premium for systemic risk, but if we set aside some % of GDP in a CO2 remediation fund and invested it in equities, would we not be able to achieve the 5-6% real return expected from such investments? If not, why not? If we would achieve those returns, then why is that not the appropriate discount rate?

I was very confused, BTW, by your comments on this issue, both here and in your personal blog. Until your comments 9 & 14 here, you didn’t actually say what discount rate is used or give any reason to justify your assertion that Stern’s treatment of discounts is correct.

I’ve been inclined to use general real GDP growth as the discount rate on the grounds that in practice, if we don’t do anything bout carbon, we won’t alternately set anything aside to take advantage of that risk premium, so we are leaving future generations a problem based on GDP growth alone. So on first cut that seems reasonable.

But if it would cost us less to create a set aside fund, why would we not do that? I don’t feel I have complete clarity on the issue.


lemuel pitkin 11.21.06 at 7:34 pm

If I’m understand him correctly, Michael Sullivan beleives that the return on equity investmetns can remain above GDP growth forever. To which one can only say: Really?


radek 11.21.06 at 7:53 pm

I don’t know about Stern, but this does go back to Pigou and Ramsey who called time discounting “an unethical failure of the imagination” – i.e. argued that the discount rate should be zero and all future generations have an equal weight in considerations of social welfare.
The problem with this, as pointed out by Koopmans is that:
“Without time preference, the utilities achieved by any finite number of generations can be “ignored” as the remaining infinity overwhelms them completely. Indeed, in continuous time, such negligibility is automatic.”

So with a zero discount rate one can justify any harm or cost to present (and a few next) generations for the sake of the utopia that will finally emerge in some distant time. So we should undertaky ANY program to stem global warning, because in generation n+k the benefits will outweight the cost and all future generations after that will benefit.

(see more here:

Of course both approaches have their problems. In practice, people have a time preference for whatever reason (intergenerational altruism, impatience) and the interest rate should reflect these preferences. So yeah, looking at rates of return on government bonds should be a good guess, in the ballpark, as John argues above.
Except. Michael Sullivan is poking about something that is also relevant and that’s the Equity Premium Puzzle. The returns on stocks exceed returns on bonds to a degree that cannot be explained by the difference in riskiness. So perhaps, we should use a higher discount rate afterall?
In any case, this would seem to give a lower and upper bound on “reasonable” discount rates and I’d be interested in how sensitive are Stern’s estimates within those ranges. I’ll get to the report itself when I’m done grading.


John Quiggin 11.21.06 at 7:55 pm

Michael, this is somewhat problematic because no-one really understands the equity premium (it’s much larger than theory suggests it should be) but the suggestion that we can use it as an opportunity cost rate raises big problems. If you think the risk can be ignored in the long run then there is an infinite arbitrage available – borrow at the bond rate and invest at the equity rate. The fact that we don’t see this happening implies that the risk is real, at least for private investors.

I think (but this is something of a side issue) that governments can exploit this arbitrage to some extent, but it’s limited by the availability of investments where the public sector has a comparative advantage or at least no significant disadvantage. This eventually gets you back to a conclusion that the government’s first-best opportunity cost rate is above the bond rate but well below the equity rate.

I can see I’m going to have to write another post on this point.


lemuel pitkin 11.21.06 at 8:01 pm

standard economic analysis says the effects will be on levels, not growth rates

But surely we can’t be satisfied with this? If the entire impact of climate change is the equivalent of loss of a few years of economic growth, why are we worried? People in the future will be much richer than us in any case. Obviously, a reduction of 20 percent in world GDP can’t capture the whole story here.


John Quiggin 11.21.06 at 8:51 pm

“Why are we worried”. First, as I mention, the environmental damage is pretty horrendous, and not properly costed.

Second, you could make the same point about almost anything covered by economics. The Great Depression involved a similar reduction in GDP. Supposing it could have been predicted well in advance it would have been worth taking some low-cost steps to avoid it.


lemuel pitkin 11.21.06 at 9:14 pm

I have no argument with your conclusion — we need to be doing far, far more to deal with climate change now. I just think this discussion could use a bit of the spirit of the earlier post on inequality. I don’t think a massive disruption in society can be regarded as simply a one-off decrease in GDP. Admittedly I don’t have a better alternative in mind.


Kenny Easwaran 11.22.06 at 3:41 am

“critics are wrong to suggest that low-probability extreme outcomes should be ignored.”

I’m professionally interested in questions like this – in an abstract decision-theoretic framework I think your probably right, but I don’t see why that clearly applies in this case. It seems that in clearly single-case circumstances we may be justified in ignoring low-probability extreme outcomes. For instance, in planning for the outcome of elections, it seems reasonable for people to ignore the chance of a third-party candidate winning an election in most US jurisdictions. But perhaps that’s just because the chance is so extremely small?

And what of the experiments at Brookhaven that had some miniscule chance of creating a black hole that would swallow the earth – was that irrational behavior?


Chris Bertram 11.22.06 at 3:57 am

Michael Sullivan:

Why wouldn’t various people exist if WW1 hadn’t taken place? I would have supposed that a large number of people might exist who don’t now if WW1 hadn’t taken place. Or do you mean if instead of a WW, we had just let germany and others overrun the land? Color me obtuse, but I don’t get the comment.

Because if WW1 hadn’t taken place, different people would have met different people, people would have had sex at different times, different sperm would have fertilized different ova.


Aaron_M 11.22.06 at 5:04 am

Chris Bertram:

How does an alternative flow of sperm negate the question from #10? Sure certain individuals would not exist given different historical paths, but the question seems to address a more general issue about intergenerational justice. You seem to be using philosophical trickery to avoid the point.


Chris Bertram 11.22.06 at 7:26 am

Not at all. The fact that no-one who actually ever exists would be worse off as a result of policies that make a future generation worse off than another one would have been, is not a reason to ignore the interests of future generations. Cf Derek Parfit, _Reasons and Persons_ on the non-identity problem.


aaron 11.22.06 at 8:17 am

18., I think it can. If you think about equity as a means of distributing income (share of GDP growth) and that the market is smaller than the ecomomy, the markets should always give a return higher than GDP growth so long as there are segments of the populations that do not participate by investing (they contribute to the economic base by working, but don’t make a claim on future returns, leaving it to the smaller group that does).


Aaron_M 11.22.06 at 9:44 am

Chris maybe you should re-read the post at #10. Leederick can accept that those that were actually born after WWI are worse off than those that would have been born had WWI never happened. In fact this seems to be precisely what Leederick thinks. What he is questioning is the claim that any generation (i.e. him and the current generation) has a moral obligation to try and bring about future outcomes such that for any possible set of future generations that potential generation where we get the best outcome (however that is defined) does come into existance. In other words he is challenging the utilitarian notion of intergenerational justice as such. A utilitarian resolution of the non-indentity problem or the problem of population numbers (which are problems that are largley generate only if you already accept utilitarian premises) does not in any way address Leederick’s question. What we need instead is an argument showing that the underlying utilitarian notion of interngenerational justice is in fact the correct one.


Chris Bertram 11.22.06 at 10:31 am

Aaron, I was reacting to his phrase: “thus preventing me from being richer than I would be if it hadn’t been fought.” It rather presupposes that the same people exist.

I agree with him, and you, in rejecting the idea that we have a duty to _maximize_ the well-being of future generations.


Michael Sullivan 11.22.06 at 12:29 pm

18: I see no reason in *principle* that equity returns cannot remain higher than *current* GDP growth indefinitely. For one thing: GDP is not the same as total wealth, and for another not all investment is equity investment. Most equity investment is leveraged by debt, whose real return may well be less than GDP growth for extended periods.

In the long run, the fundamentals are bounded above by GDP growth (earnings can only be 100% of GDP and no more), but it’s not clear to me that equity returns must come down, rather than GDP growth increase. GDP growth is on a long-run slowly accelerating trend. The unusually high equity premium could be a signal from the market that it expects GDP growth to reach 5% real on average before we start seeing untenable relationships between earnings and GDP.

I’ll be interested to read John’s further thoughts on the matter.


radek 11.22.06 at 7:53 pm

GDP growth is on a long-run slowly accelerating trend

It is? How you figure? Because of growing population?


Eli Rabett 11.23.06 at 2:45 pm

The Stern Report, long as it is, is only a summary. Stern also provides a series of commissioned position papers which are the background for his choices on many of the issues.

Comments on this entry are closed.