I learned Austrian economics over a year before I started learning regular economics. The Austrians taught me to stonewall all cost-benefit analysis with methodological objections. It took me years to see the emptiness of their approach. Cost-benefit analysis is imperfect, but so is every performance measure. We learn a lot more about policy effectiveness if we carefully measure costs and benefits, then reflect on potentially serious flaws, than if we refuse to play the cost-benefit game.
Partly as a result of this experience, I am deeply suspicious when the proponents of any policy dodge requests for cost-benefit analysis. I know cost-benefit analysis isn’t everything. But if I request estimates of net benefits, and you respond with methodological scolding, I hold it against you. Indeed, I summarily reduce my estimate of the net benefits of your policy proposal. After all, if cost-benefit analysis were on your side, you’d probably answer my question instead of questioning my question. Case in point: I asked Yoram Bauman, author of The Cartoon Introduction to Climate Change, to “review the point estimates and confidence intervals” for cost-benefit studies of global warming. Instead of satisfying my curiosity, he alluded to four methodological problems with cost-benefit analysis. This heightened my suspicion that cost-benefit analysis views climate inaction much more favorably than mainstream thinkers allow.
In sharp contrast, Ron Bailey’s new The End of Doom: Environmental Renewal in the Twenty-first Century, gave me just what I wanted: A readable summary of cost-benefit estimates. Highlights from the section on “How Much Will Global Warming Cost?”:
The IPCC reports offer cost estimates for both adaptation and mitigation. The 2014 Adaptation report reckons, assuming that the world takes no steps to deal with climate change, that “global annual economic losses for additional temperature increases of around 2°C are between 0.2 and 2.0 percent of income.” The report adds, “Losses are more likely than not to be greater, rather than smaller, than this range.”
In a 2010 Proceedings of the National Academy of Sciences article, Yale economist William Nordhaus assumed that humanity does nothing to cut greenhouse gas emissions. Nordhaus uses an integrated assessment model that combines the scientific and socioeconomic aspects of climate change to assess policy options for climate change control. His RICE-2010 integrated assessment model found that “of the estimated damages in the uncontrolled (baseline) case, those damages in 2095 are $12 trillion, or 2.8% of global output, for a global temperature increase of 3.4° above 1900 levels.”
Followup work:
In his 2013 book The Climate Casino… Nordhaus notes that a survey of studies that try to estimate the aggregated damages that climate change might inflict at 2.5° comes in at an average of about 1.5 percent of global output. The highest climate damage estimate Nordhaus cites is a 5 percent reduction in income. The much criticized 2006 Stern Review: The Economics of Climate Change suggested that the business-as-usual path of economic growth and greenhouse gas emissions could even reduce future incomes by as a much as 20 percent.
[…]
The IPCC Mitigation report notes that the optimal scenario that it sketches out for keeping greenhouse gas concentrations below 450 ppm would cut future incomes by 2100 by between 3 and 11 percent… [P]rojected IPCC income losses that would result from doing nothing to adapt to climate change appear to be roughly comparable to the losses in income that would occur following efforts to slow climate change. In other words, it appears that doing nothing about climate change now will cost future generations about the same as doing something now.
Is Bailey’s literature review accurate? If not, please name an equally responsive but higher-quality review.
READER COMMENTS
michael pettengill
Apr 28 2016 at 1:06am
The way I see it, the cost of a high effective carbon tax implementation is trillions of dollars in spending to dodge taxes, and the benefit is trillions of dollars in wages and new jobs with millions of workers spending lots more income than they would as unemployed on welfare or underemployed in low wage jobs doing menial work. An effective carbon tax will, as Laffer teaches, generate zero tax revenue.
Economies are zero sum.
All costs in a sustainable economy, when the initial condition is extremely slack labor demand, are entirely labor costs, and all labor costs become demand for gdp, thus driving gdp higher.
Waiting until sea level rises and then moving a billion people to higher ground will cost the labor costs paid to workers who spend the income driving gdp. The sea level rise will redistribute wealth – those rich from owning beach front property will see their wealth transferred to the poor who can only afford land without access to the beaches.
Other changes in environment will have similar impact, perhaps more flooding, more drought, more drought and flooding, etc.
Note that warmer temperatures in higher latitudes will not result in greater agricultural output – sun light, not temperature, is the bigger determinant of agricultural production.
But the cost of not acting today is lower incomes, higher un- and under-employment, more poverty, and lower gdp today.
Kevin Dick
Apr 28 2016 at 1:21am
This paper may be a useful supplement. It looks at the undercounted benefits of warming and examines potential overestimates of costs of warming:
http://papers.ssrn.com/sol3/papers.cfm?abstract_id=2674685
Ally
Apr 28 2016 at 3:02am
michael pettengill:
I’m pretty certain that’s not true.
If that were the case how would one explain the fact that both the median and mean human today is many many times wealthier than the median/mean human living a couple of hundred years ago, despite the fact that World’s population has also grown to several times what it was over the same period?
Nathan
Apr 28 2016 at 5:48am
The IPCC reports themselves are a literature review. They would seem to be the first place to start.
Of course, if you do look at the literature reviewed by the IPCC on the costs of climate change (as opposed to the physical science which is much more rigorous) you’ll quickly note a disturbingly high reliance on non peer-reviewed “grey literature”, especially from ideologically motivated groups like WWF.
Tim Worstall
Apr 28 2016 at 6:18am
It’s a reasonable review from Bailey, as we would expect from Bailey. However, I always take issue with this little bit:
“The much criticized 2006 Stern Review: The Economics of Climate Change suggested that the business-as-usual path of economic growth and greenhouse gas emissions could even reduce future incomes by as a much as 20 percent.”
No, just not so. Business as usual (BAU) at that time meant any and every of the 40 scenarios in the SRES, the Special Report on Emissions Scenarios. The 20 % result is from Stern’s own adaptation of just the one specific scenario. He takes A2, makes it worse, then uses that as his one and only major emissions scenario.
Much the same thing is done today with the newer emissions scenarios. Everyone uses RCP 8.5 as their BAU. When that’s the most extreme of all of the scenarios presented in the literature and almost certainly the one which simply will not happen.
James
Apr 28 2016 at 8:17am
Bryan,
Austrian economics does not reject cost benefit analysis. Every Austrian economist will agree that if a business can refinance a loan at a lower rate of interest with a one time refinancining fee, the right way to decide is to compare NPVs.
Where cost benefit analysis goes off the rails is in stuff like the global warming where, instead of comparing the costs and benefits to a single party as in the case above, the analysis tries to weigh the benefits to some people against the costs to others. It doesn’t help that in the case of global warming, projected costs and benefits are arrived at by multiplying values that are estimated with error.
I can see why people who embrace central planning would bristle at the Austrian aproach to the topic: If the Austrians are right, then there is no good way for a central planing authority to decide between different plans. For a libertarian, it should be sufficient to analyze whether a proposal is voluntary or coercive. Yes, sometimes that can go wrong too, mostly in very contrived thought experiments, but cost benefit analysis goes wrong in real life with a distinct directional bias. Cost estimates for infrastructure projects and other large scale projects consistently undershoot, not just in bull sessions but in real life.
Daublin
Apr 28 2016 at 8:27am
Cost-benefit analysis is good, and indeed should be part of the minimum bar for anyone who wants to be a leader of large numbers of people. I’d put it slightly below things like developing a balanced budged for whatever group you are trying to lead. When you refuse to do things like this, you are putting the larger group at tremendous risk.
In the case of global warming, one way to do the cost-benefit analysis is to say, okay, suppose we commit to holding temporatures steady by 2116. You can then look at the costs of doing this in different ways. The costs of doing so via a carbon tax are prohibitive; the tax will have to be high enough that people reduce carbon emissions by around 90%. I think it is fair to say there are cheaper ways to do it than that.
Andrew_FL
Apr 28 2016 at 9:30am
The costs of global warming cited are, generally, assuming future projected amounts of warming are correct (so far they are higher than the actual change occurring, on average) and that lots of bad weathery things happen because of that warming-which, generally speaking, have not been demonstrated to be occurring and lack theoretical grounding. If anything the estimates are probably too high, methodology aside.
Ray
Apr 28 2016 at 9:52am
[Comment removed pending confirmation of email address. Email the webmaster@econlib.org to request restoring this comment. A valid email address is required to post comments on EconLog and EconTalk.–Econlib Ed.]
Josiah
Apr 28 2016 at 9:55am
Bailey’s literature review is accurate, but you need to be careful about what precisely is and isn’t being said.
Adam Martin
Apr 28 2016 at 11:02am
David Friedman has taken issue with some of Nordhaus’s methods:
Here
And here
Levi R
Apr 28 2016 at 11:16am
Much of the analysis in Leland Yeager’s paper “Pareto Optimality in Policy Espousal” in the Journal of Libertarian Studies applies to cost-benefit analysis in the policy world. Interesting paper.
Bob Murphy has, over time, written what amounts to a literature review much like the passages you cite above. His stuff can be found at the Institute for Energy Research. But yeah, Austrians hate cost-benefit analysis. 😉
Ben H.
Apr 28 2016 at 12:05pm
Reading this is a perfect illustration of confirmation bias in action. The denialist position has evolved over time. First, the standard tack was to deny that climate change was even happening. Well, that turned out to be wrong. Then, the next approach was to deny that it was caused by humans. Strike two. Now the new fad is to admit that it is real, and that it is caused by humans, but to question the need to do anything at all about it. I would suggest that given the track record of the denialists thus far, and the clear evidence that those invested in denialism – from Exxon to Heritage – have been willing to lie and fabricate endlessly in pursuit of their deliberate spread of disinformation, you ought to have a very strong prior against believing anything that comes from the denialist camp. Sadly, you seem to approach books like this without the strong prior against the denialist position that you ought to have, given that history.
I have no alternative literature review to offer you, not being an expert in this field, although the suggestion above that you might wish to read the IPCC report directly, rather than trusting Bailey’s summary of it, seems like a good idea to me (while recognizing that the IPCC report is in fact a strongly conservative document, since it was produced by a consensus-based process). I’m not an economist, nor a climate scientist, and I’m not going to attempt to pose as one here. I’m sure lots of papers have been written about these things. If you truly want good references to papers from a perspective contrary to Bailey’s, you might try asking people outside of the echo chamber of this blog (you do realize this is an echo chamber, right?). One sign of confirmation bias is a tendency to seek information only from those who are likely to agree with you. Why don’t you ask, say, James Hansen?
Regarding the quotes you give from Bailey’s book, I would just note that the temperature increases cited of 2°C, 2.5°C, and even 3.4°C are toward the low end of projections. Even with a really concerted effort we will probably blow through 2°C pretty soon, and 2.5°C will fall soon after. 3.4°C by 2100 is a more defensible estimate for “business as usual”, depending upon exactly what one means by that phrase, but is still toward the low end of current predictions as I understand it. The high end of current predictions is much higher. So one question you ought to be asking is: what would the economic cost be if the high end turns out to be the reality? These are all just model projections, and as global temperatures climb we stray into realms where we don’t even know whether our models will be accurate at all or not. So the full range of projections needs to be taken seriously; cherry-picking estimates toward the low end is clearly biased. In particular, pretending that cost estimates for 2°C represent a case in which “the world takes no steps to deal with climate change”, as Bailey apparently does in the passage you quote, appears to me to be quite simply intellectually dishonest, and ought to make you distrust everything else he writes.
I would also wonder exactly what costs are being included in the cost estimates given. It has recently been in the news, for example, that massive bleaching and death of corals in the Great Barrier Reef is being observed already, due to elevated ocean temperatures. This is not unexpected, and the projection is that most of the coral reefs around the world are likely to be dead fairly soon. How exactly is the cost of that figured into these cost estimates, I wonder? I imagine they factor in the economic impact on fishing, and on tourism. But what price do they put on the loss of beauty in the world? What’s the dollar amount on that, and how is it defended? Similarly with, for example, the destruction of the Amazon rainforest due to drying and rising temperatures; many models predict that this rainforest will completely disappear, to be replaced by open savannah. An economic cost can be attached to that, no doubt; but what is the dollar value on the loss of all of the species that exist only in that rainforest? What we are talking about is the permanent loss of the planet as we know it. Once it is gone, we will not be able to get it back. You may not care about that, but a great many people do. I am all in favor of cost-benefit analysis; but it is made difficult in this case by the fact that I, and many others, would attach an essentially infinite value to the continuing existence of things like coral reefs and the Amazon rainforest. Economic analyses that I have seen, on the other hand, tend to attach a value of $0 to such considerations, apart from the direct and measurable economic benefits that they produce. That problem ought to be taken into account in your evaluation of this sort of thing. Maybe you regard that as merely a sort of “methodological scolding”, but I would say it is much more than that; it is a reasoned argument for all such economic estimates being massive underestimates of the true cost. If you want to do a good cost-benefit analysis, you can’t just dismiss that objection as “methodological scolding”, you simply must find a way to correct for it in your analysis. Methodological problems cannot always be completely eliminated, of course, and that is not a reason to throw out the methodology completely; I agree with you on that. At the same time, a good-faith effort must be made to correct for known methodological biases, to the extent possible. The temptation many conservative economists exhibit not to do so in the case of climate change is, I would say, yet another manifestation of confirmation bias.
Thomas
Apr 28 2016 at 12:13pm
James has anticipated my objection to cost-benefit analysis, as it applies to many parties (e.g., everyone in the U.S. or on the planet). One person’s costs are another person’s benefits. A classic example is longer growing season in the North vs. greater demand for electricity in the South. Cost-benefit analysis implies a god-like weighing of those disparate effects against the cost of curtailing global warming. If the cost (undoubtedly astronomical) could be levied only against those who benefit, that would be a step in the right direction. But there’s still a presumption that government should step in, instead of leaving action in the hands of those who would be harmed (e.g., voluntary migration, mitigation through market-based action).
Neil S
Apr 28 2016 at 12:38pm
Ben H.
Given that you start by denigrating those who disagree with you (“denialists”), “have no alternative literature review to offer”, and are “not an economist, nor a climate scientist”, why should anyone pay attention to your assertions?
Confirmation bias much? Beams and motes.
MikeP
Apr 28 2016 at 12:53pm
Ben H.,
The denialist position hasn’t changed over time. What has changed over time is which particular “denialists” proselytizers point to as the most vulnerable to mockery.
I, for instance, have for more than a decade had a consistent stand on climate change: science is not policy, and even if you take the IPCC at its word on the science, the economic case for significant efforts in order to reduce CO2 emissions is simply not there.
So I do take IPCC at its word — though apparently you do not. I have never denied anthropomorphic climate change. Nonetheless, reading the IPCC ARs and Nordhaus’s serial monographs leads me to the inescapable conclusion that addressing climate change at this time costs far more than it benefits.
Mark Barbieri
Apr 28 2016 at 2:05pm
Aren’t these the logical positions to take? When someone says that something catastrophic is happening, don’t your first say “prove it?” When they have done so and they also say they know the cause, don’t you say “prove it?” If they do that, and the propose some specific action, don’t you ask whether it is justified? Surely your issue is not that they have asked for proofs and justifications but that they have not been convinced as easily as you.
Mark Barbieri
Apr 28 2016 at 2:07pm
What? I think you lost virtually every reader of this site with that line.
If that is the case, explain to me why people use greenhouses.
ThaomasH
Apr 28 2016 at 2:28pm
@ Thomas and James
“One person’s costs are another person’s benefits.”
No always and no necessarily equally. Sure a longer growing season in the “north” may offset worse weather in the “south” but that just means we need to net the gains/losses, not that they are necessarily equal and offsetting. And it’s hard to see where the offsetting benefits from sea level rise will come from.
I would be pretty happy if climate change policy were being made on a cost benefit basis. I do not think its alarming, indeed it should be a very good thing, if people who are skeptical of the ability of governments to make good policies scrutinize the net costs and benefits of different proposals based on different physical and economic models.
One of the great things about a carbon tax is that the rate can be changed as estimates of the cost of marginal accumulation of CO2 changes. At the very least a carbon tax would be a lot cheaper than the combination of subsidies and mandates presently being used.
ThaomasH
Apr 28 2016 at 2:43pm
@ James
Where cost benefit analysis goes off the rails is in stuff like the global warming where, instead of comparing the costs and benefits to a single party as in the case above, the analysis tries to weigh the benefits to some people against the costs to others.
I’m not sure I understand this criticism. Is it that the distribution of the costs and benefits could be regressive? That to achieve the benefit of saving $10 worth of consumption of beach-front property services by 1 Mr Rich Guy, policy Z would cost $1 worth of consumption by 9 Poor Guys and that cost benefit analysis would say policy Z is desirable?
Or do you reject comparing benefits and costs between people, period? And if so, what other decision rule can we use?
MikeP
Apr 28 2016 at 3:25pm
Much the same thing is done today with the newer emissions scenarios. Everyone uses RCP 8.5 as their BAU. When that’s the most extreme of all of the scenarios presented in the literature and almost certainly the one which simply will not happen.
Indeed, this extreme bias of the compared future scenarios has been going on for some time. It is likely to dramatically distort and pollute the results of future analysis.
In the U.S. Global Change Research Program’s Third National Climate Assessment, we find this:
That the two scenarios used are the less-optimistic high-carbon A2 and the more-optimistic low-carbon B1 is simply fraud, perpetrated by interested state parties at a massive scale. And don’t be surprised when future assessments drop this paragraph as understood by experts and uninteresting to the public, by which time it will be almost impossible to get to the truth behind the models.
David Rager
Apr 28 2016 at 3:53pm
[Comment removed pending confirmation of email address. Email the webmaster@econlib.org to request restoring this comment. A valid email address is required to post comments on EconLog and EconTalk.–Econlib Ed.]
Mark Bahner
Apr 28 2016 at 6:21pm
I don’t think this was the particular criticism, but the huge problem I see in establishing a “cost/benefit” analysis for global warming is that the costs are born by the already-living for the benefit of the not-yet-born (or at least the extremely young).
I think most people would agree that it’s reasonable to leave a world in which the not-yet-born are *worse* off than already-living today. But I have not seen any serious scholarly analyses that make the case that people born in 2100 (for example) will be worse off, on net, than the people of today.
For example, Ron Bailey wrote, regarding analysis by William Nordhaus:
If $12 trillion represents 2.8% of global output in 2095, that means that the global output in 2095 is $12 trillion/0.028 = $428 trillion. Now, I’m too lazy to check, but I’ll assume that the Nordhaus value was adjusted for inflation back to the year of the analysis, which was apparently circa 2010. Well, in 2010, the world GDP was only $65 trillion. So the world GDP in 2095 is projected to be 428/65 = 6.6 TIMES higher than in 2010. So let’s even assume that the GDP is reduced by a ridiculous 50% (no analysis I’ve seen comes even close to that…even for the highest warmings). That still means the world GDP in 2095 is 3.3 times higher than in 2010.
So where is the moral case for the people of 2010 sacrificing for the benefit of the people in 2095?
Mark Bahner
Apr 28 2016 at 6:47pm
D-oh! I wrote:
That should have been “I think most people would agree that it’s reasonable to leave a world in which the not-yet-born are not *worse* off than the already-living today.”
Or to remove the double negatives, “I think most people would agree that it’s reasonable to leave a world in which the not-yet-born are (on net!) better off than the already-living today.”
Steve Brecher
Apr 28 2016 at 7:20pm
MikeP, w/r:
Sorry, I don’t understand why using those two scenarios is fraud. (I request a bit of elaboration as a student, not as a critic.)
Glen Raphael
Apr 28 2016 at 11:49pm
Ben H.,
Suppose you are a student trying to publish a mathematical proof as your thesis. The proof involves a dozen steps but if it all holds up you will have established an important result. Your thesis advisor thinks you haven’t sufficiently established step 1. So you go off and work on it for a year or so and come back having shored up the logic and evidence and argument such that step 1 is finally something you could defend before a committee. Your adviser moves on to the next possibly-weak link in your chain of reasoning which happens to be step 2. He says “I don’t think you’ve proven step 2 yet.” So you go back to work and successfully figure out how to prove step 2.
Your advisor now accepts that steps 1 and 2 are proven but asks you to provide better evidence for step 3.
How do we feel about the advisor’s behavior so far?
Do we call the second objection “strike two“? Do we call the advisor a dishonest liar because he changed his main objection as your argument developed? In a proof with a dozen parts, does the fact that you proved the first two when challenged mean we ought to revise our priors to assume “given the track record” you’re probably right about the whole thing?
Or is the advisor simply being reasonable?
Getting back to climate, the claim “we have to DO SOMETHING NOW about global warming” typically is based on a set of assumptions along these lines:
Remember, your thesis adviser doubted EVERY STEP of your math thesis and wouldn’t let you defend it until EVERY STEP was well established because to do otherwise wouldn’t be science – it’d mean accepting the bulk of the logical case just on your say-so. Similarly, climate skeptics applied rigorous and appropriate doubt to EVERY STEP behind the alarmist assertion “we need to do something now about global warming”.
That’s not irrational, it’s not dishonest, it’s just science. In science, you don’t get to extrapolate from “I proved these two points you doubted” to “therefore I am right about everything!” You still have to prove the whole thing.
So no, the fact that some people were wrong to doubt earlier steps doesn’t let you skip past the cost-benefit analysis now.
Mark Bahner
Apr 29 2016 at 12:19am
Yes, I laugh at the naivete of who say there can’t possibly be a conspiracy of thousands of scientists.
To see that there can be such a conspiracy, just look at the fact that the all the scientists that worked on the IPCC AR5 never made any attempt to point out how very unlikely emissions are to be as high as in RCP8.5.
“Science is self-correcting.” LOL.
MikeP
Apr 29 2016 at 12:35am
Sorry, I don’t understand why using those two scenarios is fraud.
The IPCC SRES scenarios were developed as inputs to climate change models: something needs to drive predicted future economic growth and future choices of high-carbon or low-carbon energy sources.
There are four main scenarios: A1 is high-growth high-carbon, A2 is low-growth high-carbon, B1 is high-growth low-carbon, and B2 is low-growth low-carbon.
Here is how the SRES describes A2 in a nutshell:
Comparing A2 and B1 is fraud because it’s comparing apples and oranges. It contrasts a globally harmonious low-carbon future against a globally contentious high-carbon future. The reason they do this is that the higher growth high-carbon A1 scenario fares quite well under climate models since under that scenario wealth increases faster than damages from climate change. The only way that measures such as carbon taxes can demonstrate significant NPV benefits is by pitting them against business as usual A2 instead of A1.
Unlearning
Apr 29 2016 at 4:07am
It’s pretty weird that you reject Bauman’s critique of CBA but don’t actually give any reasons why. What if it’s just not suitable for analysing climate change? I don’t see why you’ve given any reason for opponents of CBA to accept your arguments.
Nathan W
Apr 29 2016 at 7:32am
The benefits’ analyses focus on carbon fertilizer extrapolation without considering the effects of heat stress on yield. To date I have not seen a credible accounting of the benefits side of the equation.
I have also not seen any reports which discuss the benefits side of the equation which address the costs side of the equation in any manner other than with an obvious agenda to discredit the calculations (without considering the legitimacy of the confidence intervals in the forecasts/estimates).
I perceive the benefits accounting to be even less credible than the costs accounting on this question.
Steve Brecher
Apr 29 2016 at 9:06am
MikeP, I had naively misinterpreted less- and more-optimistic to refer to carbon emission level rather than to economic growth. Thanks very much for your elaboration.
Floccina
Apr 29 2016 at 11:55am
michael pettengill wrote:
I think that you are wrong on that. At the latitudes that we are likely going to see because in the growing season the length of the day condensates for the angle of the sun.
Here is something about the subject. Too busy with work to look for more right now.
James
Apr 29 2016 at 2:45pm
ThaomasH,
I’m not very concerned with regressiveness.
My objection is to any use of estimates which doesn’t account for estimation error. When we net my gains against your loss, the result depends on an assumed rate of substitution between your welfare and mine. Most cost benefit analysis assumes that all persons have the same utility function. That assumption is just a guess (as opposed to a measurement) and is therefore subject to error. It would have to be possible to measure the size of this error (estimated net utility vs actual) to get a reliable confidence interval around the estimated net gain or loss.
I think it’s telling that most advocates of cost benefit analysis do not even regard this estimation error as a real quantity. They get to sounding very Austrian when you ask them how big is the estimation error around their assumed marginal rate of substitution.
In my earlier comment I offer an alternative in the last paragraph.
Yoguy
May 1 2016 at 9:12pm
“In other words, it appears that doing nothing about climate change now will cost future generations about the same as doing something now.”
Please, this is a huge and common mistake that you should debunk and not publicize! “Doing something now” is not a binary decision, the question you should ask is “What is the cost of the marginal tCO2 emitted today?” and to find an answer you can look at various places but Nordhaus is a reference and computes 12$.
See here : http://aida.econ.yale.edu/~nordhaus/homepage/documents/CFDP1826.pdf
There is review at the end
Comments are closed.