The National Review: Your pocket-change or your life

David Brooks recently wrote an excellent column in the New York Times explaining why it is so important for the Senate to pass comprehensive climate and energy legislation. In passing, however, Brooks asks whether we know that the benefits of carbon mitigation exceed the costs, and he mentions an article in the National Review by Jim Manzi of the Manhattan Institute, entitled “Dunce Cap and Trade.”   

But where does the dunce cap belong?  A close look at Manzi’s paper shows that it is riddled with factual errors and distortions, and reveals an embarrassing ignorance of the economic models on which it relies.

Manzi’s article is so problematic it is hard to know where to begin. So I will start with a high level critique of his argument, followed by a miscellaneous list of errors and distortions for those of you who crave more details.

His basic argument goes something like this:  

1. Cost-benefit analysis is the right decision-making tool for addressing climate change.
2. Economists generally agree that the expected costs of Waxman-Markey (and other bills like it) far exceed benefits that are highly uncertain.
3. By implication, economists would therefore object to passing a bill like Waxman-Markey.
4. In situations with great uncertainty, the best strategy is to “keep your options open.” Specifically, keep the option open to have a little more GDP to spend on technology and adaptation in the future, rather than spending resources now on mitigation.

Points 2 and 3 are factually incorrect and misleading, and 1 and 4 conceptually problematic: 

1. There is a broad consensus (click here, here, and here) among economists that such legislation needs to be passed, and that the cost of doing so will be modest. Some economists, including two Nobel Prize winners and other prominent economists, even expect that passing legislation will generate economic growth. The consensus? Economic models consistently project that the cost of climate legislation will be insignificant when compared to projected economic growth. These models unanimously predict carbon mitigation will cost roughly 1% to 3.5% of GDP by 2050, but far greater total GDP by 2050: approximately 150% percent higher than today, in fact. Surprisingly, these results are true of both non-partisan studies and studies funded by the opposition. 
2. Manzi’s view of climate mitigation might seem odd to normal people, who tend to think differently about low-probability/high-risk disasters than he does. Manzi believes we should reduce carbon emissions only insofar as the expected benefits of doing so exceed the expected costs. In other words, he is using a simplistic cost-benefit framework that does not account for catastrophic risks.  A more sophisticated approach would view climate mitigation as an insurance policy against catastrophic impacts, even of uncertain probability, an insurance package itself a benefit worth buying and accounting for on the benefits side. Manzi derides this perspective: “This concept has been called, somewhat grandiosely, the ‘precautionary principle.’ Once you get past all the table-pounding, this is the crux of the argument for emissions abatement. It is an emotionally appealing political position, as it is easy to argue that we should oppose some consumption now to head off even a low-odds possibility of disaster. ” Yes, that’s exactly the point: a disaster. What kind of catastrophic risks are we looking at with climate change? If we do nothing to mitigate emissions, we are looking at CO2 concentrations as high as eight to fifteen (see Figure 2) times the historical range of CO2 that differentiates an ice age from today. Over the last 600,000 years, CO2 has been 100 ppm higher in warming periods, such as the one in which we are now are, than during ice ages. During this time, the natural historical range is roughly 180 ppm to 280 pmm, and average global temperatures only 12 degrees C apart. Today we are at 380 ppm. If we do nothing, we are looking at 800 to 1,500 ppm by the end of the century. Given that only 12 degrees and 100 ppm separates us from an ice age, you might even wonder if the science is too conservative. Not a wild speculation, given that every report by the IPCC has turned out more dire than the previous.

Risk management in the face of potentially low-probability/high-catastrophe outcomes is an approach that is familiar to, and probably preferred by, many of us. Consider, for example, life insurance. While not a perfect analogy (death ultimately cannot be avoided, but we still have a shot at reducing carbon emissions), life insurance is an investment with huge negative expected returns (that’s why insurance companies make so much money). Premiums can range from $10,000 to $20,000 per year for a 70 year old woman, yet many women buy it. My mom, bless her heart, forks over 25% of her annual after-tax income on such premiums, and she is not alone. How does 25% for an outcome we ultimately cannot prevent (death) stack up against 1-3.5% of GDP for one we potentially can? Manzi argues that we should save this premium in order to have more money now and in future years to invest in technology and adaptation. But this closes off the most important option of all: avoiding an irreversible increase in atmospheric concentration levels of 800 to 1,500 parts per million, truly astounding numbers when put into their proper historical perspective.

Hopefully points A and B are sobering enough for you to feel satisfied stopping here. But for those of you who favor a more cost-benefit approach, it is important to understand the errors in Manzi’s article. While he makes some valid points about the difficulty of getting an effective international agreement, his net benefit calculation is fundamentally flawed:

Expected Costs:

• Manzi raises the spectre of high consumer costs, but he ignores the measures in pending legislation that protect consumers. He leaves out the fact that most of the valuable pollution allowances will be returned to consumers to protect them from higher energy costs. As the World Resources Institute has found: “The amount of allowance value directed to consumers or for other public benefit totals 76 percent of the cumulative allowance pool from 2012 through 2025 and 83 percent from 2012 through 2050. This is value that is not directly distributed to regulated entities in any way… ”
• Manzi asserts that EPA has underestimated household costs – averaging $160 per year out to 2050.  But in fact, EPA ran other scenarios with higher cost assumptions, and even EPA’s worse-case scenario for household costs isn’t a show stopper: $417 per year ($1.17 per day per household), against a median household income $6,300 higher on average than today’s levels. (These are discounted numbers; undiscounted, however, does not qualitatively change the story).
• Manzi also doesn’t acknowledge that economic models of climate legislation do not account for innovations in the future. How likely is it that there won’t be any cost-reducing innovations in low carbon energy over the course of forty years? Just think about cell phones over only the last decade. The profit motive is a very powerful thing indeed, and putting a price on carbon creates just that for clean energy.

Expected benefits:

• Manzi refers to a benefits estimate by Yale’s William Nordhaus. Nordhaus projects a loss of 3% of world GDP by 2100 without emission reductions, mostly in other countries. But Nordhaus’s numbers stand at the low end of the spectrum. A different study, done for the UK government by Nicholas Stern, finds that the cost of inaction would be 5-20% of world GDP. Another finds that only four impacts (hurricane damages, real estate, energy-sector expenditures, water expenditures) could cost the U.S. alone 1.8% of GDP by 2100.
• A critical explanation for the difference between the Nordhaus and Stern estimates is the discount rate, a number used to translate (more exactly, “discount”) the value a person assigns to getting a lump sum of money today versus waiting for the same (i.e. inflation-adjusted) amount in some future year. In Nordhaus’s model, future damages are discounted at 4% (p. 10, A Question of Balance), which means that $1,000 worth of damage 100 years from now is valued at $20. I wouldn’t want to be Nordhaus’s grandchild. Stern used a discount rate of 1.4%, putting $1000 at $250 today. Not great either, but certainly not as bad. 

How should we evaluate these two very different choices? Economists consider three factors to be relevant in choosing a discount rate: 1) people tend to be myopic, preferring to consume now than wait until later; 2) if there is economic growth, there will be more money available to offset future costs; and 3) there is the possibility, even if remote, that the planet will be destroyed by some other event (e.g. a meteor).

Nordhaus’s discount rate includes a value for 1). The problem with this is that the future person receiving the damages is not the one making the decision today, so this component of the discount rate is not relevant to climate change and should be removed. In regards to 2), there are two issues. First, can we assume there will be economic growth with absolute certainty? Is it possible that climate change damages will result in net negative growth? Second, the justification assumes that money is a perfect substitute for nature and human lives. Can money make up for things like collapsed ecosystems, or deaths cause by climate change? There is a long contentious debate on discount rates in the arcane halls of academia that is beyond the scope of a blog, but many people view Stern’s discount rate as more justified than Nordhaus’s. Perhaps it is no coincidence that Stern’s discount rate falls squarely within the ranges that OMB and EPA suggest (p. 9) might be appropriate for discounting intergenerational costs and benefits (OMB suggests 1-3%, and the Environmental Protection Agency .5%-3%).

• Manzi seems to be very uninformed about what models like Nordhaus’s are and are not, capable of, the latter being the operative factor. Such models, called integrated assessment models, or IAMs, do not estimate a very large number of very bad impacts. To name only a few: social and political unrest abroad that affects U.S. national security (e.g., violent conflict, mass migration due to infectious disease or extreme weather events, humanitarian crises), mass species extinction, the loss of coral reefs and other valued ecosystems and ecosystem services.
• These models also assume very low costs for adaptation. For example, even in the model used by Stern, in OECD countries 100 percent of the economic damages resulting from the first 2 degree C of warming are eliminated, and 90 percent of those above 2 degree C. For non-economic (e.g. impacts on wilderness areas and animals) and non-catastrophic damages, adaptation is assumed to remove 25 percent of the impact everywhere. No adaptation is assumed for catastrophic damages
• Another reason for low damage estimates from these models is that they often include questionable benefits from climate change. For example, Nordhaus assumes (on very thin evidence) that large benefits (i.e. negative damages) will accrue to people who live in areas where average annual temperatures are projected to rise to 70 degrees. Nordhaus monetizes these benefits using estimates of how much money wealthy people from Northern latitudes spend to vacation in places whose temperature is 70 degrees. But does an ideal average year-round temperature of 70 degrees make sense? Perhaps, if you think the temperatures in Houston make it an ideal city to live in.
• Finally, Manzi mistakenly asserts that Nordhaus’s model takes into account people’s concerns for high-risk low-probability events. He is wrong both technically and conceptually. Technically, Nordhaus allows for a higher-than-expected temperature increase (and a lower-than-expected one), as do many modelers. But the high end values are lower than the 5% likelihood of 7 degree C  (12.6 F) and 1% likelihood of 10 degree C (18 F) implied by IPCC estimates (Weitzman, 2009, Review of Economics and Statistics). Conceptually, by calculating expected, or average, damages, Nordhaus’s approach is a cost-benefit analysis, but people do not value the kinds of risks presented by climate change from this perspective. The problem is more properly framed, as discussed in point B above, as one of minimizing risk, not maximizing returns from climate mitigation.

Alas, If I had more time, I could discuss many more problems with Manzi’s analysis. This list is exhausting, but by no means exhaustive... 

Comments