Global Warming and Discount Rates

The latest report of the Intergovernmental Panel on Climate Change, issued on Friday, confirms the scientific consensus that the emission of carbon dioxide and other greenhouse gases, as a result of the combustion of fossil fuels such as oil and gas, and other human activities (such as deforestation by burning), is having significant and on the whole negative effects by causing global temperatures and sea levels to rise. See http//ipcc-wg1.ucar.edu/wg1/docs/WG1AR4_SPM_PlenaryApproved.pdf. I discussed global warming in my book Catastrophe: Risk and Response (2004), I considered the evidence that global warming was a serious problem for which man-made emissions were the principal cause altogether convincing--and since then more evidence has accumulated and the voices of the dissenters are growing weaker. The global-warming skeptics are beginning to sound like the people who for so many years, in the face of compelling evidence, denied that cigarette smoking had serious adverse effects on health.

What has changed since I wrote my book is that not only is the evidence that our activities (primarily the production of energy) are causing serious harm even more convincing, but also that the scientists are increasingly pessimistic. It is now thought likely that by the end of the century global temperatures will have risen by an average of 7 degrees Fahrenheit and that the sea level will have risen by almost 2 feet. Besides inundation of low-lying land areas, desertification of tropical farms, and migration of tropical diseases north, global warming is expected to produce ever more violent weather patterns--typhoons, cyclones, floods, and so forth.

There is much uncertainty in climate science, and climate scientists concede that their predictions may be off--but they may be off in either direction. Far worse consequences are possible than those thought highly likely by the authors of the report, including a temperature increase of 12 rather than 7 degrees Fahrenheit, higher sea levels that could force the migration inland of tens of millions of people (or more), the deflection of the path of the Gulf Stream, causing Europe's climate to become Siberian, and abrupt, catastrophic sea-level rises due to the sliding of the Antarctic ice shelf into the ocean. Not only has the consensus among scientists concerning the harmful anthropogenic (human-caused) character of global warming grown, but the scientific consensus is increasingly pessimistic: recent evidence indicates that the global-warming problem is more serious than scientists thought just a few years ago.

My own view, argued in the book, is that the risk of abrupt global warming--a catastrophe that could strike us at any time, with unknown though presumably low probability--is sufficiently costly in expected-cost terms (that is, multiplying the cost of the catastrophe by its probability) to warrant taking costly measures today to reduce emissions of carbon dioxide and other greenhouse gases. Both the scientists and the policymakers, however, are mainly focused on the long-term costs of global warming--costs that will unfold over the remainder of this century. That focus makes the choice of the discount rate important, and potentially decisive.

A discount rate is an interest rate used to equate a future cost or value to a present cost or value. As a simple illustration (and ignoring complications such as risk aversion), if the interest rate is 5 percent, the present value of $1.05 to be received in a year is $1, because if you are given $1 today you can invest it and have $1.05 in a year. That is financial discounting. But discounting is important even when financial considerations are not the only ones involved in a choice. If you have a very strong preference for spending money now rather than a year from now, you might prefer $1 today to $1.50 a year from now.

These approaches don't work well when the question is how much we should spend today to avert costs that global warming will impose in the year 2107. Suppose we estimated that those costs would be $1 trillion. Then at a discount rate of 5 percent, the present-value equivalent of the costs is only $7.6 billion, for that is the amount that, invested at 5 percent, would grow to $1 trillion in 100 years. At 10 percent, the present value shrinks to $73 million.

So it is possible to argue that, rather than spending a substantial amount of money today to try to prevent losses from global warming in the future, we should be setting aside a modest amount of money every year--$73 million this year to deal with global warming in 2007, the same amount next year to deal with global warming in 2008, and so on. Of course we would also want to spend money to prevent the lesser losses from global warming that we anticipate in earlier years. For example, suppose we estimate that the loss in the year 2057 will be $100 billion. Then at the same 10 percent interest rate, we would want to spend $852 million this year.

Thus two effects are being balanced in computing the present equivalent of future losses from global warming--the larger loss in the more distant future, and the greater shrinkage of the larger loss, because of its remoteness from today, by the operation of discounting. The latter effect will often dominate, as in the examples, but of course this depends critically on the choice of discount rate. At an interest rate of 3 percent, a $1 trillion loss in 2007 has a present value not of $73 million or $7.6 billion, but of $52 billion. However, when either of the latter two figures is added to figures representing the present value of losses in intermediate years, the sum will be formidable.

A very high discount rate, implying that optimal current expenditures to avert the future consequences of global warming are slight, could be defended on the ground that the march of science is likely to deliver us from the consequences of global warming long before the end of the century. Clean fuels for automobiles as well as for electrical plants (where already there is a clean substitute for oil or coal--nuclear power, though it is more expensive) will be developed, or carbon dioxide emissions from electrical plants will be piped underground, or artificial bacteria will be developed that "eat" atmospheric carbon dioxide. These are not certainties but they are likely, and so they provide a good argument for using a high discount rate, such as 10 percent--and perhaps for considering no losses after 2107, on the theory that the problem of global warming is almost certain to be completely solved by then.

Nevertheless there are at least three arguments for incurring hefty current expenditures on trying to reduce carbon dioxide emissions in the near term. The first is that global warming is already imposing costs, and these will probably increase steadily in the years ahead. Discounting does not much affect those costs. They may well be great enough to warrant remedial action now.

The second argument for incurring heavy expenditures today to reduce global warming is that there is a small risk of abrupt, catastrophic global warming at any time, and a small risk of a huge catastrophe can compute as a very large expected cost. "Any time" could of course be well into the future, and so there is still a role for discounting, but it is minimized when the focus is on imminent dangers.

The third argument is that reducing our consumption of energy by a heavy energy tax would confer national security benefits by reducing our dependence on imported oil. Our costly involvement in the Middle East is due in significant part to our economic interest in maintaining the flow of oil from there. It is true that because our own oil is costly to extract, a heavy energy tax would not cause much if any substitution of domestic for foreign oil. But that is fine; our oil would remain in the ground, available for consumption if we decide to take measures abroad, such as withdrawing from Iraq, that might reduce our oil imports.

Heavy U.S. energy taxes would induce greater expenditures by industry on developing clean fuels and techniques for carbon sequestration; might persuade other big emitters like China and India to follow suit; and by reducing emissions of carbon dioxide slow the increase in the atmospheric concentration of the gas. Drastic reductions might actually reduce that concentration, because carbon dioxide does eventually leach out of the atmosphere, though at a slower rate than it is built up by emissions.

I cannot do justice to the 102 comments that my post evoked, but I will try to respond to the recurrent themes in them.

One comment suggests the use of the "minimax" (meaning, in this context, minimizing the maximum loss) decision rule to guide response to the risk of abrupt global warming, since no probability can be assigned to that risk. This raises the interesting and important question of how if at all to adapt cost-benefit analysis to risks that cannot be quantified. The problem with minimax is that it provides no definite guidance. I prefer, as argued in my book Catastrophe: Risk and Response (2004), to retain as much of the structure of cost-benefit analysis as possible in situations where the probability of a catastrophe cannot be quantified. In the case of abrupt global warming, this means trying to quantify the loss should such warming occur and the cost of averting the loss, and then see whether the probability implied by assuming that incurring the cost would be cost-justified is reasonable. So suppose a loss of $1 trillion could be averted at a cost of $100 billion. That cost would be worth incurring (ignoring a number of refinements and qualifications that would be necessary in a rigorous analysis) if the probability of the catastrophe were at least 10 percent. The question would then be whether 10 percent was in the probability ballpark--assuming we could estimate the ballpark, though not the location of the ball.

I agree with the suggestion that prizes are a good way of motivating research; recently Al Gore teamed with the British billionaire Richard Branson to offer a $25 prize for the development of a workable method of removing carbon dioxide from the atmosphere.

Most of the comments, as I expected, are in a state of denial about global warming. (Indeed, as I would have expected, one of them denies that cigarette smoking has been shown to have adverse health effects.) To many conservatives, global warming is a red flag.

The global warming skeptics point out that there are natural climate fluctuations, that anticapitalists are enthusiastic beaters of the drum for action against global warming, and that global warming would have good effects on agriculture in northern climes. These points are correct, but do not support the skeptical position. The existence of natural climate fluctuations increases the risk from human-caused global warming, because increased atmospheric concentrations of carbon dioxide increase the amplitude of the fluctuations. The fact that the motives of some of the people who are worried about global warming are political is irrelevant to the scientific issues, not only because scientists use apolitical methods of testing their hypotheses, but also because there are politics on both sides of the global warming debate: if leftwingers exaggerate the danger of global warming, rightwingers belittle them excessively. As for improving agricultural yields in northern climes, the transitional costs of relocating agriculture from (at present) tropical to arctic climes would be immense. Nor would improvements in agricultural yields respond to the effects of inundation of low-lying land areas, the migration of tropical diseases to temperate climates, the effects of increasingly violent weather, and the possible deflection of the Gulf Stream, causing Europe's climate to become Siberian.

It is also untrue that a 7 degree Fahrenheit increase in average global temperatures by the end of the century is a "worst case" prediction. That would imply a degree of certainty that we clearly do not have. And it is untrue that warming and cooling in millennia prior to the Industrial Revolution were unrelated to human activity. Substantial deforestation through burning, releasing large quantities of carbon dioxide into the atmosphere, began with the invention of agriculture some 8,000 years ago, and periods of reforestation (e.g., after the Black Death reduced the European population by a third) are correlated with global cooling. So at least the paleoclimatologist William Ruddiman argues, and I do not sense that the skeptics have read his work.

Some of the skeptics believe that Becker and I are part of a leftwing conspiracy to foist a false belief in global warming on the world. Anyone familiar with our work would know that we are conservatives. What is true and important is that there is considerable uncertainty about predictions of climate change. The climatologists' consensus may prove incorrect. What is striking however is the thinning of the ranks of the dissenters over time.

Many of the skeptical commenters appear to have visceral rather than a reasoned hostility to the idea that global warming is a problem that might require costly solutions. They and are not impartial readers of the scientific evidence. One commenter describes global warming as "another cult religion just like Marxism or Lysenkoism." But neither Marx nor Lysenko ever commanded a scientific consensus for their views. But I do agree with this otherwise rather intemperate commenter that Paul Ehrlich's The Population Bomb was total nonsense--and I have so said in my book* Public Intellectuals*.

One comment questions how heavy gasoline taxes could reduce our reliance on imported oil, since the cost of production of Middle Eastern oil is lower than that of oil produced in the United States. Depending on how stiff the taxes were, however, our total consumption of oil would fall, including consumption of foreign oil, though the mix would indeed shift (as I said in my post) toward imported oil. Another effect would therefore be to conserve our own oil--it would remain in the ground, available for future pumping, and a check on the behavior of foreign producers, such as threats by Iran to embargo oil. Our dependence on foreign oil would diminish in still another sense: the incomes of the foreign oil producers would fall, reducing those countries' geopolitical influence, including influence over us. We would be less dependent on their political whims.