Why current practices of regulatory analysis undervalue future benefits.
Uncertainty about a possible harm is obviously relevant in deciding how much to regulate. More surprisingly, however, risks that are completely unrelated to the subject of the regulation can also be important in formulating regulatory policy. In particular, uncertainties about the future trajectory and distribution of economic growth have a substantial impact on the discount rate in cost-benefit analysis. This insight favors additional investment in the short-term in order to prevent long-term harms.
For those who are unaccustomed to the technicalities of cost-benefit analysis, discounting is a technique for taking time delays into account. For instance, if we build a dam today, it may continue producing benefits years in the future. The question is how to compare today’s costs with these future benefits. There are several reasons not to treat a dollar received in 2050 as equivalent to a dollar received today, even putting aside inflation. People generally prefer to receive benefits sooner rather than later, so the delay itself may be a reason to discount the future payment. Moreover, economic growth is likely to make America richer in future decades. An extra dollar of income doesn’t translate into as much improvement in welfare if society as a whole is richer. So it may be better to spend today’s dollar on some immediate benefit now, while society is poorer, rather than to receive the same dollar in benefits returned in the future when society will probably be better off anyway. The discount rate reflects the relative desirability of future versus immediate benefits, allowing us to reduce future benefits to their present value.
Although discounting issues are notoriously difficult and controversial, an increasingly prevalent view among economists supports the use of declining interest rates because of uncertainties about future economic growth. A similar argument supports declining rates when the distribution of growth is unequal or the future distribution is uncertain.
As I discuss in more detail in a recent paper, the rationale for the use of a declining discount rate applies to a broad range of regulatory decisions involving climate change, carcinogens, and infrastructure safety. Unfortunately, current U.S. practices within regulatory agencies tend to undervalue future regulatory benefits due to improper discounting. Regulations to protect environmental quality, public health, and safety may have additional value than current estimates suggest because they provide hedges against uncertainties and disparities involving future economic growth.
The argument is a bit complicated but an example may help. To simplify things, let’s ignore the part of the discount rate that simply stems from the human desire to get benefits sooner rather than later, and focus on the part that relates to future income growth. Suppose we expect half the population to experience income growth of 10% per year but anticipate the other half to see no changes in income. Now suppose the government is considering a regulation that will improve air quality, providing benefits to every person equal to $100 per year on average. At a 10% growth rate, incomes in the half of the population experiencing income growth will double in less than a decade because of compounding. After thirty years, these people will be making about ten times their current incomes, so someone who is currently making $50,000 a year will be making $500,000 per year. Therefore, for the half of the population experiencing annual growth, an extra $100 won’t make a big difference in their lives.
For the other half of the population, those who will experience no increase in income, the $100 per person will be worth just as much in thirty years or fifty years as it’s worth today. Thus, as time goes on, the value of the overall benefits of the hypothetical air quality regulation will be determined almost entirely by its value to the zero-growth part of the population. As the value to this group will stay constant over time, the appropriate discount rate to use would be zero percent. Therefore, over time, the discount rate we should use for society as a whole in this hypothetical would also tend toward zero. This is an artificial example, to be sure, but it turns out to be generally true—the discount rate will converge toward the rate of the group expecting the lowest growth in income. As such, using a declining discount rate would provide the most accurate estimate of the value of a regulation’s anticipated future benefits.
A similar argument applies in cases where the growth rate for society as a whole is uncertain. Instead of thinking about the issue in terms of separate high-growth and low-growth sub-populations in society, we can think in terms of high-growth and low-growth as two different possible economic trajectories for the future. From a mathematical point of view, the two situations are equivalent. As a result, while spending to solve a societal problem now may have negligible impacts if the society experiences a high level of economic growth in the future, such an allocation may have significant impacts if society were to turn out not to experience much economic growth at all. As such, policy analysts again would do well to use a declining discount rate when estimating the value of future regulatory benefits.
For those of us who are not specialists in environmental economics, it can sometimes be treacherous to evaluate the latest economic research. But there exists an impressive consensus among a broad group of economists about the need for declining discount rates. This group includes prominent economists with varying views about climate policy, along with economic theorists such as Nobel laureate Kenneth Arrow.
Unfortunately, most economic models of climate change still assume that the average global growth rate of the past few centuries will continue into the future. That’s certainly a reasonable projection, but we have no real way of knowing whether it will hold true. Our uncertainty about future growth should clearly affect the discount rate we use in cost-benefit analyses that support regulatory decision making.
Similarly, we can’t be sure whether future growth will be equally distributed. There may be part of the globe where growth will sputter and halt. Even within advanced countries such as the United States, recent experience shows that much of the population may suffer stagnant income growth while incomes skyrocket at the upper end.
It would be unrealistic to ignore the possibilities of either stagnant or disparate growth patterns when considering government policies with long-term benefits to the entire population. These benefits will be especially valuable in in the future if global economic growth turns out to be disappointing or if parts of the population are left behind as others progress.
This insight will not significantly affect the cost-benefit analysis of regulations that deal only with short-term issues. But it becomes extremely important when considering long-term issues like climate change. Today’s carbon dioxide emissions will stay in the atmosphere for two centuries or more, and even after emissions stop warming the earth, climate change will continue as a warmer ocean gradually releases heat to the land. So climate change is a very long-term problem, and the discount rate is a major factor in determining how much we should be willing to spend today to reduce carbon emissions now. It may seem like the ultimate in wonkish technicalities, but this adjustment to discount rates implies bolder regulatory action on climate change and other long-term problems in order to produce meaningful future benefits.