Discounting, EPA's nonroad spark-engine rule, and the hidden anti-regulatory agenda of cost-benefit analysis.

• Author: Corey, Jeffrey C.

I.

INTRODUCTION

In the past twenty years, it has become increasingly common for administrative agencies to use cost-benefit analysis to evaluate proposed regulatory programs. (1) As part of their cost-benefit analysis, agency policymakers typically assign a monetary value to the health benefits of a given regulatory program. This monetization of human health benefits allows agencies to assess a proposed regulation by comparing the theoretical monetary benefits of a new regulation with its projected costs. Scholars have criticized many aspects of this cost-benefit analysis, with perhaps the most attention given to the ethical implications associated with assigning a dollar value to a life saved by a regulation. (2) However, less attention has focused on a potentially more controversial component of cost-benefit analysis known as discounting. (3)

On a basic level, discounting is a way of assessing the present value of a cost or benefit that will take place in the future. Discounting is founded on two assumptions that, when mathematically factored into a cost-benefit analysis, reduce the value of both costs and benefits that occur in the future. The first assumption, generally referred to as the "time value" of money, (4) views money as worth less in the future than today because money can be invested to increase its value over time. In this sense, discounting is simply the application of a reverse interest rate, minimizing costs or benefits depending on how far they occur in the future. The second assumption justifies discounting based on people's preferences. (5) Discounting assumes that people prefer to have a benefit now instead of in the future and incorporates this preference into cost-benefit analysis by reducing the present-day valuation of a future benefit. Similarly, discounting assumes that people prefer to incur costs in the future rather than today, and therefore reduces the present-day valuation of a future cost.

The logic of discounting is deceptively simple. In the abstract, discounting may seem like a reasonable method for evaluating future benefits. However, when applied to a regulation that imposes present costs but does not provide benefits until many years into the future, discounting skews the analysis against regulation. (6) In other words, discounting acts as a default presumption in favor of small benefits today over much larger benefits far in the future. (7) The way in which discounting skews cost-benefit analysis is perhaps best seen in the field of environmental regulation in general, and, more specifically, in efforts to improve air quality. (8) Air pollution is largely created by the bedrock components of modern society: the internal combustion engine and industrial production processes. (9) Any effort to improve air quality by changing the internal combustion engine and industrial production processes will naturally require significant up-front expenditures such as research and development costs, new technology costs, and implementation costs. In contrast, the sheer size of the task of improving air quality means that widespread improvements in air quality may not occur for many years, perhaps many generations, into the future. Although cost-benefit analysis of air quality regulations are highly susceptible to being skewed against regulation, the U.S. Environmental Protection Agency (EPA) has continued to employ discounting in its three most recent proposals to improve air quality: the Clear Skies Initiative, (10) the Nonroad Diesel Engine Rule, (11) and the Nonroad Spark-Engine Rule. (12)

This paper focuses specifically on EPA's use of discounting in support of its Nonroad Spark-Engine Rule. I focus on this particular rule for two reasons. First, it provides a good example of how discounting can skew cost-benefit analysis against regulation, particularly when the regulation is aimed at a substantial problem where benefits may not be realized until many years into the future. (13) Second, it shows how discounting can play a supporting role in a cost-benefit analysis using several unorthodox valuation practices to produce a highly questionable comparison of costs and benefits. In examining this particular application of discounting in the context of a highly unusual cost-benefit analysis, this paper also asks why EPA would engage in such suspect valuation practices. In asking this question, I focus particularly on the relationship between EPA and the Office of Information and Regulatory Affairs (OIRA), an office within the Office of Management and Budget (OMB). As the office charged with reviewing proposed and final agency rules to ensure that agency administrators are keeping in line with executive branch policies, OIRA has substantial power to influence agency action. (14) In the case of the Nonroad Spark-Engine Rule, OIRA exerted that power in an extreme manner by demanding that EPA go back to the drawing board and engage in an alternative cost-benefit analysis that employed suspect discounting calculations, despite the fact that the agency's initial review of costs and benefits concluded that the regulation would more than pay for itself.

Before delving into the details of the cost-benefit analysis of the Nonroad Spark-Engine Rule and OIRA's influence on this analysis, Part II of this paper provides general information about discounting. It explains the basic mechanics of discounting, addresses some of the theoretical problems of discounting, and examines the government's official positions on discounting. Part III focuses on the cost-benefit analysis used to support EPA's Nonroad Spark-Engine Rule. My goal in Part III is to show how, in a real-word context, discounting can be used in conjunction with other suspect valuation practices to bias cost-benefit analysis against regulation. Finally, I use Part III's discussion of the Nonroad Spark-Engine Rule's cost-benefit analysis as a basis for my cautionary conclusion that both the practice of discounting and, more generally, cost-benefit analysis, are far less scientific than their mathematical formulas indicate. Instead, as shown in the case of the Nonroad Spark-Engine Rule, discounting can allow manipulation of any cost-benefit analysis to the point where the analysis is more of a policy statement against regulation than an even-handed consideration of benefits and costs.

II.

UNDERSTANDING DISCOUNTING

This Part provides background information on discounting by considering discounting from three different perspectives. Section A looks at discounting from a mathematical point of view by demonstrating how the basic discounting equation minimizes the value of costs or benefits that occur in the future. Section B looks at the theory underlying the discounting formula and highlights some of the logical flaws hiding behind any discounting calculation. Section C examines federal government policies on discounting. Finally, Section D provides a brief conclusion for this Part's overview of discounting.

1. The Mechanics of Discounting

   As noted above, discounting is founded on two premises: (1) the time value of money; and (2) people's preference to receive a benefit today rather than to wait for that benefit to occur at a later date. The following formula incorporates these two premises, allowing policymakers to mathematically reduce the value of future costs and benefits:

   "present value" = "actual value"/[(1 + discount rate).sup.years]

   In this discounting formula, the "present value" represents the discounted value of the future benefit. (15) Proponents of discounting have labeled this the present value because, in accordance with the theory that a benefit in the future is not as valuable as a benefit today, they believe that the present value is substantially less than the actual value. (16) The "actual value" is the thing being discounted. When evaluating life-saving regulation, policymakers can use the discounting formula to discount projected monetary benefits or costs of a proposed regulation. Alternatively, they can use the discounting formula to directly discount the number of lives saved by the proposed regulation. In other words, the formula can be used to discount any form of future benefit or cost, whether that benefit or cost is expressed in monetary terms or in some other unit, such as the number of lives saved.

   For illustrative purposes, suppose an agency is considering a regulation that is projected to prevent 200 acute asthma attacks in ten years and the agency wants to determine the monetary value of those future asthma attacks. If an emergency room visit to treat an acute asthma attack costs $300, then the present (i.e. discounted) value of those 200 asthma attacks would be calculated according to the following formula:

   "present value" = [(200 asthma attacks) ($300/attack)]/[(1 + discount rate).sup.10]

   To complete this calculation, the agency would have to choose a particular discount rate. This is the most critical step in the calculation because the discount rate determines how little or how much any future benefits will be discounted. (17) Even slight differences in discount rates can have a dramatic impact on the calculation of the present value of a future benefit. For example, if a 7% discount rate is used, the above calculation would result in a determination that 200 asthma attacks in ten years have a present value of about $30,000. In contrast, the same assessment using a 3% discount rate would value the same 200 asthma attacks at approximately $45,000. The difference between these two calculations becomes even more extreme as the time horizon is expanded. A policymaker calculating the present value of avoiding 200 asthma attacks in 100 years would find that those asthma attacks are worth just $69 if a 7% discount rate is used, whereas they are worth over $3000 if a 3% discount rate is used.

   Given this disparity, how is one supposed to determine whether a few hundred asthma attacks in the next...