How accurate are these estimates? A closer examination reveals that the benefit figures are highly dependent on a few assumptions and that the ranges presented are unlikely to reflect the true uncertainty surrounding them.
The benefits and costs of regulations, individually and in the aggregate, are notoriously hard to measure. There is no mechanism like the fiscal budget for keeping track of regulatory compliance spending by individuals and businesses, nor of the benefits that regulation brings. The OMB’s annual report probably offers one of the most comprehensive estimates available on the expected benefits and net benefits (benefits minus costs) of federal regulation, but as the agency acknowledges, it has limitations.
Federal agencies publish between 3,000 and 4,000 regulations each year. To keep its task manageable, the OMB confines its aggregate estimates to “economically significant” rules (those with estimated impacts of $100 million or more in a year) that are issued by executive branch agencies over the previous 10 years for which the agencies have estimated both costs and benefits. Thus the estimates exclude the effects of regulations issued by independent regulatory agencies such as the Securities and Exchange Commission, the Federal Communications Commission, and the new Consumer Financial Protection Bureau, as well as high-impact regulations for which agencies did not estimate either benefits or costs. In the most recent report, for example, the OMB bases benefits and costs for the last 10 years on agency estimates for 115 regulations—less than one-quarter of the 536 economically significant final regulations issued during that period and a fraction of the 3,203 significant regulations and almost 38,000 total regulations published since 2003.
The reported benefits and costs are based on ex ante estimates developed by the agencies themselves before the regulations went into effect. The OMB recognizes that “prospective estimates may contain erroneous assumptions, producing inaccurate predictions” and cautions that its “reliance on [agencies’] estimates should not necessarily be taken as an OMB endorsement of all the varied methodologies used by agencies to estimate benefits and costs.” Those caveats often get lost in public discourse and the aggregate estimates are widely reported, without qualification, as evidence of the net benefits of federal regulatory activity (and recently, of the Obama administration’s skill at identifying new opportunities for beneficial market intervention).
How the Benefits Stack Up
A look at how the OMB’s reported benefits break down is revealing. Figure 1 presents its upper-bound estimates of the benefits of regulations between January 21, 2001 and September 30, 2012. These data reflect “presidential years,” or regulations issued from January 21st in one year until January 20th of the following calendar year. The stacked column within each year distinguishes benefits attributed to three types of regulation:
- reductions in a single pollutant, fine particulate matter (PM2.5)
- regulations that yield private consumer savings
- all other actions
Since the first two categories of benefits comprise almost 80 percent of total reported benefits since 2001 (and over 90 percent of reported benefits in 2012), we examine them below.
Fine particulate matter | The majority of the OMB’s reported benefits derive from regulations that reduce PM2.5, which the Environmental Protection Agency predicts will reduce premature mortality. The EPA derives dollar benefits by multiplying an estimated reduction in premature deaths by a value per statistical life saved (VSL). In reporting these estimates in its annual report, the OMB recognizes the “significant uncertainty” associated with both “the reduction of premature deaths associated with reduction in particulate matter and … the monetary value of reducing mortality risk.”
The OMB identifies six key assumptions that contribute to this uncertainty in PM2.5 benefits estimates. One assumption is that “inhalation of fine particles is causally associated with premature death at concentrations near those experienced by most Americans on a daily basis.” The EPA bases this assumption on epidemiological evidence of an association between particulate matter concentrations and mortality; however, as all students are taught, correlation does not imply causation (cum hoc non propter hoc), and the agency cannot identify a biological mechanism that explains the observed correlation. Risk expert Louis Anthony Cox raises questions as to whether the correlation the EPA claims is real. His statistical analysis (published in the journal Risk Analysis) concludes with a greater than 95 percent probability that no association exists and that, instead, the EPA’s results are a product of its choice of models and selected data rather than a real, measured correlation.
Another key assumption on which the EPA’s (and therefore the OMB’s) benefit estimates hinge is that “the impact function for fine particles is approximately linear within the range of ambient concentrations under consideration, which includes concentrations below the National Ambient Air Quality Standard” (NAAQS). Both theory and data suggest that thresholds exist below which further reductions in exposure to PM2.5 do not yield changes in mortality response and that one should expect diminishing returns as exposures are reduced to lower and lower levels. However, the EPA assumes a linear concentration-response impact function that extends to concentrations below background levels. The OMB observes, “indeed, a significant portion of the benefits associated with more recent rules are from potential health benefits in regions that are in attainment with the fine particle standard.”
Based on its assumptions of a causal, linear, no-threshold relationship between PM2.5 exposure and premature mortality, the EPA quantifies a number of “statistical lives” that will be “saved” when concentrations of PM2.5 decline as a result of regulation. If any of those assumptions are false (in other words, if no association exists, if the relationship is not causal, or if the concentration-response relationship is not linear at low doses), the benefits of reducing PM2.5 would be less than estimated and perhaps even zero.
Further, as the OMB notes, “the value of mortality risk reduction is taken largely from studies of the willingness to accept risk in the labor market [where the relevant population is healthy and has a long remaining life expectancy] and might not necessarily apply to people in different stages of life or health status.” This caveat is particularly important in the case of PM2.5 because, as the EPA’s 2011 analysis reports, the median age of the beneficiaries of these regulations is around 80 years old, and the average extension in life expectancy attributable to lower PM2.5 levels is less than six months.
PM2.5 benefits also figure prominently in regulations whose purpose is not to reduce PM2.5. The EPA refers to these as “co-benefits” because they arise not directly from reducing the pollution targeted by the particular regulation, but from coincidental reductions in PM2.5. Figure 2 illustrates that in 2008, 2010, and 2012 in particular, co-benefits from PM2.5 reductions represent significant portions of total upper-bound benefits. (In 2008, the NAAQS for another criteria pollutant, ozone, derived over 70 percent of its benefits from reductions in PM2.5. In 2010, four regulations claimed 100 percent of their benefits from ancillary reductions in PM2.5. Three of those regulations targeted emissions of toxic air pollutants and the fourth established NAAQS for sulfur dioxide, another criteria pollutant. In 2012, 99 percent of the reported benefits from the EPA’s mercury and air toxics rule, discussed below, were co-benefits.)