TCS Daily

Holes in EPA's Ozone Policy

By Joel Schwartz - November 18, 2005 12:00 AM

No sooner has EPA implemented its tough new 8-hour ozone standard, than the agency is considering a substantial new clampdown on federal smog limits.[1] The Clinton administration concluded back in 1996 that trying to attain the 8-hour standard would cause net harm to Americans.[2] Bearing down still further on ozone would guarantee permanent employment for EPA's air pollution warriors, but at a high price for the American public, who will foot the bill for EPA's regulatory tunnel vision.

A new EPA Staff Paper on ozone claims that meeting the current 8-hour ozone standard of 0.085 parts per million (ppm) "would likely result in substantial reductions in exposures of concern and associated risks of serious health effects."[3] But EPA's staff plans to assess the benefits of going even lower -- down to 0.07 or 0.06 ppm, because "there is risk of moderate or greater lung function decrements in children, hospital admissions, and mortality from [ozone] resulting from exposures across the range of levels allowed by the current standard." The Staff Paper is a follow-up to EPA's Air Quality Criteria Document (AQCD) for ozone.[4] The AQCD is an in-depth analysis of health effects and related issues that the agency must conclude as a prelude to tightening pollution standards.

Despite its sober tone and measured language, the Staff Paper nevertheless substantially exaggerates the health risks of ozone at contemporary U.S. levels. If anything, what has become clear over the last several years is that ozone is having, at worst, a tiny effect on Americans' health and welfare. Ironically, it is EPA's own research staff, along with their counterparts at the California Air Resources Board (CARB), who have provided much of the analysis demonstrating the mildness of ozone's public health impacts.

For example, based on recent estimates by EPA scientists, going from ozone levels during 2002, which were by far the highest of the last several years, down to national attainment of the federal 8-hour standard -- typically about a 10 to 20% reduction for most non-attainment areas -- would reduce premature mortality by 0.06%, respiratory hospital admissions by 0.07% and asthma emergency room (ER) visits by 0.04%.[5] The improvement would be a few times greater if benefits continue to accrue when ozone is reduced from levels that already comply with the federal standard.[6]

Scientists at the California Air Resources Board (CARB) drew a similar conclusion. Based on CARB's estimates, reducing ozone by about 40 to 50% in California -- which essentially means eliminating all human-caused ozone -- would reduce premature mortality by 0.05%, respiratory hospital admissions by 0.23%, and asthma ER visits by 0.35%.[7] This is despite the fact that millions of Californians live in areas that have ozone levels far higher than any other part of the U.S.

In other words, large ozone reductions will at best result in tiny reductions in short-term health risks. The story is similar on long-term ozone exposure. CARB's Children's Health Study (CHS) followed nearly 1,800 children in 12 California communities with air pollution levels running from near background up to the highest levels in the country. After tracking the children from ages 10 to 18, researchers from the University of Southern California, who carried out the study, reported that there was no relationship between ozone and children's lung development.[8] In a separate CHS analysis that followed more than 3,500 children for up to 5 years, the USC researchers reported that children in high-ozone communities were 30% less likely to develop asthma, when compared with children in low-ozone communities.[9]

But the risks may be even smaller still. The reason is publication bias and a related problem known as data mining. Publication bias refers to the tendency of researchers to seek publication of and for journals to accept mainly those studies that find a statistically significant effect, while not publishing studies that don't find an effect.[10] Data mining refers to the risk that analyses of noisy, multivariate statistical data can become fishing expeditions that turn up chance correlations, rather than real causal relationships. As a recent review of air pollution epidemiology studies concluded:

        Estimation of very weak associations in the presence of measurement error and 
        strong confounding is inherently challenging. In this situation, prudent epidemiologists 
        should recognize that residual bias can dominate their results. Because the possible 
        mechanisms of action and their latencies are uncertain, the biologically correct 
        models are unknown. This model selection problem is exacerbated by the common 
        practice of screening multiple analyses and then selectively reporting only a few 
        important results.

A number of recent studies have assessed the effects of publication bias and data mining on estimates of air pollution and mortality. Taken together, they suggest that ozone is unlikely to be increasing the risk of premature death.[12] Perhaps not surprisingly, EPA has brushed off these concerns.[13]

The case of hormone replacement therapy (HRT) points up the risk of relying on epidemiological studies for drawing conclusions about small health risks. Based on epidemiological studies of HRT, researchers concluded that not being on HRT increases a woman's risk of heart disease by a factor of 2.[14] An influential meta-analysis of these studies, published in 1991, helped make HRT one of the most prescribed therapies in the United States. But more recently, randomized controlled trials, which eliminate the possibility of confounding by unobserved factors that affect health, showed that HRT doesn't reduce heart disease risk and might even increase risk.

With HRT, even a 100% increase in risk based on observational epidemiological studies turned out to be spurious once all confounding effects were genuinely eliminated through a randomized, controlled trial. The putative risks that air pollution studies are attempting to tease out are tiny by comparison -- at most a few percent increase in risk -- making their results even more precarious. A number of epidemiologists have suggested that epidemiological studies are inherently unreliable for assessing the existence of such small risks.[15]

Even with all the caveats on the health research, let's assume for the sake of argument that ozone is nevertheless really causing thousands of hospital visits and premature deaths each year. Of course we would all choose to prevent death and disease if we could. But attaining just the current 8-hour standard nationwide will cost at least tens of billions of dollars per year, and likely more than $100 billion per year.[16] We're talking about costs on the order of about $1,000 per year per American household to reduce total respiratory disease and premature death by at most a few tenths of a percent. Surely additional ozone reductions would have to be near the bottom of any rational list of priorities for improving Americans' health.

Even the EPA under President Clinton concluded that the costs of attaining the 8-hour standard would outweigh the benefits by a factor of 2, and that was based on the implausible assumption that attaining the standard would cost only $10 billion per year.[17] Nevertheless, EPA's air warriors are now considering a far tougher ozone standard.

As if the costs of reducing ozone weren't bad enough, it turns out that ground-level ozone actually has health benefits. Ozone up in the stratosphere protects us from solar ultra-violet (UV) light. Ozone at ground level adds a small increment of additional protection. Reducing this ozone shielding marginally increases people's risk of developing skin cancer and cataracts. EPA's ozone Criteria Document claims these effects are too uncertain to bother about:

        Within the uncertain context of presently available information on UV-B surface 
        fluxes, a risk assessment of UV-B-related health effects would need to factor 
        in human habits (e.g., daily activities, recreation, dress, and skin care) in order 
        to adequately estimate UV-B exposure levels. Little is known about the impact 
        of variability in these human factors on individual exposure to UV radiation. 
        Furthermore, detailed information does not exist regarding the relevant type 
        (e.g., peak or cumulative) and time period (e.g., childhood, lifetime, or current) 
        of exposure, wavelength dependency of biological responses, and interindividual 
        variability in UV resistance...In conclusion, the effect of changes in surface-level 
        O3 concentrations on UV-induced health outcomes cannot yet be critically assessed 
        within reasonable uncertainty.

It's almost laughable that EPA would claim uncertainty as a reason not to address a potential health risk. EPA normally uses uncertainty as the justification for more stringent regulatory limits. And EPA has for decades not only tolerated, but required the use of computer models with huge known errors and uncertainties as the driver for multi-billion-dollar regulatory and transportation infrastructure decisions.[19]

EPA's uncertainty claim is bogus in any case. Back in 1997 an internal EPA analysis concluded "any decrease in atmospheric ozone (tropospheric or stratospheric) causes ... an increase in the incidence of non-melanoma skin cancers. ...The methodology for estimating such increases (of both UV levels and skin cancer incidence) is well established."[20] In fact, the ozone reductions necessary to attain the current 8-hour ozone standard could easily result in a few thousand new cases of skin cancer and tens of thousands of cases of cataracts each year.[21] The much larger ozone reductions necessary to attain the more stringent standards EPA is now considering could double or treble this toll. The harm from lower ozone levels are sufficient to offset much or perhaps even all of the benefits from further ozone reductions, depending on how large those benefits really are. It seems safe to conclude that EPA has ignored the risks of marginal increases in UV exposure not because of a lack of scientific support, but because of the political and bureaucratic inconvenience of acknowledging the issue.

Why would EPA pursue policies that will make Americans worse off? Single-purpose regulatory agencies inevitably suffer from tunnel-vision -- the pursuit of an otherwise worthy goal to the point where it does more harm than good.[22] We have many needs and aspirations, but limited resources of money, time, knowledge, and attention. This forces us to make implicit and explicit tradeoffs every day based on our goals, tastes, circumstances, and financial means. But EPA's air regulators deal only in air-pollution reduction, and they place their mission ahead of other people's particular needs and desires.

EPA employees' jobs, budgets, and power also depend on having a serious problem to solve, and it is these same employees who get to decide when their job is finished. Despite already stringent standards and the small effects of current air pollution levels, EPA will pursue the next increment of air-pollution reduction, and the next, regardless of whether the increasingly marginal benefits are worth having or the costs worth bearing. By pursuing small health improvements at great cost, EPA continues to make Americans worse off overall.

[1] EPA adopted the 8-hour standard in 1996. However, it was not formally implemented until 2004 after legal challenges to the standard were finally resolved.

[2] EPA under President Clinton put the cost-benefit ratio at 2-to-1. As will be discussed below, neither independent analysts nor other federal agencies took even this poor ratio seriously, and instead concluded that the costs of attaining the 8-hour standard would be many times greater than EPA claimed.

[3] Environmental Protection Agency, Review of the National Ambient Air Quality Standards for Ozone: Policy Assessment of Scientific and Technical Information (Washington, DC: November 2005),

[4] Environmental Protection Agency, Air Quality Criteria for Ozone and Related Photochemical Oxidants (Second External Review Draft) Volumes I-III (Washington, DC: August 2005),

[5] EPA's scientists didn't put the results in terms of the percentage reduction from the baseline level. Instead, for each condition, they provided estimates of the number of events (e.g., hospital admissions) prevented by attaining the 8-hour standard. But they also provided estimates of the baseline rate of each event in the U.S. population. Given this underlying rate, you can calculate the percentage change due to ozone reductions. First, multiply the underlying rate of a given condition by the U.S. population, to get the total number of events for that condition. Then take EPA's estimate of the number of events prevented by reducing ozone and divide by the total number of events in the U.S. population to get the percentage reduction. B. J. Hubbell, A. Hallberg, D. R. McCubbin et al., "Health-Related Benefits of Attaining the 8-Hr Ozone Standard," Environmental Health Perspectives 113 (2005): 73-82. The paper is available free at

[6] This is because attaining the ozone standard requires ozone to be reduced below the level of the standard on the day with the fourth-highest ozone level of the year at the most polluted location in a given region. But this means that ozone will also be reduced on all days and all locations within the region. Assuming ozone causes harm at levels below the federal standard, then t is the reductions on these non-peak days at non-peak locations that would account for most of the benefits.

[7] Like EPA, CARB provided only absolute numbers of cases avoided, as well as underlying rates of each type of event in the general population. I calculated the percentage change from CARB's values. And just as with the EPA estimates, these numbers assume that ozone benefits accrue only until ozone is reduced down to 0.07 ppm. If benefits accrue below this level, then CARB staff estimated that the total benefits would be about 5 times greater. J. Schwartz, Rethinking the California Air Resources Board's Ozone Standards (Washington, DC: American Enterprise Institute, September 2005), California Air Resources Board, Review of the California Ambient Air Quality Standard for Ozone (Sacramento: March 2005),

[8] W. J. Gauderman, E. Avol, F. Gilliland et al., "The Effect of Air Pollution on Lung Development from 10 to 18 Years of Age," New England Journal of Medicine 351 (2004): 1057-67.

[9] This is based on 1-hour peak ozone levels. Asthma risk was 20% lower based on 8-hour ozone levels, though this result was only borderline statistically significant (relative risk of developing asthma was equal to 1.0 at the top of the 95% confidence interval). The study also reported that when the sample was limited to just the 8% of children who played three or more team sports, asthma incidence was 3.3 times greater in the high-ozone communities. News reports on the study, as well as the CARB/USC press release, mentioned only the results for these athletes, but omitted the overall lower asthma risk in the high-ozone areas.

It is worth noting that the results for children playing three or more team sports are based on a small sample -- 273 children total, of which 29 developed asthma. It is also worth noting that even if the ozone-asthma association in these athletes was causal, it is irrelevant for national ozone policy. No area outside California has ever had ozone levels anywhere near as high as the areas where the CHS was performed. In fact, by the time the study was published in 2002, ozone had already declined substantially in the areas where the study was performed.

EPA's ozone Staff Paper incorrectly states that asthma incidence was "no higher" in the high-ozone communities. In fact, it was substantially lower. EPA's staff can't take all the blame for this misleading claim. In their Lancet journal article reporting the study's results, the USC researchers stated "Risk of developing asthma was not greater overall in children living in the six high pollution communities than children living in the six low pollution communities" (my italics). R. McConnell, K. T. Berhane, F. Gilliland et al., "Asthma in Exercising Children Exposed to Ozone: A Cohort Study," Lancet 359 (2002): 386-91.

[10] Publication bias is a well-documented problem in a range of disciplines. See, for example, V. M. Montori, M. Smieja and G. H. Guyatt, "Publication Bias: A Brief Review for Clinicians," Mayo Clinic Proceedings 75 (2000): 1284-8; A. Thornton and P. Lee, "Publication Bias in Meta-Analysis: Its Causes and Consequences," Journal of Clinical Epidemiology 53 (2000): 207-16.

[11] T. Lumley and L. Sheppard, "Time Series Analyses of Air Pollution and Health: Straining at Gnats and Swallowing Camels?" Epidemiology 14 (2003): 13-4.

[12] M. L. Bell, F. Dominici and J. M. Samet, "A Meta-Analysis of Time-Series Studies of Ozone and Mortality with Comparison to the National Morbidity, Mortality, and Air Pollution Study," Epidemiology 16 (2005): 436-45; H. Anderson, R. Atkinson, J. Peacock et al., Meta-Analysis of Time-Series Studies and Panel Studies of Particulate Matter (PM) and Ozone (World Health Organization, 2004),; G. Koop and L. Tole, "Measuring the Health Effects of Air Pollution: To What Extent Can We Really Say That People Are Dying from Bad Air?" Journal of Environmental Economics and Management 47 (2004): 30-54.

[13] For additional concerns with the basis for EPA's air pollution standards, see S. H. Moolgavkar, "A Review and Critique of the EPA's Rationale for a Fine Particle Standard," Regulatory Toxicology and Pharmacology 42 (2005): 123-44. The paper addresses EPA's standard for fine particulates, but the critique addresses problems of analysis and process that are applicable to ozone as well. On ozone in particular, also see Schwartz, Rethinking the California Air Resources Board's Ozone Standards.

[14] This discussion of the implications of HRT studies for air pollution epidemiology is summarized from S. H. Moolgavkar, "A Review and Critique of the EPA's Rationale for a Fine Particle Standard," Regulatory Toxicology and Pharmacology 42 (2005): 123-44.

[15] G. Taubes, "Epidemiology Faces Its Limits," Science 269 (1995): 164-69.

[16] R. Lutter, "Head in the Clouds Decision-Making; S. E. Dudley, National Ambient Air Quality Standard for Ozone (Arlington, VA: Mercatus Center, George Mason University, March 12 1997),; R. Lutter, Is EPA's Ozone Standard Feasible? (Washington, DC: AEI-Brookings Joint Center for Regulatory Studies, December 1999), It isn't even clear that the 8-hour standard can be attained at all in some parts of the country. See, for example, S. Reynolds, C. L. Blanchard, and S. D. Ziman, "Understanding the Effectiveness of Precursor Reductions in Lowering 8-Hr Ozone Concentrations," Journal of the Air & Waste Management Association 53 (2003): 195-205; S. Reynolds, C. L. Blanchard, and S. D. Ziman, "Understanding The Effectiveness of Precursor Reductions in Lowering 8-Hr Ozone Concentrations--Part II. The Eastern United States," Journal of the Air & Waste Management Association 54 (2004):1452-70.

[17] Environmental Protection Agency, "Regulatory Impact Analyses for the Particulate Matter and Ozone National Ambient Air Quality Standards and Proposed Regional Haze Rule,"(Washington, DC: 1996), For a fascinating inside account of EPA's behavior during the development of the 8-hour standard, see R. Lutter, "Head in the Clouds Decision-Making: EPA's Air Quality Standards for Ozone," in Painting the White House Green: Rationalizing Environmental Policy inside the Executive Office of the President, eds. R. Lutter and J. F. Shogren (Washington, DC: Resources for the Future, 2004).

[18] Environmental Protection Agency, Air Quality Criteria for Ozone and Related Photochemical Oxidants (Second External Review Draft), p. 10-35.

[19] See, for example, J. Schwartz, Finding Better Ways to Achieve Cleaner Air (Washington, DC: AEI Environmental Policy Outlook, September-October 2004),

[20] EPA never made this analysis public, but it was anonymously placed in an OMB docket and is now available at

[21] See Lutter, "Head in the Clouds Decisionmaking;" Department of Energy, EPA Docket A-95-54, IV-D-2694, Appendix B-9 (Washington, DC: March 21, 1995); R. Lutter and H. Gruenspect, "Assessing Benefits of Ground Level Ozone: What Role for Science in Setting National Ambient Air Quality Standards?" Tulane Environmental Law Journal 15 (2001): 85-96. The United Nations Environment Program (UNEP) also had no trouble with the idea that decreasing ozone increases harm from UV exposure. UNEP estimated that each 1% decrease in total atmospheric ozone results in a 1 to 2% increase in human exposure to UV light, and thereby to increased cancer incidence. According to UNEP, each one percent increase in UV exposure would likely result in 10 to 20 new skin cancers per year per million people, or 3,000 to 6,000 new cancers in a population the size of the U.S. United Nations Environment Program, Environmental Effects of Ozone Depletion: 1998 Assessment (Nairobi, Kenya, 1998), The chapters of this report were also published as separate papers in the October 1998 issue of Journal of Photochemistry and Photobiology B, available at

[22] S. Breyer, Breaking the Vicious Circle: Toward Effective Risk Regulation (Cambridge, Mass.: Harvard University Press, 1993).


TCS Daily Archives