Many states impose regulations on fuel refineries to reduce emissions of volatile organic compounds (VOCs), though the policies vary considerably in both stringency and flexibility. How effective are these regulations at reducing pollution and at what cost? And do flexible regulations necessarily make sense?
Ozone is an odorless, colorless gas that has been linked to asthma, pneumonia, bronchitis, and damage to crops and natural vegetation. Even short-term increases in ambient ozone concentrations can cause a significant increase in deaths. Ozone is not emitted directly by any source; rather, it forms in the atmosphere through complex reactions involving two classes of chemicals—volatile organic compounds (VOCs) and oxides of nitrogen (NOx), both produced in part through human activity.
Since passage of the original Clean Air Act in 1963, U.S. state and federal governments have sought to reduce human exposure to ground-level ozone pollution by regulating emissions of ozone’s precursors. Our recent work (see Further Reading) focuses on a set of regulations that target emissions from motor vehicle use—a major source of VOCs and NOx—by restricting the chemical composition of gasoline. These regulations primarily target VOCs, which include a large number of chemicals with varying degrees of ozone-forming potential. Some VOCs are almost 80 times more reactive than others. Other major sources of VOCs include chemical solvents and natural emissions from plants, particularly deciduous trees. In rural areas, natural emissions may be substantially larger than emissions from anthropogenic sources.
Two basic types of policies have been designed: flexible performance standards that allow refiners to choose how to meet an overall VOC emissions limit, and stricter standards that precisely limit the concentrations of particular VOCs. The earliest federal regulation required refiners to reduce overall VOC emissions. During the first phase of implementation (1989–1991), the standards varied by place and month (because ozone is largely a summertime problem). Beginning in 1992, the most stringent of these requirements were imposed on all states.
In 1995, in areas that were still failing to meet national air quality standards, the regulations were replaced by “federal reformulated gasoline” rules that further tightened the VOC emissions limit. Arizona and California went further and adopted the California Air Resources Board (CARB) standards, which both impose tighter VOC emissions standards and limit the concentrations of the most reactive VOCs.
All these regulations have the side effect of increasing gasoline prices and price volatility by raising production costs and segmenting the U.S. gasoline market. Because some states and counties are regulated more tightly than others, the resulting patchwork of standards prevents gasoline transporters from arbitraging price differences across areas.
The cost to consumers is considerable. Leaving out California, gasoline consumption in the United States during the summer of 2008 was 47 billion gallons. Assuming a $0.01 to $0.015 per gallon price effect of the federal VOC standard (Brown et al. 2008), ozone regulations increased U.S. annual gasoline expenditures by an estimated $524 million to $784 million. The U.S. Congress, concerned about such price effects, inserted language into the Energy Policy Act of 2005 (section 1541[b]) that constrained the ability of the U.S. Environmental Protection Agency (EPA) to impose new gasoline standards that could further segment the U.S. gasoline market and exacerbate local price increases.
And yet, despite 40 years of costly emissions regulation, ambient air concentrations of ozone in many areas of the nation continue to exceed EPA standards. Are any of the regulations effective? The simultaneous presence of different regulations of differing stringency—combined with detailed data on ozone concentrations across the country—creates a natural experiment to answer that question.
We looked for changes in ambient ozone concentrations when and where the regulations were imposed. The nature of both the pollution and the policies conferred an advantage to the statistical analysis: gasoline regulation immediately involves all on-road vehicles and thus affects ground-level ozone concentrations right away. Thus, our analysis examined changes in ozone concentrations at each location immediately before and after the regulations took effect. We also compared changes in ozone concentrations in areas affected by the policies with changes in ozone in unaffected areas.
Comparing ozone concentrations in counties subject to the different phases of federal regulation, we found only slight improvements in air quality despite the increasing stringency of the requirements. The federal regulations limit the total evaporation of VOCs from gasoline, without focusing on the VOCs most likely to form ozone. We found that these regulations, on average, have no economically or statistically significant effect on ambient ozone concentrations.
This result is likely due to the rational response of refiners: they minimize the cost of producing fuel by removing the VOCs that are easiest to take out (particularly butane). These are less responsible for ozone formation than the more highly reactive VOCs (olefins and aromatic hydrocarbons), which are more expensive to remove. The overall VOC standard provides no compensating incentive or mandate for reducing concentrations of the highly reactive compounds.
California’s gasoline regulations, however, strictly limit the VOCs most important in forming ozone and thus prevent refiners from avoiding the costly abatement of these substances. As a result, California has enjoyed a large improvement in air quality. We see a substantial improvement in 1996, the year when CARB gasoline was introduced in all California counties. Our estimates suggest that CARB gas reduced ozone concentrations by 16 percent in California’s worst air-quality area, Los Angeles–San Diego. A back-of-the-envelope calculation indicates that just in terms of human mortality, the benefits of the regulation outweigh its costs.
Our simple cost–benefit analysis suggests that CARB should be required in southern California counties because ozone concentrations in this area are high and sensitive to changes in the emissions of reactive VOCs. CARB’s ozone benefits appear to be limited in other parts of the state. However, a comprehensive policy analysis of spatial gasoline regulation must allow for the possibility that county-by-county regulation may further segment the gasoline market, potentially increasing gasoline prices and volatility. Thus, a statewide policy may be optimal despite the fact that the regulation does not decrease ozone concentrations in all areas. In addition, by reducing emissions of toxic air pollutants, CARB may convey health benefits beyond those attributed to reductions in ozone formation. These benefits may be significant even where we do not observe substantial decreases in ozone levels.
The very different outcomes suggest a potential trade-off when setting the degree of flexibility in environmental regulations. More flexible regulatory approaches minimize the cost to the regulated firms, but for ozone, they lowered the environmental benefits. Cap-and-trade systems, which cap total emissions while allowing actual abatement to be traded across firms, may suffer from a similar problem. The intent is to minimize the cost of meeting the emissions standard by allowing firms with high-cost abatement to buy pollution permits from firms with lower costs. But if the marginal benefits of the emissions reductions differ across space because of the location of the exposed populations, permit-trading systems may not achieve the best welfare outcome. Fowlie (2010) finds that the benefits of one major cap-and-trade program were undercut because the pollution abatement was concentrated in areas with low marginal damage rather than in dense urban centers. She argues that less flexible regulation—perhaps weighting emissions by local marginal damages—would yield a better outcome.
In our study, the consequences of regulatory flexibility were severe. True, the abatement cost for the flexible standards was $0.01 to $0.015 cents per gallon, versus $0.08 to $0.11 cents for the CARB standard, but the flexible regulations appear to have had no measurable effect, while the California standards significantly improved air quality.
Anticipating firms’ probable response is critical, then, when designing flexibility mechanisms in regulation, even in command-and-control systems. Regulators should weigh the benefits from increased flexibility against the reductions in compliance costs to the regulated firms. Flexibility may be desirable, but not if it thwarts the goal of the regulation.
Maximilian Auffhammer is an associate professor in the Agricultural and Resource Economics Department at the University of California–Berkeley. Ryan Kellogg is an assistant professor in the Economics Department at the University of Michigan.