Air pollution abatement costs under the Clean Air Act: evidence from the PACE survey

Abstract

This paper uses establishment-level data from the US Census Bureau's Pollution Abatement Costs and Expenditures (PACE) survey to investigate the effects of the Clean Air Act and its amendments on the air pollution abatement (APA) capital expenditures and operating costs of manufacturing plants from 1979 to 1988. Results, based on some 90,000 observations, show that heavy emitters of the “criteria” air pollutants that were subject to more stringent regulation (due to county non-attainment of national ambient air quality standards) generally had higher APA expenditures, with estimates that imply hundreds of thousands of dollars of additional annual costs for the abatement of a specific pollutant for the average affected plant. Establishment characteristics, such as the size of the facility, appear to affect the intensity of this regulation and enforcement. While this study validates the PACE data to a certain extent, potential limitations are also revealed. The findings of this paper support those of a number of recent studies.

Introduction

The Clean Air Act (CAA) strictly prohibits the Environmental Protection Agency (EPA) from considering potential costs—in conjunction with benefits—when setting the national ambient air quality standards (NAAQS) that all areas of the country are required to meet.¹ Nonetheless, compliance costs are of significant interest,² and a compelling case has been made (e.g., [1]) for allowing cost-benefit analyses to be conducted here, as is required for other types of regulation. This paper examines some of the costs to manufacturing plants of the CAA (and its amendments) from 1979 to 1988, using establishment-level data on environmental expenditures from the Census Bureau's Pollution Abatement Costs and Expenditures (PACE) survey. The effects of air quality regulation, in terms of higher operating costs and additional capital expenditure for air pollution abatement (APA), are identified by using the variation in regulatory stringency, between counties that are in non-attainment versus attainment of federal air quality standards, for establishments that are and are not heavy emitters of the targeted pollutants. The analyses also control for location-specific effects as well as various establishment characteristics thought to be important determinants of plant-level environmental expenditures.

If no significant effects are found here, a number of possibilities might be implied. First, it might suggest that these regulations are not enforced or that it is relatively inexpensive to comply with them. This would counter mounting evidence of significant impacts of the CAA on manufacturing industries (e.g., [4], [12], [15], [17]). Second, it could mean that the most noteworthy regulatory expenses are the ones that are not explicitly covered by the PACE survey. For example, no attempt is made to capture items such as taxes, fines, permits, pollution offsets, and the like, let alone lost or sacrificed output. Furthermore, arguably the most affected segment of the population—plants under construction—is not canvassed by the survey at all. Finally, it has been suggested that the PACE survey fails to measure all that it was designed to measure (e.g., [5], [11], [16]). In particular, respondents may face difficulty in separating out their firms’ environmental costs from normal production costs and/or deciding on the relevant baseline with which to compare their expenditures. Given the frequent use of the published PACE data in proxies for regulatory stringency, knowing whether the underlying PACE responses actually correlate with precise measures of environmental regulation is of great significance. The analyses in this paper can be viewed as a test of whether these various claims can be refuted.

Results, based on some 90,000 observations, suggest that establishments subject to CAA regulation during this period generally did have higher APA capital expenditures and operating costs. Estimates imply hundreds of thousands of dollars of additional annual costs for the abatement of a specific pollutant for the average affected plant in this sample. In general, however, this reflects just a minority of the additional abatement costs incurred by such facilities. Analyses also show that establishment characteristics, such as plant size, can have an impact on the intensity of CAA-related regulation and enforcement. In the end, while many of the results suggest significant regulatory enforcement and compliance costs associated with county non-attainment status—corroborating the findings of a number of recent studies and validating the PACE data to a certain extent—there is also evidence that the PACE survey may indeed have limitations that prevent the full costs of regulation from being captured.

The paper proceeds as follows. In Section 2, I offer some background discussion, including an overview of air quality regulation in the United States, a review of the recent literature, and the motivation for this current study. This is followed by Section 3, which describes data sources and empirical specifications. Section 4 contains my empirical results, and Section 5 offers some concluding remarks.