‘Optimal’ Pollution Abatement – Whose Benefits Matter, and How Much? Wayne B. Gray Ronald J. Shadbegian Working Paper Series Working Paper # 02-05 September, 2002 U.S. Environmental Protection Agency National Center for Environmental Economics 1200 Pennsylvania Avenue, NW (MC 1809) Washington, DC 20460 http://www.epa.gov/economics ‘Optimal’ Pollution Abatement – Whose Benefits Matter, and How Much? Wayne B. Gray Ronald J. Shadbegian Corresponding Author: Ronald J. Shadbegian Economics Department University of Massachusetts at Dartmouth and U.S. Environmental Protection Agency National Center for Environmental Economics 1200 Pennsylvania Avenue (1809T), NW Washington, DC 20460 [email protected] NCEE Working Paper Series Working Paper # 02-05 September 2002 DISCLAIMER The views expressed in this paper are those of the author(s) and do not necessarily represent those of the U.S. Environmental Protection Agency. In addition, although the research described in this paper may have been funded entirely or in part by the U.S. Environmental Protection Agency, it has not been subjected to the Agency's required peer and policy review. No official Agency endorsement should be inferred. ‘Optimal’ Pollution Abatement – Whose Benefits Matter, and How Much? Wayne B. Gray and Ronald J. Shadbegian ABSTRACT: In this paper we examine the allocation of environmental regulatory effort across U.S. pulp and paper mills, looking at measures of regulatory activity (inspections and enforcement actions) and levels of air and water pollution from those mills. We combine measures of the marginal benefits of air and water pollution abatement at each mill with measures of the characteristics of the people living near the mill. This allows for the possibility that some people may count less in the calculations of regulators (and polluters), either because they have less political clout or because they live in another jurisdiction. We perform the analyses using a plant-level panel data set with approximately 300 pulp and paper mills from 1985-1997. We find support for the importance of both the benefits from pollution abatement and political factors related to the people affected, particularly related to the amount of air and water pollution being emitted. The results suggest substantial differences in the weights assigned to different types of people. In some models the benefits received by out-of- state people seem to count only half as much as benefits received in-state, but their weight increases if the bordering state’s Congressmen are strongly pro-environment. A few of these variables are also associated with greater regulatory activity being directed towards the plant, although those results are less consistent with our hypotheses than the pollution results. One set of results was consistently contrary to expectations: plants with more nonwhites nearby emit less pollution. Some of our results might be due to endogenous sorting of people based on pollution, but an attempt to examine this using the local population turnover rate found evidence of sorting for only one of four pollutants. Subject Area: Air Pollution, Water Pollution, and Environmental Policy Key Words: 1) Environmental Regulation; 2) Environmental Justice ; 3) Pollution Abatement Benefits; 4) Transboundary Pollution Financial support for the research from the National Science Foundation (grant # SBR-9410059) and the Environmental Protection Agency (grant # R-828824-01-0) is gratefully acknowledged. We are also grateful to the many people in the paper industry who have been willing to share their knowledge with us. Thomas McMullen, Charles Griffith, George Van Houtven, and Tim Bondelid helped us work with the EPA datasets; thanks to Mahesh Podar and John Powers in the EPA Office of Water, sponsors of the NWPCAM model. We received helpful comments on earlier drafts of the paper from Eli Berman, Matt Kahn, Arik Levinson, Gilbert Metcalf, and seminar participants at the EPA’s National Center for Environmental Economics, the Public Choice Meetings, the Second World Congress of Environmental and Resource Economists, and the NBER Empirical Environmental Policy Research Conference. Capable research assistance was provided by Aleksandra Simic, Martha Grotpeter, and Bhramar Dey. Any remaining errors or omissions are the authors’. The views expressed herein are those of the authors and not necessarily those of EPA or NSF. 1. INTRODUCTION In this paper we examine the optimal allocation of environmental regulation across pulp and paper mills. The optimal allocation depends on the costs and benefits of pollution abatement at the plant, as seen by the regulator. The direct costs of pollution abatement at a particular plant are related to the plant’s age, size, and technology, while the benefits are related to the extent of the pollution being generated and the number of people affected. Past studies comparing benefits and costs have focused on fairly simple measures of abatement benefits. In this study we develop more sophisticated measures of air and water benefits from pollution abatement based on the SLIM-3 Air Dispersion Model and the Environmental Protection Agency’s (EPA) National Water Pollution Control Assessment Model (NWPCAM) respectively. We expect that regulators should impose stricter regulation on plants located in areas with greater benefits from pollution abatement. However, we also consider political factors that may influence the allocation of pollution abatement. The focus of our paper is on spatial differences across plants in the distribution of benefits from pollution abatement and in the characteristics of the population living nearby. Responding to some of these population measures may be socially optimal if certain population groups are more sensitive to pollution, but in many cases these measures suggest self-interested behavior by regulators seeking to maximize the political support for their actions. We perform our analyses using a plant-level panel data set on approximately 300 pulp and paper mills from 1985-1997. We find substantial supporting evidence for both benefits and population characteristics affecting environmental outcomes. Plants with 1 larger benefits to the overall population emit less air and water pollution, and those with more kids and elders nearby emit less air pollution. Plants located in poor neighborhoods get less regulatory attention and emit more pollution. Plants located near state boundaries emit more pollution, with these effects reduced if the nearby states have more pro-environment Congressmen. Not every result fits those predicted by theory: the percentage nonwhite near the plant, expected to reduce regulatory attention (assuming nonwhites have less political clout), is actually associated with lower emissions. The results for our measures of regulatory activity tend to be less often significant, and sometimes carry unexpected signs. One important caveat on our results is the cross-sectional nature of our demographic data. Some of the results could be explained as reverse causation or sorting: poor people move towards dirty neighborhoods because housing is cheaper there; families with sensitive individuals such as kids and elders avoid dirty neighborhoods. It is difficult for us to control for such endogeneity because most paper mills are very old, so we cannot include pre-siting demographics in the analysis. Our attempt to test for sorting (using the degree of population turnover near the plant) finds significant evidence in favor of sorting for only one of the four pollutants (particulates), while the two water pollutants find significant evidence against sorting. Some of the differences in results for different regulatory measures pose further research questions. There is a pattern of unexpected signs for regulatory actions, where factors associated with fewer regulatory actions are often associated with less, not more, pollution. We would have expected opposite signs on these coefficients, and do find opposite signs in some cases. Is this an artifact of the data, or does it represent a real 2 difference in the process by which regulatory activity is allocated in different situations? Similarly, we find different effects on air and water pollution of being near the Canadian border: do these reflect real differences across pollution media in the mechanisms for ensuring international cooperation on pollution control? The remainder of the paper is organized as follows. Section 2 provides a brief survey of the relevant literature. In section 3 we provide some background on the pulp and paper industry. Section 4 outlines our model of the regulator’s allocation of pollution abatement across plants. In section 5 we present our empirical methodology and a description of our data. Section 6 contains our results and finally we present some concluding remarks and possible extensions in section 7. 2. PREVIOUS STUDIES A few studies have addressed the issues raised above, providing empirical estimates of the impact of political boundaries, demographics, and political activism on exposure to pollution. For example, Helland and Whitford (2001), using annual county- level data from the Toxic Release Inventory (1987-1996), find that facilities located in counties on state borders (border counties) have systematically higher air and water pollution releases than facilities located in non-border counties: facilities in border counties emit 18 percent more air pollution and 10 percent more water pollution than facilities in non-border counties. Kahn (1999) also finds some evidence of a transboundary externality problem with particulates. Kreisel et al (1996) find that minorities are not disproportionately exposed to TRI emissions, but find some evidence that the poor are disproportionately exposed to TRI emissions. Arora and Cason (1999) 3 find evidence of racial injustice only in the south. In particular, Arora and Cason find that race is a significantly positive determinant of TRI releases in non-urban areas of the south. Hamilton (1993, 1995) examines whether exposure to environmental risk varies by demographics and political activism. Using data at the ‘zip-code neighborhood level,’ he relates the capacity expansion/contraction decisions of commercial hazardous waste facilities to race, income, education, and level of political activity (voter turnout), finding that capacity expansions are negatively correlated with voter turnout. Jenkins, Maguire, and Morgan (2002) show that minority communities receive lower ‘host’ fees for the siting of land fills while richer communities receive higher ‘host’ fees. Wolverton (2002) examines the issue of the location decision of ‘polluting’ plants. Previous studies indicate that ‘polluting’ plants tend to locate in poor and minority neighborhoods. However, Wolverton shows that once you consider the characteristics of the community at the time the plant is sited that contrary to popular opinion race no longer matters and that poor neighborhoods actually attract disproportionately less ‘polluting’ plants.1 3. PULP AND PAPER INDUSTRY BACKGROUND During the past 30 years environmental regulation has increased considerably both in terms of stringency and levels of enforcement. In the late 1960s environmental rules were primarily enacted at the state level, and were not vigorously enforced. Since the creation of the Environmental Protection Agency (EPA) in the early 1970’s the 1 Been (1994) and Been and Gupta (1997) find mixed results for environmental injustice in the citing of hazardous waste facilities when considering the characteristics of the neighborhood at the time of citing – 4 federal government has been the lead player in proposing and developing stricter regulations, and in encouraging greater emphasis on enforcement (much of which is still performed by state agencies, following federal guidelines). The expansion in environmental regulation has imposed large costs on traditional ‘smokestack’ industries, like the pulp and paper industry, which is one of the most impacted industries due to its sizable generation of both air and water pollution. The pulp and paper industry as a whole faces a high degree of environmental regulation. However, plants within the industry can face very different impacts from regulation, depending in part on the technology being used (pulp and integrated mills vs. non-integrated mills2), the plant's age, the plant’s location, and the level of regulatory effort directed at the plant. The most important determinant of the regulatory impact is whether or not the plant contains a pulping process. Pulp mills begin with raw wood (chips or entire trees) and use a variety of techniques to separate out the wood fibers, which are then used to produce paper. The most common form of pulping in the U.S. is the Kraft technique, which separates the wood into fibers using chemicals. A large number of plants also use mechanical pulping (giant grinders separating out the fibers), while still others use some combination of heat, other chemicals, and mechanical methods. Once the fibers are separated out, they can be bleached and combined with water to produce a slurry. After the pulping stage is complete, residual matter remains which historically was released directly into rivers (hence water pollution), but now must first be treated. The pulping process is energy intensive, so most pulp mills have their however, their results are based on data sets with only 4 and 10 observations respectively. 2 Integrated mills produce their own pulp and non-integrated mills purchase pulp or use recycled 5 own power plant, and thus are significant sources of air pollution. The pulping processes may also involve hazardous chemicals, such as the use of chlorine bleaching in Kraft pulp mills, which can create trace amounts of dioxin, raising the concern over toxic releases. The paper-making process is not nearly as pollution intensive as pulping. Non- integrated mills either purchase pulp from other mills or use recycled wastepaper. During the paper-making process, the slurry (more than 90% water at the beginning) is laid on a rapidly-moving wire mesh which progresses through a succession of dryers in order to remove the water, thereby creating a continuous sheet of paper. The energy required during this stage is less than during the pulping stage, but it can still cause air pollution concerns if the mill produces its own power. Finally, during the drying process some residual water pollution is created. However, both of these pollution concerns are much smaller than those created during the pulping process. The past 30 years has seen large reductions in pollution from the paper industry, with the advent of secondary wastewater treatment, electrostatic precipitators, and scrubbers. In addition to these end-of-pipe control technologies, some mills have altered their production process, more closely monitoring material flows to lower emissions. Overall these alterations have been much more prevalent at newer plants, which were at least partly designed with pollution controls in mind – some old pulp mills were deliberately built on top of the river, so that any spills or leaks could flow through holes in the floor for ‘easy disposal.’ These rigidities can be partially or completely offset by the tendency for most regulations to include grandfather clauses exempting existing wastepaper. 6 plants from the most stringent requirements – e.g. until recent standards limiting NOx emissions, most small existing boilers were exempt from air pollution regulations. 4. MODEL OF POLLUTION ABATEMENT REGULATION Why do profit-maximizing plants employ resources to abate pollution emissions? If pollution were a pure externality, with all the burden falling on those who live downwind or downstream, we would not expect to see any profit-maximizing plant spend money on pollution abatement. Some market-based mechanisms like consumer demand for ‘green’ products or managerial taste for ‘good citizenship’ may provide incentives for plants to abate pollution. However, we believe that the main motivation for controlling pollution emissions in the U.S. is government regulation of pollution, especially for the air and water pollutants being considered in this paper, so we model the amount of pollution abatement as being determined by regulators rather than by the polluting firms. One could instead employ other models, in which pressure from regulators is supplemented by pressure from customers and community groups, or in which the polluting firms are concerned about some groups of people but not others, affecting where pollution levels are greater. These alternative models could lead to analyses similar to those presented here (explaining why pollution levels from paper mills differ depending on which groups of people are affected by the pollution). We are also assuming that differences in regulatory pressures among U.S. paper mills are primarily determined at the state level, so we view the state as the relevant jurisdiction for political concerns. A socially optimal government regulator maximizes social welfare by increasing 7