![[logo] US EPA](https://webarchive.library.unt.edu/web/20081107095925im_/http://www.epa.gov/epafiles/images/logo_epaseal.gif){kind=logo}
Final Report: Determinants of Environmental Compliance: Plant, Firm, and Enforcement Factors
EPA Grant Number: R828824Title: Determinants of Environmental Compliance: Plant, Firm, and Enforcement Factors
Investigators: Gray, Wayne B. , Shadbegian, Ronald J.
Institution: Clark University
EPA Project Officer: Clark, Matthew
Project Period: May 1, 2001 through April 30, 2004 (Extended to September 30, 2004)
Project Amount: $276,883
RFA: Corporate Environmental Performance and the Effectiveness of Government Interventions (2000)
Research Category: Economics and Decision Sciences
Description:
Objective:The objective of this research project was to examine the determinants of environmental compliance with air and water pollution regulation for three industries: pulp and paper mills, oil refineries, and steel mills. The analysis incorporated data on the plant, its owning firm, and regulatory activity. We addressed four related questions:
(1) What makes plants differ in their compliance and their sensitivity to enforcement activity?
(2) How are compliance and emissions performance related, and how is compliance related across pollution media?
(3) Does enforcement’s effectiveness differ across states or between state and federal regulators?
(4) Do different statistical models give different results, comparing the determinants of compliance status, changes in compliance status, and duration of noncompliance?
Summary/Accomplishments (Outputs/Outcomes):The major initial task of the project was to organize an extensive database of plant-level information for the three industries being studied: paper, oil, and steel. The paper industry database was updated from an earlier version created with prior U.S. Environmental Protection Agency (EPA) support (Science To Achieve Results [STAR] Grant R286155). We used various industry directories to help identify plants in each industry, as well as to gather data on each plant. These directories identified the products produced by the plant, the plant’s overall production capacity, and some of the production technology in use at the plant. They also contained information on the ownership of plants; combining these data from directories in several years enables us to identify changes in ownership. If the owning company was publicly traded, we linked it to firm-level data from the Compustat database.
We then developed linkages from these data to a number of EPA databases, providing information on water pollution (PCS), air pollution (AIRS), and toxic releases (TRI), along with enforcement and compliance history (IDEA). The latitude and longitude of each plant were identified, enabling us to add demographic characteristics of the surrounding population, proximity to borders, and political activity to the database. We also prepared links from our plants to the Census Bureau’s Longitudinal Research Database; that work was done at the Census Bureau’s Boston Research Data Center because of the confidential restrictions on working with the plant-level Census data.
The initial paper for the project examined environmental compliance in the pulp and paper industry (Gray and Shadbegian, 2005). This analysis, using air pollution data from the 1980s, shows that plant characteristics (production technology, plant age, and plant size) are more important than firm characteristics (size, profitability, and industry focus) for determining compliance. Plants that are older, larger, or incorporate a pulping process are less likely to be in compliance than other plants. We also tested interactions between different regulatory areas, finding that plants that violate water pollution or Occupational Safety and Health Administration regulations are less likely to be in compliance on air pollution regulations.
We tested for the effectiveness of regulatory activity (inspections and other enforcement actions) in inducing compliance, finding some positive effects on compliance from both inspections and other enforcement actions. Finally, we tested if different plants were affected differently by regulatory activity. It appeared that the effects differed more by firm-level characteristics than plant-level ones: inspections were more effective in improving compliance at plants owned by smaller firms, whether firm size is measured by the number of paper mills operated by the firm or by total firm employment, whereas other actions were more effective in larger firms. This is consistent with a model that large firms are less likely to be surprised by regular inspections but are more concerned about negative publicity from other enforcement actions.
We tested similar models of air pollution compliance using data from the 1990s for plants in the paper, oil, and steel industries. These results are preliminary but suggest some differences across industries in their compliance behavior. As in the earlier data, plant characteristics play a greater role than firm characteristics in determining compliance, and plants with water pollution violations are less likely to be in compliance on air regulations. In the 1990s data, however, the impacts of regulatory activity seem to follow different patterns: oil and steel plants owned by larger firms seem to be more responsive to inspections and less responsive to other enforcement actions. We plan to explore this issue further in future work.
Our second paper (Gray and Shadbegian, 2004) examined the differences in regulatory activity and pollution emissions across pulp and paper mills. Our results indicate that both aggregate benefits and the demographic characteristics of neighboring populations (within 50 miles of the plant) affect emissions. In particular, plants with larger benefits to the overall population emit less air and water pollution, and those with more kids and elders nearby emit less air pollution, whereas plants located in poor neighborhoods emit more pollution. Plants whose pollution affects residents of other states emit more pollution, but these effects are reduced if the nearby states have more pro-environment Congressmen. Not every result fits those predicted by theory: the percentage of nonwhites near the plant, expected to reduce regulatory attention (assuming nonwhites have less political clout), turns out to be associated with lower emissions. Also, the overall results for our direct measures of regulatory activity, inspections, and enforcement actions, are less often significant than our results for emissions.
One key element of our research was the creation of detailed measures of the benefits from air and water pollution abatement at the plant level, dependent on models of the physical processes by which the pollution moves from the plant through the environment. The air pollution abatement benefit, based on the SLIM-3 Air Dispersion Model, was created during our prior EPA-funded research. We have begun experimenting with the measures of air pollution flows between U.S. counties provided by the EPA’s Source-Receptor Matrix, and anticipate using that model in future research. The water pollution benefit numbers were newly created for this research using the EPA National Water Pollution Control Assessment Model (NWPCAM) with the assistance of Tim Bondelid and George Van Houtven of Research Triangle Institute. We used NWPCAM to estimate the impact of an increase in a plant’s water pollution discharges on water quality downstream. These changes were evaluated using a continuous measure of water quality, rather than the four-step water quality ladder more commonly used. Our research demonstrates the value of a continuous quality measure, since it lets us measure the value of the small changes in water quality that result from changes in a single plant’s water pollution discharges.
This paper also addresses a key methodological issue for this research. A concern with our results is the potential for reverse causation or sorting: poor people could move into dirty neighborhoods because housing is cheaper there and/or families with sensitive individuals such as kids and elders might avoid dirty neighborhoods. If so, the pollution might ‘cause’ the neighborhood demographics, rather than vice versa. Some previous research used pre-siting demographics to control for this potential endogeneity. Because our sample of paper mills is quite old (most commencing operations before 1960) we could not use this approach. Instead, we use the demographic characteristics for people living between 50 and 100 miles from the plant as ‘spatially lagged’ instruments for the demographic characteristics nearer the plant. As long as the effects of a plant’s pollution decline with distance, this procedure should purge most of the endogeneity from the demographic variables, and it seemed to work well in this case.
Our third paper (Gray and Shadbegian, submitted, 2005) uses spatial econometric techniques to examine compliance and emissions behavior. To get sufficient numbers of nearby plants, we use data for all manufacturing plants within 50 miles of three cities near state boundaries (St. Louis, Cincinnati, and Charlotte). The analysis allows us to test if enforcement activity directed towards one plant has any impact on nearby plants, giving us a more precise measure of ‘general’ deterrence (as distinct from ‘specific’ deterrence, where activity directed toward one plant affects its own future behavior) and if these impacts cross state boundaries. We can also test if accounting for spatial relationships in compliance across plants affects the rest of the model.
Our results indicate a significant, but limited, role for spatial factors in modeling environmental performance. We find some evidence of positive spatial autocorrelation in the analysis of compliance status, but no such spatial effects are observed for toxic releases or emissions. Much of the explanatory power of the compliance models comes from plant-specific characteristics, with larger, older, more pollution-abatement-intensive plants and those in single-plant firms showing less compliance. Our measures of inspection activity tend to have the expected signs ¾ having more inspections at the plant, at nearby plants, and at plants in the same state is associated with greater compliance ¾ but these effects are not always significant. As expected, inspections at nearby plants in other states do not seem to increase compliance, thereby showing a significantly different effect from inspections at nearby plants in the same state.
In a fourth research paper, we examine emissions performance for pulp and paper mills, oil refineries, and steel mills using data for 1990-2000. The results indicate a substantial degree of heterogeneity across industries and pollutants, with five different emission measures tested. Plants with higher labor productivity tend to have lower emissions. Estimates of Seemingly Unrelated Regressions (SUR) models find a negative correlation between a plant’s productivity and its emissions and a positive correlation among the different emissions measures. Plants facing more enforcement actions have lower emissions (as expected), but the reverse is true of inspections. Plants in states with greater political support for regulation have lower emissions of those pollutants with more localized health effects (water and toxic pollution). We do not find much evidence, however, for consistent differences in environmental performance based on plant characteristics, unlike some of our earlier results.
The database created for this research has already been useful on other projects, including an National Science Foundation-funded research project on which Gray served as Principal Investigator, “Industrial Restructuring and Corporate Risk Management” (grant SBR-9809204). That research examined the relationship between corporate restructuring and the management of environmental, health, and safety risks in the pulp and paper industry. We carried out econometric analyses by adding measures of corporate restructuring to the plant-level paper industry dataset. We also conducted a survey of paper companies and did case study interviews at selected companies to gather information about the determinants of their environmental behavior, both regulatory compliance and pollution emissions. We concluded that regulatory compliance and other environmental functions have matured in most companies, with large firms having developed systems to handle basic compliance issues as part of their routine. Perhaps because of this, corporate restructuring seems to have had little impact on compliance or emissions at paper mills. These results were confirmed in all three analyses (econometric, survey, and case study).
This database will also be useful in our future research, for which we have recently received EPA funding (STAR Grant R832155, “Why Do Plants Comply with Environmental Regulations? The Importance of Enforcement Activity, Abatement Costs, and Community Pressure”). This research will continue our work on the determinants of compliance and emissions levels, combining EPA and Census data in a wide range of analyses of the determinants of environmental performance. This will allow us to make further progress in those areas of the current project that were not completed during the project period, most notably the testing of different statistical models of compliance. New elements in the research include extending the database to electric utilities, collecting data on community and political pressures, and conducting more detailed spatial econometric analyses, including an examination of spatial interactions across different pollution media.
Research Benefits
Our research results provide valuable information to environmental regulators targeting enforcement activity. Encouraging communication between regulators responsible for different pollution media could be worthwhile, as plants that are in compliance in one regulatory area tend to be in compliance in other areas (e.g., air and water pollution regulations). Regulatory activity directed at paper mills (inspections and other enforcement actions) seems to encourage better compliance with air pollution regulation. As inspections seem to be less effective and other enforcement actions more effective at paper mills owned by large firms, it may be efficient to consider the size of the firm when allocating regulatory activity across different activity types.
Spatial issues play a role in environmental performance. Plants located near each other tend to have similar compliance behavior. Inspections at one plant seem to improve compliance at nearby plants, though this effect does not carry across state borders. Borders also matter in our second paper. Plants where the benefits of pollution reductions accrue to out-of-state people (i.e., plants located near state borders) seem to have higher emissions levels. This effect seems to be lessened when the neighboring state is strongly pro-environmental politically, suggesting that those states put more effort into reducing cross-boundary pollution entering their state. These results suggest that federal regulatory attention (either direct inspections or heightened oversight of state regulators) may be needed to ensure reductions in emissions from plants near state borders.
Journal Articles on this Report: 4 Displayed | Download in RIS Format
| Other project views: | All 21 publications | 4 publications in selected types | All 4 journal articles |
| Type | Citation | ||
|---|---|---|---|
|
|
Gray WB, Shadbegian RJ. The environmental performance of polluting plants: a spatial econometric approach. Journal of Regional Science (submitted, 2005). |
R828824 (Final) |
not available |
|
|
Gray WB, Shadbegian RJ. When and why do plants comply? Paper mills in the 1980s. Law and Policy 2005;27(2):238-261. |
R828824 (Final) |
not available |
|
|
Gray WB, Shadbegian RJ. Optimal pollution abatement: whose benefits matter and how much? Journal of Environmental Economics and Management 2004;47(3):510-534. |
R828824 (2002) R828824 (Final) |
not available |
|
|
Shadbegian RJ, Gray WB. Assessing multi-dimensional performance: environmental and economic outcomes. Journal of Productivity Analysis 2006;26(3):213-234. |
R828824 (Final) R832155 (2006) |
|
regulatory impact, productivity, benefits analysis, petroleum refining, steel, pulp and paper industry, public policy, environmental justice, corporate performance, air and water pollution regulations, enforcement strategy, environmental compliance determinants, paper mills, statistical model,
,
Economic, Social, & Behavioral Science Research Program, Air, Sustainable Industry/Business, Scientific Discipline, RFA, Corporate Performance, Social Science, Ecological Risk Assessment, air toxics, Economics and Business, tropospheric ozone, ownership status, air & water pollution regulations, OSHA compliance, policy making, petroleum refining, paper mills, corporate environmental behavior, stratospheric ozone, enforcement strategy, statistical model
Progress and Final Reports:
2001 Progress Report
2002 Progress Report
Original Abstract