Grantee Research Project Results
Final Report: Regulating Pollution Through Information Disclosure: Facility Response to the Toxics Release Inventory
EPA Grant Number: R829689Title: Regulating Pollution Through Information Disclosure: Facility Response to the Toxics Release Inventory
Investigators: Stavins, Robert N. , Miller, Nolan
Institution: Harvard University
EPA Project Officer: Hahn, Intaek
Project Period: July 1, 2002 through June 30, 2005 (Extended to June 30, 2006)
Project Amount: $396,235
RFA: Corporate Environmental Behavior: Examining the Effectiveness of Government Interventions and Voluntary Initiatives (2001) RFA Text | Recipients Lists
Research Category: Environmental Justice
Objective:
There is increasing interest in the United States and other countries in the potential role that can be played by information disclosure programs as substitutes for or complements to conventional command-and-control or market-based environmental policy instruments. Much of this interest can be attributed to the apparent success of the Toxics Release Inventory (TRI) program, which requires large manufacturing facilities to report publicly their annual releases of certain chemicals. Since the inception of the TRI program in 1986, reported releases of over 300 regulated chemicals have fallen by more than 45%.
The focus of research conducted under this grant is on estimating the causal effect of disclosure requirements on plant behavior. The research capitalizes on several different sources of variation to make these causal estimates, including both time-series variation and variation across states.
Summary/Accomplishments (Outputs/Outcomes):
We examined three specific research questions, summarized below.
What Effect Do TRI and Other Regulatory Programs Have on the Rate of Environmentally Friendly Technology Adoption?
We examined the effect of various regulatory regimes on technological change in the chlorine manufacturing industry. We focused on the diffusion of membrane cell technology, widely viewed as environmentally superior to both mercury cell and diaphragm cell technologies. The chlorine manufacturing industry has experienced a substantial shift over time towards the membrane technology, but only a relatively small fraction of this change has come about through adoption of cleaner technologies at existing plants. Most of the change has been due to the use of membrane cells by newly constructed facilities (entry) and the closing of facilities using diaphragm and mercury cells (exit). Therefore, in this context, technology diffusion can be viewed as a combination of adoption, entry, and exit decisions.
The research team examined econometrically two of these aspects of diffusion—adoption and exit behavior—to assess the effects of environmental regulation. Employing plant-level data on technology choice, economic variables, and regulatory variables from 1976 to 2001, we examined these adoption and exit decisions with a hazard model, considering the effects of both direct regulation of chlorine manufacturing and regulation of downstream users of chlorine.
We found that environmental regulation did affect the diffusion of the cleaner technology in the chlorine industry. However, it did so not by encouraging the adoption of membrane cells by existing facilities, but by reducing the demand for chlorine and hence encouraging the shutdown of facilities using the environmentally inferior options.
This research resulted in a paper titled “The Effects of Environmental Regulation on Technology Diffusion: The Case of Chlorine Manufacturing,” which was published in 2003 in The American Economic Review.
How Do Facilities Report to Mandatory Pollution Prevention Planning and Reporting Programs?
We examined this question both theoretically and empirically. To examine this question empirically, we utilized state-level variation in pollution prevention program design. During the 1990s, fourteen states adopted pollution prevention programs that required mandatory pollution prevention planning and public filing of pollution prevention progress reports. In addition, four states required annual public disclosure of toxics use in addition to toxics releases. These state-level programs provide a quasi-experiment of the effects of mandatory pollution prevention planning and public disclosure on environmental performance by firms. The data collected from the TRI provide natural outcome measures of the policies’ effects. In particular, we investigated whether facilities located in states with mandatory planning and reporting requirements had larger changes in total quantities of TRI chemical releases, reported for fewer chemicals, or engaged in more source reduction activities or more elaborate source reduction activities (as reported in Section 8 of the TRI). We control for other factors that might explain differences in release levels across states, including differences in environmental preferences of state residents, differences in other regulatory regimes, differences in plant- and firm-level characteristics, and temporal changes that are not related to the policy programs. This research strand resulted in two papers.
The first paper contains the results of the empirical analysis. This paper is titled: “Are Management-Based Regulations Effective: Evidence from State Pollution Prevention Programs.” This paper has been conditionally accepted for publication in the Journal of Policy Analysis and Management. Using panel data for just over 31,000 manufacturing plants in the United States, we investigated whether facilities subject to management-based regulations (MBR) had larger changes in total quantities of toxic chemical releases, engaged in more pollution prevention activities, or reported fewer toxic chemicals to the TRI. The analysis suggests that MBR has had a measurable positive effect on the environmental performance of manufacturing plants. In particular, plants subject to MBR experienced larger decreases in total pounds of toxic chemicals released and were more likely to engage in source reduction activities.
A more theoretical analysis of MBR is contained in the second paper, “Evaluating Management-Based Regulation: A Valuable Tool in The Regulatory Tool Box?” This paper appears as a chapter in Leveraging the Private Sector: Management-Based Strategies for Improving Environmental Performance, published by Resources for the Future Press in 2006 and edited by Cary Coglianese and Jennifer Nash. This paper focuses on two important sets of questions. First, does MBR actually work? Second, assuming MBR does work, under what circumstances does it work better than other regulatory alternatives? More specifically, when does MBR attain a given level of risk reduction at lower cost than other policy alternatives, where cost includes both the cost of compliance by the regulated entities and the cost of implementation and enforcement by regulators? The evidence for effectiveness draws on the results from the first paper. The unique contribution of this paper is an investigation of the circumstances under which MBR is likely to be more cost-effective than other regulatory instruments. We developed a model that suggests that MBR is likely to be more cost-effective when: (1) there is a large degree of heterogeneity in the population of regulated entities; (2) the risk being regulated is not easily measurable; (3) traditional regulations would impose high information burdens on the regulatory agency; and (4) there is some uncertainty with regard to the extent of the risk being regulated.
Do Facilities Strategically Respond to the Reporting Thresholds in TRI by Reducing Chemical Use to Levels Below Reporting Threshold? And How Does This Strategic Behavior Effect the Use of TRI Data?
Analysis of right-to-know programs has often focused on their effect on pollution levels. However, the nature of these programs may result in unanticipated strategic responses by regulated facilities that do not necessarily improve environmental quality. We were particularly interested in the possibility that facilities respond to the existence of “reporting thresholds,” levels of toxic chemical use below which facilities are not subject to regulation. Do facilities strategically reduce their chemical use to levels just below the reporting thresholds?
We used data from the Massachusetts Toxics Use Reduction Act (TURA) to evaluate these strategic responses. The TURA data provide two sources of variation that allow for the identification of bounds on the strategic bias introduced by regulatory thresholds. First, if a facility does not report for a chemical in year t for which it had reported in year t-1, the TURA reporting forms include an optional question that asks the facility to explain the change. Approximately one-third of facilities that cease reporting for a chemical answer the optional question. Responses to the optional question are used to gauge the degree to which facilities are not reporting for chemicals they use in positive quantities, and sensitivity analysis is done to determine how responsive the results are to changes in this imputation method.
The second source of variation in the TURA data that can be used to help estimate bounds on the strategic response bias is that facilities are required to report how much chemical they use in addition to how much they release. The federal TRI has reporting thresholds based on use, but only requires reporting on releases. The use data combined with the responses to the optional question about why facilities ceased reporting reveal that the distance from the reporting threshold is a good predictor of whether a facility ceases reporting because it went below the reporting threshold, but still used the chemical in positive quantities. We use this relationship to predict for non-responders the reason why they ceased reporting.
For facilities that directly reveal that they ceased reporting because they went below the reporting threshold, and for facilities that are predicted to have ceased reporting because they went below the reporting threshold, we then estimate the effect of these missing releases on trends in releases over time and on cross-sectional comparisons of facilities within a given year. Missing releases are estimated using three different procedures. The first is a lower bound estimate on total releases at the facility. The lower bound estimate assumes that when a facility ceases reporting its true releases are zero. This is the implicit assumption currently made by government agencies, non-profits, and academic researchers when aggregating releases data to the facility level (or higher levels). The second estimate of missing releases can be considered an upper bound estimate. In the upper bound scenario, if a facility ceases reporting a chemical because it went below the reporting threshold, releases are set equal to the most recent level of releases reported for that chemical at that facility. Moreover, the facility is assumed to continue releasing the chemical at the same level in perpetuity. The final estimate of missing releases is one that extrapolates the value of non-reported releases based upon linear trends in reported releases. These three scenarios present an upper and lower bound and an intermediate estimate of the degree of bias introduced by the reporting thresholds.
To see how these estimates differ, let us examine a hypothetical facility that reports for two chemicals in 1992 and only one of those chemicals in 1993. Furthermore, we either know from their answer to the optional questions, or we have predicted using the logit equation, that this facility stopped reporting because they went below the reporting threshold. The facility reported releases of 500 pounds of each of the two chemicals in 1992 and 495 pounds of chemical number one in 1993. What would happen to their total releases under each of our three estimation strategies?
Under the lower bound estimation, we assume the facility’s true releases of chemical number two in 1993 is equal to zero. So total releases from this facility fell from 1000 pounds in 1992 to 495 pounds in 1993. Under the upper bound estimation, we assume the facility’s true releases of chemical number two are equal to its last reported releases. The last reported release information for chemical number two was 500 pounds. So total releases in 1993 under the upper bound assumption is 495 + 500 = 995 pounds. The final estimate assumes that the facility would decrease its releases of the non-reported chemical at the same rate as observed in the whole population of facilities. On average, total releases fall by 5% each year. The projected estimate of total releases for this hypothetical facility in 1993 would then be equal to 495 + (500*0.95) = 495 + 475 = 970.
This research project resulted in the paper “Strategic Response to Regulatory Thresholds: Evidence from the Massachusetts Toxics Use Reduction Act,” which is currently in review for publication. The results of the analysis suggest that the bias introduced by the reporting thresholds may be significant. Up to 40 percent of the observed decline in TRI releases in Massachusetts may be attributed to truncation at the reporting thresholds. In addition, quartile rankings of facilities based on reported releases may be in error 45 percent of the time when behavior around the reporting thresholds is not taken into account. Because the regulatory thresholds for TURA are similar to many other state and federal “right-to-know” programs, including the widely used TRI, the fact that strategic response to regulatory thresholds in TURA are significant suggests that such strategic responses may affect validity of environmental right-to-know data more broadly.
Collectively, this research affects our understanding of how facilities respond to information disclosure and the relative effectiveness of information disclosure as a policy instrument. First, we found no evidence that chlorine manufacturing facilities respond to information disclosure requirements by choosing cleaner technologies. Facilities required to report to TRI were no more likely to adopt cleaner technology and were no more likely to shutdown if they used older, dirtier technology. However, we did find evidence in the broader manufacturing sector of some response to information disclosure programs. Facilities that are required to develop detailed management plans for reducing toxic chemicals and report their progress annually to the state were more likely to reduce pollution than facilities that were not subject to those requirements. The ability to benchmark facilities in an industry against one another is facilitated by these annual progress reports. We show that, theoretically, this can lead to the types of improvements that we find empirically. Finally, we show that facilities may respond to information disclosure requirements strategically by lower reported releases in ways that may not accurately reflect their risk profile. We show that facilities in Massachusetts stopped reporting for a significant number of chemicals, but continued to use these chemicals. The reported stopped because the facilities reduced their use of the chemical below the regulatory threshold for reporting. However, releases of these chemicals may have continued. This type of strategic response is common in other policy areas and we should not be surprised to find these strategic responses to environmental disclosure requirements as well.
Conclusions:
Policy Conclusion #1: Facilities Do Respond to Information Disclosure Requirements, but Some of These Responses May Be Strategic and May Confound Policy Conclusions. We focused on strategic responses to one feature of typical right-to-know disclosure regimes, such as the TRI, namely the existence of reporting thresholds. We showed that strategic responses to these reporting thresholds may have non-trivial impacts on the validity of TRI data. Specifically, the failure to report releases of chemicals used or produced below the reporting threshold can significantly impact inferences about changes in environmental performance over time and inferences about relative performance among some subset of facilities.
This finding is timely and policy relevant. The U.S. Environmental Protection Agency (EPA) is currently proposing to raise the reporting threshold for certain categories of facilities. Naturally there exist tradeoffs between the benefits of information disclosure and the costs. The costs involve significant time on the part of private facilities in monitoring, collecting, and reporting data to the TRI.
The benefits of disclosure may include providing the public information about environmental hazards so that they can improve their private decisions (e.g., where to buy a home); providing data to EPA and researchers to help understand changes in environmental performance (i.e., an environmental census); and serving as a regulatory instrument designed to put public pressure on plants to improve. Our work suggests that the loss of information from changing these reporting thresholds could be substantial and result in a significant decrease in the benefits of right-to-know programs. The proposed changes are likely to significantly affect our ability as researchers, governments, and citizens to assess changes in environmental performance over time or compare relative environmental performance among facilities.
Policy Conclusion #2: Non-traditional Regulation That Requires Facilities To Develop Detailed Risk Management Plans Can Be an Effective Policy Tool. We found both theoretical and empirical support for the use of MBR—regulations that require facilities to develop detailed management plans. State-level MBR of toxic chemical use in the 1990s did have an impact on toxic chemical releases. Facilities subject to these regulations reduced toxic chemical releases by 30% more than facilities that were not subject to such regulations. Regulated facilities were also much more likely to engage in pollution prevention activities.
Policy Conclusion #3: Non-traditional Regulation That Requires Facilities To Develop Detailed Risk Management Plan, Can Be Cost-Effective in Some Circumstances. MBR may be a relatively cost-effective means of attaining changes in environmental performance. Facilities that have high marginal costs of toxic chemical control will not find it in their interests to do more abatement, despite the additional planning effort. Facilities that have lower marginal costs are more likely to find it in their private interests to do more abatement. Thus, the abatement that does occur will occur at the lower marginal cost plants. That is what is desired for cost-effectiveness.
However, there are other policy instruments that can lead to cost-effective pollution reductions—in particular market-based instruments. Under what circumstances do we expect MBR to be a preferable regulatory alternative? The theoretical model indicates that MBR is more likely to be an effective risk regulatory tool when there is a strong connection between risk management effort and risk reductions. Policy makers should consider the likely degree of complementarity between planning and risk reductions when considering MBR as a regulatory tool. This further suggests that MBR is potentially most useful in situations where risk reductions themselves are difficult to measure but planning effort is easy to measure, and in situations where complementarity between management and reductions exists. These are the cases where direct regulation is hard and MBR is most likely to be effective.
Another potential use of MBR is when standards (including market-based instruments) are costly and the benefits of the standard are quite uncertain. In this context, MBR may provide a relatively inexpensive way of achieving some improvements while more information on the social benefits is obtained.
Policy Conclusion #4: Indirect Regulation—Regulation of Upstream or Downstream Industries—Can Have Significant Impacts on Environmental Performance. In the chlorine industry it appears that regulation of the downstream uses of chlorine, particularly in the pulp and paper and organic chemicals industries, is accelerating the closure of plants using the older technologies. This finding may be relatively specific to chlorine manufacturing. In particular, due to the hazardous nature of chlorine transportation, chlorine manufacturing tends to locate near users of the product. Thus changes in market forces in the chlorine using industries may have disproportionately large impacts on the survivability of particular chlorine plants.
However, the general idea that regulation of users of a product may be important to technological change seems intuitive and likely to apply broadly to other industries. Recently GM and Ford began requiring their suppliers to adopt environmental management systems that are certified to an international set of standards called ISO 14001. This is another example of downstream regulation, namely from the automobile industry, impacting technological change and environmental performance in the upstream industries (namely the auto parts suppliers). The cross-industry effects of regulation are potentially quite interesting and warrant further study, both by academics and policy makers.
Journal Articles on this Report : 2 Displayed | Download in RIS Format
Other project views: | All 16 publications | 3 publications in selected types | All 2 journal articles |
---|
Type | Citation | ||
---|---|---|---|
|
Bennear LS. Are management-based regulations effective? Evidence from state pollution prevention programs. Journal of Policy Analysis and Management 2007;26(2):327-348. |
R829689 (Final) |
Exit Exit |
|
Snyder LD, Miller NH, Stavins RN. The effects of environmental regulation on technology diffusion: the case of chlorine manufacturing. The American Economic Review 2003;93(2):431-435. |
R829689 (2002) R829689 (Final) |
Exit Exit |
Supplemental Keywords:
regulation, behavioral responses to regulation, public policy, State-level environmental policy, program evaluation,, RFA, Scientific Discipline, Sustainable Industry/Business, Corporate Performance, Economics and Business, Social Science, environmental performance, community involvement, corporate environmental policy, policy making, toxic release inventory, plant emissions, information dissemination, public reporting, community relations, right-to-know programs, environmental management strategies, green technologyProgress and Final Reports:
Original AbstractThe perspectives, information and conclusions conveyed in research project abstracts, progress reports, final reports, journal abstracts and journal publications convey the viewpoints of the principal investigator and may not represent the views and policies of ORD and EPA. Conclusions drawn by the principal investigators have not been reviewed by the Agency.