Grantee Research Project Results
Final Report: Over-compliance in Point Source Water Pollution
EPA Grant Number: R827972Title: Over-compliance in Point Source Water Pollution
Investigators: Horowitz, John K.
Institution: University of Maryland - College Park
EPA Project Officer: Chung, Serena
Project Period: December 15, 1999 through December 15, 2000 (Extended to December 15, 2002)
Project Amount: $59,316
RFA: Decision-Making and Valuation for Environmental Policy (1999) RFA Text | Recipients Lists
Research Category: Environmental Justice
Objective:
The objectives of this research project examined the discharges of conventional water pollutants from major dischargers. There are low levels of discharges, on average, for these pollutants relative to the dischargers' National Pollutant Discharge Elimination System (NPDES) permits. The research attempted to explain and understand this apparent "overcompliance." We focused particularly on: (1) discharge variability and uncontrollability and the effects these have on the pollution decisions by plant operators; and (2) the relationship between discharges, violations, and the characteristics of downstream communities.
Summary/Accomplishments (Outputs/Outcomes):
This research project examined point-source water pollution discharges in the United States. We focused the analysis on monthly average biochemical oxygen demand (BOD), a common organic pollutant, concentrations in wastewater, on a plant-level basis, from 1992 to 1999, for sources regulated under the Clean Water Act. Most of the data are from municipal wastewater treatment plants. A smaller portion of the data is from the pulp-and-paper and food-processing industries.
Point-source water pollution can be best understood from Figure 1, which shows a typical pattern for a sewage treatment plant in Maryland. Monthly concentrations are depicted relative to their permitted value of 30 mg/L. This figure shows two patterns that are common across regions and industries:
1. Plant-level discharges show tremendous variability from month to month. This variability is due to weather, human error, and mechanical breakdown as well as a basic randomness in the biological process used to treat organic waste.
2. Concentrations are far below their permitted levels, often in the range of 9 mg/L. In other words, effluent from this plant is far cleaner than it needs to be. In Figure 1, the median ratio of reported discharges to permitted discharges is 0.3, well below the benchmark of 1. In our full data set, the median ratio over all plants is 0.32. There also are very few instances in which discharges exceed the permitted level; in the case of this representative plant, none. In the worst months of an 8-year period, this plant's concentration levels are still 20 percent below the permitted levels.
These phenomena are likely related. Regulators and plant managers claim that such low discharge levels are warranted by discharge variability. Plants are believed to pollute below their permitted levels, on average, to compensate for the possibility of an unexpectedly large discharge.
Figure 1. Compliance Ratios, 1992-1999, for a Representative Sewage Treatment Plant
Another factor mentioned frequently by plant managers is the public pressure they face for clean discharges. Kagan, et al. (2003), for example, quotes a manager of a pulp and paper plant as saying, "We spent a lot of money to achieve this [reduction in complaints]. The driver is to pacify the community."
These issues—compensation-for-variability and public pressure—lead to two research questions: (1) Does variability truly lead plants to reduce their average discharges, and if so, to what extent? and (2) Does public pressure lead to actual overcompliance?
We first examined the compensating-for-variability explanation. If this explanation is true, then plants with higher variability should pollute less on average. This is a straightforward, if underappreciated, implication.
To test this hypothesis, we constructed a measure of variablity for each plant in our data set, being careful to isolate other controllable factors that might be affecting discharges. If plants are attempting to compensate for discharge randomness, then the higher a plant’s variability, the lower its average discharges. This is an intuitively appealing test that has not been conducted before, for any pollutant, to our knowledge. We tested and rejected the hypothesis of "no-relationship" between the median and standard deviation of discharges. There is a statistically significant, negative relationship between the variability of discharges by the plant and the average discharges.
Furthermore, the estimated size of the effect is large. The plant in our data with the least variability is predicted to pollute on average at 57 percent of its permitted level. The plant with the greatest variability is predicted to pollute at only 10 percent of its permitted level.
We examined the question of whether plants are reducing discharges beyond what is required to compensate for variability; in other words, they are truly overcomplying with their permits, even when discharge variability is considered.
We predicted that even plants with very little variability pollute at about 60 percent of their permitted level. In other words, we estimated that plants are reducing their discharges by about 30 percent to compensate for variability, and another 40 percent due to some other factors. This latter quantity is our measure of overcompliance. We refer to this concept as "numeric overcompliance." It refers to compliance in terms of actual discharge levels, not the other forms of compliance such as having the correct permits and making reports in the proper manner.
The third research question attempted to determine which plants are reducing discharges so far below the permitted levels. We examined the effects of plant, regulator, and community characteristics on the estimated frequency of violation. This analysis is important, because the preceding results suggest that factors other than variability are affecting discharges.
To address this question, it is necessary to recognize that differences in discharge variability across plants means that plants' polluting behavior cannot be fully understood based only on analysis of average discharges. Therefore, to analyze compliance, we constructed a plant-level measure of the probability that the plant's discharges will exceed the permitted level. This measure considers both the plant's average discharge and its variability. We divided each plant's median compliance by its standard deviation, which can be converted into a predicted violation rate.
Our research shows that this measure is quite close to observed violation rates. The U.S. Environmental Protection Agency (EPA) assumes that discharges are randomly distributed log normal. We confirmed this assumption. (We divided the median log compliance by its standard deviation; under log normality, this expression is normally distributed and can be used in ordinary least squares regressions).
Using this measure, we related the predicted probability of a violation to the community characteristics. Our best estimate is that community characteristics have large effects on plant behavior, although the effects are measured imprecisely. Small plants in poorer communities tended to have higher probabilities of violation. Poorer, nonwhite communities have violation rates 2 to 22 percentage points higher than their richer counterparts. Wealthy communities are predicted to violate almost never; poorer communities are predicted to violate anywhere from once every 4 years to more than twice a year. These are economically and environmentally large differences.
This research also shows, along with our other results, that a relatively small number of plants are responsible for most of the violations that occur. The majority of plants are indeed remarkably clean. An important implication of our findings is that plants no longer exhibit a one-to-one relationship between discharges and permitted levels. This feature must be considered when predicting the effect of any change in regulation. A move to make regulation more stringent by reducing permitted concentrations would not necessarily lead to an equal reduction in discharges. We predict that a 1-percent decrease in the permitted BOD concentration will lead to between a 0.60- and 1.0-percent decrease in actual discharges.
The final section of the paper discusses the implications of these findings for three types of regulatory issues: (1) efficacy of current regulations and possible enforcement strategies; (2) implications for possible introducing of tradeable water pollution reduction credits; and (3) role of informal regulation.
Current Regulatory Environment
Because of discharge variability and community pressure, together and possibly separately, plants no longer exhibit a one-to-one relationship between discharges and permitted levels. This feature must be considered when predicting the effect of any change in the regulation. A move to make the regulations more stringent by reducing the permitted concentrations would not necessarily lead to an equal reduction in discharges. This research predicts that a 1-percent "tightening" of the regulation would lead to between a 0.6- and 1-percent decrease in average discharges.
The importance of discharge variability depends greatly on the U.S. Environmental Protection Agency's enforcement policy. A move to enforce standards over a shorter time (say, the daily limits) would likely reduce discharges further; a move to enforce standards over a longer time (say, annual average concentration or total annual quantity) would likely raise discharges. These predictions arise because of differences in the degree of randomness of these measures.
Tradeable Credits
Tradeable credits, a commonly proposed direction for pollution regulations, would likely, in the case of random water pollution discharges, substantially raise overall pollution levels. Currently, plants pollute well below their permitted levels on average, at least in part because of the risk of a random violation. If a fully tradeable, nonbankable pollution credit system were in place, plants could, in principle, discharge exactly at their permitted level. Whenever low discharges occurred, plants would sell the excess credits; when high discharges occurred, they would buy credits from other plants. As long as variability was uncorrelated across plants, they could pollute exactly at their permitted level on average and incur no net penalty. The difference between this and the current levels of discharges shows the potential consequences of allowing pollution trading. Of course, there are many other issues involved in the design of trading schemes for water pollution.
Informal Regulation
The large role that we found for community pressure also leads us to wonder whether the formal legal system of regulation could be modified or somehow "relaxed." There are two answers to this question. First, between 10 and 25 percent of the plants in our sample are not overcomplying, and some of them are rather seriously out of compliance. Community pressure is a feature primarily of well-to-do communities. Our evidence, preliminary though it is, does not suggest that community pressure would be a successful regulatory tool in poorer or minority communities.
Second, the interaction between community pressure and formal regulation is unknown. Do plant managers, in trying to convince the community (or themselves) that they are doing a good job, point to their permitted levels as a benchmark against which their success can be measured? If so, are there alternative nonregulatory benchmarks that could play this role? Answers to these questions would seem to be required before greater reliance on “informal” regulation is prescribed. It is worth noting that community pressure’s effects on local plants’ pollution decisions under tradeable permits systems also are unknown.
Variability creates an agency problem for engineers and plant managers; it makes it more difficult for them to show that they are doing a good job in properly balancing treatment costs with the penalty costs of violations. Excessively low probabilities of violation may be their optimal response in this situation. For regulators and the public, randomness means they must judge, from the highly variable discharge pattern, the abatement efforts being made by plants. Indeed, we believe that managers may not realize they are overcomplying to the degree that they are; their "true" compliance is masked by randomness, which makes inference difficult. The policy implications of this problem have not yet been fully explored.
Journal Articles:
No journal articles submitted with this report: View all 8 publications for this projectSupplemental Keywords:
biochemical oxygen demand, BOD, community pressure, informal regulation, economics, sewage treatment plants, water, chemicals, toxics, effluent, discharge, public policy, cost benefit, industry, clean technology, emission level., RFA, Economic, Social, & Behavioral Science Research Program, decision-making, Economics & Decision Making, emission levels, decision analysis, economic benefits, Clean Water Act, cost benefit, economic incentives, environmental values, non-regulatory benefits, cost/benefit analysis, environmental policy, effluent, compliance costs, legal and policy choices, public policy, regulations, cost effectiveness, over-compliance, point source waterProgress 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.