Safe Drinking Water

While the conference ranged over the scientific, economic, and administrative aspects of the public policy problems raised by our new knowledge in the drinking water area, this summary will concentrate on two subjects: first, the costs and benefits implied by existing policy as represented by the Safe Drinking Water Act (SDWA) and the implementing regulations being promulgated by the Environmental Protection Agency (EPA), and second, the problems attached to the present mechanisms for implementation and suggestions for change.

Costs and benefits. To a great extent, researchers disagree as to what the costs and benefits of SDWA regulations actually are. While some of this divergence of opinion may be due to misinterpretation of the figures, it seems likely that tile main reason is that we do not begin to know enough to estimate dollar benefits of the SDWA, nor, though the range of disagreement is much smaller, to measure costs. Why do we have such grave difficulties?

First, we have only incomplete knowledge of what potentially harmful contaminants are in drinking water. While our knowledge seems to be growing at a rapid pace, it also seems clear that we are going to go on discovering how much we did not know yesterday as we increase the sophistication of our analytical methods tomorrow. If we do not know what is present, we may not be attributing observed morbidity and mortality effects correctly, and, potentially as important, we may not be setting out to reduce the concentrations of the "right" substances.

Second, even given complete knowledge of contaminant levels, we would not have the ability to estimate the effects of chronic exposure to those substances. Consider only cancer rates, to single out the disease which dominated discussion at the conference. The difficulties in the way of obtaining the desired knowledge are immense. Manageable toxicological experiments require high doses to keep down the number of animals. This dosage leaves the knowledge gained open to objections concerning not only the appropriateness of the animal "models" but also the possibility that very high doses overload some natural defense mechanisms and are thus qualitatively different from the low, chronic doses actually at issue. But beyond that, we have a devilish problem deciding what to feed our animals, for if we take seriously the arguments about synergistic and antagonistic effects of chemicals, we are not faced with the "simple" task of testing 400 or 700 substances, but of exploring an essentially infinite space of combinations of those substances. Thus, as a practical matter, toxicological tests may never be able to tell us quantitatively what the effects of observed contamination of drinking water are, though, as William Lijinsky said, it may be that laboratory studies of toxicity are "the best we have now" as a basis for trying to avoid risks.

Epidemiology has its own problems when it comes to estimating effects. It has to work with the results of natural experiments which, while involving the desired human model and the actual chronic, low-dose contamination, also involve the inevitable noise associated with people who change locations over the crucial time periods, who eat tainted foods as well as drink contaminated water, and who change their habits over time; and with environments which also change, sometimes without recorded evidence. Thus, we turn in vain to epidemiologists for quantification of effects.

This is not to say that there is no agreement on the general outline of the problem. We do know that if we postpone action while waiting for more convincing evidence, we may be generating many times the currently observable cancer rates. This is because we are now seeing the effects of twenty or thirty years' exposure beginning in the late 1940s, before the production of synthetic organic chemicals began its tremendous growth. Since there has almost certainly been a continuing sharp increase in exposure to the products and residuals of this industry over these years, even if morbidity and mortality are only linearly related to exposure, we can expect nonlinear increases in these measures of effect. On the other hand, it simply does not seem possible to reject out of hand the possibility that the effects of doing nothing now would be quite small. Here we have the makings of a most difficult and confusing policy debate. The scientists to whom we are accustomed to turn for answers do not really have any. They give us impressions of fragmentary evidence in which descriptions tend to get mixed up with policy prescriptions that are based on personal judgments about the appropriate degree of risk aversion for society.

The debate also involves costs, of course, for if removal of trace contaminants were very cheap, we would be much less concerned about the elusive benefit measurements. The cost side is easier to pin down, or, more accurately, to peg within a range which is much narrower than the range of estimated benefits. We know that annual costs for a family of four implied by EPA's proposed organic contaminant control relations can vary enormously depending on what specific actions are required and what size plant is appropriate. A large system with a pure, up-land source, requiring only monitoring and perhaps a change to a new disinfectant, might only have to increase the annual bill of this family by less than one dollar. But a small, 1 million gallon per day (MGD), system with turbidity and organic problems might be increasing annual bills by $114 for the same average family. Estimates of the aggregate cost of SDWA are, therefore, difficult—requiring case-by-case estimation of actions required—and are of very little use in understanding how the act will appear to particular utility managers and customers. It is entirely conceivable, for example, that if we knew the aggregate benefits of the act, we would find that they exceeded the aggregate costs, but that a majority of water system customers were in fact incurring costs greater than their expected benefits.

The range of costs discussed above does not, however, cover all the relevant ground, for the EPA figures on which the remarks were based are not universally accepted as accurate. The reasons for this include differing engineering judgments on such items as the appropriate peak-to-average flow rate, required carbon contact time, and required regeneration frequencies for carbon. The differences also reflect guesses about future system growth.

Problems of implementation. The Safe Drinking Water Act establishes a national goal which is to be achieved through national standards and treatment requirements promulgated by a federal agency. But putting these standards into effect at the level of the water system and enforcing them over the long haul are supposed to be the responsibilities of the states and of the water suppliers themselves. Furthermore, the costs of meeting these national standards are to be borne by the customers of the utilities directly. A key instrument of enforcement is the notification requirement, whereby water suppliers who fail to meet the quality or monitoring requirements of the act must so inform their customers. There is also provision for state action to call attention to failures, if the local systems do not. This approach came in for repeated criticism during the conference, particularly, as is hardly surprising, from state and local officials.

The problem of carrying out the provisions of the SDWA is a complex one, involving attitudes, institutions, and incentives. First, there is a tradition of local control of the water supply with only minimal interference from the state. This creates a bias against federally imposed standards, however clear their justification. Second, as just discussed, justification is still a matter of some dispute and uncertainty among scientists, and is far from self-evident to the customer at the tap. Third, even were there an agreed level of benefits from the act, many customers of small or badly placed systems might oppose the local implementation of federal standards because their costs would be very high relative to benefits. Fourth, this requirement for increased spending on what might be called an environmental issue comes on top of a long period of publicity of other, similar increases, and at a time of spreading local tax revolt.

In addition, the Safe Drinking Water Act is only one room in what is by now a rather large and rambling edifice of environmental policy. In some cases there are inconsistencies of treatment and jurisdictional overlaps among the various agencies designated to administer various programs (see the discussion of toxic substances legislation in Resources No. 59, April—July 1978).

Several possibilities for changes in drinking water policies warrant attention. No single complaint was more often heard at the conference than that of federal government arrogance and inflexibility in the pursuit of what all recognized was a worthy goal. It seems clear that this particular sore can only continue to fester and poison other aspects of federal relations. There seems to be particular merit in state complaints about the SDWA and no strong arguments for federal direction, because no externalities of the sort connected with water and air pollution and pesticide application are involved. The effects of local drinking-water-quality decisions are felt locally, not 10 miles downstream or 100 miles along the prevailing wind track.

There are, in addition, very good reasons for expecting local judgments about the appropriate level of drinking water purity to differ across the country. The most straightforward of these is that treatment costs will vary enormously with system size. Even if everyone had the same (correct) understanding of the risks of exposure and valued those risks in the same way, one would expect that, in general, smaller communities would opt for less pure water than large ones. It is important to stress that there is nothing necessarily wrong about such an outcome. No one is forced to live in any particular community, and, if a person's judgment about drinking water quality differs from the local majority—and if this quality seems sufficiently important—there is no legal bar to a vote with the feet.

Federal intervention is now a fact of life in the drinking water field, however, and there seems to be no possibility that this situation will be reversed after a review of progress to date. What, then, can be done to mitigate the worst strains? One possibility is that the law and associated regulations could be changed to stress the guideline nature of the maximum contaminant level (MCL) and treatment specifications. These guidelines could be widely publicized, and then local communities could be given a chance to decide whether to adopt them or not. Those systems opting for the standards could be given primary responsibility for self-policing, subject to state (or probably federal) audit and random checks. Those systems opting for higher contaminant levels might be required to remind their customers of that decision on a regular basis, to give the customers periodic chances to change the decision, to send out easily read summaries of relevant evidence on costs and damages as it changes, and to monitor actual contaminant levels and be subject to the same audits and checks.

Another sticking point with the current law for water system managers is the notification requirement. No one seems to object to being responsible for rapid notification of emergency conditions, such as an upstream spill or an equipment malfunction. But there were complaints about being required to notify the public of a condition which had already occurred. Tests for certain problems may not be fast enough to allow notification before the "slug" of offending contaminant has passed through the system. There was some feeling that notices after the fact merely create distrust and confusion. Further, local managers would prefer to have all but emergency notices come from higher jurisdictions, state or federal. Clearly, one does not want to abandon an important and rather cleverly designed instrument for motivating compliance with the goals of the SDWA, but fine tuning might do much to reduce ill will.

A concluding comment. The policy issues and opportunities relevant to the Safe Drinking Water Act review are not only interesting in themselves. They are representative of important problems facing our society in many areas as it attempts to come to grips with the dangers it has created through efforts to control other problems. Making decisions in the face of extreme uncertainty is only the most obvious of the issues. Others include the "correct" balance between paternalistic concern and individual freedom to take risks; and the proper balance between federal, state, and local governments in these areas. It is important that the participants in the continuing policy debate keep in mind that much more than science (or engineering or economics) is involved and that the actions taken on this issue have implications for our collective life at least as important as the lengthening of average life expectancy by six days or six months.

Excerpt from Clifford S. Russell