On Assessing Risk

Risk analysis is a formal extension of the almost automatic calculation of risks that everyone performs when preparing, say, to cross a busy street. Assessing the risks of one action and comparing the risks of several make sense for the individual, and risk analysis makes sense for governments and private firms.

Indeed, the need to analyze and compare the risks of proposed policies and actions makes such good sense that too much faith now may be invested in risk analysis. It is a valuable analytical tool in making sensible policy decisions, but it is by no means infallible, nor even very precise.

Both analysts and policymakers need to recognize that a great many uncertainties plague the application of risk analysis. These uncertainties generally fall into three groups—incomplete or inaccurate raw data; ignorance about cause and effect; and value judgments, with value questions probably the most serious and least tractable.

Uncertain data

The simple-sounding task of determining the level of carcinogens produced by a new coal synfuels plant, for example, is not easy at all. For one thing, while it is difficult enough to measure certain pollutants at their source, it may be exceedingly tricky to do so at some later point in time and space. And even if it is known what combination of pollutant compounds are released at a power plant smokestack, for example, the mixture of compounds may turn out to be quite different after being carried long distances through the atmosphere. Finally, whatever is known about present pollution patterns offers no guarantees about what will happen in the future.

Consider Three Mile Island and the question of nuclear reactor accidents: we now realize we know much less about the probability of nuclear accidents than some had thought. And this brings up a point that needs emphasis—the "uninformed" public may not always be wrong. In some nuclear circles, the feeling long had been that the public is unaware of comparative risks and unable to think rationally about tradeoffs between energy technologies. There undoubtedly is some truth to this proposition, but one might also bear in mind that the public may have a healthy skepticism about the general state of scientific knowledge on the risks of energy technologies, and of nuclear power in particular.

Causes and effects

A pollutant released into the environment may be identified, and a general process such as strip mining observed, but linking both discrete and general probable causes with specific impacts on health and environment still can be more speculation than science. The problem in assessing the damage from Love Canal—if any—is a typical example.

Or, growing more grain to produce fuel alcohol probably would increase water pollution. This can be estimated and perhaps even measured for a particular piece of cropland, but what is the net impact on the downstream environment?

And what about low-level radiation from nuclear power, or extra low frequency radiation from microwave ovens and other devices? Controversies over these questions have to do not so much with the amounts of radiation emitted, but with what that radiation actually does to the human body. We are still a long way from understanding all of these phenomena.

Questions of value

Knowledge grows. More raw data on health and environmental risks will be gathered and connected with definite effects on the human body or on the ecosystem. Some emissions from coal may be shown to actually cause cancer in human beings; certain amounts of strip mining demonstrably may lead to a loss of species variety in a particular environment. However, even if this kind of knowledge increases, the third and probably worst uncertainty remains—the question of differences in values.

Suppose there are two different effects, one on health, and one on the nonhuman environment: which effect is more important? Or, confining the dilemma to human health alone, are fatal falls from roofs while installing solar collectors--surprisingly, a considerable health risk, as these things go—more important than chronic lung disease that may (or may not) be caused or aggravated by emissions from coal-fired power plants? Are chronic diseases like asthma less serious than acute diseases like pneumonia? They probably are, but, if so, how many asthma attacks are equivalent to one case of pneumonia?

Why is clean air important? Does the answer lie in better visibility or in better health? If the answer is health, then why do the Clean Air Act Amendments of 1977 tend to put more pollution where pollution exists now, rather than spreading it around more evenly?

The difficulties in deciding such questions are no secret; probably everyone working in the field has been guilty of giving them short shrift. Nevertheless, most analysts have given most importance to death rather than illness, even though one can note that death is inevitable but illness not always so. One can go even further, and worry about at what time the death or illnesses are caused.*

These questions involve important human values about which most people probably possess a greater depth of understanding than some analysts give them credit for. For example, there is the key comparison of catastrophes with routine events. How do 200 coal miner deaths, say, every year compare, with 200,000 possible deaths from some sort of nuclear power impact spread out over 1,000 years? Is there any difference between having a small chance of a large disaster or a large chance of a small disaster? How does the real acid rain problem, which though important, probably can be lived through, compare with the more hypothetical carbon dioxide accumulation, which could lead to serious worldwide environmental catastrophes?

Finally, there is the question of equity—or whose ox is gored. does it make any difference if an energy technology visits most of its harm on workers, as opposed to members of the general public? This is possibly the case for the synfuels industry, where there may be some dangers to workers of developing cancer from compounds produced in the process of turning coal to fluids, but relatively little danger to the general public. Is this type of arrangement "fair?" Similarly, city residents are more exposed to air pollution than people in the country. In either case, of course, one can say that there is some freedom to choose: presumably, one does not have to take a dangerous job, and one does not have to live in a city if one fears the effects of bad air. On the other hand, freedom of job and location choice, while great in America, certainly is not perfect.

Equity includes fairness to future generations. Not all scientists are persuaded that the nuclear waste problem is very difficult, let alone insolvable. Nevertheless, to the extent that nuclear wastes are hazardous, their impacts will fall not on us, but on inture generations. One may ask, of course, what have future generations ever done for us?—a question not quite as cynical as it may seem. Regardless, there seems to be within the human animal some feeling of responsibility for transmitting an environment as relatively clean as possible to those who follow us.

Common sense

Risk analysis is a useful tool, but the role of key uncertainties—in data, cause and effect, and in human value judgments—must be recognized if it is to gain respect as a serious element in making policy decisions. In particular, analysts should take into account that some attitudes of the public, such as preferring small routine risks to catastrophic nonroutine risks, may not be "irrational" but rather provide clues about how policy risk analysis should seek to incorporate human value preferences. In other words, much wisdom already resident in "common sense"—that uncommon quantity—must be combined with scientific advances in the field of quantitative evaluation of risks if risk analysis is to become a widely accepted and truly useful device for developing future policy.