Reducing the Cost of Reducing Pollution

As we discussed last month, the efficient amount of pollution is not likely to please many. The problem is that everyone in an area has to consume the same amount of environmental quality while the value of that quality and the price paid for it vary from person to person. Some will want less than the efficient amount of pollution; some will want more; and almost no one will want the same amount.

The efficient amount is what everyone would want if there were no cost to negotiating and each individual faced payments that honestly reflected the value he receives from pollution reduction. Unfortunately, reaching agreement on and enforcing such a payment scheme is impossible when many people are involved. The result is that we cannot determine the efficient amount of pollution or reach agreement on the desirability of any amount of pollution. (See "Air Pollution Regulation Isn’t the Answer" for more.)

But a pollution policy requires a decision on how much pollution to reduce. And even if we cannot all agree on the desirable pollution level, we should all be able to agree on one thing: no matter what pollution level is decided on, we want to achieve it at least cost–at least sacrifice of other things we value. Equivalently, we can all agree that no matter how much cost we incur, we want to reduce as much pollution as possible.

What does it take to reduce pollution at least cost? More than it may seem. Clearly, it requires that every polluter reduce pollution as cheaply as possible. There are many ways to reduce pollution, and doing so in the least expensive way requires a lot of local information–information available only to those familiar with local conditions and circumstances–that is difficult, if not impossible, to communicate to others.

For example, there are many ways to reduce the emission of sulfur dioxide from an electric generating plant–substitute low-sulfur western coal for high-sulfur eastern coal, substitute natural gas for coal or petroleum, install a stack scrubber to filter out some of the sulfur dioxide, substitute more costly but nonpolluting pump-storage generation (see my March column) to serve peak-load demands, or shut down the plant. The costs associated with each of these actions vary among generating plants. A plant in Kansas may find it less costly to switch to western coal, while a plant in New Jersey is more likely to find it cheaper to install a stack scrubber. The cheapest way to reduce emissions by a small inefficient generating plant that is barely covering its costs may be to shut it down, something that would be very costly for a large efficient plant that is producing electricity worth far more than it costs to generate. Or the least-cost action may be some combination of approaches, such as relying more on pump-storage generation for peak load demand and substituting natural gas for coal. The possibilities are endless for every type of polluting activity, and the only hope for choosing the least-cost reduction in demand requires the use of information known only to those closely involved in each situation.

But even if every polluting firm were reducing its pollution at least cost, pollution would not necessarily be reduced at least cost. Having everyone reduce pollution as cheaply as possible is clearly necessary for least-cost pollution control, but it is not sufficient. We also need the right pattern of pollution reduction over all polluters. Some polluters can reduce pollution at a lower cost than others. Clearly, those who can reduce pollution at low cost have to reduce by more than those who can reduce only at high cost if we are to achieve the least?cost pattern of reduction. But to determine exactly what this least?cost pattern is, we have to consider the marginal costs of pollution reduction.

Equating at the Margin

It does not tell us much to say that one firm can reduce pollution at low cost and another at high cost. If this were true at all levels of reduction, then the first firm should reduce its pollution all the way to zero before the second begins any reduction at all. But as a firm reduces more of its pollution, the marginal cost of reduction will begin to increase (it makes sense to start with the pollution easiest to reduce and then move to that which is progressively more costly), and long before the low-cost firm has reduced its pollution to zero, its marginal cost of reduction will exceed the high-cost firm’s marginal cost of its first unit of reduction. Clearly, reducing another unit of pollution at minimum cost requires that it be reduced by the firm with the lowest marginal cost. But this increases that firm’s marginal cost, and soon additional reduction is more cheaply done by another firm. No matter what the level of pollution reduction, it is not occurring at least cost unless the marginal cost of reduction is the same for all firms.

For example, if the marginal cost of reducing pollution is $50 in one firm and $25 in another, then the first firm could increase pollution by one unit (saving $50) while the second firm reduces pollution by another unit (costing $25). This would result in the same amount of reduction at a saving of $25. This increase in pollution by the first firm and offsetting reduction by the second continues to reduce the cost of a given amount of pollution until the marginal cost of reduction is the same for both.

It obviously requires a lot of information to reduce pollution at least cost, and this information is widely dispersed. Each polluter knows more than anyone else about how to reduce his pollution as cheaply as possible. Even if this information could be communicated to a central authority, it would soon be rendered obsolete by changing circumstances. There is simply no way remote government authorities can acquire the knowledge necessary to dictate to firms how each should reduce pollution and how much each should reduce to protect the environment efficiently and effectively. Furthermore, even if policymakers had all the information necessary for reducing pollution at least cost, they would have little motivation to use it appropriately.

But the current centralized command-and-control approach assumes government can do those things. I’ll discuss this approach in my next column, emphasizing why it often does more to protect special interests than to protect the environment.