Accounting for Costs and Cost Biases

Accounting for Costs and Cost Biases

Benjamin Zycher
Benjamin Zycher is vice president for the Research
Milken Institute for Job and Capital Formation, Sherman Oaks, California.




To the editor:

Geoffrey Rothwell asks, "Can Nuclear Power Compete?" (Regulation, Winter 1992), but two sets of analytic flaws prevent his article from answering that question.

Price-Anderson and the Coase Theorem. Rothwell adopts the common view that the Price-Anderson liability limit constitutes an implicit subsidy for the nuclear industry, in that the industry in the event of an accident would not have to pay for damage above the limit. That conventional view is incorrect. Consider a train that generates sparks, which in turn damage corn adjacent to the railroad right-of-way. If the crop damage is, say, $100, but the cost of spark suppression (either by installing equipment or discontinuing train service through the corn field) is $150, the railroad will continue service and corn destruction if transactions costs are low, regardless of the allocation of property rights, if the farmer has property rights in the land adjacent to the right-of-way, the railroad will offer a payment between $100 and $150 for the right to continue operations. If the railroad has the property rights, the farmer will offer no more than $100 to induce the railroad to avoid damaging the corn.

If the railroad can avoid $90 of the crop damage at a cost less than $90, it will do so (assuming that the farmer cannot avoid this damage even more cheaply), regardless of the allocation of property rights. If avoiding the remaining $10 of crop damage would cost the railroad more than $10, it will inflict that increment of damage, again regardless of the allocation of property rights. In the absence of transactions costs, the party that can avoid given damage most cheaply will be induced to do so, regardless of the allocation of property rights.

The allocation of property rights determines the direction of payment, that is, the distribution of wealth; but it is not a "subsidy" in that it does not affect the marginal cost of either railroad operations or corn production, because the cost of, say, a payment to the farmer is the same as the opportunity cost of a forgone payment from the farmer.

If transactions costs are significant, the allocation of rights does affect the allocation of resources. Coase demonstrated that the efficient allocation of rights is that yielding the allocation of resources that would obtain were transactions costs zero. Thus, the efficient liability rule is one that minimizes the sum of the costs caused by accidents and the costs of avoiding accidents. In our example, if it is cheap for the farmer to forgo crop production near the right-of-way, but expensive for the railroad to suppress its sparks, it would be efficient to give the railroad the right to use the adjacent land. That is not a "subsidy" from the farmer to the railroad; it is the efficient allocation of a scarce resource under conditions of significant transactions costs.

The prospective costs of potential nuclear accidents are identical analytically. Individuals and businesses near a nuclear generating station want to use the area for myriad purposes, while the nuclear plant would like to "use" it, in the event of an accident, to deposit radioactivity. Were transactions costs zero, the plant would find it cheaper to install safety equipment and to make other damage-limiting investments than to induce many or most of those living near the plant to move or otherwise to reduce the damage caused by future accidents. But surely it is relatively cheap for some people and businesses to take actions limiting future damage; at some margin, the incremental cost of damage avoidance by nearby residents falls below the incremental cost of damage avoidance by the plant. The plant would, therefore, pay some people and businesses to leave or to take other actions limiting future damages. That is particularly true for those who for whatever reasons would bear especially high costs in the event of an accident.

Since transactions costs in reality are high, the efficient liability rule is the one yielding that efficient allocation of resources. The Price-Anderson liability limit, by reducing expected damage payments, induces those who can limit damage cheaply, perhaps by moving, or those who would suffer disproportionately in the event of a serious accident to leave the area or take other actions limiting future damage. Thus, Price-Anderson minimizes the sum of accident and accident-avoidance costs. That is not to say that $7 billion is the correct limit. But the limit conceptually is an efficient policy and is not a "subsidy" any more than full liability would be a subsidy for those living or moving near the plant.

The basic problem with the Roth-well view is its confusion of analytic and normative issues; neither the railroad nor the nuclear plant alone is the "cause" of the problem. The general problem of externality results instead from the competition for the use of scarce resources, in this case, the area surrounding the nuclear plant. Were there no other human activity near the power plant, no externality problem would exist. Any externality problem is dual in nature, regardless of the allocation of property rights. The nuclear plant and its neighbors impose costs upon each other in the form of externalities, explicit payments for rights, and opportunity costs borne in thc pursuit of given activities.

The central issue is the most valuable use of the resource, and the liability limit is consistent with that end. Rothwell is confusing an assignment of property rights-which certainly affects the distribution of wealth-with a subsidy for particular activities. The $7 billion damage limitation in effect bestows upon the inhabitants of the area surrounding the nuclear plant a property right not to be damaged in the event of an accident up to $7 billion; and it gives to the plant a property right to impose damage above $7 billion. That is not a subsidy, although it is a transfer of wealth; it is an allocation of property rights that attempts to achieve the same allocation of resources that would obtain in the absence of transactions costs, that is, a minimization of the sum of accident and accident avoidance costs. A subsidy, properly defined, would change the allocation of resources from that prevailing in the absence of transactions costs. With significant transactions costs, the efficient allocation of rights yields the efficient allocation of resources; that is why it is not a "subsidy."

Analytic Asymmetries. Can anyone believe that regulated safety costs for nuclear generation stations are not far higher than those for coal-fired plants? Bernard Cohen's estimate of premature deaths caused by coal-fired generation is seventy-two per 1,000 megawatts per year, even with scrubbers and other mandated pollution equipment. That works out to about 38,000 premature deaths per year, most of which are among the elderly. Either nuclear plants are excessively safe or coal-fired plants are inordinately dangerous.

With respect to the waste issue, Rothwell claims that radioactive waste creates an "acute" negative externality, but the precise nature of that externality remains entirely obscure. The disposal of low-level waste-I percent of the radioactivity, but 99 percent of the volume-is trivial, and only the most shrill and dishonest of the antinuclear activists make an issue of it. The disposal of high-level waste, from a purely technical standpoint, is a non-problem, as the waste can be transformed and sealed into an inert glass and then buried in stable geologic formations. Cohen and Petr Beckmann estimate that radioactivity from such waste repositories at the earth's surface would be far below the natural background level and so would be immeasurable.

Those technical conditions ought to be compared with the huge volume of toxic sludge that is engendered by coal-fired generation each year but that is hardly mentioned as a cost of such generation. That is one manifestation of the Alice-in-Wonderland politics of nuclear waste, in which the citizenry is misled by politicians and the media about the dangers of nuclear waste and then is asked to accept repositories in their backyards without pecuniary compensation.

Regulation creates other serious biases, of which I mention briefly only two here. Relative construction periods for nuclear and coal-fired plants affect relative costs crucially, but are heavily a function of regulatory politics. After all, how is it that kilowatt hours from, say, Commonwealth Edison's nuclear plants- completed on schedule-are so much cheaper than those of most coal-fired generation? Moreover, the steady erosion of electricity rate regulation into a tax-transfer game has increased the riskiness of all base-load investment. While difficult to measure, it is at least plausible that the effect has been more pronounced with respect to nuclear investment precisely because of the political environment.

Rothwell's article simply does not examine the cost biases regulation inflicts on nuclear generation and so sheds little light on the issue of relative cost. That general problem is exacerbated by a series of narrower errors. Rothwell argues that cost-plus contracts be replaced with fixed-price contracts in plant construction, but nowhere does he demonstrate that the risk allocation inherent in the latter is superior to that of the former. He advocates standardized designs for plants but does not consider the ensuing effects over time for technical evolution and competition in plant engineering. He argues that an informational asymmetry exists between regulators and plant operators and advocates the use of incentive prices based on plant performance; but since the target performance must be negotiated, it is difficult to see how such a system would remove the informational problem. All in all, the Rothwell article is not a useful guide to the issue of relative cost.