Electric utilities now face the risk that existing assets, costs, or contract commitments may be "stranded" by increased competition, leaving shareholders rather than customers to bear the costs. Have shareholders already been compensated for this risk?

Some argue that shareholders have automatically been compensated for this risk by an allowed rate of return equal to the cost of equity capital determined in efficient capital markets.1 If so, forcing shareholders to bear stranded costs may seem fair. However, this argument does not stand up to scrutiny, because the cost of capital by definition does not include such compensation.

Loose Definitions Create

the Confusion

Standard regulatory practice is to equate the allowed rate of return to the cost of capital and to set the rest of the revenue requirement so that if costs and sales are as anticipated, the utility will earn the allowed rate of return exactly. Differences between actual and anticipated costs and sales mean that the rate of return the utility actually realizes typically differs from the allowed rate of return, but the allowed and actual rates of return are supposed to be equal on average.

For the "automatic compensation" theory to work, it must hold true within this standard framework. The essence of the automatic compensation theory is that the cost of capital in and of itself offers adequate compensation for the risk of stranded costs.

Most of the confusion arises from failure to pay close attention to what the cost of capital used to set the allowed rate of return really is. Many assume that because capital markets are aware of a given risk when setting the cost of capital, compensation for that risk must somehow be "in there" when the cost of capital is measured. Straightening out the confusion involves only a pair of definitions (for the "cost of capital" and "stranded costs") and the law of averages (see sidebar).

The "cost of capital" has a standard meaning in corporate finance theory, which may be specified in two parallel ways: (a) The expected rate of return prevailing in capital markets on alternative investments of equivalent risk; and (b) The discount rate for determining the net present value of future uncertain cash flows by discounting their expected value. The key term in both definitions is

"expected," which is used in the statistical sense: the probability- weighted average over all possible outcomes. Thus, the cost of capital is not the "due course" or "single most likely" rate of return investors anticipate (em it is the overall average.

Standard definitions of stranded costs are harder to come by, but a simple definition can capture the essence of the concept: Costs are "stranded" when investments made under cost-of-service regulation cannot expect to earn the cost of capital due to a transition to greater competition, because either 1) the investments themselves cannot earn a sufficient return, or 2) other costs or prior commitments cannot be recovered.

Again, the term "expect" operates in the statistical sense: If greater competition creates "stranded costs," then by definition the utility's probability-weighted average rate of return over all possible remaining outcomes (that is, given that costs are to be stranded) lies below the cost of capital.

Refuting the Automatic Compensation Theory

We now have all the ingredients needed to refute the automatic compensation theory. The reasoning is simple, given the proper definition of the cost of capital.

Under the automatic compensation theory, the utility will expect, on average, to earn the cost of capital (the allowed rate of return under traditional regulation) during the period prior to a transition to greater competition; if costs are to be stranded, the utility will expect to earn less.

From the law of averages, the average of these two expectations

(em (a) the rate of return expected prior to the transition (equal to the cost of capital), and (b) the rate of return expected after a transition with stranded costs (less than the cost of capital) (em must itself fall below the cost of capital. Since the utility cannot expect to earn the cost of capital averaged over periods when costs are and are not stranded, shareholders cannot have been automatically compensated for the risk of stranded costs by an allowed rate of return equal merely to the cost of capital.

The heart of the flaw in the automatic compensation theory is a failure to distinguish between 1) the cost of capital, the return investors require on average, and 2) whether the regulatory process affords investors an unbiased opportunity to actually earn the cost of capital on average. Put differently, setting the allowed rate of return equal to the cost of capital is compensatory only if that figure represents the rate of return the utility actually expects on average. With the risk of stranded costs and an allowed rate of return equal merely to the cost of capital (and regulation that otherwise is unbiased), this condition fails by definition.

A "junk" bond offers a good analogy. The bond's cost of capital is its expected rate of return (em the weighted average with and without default. However, the bond must promise a rate of return above its cost of capital if it does not default.

No Loopholes

Some additional points require discussion.2

First, the conclusion that the expected return (averaged over all time periods) is below the cost of capital holds regardless of the specific mechanism by which investors determine the cost of capital. In particular, it holds even if the cost of capital is determined in efficient capital markets that (by now, at least) have full knowledge of the risk of stranded costs. No market imperfections are necessary for our conclusion.

Second, the automatic compensation theory fails independently of whether the risk of stranded costs also increases the cost of capital itself. For example, increased competition ordinarily increases the sensitivity of a company's earnings to the business cycle. The result may be a bigger risk of stranded costs in bad times. That higher market-correlated risk should lead to a higher cost of capital. But an allowed rate of return equal to the new, higher cost of capital by itself remains insufficient to ensure that the utility has an unbiased opportunity to earn that new cost of capital. The only way to compensate shareholders for the risk of stranded costs is to permit the utility to expect to earn more than the cost of capital if no costs are stranded, whatever the cost of capital is and regardless of whether it has gone up.

Finally, stranded-cost compensation does not disappear in the ordinary "noise" of rate regulation. The difference between the cost of capital and an allowed rate of return high enough to provide compensation for stranded-cost risk can easily get quite large. For example, suppose that stranded-cost risk for a particular utility can be summarized as a 0.25 probability that a transition to more competition would result in a write-off that reduces the utility's actual return on equity by 30 percentage points in the year the transition occurs. Once this happens, if it does, the stranded-cost risk is resolved once and for all. Finally, suppose that investors are fully aware of the risk, and that the utility's cost of equity capital given this knowledge, and measured perfectly, is 12.5 percent (which, without loss of generality, we could suppose had moved up from, say, 11 percent as a result of the higher market-correlated risk). The issue is, what allowed rate of return provides adequate compensation for the stranded-cost risk?

So What Would it Take?

As the Figure illustrates, 12.5 per- cent (the cost of capital) clearly does not offset the risk. If the allowed rate of return equals only the cost of capital, given the 0.25 chance of stranding, the weighted average rate of return expected over both the "good" (12.5 percent) and "loss" (12.5 percent - 30 percent = -17.5 percent) outcomes is only 5 percent, well below the cost of capital. Adequate compensation instead requires an allowed rate of return of 20 percent. With a 20-percent rate of return in the "good" outcome and a -10-percent (20 percent - 30 percent = -10 percent) rate of return in the "loss" outcome, investors expect a 12.5-percent rate of return overall.

And so, have investors automatically been compensated for the risk of stranded costs when regulators equate the allowed rate of return to the cost of capital? The answer is "no," because stranded costs affect not only (potentially) the size of the cost of capital, but also (definitely) the odds that investors will earn the cost of capital. Stranded-cost risk compensation "just ain't in there." t

Lawrence Kolbe and William Tye are principals of Brattle/IRI, an economic and management consulting firm headquartered in Cambridge, MA, which was recently formed by the merger of The Brattle Group and Incentives Research Inc. Over the years, Messrs. Kolbe and Tye have contributed frequently to PUBLIC UTILITIES FORTNIGHTLY. Their last article, "Who Pays for Prudence Risks?," appeared Aug. 1, 1992, p. 13.

Articles found on this page are available to Internet subscribers only. For more information about obtaining a username and password, please call our Customer Service Department at 1-800-368-5001.