In the alphabet soup of regulatory acronyms, performance-based ratemaking (PBR) may help shape events well into the next century. At present, PBR is being implemented, or considered by, public utility commissions (PUCs) in over 20 states. By 2000, PBR is likely to reach most of the 50 states as well as the Federal Energy Regulatory Commission. The pressures of a global economy have raised the stakes. State regulators as well as politicians are increasingly aware that major industrial sectors such as electricity, gas, and telecommunications can play a key role in creating (em or destroying (em competitive advantage. This concern over economic competitiveness is fueling the PBR experiment.
Proponents tout PBR as an evolutionary reform of traditional regulation for so-called "natural monopoly" industries such as electricity and gas distribution. PBR is also being promoted as a useful transitional step or "bridge" toward complete deregulation of electricity generation.
The basic premise of PBR is that traditional, cost-plus regulation does not teach utility managers to minimize costs but rather to strategically conceal their firm's true minimum cost curve. Why?
Because a set of "perverse incentives" encourages managers to inflate the firm's operation and maintenance expenses, "goldplate" or overinvest in capital, avoid optimal risk-taking, and otherwise operate inefficiently.
Though touted as a solution to this problem, PBR presents three paradoxes:
1) The PBR regulator seeks to encourage a utility to operate at minimum cost, but does not know the firm's true cost structure.
2) The best way to promote maximum cost savings is to give all the savings to the utility, but this defeats the goal of reducing the cost of service to customers and thereby enhancing their economic competitiveness.
3) Any effective PBR system must prevent utilities from cutting quality to achieve false "cost savings"; however, such mechanisms are expensive and difficult to implement, and changes in quality (em though easy to measure (em are difficult to value in the market.
The policy question is whether PBR will be a positive force for change or, as one critic has put it, simply "a luxury trip to nowhere, at ratepayer expense." A poorly designed PBR system will exacerbate, rather than eliminate, regulatory inefficiencies. It may also destroy, rather than help create, competitive trade advantage.
Designing a PBR System
To design a system of incentives that encourages utility managers to pursue savings both in the short and longer run, the PBR regulator should set a "baseline" revenue requirement that will permit adjustments for inflation, productivity, and other factors over time. Next, the regulator should provide utility managers with a package of incentives that encourages them to produce at a cost below this baseline. Any sharing of cost savings between ratepayers and shareholders must preserve the manager's incentive to pursue such savings. Finally, the regulator should include a "quality control" mechanism to ensure that the utility does not pursue cost savings at the expense of system reliability, safety, customer satisfaction, and other measures of quality.
In taking these three steps, the PBR regulator faces several potential pitfalls. First, the baseline revenue requirement must not be set too high or too low. Second, the sharing mechanism must encourage the utility to pursue the maximum achievable cost-effective cost savings rather than simply allowing the utility to reap the lion's share of the most easily achieved cost savings and then stop. Third, there must be a reasonable link between the penalty system for quality deterioration and the reward system established by the sharing mechanism. In this regard, the PBR system must not include any additional alleged "incentives" that, in reality, may have little or no relation to the utility's strategic behavior.
Setting a Baseline
Setting the baseline correctly is absolutely critical. The regulator faces the same problems of gamesmanship, incomplete information, and cost revelation as it does under rate-base regulation. Specifically, the utility will attempt to inflate the initial baseline, build in a generous escalation factor, and minimize any offsets such as "productivity factors."
Ratepayer advocates will have a tendency to do just the opposite: Understate the baseline, minimize escalation, and boost the productivity factor. To cope with this strategic gaming problem, the PBR regulator has three options: 1) use the same method used in rate-base regulation, 2) use a statistical benchmark model approach, or 3) apply some combination of the two methods.
There is nothing a priori to suggest that PBR will set a more bloated baseline than rate-base regulation. Nevertheless, statistical benchmark modeling can help the regulator reach a truer approximation of a firm's minimum cost curve.
The basic unit of measure in traditional rate-base regulation is the firm itself or, perhaps, a small cluster of firms operating within the regulatory jurisdiction. In contrast, statistical benchmark modeling examines the price and cost structure of a much wider sample of utilities. It then normalizes this data by adjusting for geography, weather, fuel mix, and other operating conditions and characteristics. A well-executed statistical benchmark modeling procedure can thus be used in a PBR proceeding both as an independent check on the traditional, firm-specific method of determining the baseline revenue requirement as well as a guidepost to the target ending point of the PBR experiment. If PBR works, this ending point should be the firm's true minimum average cost.
Choosing a Cost-Sharing Method
The major purpose of PBR is to encourage utilities to voluntarily undertake cost savings and then distribute a portion of those savings to ratepayers. If we assume that the profit-maximizing utility will pursue cost savings up to the point they no longer pay, then equilibrium in this PBR "market for cost savings" will occur at the intersection of supply (which reflects marginal cost) and demand (which reflects the utility's marginal revenue).
In a world of increasing costs, each additional increment of savings will be more expensive to achieve than the last. This means that the best cost-sharing mechanism will be "progressive" rather than "regressive." That is, the utility's share should increase, not decrease, with the amount of cost savings achieved.
In a world of perfect information, the supply curve for cost savings will be known, and this "progressive" mechanism will smoothly and continuously track up the supply curve. Maximum savings will be achieved along with maximum rate reductions for customers, thereby enhancing economic competitiveness.
The obvious problem, however, is that in a world of incomplete information, the precise savings cost curve is unknown. Therefore, a multi-tiered progressive mechanism may be a more practical compromise. The figure on the previous page illustrates the key differences between such a progressive sharing mechanism and an alternative regressive sharing mechanism in a world of either scarce or abundant cost savings.
Supply curve "A" represents a world in which cost savings are relatively scarce and expensive to achieve. Supply curve "B" represents a world in which cost savings are relatively abundant and inexpensive to achieve. (Note that both curves slope upward, reflecting the principle of increasing costs.)
The "ascending staircase" in green represents a multi-tiered progressive sharing mechanism in which the utility receives 25 percent of the first 100 basis points, 50 percent of the next 100 basis points, 75 percent of the next 100 basis points, and 100 percent thereafter. The solid portions of the staircase trace the utility's share of savings. Equilibrium in the abundant cost savings world occurs at point D. Maximum cost savings are achieved, and ratepayers receive a substantial portion of the savings.
In contrast, the "descending staircase" in gray represents a regressive sharing mechanism in which the utility receives 100 percent of the first 100 basis points, 75 percent of the next 100 basis points, 50 percent of the next 100 basis points, and 25 percent thereafter. With a regressive mechanism, equilibrium occurs at point C. Significantly less cost savings are achieved than with the progressive mechanism (em almost 200 basis points less.
Note, however, that the progressive sharing mechanism does not perform as well in the scarce cost savings world. Equilibrium is at point A, with 30 basis points of cost savings. The regressive sharing mechanism, on the other hand reaches equilibrium at point B, a savings of 70 basis points (em although the ratepayer receives none of these savings.
However, this scarce savings scenario belies the basic premise of PBR (em namely, that traditional rate-base regulation has bred a bloated, inefficient cost structure. Moreover, the regressive mechanism fails to meet PBR's goal of increasing economic competitiveness. As the figure shows, PBR's only benefit is a small reduction in technical inefficiency with no compensation to the ratepayer and a small reward to shareholders.
If PBR regulators believe that potential cost savings are small and potential rate reductions are negligible, there really is no point in embarking upon an admittedly speculative PBR experiment. On the other hand, if PBR regulators believe that utility cost structures are bloated, a regressive sharing mechanism is unambiguously undesirable and a progressive sharing mechanism will always be preferable.
Choosing a Quality Control Safeguard
The function of the quality control mechanism is to establish a clear link between any utility cost savings achieved under PBR incentives and the maintenance of various measures of utility performance. The potential problem is obvious: The utility may be tempted to achieve false cost savings by deferring necessary maintenance, reducing service personnel, or engaging in some other type of cost-cutting that reduces some measure of performance. The solution is to devise a system that directly links the sharing of cost savings to quality standards. In designing this third component of the PBR system, the regulator must:
Determine Parameters. The relevant quality parameters should include, but perhaps not be limited to, system reliability, customer service, and employee safety. Each of these parameters is regularly measured by utilities and therefore easy to monitor. For example, one common measure of system reliability in the electricity industry is the "average number of customer interruption minutes," while employee safety can be measured by the accident rate. Similarly, customer satisfaction is typically measured, albeit less precisely, through annual customer surveys for activities and areas such as field service and meter reading, local office, telephone center, service planning, and energy services.
While these various quality parameters may be easy to measure, they are difficult to value. For example, what is the dollar value of a 5-percent decrease in service reliability or customer satisfaction?
Set Thresholds. The PBR regulator may be tempted to simply peg quality at its existing level; but these levels may not, in fact, be optimal. For example, under traditional rate-base regulation, the utility may have padded its service force, or overbuilt its generation system. However, in cutting levels of target quality, the PBR regulator runs the risk of political criticism when levels of service or reliability fall.
Establish Penalties. In theory, the optimal penalty system is straightforward: Set the penalty high enough to outweigh gains from cutting service quality.
The easiest and toughest penalty would be to deny the utility its share of cost savings when a quality parameter is breached. Under such a rule, the quality threshold is inviolable. The danger here is that such a rule would discourage risk-taking on the part of utility managers and likely lead to a non-optimal "quality cushion" well above the quality threshold. A second, more flexible, approach is to assess the penalty as some fraction of the cost savings that increases as quality falls. Regardless of the method used, the most important rule is "do not impose small penalties for big violations and vice versa."
In the "one period and deregulate" framework, the PBR regulator sets a baseline, provides the utility with incentives to beat the baseline, and then deregulates the utility at the end of the period. Under this framework, utilities have two unambiguous incentives to minimize costs.
The first incentive is the potential savings available from the sharing mechanism. Utility managers can be expected to respond appropriately to a well-designed sharing mechanism. The second incentive is the competitive pressures of the market place: Utility managers can use the PBR period to "get into shape" for the rigors of competition waiting at the end of the experiment.
Under a recurring, multi-period framework, however, PBR may not provide the same incentives or results. Utility managers will weigh the benefits of achieving costs savings in any given period against the loss of utility savings and degrees of freedom in all future periods. Suppose that in the first regulatory period, utility managers undertake the appropriate actions and investments, and the utility eliminates all possible waste from its operations. The question is: What have utility managers given up?
If we assume that this new cost structure will serve as the initial baseline for the next period in the PBR cycle, the utility will have nowhere further to go to reduce its costs. It will also run a greater risk of being "low-balled" by the PUC (em that is, having its baseline set at a level where it could actually lose money. Perhaps most
importantly, in exchange for the cost savings achieved in the first period, the utility and its managers will forgo all the benefits associated with operating with a more inflated cost curve in future periods (em larger and more plush offices, high salaries, company cars, excessive staff, and so on.
The bottom line: PBR is unlikely to be a panacea for the ills of traditional rate-base regulation, particularly when applied in a multi-period model of continuing regulation. Moreover, PBR is clearly unlikely to reduce administrative costs significantly. The analytical problems associated with setting the baseline revenue requirement alone are formidable and resource intensive. They will require a type of proceeding similar to the general rate case format. Further, the information requirements for designing an appropriate sharing mechanism are demanding, as are the resource requirements for adequately valuing nonmarketed amenities such as customer service and employee safety. All in all, PBR is not an experiment to be undertaken lightly. t
Peter Navarro is a professor of economics and public policy at the Graduate School of Management, University of California-Irvine. Mr. Navarro offers a more indepth and technical treatment of PBR in "The Simple Analytics of Performance-based Ratemaking," an article that will appear in the winter issue of the Yale Journal on Regulation.
PBR in a Nutshell
The mechanics of implementing PBR are relatively straightforward.
First, avoid setting the baseline revenue requirement too high. Do not incorporate excessive escalators, indexing factors, or passthrough mechanisms. They can remove incentives in specific areas.
Second, choose a progressive rather than regressive sharing mechanism with as many tiers as practical. Forgo PBR if potential gains are small relative to risk.
Third, design a quality control mechanism that includes worker safety, system reliability, and customer service. Link the rewards from cutting costs to penalties for cutting quality.
California Scheming: PBR in Practice
In October of 1992, citing the need to "reduce the significant burden and regulatory inefficiency that arise from traditional regulatory oversight," San Diego Gas & Electric Co. (SDG&E) asked the California PUC for permission to convert to a PBR framework. That permission was granted on August 3, 1994, in Decision 94-08-023. Unfortunately, the CPUC adopted a precedent-setting PBR framework that appears to violate the basic principles of sound PBR.
First, the CPUC rejected statistical benchmark modeling and instead approved a baseline revenue requirement proposed by SDG&E that was arguably too high and based solely on firm-specific data.
Second, the CPUC endorsed the highly regressive sharing mechanism recommended by SDG&E.
Third, although it quite appropriately approved employee safety, customer satisfaction, and system reliability as quality control parameters, the CPUC incorporated an inappropriate fourth criterion involving a comparison of SDG&E's rates to a national rate index-inappropriate because the movement of SDG&E;s rates relative to such an index is driven more by exogenous events than by any cost-cutting behavior on the part of the utility.
The initial results of this experiment are hardly encouraging. In 1994, SDG&E was able to earn a rate of return 114 basis points above the baseline PBR return granted by the CPUC (10.17 percent earned versus 9.03 percent allowed). Under the regressive sharing mechanism, SDG&E shareholders captured roughly $32 million in shared savings, and another $7 million in rewards related to the quality control and national price index parameters. In contrast, SDG&E customers actually experienced a net rate increase of roughly $6 million, because after receiving $1 million in shared savings they had to finance the $7 million in rewards distributed to SDG&E.
This rate hike strongly suggests that, at least thus far, PBR is an abject failure in California if the CPUC's primary goal is to increase economic competitiveness by reducing electricity costs. Regulators in other states would be wise to regard the California PBR experiment more as a warning sign of the dangers of PBR than as a model.
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.