Revisiting performance-based rates with endogenous market designs.
More than 20 years ago in the pages of this publication, economist William Baumol outlined a method by which the regulation of public utility monopolies could be streamlined while simultaneously providing incentives for efficiency and productivity growth.1 Baumol proposed a productivity incentive clause that adjusts rates automatically according to the formula,
where, is the allowable annual rate change, is the annual change in input prices, and is the annual productivity requirement.
This now familiar formula describes a price-cap mechanism. Such rate adjustment mechanisms have found widespread applications in performance-based regulation (PBR) of telecommunication monopolies (for which Baumol originally proposed the mechanism), as well as restructured electricity and water utilities. However, is it possible that regulators selected the wrong market design in their restructurings during the past two decades? If so, what were the consequences? What market design should regulators have employed to mitigate these problems?
The Hope and the Reality
Baumol identified several advantages to this method that still are used to promote price-cap regulation: increased productivity growth, decreased regulatory lag, and reduced regulatory burden. Unfortunately, what appeared at first glance to be a simpler and more straightforward approach to ratemaking has not worked out that way. Unanticipated and unforeseen complexities in setting the inflation and productivity parameters have made the regulatory process for determining the price-cap formula complex and contentious. Add-ons, off-ramps, and other side agreements (such as earnings-sharing mechanisms or cost pass-through mechanisms) imposed by regulators or demanded by interested parties limits the efficiency incentives and further complicates the process. Finally, and perhaps most importantly, errors in specifying plan parameters are not market neutral and can be costly to consumers and other regulated firms.
Due to the nature of the process by which regulators and the monopolies they regulate react to each others' conflicting interests (the principal-agent problem), regulators do not have all of the information necessary to correctly set the parameters in these exogenously determined PBR plans, and the regulated firms have little incentive to give them this information, even if they know it. In such an asymmetric environment, determining the appropriate inflation escalator and productivity offset can be complicated, confusing, time-consuming, expensive, and divisive. Often, the necessary data is as difficult, or more difficult, to obtain than the process of determining the firm's cost-of-service. Thus, exogenously determined PBR plans often suffer from the same shortcomings as rate-of-return/cost-of-service plans, and may in fact result in worse outcomes, if plan parameters are significantly off the mark.
Good News and Bad News
The good news is that regulated firms do respond to the incentives presented to them. For example, research on 48 electricity distributors in the province of Ontario finds that utilities increased their average annual growth in total factor productivity (TFP) by 2.3 percent after the imposition of a variable offset PBR (, price freeze), rising from -0.2 percent prior to the freeze to 2.1 percent after its imposition.2 This increase was pervasive among individual distributors and consistent across size classes. On average, small, medium and large utilities had relatively similar increases in productivity growth. Firms understood the incentives and responded to them (see Table 1).
The bad news is that even after two decades of experience, critical design issues such as asymmetric information, allocative inefficiency, initial efficiency differences, and adjustment dynamics have not been resolved, or in some cases, even addressed. Despite the opportunity provided by group restructurings in network industries in the United States, Canada, and the UK, regulators opted for exogenously imposed, fixed factors (except for the Interstate Commerce Commission [ICC] with rails) based only on estimates of technical efficiency improvements, rather than PBR predicated on inter firm-based competition. Such unresolved issues are the direct legacy of decisions to implement PBR in a data-intensive, regulator-imposed, exogenous framework, rather than the minimalist information, endogenous shadow-market advocated by some researchers. In some cases (, adjustment dynamics/periods), regulators have opted to impose productivity offsets and performance periods with little to no research upon which to base their parameters;3 in other cases, (, the extent and sources of initial inefficiency), regulators have seemingly ignored earlier, critical research findings.
Critical Research Ignored
In the late 1970s and early 1980s, researchers studying the relative efficiency of public and private utilities found that utilities were operating at significantly different levels of both technical (i.e., output per input ratios) as well as allocative (i.e., input mix costs) efficiency, some utilities being only 25 percent as efficient as the best practice firm. Furthermore, two-thirds of the inefficiency was due to non-optimal factor mix choices.4 Yet, not only do most PBR restructurings ignore the issue of varying levels of initial efficiency among utilities, the process of setting the productivity offset (sometimes called the X factor) relies only on changes in technical efficiency, leaving moot the extent, or even existence, of allocative inefficiency ().
In 1999, the Ontario Energy Board (OEB), the gas and electric distributor regulator, undertook the stakeholdering process associated with the restructuring of more than 275 electric distributors. Some stakeholders in the Ontario restructuring maintained that utilities varied significantly in their initial efficiency. Furthermore, stakeholders argued that a number of utilities were overcapitalized due to an over-reliance on customer contributed capital. Unfortunately, deadlines in the restructuring process did not allow the time necessary to properly examine these contentions by benchmarking initial levels of efficiency.
The Road Not Taken
It did not have to be this way; alternative paradigms of PBR design also were proposed. In 1985, Schleifer introduced the concept of Yardstick Competition (YC).5 YC describes the simultaneous regulation of identical or similar firms. Rewards to an individual firm depend on its standing vis-à-vis a shadow firm. Participants in the tournament (the competition with the shadow firm) reveal information about their cost structure and productivity potential to the regulator. With correlated underlying costs, the regulator can use performance information from a group of firms to better define the efficiency frontier and the time path necessary to attain it for inefficient competitors.
Such YC effectively handles the principal-agent problem of regulators with asymmetric information. Indeed, this approach requires a minimum amount of information-just service prices for each competitor. Initial efficiency differences among utilities are revealed by using correlated cost information from like firms. Allocative inefficiency is benchmarked by using total cost comparisons, not just data on changes in output/input ratios. And, rather than regulator-imposed productivity factors and adjustment periods, potential efficiency improvements and the adjustment periods are revealed by endogenizing the process through the tournament ().
It is illustrative to note Scheifer's comments in the original paper outlining the concept. Schleifer states that YC works because it "does not let an inefficient cost choice by a firm influence the price and transfer payment that that firm receives. It is essential for the regulator to commit himself not to pay attention to the firms' complaints. … Unless the regulator can credibly threaten to make inefficient firms lose money … cost reduction cannot be enforced."6
Recent Misapplications of 'Yardstick Competition'
Recently, regulators in Europe, the UK, Australia, and the United States have structured with "so-called" yardstick paradigms based on a limited time series of partial costs (generally ignoring capital costs and line losses). In many of these jurisdictions deterministic frontier techniques are employed. Unfortunately, such PBR implementations have selected the worst of both paradigms. Using data-intensive but "quick fix" applications, regulators have opted to preserve the traditional exogenous process, with a new twist: benchmarking on partial costs that produces unstable peers and significant biases in inter-utility efficiency rankings.
In these applications, utilities have been benchmarked for differences in technical efficiency but not allocative (due to the difficulty of calculating input prices) and regulators have opted for notable efficiency improvements as mentioned above. Our research has examined such applications in detail: First, we find such frontier paradigms are unstable. Firms judged to be efficient in one period are judged inefficient in the next period and so on. Second, we find that a utility's technical inefficiency is not necessarily correlated with its allocative inefficiency. Thus, one cannot assume one can over-penalize based only on the calculated technical inefficiency (assuming it was done correctly) to compensate for unknown but suspected allocative inefficiency. Third, we find significant differences in calculated efficiency between the partial-cost approach and one based on total costs, which produces a better estimate of both technical and allocative efficiency. Not surprisingly, such attempts to yardstick benchmark on partial costs have created sizeable distortions in efficiency rankings and X factors.7
The PBR toolkit available to regulators is a powerful regulatory incentive mechanism and can work to achieve a more socially optimal outcome if applied correctly. Proper efficiency benchmarking is critical. Rather than the data-intensive, regulator-imposed, exogenous framework traditionally implemented in North America, or the partial cost/partial efficiency "quick fix" benchmarking recently employed in Australia, Holland, and California, among other jurisdictions, we suggest the adoption of endogenous shadow-market, yardstick competition. This approach simplifies the regulatory process and handles issues such as the principal-agent problem and allocative inefficiency and initial efficiency differences with elegant simplicity. Exogenously determined YC as practiced recently only should be implemented with cost-per-unit comparisons based on full costs, including capital and line losses, as well as calculated input prices. This would reduce the bias in inter-utility efficiency benchmarkings and allow total efficiency comparisons, including both technical as well as allocative efficiencies.
- William J. Baumol, "Productivity Incentive Clauses and Rate Adjustment for Inflation," , July 22, 1982, pp. 11-18.
- Cronin, F.J., et al., 1999, "Productivity and Price Performance for Electric Distributors in Ontario," Ontario Energy Board Staff Report, available at http://www.oeb.gov.on.ca/documents/cases/RP-1999-0034/ppp1.html.
- Both the British regulator, Ofgem, and the Dutch regulator, Dte, have imposed what some participants viewed as relatively sizeable efficiency improvements without any empirical basis for these dynamics/periods. In the case of Dte, some LDCs had X factors of 9 percent per year imposed.
- Fare, R., S. Grosskopf, and J. Logan, "The Relative Performance of Publicly-Owned and Privately-Owned Electric Utilities," 26 (1985): 89-106
- Shleifer, A., 1985, "A Theory of Yardstick Regulation," , Autumn: 319-27.
- Ibid., p.323.
- Cronin, F.J., S.A. Motluk, "Examining the (Mis) Specification of Peer Group Performance Benchmarks for Electric Utilities," presented at the North American Productivity Workshop, Union College, June 2002.
An Example of Asymmetric Information
The FCC's initial review of the three-year-old PBR plan for telephone local exchange carriers (LECs) took four years. LECs submitted evidence during the review that their X factor had actually fallen, dropping to only 1.7 percent. In a presumed effort to draw out potential through revealed preferences, the FCC offered new interim X options ranging from 4.3 to 5.3. Despite having just argued for 1.7, the majority of LECs selected the new, no-sharing X option of 5.3 rather than the lower, earnings-constrained X options starting at 4.3. In their follow-on 1997 price cap decision, the FCC raised the X factor to a single option 6.5, a doubling over the initial X factor. During the 1990s major price cap LECs' return on equity (ROE) also more than doubled, reaching 29 percent in 1999. Over this same period, inter-exchange carriers' (IXCs') ROE fell to 2.1 percent. LECs' ROE would have ranked third highest out of 42 industries and IXCs' would have ranked last.1 -F.J.C. and S.A.M.
- Cronin, F.J. and M. Gold, "Prices, Profits and Productivity: Analytical Miscues and their Effects on Telecommunications Deregulation," presented at the Southern Economic Association meetings. Given the LEC response, the FCC's design would have been more effective (i.e., would have obtained more information on productivity potential) with a higher ceiling.
Technical vs. Allocative Efficiency
Productive efficiency combines both technical efficiency and allocative efficiency. Technical efficiency refers to the minimum combination of inputs to produce outputs. A firm that is technically efficient is operating on the production possibility frontier SS'. A firm that is technically inefficient would be operating somewhere in the interior of SS'. Though Q and Q' represent different combinations of inputs, both are technically efficient. P is technically inefficient. Allocative efficiency refers to the mix of inputs selected, given their relative prices. The input mix at Q is non-optimal and therefore allocatively inefficient; given its non-optimal input mix, however, Q does maximize its output. Total economic efficiency is maximized at Q', where both technical and allocative inefficiency is zero. Some regulators have explicitly ignored allocative inefficiency because of the purported difficulty in measuring input prices; in general, the issue of non-optimal factor mix has typically been ignored in many regulated industry studies. (Economist M.J. Farrell originally illustrated the example in the 1950s.)1
- Farrell, M.J., " The Measurement of Productive Efficiency," , Series A, Part III, Vol. CXX, 253-290.
Growth vs. Income: A Profit-Productivity Menu
Given the asymmetric information on initial efficiency levels and a desire to maximize the efficiency response among distributors, the authors1 and others proposed a variant of the FCC's productivity menu. Distributors could choose a combination of productivity growth and earnings return: higher productivity growth would be permitted higher returns on equity. The menu offered combinations from 1.25 percent productivity growth and 10 percent ROE up to 2.5 percent productivity and 15 percent ROE.2 Some observers argued that such a menu would disadvantage more efficient utilities by requiring them to make the same percentage improvement as inefficient utilities for the same ROE.3 In its PBR decision, the board rejected the menu and opted for a single, fixed factor of 1.5 percent.4
In research undertaken after the initial PBR process, we found electricity distribution monopolies in Ontario operating at significantly different levels of efficiency: on average, utilities are 35 to 40 percent less efficient than the best practice firms.5 And, similar to the earlier research from the 1980s, we too find that allocative inefficiency is about twice as large as technical inefficiency (i.e., about two thirds of the mean total inefficiency). Indeed, just as some stakeholders noted, we find a key determinant of allocative inefficiency is overcapitalization induced by a utility's reliance on third-party contributed capital to fund infrastructure requirements.6
- Ontario Energy Board consultant and staff, respectively.
- Ontario Energy Board Staff Report, , June 1999. http://www.oeb.gov.on.ca/documents/cases/RP-1999-0034/handbook0.html.
- Indeed, in an unregulated market, the profits of the most efficient utility would be determined by its cost level relative to the cost level of less efficient utilities.
- Ontario Energy Board Decision with Reasons, , Jan. 18, 2000. http://www.oeb.gov.on.ca/documents/cases/RP-1999-0034/dec.pdf.
- Specifically, we find technical inefficiency averages 18 percent, with some distributors almost 40 percent less efficient than the best producers. Allocative inefficiency averages 26 percent, with some distributors almost 50 percent less efficient than the best producers. Total inefficiency was found to exceed 40 percent among many of the utilities examined (, they would need to reduce total costs by more than 40 percent while holding output constant to match the efficiency of the best practice firms). Cronin, F.J. and S.A. Motluk, "Inter-Utility Differences in Efficiency," presented at the Canadian Economic Association meetings, May 2001.
- Cronin, F.J., S.A. Motluk, "Agency Costs of Third Party Financing and the Effects of Regulatory Change on Utility Costs and Factor Choices," submitted for publication.
Articles found on this page are available to subscribers only. For more information about obtaining a username and password, please call our Customer Service Department at 1-800-368-5001.