III. Electric Distribution Utility Risk

The one new issue in this case is the question of whether and how unbundling and restructuring have changed or are changing the risk of the UDC, and the degree of change relative to the traditional integrated electric utility. The utilities see new and increasing risk for the UDC as a result of unbundling, restructuring, and the increase in competition which has begun to emerge as restructuring gets underway. The Office of Ratepayer Advocates (ORA) and the Utility Reform Network, Utility Consumers Action Network, and James Weil (together TURN) argue that the risk of the UDC is less than that for the integrated electric utility. The Federal Executive Agencies (FEA) and Czahar and Knecht (C and K) respond that they can't determine whether the risk is more, less, or the same, and support a review of the situation in a few years when more market data is available.

PG&E and SDG&E each recommend that in determining their 1999 ROE using traditional financial models we increase our result by 100 basis points and 20 to 100 basis points, respectively, to compensate for the increased risk caused by their becoming a distribution-only electric utility. Because they believe a distribution-only electric utility is less risky than an integrated one, TURN and ORA recommend a distribution risk discount of 30 to 124 basis points and 49 basis points, respectively. FEA and C and K recommend no adjustment for distribution risk. Edison takes no position, as it insists its ROE was determined in its last PBR decision (D. 96-09-092) and is not at issue in this proceeding.

Table 1 and Table 2 show the recommendations and the current ROE.

Table 1

ROE RECOMMENDATIONS

Recommendations A /
Party	Electric	Gas	Basis Points for Electric Distribution Risk (included in ROE)
PG&E	12.10	12.10	+ 100
SDG&E	12.00	12.00	+ 20 to + 100
Edison	11.60	NA	0
FEA (all)	10.85	10.85	0
Knecht-Czahar (all)	10.80	10.80	0
Weil-TURN-PG&E	9.00	9.10	- 30 to - 124
-SDG&E	9.10	9.20	- 30 to - 124
-Edison	8.80	NA	- 30 to - 124
ORA (all)	8.64	9.32	- 49

^A/ Before adjusting for the October 1998 DRI forecast.

Table 2

Current Authorized ROE
Party	Electric	Gas
PG&E	11.20	11.20
SDG&E	11.60	11.60
Edison	11.60	-
CPUC Historical BenchmarkB/	9.47	9.47

^B/ October 1998 DRI forecast 30 year T-Bonds 4.71 + 4.76 (The average Commission authorized risk premium as computed by ORA).

A. PG&E

PG&E argues that there have been changes in both regulatory policy and market activity which point to growing distribution competition which was not present before. PG&E asserts that until 1998, this Commission did not show support for the introduction of competition into the electric distribution business. For instance, in late 1996 the Commission decided that if a proposed irrigation district built duplicative distribution facilities, PG&E's remaining ratepayers would be adversely impacted (Resolution E-3472, November 26, 1996). Other Commission decisions from the mid-1990's state that a primary purpose of Commission regulation has been to avoid unnecessary duplication. In Pacific Corp. v. Surprise Valley Electrification Corp., D.95-10-040, 62 CPUC2d 135, the Commission stated:

"This Commission's specific constitutionally derived duty is the regulation of public utilities in California. As to electric utilities, whether they be investor-owned or cooperatives, our regulatory authority includes the structure and extent of service territories. This regulation is necessary to avoid unnecessary and wasteful duplication. From the inception of the Commission, a feature of its regulation has been the Commission's early determination that direct competition in the same geographic area where it would involve duplicating service facilities would be contrary to the public interest." (62 CPUC2d 135, 139).

As recently as 1997, the Commission reiterated the concept disfavoring distribution competition in the Application of Mather Field Utilities, Inc., D.97-04-084, mimeo., pp. 19-20, where it stated "[e]xclusivity, or freedom from competition, traditionally has been part of certificates granted by the Commission," in granting Mather Field Utilities an exclusive gas distribution franchise.

However, in 1998, PG&E sees an apparent change in regulatory policy regarding distribution competition involving distribution facilities. In June 1998, the Commission issued D.98-06-020 on the proposed PG&E/Modesto Irrigation District (MID) sale. The Commission in that decision indicated that it finds distribution facilities competition acceptable, stating:

"...in general the Commission's policy is to promote competition in all markets where competition may be economic. Apparently, competition in transmission and distribution markets may be possible in some areas of the state...Where economic competition is possible, and where other public policy goals are not unduly compromised, our policies will promote competition in utility markets." (D.98-06-020, mimeo., pp. 7-8).

Besides this change in policy in the MID decision, PG&E refers to Resolution E-3528 where the Commission found that potential construction of duplicative distribution facilities provides a competitive check on the ability of the utility to pass through costs (Resolution E-3528, discussion Paragraph 7, April 23, 1998).

PG&E's witness testified about signs PG&E is seeing in the marketplace that herald the beginning of distribution competition, including market activity by irrigation districts and important developments in distributed generation. He said MID has used its competition transition charge (CTC) exemptions to take large customers in Oakdale, Escalon, and Ripon away from PG&E. In a matter of months, MID has installed duplicative facilities and started to serve customers. MID has been selected by the San Joaquin County Board of Supervisors to provide electric service to a new development near Tracy. Meanwhile, Merced Irrigation District has begun construction of a 33-mile 115 kV transmission loop. And other irrigation districts, both existing and proposed, have expressed their interest in acquiring significant customer loads for electric distribution service.

PG&E does not expect the competition for distribution facilities service to extend to all customers or cause wholesale duplication of entire distribution systems. Instead, it believes competitors are concentrating on major customers who provide high margins. What is developing is a patchwork of small distribution systems which are situated to cherrypick the most profitable customers. This aspect of distribution competition would have at least two effects. First, the distribution utility may lose customers who provide significant contributions to margin. Second, the competition for customers may provide broad price signals having the potential to affect margins systemwide, most likely downward.

In addition, PG&E believes distributed generation (DG) is poised to compete with electric distribution service for the customer's business.² PG&E's witness testified that DG is an emerging technology which is expected to increase over the next several years as the deregulation of the U.S. electric industry drives demand for distributed generation. Recent technological developments have now made it possible to install DG in modular units.

PG&E contends that the changes in regulatory policy and market activity which it has seen this year are altering the landscape in favor of distribution competition; it points to increased risk for the UDC as it moves into an unbundled, restructured environment. PG&E believes that the direction of change in risk for the stand-alone UDC is upward. PG&E's study of restructuring impacts in other industries indicates that the return required for the regulated distribution system post-restructuring increases substantially above the level required for the formerly integrated utility. PG&E seeks an upward adjustment of 100 basis points in the return on equity to recognize the direction of the change in UDC risk.

B. SDG&E

SDG&E asserts that it is not a pure distribution utility; its unbundled UDC business is not purely a "wires" business, although it has a substantial wires service component. It argues that for purposes of describing risk that drives an appropriate return for investors, the business realities facing SDG&E involve not only wires service but more broadly a combination of public utility activities that remain subject to this Commission's jurisdiction. In addition to distributing electricity, SDG&E provides bundled commodity services to all customers that desire such service, whether by choice or on a default basis. SDG&E owns a 20% share in the San Onofre Nuclear Generating Station. SDG&E retains its contractual obligations to purchase power from other utilities and qualifying facilities (QFs). SDG&E also provides revenue cycle services. It notes that approximately 93% of its total revenue on a 1999 forecast basis remains subject to this Commission's jurisdiction. The balance, comprising forecast revenues from SDG&E's fossil generator and transmission business components, is subject to FERC's jurisdiction. SDG&E anticipates that during 1999 it will divest itself of ownership in all of its fossil plants, but until such time SDG&E owns and operates these fossil plants. Further, after divestiture it will continue to operate these plants for at least a two-year period. It is the risk of variability or volatility of earnings that is associated with this combination of business activities - all performed under a rate freeze - that an investor examines and what this Commission is legally obligated to consider when authorizing a return on equity.

SDG&E's witness testified that under the rate freeze, commodity risks associated with ISO and PX market operations are substantial. The UDC remains obligated to provide commodity service to all customers that continue to be served by the UDC either by choice or through default of an energy service provider. The commodity is acquired through the PX during the transition period (the period of the rate freeze). Purchases of commodity always involve risks to the utility providing service.

Focusing on the transition period, high PX and ISO market prices will cause a risk of cost underrecovery, including commodity costs, because of the rate freeze. SDG&E believes the risk is more serious than originally contemplated. Its witness testified that during the first several months of operations, the ISO and PX operations have created substantial price volatility in the energy and the ancillary services markets. Both the PX hourly day-ahead prices and the ISO ancillary service prices continue to experience severe upward price fluctuations. The hour ahead and ex-post prices have reached $250/MWhr during approximately 40 hours. The ISO's ancillary service prices have also been extremely volatile with prices as high as $9999/MW during July. Subsequently, the ISO initially capped the rates it was willing to pay at $500/MW, which it later reduced to $250/MW, to mitigate what the ISO characterized as market dysfunction. The ISO is considering eliminating these caps.

SDG&E contends that the ISO dispatches SDG&E's generation in an extremely inefficient manner in contrast to SDG&E's historic operation of its generation. This inefficiency has raised the cost of energy and ancillary services. ISO operations include running inefficient gas turbines much more than was appropriate; running several more units than were required to maintain system reliability; dispatching must-run generation before exhausting available market bids; going out of market to buy generation at excessive prices of more than $200/MWhr; going out of market to sell excess generation at negative prices, i.e. paying out of state utilities to take power; and restricting access to ancillary services markets which has had the effect of making these markets even thinner and less competitive with resulting higher, more volatile prices.

SDG&E's witness testified that the risk associated with the substantial commodity price volatility that the market has experienced to date creates a material risk to the UDC's ability to recover revenue requirements associated not only with the transmission and distribution components of rates but also CTC. This exposure is substantial, in SDG&E's opinion. For example, if SDG&E's Schedule PX rate was 5 cents/kwhr and SDG&E was charging an average of 3 cents/kwhr for commodity, then SDG&E would be paying the PX more for the energy and ancillary services delivered to customers than what it was charging its customers. When this occurs, as it did during August, September, and October, 1998, it could impact SDG&E's ability to fully collect CTCs during the transition period. SDG&E claims that its undercollection exposure of at least 2 cents/kwhr during a three-month period, such as occurred during August through October, is about $80,000,000.

SDG&E is concerned that in the event a direct access customer's energy service provider fails to have a scheduling coordinator, then the UDC as the default provider is exposed to increased costs due to unaccounted-for energy. The period of time this risk may last can extend for weeks if adequate notice is not provided to the UDC. Currently, approximately 14% of SDG&E's customers have selected direct access service. Therefore, while the risk associated with unaccounted-for energy is presently small, it is growing and, incrementally, it is one more cost that places greater risk on earnings volatility.

SDG&E says that it is currently experiencing competitive pressures in its distribution business by the active marketing efforts of some 59 new marketers in its service area. Of these, 13 electric service providers (ESPs) representing approximately 14% of SDG&E's electric load provide a variety of services, including commodity, billing, meter installation, and meter data management services. Such market penetration over an eight-month period strongly suggests to SDG&E that it will see many more competitors in its service area. Subsequent to April 1, 1998, SDG&E has received five inquiries from customers seeking to bypass its UDC system by means of constructing their own substations which permit direct access to the transmission grid. One large industrial customer has already contracted with an ESP to provide for the installation of such a substation. This customer tentatively could reduce its bill from SDG&E by $250,000 to $300,000 per year, which represents distribution bypass. Prior to the commencement of restructuring, SDG&E experienced three similar inquiries over a period of 20 years. If SDG&E were to attempt to maintain these customers by means of a rate discount, SDG&E would bear the cost of such discount, which would reduce realized return on equity.

SDG&E contends that there is a large degree of uncertainty as to the likelihood of full stranded cost recovery as a result of activities beyond its direct control.³ It says this uncertainty, as well as uncertainty over the recoverability of the cost of commodity or revenue requirements for distribution and transmission, is even greater for SDG&E than for Edison and PG&E since these companies can more rapidly recover their stranded costs due to the QF "cliff". This "cliff" represents a reduction of QF payment exposure by Edison and PG&E and, thus, substantially improves the probability that these two utilities may be able fully to recover their CTC. SDG&E will not see such a steep drop off in QF payments because it does not have anywhere near as large a commitment to QFs. SDG&E attributes approximately 50 basis points of the 100 basis point premium resulting from unbundling to increased risk resulting from SDG&E's having such a small drop off in QF payments during the transition period.

SDG&E believes the risks it has described and is experiencing in the distribution business require a 20 to 100 basis point premium.

C. Edison

Edison, although claiming that its rate of return is not subject to this proceeding, out of an abundance of caution has presented testimony on distribution risk.

Edison argues that the risks of an unbundled UDC are equal to or greater than integrated utility operations; the UDC bears much greater risk during the transition period and thereafter. Edison believes that the UDC bears a significant energy procurement risk. During the transition period, utility rates are frozen at the June 10, 1996 level. Within the frozen rate level, the utility must recover its operating costs, the costs of procuring sufficient energy and capacity to meet its load, pay for mandated public purpose programs, and recover its transition costs. If its operating or energy procurement costs rise, the UDC's shareholders may not be able to fully recover transition costs. The energy procurement cost is the most highly variable component of the utility's frozen rate and is completely outside the control of the utility. Customers are shielded from the risk of price increases during the transition period; utility shareholders bear the entire risk. This risk is not a generation-related risk and therefore cannot be ignored in setting the UDC's return. Utility shareholders bear the risk of recovering transition costs through the UDC's rates during the transition period. While the transition costs may have been largely related to generation assets, it is the UDC that is at risk if these costs are not recovered during the transition period.

Edison contends that there are also greater risks for the UDC during the restructuring process. The legislative underpinnings of the restructured industry have been subject to challenge in the initiative process. There may be other efforts to deny UDCs the opportunity to fully recover their transition costs. The ISO and PX are new market structures that are evolving in ways that create uncertainty for investors. There are uncertainties regarding the recovery of capital additions Edison made in 1997 and 1998 prior to the commencement of the generation market. These aspects of the restructuring process are viewed by investors as significant potential risks that must be compensated for in the allowed return on equity, in Edison's opinion.

Edison claims the new industry structure has risks the UDC has not previously borne. It notes that the Commission is opening revenue cycle services to competition from energy service providers. How revenue cycle service costs are allocated and the utility's ability to recover its costs are much more uncertain now than under integrated utility operations. The UDC bears some risk of default by energy service providers, despite the efforts of the Commission to mitigate these risks. Because Edison's total rate level is frozen , there is the potential for the Commission or other governmental agencies to mandate new utility activities with no opportunity to collect offsetting revenue to recover the associated costs. Finally, there is a growing risk of competition from distributed generation, cross-fuel competition from natural gas, and bypass of the utility as competitors seek ways to exploit the newly created market.

D. TURN

TURN argues that unbundled distribution risks are lower than integrated utility risks. It cites five areas: (1) Wall Street assessments of distribution risks, (2) the likelihood of stranded distribution costs under competition, (3) engineering and economic fundamentals regarding distribution and generation facilities, (4) the history of regulatory risks, and (5) measured variability of distribution and generation costs.

TURN believes that financial community reaction to the unbundling of distribution service provides an important reality check on business risks. It says Wall Street plainly disagrees with the testimony of the applicants. In October 1995, Fitch Investors Service issued a special report on unbundling electric utilities. The report concludes that the generation sector is likely to be the most volatile. The Fitch report states:

"Under the current cost-of-service regulation, utilities have experienced greater regulatory risk associated with generation than in their distribution and transmission activities.

In October 1996, Duff & Phelps, a credit rating agency, issued a special report on credit quality implications of electric industry disaggregation. The report states:

"In general, it is reasonable to expect that within a given rating category companies involved in only the distribution and transmission segments of the electric utility business will have a lower business risk profile."

In May 1997, Standard & Poor's CreditWeek published a feature article that concluded that electric transmission and distribution companies have relatively low business risk. The article states:

"Tightly regulated transmission and distribution utilities generally face limited business risk and can operate with relatively low operating margins and high leverage. Conversely, generating companies operating in a very competitive environment face much higher business risk and attendant cash flow volatility, and therefore generally can sustain only modest levels of debt."

In October 1997, Moody's Investors Service, a debt rating agency, issued a Special Comment report on electric distribution providers. Moody's concluded that distribution firms are more stable than generation service providers and that cash flow coverage ratios will remain the most important measure of financial risk.

TURN says it is unlikely that significant stranded distribution costs will appear. PG&E's estimate of new distribution bypass in the test year is less than 0.4% of sales. SDG&E and Edison have not even tried to forecast distribution bypass. The primary generation risk during the transition from cost of service regulation to competition is the disposition of stranded or uneconomic assets. The Commission authorized a low ROE to reflect reduced risks.

TURN contends that distribution facilities will not be stranded in the same way that expensive utility generation facilities were stranded because generation could not compete in the open market. Commodity electricity can be readily transported over the transmission grid, and all generators can compete against each other. On the other hand, distribution service is not a commodity that can be transported from one place to another, and distribution facilities will see few if any competitors. Distribution competition will be limited to service territory threats from a few alternate providers. Loss of service territory alone does not cause stranded costs.

TURN explains that on the distribution side, CTC exemptions and irrigation district tax advantages are the primary drivers of competition. These factors lead districts to take over utility facilities and service territory, but they do not encourage construction of duplicative facilities. In fact, building duplicative facilities is counter-productive because it introduces competition that would not exist if the district buys or leases utility facilities. Unlike generation, duplicative distribution service cannot be shipped elsewhere in search of sales. Utility resistance to takeovers may cause some duplication of distribution facilities, but duplication will be limited overall. When and if irrigation districts buy utility distribution facilities, sales prices will likely exceed book value, without stranded cost risk to investors.

TURN maintains that the risks of mechanical failures and consequent financial harm are lower for distribution service. Distribution technology is less complex than generation technology, and complexity is directly related to the risk of failure and increased earnings variability. Distribution systems are collections of standard, off-the-shelf components like poles, wires, transformers, and circuit breakers. Generation systems also include standard components, but individual power plants have unique designs, and major components are manufactured one at a time with long lead times.

Mechanical and electrical failure within distribution systems carry lower potential to affect utility earnings, in TURN's opinion. A single distribution component failure may affect utility service in a neighborhood or local area, but repairs can usually be made quickly. A single generation component failure can take billions of dollars of assets out of service and cause substantial repair and replacement power costs. Overall, distribution failures have a reduced impact on earnings, compared to generation failures.

TURN suggests that distribution systems are less vulnerable to cost disallowances ordered in Commission reasonableness reviews. The most significant disallowances and related settlements in recent years have arisen from reviews of large capital projects and gas transmission practices. Examples include PG&E's Diablo Canyon Nuclear Power Plant, the San Onofre Nuclear Generating Station owned in part by Edison and SDG&E, and PG&E's Canadian gas transactions. There has been no distribution disallowance of comparable importance.

Finally, TURN says that the variability of costs for distribution service is markedly lower than for generation. Investors are rewarded for the earnings risks they undertake, and risk is defined as the uncertainty or variability of outcomes. (D.94-11-076, Finding of Fact 21, 57 CPUC2d 533, 561.) There is no convenient method for unbundling past utility earnings into generation, transmission, and distribution components, but accounting records contain information about the variability of expenses. In order to test the variability of cost streams alone, TURN reviewed PG&E's electric generation, transmission, and distribution expenses, and approximate returns on rate base. TURN computed cost variability for 10 years of recorded data. The results show that the variability of generation costs is roughly four times the variability of distribution costs, and the variability of transmission costs is roughly 1.3 times the variability of distribution costs. TURN's study is limited to utility costs, but the results strongly suggest that distribution and generation earnings variability have followed cost variability. Bundled service in past years implies that variations in revenues assigned to individual services will track one another.

As a result of its analysis, TURN recommends a 30 basis point reduction in ROE.

E. ORA

ORA argues that applicants should be considered distribution only utilities and found less risky than integrated utilities. ORA supports TURN's reasoning. ORA makes the distinction between diversifiable risks and nondiversifiable risks, with only nondiversifiable risks requiring compensation in return on equity. It provides an example of a nondiversifiable risk: the state of the economy. If the general economy is bad, an investor cannot diversify that risk by diversifying his investment. Virtually all companies are affected by a weak economy. ORA asserts that every risk which the utilities have identified is diversifiable. That is, a prospective investor facing such utility risk can diversify away the risk by purchasing other securities. The risks identified by the utilities are either unique to one utility, unique to California utilities, or are symptomatic of utilities generally. Each such risk can be diversified by purchasing stock of utilities outside California, or by purchasing non-utility stock. As such, this Commission cannot compensate the utilities for these risks, regardless of their degree.

ORA maintains that the risks identified by the utilities are relatively minor ones, whether they are diversifiable or not. The three utilities are in better shape than ever. Generation, the most risky element of their business, has been dealt with by divestiture. Their remaining business is the distribution of electricity and gas. The distribution system remains regulated. Rating agencies find California distribution utilities to be in a strong position. The California utilities are also in a strong position to collect all of their transition costs. The utilities themselves believe this, and they provide such information to their investors.

But, regardless of definition, ORA states that the Commission does not need to judge whether a risk is minor or diversifiable. The risks, and investors' views on the risks, are captured by financial models. Thus investors' views of whether a company or industry is risky, the degree of the risk, and its diversifiability, are contained in the model results. That presents yet another reason to trust the model results, rather than torturing the results to increase return by identification of risks which are conjectural.

ORA believes the commodity price risk is a risk of generation, not distribution. It is a transition cost recovery risk. Given this general foundation, ORA recommends that the Commission find that consideration of changes in ROE which are related to transition cost recovery would violate Assembly Bill (AB) 1890, effectively modify the Commission's own policy decision of this issue, result in double recovery for transition cost risk, and undermine the basis for the competitive generation market and restructuring itself. ORA states that it is difficult to envision a set of findings that could wreak greater damage.

ORA contends that while AB 1890 is complex, the central structure is this: utilities have an opportunity to recover 100% of their uneconomic costs at a reduced rate of return by the end of 2001 within frozen rates. If utilities successfully manage their costs, they will recover their investment in generation-related assets plus a reduced rate of return on those assets. Thus, under AB 1890 all risks associated with the rate freeze were incorporated in the reduced rate of return. To grant an increase in the UDC's ROE for risks related to the rate freeze has precisely the same effect as increasing the reduced rate of return on utility generating assets.

ORA argues that allowing a particular form of transition cost risk to be reflected in ROE constitutes no less than double payment by ratepayers. Utility testimony clearly associates commodity price risk with the risk of asset recovery under the rate freeze. Clearly, both the Commission and Legislature have considered and explicitly determined the appropriate level of return for transition cost recovery. Reflecting commodity price risk in the distribution rate is profoundly anticompetitive. Commodity price risk affects the pricing of, and competition for, competitive generation services. Just as the commodity price risk varies for purchases from the PX, it will vary for other producers and providers of energy. Those providers must recover their commodity price plus a profit from the marketplace. The utility would not have to recover its profits from the marketplace. Those profits would be in the regulated cost of service distribution rate. If all else were equal, non-utility providers could not stay in business. The utility would buy and sell at the PX price, and earn a regulated profit. A non-utility provider would have to buy and sell at the PX price as well, or lose the customer's business. The result: no profit opportunity, no competition, and no direct access market. That outcome could not be more at odds with the Commission's overall restructuring policy.

ORA concludes its analysis of risk with the comment that volumes could be written about competitive risks, but there is no need to do so here. Utilities have faced and continued to face bypass risk in pockets of their system. Irrigation districts have long had the ability to expand into the utility's franchise territory. PG&E has provided evidence of duplication and bypass in three instances, totaling less than one million dollars. While there are clearly shallow pockets of competition, competition hardly threatens the utility's remaining services. Granting the utilities even one basis point for distribution competition would outweigh the lessened competition the utilities now face.

ORA recommends a 49 basis point reduction in ROE to compensate for reduced risk.

F. Czahar and Knecht (C and K)

C and K put little credence in the utilities' request for a risk premium because of procurement risks, or for that matter, any risks peculiar to a distribution-only utility. In regard to the utilities' claim that their UDC operations require a procurement risk premium for the possibility that high fuel costs in the next few years may keep them from fully recovering their stranded costs, C and K argue that the claim misses the symmetry of the situation, which mitigates the risk greatly: fuel prices may be low in the next few years, in which case they would tend to assure full recovery of the stranded costs, not diminish that likelihood. Further, they note that all three utilities have indicated on the public record that they expect full recovery. Thus, there is small procurement risk.

On the opposite tack, C and K reject ORA's and TURN's call for a distribution risk discount. C and K performed a multi-variate statistical analysis using a wide sample of electric utilities. Their regression analysis tested whether the result of each method for estimating the ROE (four discounted cash flow (DCF), three capital asset pricing model (CAPM), and two risk premium (RP) methods) varies with the percentages of generation, transmission, or distribution (%G, %T, or %D) in the business mix of a utility. Their results show that %G, %T, or %D never appear as significant determinants for each of the nine methods. That is, the ROE, as estimated by each of these models, is invariant with the relative fractions of generation, transmission, and distribution in an electric utility's business mix under utility regulation in the United States. This result means that estimates based on the universe of domestic electric utilities for which data are available are good proxies for the UDC. They conclude:

"No sound basis has yet been shown for different ROEs between electric-utility G, T and D sectors, as such. If efficient capital markets required differentials due to different levels of business risk inhering in G, T and D, our multiple-regression analyses almost certainly would have revealed that. Thus, we reject such differentials at this time as unfounded and unsound. The G, T and D factors did not play a significant role in any regression equation." (Exh. 23, p. 26, II. 3-7.)

G. FEA

FEA takes the position that it is too early to determine whether a distribution-only UDC has functions more risky or less risky than an integrated electric utility, with a concomitant upward or downward adjustment to the return on equity.

In calculating the appropriate cost of common equity, FEA applied the same financial models as it did in previous years, and used the same versions of those models. However, those financial models have been applied to different groups of proxy companies necessitated by the change in focus to estimating the cost of equity for the distribution function. FEA has not used the occasion of the change in focus to change its basic approach to estimating the cost of equity. FEA does recognize that many of the issues raised are new to this proceeding with uncertain outcomes.

FEA says that the changes brought about by unbundling are so new and uncertain in result that it cannot be known at this time whether they will increase or decrease risk. Plausible arguments can be made on both sides of the argument. What does seem clear to FEA is that ultimately it will be the reaction of the financial markets to these issues, and its perception of the risk associated with the Commission decisions, that will determine the effects on the cost of equity. It is too early to tell how the financial markets will react. It is also unfortunate that the positions adopted on the issues, although predictable, have resulted in a broad range of cost of equity recommendations.

FEA makes no adjustment to its ROE for exogenous changes in risk.

H. Discussion

California formally began its quest to introduce competition into the electric services industry in April 1994, when this Commission instituted an investigation and rulemaking into that industry. (I.94-04-032, R.94-04-031.)⁴ After more than a year and half of receiving evidence and comments from almost 500 persons and entities, we issued D.95-12-063, as modified by D.96-01-009, which enunciated our views of a restructured electric services industry which is expected to provide competition and downward pressure on the cost of electricity. We said "Our proposal today unbundles traditional utility services into generation, transmission, and distribution functions. . . . In the restructured industry, [utilities] would continue their obligation to provide distribution services to all customers, including direct access customers, in their service territories." (D.95-12-063, D.96-01-009 at pp. 84-85.)

We pursued our restructuring effort on many fronts, at the federal level, in the State Legislature, and in numerous decisions. But, for the purposes of this cost of capital decision, our pertinent decisions are few. In D.96-09-092, Edison's PBR decision, we said "As a part of our unbundling proceeding in electric restructuring and with coordination in the cost of capital proceeding, we intend to order separate and distinct authorized equity returns for the generation, transmission and distribution operations." (Id. p. 42.)

In D.97-08-056 (the unbundling proceeding to accomplish the policy set forth in D.95-12-063 and D.96-01-009) we said "We will consider unbundling utility cost of capital in the generic cost of capital review proceedings as proposed by PG&E and SDG&E in their comments on the proposed decision and will direct the utilities to file applications on May 8, 1998." (Id. p. 19.)

We ordered Edison, SDG&E , and PG&E to file their applications seeking review of their cost of capital for the 1999 test year. (D.97-08-056, p. 62, Ordering Paragraphs 6, 7, and 8.) Finally, in D.97-12-089 (PG&E's last cost of capital decision) we said in reference to PG&E, SDG&E , and Edison,

"For 1998, the utilities' filings for ROR and ROE will not utilize the incremental basis we apply in this decision, but will propose unbundling of long-term debt, preferred stock, and shareholders' equity to correspond to the business realities of 1998 when largely regulated distribution assets must be separated from largely deregulated generation assets. Thus, next year's cost of capital proceeding will be substantially different from those of recent years." (Id. p. 16.)

We have reviewed our decisions on electric restructuring and unbundling for two reasons: 1) to show that restructuring and unbundling are procedures well known in California and the United States since at least 1992; and 2) to show that seeking the appropriate cost of capital for an unbundled distribution system was not intended to be a mere intellectual exercise, but was to "correspond to the business realities of 1998."

PG&E and SDG&E and all intervenors other than TURN have approached the determination of the appropriate return on equity in the same manner. They determine the ROE using traditional financial modeling; then the parties that find it, adjust their result by a "distribution adjustment."

· PG&E and SDG&E add basis points because in their opinion a distribution electric company is more risky than an integrated electric company;

· ORA subtracts basis points because in its opinion a distribution electric company is less risky than an integrated electric company;

· TURN, using an incremental approach based on its filings in earlier cases, subtracts basis points for the same reasons as ORA;

· FEA and C and K make no adjustment because their analysis shows no difference;

· Edison makes no recommendation because it believes it is only a spectator in this proceeding.

The distribution adjustment is the overriding issue in this proceeding. With the adjustment the spread in ROE reaches from a low of 8.64% (ORA) to a high of 12.1% (PG&E). Without the adjustment the spread is a more manageable 9. 13% to 11.6%.

An integrated electric utility is often described as consisting of three distinct components - generation, transmission, and distribution. Prior to electric restructuring rate of return was determined on the basis of the integrated unit, not the sum of its parts. After electric restructuring the generation function and the transmission function have been considered separate functions to be treated, in an economic sense, apart from each of the other functions. Generation has been deregulated (AB 1890); transmission is now regulated by the FERC; leaving, residually, the distribution function to be regulated by this Commission.

Conventional wisdom has it that in the integrated unit the generation function is considered the most risky function, transmission and distribution less risky. Therefore, when considered separately, whatever the rate of return was for the integrated unit, the less risky distribution-only rate of return should be lower. This is the contention of ORA and TURN. The utilities see it differently. They contend that the distribution-only function, bereft of the support of its generation and transmission balance is naked to the buffeting winds of competition and, therefore, requires a higher rate of return. In this proceeding there is little controversy over debt and preferred stock; the entire thrust of each party is on return on equity.

All parties agree that the riskier a company appears to be, the higher the return on equity will be demanded by investors. It is on that basis the PG&E expert would add 300 basis points to the 7.5% after-tax weighted average cost of capital of an integrated electric utility to compensate for its loss of generation and transmission. (Exh. 1, pp. 1-5.) However, the expert tempered his estimate by actually recommending a 100 basis point upward adjustment to his benchmark 7.5% to arrive at an 8.5% after-tax weighted average cost of capital, which yields an ROE for PG&E of 13.1%. PG&E's policy witness, recognizing the uncertainties of the times and to balance shareholder and customer interests, requests an ROE of only 12.1%. (Exh. 1, pp. 1-5.)

SDG&E `s expert, using more conventional methods, recommends that SDG&E `s unbundled distribution business be allowed an ROE in the range 11.6% to 12.8% (Exh. 6, p. JVW-34). SDG&E `s policy witness recommends an ROE of 12.0% (Exh. 6, p. CAM-12), although he believes a 100 basis point upward adjustment for the new regulatory scheme would be reasonable (Exh. 6, p. CAM-10).

Edison, for reasons discussed elsewhere in this opinion, requests a continuation of its 11.6% ROE. It did, however, present an expert to discuss the risks of the new regulatory scheme. His analysis found a risk-return differential of 30 basis points between a distribution business and a vertically integrated operation, with distribution the less risky component (Exh. 10, p. 47). Nevertheless, he says that his estimates of the risk of the wires business are conservative "because they ignore the lost benefits of vertical integration. The mere act of unbundling the business will make each of the newly formed independent businesses riskier in the future." (Exh. 10, p. 39.)

The experts of all three utilities cite the same competitive threats and risks inherent in the distribution business which, in their opinions, require an increase in ROE.

· Competition releases competitive energies; one cannot predict how non-utilities will react to the deregulated industry,

· Metering and billing functions are currently under attack and high margin customers are the target,

· UDC's costs are fixed; they cannot be reduced if demand is reduced,

· There are no balancing accounts to match costs and revenues; there is a commodity price risk,

· Bypass is a distinct possibility; municipal districts are entering the electric distribution business,

· Distributed generation may be the first step to bypass,

· Regulatory risk - rapid changes by legislatures and commissions create uncertainty,

· Procurement risk - the UDC may not be able to recover the full cost of energy purchased on behalf of its customers,

· The risks and rewards of performance based regulation, and

· For SDG&E , its QF cliff.

ORA and TURN do not accept the utilities' arguments, claiming they are more theoretical than real. They believe the evidence shows that in the actual world of investors and current regulation the distribution business is less risky than the integrated utility and, therefore, should have a substantially reduced ROE.

We will not discuss each potential risk to determine its viability. For the reasons set forth below we find that a distribution only UDC is neither more nor less risky than a vertically integrated electric utility. Our starting point is Bluefield and Hope. The return should be commensurate with the expected return on investments with similar risks.

ORA asserts that "the Commission is not and should not set a rate of return for the firm or for the utility as a whole, but for the property that is the subject of this proceeding," which are "the risks of activities included in those unbundled distribution rates." (ORA Reply Brief p. 17.)

Here we are setting a return so that shareholders have the opportunity for earnings commensurate with investments of similar risks. We cannot, by fiat, say that some risks do not exist, e.g., procurement, commodity price volatility. The focus of this proceeding is the appropriate return for UDC operations. Although, the three utilities are not pure distribution utilities now and will not be pure distribution utilities for the foreseeable future, a separate return has been established for generation assets. However, at this time, the UDC is more than a "wires and meters" business. For example, the utilities retain significant responsibilities serving as the electricity providers of last resort.To properly reflect the Bluefield and Hope criteria we cannot base our result on less than the actual operations of the utility (recognizing that the FERC has set the transmission return and that the return for generation assets has also been previously set). This does not contradict our decision to consider the cost of capital for unbundled operations (D.97-08-056 at p.19). We have considered that cost and find that it is comparable to bundled operations.

The evidence that PG&E and SDG&E require a premium on ROE because of increased risks is not persuasive at this time. We accept that the distribution function is less risky than competitive generation functions. TURN's and ORA's discussion of rating agencies' opinion is pertinent. We note that financial rating agencies advise clients that distribution companies have less risk than generation companies. Moody's is directly on point: "The wires business will entail the lowest business risk of the future distribution business lines. . . .Although performance-based ratemaking and the effects of regulatory lags to recoup weather-related expenditures for example, may add some slight volatility to its cash flows, the wires business's prospects will remain highly predictable." (Exh. 12, Att. 7, p. 3.) S&P states "Standard & Poor's measures financial strength by a utility's ability to generate consistent cash flow to service debt, finance operations, and fund investment. . . .Tightly regulated transmission and distribution utilities generally face limited business risk and can operate with relatively lower operating margins and high leverage." (Exh. 12, Att. 8, p. 28.)

The utilities' argument that Moody's and S&P's recommendations are only valid for bond purchasers is unpersuasive. If anything, investors in equities are more concerned about risk than bond investors. We cannot envision how a company's risk could be lessened for prospective bond purchasers at the same time it is rising for prospective equity purchasers. Based on this review we find no premium is warranted.

In regard to the litany of risks proffered by the utilities we are of the opinion that although real, they are exaggerated. TURN has described that exaggeration. Distribution competition is limited relative to generation competition because generation can be transported wherever there are wires, but distribution competition is localized. Irrigation districts have little incentive to build duplicative systems because that would put them in competition with massive utility companies. Distributed generation and other forms of bypass have potential, but are in their formative stage and their impact will be further assessed in the Distribution Rulemaking, R.98-12-015. Whether they will be a serious threat is too early to tell. There is less variability in distribution costs relative to generation. And at present any loss of revenue because of price and procurement risks appears minuscule when compared to total revenue.

More persuasive to our conclusion that no premium is due are the views of FEA and C and K, as well as the testimony of ORA's expert and the market fundamentals espoused by PG&E's, SDG&E's, and Edison's experts.

FEA's expert refused to base his opinion on his subjective perception of whether a distribution-only electric utility was more or less risky than an integrated electric utility considering the risk-affecting factors assumed by the other witnesses. He based much of his analysis using a group of natural gas local distribution companies as a proxy for the unbundled electric distribution company. Both SDG&E's expert (at Ex 7, p. JVW-7) and Edison's expert (at Ex 10, p. 24) agree that LDC's are a valid proxy. FEA's expert believes the proper approach is to use traditional financial methods applied to companies closely comparable to a distribution-only utility. The results of that exercise would serve as the foundation for the ultimate judgment of the ROE. In his opinion, it is inappropriate to add or subtract basis points for perceived changes that cannot be gleaned from financial models. (Although he, as do all the witnesses, states that financial models are the basis for judgment, not a substitute.)

We have set forth C and K's opinion above. They conclude that their statistical analysis does not show variations for companies having more or less distribution risk in relation to generation and transmission. They would neither add nor subtract basis points based on perceived subjective changes for a distribution-only electric utility. They point out that years have passed to permit the financial markets to have absorbed the effects of restructuring changes and expectations, yet their analyses of financial data could disclose no difference in risk between electric companies with different distribution, transmission, and generation ratios.

PG&E's expert testified "the cost of capital is determined in capital markets, market values both determine and reflect its risks, and market values must be used to calculate it." (Ex 2., p. 2-8, 2-9.) Then he attempted to draw parallels between (i) distribution-only electric companies and (ii) electric utilities recently divested from state ownership in Great Britain and the restructured telephone industry in the United States to support his opinion that investors will demand more return from a distribution-only UDC. We see no meaningful comparison between U.K. electrics and U.S. electrics, nor do we see a meaningful relationship between the deregulation of the telephone industry and the deregulation of the electric distribution industry based on the evidence presented by PG&E. It is difficult to understand why the U.K. capital market and the U.S. telephone market are reasonable proxies for California electric companies, but the U.S. capital markets of electric companies and gas companies are not.

ORA's expert said in regard to competition "this has been a hotbed of investor interest, particularly utility investor interest for three or four years. So if the financial models work right, they will incorporate investors' expectations of the risks associated with any source, including competition. So I don't think the fact that there may be competitive risk requires any adjustments in the models because the investors have already incorporated that into their thinking." (Tr. Vol. 10 p. 12, 92-93.)

On balance, we agree that no basis points should be added to or subtracted from a financial model to account for subjective perceived changes in risk for a distribution-only electric utility. Electric utility restructuring has been well known in California and in the United States since at least 1994. California has taken the lead and other states are following, as is the federal government. Investors are quick to react to changes, and potential changes, in the market place. We have every reason to believe that the financial community has factored into its activities its expectations regarding restructuring. Changes in economic expectations are usually reflected immediately in financial markets; four years is more than enough time to reflect competitive risk. The commentators - Moody's, S&P - and the public pronouncements of the experts testifying in this proceeding are proof of that. At present we will not modify our ROE finding for a distribution discount or premium. However, as electric restructuring unfolds, we anticipate investors expectations may also change. Therefore, it would be appropriate to reevaluate the risk associated with the UDC no later than the 2002 cost of capital.