V. Rate of Return

The rate of return (ROR) is the amount earned, or allowed to be earned, by a utility, expressed as a percentage of the utility's rate base. In our proceedings it is calculated as weighted average of the utility's cost of capital: the cost of long-term debt, the cost of preferred stock, and the return on common stock equity (ROE). In this proceeding there is little dispute over the debt and preferred stock portions of the cost of capital.⁷ It is the ROE that captures our attention.

A. ROE

The ROE is the most contentious issue in a cost of capital proceeding. All commentators have emphasized the need for the decision maker to apply judgment to whatever financial models are used, and those commentators have been forthright in rendering their judgment on the reasonable ROE. We begin our analysis with the financial models proposed by the parties in light of our finding that we can find no discernible risk difference between an unbundled electric utility and an integrated electric utility. Stated another way, we believe that whatever the risk differences, they have been noted by investors and are incorporated in the financial models.

We approach the financial models gingerly. We cannot repeat too often our concern.

"We have often expressed our opinion that the financial models employed in our cost of capital proceedings should not be determinative and must be tempered with a great deal of judgment. (38 CPUC2d 233, 238 (1990).) The Discounted Cash Flow (DCF) Model, Risk Premium (RP) Model, and Capital Asset Pricing Model (CAPM) cannot be relied upon exclusively to develop a particular ROE, but may be helpful in developing a range of reasonable values. (Id.) `Our consideration of these three models has always been accompanied with considerable reservation.' (Id.) First, `[t]he application and interpretation of these financial models may not accurately reflect all of the intricacies of the financial market.' (26 CPUC2d 392, 426 (1987).) Second, `[a]lthough the quantitative financial models are objective, the results are dependent on subjective inputs.' (D.91-11-059 mimeo at p. 25.) We have also recognized that the CAPM and RP models currently provide higher results than does the DCF model (33 CPUC2d 233, 238 (1990)). This continues to be true in this year's proceeding." (D.93-12-022, 52 CPUC2d 390, 406.)

What stands out immediately in the DCF and CAPM tables (above, pp. 40, 42) is the wide disparity in result. The DCF ranges from 8.7% to 11.5%; the CAPM from 8.30% to 11.35%. This disparity is based on the choice of proxies and inputs used by the experts. Much of the hearing time was consumed by the utility experts' criticizing the modeling inputs of the non-utility experts and vice versa. We see no useful purpose in attempting to resolve differences in modeling inputs because there is no satisfactory resolution that commands the approbation of either the experts or the investment community.

When we look at each party's ROE recommendation based on that party's financial model results without considering the party's adjustment for unbundling we get a more manageable value.

Table 7

For PG&E For SDG&E For Edison
PG&E	11.1	SDG&E	11.08	Edison	11.6
FEA	10.85	FEA	10.85	FEA	10.85
C-K	10.8	C-K	10.8	C-K	10.8
ORA	9.13	ORA	9.13	ORA	9.13
TURN9	9.85	TURN	9.85	TURN	9.85

The above values should be adjusted by the October 1998 DRI forecastfor AA utility bonds. However, the forecast should not be applied 1 to 1, but as discussed in Section II E, we will apply a .6 ratio (Table 8).

Table 8

For PG&E For SDG&E For Edison
PG&E	10.67	SDG&E	10.57	Edison	11.17
FEA	10.42	FEA	10.42	FEA	10.42
C-K	10.37	C-K	10.37	C-K	10.37
ORA	8.70	ORA	8.70	ORA	8.70
TURN	9.42	TURN	9.42	TURN	9.42

A final consideration is the magnitude of the change, regardless of a cold reading of the numbers. A precipitous drop would be unfair to investors and would send the wrong message to all stakeholders - the ratepayer, the utility and its employees, and the investment community. The long-term financial health of a utility should not be hostage to sudden fluctuations in the market. As we have expressed the view in the past, "We have moderated ROE increases during inflationary periods, and have declined to lower ROE abruptly when inflation is low." (D.93-12-022 at 41, 52 CPUC2d 390, 411, D.92-11-047 at 103, 46 CPUC2d 319, 370.) Therefore, we cannot adopt the recommendations of ORA or TURN.

Considering the evidence, and based on our judgment, we find that the reasonable ROE for PG&E and SDG&E is 10.6%.¹⁰ Our analysis of the model results also indicates that we should adopt an ROE of 10.6% for Edison, all else being equal. We must now evaluate whether PG&E, SDG&E, and Edison have the same circumstances by addressing the merits of Edison's argument that to reduce its ROE and undo the cost of capital trigger mechanism authorized in its PBR decision is unfair and biased because the PBR mechanism sets and adjusts its cost of capital through 2001.

There is no dispute, that, although SDG&E has a PBR mechanism, its rate of return is the subject of review and adjustment in this cost of capital proceeding. There is also no dispute that PG&E's rate of return is the subject of this proceeding, in that it does not currently have a PBR mechanism. However, in late 1996 the Commission concluded its review of a comprehensive PBR procedure for Edison (D.96-09-092). The PBR established Edison's base rates for nongeneration and distribution service for the period 1997 through 2001. The Commission excluded consideration of generation costs and anticipated removal of transmission costs to FERC jurisdiction. Edison's base rates were fixed during the PBR period, subject to adjustment for inflation, assumed levels of increases in productivity, and a cost of capital trigger mechanism to track changes in economic conditions affecting Edison's return on equity. Cost savings achieved by Edison under the PBR are shared with customers to provide an incentive to achieve efficiencies to lower rates.

Edison asserts that a key element of the PBR mechanism is the cost of capital trigger mechanism. The trigger mechanism governs Edison's return on common equity which is the benchmark for net revenue sharing with Edison's customers between 1997 and 2001. The trigger mechanism included a procedure to update the return on equity to reflect changes in economic conditions. The PBR mechanism was implemented by Commission Resolution E-3478 and became effective January 1, 1997.¹¹ The Commission accepted Edison's PBR rates based on the adopted 1997 return on equity, capital structure, and preferred stock and embedded debt costs. These facts makes Edison's circumstances distinct from those of PG&E and SDG&E.

Edison contends that changing its ROE in this proceeding would completely undo the cost of capital trigger mechanism by changing the balance of risks that was adopted in its PBR. The adjustments proposed by ORA and TURN reflect unbundling discounts and updates for changes in economic conditions which have occurred since the PBR was adopted. In Edison's opinion, those recommendations violate both the letter and the spirit of the Commission's PBR decision because the PBR mechanism already contained procedures to update the cost of capital based on changes in economic conditions.

Because we conclude that no risk premium is warranted as made clear in our discussion above, our recommended changes to the ROE are primarily based on the changed market conditions since the last cost of capital proceeding. Edison's trigger mechanism is specifically designed to adjust for these types of changes. We must ask ourselves whether it is appropriate to modify one element of a balanced PBR mechanism based on changed market conditions, especially when that mechanism specifically takes such changes into consideration. After much consideration, we agree with Edison that, at this point in time, it would be inappropriate to modify its ROE rather than continuing to allow the trigger mechanism to operate. Had we concluded that a risk premium, up or down, was appropriate as a result of unbundling, we would likely have reached a different conclusion. Therefore, we will not adjust Edison's ROE of 11.6%.

Interest coverage for each utility based on their individual capital structure and cost of preferred stock and long-term debt (updated in Exh. 88) and adopted ROE is:

Table 9

Pre-Tax Interest Coverage

PG&E	3.8x
SDG&E	4.2x
Edison	3.74x

This coverage is more than adequate and should not negatively affect the utilities' bond ratings.

B. Debt, Preferred Stock, and Rate of Return

The capital structure and costs of debt and preferred stock of the utilities have not been contested in this proceeding.¹² Consequently, we will adopt PG&E and SDG&E's proposed capital ratios and costs, with updated information from Exh. 88. PG&E's capital structure and costs are the same for its electric distribution operations and gas operations. SDG&E's capital structure and costs are also the same for its electric distribution operations and gas operations. We make no change to Edison's capital structure and costs of debt and preferred stock which were fixed in Edison's PBR mechanism.

Table 10

Capital Structures

PG&E
	Capital Ratio	Cost	Weighted Cost of Capital
Debt	46.20%	7.09%	3.28%
Preferred	5.80%	6.55%	0.38%
Common Equity	48.00%	10.60%	5.09%
Total	100.00%		8.75%

SDG&E
	Capital Ratio	Cost	Weighted Cost of Capital
Debt	45.25%	6.87%	3.11%
Preferred	5.75%	7.76%	0.45%
Common Equity	49.00%	10.60%	5.19%
Total	100.00%		8.75%

⁷ With the exception of ORA, there was no dispute over the cost of debt and preferred stock. ORA's derivation of the cost of debt and preferred stock resulted in slightly different costs than recommended by the utilities, see for example Exh. 13, pp. 1-5. ORA also believes Edison's cost of debt should be adjusted to reflect directives in the rate reduction bond decision, D.97-09-054, see Exh 13, p. 15. The maximum difference between ORA's proposal and the utilities is 10 basis points, see Exh 13, pp. 1-7. ⁸ SDG&E's testimony is confusing on whether SDG&E recommended a 20 or 100 basis point adjustment. We do not adjust the figure in Table 7 after our review of the record because SDG&E did not clearly state its position. ⁹ TURN did not use financial models to support its recommendations. We use the midpoint of its recommended range of 9.5% to 10.2% for each utility. ¹⁰ All three utilities have cited our recent Southwest Gas Corporation (SGC) decision (D.98-09-030 in A.98-05-003) for the fact that there we authorized an ROE of 11.35%. From that they argue an ROE of less than 11.35% would contradict our most recent holding. The utilities' confidence in D.98-09-030 is misplaced. That decision merely continued without modification an ROE adopted in late 1994, pending SGC's next general rate case, expected in 1999. We noted that "the trend of Commission authorized returns on equity has been downward." (Id. p. 3.) D.98-09-030 was issued ex parte, without protest. ¹¹ Resolution E-3478, Finding 8, p. 7 ("Edison should modify its Cost of Capital Trigger Mechanism to reflect the authorized return on rate base. With that adjustment, it is reasonable to approve the Cost of Capital Trigger Mechanism for adjusting nongeneration base rates.") ¹² With the exception of ORA, there was no dispute over the cost of debt and preferred stock. ORA's derivation of the cost of debt and preferred stock resulted in slightly different costs than recommended by the utilities, see for example Exh. 13, pp. 1-5. ORA also believes Edison's cost of debt should be adjusted to reflect directives in the rate reduction bond decision, D.97-09-054, see Exh. 13, p. 15. The maximum difference between ORA's proposal and the utilities is 10 basis points, see Exh 13, pp. 1-7.