XV. Cost of Capital

Park provides financial, administrative, accounting, engineering and data processing support for AVR and its other subsidiaries. While Park has external debt, AVR and its affiliates do not. Park serves as a common source of any necessary debt capital for its subsidiaries because with its size it can acquire debt at more favorable rates than could any of its subsidiaries. Since Park serves as the de-facto borrower for these subsidiaries by providing a source of capital through inter-company transactions, there is in effect one common capitalization for Park and its subsidiaries. Thus Park uses a consolidated capital structure applicable to all its subsidiaries. We use the same capital structure here on an estimated basis.

AVR and ORA agreed on the capital structure and long-term debt cost for AVR's test years. However, they disagreed on the appropriate return on equity (ROE). Their recommended capital structure and ROE for the test years are shown in the following table.

AVR ORA

	Capital Ratio	Cost Factor	Weighted Cost	Capital Ratio	Cost Factor	Weighted Cost
TEST YEAR 2003 Long Term Debt Common Equity Total	40.09% 59.91 100.00%	8.58% 12.00	3.44% 7.19 10.63%	40.09% 59.91 100.00%	8.58% 9.53	3.44% 5.71 9.15%
TEST YEAR 2004 Long Term Debt Common Equity Total	40.09% 59.91 100.00%	8.56% 12.00	3.43% 7.19 10.62%	40.09% 59.91 100.00%	8.56% 9.53	3.43% 5.71 9.14%
Attrition Year 2005 Long Term Debt Common Equity Total	40.09% 59.91 100.00%	8.54% 12.00	3.42% 7.19 10.61%	40.09% 59.91 100.00%	8.54% 9.53	3.42% 5.71 9.13%

Equity cost is a direct measure of the utility's after-tax ROE investment. Its determination is based on subjective measurement, and not susceptible to direct measurement in the same way capital structure and embedded long-term debt costs are.

Both AVR and ORA acknowledged the well-established legal standard for determining a fair ROE, and we have many times cited that same legal standard. In the Bluefield Water Works case, the Supreme Court stated that a public utility is entitled to earn a return on the value of its property employed for the convenience of the public, and set forth parameters to assess a reasonable return.³² That return should be "...reasonably sufficient to assure confidence in the financial soundness of the utility and should be adequate, under efficient and economic management, to maintain and support its credit and enable it to raise the money necessary for the proper discharge of its public duties."

As the Supreme Court also noted in that case, a utility has no constitutional right to profits such as are realized or anticipated in highly profitable enterprises or speculative ventures. In 1944, the Court again considered the rate or return issue in the Hope Natural Gas Company case, stating, "[T]he return to the equity owner should be commensurate with returns on investments in other enterprises sharing corresponding risks. That return, moreover, should be sufficient to assure confidence in the financial integrity of the enterprise, so as to maintain its credit and to attract capital."³³

Hence, we set the ROE at a level of return commensurate with market returns on investments having corresponding risks, and adequate to enable a utility to attract investors to finance the replacement and expansion of a utility's facilities to fulfill its public utility service obligation. To accomplish this objective we have consistently evaluated quantitative financial models and risk factors prior to exercising informed judgment to arrive at a fair ROE.

A. Financial Models

The quantitative models commonly used in ROE proceedings as a starting point to estimate investors' expectations for ROE are the discounted cash flow (DCF), risk premium (RP) and capital asset pricing model (CAPM). Although the parties agreed that the financial models are objective, the results are dependent on subjective inputs. Detailed description of the CAPM, DCF, and RP models are contained in the record and are not repeated here.

Although the parties agree that the models are objective, the results are dependent on subjective inputs. For example, each party used different proxy groups, betas, growth rates, and calculations of market returns. It is the application of these subjective inputs that resulted in a wide range of ROEs being recommended by the parties as shown by the results of their individual DCR, RP, and CAPM models. From these subjective inputs the parties advance arguments in support of their respective analyses and in criticism of the input assumptions used by the other party. These arguments will not be addressed extensively in this opinion, since they do not materially alter the model results. In the final analysis, it is the application of judgment, not the precision of these models, which is the key to selecting a specific ROE estimate within the range predicted by analysis.

1. AVR's Financial Models

AVR estimated the ROE that Park's investors expect to earn by applying the DCF, RP, and CAPM models to a selected proxy group of three water utilities. The criteria it used to select this proxy group were that the water utilities have bond ratings by Moody's or Standards & Poor's (S&P); that water operations account for at lest 70% of the utilities' operations; and that there be analysts' forecasts of future earnings, dividends and returns on equity.

AVR supplemented its small sample of water utilities with a separate proxy group of eight gas distribution (gas) utilities by applying the DCF and RP models. The criteria used by AVR to select this proxy group were that the gas utilities paid dividends; operations accounted for at least 70% of the utilities' operations; they had at least one bond rating of A or better; and data required to make a DCF analysis were available.

AVR derived an overall 10.6% to 12.0% ROE range from the results of its DCF, RP, and CAPM models applied to its water utilities proxy group and a 11.40% to 12.10% ROE range from the results of its DCF and RP models applied to its gas proxy, as summarized in the following table.

2. ORA's Financial Models

ORA estimated the ROE that investors expect to earn from Park by applying a selected proxy group of seven water utilities to the DCF and RP model. The criteria used by ORA to select this proxy group were that water operations accounted for at least 70% of the utility's revenues and that the utility's stock is publicly traded.

ORA applied three variations of the DCF model to mitigate period specific biases and to consider both current and long-term trends. It also applied two variations of the RP model to its same proxy group. ORA derived an overall simplified 7.82% to 11.24% average ROE range from the results of its DCF and RP models applied to its water utilities' proxy group, as summarized in the following table.

3. Discussion

We view the output of the quantitative financial models provided by the parties with some skepticism. Although AVR acknowledged that the Commission has found that energy utilities require higher ROEs than large water utilities in the past, it contends that new evidence "indicates" that the risk differential between water and gas utilities has been reduced or no longer exists.³⁸ That evidence consisted of a comparison of two risk measurements, beta and Value Line's safety rank, and notice of a June 21, 1999 Utilities & Perspectives in which S&P announced that it has created a single set of financial targets that can be applied across the different utility segments.

Beta is used in the CAPM model as a measurement of risks of holding a stock in a diversified portfolio. Value Line's safety rank is a measurement of risk by Value Line of the risk an investor incurs when holding an individual stock as opposed to the risk of holding the stock in a portfolio. Both measurements are based on subjective inputs and not necessarily uniformly agreed to. The S&P announcement relied upon by AVR was issued prior to D.01-04-034 in which we concluded that water utilities' risks were not comparable to those faced by energy utilities.

AVR further reduced its gas proxy results by 50 basis points to "provide a conservative adjustment for potential differences in required ROEs for gas distribution utilities and large water utilities."³⁹ This arbitrary reduction tends to confirm that risk differentials between water and gas utilities continue to exist. No weight is given to AVR's gas proxy results because AVR has not substantiated that our prior determinations that water utilities risks are not comparable to energy utilities' risks should be overturned.⁴⁰

We are also skeptical of AVR's CAPM model result based on a proxy sample of only two water utilities and a comparison of that proxy group to industrial companies operating in an environment foreign to regulated water utilities. For example, industrial companies do not have the benefit of recovering their operating expenses through a 50% service charge, balancing accounts, and memorandum accounts like regulated water utilities do. Hence, we give no weight to AVR's water proxy CAPM results.

We also view ORA`s DCF result with skepticism. ORA's witness acknowledged that common stock investments are riskier than long-term debt investments, thus leading investors holding common stock to expect higher returns.⁴¹ As AVR has, we observe that ORA's average DCF result appears to demonstrate the opposite. This is because ORA's 7.82% DCF result is 76 basis points lower than the 8.58% long-term debt agreed to by the parties. We recognize that each model may have its own individual bias, whether high or low, making it appropriate to average the various model results as ORA did to arrive at its recommended 9.53% ROE for AVR. We also recognize that AVR's common equity ratio is higher than ORA's proxy group, giving the appearance that its risk is lower than the proxy group because its percentage of debt is less and risk premiums should not be based on company-specific data.

From these broad ROE ranges based on subjective inputs and our skepticism identified above, we apply informed judgment in adopting a 9.21% to 11.22% ROE range we deem fair and reasonable for AVR. We derived the floor rate by taking the simple average each party's lowest water proxy financial model results.⁴² The ceiling rate was derived by the same basis, by taking the simple average of each party's highest water proxy financial model results.⁴³

We next assess financial, business and regulatory risk factors to determine whether a higher range of ROE is warranted for AVR so that it may continue to attract investors and fulfill its public utility service obligation.

B. Risk Factors

Risk factors consist of financial, business and regulatory risk. Financial risk is tied to the utility's capital structure. The proportion of its debt to permanent capital determines the level of financial risk that a utility faces. As a utility's debt ratio increases, a higher return on equity may be needed to compensate for that increased risk.

Business risk pertains to uncertainties resulting from competition and the economy. That is, a utility that has the most variability in operating results has the most business risk. An increase in business risk can be caused by a variety of events that include poor management, and greater fixed costs in relationship to sales volume.

Regulatory risks pertain to new risks that investors may face from future regulatory actions that we, and other regulatory agencies, might take. Assessments of these risks are conduced to determine whether there is a need to increase a ROE to compensate investors for added risks.

1. AVR's Risk Assessment

AVR concluded from its risk assessment analysis that Park's investors need to be compensated for additional risk due to factors such as its small size and uncertainty of recovering investments and expenses due to weather, potential disallowance of investments and expenses, and Resolution W-4294. AVR seeks at least 90 basis points added to the ROE determined fair and reasonable so that investors are compensated for this additional risk. This results in its recommended 11.90% to 12.20% ROE range, of which it seeks a 12.00% ROE as fair and reasonable for its test years.

2. ORA's Risk Assessment

ORA disagreed with AVR's risk assessment and instead found from its own risk assessment that risk was low. This was based on the fact that Commission provides AVR a multitude of mechanisms designed to minimize risk. These mechanisms included balancing accounts for purchased water and power, and pump tax; memorandum accounts for catastrophic events and Safe Drinking Water Bond Act compliance; a service charge to recover 50% of its fixed cost; and, inclusion of construction work in progress in rate base.

ORA balanced these mechanisms with Park's high common equity percentage (59.91% compared to 47.82% of ORA's proxy group) to demonstrate that it faces significantly less financial risk than ORA's financial model proxy group. To confirm its low risk conclusion, ORA compared S&P benchmark financial ratios (even though S&P doesn't rate AVR or its parent Park) to see what Park's overall rating would be. Its result was a financially healthy "AA" rating.

Hence, ORA opposed any increase in AVR's test year ROE for added risk. Applying equal weight to its DCF and RP model results, ORA recommended that AVR be authorized a 9.53% ROE.

3. Discussion

AVR has previously earned a premium on its ROE due to its small size, limited sources of external financing, and the fact that its stock is not publicly traded. In D.99-03-032, its prior GRC, we authorized a 30 basis premium in AVR's ROE to fairly compensate its investors for this overall perceived risk. AVR now seeks a 90 basis point premium based on its current risk assessment.

Some risks are inherent to the water industry. Two of these risks are cited by AVR as justification for increasing its ROE above the quantitative model ROE results. They are weather uncertainty and potential disallowance of investments and expenses.

A premium on the authorized ROE is not appropriate for inherent risks. This is because the effect of these risks should already be incorporated into the model results, to the extent that water utilities are properly included in the model proxy groups. This is affirmed by statements ORA cited in its testimony. For example, the Middlesex Water Chairman of the Board and President's annual report was quoted as saying that "The weather played a significant role, as the lack of rainfall made the second half of 1998 the driest in New Jersey in over 100 years." Also, the Connecticut Water 1998 Annual Report to Shareholders stated that the Company's profitability is primarily attributable to the sale and distribution of water, the amount of which is dependent on seasonal weather fluctuations.⁴⁴

Clearly, weather has an impact on water utilities and is reflected in their financial statements and risk assessments by investors. Weather variation is also one of the reasons California regulated water utilities are able to recover 50% of their fixed cost through a service charge and authorized balancing and memorandum accounts.

Resolution W-4294 is a recent event. Order Instituting Rulemaking (OIR) 01-12-009 through the Commission is exploring whether purchased power and water costs should become contingent expenses afforded balancing account treatment only if the utility is earning a weather-normalized ROE less than its authorized ROE or if actual purchased power and water costs were less than anticipated.

AVR sees the Commission's pending OIR decision on Resolution W-4294 as imposing additional risk on AVR because its authorized ROE would be treated as a ceiling rather than a target ROE where it may earn more or less that its authorized ROE. We disagree, as did ORA. Irrespective of AVR's premature position on this issue, to the extent a decision is issued providing the restrictions AVR believes will be imposed, that condition would not restrict AVR from earning its authorized ROE.

Although Park's approximate 60% equity ratio is slightly higher than the average of AVR's proxy groups and the 48% average of ORA's proxy group, AVR still has a limited source of external financing and its stock is still not publicly traded, justifying a premium ROE. The evidence in this proceeding continues to support a 30 basis point premium ROE for AVR.

After considering all the evidence on the quantitative financial models based on subjective inputs, risk factors, limited source of external financing non-publicly traded stock, interest rate trends, current economy, and our informed judgment, we authorize AVR an 10.10 % ROE for its test years. This ROE is based on the lower second quarter (or 9.80%) of the 9.21% to 11.22% ROE range found reasonable for AVR and a 30 basis point ROE premium for added risk perceived by investors. This constant ROE equates to a 9.49% return on AVR's test year 2003 rate base, 9.48% for test year 2004 rate base, and 9.47% for attrition year 2005 rate base as follows.

	Capital Ratio	Cost Factor	Weighted Cost
TEST YEAR 2003 Long Term Debt Common Equity Total	40.09% 59.91 100.00%	8.58% 10.10	3.44% 6.05 9.49%
TEST YEAR 2004 Long Term Debt Common Equity Total	40.09% 59.91 100.00%	8.56% 10.10	3.43% 6.05 9.48%
Attrition Year 2005 Long Term Debt Common Equity Total	40.09% 59.91 100.00%	8.54% 10.10	3.42% 6.05 9.47%

When we review the historical authorized ROEs for California's Class A water utilities we find the adopted 10.10% ROE for AVR is within a reasonable range of ROEs recently authorized other California Class A water utilities. An exception to this observation is the 10.25% ROE authorized California-American Water Company (Cal-Am) in D.03-02-030. However, that higher rate was authorized because no party made any reasoned analysis on the record that would lead to a lower ROE even though subjective judgment alone justified a lower ROE. The adopted ROE is also below recent ROEs authorized the more risky energy utilities with substantially less equity.⁴⁵