Costs and Funding Sources - Project 14318

CalWater sought to include $406,350 for Project 14318 in its 2005 plant in service, and thus in rate base beginning in the first test year. DRA proposed to exclude it from rate base but would allow CalWater to file an advice letter after the final costs have been accounted for. In its opening brief, CalWater agreed to advice letter treatment. DRA would cap the project's GRC cost at its $227,800 estimate; CalWater's closing brief is silent on Project 14318, stating only that the reasonable costs of all plant construction projects should be included in rate base. DRA's figures are the more recently prepared, better reflect both the cost and tentative sources and availability of funds, and present a more consistent, credible overall picture. As such, we accept them for our purposes today. We will authorize CalWater to file an advice letter to recover the incremental revenue requirement associated with the project, including up to $227,800 in company-funded plant costs (i.e., exclusive of SRF and SDWBA loan funding) when Project 14318 and its companion projects discussed below are completed and placed in service.

CWS asked to include $242,800 in plant for Project 14319 in its late-amended application.²⁶ Because the project was said to have been completed but the accounting was not closed, CalWater was unable to provide the parties or this record with sufficient information for a proper review. DRA's initial proposal was advice letter recovery when all the costs are in, with a cap of $239,800 in plant cost to reflect CalWater data responses that indicated $114,250 in direct expenses and $125,950 in unexplained CalWater labor and overhead charges. When CalWater was unable to provide a clear and consistent explanation before and during the evidentiary hearings, DRA revised its advice letter recovery cap downward, to $114,000. In support, DRA points to the many potentially serious issues raised by intervenors Young and Pareas and outlined briefly below.

Intervenors Young and Pareas take issue with numerous aspects of Projects 14318 and 14319. Among them are the fact that these projects and others were awarded to a single contractor at prices sometimes far in excess of CalWater's own estimates, without obtaining competing bids and in apparent violation of its own competitive bidding guidelines. They also allege that CalWater inappropriately repackaged its projects to keep the amount of low cost SRF funding below $500,000 in an admitted effort to avoid federal women and minority owned business contracting requirements, substituting instead higher cost capital at ratepayers' expense. Young, Pareas and DRA question whether some specific plant items included in these projects may be unnecessary, or their costs inflated or double counted from earlier Project 8087. DRA gives examples of specific suspected overruns: treatment plant inside piping increasing to $34,000 from $15,000; external piping to $77,000 from $40,000; electrical costs to $135,850 from $50,000; and the treatment plant building increasing to $200,000 from $27,000. For some items, CalWater had plausible explanations during the evidentiary hearings, but for others it did not, and it could not provide a consistent, understandable summary of these projects' costs overall.

In the end, Young and Pareas argue for strict limits on recovery from customers, through either immediate disallowances or caps on later advice letters. Both, along with DRA, advocate a disallowance equivalent to the excess cost ratepayers would bear as a result of CalWater's decision to forego SRF funds in excess of $500,000. As Young points out in his closing brief, the Commission left the door open to just such an outcome when it addressed his similar arguments in CalWater's 2002 Coast Springs GRC:

Young argues that should SDWSRF funding for the Coast Springs and Lucerne treatment plants prove unavailable, Cal Water should bear the burden of proving that the cause of the denial was solely due to [Department of Health Services] or others. The argument is premature. Issues regarding ratemaking for these projects, if SDWSRF funding is denied, can be explored in an application CalWater may file to request such ratemaking.²⁷

However, no party developed data in the record here to calculate an SRF-based capital disallowance, and in any case we would first need to decide the level of plant investment allowed against which to apply the SRF loan criteria.

We conclude that CalWater has failed to demonstrate the reasonableness of its plant in service estimates for Projects 14318 and 14319. DRA, Young and Pareas have shown a likelihood that, when the projects are completed and a full cost accounting is made available, some part of CalWater's costs for these projects will be found to be unnecessary and unreasonable. In the meantime, we will adopt DRA's recommendation and allow CalWater to file a single offset advice letter when all three projects (8087, 14318, and 14319) are complete and in service, and all of their associated costs properly booked. For now, DRA's estimates of these projects' costs are more reliable for that purpose than are CalWater's. The maximum company-funded plant costs (i.e., exclusive of SRF and SDWBA loan funding) will be capped at $227,800 for Project 14318 and $114,000 for Project 14319. Young's concern that finalizing the rate base value through an advice letter filing could deprive him and other customer representatives of sufficient notice and opportunity to participate has merit. Given the amounts at stake in relation to the small number of customers in the Coast Springs system, adjudicating the issue is likely to be highly contentious and beyond the proper scope of a simple advice letter filing. Accordingly, the interim, capped rate base amounts determined in the advice letter filing will become final only after the next CalWater Coast Springs GRC, unless CalWater and/or its counterparties there present persuasive evidence supporting a different final outcome.

Young and Pareas take issue with two other Coast Springs projects: a glass-fused storage tank, and a solar-powered pump to prevent storage tank nitrification.

In CalWater's last Coast Springs GRC, it projected adding a 125,000 gallon steel storage tank in 2003 at a cost of $104,800. That tank was never installed, and CalWater now seeks approval instead to include in rates a glass-fused 125,000 gallon tank, competitively bid, at a cost of $216,840. CalWater's witness cited the company's experience with a previous steel tank in the area that failed after only 15 years, and another that required extensive maintenance at a cost of $75,000 to $80,000 after about five years. CalWater claims that its glass-fused tank will stand up better to the harsh coastal weather conditions in the area and cost less in the long run than the steel tank would have. Young, supported by Pareas, takes issue with that conclusion and advocates limiting CalWater's recovery in rates to the earlier $104,800 estimated cost for a steel tank. Young believes glass-fused tanks are costly, untested technology for water storage in the coastal area; CalWater has no other glass-fused tanks in its California service territories and no actual experience by which to judge their expected performance, and the steel vs. glass-fused tank economic comparison study CalWater conducted was faulty. Young's cross-examination of two CalWater witnesses pointed out several shortcomings in CalWater's economic comparison, but he did not establish whether a more complete study would have tipped the balance to favor steel. CalWater claims to have been favorably influenced by a glass-fused tank on a farm in the area, but that tank was used for silage rather than water storage. The parties sparred over which use was the more damaging, and the role cathodic protection might play in prolonging tank life.

Young and Pareas fail to convince us to penalize CalWater for its selection of tank type. Absent solid proof to the contrary, choices of what equipment and materials will perform best in utility water systems are best left up to the engineers and other water company experts to whom that responsibility has been assigned.

Pareas advocates disallowing the $10,800 cost of a solar-powered pump installed to prevent nitrification in the Ravine Tank. Pareas contends the pump is useless because solar panels are ineffective in Coast Springs' foggy coastal climate. CalWater responds that it used a solar contractor who resides in the local area and was experienced with local conditions, and the pump has proved to be an economic and effective choice. Young pointed out that the pump cost had apparently been included in two plant projects. CalWater confirmed that it had erroneously been included twice, but that error had been corrected and the pump is now only in Project 12505. We conclude that the solar-powered pump is used and useful and should be allowed in plant for ratemaking.

Finally, Pareas would have the Commission order CalWater to retain an independent firm chosen by others "to analyze the validity of the top dollar paid on CWSC's projects at Coast Springs." While we agree with Pareas that CalWater's Coast Springs plant showing left much to be desired, that does not lead us to conclude that ordering an outside audit would be justified. On the contrary, we believe all of the opposing parties involved in the Coast Springs rate area issues were very effective in making their cases, as evidenced by the caps we have placed on CalWater's major Coast Springs construction projects for this rate case cycle. DRA's plant witness, Clement Lan, is a licensed engineer and highly qualified and experienced for this undertaking. We were impressed by his thorough, competent and comprehensive plant and rate base showing for this district and others. CalWater would undoubtedly seek to recover from Coast Springs ratepayers the additional costs of an outside auditor, and we are mindful that Coast Springs is a small district with already-high rates. Instead, we cap the plant expenditures we allow in rates for this GRC cycle, and put CalWater on notice that we will require a persuasive showing that any Coast Springs plant costs above those caps were reasonable and not reasonably avoidable before we will allow them in future rates.

Cost of Capital

In order to determine a fair rate of return for a utility, we project the proportions of debt and equity in its capital structure, estimate what the effective cost of each will be, and take a weighted average. The resulting rate of return is used to determine the revenue requirement in the summary of earnings for each test year.

Capital Structure and Cost of Debt

While capital structure and cost of debt are not mentioned in the DRA/CalWater stipulation on remaining issues, they agreed on both at the time of the evidentiary hearing. CalWater estimated it would issue $20 million in new debt in 2005, $15 million in 2006, $30 million in 2007, and $20 million in 2008. For 2005 through 2009, it projected the cost of new debt to be 6.59%, 7.13%, 7.61%, 7.85% and 8.14% respectively. It had no plans to issue or retire preferred stock during the period. DRA's cost of capital witness reviewed the embedded cost of debt and preferred stock and the projected new debt issuances set forth in the company's applications and found them reasonable. The resulting capital structure is reflected in our adopted rate of return for each year, Table 3 below.

Cost of Equity

Cost of equity is typically the most contested component of rate of return in water GRCs. It is a direct measure of the company's after-tax return on equity investment (ROE), and its determination is by necessity somewhat subjective and not susceptible to direct measurement in the same way capital structure and embedded cost of debt are.

We have many times over the years cited the well established legal standard for determining a fair ROE. In the Bluefield Water Works case,²⁸ 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. 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 of return issue in the Hope Natural Gas Company case,²⁹ 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."

The Court went on to affirm the general principle that, in establishing a just and reasonable rate of return, consideration must be given to the interests of both consumers and investors.

CalWater and DRA each made a showing to support an ROE recommendation. With the principles above in mind, we first describe the methods each used, and then discuss our evaluation of them.

CalWater's Recommended Return on Equity

CalWater used DCF (discounted cash flow) and RP (risk premium) models to determine its recommended ROE, but ran them on different, more varied sets of data than did DRA.³⁰ CalWater also made various adjustments that DRA did not. Using data available in mid-2005, CalWater prepared a single DCF estimate (based on a sample of five water utilities and five gas utilities, all said to be comparable to CalWater for ROE purposes), and a single RP analysis based on the same group of comparables. CalWater's witness averaged the 12.11% ROE from the DCF analysis with 11.79% from the RP analysis, and then added 28 basis points (0.28%) to recover what he calculated was the company-wide average balancing account losses due to overearnings in 2002, 2003 and 2004, to generate a final recommended ROE of 12.23%.

In support, CalWater cites risks due to tightening of water quality standards in recent years, lowered credit ratings in November 2002 and February 2004,³¹ failure to earn its authorized return in recent years,³² potential electric power shortages in California, an unfavorable association in investors' minds between California's water utilities and the extreme financial distress of its largest energy utilities earlier in the decade, and a host of other factors.

DRA's Recommended Return on Equity

DRA also used DCF and RP models to estimate investors' expected ROE, and applied both models to a group of comparable water utilities selected based on two criteria: (1) water operations account for at least 70% of their revenues, and (2) their stocks are publicly traded. The comparable group comprised five companies: American States Water, Connecticut Water Service, Middlesex Water Service, Philadelphia (Aqua American), and San Jose Water. DRA's DCF analysis yielded an average expected ROE of 9.35%. Its RP analysis produced 10.22%. It averaged the two results to produce its 9.78% final recommended ROE for CalWater.

DRA concluded that CalWater's business risk, which DRA related primarily to regulatory risk, was low, citing the Commission's many risk-reducing mechanisms available to water utilities. Those mechanisms include balancing accounts (now redesignated as balancing-type memorandum accounts) for purchased water, purchased power, and pump taxes; memorandum accounts for catastrophic events and waste contamination and for Safe Drinking Water Act compliance; 50% fixed cost recovery in the service charge; and construction work in progress in rate base.³³ DRA also noted that CalWater's average equity ratio during the test years will be higher than the comparable group's average, and its debt ratio lower, factors which tend to reduce financial risk. In the end, however, DRA's witness characterized CalWater's financial risk as "high," based on his belief that CalWater held a Standard & Poor's bond rating of B.³⁴

Return on Equity Discussion

Both CalWater and DRA recommend ROEs generated mathematically from their evaluation of the same two models, DCF and RP. We are left to sort through their underlying assumptions, data sets and biases to ferret out the root causes of their considerably different ROE results (12.23% versus 9.78%).

Several difficulties leap immediately to the fore in CalWater's calculations: The inappropriate use of gas utilities in the comparables group, biased culling of intermediate results that would otherwise work against its interests, extreme variations in some of the data sets it chose, and two arithmetic errors that cause large shifts in its favor in the final result. We also agree with CalWater where it points to the need to make a series of corrections to DRA's figures.

We begin by noting our disappointment that CalWater has relied so heavily in its analysis on comparisons with gas utilities, a practice the Commission has repeatedly rejected in the past because water utilities are less risky.³⁵ A simple inspection of the data in CalWater's Table 4-1 demonstrates that the gas utilities it selected to include in the comparables group have in the aggregate much higher common equity ratios than the corresponding water utilities (average 55.7% vs. 49.5%), very much higher returns on common equity (average 13.5% vs. 9.1%), and higher authorized returns on equity (average 11.7% vs. 10.6%). Each of these differences illustrates that gas and water utilities are not comparable in important ways that matter here. We would consider backing out the gas utility data to see just how much they inflate the resulting ROE, but CalWater has made that impossible by aggregating gas and water companies in one key analysis sheet, the historical growth rates in Table 4-3.

Next, we note CalWater's culling figures from its calculations where they did not promote the end it desired. The single most influential figure in the DCF model is the estimated growth rate. Both CalWater and DRA properly tabulated earnings growth, dividend growth, and sustainable growth percentages (CalWater's Table 4-3) for their comparable groups for 1995 through 2004. CalWater found two of those three measures to be markedly lower than the third, whereupon it discarded the entire data set (as opposed to averaging the figures to produce a seemingly reasonable result as DRA did), saying the low figures produced "implausible estimates" of the ROE.³⁶ It then proceeded to rely entirely on the simple average of four analysts' subjective growth rate projections for its ten comparable utilities (gas and water) for an unspecified period. Even in the face of extreme variations (one analyst estimated those growth rates between minus 6% and plus 32.9% while the others' estimates were generally confined within a reasonably narrow range), CalWater accepted subjective figures that suited its desired end (e.g., the plus 32.9% growth rate estimate was included, while a minus 18% growth rate was omitted because it "would skew data to unreal averages").

Lastly, CalWater's cost of capital showing included two very significant arithmetic errors in calculating the average current dividend yields for its ten comparable utilities.³⁷ Correcting those two errors alone would have reduced its DCF ROE result by approximately 0.8%.

For its part, CalWater recommends a series of revisions to DRA's DCF and RP estimates. We summarize here the four changes we agree should be made.

First, CalWater points out that in calculating its current annualized dividend yield, DRA used dividends from 2004 and stock market prices during September 2004 through August 2005. Realigning and recalculating as CalWater recommends boosts each current yield figure by approximately 0.1%.³⁸

Next, CalWater notes that DRA's average year forecast of interest rates uses 2005 through 2008 estimated data rather than 2006 through 2009 to most closely match this GRC cycle. We agree, but note that the 2009 data is not in the record. We will revise the calculation to include only the 2006 through 2008 data we do have available, thus raising the interest rate forecast by approximately 0.2%.³⁹

We accept CalWater's suggestion to remove historical dividend growth from the overall average growth rate calculation, for some of the reasons CalWater states on rebuttal.⁴⁰ We also note that DRA used 11 years' of data to calculate its 10-year average, so we make that correction as well. These two changes raise the overall average growth rate by approximately 0.9%.

The result of these changes is to raise DRA's DCF ROE result to 9.91% from 9.35%, the RP ROE to 10.41% from 10.22%, and the average ROE to 10.16% from 9.78%.

After evaluating the parties' ROE presentations, we find that CalWater's analysis was incorrect and less credible than DRA's, and produces results that fail a reasonability test. In addition to the infirmities we note above, investors in today's economic climate and over the forthcoming rate case cycle will not, as CalWater claims, require it to earn after-tax returns at or above 12% to purchase its debt offerings or make equity investments in it. Thus, we will adopt the 10.16% ROE that results from our revisions to DRA's analysis, as set forth above. This represents a small increase over the 10.10% ROE the parties settled on and we adopted for CalWater in its last round of GRCs one year ago.⁴¹

Rate of Return on Rate Base

With the capital structure, cost of debt, and cost of equity components defined, the straightforward calculations in Table 3 derive the rates of return on rate base for each test year:

²⁶ DRA opening brief, page 4, citing CalWater Exhibit P-RV-CS, page 3.

²⁷ D.04-03-040, Section IV.B.

²⁸ Bluefield Water Works & Improvement Company v. Public Service Commission of the State of Virginia (1923) 262 US 679.

²⁹ Federal Power Commission v. Hope Natural Gas Company (1944) 320 US 591.

³⁰ The DCF model is a financial market value technique based on the premise that the current market price of a share of common stock equals the present value of the expected future stream of dividends and the future sale price of a share of stock, discounted at the investor's discount rate. By translating this premise into a mathematical equation, the investor's expected rate of return can be found as the expected dividend yield (the next expected dividend divided by the current market price) plus the future dividend growth rate.

The RP model is a risk-oriented financial market value technique which recognizes that there are differences in the risk and return requirements for investors holding common stock as compared to bonds. An RP analysis determines the extent to which the historical return received by equity investors in utilities comparable to the utility at issue exceeds the historical return earned by investors in stable, long-term bonds. This difference, or "risk premium," is then added as a premium to the estimated cost of long term debt to derive average expected return on equity for the test period. (D.03-02-030).

³¹ CalWater's backup documentation shows that Standard & Poor's gave CalWater an improved outlook rating in January 2005, to "stable," based in part on "its regulatory insulation from competition... and the relatively low operating risk of managing water utility systems." Standard & Poor's also credited large equity issuances in 2003 and 2004 for "bringing debt leverage comfortably within the A+ rating benchmark" (during 2003 and 2004, CalWater increased its equity ratio dramatically, to 50.2% from 41%). On the down side, Standard & Poor's cites CalWater's reliance on rate relief from the Commission. A Value Line report issued that same month cited some of the same factors for the water industry generally, and singled out CalWater thusly: "More risk-averse individuals may find added appeal in California Water, given its 2 (Above Average) rank for safety." (Exhibit L, page 8 and Enclosures 6 and 10).

³² CalWater's cost of capital witness claimed that the company earned far below its Commission authorized ROE each year from 2001 through 2004. In the same exhibit, he cites approximately $3.5 million in lost balancing account revenues in 2002, 2003 and 2004 through the Commission's application of the earnings test. There was no explanation for this seeming contradiction, except that the witness may not understand the mechanics of the Commission's balancing account procedures. In an attempt to bolster CalWater's regulatory risk profile, his prepared direct testimony four times mischaracterizes the working of the D.03-06-072 earnings test as placing a upper limit on water utility earnings (Exhibit L, pages 15, 16, and 22; and Exhibit CWS-5, page 11). That is not the case, as we explained when we finalized those procedures: "The proposal will not deny the utility a right to earn a fair rate of return on investment.... [T]he revised procedures will permit the utility to earn at least up to its authorized rate of return, and even more than the authorized rate of return through any means other than the collection of these balancing-type memorandum accounts." (D.03-06-072, Section V.B.2. Emphasis added).

³³ DRA's evaluation of business risk is consistent with our own observation in CalWater's 2005 GRC: "[W]e note that many aspects of this decision provide significant protections to CWS against erosion of earnings, including the use of recent expense estimates, provision for future advice letter filings regarding major plant additions and expenses, and allowance for escalation effects in escalation years 2006-2007 and 2007-2008. In addition, CWS is protected through separate balancing accounts for purchased water, purchased power and pump taxes, and memorandum accounts for catastrophic events and waste contamination. The result of these protections is to reduce the risk that CWS faces with regard to its opportunity to earn its return on equity. Consequently, we expect that in future proceedings all of these existing and adopted protections against erosion of future earnings will be given their proper weight in the determination of risk and consequently return on equity." (D.05-07-022, Section VII.G.).

³⁴ Exhibit DRA-9, pages 3-3 and 3-4. Note, however, that may not be the case. CalWater's cost of capital witness testified that its debt was rated by Standard & Poor's as A+. (Exhibit L, page 6, reiterated in Exhibit CWS-5, page 2).

³⁵ See, e.g., D.90-02-042, D.92-01-025, D.01-04-034, and D.04-05-023.

³⁶ Exhibit L, page 13. Returning those figures to the calculation in Table 4-5 would decrease CalWater's DCF ROE result from 12.11% to approximately 9.1%.

³⁷ Exhibit L, Table 4-2. The average 6-month and 12-month dividend yields are arithmetically incorrect. Those errors carry over to the final DCF ROE result on Table 4-5.

³⁸ Exhibit CWS-5, CalWater's cost of capital rebuttal testimony, pages 4 and 5; and Exhibit DRA-9, DRA's cost of capital report, Table 2-2.

³⁹ Exhibit CWS-5, page 5; and Exhibit DRA-9, Table 2-6.

⁴⁰ Exhibit CWS-5, page 6; and Exhibit DRA-9, Table 2-3.

⁴¹ D.05-07-022 (July 21, 2005), as modified by D.05-11-022 (November 18, 2005).