8. Cost of Debt and Capital Structure
Ratemaking capital structure is long-term debt, preferred stock, and common equity.15 Because the level of financial risk that the utilities face is determined in part by the proportion of their debt to equity capital, or the degree of financial leverage, we must ensure that the utilities' adopted equity ratios are sufficient to maintain reasonable credit ratings and to attract capital without incurring unnecessary costs for an excessive amount of expensive equity.
Generally, long-term debt is the least expensive form of capital but the utility must ensure that it timely meets every interest payment and maintains any required terms or conditions of the loan agreements or mortgage indentures, and that it can refinance or refund the debt when it matures. Preferred stock is generally more expensive than debt and may or may not have a maturity or refund provision. Interest may usually be deferred but it then accumulates and takes preference over payment of dividends to common equity owners. Thus, equity owners assume more risk than either debt holders or preferred stock owners, including the risk of losing their entire investment, and therefore equity investors require the highest return over the long run.
The equity returns adopted in this proceeding are summarized below and discussed later. We are not persuaded to use an average capital structure to reflect the proxy groups as proposed by DRA (DRA Opening Brief at 45) nor will we be beguiled by a false precision of four decimal places. Therefore, we adopt and round to the nearest full percentage the proportion of debt requested by each applicant. As also noted in this decision, the proportion of equity has a direct bearing on risk, and we consider the assertions of company-specific risk (and requests for as much as a 90 basis point "adder" by Park/Apple16) in determining the fair return on equity for each company. We do find that the companies reasonably calculated their actual and 2010 forecast embedded costs of debt and preferred stock and therefore we adopt (after rounding) Applicants' 2010 capital structures and embedded cost. In adopting these costs and ratios, we note, as discussed elsewhere, our long-term concern over the appropriate range of equity ratios for regulated Class A water utilities.
We noted our concern recently when a company has a high equity ratio:
We find equity components [for large Class A water companies] in excess of 50% to be problematic and have concerns about equity ratios less than 45%. It is this Commission's responsibility to establish a safe range within which a company's capital ratio may move and against which the cost of capital may be measured. In [A.08-05-002 et al.], there is a significant cost differential, compounded by the tax consequences of equity. (D.09-05-019 at 9.)
We again address the issue, noting in particular that San Jose is much larger than the other companies in this consolidated proceeding and closer in size to the three companies addressed in D.09-05-019. Further, Valencia, Park/Apple, and Suburban are very much smaller than San Jose while
San Gabriel, is in-between in size. However, as a general rule, as the companies become smaller, we are more likely to find persuasive financially sound cost-based justifications for a higher equity ratio, as discussed in more detail below.
DRA argues that we should adopt an adjusted capital structure to reflect the return on equity derived from its proxy groups-that is, fit Applicants to the proxy rather than adapt the proxy results to Applicants. DRA proposed:
DRA has developed capital structures for the water companies that reflect both the individual company capitalizations as well as those of the proxy group of publicly-held water companies. This is necessary since the capitalizations of the water companies have higher common equity ratios than the companies in the proxy group which are used to determine [DRA's proposed baseline] equity cost rate of 9.75%. (DRA Opening Brief at 45.)
We believe that the companies do have a significant control over the mix of debt and equity and that ratepayers should not bear unnecessary costs as a result of management discretion-the equity return is a market return for the assumption of like-risk in comparable investment choices. Thus, even if, for the sake of argument, a smaller water company as a matter of course has a higher transaction cost and even a higher interest cost for debt, debt is almost always cheaper than equity and does not carry the added loading of an income tax allowance in rates. Therefore, if a company carries a high equity ratio, we should necessarily consider adjusting the return on equity. We are concerned here, however, that DRA is adjusting capital structure to fit its proxy group equity return recommendation rather than adjusting the study result to the companies' capital structures. DRA does not argue that the companies' proposed structures are inherently wrong, but explicitly argues for the change to fit its equity return recommendation. We reject such an adjustment here for all of the companies. As discussed elsewhere, we will, however, consider the actual equity ratios as a part of adopting a final return on equity.
For Valencia we adopt a 2011 capital structure using its recent long-term debt, which substantially lowers its equity ratio going forward. We use 10.20% on equity subject to any adjustment due to the attrition mechanism adopted in this decision. Valencia issued $12 million at a cost of 4.62%.17 This doubled the debt ratio and we compute a weighted cost of debt of 6% (the average of the 2010 weighted cost of 7.37% plus the new debt's cost of 4.62%).
Valencia 2011 Cost of Capital | |||
Ratio |
Cost |
Weighted | |
Debt |
46% |
6.00% |
2.76% |
Preferred |
2% |
9.50% |
0.19% |
Equity |
52% |
10.20% |
5.30% |
Total |
100% |
8.28% | |
8.1. Costs of Long-Term Debt and Preferred Stock
Long-term debt and preferred stock costs are based on actual, or embedded, costs. Future interest rates must be anticipated to reflect projected changes in a utility's cost caused by the issuance and retirement of long-term debt and preferred stock during the year. This is because the rate of return is established on a forecast basis.
We recognize that actual interest rates do vary and that our task is to determine "reasonable" debt cost rather than actual cost based on an arbitrary selection of a past figure.18 In this regard, we conclude that the latest available interest rate forecast should be used to determine the forecast of additional debt included in the embedded debt for the forecast period. (See recently, D.07-12-049, and 38 CPUC2d 233, where 18 years ago, the Commission definitively discussed the need for, and use of, a reliable forecast of future interest costs.)
8.1.1. DRA's Proposed Cost of Debt
DRA used the proposed debt cost rates from Applicants for the rate year 2010. It argued that to a large extent, forecasts of future interest rates are not accurate, and therefore the projected debt cost rates beyond 2010 are not useful. Thus, it would be improper to rely on the utilities' proposed debt cost beyond 2010. For example, DRA indicates that San Gabriel did not have any bids at the time it served testimony or quotes for pricing out its anticipated debt issues. (Ex. SG-3, Table B, at 2 and 4.) DRA also notes that San Gabriel historically used a spread of 246 basis points to forecast its long-term debt rates for anticipated issues.19 DRA notes that San Gabriel imputed 492 basis points in forecasting the interest rate for the mortgage bonds it plans to issue in 2010 and 2012.
DRA argues that its testimony (Exhibit DRA-1, Page 2 of Attachment JRW-3) shows the spread for Utility BBB rated bonds peaked at 450 basis points in December 2008 and as of July 2009 were 250 basis points above treasury rates. DRA argues the financial markets have stabilized since the peak of the crisis and spreads are continuing to move downward towards historical trends.
DRA's review of the 2010 through 2012 forecasted spreads between
Baa Corporate Bonds and 30-Year Treasury Bonds is based on the May 2009 Global Insight Forecast. As shown in Exhibit DRA-1, Attachment JRW-22, the average spread over the three-year period (2010-2012) is 334 basis points based on the most recent Global Insight forecast. Also, the most recent Federal Reserve data shows that as of June 2009, the spread between Baa Corporate Bonds and 30-Year Treasury Bonds reached 300 basis points. (Attachment JRW-23.) This is a 256 basis point drop since its peak of 556 basis points reached in December 2008. This spread has continued to drop. As of August 2009, the spread between the Baa Corporate Bonds and 30-Year Treasury Bonds yields was 221 basis points. DRA requests, and we grant, that the Commission take notice of this information.
Other utilities have recently issued secured debt issues with spreads significantly lower than San Gabriel's proposed 492 basis points. For example, in June 2009, Valencia issued $12 million in 30-year Senior Secured Notes at a rate of 7.73%,20 a spread of 321 basis points based on the historical June 2009, 30-Year Treasury Bonds yield of 4.5%. Also, Park issued two new first mortgage bonds in June 2008 with spreads of 285 and 300 basis points.
Thus, if San Gabriel's spread is just slightly higher than its historical 246 basis points, recalculating the company's weighted debt cost results in an average debt cost of 7.55% over the three-year period (Attachment JRW-24), compared to San Gabriel's 7.81%. Therefore, DRA continues to assert that the weighted average debt cost for the utilities should be based in this case on 2010 projections rather than relying on forecasts beyond this period, which are inaccurate and will change substantially.
8.2. Discussion
We agree with DRA that the utilities' projected interest rate spreads beyond 2010 are very high and rely on the early impacts of the financial market crisis. As noted elsewhere, we grant Valencia's motion to use its large 2010 debt issuance to adjust its weighted cost and debt ratio for 2011.
8.3. Actual and Imputed Capital Structures
DRA proposed capital structures for the water companies that it argued reflected both the individual company capitalizations as well as those of the proxy group of publicly-held water companies. This is necessary according to DRA because the capitalizations of the water companies have higher common equity ratios than the companies in the proxy group used to determine DRA's recommended equity cost rate of 9.75%.
The companies used their existing capital structures and forecast likely retirements, refinancing, and new debt issues for 2010 through 2012. We find the Applicants' forecasts of debt and preferred costs for 2010 to be reasonable.
We are concerned, however, that Valencia has a very high 2010 equity ratio over 70% and San Gabriel comes in over 60%. These high equity ratios significantly drive total cost to ratepayers higher because of both the higher return applied to equity over debt and the required allowance for income taxes. We have noted this before when we stated that we "find equity components in excess of 50% to be problematic and have concerns about equity ratios less than 45%." (D.09-05-019 at 9.) We therefore expect these two companies to actively look for ways to lower their equity ratio before their next cost of capital proceeding and for all of the applicants to make a substantial showing to justify their proposed capital structures in their next cost of capital proceeding. In fact, Valencia has lowered its equity ratio by a large debt issuance and we use this to reflect the actual structure and cost on 2011. (See June 10, 2010 Motion.) We will not impute a different capital structure at this time and we will not impute DRA's proxy structure based on the very limited justifications offered in this record.
15 Short-term debt due within one year is excluded.
16 Park/Apple Opening Brief at 1.
17 June 16, 2010 Motion at 2, at Table 1, and attached Declaration.
18 38 CPUC2d 233 at 242 and 243 (1990).
19 DRA citing to Ex. SG-3 at 5.
20 Exhibit VW-1 at 3.
