20040712Avera Rebuttal.pdf

-- f- '"
. .
\ J Hc.L-. u
'--"
f'lL_-
:~-
!LED
DAVID J. MEYER nrh
. -
uUt\.J-' HIII-VICE PRESIDENT AND CHIEF COUNSEL OF
REGULATORY AND GOVERNMENTAL AFFAIRS;;) j U;3 LieVISTA CORPORATION U T It I TIE S COt-if"!! S 51 ON
O. BOX 3727
1411 EAST MISSION AVENUE
SPOKANE, WASHINGTON 99220-3727
TELEPHONE: [phone]
FACSIMILE: [phone]
BEFORE THE IDAHO PUBLIC UTILITIES COMMISSION
IN THE MATTER OF THE APPLICATION
OF A VISTA CORPORATION FOR THE
AUTHORITY TO INCREASE ITS RATES
AND CHARGES FOR ELECTRIC AND
NATURAL GAS SERVICE TO ELECTRIC AND
NATURAL GAS CUSTOMERS IN THE ST ATEOF IDAHO
CASE NO. AVU-04-
CASE NO. A VU-04-
REBUTTAL TESTIMONY
WILLIAM E. AVERA
FOR A VISTA CORPORATION
(ELECTRIC AND NATURAL GAS)
II.
ID.
IV.
CONTENTS
Section Pa2e
INTRODUCTION............................................. ....................................................
TERRI CARLOCK ......
..................... ............ ................... ............... .... ..................
DENNIS E. PESEAU ...........................................................................................
JOHN S. THORNTON, JR. ................................... ............ ................. ..................
INTRODUCTION
Please state your name and business address.
William E. Avera, 3907 Red River, Austin, Texas, 78751.
Are you the same William E. Avera that previously submitted direct
testimony in this case?
Yes, I am.
What is the purpose of your rebuttal?
The purpose of my testimony is to respond to the direct testimony of Ms. Terri
Carlock, submitted on behalf of the staff of the Idaho Public Utilities Commission ("IPUC"
In addition, I will also rebut the recommendations contained in the direct testimony of Dr.
Dennis E. Peseau and Mr. John S. Thornton, Jr., on behalf ofPotIach Corporation, concerning
the cost of equity for the jurisdictional utility operations of Avista Corporation. ("Avista
Please summarize the conclusions of your testimony.
With respect to the testimony of Ms. Carlock, I concluded that her
recommendations were biased downward because of her failure to consider the results of
other accepted methods of estimating the cost of equity. Additionally, Ms. Carlock'
assessment of relative risks focused exclusively on Avista s relatively low rates, while
ignoring the substantial uncertainties and higher investment risks that investors must bear to
provide the benefits of lower electricity costs to Avista s customers. Finally, her flotation cost
adjustment understates the costs necessary to raise the equity capital invested in Avista
jurisdictional utility operations in Idaho. At a minimum, considering the results of risk
premium approaches, investors' risk perceptions, and correcting Ms. Carlock's flotation
Avera, Di - Reb
A vista Corporation
adjustment would support a rate of return at the very top of the range of her results, or 11.3
percent.
Meanwhile, Dr. Peseau did not conduct any independent analyses of the cost of equity
to Avista. Instead, his recommendations were based entirely on flawed "updates" and
revisions" to my analyses, which should be rejected in their entirety. Similarly, Mr.
Thornton s recommended 8.5 percent cost of equity is woefully inadequate and, by any
reasonable benchmark, falls well short of investors' required rate of return from an electric
utility, especially considering Avista s unique risks and weakened credit standing. Mr.
Thornton s recommendations do not pass the financial end-result test fundamental to
regulation and would preclude Avista from restoring its financial integrity and attracting
capital on reasonable terms.
Q. Would you please summarize the principal shortcomings in the testimony
of Ms. Carlock, Dr. Peseau, and Mr. Thornton that you address in rebuttal?A. Yes. The major issues addressed in my rebuttal testimony are as follows:
Ms. Carlock
While the risks premium approach is widely recognized as a meaningful approach
to estimate the cost of equity, Ms. Carlock did not use this method;
. No methodology provides a foolproof guide to investors' required rate of return
and it is important to consider alternative approaches and evaluate the results of
accepted methods;
The results of risk premium analyses are consistent with a rate of return at the top
of Ms. Carlock's discounted cash flow ("DCF") and comparable earnings ranges;
Ms. Carlock's recommendation does not fully reflect the investment risks
associated with A vista s weakened credit profile and exposure to market
uncertainties;
The pre-tax coverage ratio implied by Ms. Carlock's recommendation is only
marginally above the minimum benchmark for a triple-B bond rating;
Ms. Carlock's flotation cost adjustment is biased downward and she failed to
adjust the results of her comparable earnings approach to incorporate issuance
Avera, Di - Reb
A vista Corporation
costs.
Dr. Peseau
Dr. Peseau performed no independent analyses of the cost of equity;
His decision to "update" my DCF analysis by ignoring historical growth trends is
unsupported and contradicts the advise and conclusions of his own sources;
In contrast to Dr. Peseau s allegations, there are no inconsistencies in my risk
premium analyses and his use of single-A bond yields as a benchmark for Avista
investment risks understates investors' required return;
Dr. Peseau did not update my application of the capital asset pricing model
CAPM"); instead, he substituted a market risk premium that does not reflect
expectations in today s capital markets; and
Dr. Peseau ignored Avista s greater investment risks and the need to adjust the
cost of equity to account for flotation costs.
Mr. Thornton
The extreme downward bias of Mr. Thornton s recommended cost of equity is
illustrated when compared against the returns on equity authorized by regulators
including the IPUC;
Mr. Thornton s recommendations are divorced from the requirements of real-
world capital markets and the inputs to his analyses do not reflect the expectations
of investors;
Mr. Thornton s criticisms of my analyses lack any reasonable basis, as does his
rejection of arithmetic mean returns and long-term bond yields in applying the
CAPM;
Like Dr. Peseau, Mr. Thornton ignored Avista s greater investment risks and the
need to adjust the cost of equity to account for flotation costs; and
Correcting Mr. Thornton s flawed calculations results in a coverage ratio that falls
below the minimum guidelines for an investment grade rating and demonstrate
that his recommendations would not allow A vista the opportunity to maintain its
financial integrity.
Avera, Di - Reb
A vista Corporation
II.TERRI CARLOCK
First, does the capital structure proposed by Ms. Carlock provide a
reasonable basis on which to calculate an overall rate of return for Avista?
Yes. Ms. Carlock recommended a capital structure composed of 50.08 percent
long-term debt, 5.57 percent trust preferred securities, 1.76 percent preferred stock, and 42.
percent common equity based on Avista s actual capitalization at December 31 2003. As
discussed in my direct testimony, the average capitalization for the firms in my comparable
group was composed of 44.7 percent common equity. Meanwhile, revised financial guideline
ratios published by Standard & Poor s Corporation ("S&P") imply a total equity ratio in the
range of 42 to 52 percent for Avista to qualify for a triple-B rating.1 Accordingly, I concluded
that the capital structure used by Ms. Carlock is in-line with industry standards.
How did Ms. Carlock arrive at her 10.4 percent cost of equity
recommendation for Avista?
Ms. Carlock estimated the cost of equity by applying the constant growth DCF
model directly to Avista. She concluded that the results of this single DCF application
indicated a cost of equity in the 8.8 to 11.3 percent range. Ms. Carlock also conducted a
comparable earnings analysis, which resulted in an indicated cost of equity in the 10.0 to 11.
percent range. Based on these two analyses, Ms. Carlock concluded that the cost of equity
was in the 9.5 to 10.9 percent range, selecting 10.4 percent as her point estimate and
recommendation for Avista.
1 Standard & Poor s Corporation
, "
New Business Profile Scores Assigned for U.S. Utility and PowerCompanies; Financial Guidelines Revised RatingsDirect (Jun. 2, 2004) at Table 1. For a utility with Avista
business profile ranking of ", S&P reported a guideline total debt ratio ranging from 58 to 48 percent for a
triple-B rating, which equates to a total equity ratio of 42 to 52 percent.
Avera, Di - Reb
A vista Corporation
Did Ms. Carlock apply the risk premium approach to estimate the cost of
equity for Avista?A. No. While Ms. Carlock stated that "much of the theoretical approach" that
she used was consistent with my testimony, Ms. Carlock did not use the risk premium
method to estimate the cost of equity. The risk premium method is widely recognized as a
meaningful approach to estimate investors' required rate of return. Unlike the comparable
earnings method, which depends on earned returns derived from accounting information, the
risk premium approach is based on capital market data indicative of investors' current
expectations. The IPUC has noted the importance of "evaluating all the methods" and "using
each as a check on the other when setting the allowed rate of return.,,2 This is especially the
case in light of the fact that Ms. Carlock's DCF range was based on the results of a single
company and her comparable earnings approach is not capital market oriented.
Why is the use of multiple methods so important when estimating the cost
of equity?
Investors ' expectations are unobservable, and there is no methodology that
provides a foolproof guide to their required rate of return. Each method provides another
facet of examining investor behavior, ~ith different assumptions and premises. Investors do
not necessarily subscribe to anyone method, and no model can conclusively determine or
estimate the required return for an individual firm. If the cost of equity estimation is
restricted to certain methodologies, while the results of other approaches are ignored, it may
significantly bias the outcome. Rather, all relevant evidence should be weighed and
evaluated in order to minimize the potential for error. The importance of considering the
2 Idaho Public Utilities Commission, Order No. 29505 (May 25, 2004) at 38.
Avera, Di - Reb
A vista Corporation
results of multiple methods has been widely noted in the financial literature, as evidenced in
this quote from two noted financial scholars:
In practical work, it is often best to use all three methods - CAPM, bond yield
plus risk premium, and DCF - and then apply judgement when the methods
produce different results. People experienced in estimating capital costs
recognize that both careful analysis and some very fine judgements are
required. It would be nice to pretend that these judgements are unnecessary
and to specify an easy, precise way of determining the exact cost of equity
capital. Unfortunately, this is not possible.
Q. Has the IPUC expressed reluctance to consider the results of the Capital
Asset Pricing Model ("CAPM") approach?
Yes. I am aware that the IPUC has continuing concerns over the measurement
and proper use of the beta value necessary to apply the CAPM and has not routinely focused
on the results of this method.4 Nevertheless, the CAPM is a rigorous conceptual framework
at the heart of modern financial theory and it is widely used and referenced in the investment
community. Of course, the CAPM is based on restrictive assumptions and does not describe
security returns perfectly and there are controversies surrounding the measurement of key
variables, such as beta. But then exactly the same could be said for the constant growth DCF
model, which assumes a single, static growth rate into perpetuity that has no observable
proxy in the capital markets.
What cost of equity is implied if the risk premium method is used to
check the results of Ms. Carlock's analyses?A. Application of alternative risk premium approaches based on 1) surveys of
previously authorized rates of return on common equity for electric utilities, 2) realized rates
3 Brigham, E.F. and Gapenski, LC.Financial Management: Theory and Practice 6th ed., Dryden Press (1991)at 256, as referenced in "Regulatory Finance: Utilities' Cost of Capital" at 239-240.
Avera, Di - Reb
A vista Corporation
ofretum on electric utility common stocks, and 3) forward-looking applications of the
Capital Asset Pricing Model ("CAPM") were discussed in detail in my direct testimony (pp.
45-52). The results of these analyses, which are not adjusted to incorporate flotation costs
are summarized in the following table:
Risk Premium Method
Authorized Returns
Realized Rates of Return
CAPM
Cost of Eqyj!y Estimate
11.
10.
11. 7%
Taken together, applications of the risk premium approach to estimate the cost of equity for
an electric utility are consistent with a rate of return from the top of Ms. Carlock's DCF and
comparable earnings ranges.
What other risk premium evidence confirms that Ms. Carlock'
recommendation is well below investors' required rate of return for Avista?A. While the IPUC has expressed concern regarding the assumptions and inputs
necessary to apply certain forms of the risk premium approach (i.e., beta) it need look no
farther than its recent decision in Case No. IPC-03-13 involving Idaho Power Company
Idaho Power ). In that case, the IPUC approved a cost of equity of 10.25 percent and a
component cost of long-term debt of5.769 percentS Thus, the IPUC's findings imply an
equity risk premium for single-A rated Idaho Power of approximately 4.48 percent. Adding
this equity risk premium to Ms. Carlock's recommended long-term cost of debt of 8.
percent suggests a cost of equity to Avista of 13.16 percent. Alternatively, combining the
4.48 percent risk premium from the IPUC's May 2004 decision with the average yield on
4 See Order No. 29505 at 38.5 Idaho Public Utilities Commission, Order No. 29505 (May 25 2004) at 43.
Avera, Di - Reb
A vista Corporation
triple-B public utility bonds for May 2004 of 6.75 percent6 results in an implied cost of equity
for a utility with the lowest investment grade credit rating of 11.23 percent. This evidence
confirms the reasonableness of selecting a rate of return from the very top of Ms. Carlock'
DCF and comparable earnings ranges.
What other evidence indicates that a return from the top end of Ms.
Carlock's range of results is warranted?
While Ms. Carlock did not provide the analyses underlying her 10.0 to 11.
percent comparable earnings range, this method is typically implemented based on a review
of historical earned rates of return on book equity for the companies or industry in question.
But earned rates of return based on historical information are not necessarily indicative of
investors' long-run perceptions of risk and expectations for return going forward.
Alternatively, reference to earned rates of return expected from firms of comparable risk can
also provide a useful guide that may better reflect the ongoing returns necessary to assure
financial integrity and attract capital. The most recent projections from the Value Line
Investment Survey (Value Line), which is the largest and most widely circulated independent
investment advisory service, indicate that its analysts expect average earned rates of return on
book equity for the electric and natural gas utility industries over the next three to five years
of 11.0 percent.7 Based on Value Line s estimates, investors would anticipate a return on
equity from the average electric and gas utility at the top of Ms. Carlock's comparable
eanungs range.
6 Moody s Investors Service Credit Perspectives (Jun. 14 2004) at 49.7 The Value Line Investment Survey, Jun. 4, 2004 at 154, Jun. 18 2004 at 458.
Avera, Di - Reb
A vista Corporation
Do you and Ms. Carlock agree on the benchmark for a fair rate of
return?
Yes. We agree that the authorized rate of return should be competitive with
returns available to investors from investments of corresponding risk, as directed by
landmark Supreme Court decisions. Ms. Carlock also correctly noted that the opportunity to
earn a return at least equal to those expected in the capital markets for comparable
investments is required if a utility is to be able to attract capital. As stated my Ms. Carlock:
. . .
if the return earned by a firm is not equal to the return being earned on other
investment of similar risk, the flow of funds will be toward those investments
earning the higher returns. Therefore, for a utility to be competitive in the
financial markets, it should be allowed to earn a return on equity equal to the
average return earned by other firms of similar risk.
Ms. Carlock also noted the importance of testing any cost of equity estimate against
applicable standards:
. . .
three standards have evolved for determining a fair and reasonable rate of
return: (1) the Financial Integrity or Credit Maintenance Standard; (2) the
Capital Attraction Standard; and (3) the Comparable Earnings Standard.
This is absolutely correct. If Avista s return on equity does not fully reflect the level of
investment risks that investors perceive, it will violate the risk-return tradeoff, breach
applicable standards, and impair Avista s ability to attract necessary capital.
Did Ms. Carlock recognize that the investment risks associated with
electric utilities have increased?
Yes. Ms. Carlock noted that a plethora of changes have impacted investors
risk perceptions, observing that:
8 Carlock Direct at 6 (emphasis added).
Id. at5.
Avera, Di - Reb
A vista Corporation
The competitive risks for electric utilities have changed with increasing non-
utility generation, deregulation in some states, open transmission access, and
changes in electricity markets.
Ms. Carlock concluded that, because of these greater uncertainties, the difference in the risk
between industrial firms operating in the competitive market and electric utilities "is not as
great as it used to be."ll
Did Ms. Carlock consider this increase in risk in her analysis of the cost
of equity for Avista s jurisdictional utility operations?A. No. Ms. Carlock ignored the implications of this trend in investment risks for
utilities, asserting instead that Avista s "competitive risks" are lower because of its "low-cost
source of power and the low retail rates.,,12 Ms. Carlock also asserted that the Pqwer Cost
Adjustment Mechanism ("PCA") reduces Avista s risks relative to other electric utilities.
Does this represent an accurate assessment of the investment risks
investors' associate with Avista?
No. While I agree with Ms. Carlock that relatively low rates provide benefits
to customers and may improve Avista s competitive position, this narrow view ignores the
substantial uncertainties that Avista s investors assume to realize these benefits. As explained
in detail in my direct testimony, because a high proportion of Avista s energy needs is
provided by hydroelectric facilities, Avista is exposed to a level of uncertainty not faced by
other utilities, which are less dependent on hydro generation.
Reduced hydroelectric generation due to below-average water conditions forces
Avista to rely on less efficient thermal generating capacity and purchased power to meet its
10 Id. at 8.
11 Id.
12 Id. at 8-
Avera, Di - Reb
A vista Corporation
resource needs. As the IPUC has noted
, "
there are no guarantees about future stream flows or
market prices ,14 and in light of the recent past, this dependence on wholesale markets entails
significant risk in the minds of investors, especially for a utility located in the west. Investors
recognize that volatile markets, unpredictable stream flows, and Avista s dependence on
wholesale purchases to meet the needs of its customers expose Avista to the risk of reduced
cash flows, increased need for financing, and unrecovered power supply costs.
Apart from exposure to market uncertainties, Avista also confronts the complexities
associated with maintaining the necessary licenses to operate its hydroelectric stations. The
process of relicensing is prolonged and involved and often includes the implementation of
various studies and measures to address environmental and stakeholder concerns. For
example, a federal court recently ordered the Federal Energy Regulatory Commission
FERC") to respond to a request for a formal review of Idaho Power Company s ("Idaho
Power ) Hells Canyon hydroelectric complex under the Endangered Species Act. 15 These
measures can impose significant additional costs and/or lead to reduced generating capacity
and flexibility.
Does the fact that Avista has a PCA absolve investors from risk
volatility in wholesale power markets, as Ms. Carlock seems to imply?A. No. The fact that Avista had been granted a PCA does not translate into lower
risk vis-ii-vis other electric utilities. First, adjustment mechanisms to account for changes in
power supply costs are the rule, rather than the exception, so that Avista s PCA merely moves
13 Id. at 9.
14 Idaho Power Granted $256 million deferral, but bond plan denied Idaho Public Utilities Commission (May, 2002).
Avera, Di - Reb
A vista Corporation
its risks closer to those of other utilities. Second, the PCA does not prevent the lag between
the time that Avista actually incurs power supply expenses and when it is actually recovered
from ratepayers. Investors are well aware that the significant reduction in cash flows
associated with mounting deferrals can have a debilitating impact on a utility's financial
position.
Moreover, investors are aware that the PCA does not apply to 100 percent of the
difference between the actual cost of purchased power and the amount collected through
rates, with Avista s shareholders remaining at risk for 10 percent of any discrepancy. Indeed
Avista and its investors have already experienced the impact that chaotic market conditions
can have when the utility is forced to rely on wholesale purchases to meet the gap in its
resource needs created by reduced hydro generation. Investors cannot afford to discount the
continuing prospect of further turmoil in western power markets, with S&P recently
emphasizing the record high wholesale prices for both peak and off-peak power:
For 2003, record-high wholesale power prices were the defining feature of the
S. merchant power markets. ... Power prices in the western regions were
also the highest on record outside of the 2000-2001 California energy crisis.
... Off-peak prices also rose about 50% across the U.S. and set record highs
along the way in most regions. 16
Is Ms. Carlock's recommended cost of equity compatible with the level of
investment risks associated with Avista?
No. Avista s weakened financial position, as evidenced by its below-
investment grade corporate credit ratings, place it on an altogether different risk plateau. The
15 "Court orders FERC to answer seven-year-old request for study of Idaho dams' fish impact Electric Utility
Week (Jun. 28, 2004) at 14.
16 Standard & Poor s Corporation
, "
Energy Commodity Report: U.S. Power Prices Record High in 2003
RatingsDirect (Jan. 15 2004).
Avera, Di - Reb
A vista Corporation
speculative grade credit rating assigned to Avista confirms that investors perceive its
investment risks to be higher than for the average utility. Investors rely greatly on bond
ratings as a source of information regarding investment risk and bond ratings and the risk of
common stock investment are closely related. Indeed, the higher risk associated with Avista
is mirrored in its Value Line beta of 0.80. As Mr. Thornton recognized:
. . .
the average risk security has a capital asset pricing model beta of 1., while
the average electric utility from my sample has a Value Line beta of ., which
is 28 percent less risky than the average-risk security.
The corollary of Mr. Thornton s conclusion is that Avista s risk is higher than the average
utility and that its expected returns need to be correspondingly greater to attract investment.
Does Ms. Carlock's recommended cost of equity adequately compensate
investors for Avista s greater risks?
No. While Ms. Carlock asserted that her recommendation considered the
risk characteristics for Avista ,18 she failed to look directly at other capital markets data to
assess the level of return investors require to compensate them for Avista s greater investment
uncertainties. Considering the IPUC's recent decision in Case No. IPC-03-13 to authorize
single-A rated Idaho Power a return on equity of 10.25 percent 19 Ms. Carlock's proposed
10.4 percent cost of equity in this case implies an adjustment of 15 basis points to account for
Avista s below-investment grade credit rating. But as discussed in my direct testimony, the
dramatically greater investment risk imposed by a weakened credit standing implies a
significant premium for Avista above the return required for an investment grade utility.
Indeed, reference to bond yield spreads suggests that the capital markets would require a
17 Thornton Direct at 11.
18 Carlock Direct at 14.
Avera, Di - Reb
A vista Corporation
minimum of2.8 percent in additional return to compensate for the greater risk associated
with a speculative credit rating.
What are the implications of disregarding Avista s investment risks in
setting the allowed rate of return on equity?A. If the greater risks associated with Avista s speculative grade credit standing
are not incorporated in the allowed rate of return on equity, the results will fail to meet the
comparable earnings standard that Ms. Carlock agrees is fundamental in determining the cost
of capital. From a more practical perspective, failing to provide investors with the
opportunity to earn a rate of return commensurate with Avista s risks will only serve to
perpetuate its impaired financial integrity, while hampering Avista s ability to attract the
capital needed to meet the economic and reliability needs of its service area.
How is a utility's financial integrity typically evaluated?
Bond ratings provide the most objective guide to a utility's financial integrity
and prospects for capital attraction. Bond ratings are assigned by independent agencies, such
as S&P and Moody s Investors Service ("Moody ), for the purpose of providing investors
with an overall assessment of the creditworthiness of a firm. As discussed in my direct
testimony, an investment grade bond rating (i.e. triple-B or above) indicates that a utility has
some measure of financial integrity. A below-investment grade rating, such as the double-
corporate ratings S&P has assigned to Avista, generally evidences a relative lack of
creditworthiness and an inability to attract capital except on more speculative terms.
19 Idaho Public Utilities Commission, Order No. 29505 (May 25, 2004).
Avera, Di - Reb
A vista Corporation
How do the rating agencies decide what ratings to assign to a utility such
as Avista?
The ratings assigned to a utility by the rating agencies are based typically on
an evaluation of the utility's business and financial risks. One of the most important of the
qualitative factors in determining a utility s bond ratings is its pre-tax interest coverage ratio
which is a measure of the protection available to pay interest expense from operational cash
flow. The financial ratio guidelines published by S&P specify a range for a utility s pre-tax
coverage ratio that corresponds to each specific bond rating. Widely cited in the investment
community, applicable ratios are determined by aligning the bond rating with the utility'
business profile ranking, which ranges from 1 (strong) to 10 (weak) depending on a utility'
relative business risks. Thus, S&P's guideline financial ratios for a given rating category
(e.triple-B) vary with the business or operating risk of the utility. A firm with a business
profile of"2" (i.relatively lower business risk) could presumably maintain lower coverage
ratios than a utility with a business profile assessment of "9" while maintaining the same
credit rating. S&P has currently assigned a business profile ranking of "6" to Avista.
What pre-tax coverage ratio would Avista require to qualify for the lowest
investment grade bond rating?
Consistent with Avista business profile ranking of"6" and S&P's available
published guidelines, Avista would be required to achieve and maintain a pre-tax interest
coverage ratio in the range of2.6 to 4.0 times to qualify for a triple-B bond rating.
20 Standard & Poor s Corporation
, "
New Business Profile Scores Assigned for U.S. Utility and Power
Companies; Financial Guidelines Revised RatingsDirect (Jun. 2, 2004).
Avera, Di - Reb
A vista Corporation
Is it clear that the coverage ratio implied by Ms. Carlocks
recommendations would grant Avista the financial strength necessary to achieve an
investment grade bond rating?A. No. As shown below, the pre-tax interest coverage implied by Ms. Carlock'
recommendations is 2.71 times:
Weighted Pre-tax
Component Percent Cost Rate Cost Cost
Debt 50.08%68%35%35%
Trust Preferred 57%15%0.34%0.34%
Preferred Stock 76%7.35%0.13%20%
Equity 42.59%10.40%4.43%89%
100.00%25%11.78%
Pre-tax Interest Coverage
Covera2e
35%
11.78%
71 X
This 2.71 times coverage is only marginally above the very bottom end of the 2.6 to 4.0 times
specified by S&P's financial benchmarks for a triple-B bond rating for a utility with Avista
business risks. To restore a company s rating to a previous, higher level, rating agencies
generally require a company to maintain financial indicators above the minimum levels
required for the higher rating over a period of time. Considering Avista s already weakened
credit standing, it is unlikely that Ms. Carlock's proposed rate of return w~uld be adequate to
allow Avista the opportunity, under efficient and economical management, to restore basic
financial integrity and implies a continuation of its current junk bond ratings.
Avera, Di - Reb
A vista Corporation
What other evidence indicates the importance of reasonable regulatory
decisions on Avista s ability to maintain its financial integrity?
Following the IPUC's decision in Case No. IPC-03-, S&P placed the
utility' credit ratings on CreditWatch , indicating the potential for a future downgrades.21 In
explaining this action, S&P noted:
Standard & Poor s Ratings Services today placed the corporate credit rating
and all long-term ratings on IDACORP Inc. ('/A-) and subsidiary Idaho
Power Co. ('A-/A2') on CreditWatch with negative implications following the
May 25 2004, Idaho Public Utilities Commission (IPUC) ruling authorizing
only a $25.3 million (5.2%) permanent electric base rate increase for the
utility, which had requested an $85.6 million (17.7%) increase. ... Following
the IPUC staff's 3.1 % rate increase recommendation in February 2004
Standard & Poor s said that "a final decision by the commission that adopted a
rate increase akin to that proposed by the staff could have an adverse effect on
bondholder protection measures." The final IPUC ruling is indeed
substantially closer to the staff's position than the company , and will weaken
credit protection measures.
Considering the vastly greater investment risks implied by Avista s already weakened credit
profile, the perception of lack of regulatory support would undoubtedly place downward
pressure on current ratings, as is occurring for Idaho Power. Such an outcome would be
inconsistent with the IPUC's stated desire to maintain credit ratings "at or above the current
level,,23 and lends further support for a return on equity at the very top of the range of Ms.
Carlock's results.
21 Standard & Poor s Corporation
, "
IDACORP Ratings Placed on CreditWatch With Negative Implications
Following IPUC Ruling,RatingsDirect (Jun. 15 2004).
22 Id.
23 Idaho Public Utilities Commission, Order No. 29505 (May 25 2004) at 43.
Avera, Di - Reb
A vista Corporation
Is there evidence regarding the importance of regulatory support in
determining a utility's financial integrity?A. Yes. Investment publications and the trade press are replete with examples
that highlight the critical role that a constructive regulatory environment plays in investors
assessment of a utility s credit quality. In discussing the criteria used to establish a
company s bond rating, S&P noted that:
The regulatory relationship can be a benign one - or it can be adversarial. It
affects virtually all corporates to one extent or another, and is obviously
critical in the case of utilities - where it is a factor in all assessments of
business risk.
In light of challenges in the industry, investors have refocused attention on regulatory
policy. An article reporting on investment analysts' comments concerning the prolonged
financial slump in the electric utility industry noted the importance of "evenhanded
regulation " with one analyst concluding "uncertainty is the main obstacle to bolstering
energy utilities' capital. ,,25 Indeed, S&P noted that "one of the major challenges facing the
industry is the daunting task of restoring investor confidence" and recognized the importance
of regulatory support in its assessment of credit quality.26 Accordingly, it is critical to assure
investors' confidence in a balanced approach if reasonable access to capital is to be
maintained.
Did Ms. Carlock consider flotation costs in her DCF analysis?
Yes. Ms. Carlock incorporated flotation costs by increasing the dividend yield
component of her DCF analysis. While Ms. Carlock concluded that direct flotation costs
24 Standard & Poor s Corporation Corporate Ratings Criteria (Nov. 13, 2003) at 42.2S Walsh, Campion
, "
Wall Street Seeks FERC's Help for Power Sector Slump Dow Jones Newswire (January2003).
26 Standard & Poor s Corporation
, "
Regulation and Credit Quality in the U.S. Utility Sector
Avera, Di - Reb
A vista Corporation
would warrant an adjustment equal to 4 percent of the dividend yield component, she reduced
this factor to 2 percent for Avista s jurisdictional utility operations, based on her belief that
all subsidiaries of Avista Corp should be responsible for some of actual flotation costS.,,27
Is there any merit to Ms. Carlock's logic?
No. While I do not disagree with Ms. Carlock that all of Avista s operations
should share the burden of flotation costs incurred to raise equity capital, no adjustment to the
cost factor is required to accomplish this objective. This is because the allowed return on
common equity, including the full 4 percent adjustment for direct flotation costs, is only
applied to the equity used to finance jurisdictional utility operations. Thus, the only flotation
costs that will be considered are those related specifically to the equity required to provide
utility service in Idaho. By adjusting the flotation cost factor downward to 2 percent, Ms.
Carlock is essentially assuming that the costs associated with raising equity invested in Idaho
jurisdictional utility operations are one-half as much as those incurred to finance Avista
other operations. This is clearly not the case and results in a downward bias to Ms. Carlock'
recommendation.
In addition, Ms. Carlock apparently did not adjust the results of her comparable
earnings approach to incorporate flotation costs. Based on Ms. Carlock's representative
dividend yield of 3.4 percent and her 4 percent allowance for flotation costs, this would imply
an upward adjustment of approximately 10 basis points, or a comparable earnings range of
10.1 to 11.1 percent.
27 Carlock Direct at 11.
Avera, Di - Reb
A vista Corporation
In light of the shortfalls in Ms. Carlock's analysis and her failure to
meaningfully address Avista s relative investment risks, what is your conclusion
regarding her recommendations in this case?
In my opinion, Ms. Carlock's recommended 10.4 percent cost of equity falls
well short of the rate of return that investors require from Avista. In order to maintain and
expand utility infrastructure, it is both reasonable and necessary that Avista be provided the
opportunity to strengthen its credit standing and enhance its ability to attract capital. To meet
these challenges successfully and economically, it is crucial that Avista receive adequate
support for its credit standing. Because of shortfalls in her analyses, Ms. Carlock'
recommendation is inadequate to meet this goal.
At the very least, the IPUC should consider the results of risk premium analyses
along with Ms. Carlock's approaches, in evaluating the cost of equity. Ms. Carlock granted
that, in selecting a point estimate from within a range
, "
any point within (the J range is
reasonable.,,28 Coupled with the ongoing risks associated with Avista s continued exposure
to wholesale power markets and its weakened credit standing, this would suggest a minimum
cost of equity from the very top of Ms. Carlock's DCF and comparable earnings ranges.
III.DENNIS E. PESEAU
How did Dr. Peseau evaluate the cost of equity for Avista?
It is important to note that Dr. Peseau s opinions were not based on any
independent analyses of the cost of equity for Avista. Rather, he arrived at his
recommendations based on a purported "update" of my analyses and by making revisions to
my methods.
Avera, Di - Reb
A vista Corporation
What "updates" and modifications did Dr. Peseau make to your cost of
equity analyses?
Apart from conducting no analyses of his own, Dr. Peseau did not simply
update my analyses. Rather, he ignored historical trends in earnings growth in applying the
DCF model, used alternative bond yields to apply my risk premium approaches, and
substituted a lower market return in the CAPM. Finally, Dr. Peseau completely ignored the
flotation cost adjustment supported in my direct testimony.
What was the basis for Dr. Peseau s "revision" to exclude historical
growth rates from his "update" of your DCF analyses?A. In Idaho Power s recent general rate case, Dr. Peseau testified that historical
growth rates should be discarded because he did not approve of the composition of my proxy
groUp.29 Now, Dr. Peseau argues that historical growth rates should be ignored because
investment analysts "have already taken that information into account.,,30 While I agree with
Dr. Peseau that investment analysts may consider historical growth rates in arriving at their
near-term projections, this fact does not support his argument that such growth measures
should be ignored in applying the DCF model. Rather, the fact that professional analysts
consider historical growth rates in their analyses is strong evidence that such growth rates are
also of relevance to investors in assessing their expectations and required rate of return.
Indeed, Value Line and other investment advisory services routinely report historical growth
rates, along with near-term projections. Ifhistorical rates of growth were not of interest or
relevance to investors, there would be no need to compile such information and present it on
28 Carlock Direct at 14.
29 Direct Testimony of Dennis E. Peseau, Idaho Public Utilities Commission, Case No. IPC-03-, at 16.30 Peseau Direct at 51.
Avera, Di - Reb
A vista Corporation
an equivalent basis with near-term forecasts. Regulatory Finance: Utilities' Cost of Capital, a
source referenced by Dr. Peseau, concluded that:
Historical growth rates. .. are often used as proxies for investor expectations
in DCF analysis. Investors are certainly influenced to some extent by
historical growth rates in formulating their future growth expectations. In
addition, these historical growth indicators are widely used by analysts
investors, and expert witnesses. ...
Obviously, historical growth rates as well as analysts forecasts provide
relevant information to help the investor with regard to growth expectations.
But instead of heeding the advice of his own source, Dr. Peseau advocates ignoring historical
information altogether and thereby introduces a downward bias to the DCF results.
Q. Is there anything "inexplicable" about your recommended 6.0 percent
growth rate, as Dr. Peseau contends?32
Not at all. The rationale underlying my use of a 6.0 percent growth rate in the
DCF model was fully explained in my testimony (pp. 42-45). As I noted there, based on
analysts' projections and historical growth rates, but giving little weight to Value Line
projections, which deviated from consensus forecasts, I concluded that investors expect
growth in the 5.0 to 7.0 percent range for my proxy group. The 6.0 percent growth rate is the
midpoint of this range. As shown below, my 6.0 percent recommended growth rate is also
equal to the average of the remaining values after excluding Value Line s pessimistic earnings
growth projections:
Source
ffiES
Value Line
First Call
Multex
Growth Rate
31 Morin, Roger A.
, "
Regulatory Finance: Utilities Cost of Capital " Public Utility Reports (1994) at 140.32 Peseau Direct at 51.
Avera, Di - Reb
A vista Corporation
Historical 10 Yr.
Historical 5 Yr.
Value Line "bxr
7.3%
Average
Thus, the growth rate developed in my testimony is consistent with the recommendation of
Dr. Peseau s reference source; which notes that "equal weight should be accorded to DCF
results based on history and those based on analysts' forecasts. ,,33
What about Mr. Peseau s contention that your recommendation would
have been lower if you had applied a multi-stage DCF model (p. 53)?
Mr. Peseau s speculation is apparently based on his observation that dividend
growth in the electric utility industry is lagging behind earnings growth. As discussed in my
direct testimony, this observation only serves to illustrate the fact that near-term trends in
dividends are not representative of investors' long-term expectations. In any event, I
explained why there is presently no compelling arguments in favor of a multi-stage DCF
model and Mr. Peseau presented no evidence to support his remarks and candidly admitted
that "I have not presented such an analysis. ,,34
Is there any merit to Dr. Peseau s suggestion that there are inconsistencies
in your risk premium approaches that lead to an upward bias in your results (pp. 54-
56)?
No. The bond yields used in my applications of the risk premium method
were consistent with the underlying data sources used to compute equity risk premiums.
developing risk premiums based on authorized rates of return on equity in Schedule WEA-
I matched allowed rates of return in each year with the average yield on public utility bonds
33 Morin, Roger A.
, "
Regulatory Finance: Utilities Cost of Capital " Public Utility Reports (1994) at 157.34 Peseau Direct at 53.
Avera, Di - Reb
A vista Corporation
reported by Moody s Investors Service ("Moody ). This composite interest rate reflects the
risk profile of the electric utility generally over the 29 years covered by my analysis and there
is simply no basis for Dr. Peseau s insinuation that this somehow results in an upward bias.
Similarly, my analysis of realized rates of return reported on Schedule WEA-6 was based on a
consistent set of data, as reported by S&P. Because S&P does not publish an average public
utility bond yield, my analyses relied on the yield on single-A rated issues as a proxy for the
average risk profile of the industry over the study period.
Was it "incorrect" to add the equity risk premium determined in your
studies to the yield on triple-B bonds, as Mr. Peseau claims (p. 54-55)?
No. The exercise at hand is to estimate investors' required rate of return from
Avista s jurisdictional utility operations, not for the average utility. Adding the risk premium
to a triple-B bond yield, as I did, reflects the investment risks of a utility with the lowest
investment grade credit rating.35 Meanwhile, Mr. Peseau derives two of his "updated" risk
premium estimates by adding his revised equity risk premium to the yield on single-A bonds.
As a result, Mr. Peseau s "update" necessarily produces cost of equity estimate that falls
below investors' required rate of return for Avista, which has higher investment risks. As
shown in the table below, even accepting Mr. Peseau s flawed "updates " correcting his
calculation to incorporate the May 2004 average yield on triple-B bonds results in the
following cost of equity estimates:
35 In fact, this approach is likely to understate the return on equity because investors in common stock, the most
junior and riskiest of a utility's securities, undoubtedly demand a greater premium to bear the higher risk of a
triple-B bond rating than debtholders.
Avera, Di - Reb
A vista Corporation
Method
Allowed Returns -A Rated
Allowed Returns - BBB Rated
Realized Returns - Arithmetic
Peseau
Updated"
Risk Premium
72%
35%
01 %
Implied
Cost of EQ!!!!I
11.
11.1%
10.
Triple-
Yield36
75%
75%
750/0
This restatement clearly confirms the downward bias to the 9.2 to 10.8 percent cost of equity
estimates he recommends based on the same approach:
Is your application of the realized rate of return approach based on the
assumption that "investors typically have holding periods of only one year," as Dr.
Peseau asserts (p. 56)?
No. My application of the risk premium method based on realized rates of
return makes no assumption regarding the holding period of the average investors, and Dr.
Peseau s assertion that the equity risk premium is a function of investors' holding period is
wrong. In estimating the cost of equity, the goal is to replicate what investors expect going
forward, not to measure the average performance of an investment over an assumed holding
period. Under the realized rate of return approach, investors consider the equity risk
premiums in each year independently, with the arithmetic average of these annual results
providing the best estimate of what investors might expect in future periods. Dr. Roger
Morin, who Dr. Peseau referenced in his testimony (p. 51), had this to say:
One major issue relating to the use of realized returns is whether to use the
ordinary average (arithmetic mean) or the geometric mean return. Only
arithmetic means are correct for forecasting purposes and for estimating the
cost of capital. When using historical risk premiums as a surrogate for the
expected market risk premium, the relevant measure of the historical risk
36 Moody s Investors Service Credit Perspectives (Jun. 14 2004) at 49.
Avera, Di - Reb
A vista Corporation
premium is the arithmetic average of annual risk premiums over a long period
oftime.
Accordingly, Mr. Peseau s risk premium calculations using geometric means are properly
ignored and I have excluded them from the table above.
How did Dr. Peseau "update" your application of the CAPM approach (p.
57)?
Dr. Peseau did not update or otherwise address my CAPM approach. Rather
he ignored it entirely and instead substituted a market risk premium into my analysis that was
based on an entirely different method. As explained in my direct testimony, I applied the
CAPM based on a forward-looking estimate of the market risk premium that relied on
investors' current expectations in the capital markets. Meanwhile , Dr. Peseau simply asserted
that "(aJt this time, the indicated 'current market risk premium and the long-term average
market risk premium are both 7.2%.,,38 But this 7.2 percent risk premium is based on
historical returns back to 1926, not on the forward-looking expectations that drive investors
required rate ofretum in today s capital markets. The end result of Mr. Peseau s calculations
is not an "update" of my approach, but instead a CAPM cost of equity estimate that fails to
reflect investors' current required rate of return.
Did Dr. Peseau address the need to adjust the cost of equity to reflect the
greater investment risks associated with Avista?
A. No. Dr. Peseau made no mention of Avista s below-investment grade credit
standing or the additional return investors require to compensate for this greater risk. Rather
he simply observed that investors do not expect to be compensated for "non-market" or
37 Morin, Roger A.
, "
Regulatory Finance: Utilities' Cost of Capital " Public Utility Reports (1994) at 275(emphasis added).
Avera, Di - Reb
A vista Corporation
company-specific" risks.39 While Dr. Peseau s comment may apply under the narrow
strictures of modem portfolio theory, it does not alter a fundamental premise of finance that
investors require higher returns to bear higher risks. The strong link between bond ratings
and equity risk premiums has been well documented, and there is no ambiguity that investors
require substantially higher rates of return to compensate them for the risks of speculative
securities, versus those with investment grade ratings. Moreover, the overall assumption that
investors care only about systemic risk and not company-specific risk is a substantial
simplification of reality. In fact, no investor is perfectly diversified and bondholders
management, and other stakeholders have an intense interest in the fortunes of individual
companies. In the real world both macroeconomic risks (like the general economy) and
specific risks (like purchased power) absolutely factor into investors' risk perceptions.
Q. What about Dr. Peseau s allegation that such risks are "taken account by
investors" (p. 48)?
I agree wholeheartedly with Dr. Peseau that investors fully consider the
uncertainties and characteristics of Avista and that the observable share prices in the capital
markets reflect their consensus view of these risks and prospects. But stock prices are only
one component used to estimate investors' required rate of return through quantitative
analyses. To the extent that other assumptions embodied in the analysis (e.market returns
beta values, or growth rates) do not reflect the expectations that investors incorporated into
observed stock prices, the resulting cost of equity estimates will be flawed. For example, Dr.
Peseau s "update" of the CAPM is predicated solely on an historical study of equity risk
38 Peseau Direct at 5839 Peseau Direct at 48-49.
Avera, Di - Reb
A vista Corporation
premiums, which does not contain any current market data. As I noted earlier, there is every
indication that the "updates" proposed by Dr. Peseau do not capture real-world expectations
or investors' requirements for Avista. These flawed approaches and logic do not absolve Dr.
Peseau of the need to consider qualitative indicators of investment risks, including the
business and regulatory uncertainties specific to Avista and the industry in which it operates.
Did Dr. Peseau consider the need to account for past flotation costs?
No. Dr. Peseau did not take issue with my testimony that an adjustment for
flotation costs is reasonable in establishing a fair rate of return for Avista. However, Dr.
Peseau entirely ignored the issue of flotation costs in conducting his "updates" to my
analyses. As discussed in my direct testimony, flotation costs are legitimate and necessary,
and unless an adjustment is made to the cost of equity, investors will not have the opportunity
to earn their fair rate of return.
IV.JOHN S. THORNTON, JR.
Does Mr. Thornton recommend a "fair and reasonable" return on equity,
as his subtitle on page 4 would suggest?
A. No. His 8.50 percent recommendation fails all tests of reasonableness. Mr.
Thornton s claim that his return is adequate to maintain Avista s financial integrity is also
wrong because of mistakes in his coverage calculation presented on Exhibit JST-1 and his
reference to the wrong benchmarks to gauge how bond rating agencies evaluate adequacy.
Finally, Mr. Thornton s criticisms of my testimony miss the mark and are simply not credible.
Avera, Di - Reb
A vista Corporation
Do recently authorized returns for electric and gas utilities conclusively
demonstrate the extreme downward bias of Mr. Thornton s 8.5 percent cost of equity
recommendation?
Yes. This recommendation falls far short of the IPUC's recent finding of a
10.25 percent cost of equity for Idaho Power. Further, in contrast to the single-digit cost of
equity estimate proffered by Mr. Thornton, Regulatory Research Associates reported that
authorized rates of return on equity for electric and natural gas utilities averaged 11.0 percent
and 11.1 percent, respectively, for the first quarter of 2004.
What causes Mr. Thornton s analysis to fall so far from a fair and
reasonable result?
In rebutting Mr. Thornton, I will show that his views are contrary to empirical
evidence and common sense and at odds with recent reasoning by the IPUC and the opinions
of investors. The most fatal flaw in Mr. Thornton s approach is that he forgets that the
bottom line test of any rate of return recommendation is whether it is consistent with the
requirements of real world investors. Mr. Thornton s personal views and insights on risk and
return are simply irrelevant if investors don t agree.
Is Mr. Thornton correct on page 31 when he claims that his 8.49 percent
recommended overall rate of return would maintain Avista s financial integrity?
Not at all. First, Mr. Thornton miscalculates the coverage ratio by ignoring the
fact that payments to holders of trust preferred securities are tax deductible. Second, he
compared Avista s projected financial parameters to other utilities actual performance during
2000-2002, a period of unprecedented turmoil in the electric utility industry. Mr. Thornton
Avera, Di - Reb
A vista Corporation
did not compare the projected coverage to the current criteria that the rating agencies apply in
their assessment of credit standing. Indeed, Mr. Thornton criticizes me for not recognizing
the improvements in the industry over the last year (pp. 33-34), yet he measures Avista
prospective performance against those dark days for the industry.
How does impact of Mr.Thornton s recommendations on Avista
financial integrity compare with that implied by Ms. Carlock's proposals?
It is far worse. As shown on below, after properly accounting for the tax
deductibility of Avista s trust preferred securities, his recommendation really translates into a
coverage ratio of 2.52 times:
Weighted Pre-tax
Com onent Percent Cost Rate Cost Cost
Debt 48.19%70%4.19%19%
Trust Preferred 79%01%0.41%0.41%
Preferred Stock 726%34%0.13%20%
Equity 44.30%50%77%86%
100.00%8.49%10.65%
Pre-tax Interest Coverage
10.65%
54 X
Covera2e
19%
This is well below the 2.times minimum threshold specified by S&P for an investment
grade credit rating.A coverage ratio below the minimum guideline specified for a triple-
bond rating is far below the level required to allow Avista to start down the road to rebuild its
creditworthiness. The continuation of junk bond ratings, as will result if Mr. Thornton
recommendations are adopted, would fail to allow Avista an opportunity to maintain its
financial integrity or the ability to attract capital on reasonable terms on a prospective basis.
As a result, Mr. Thornton s proposals are clearly inconsistent with the financial integrity
40 Regulatory Research Associates
, "
Major Rate Case Decisions - January-March 2004"Regulatory Focus
Avera, Di - Reb
A vista Corporation
end-result" test and should be rejected. A speculative grade corporate credit rating does not
permit Avista to maintain its financial integrity or ability to attract capital on other than
speculative terms.
Should it be relevant to this Commission that Mr. Thornton does not
share your "rather gloomy outlook" for electric utilities (p. 33) and has less pessimism
in his own views?
Neither my views nor those of Mr. Thornton are as relevant as the perceptions
of investors and their willingness to provide capital to Avista on reasonable terms. The
headline of the Fitch report included in Exhibit JST-, pp. 20-21 indicates that at the end of
2003 there were finally prospects for stabilization in the industry. Stable is better than
deterioration, to be sure. This Fitch report, which Mr. ThorntQn referenced on page 34 of his
testimony, confirms that the industry is coming out of a bleak period that left many
participants weakened. Avista, with its double-B corporate rating is a prime example of a
company striving to stabilize its financial circumstances. Were this Commission to send
disturbing signal, such as adopting an unreasonable return like that recommended by Mr.
Thornton, Avista and its customers would be denied the benefits of stabilization and the
opportunity to regain an investment grade credit rating.
Is Mr. Thornton correct when he claims on page 8 that the arithmetic
mean is "spurious" so that the geometric mean should be the sole measure of average
rate of return?
(Apr. 5, 2004).
Avera, Di - Reb
A vista Corporation
, absolutely not. Both the arithmetic and geometric means are legitimate
measures of average return; they just provide different information. Each may be used
correctly or misused depending upon the inferences being drawn from the numbers. I am
particularly sensitive to Mr. Thornton s cavalier attitude toward these measures since my
Ph.D. dissertation dealt with the proper use of the geometric mean by investors.
The geometric mean of a series of returns measures the constant rate of return that
would yield the same change in the value of an investment over time. The arithmetic mean
measures what the expected return would have to be each period to achieve the realized
change in value over time. The observation on page 10 of Mr. Thornton s Exhibit JST-
recognizes the legitimate role of the arithmetic mean:
Investors can be expected to realize geometric returns only over long
periods of time. The average geometric return is always less than the
arithmetic return except when all yearly returns are exactly equal. This
difference is related to the volatility of yearly returns.
As noted earlier in my rebuttal of Mr. Peseau, the arithmetic mean is the preferred
measure when using historical data for rate of return analyses. Yet, Mr. Thornton uses the
geometric mean exclusively and criticizes me for use of the arithmetic mean. One does not
have to get deep into finance theory to see why the arithmetic mean is more consistent with
the facts of this case. The IPUC is not setting a constant return that Avista is guaranteed to
earn over a long period. Rather, the exercise is to set an expected return based on test year
data. In the real world, Avista s yearly return will be volatile, depending on many economic
and weather factors, and investors do not expect to earn the same return each year.
Did Mr. Thornton apply the conventional DCF model used by you, Ms.
Carlock, and Dr. Peseau?
Avera, Di - Reb
A vista Corporation
No. Mr. Thornton used a multi-stage DCF model of his own design.
Although Mr. Thornton discusses his thoughts on why this model makes sense to him, he
presents no evidence that this model replicates the reasoning of real world investors. Mr.
Thornton s discussions of the record of stock market returns going back two centuries and
examination of a number of economic forecasts may be an intellectual exercise of sorts, but it
doesn t inform us of what real world investors expect when they invest in utilities like Avista.
Indeed, it is particularly telling that Mr. Thornton refers to "my growth estimates" on page 18
of his testimony. What matters are investors' estimates. Mr. Thornton gives us no credible
evidence that any investors share his expectations.
Do you agree with Mr. Thornton that dividend growth rates are likely
provide a superior guide to investors' growth expectations?A. No. Dividend policies in the electric utility industry have become increasingly
conservative as business risks in the industry have become more accentuated. Thus, while
earnings may be expected to grow significantly, dividends have remained largely stagnant as
companies conserve financial resources to provide a hedge against heightened uncertainties.
In this regard, the near-term dividend growth projections understate long-term expectations
for an industry in the midst of turmoil. S&P observed that, while over the past few years
many utilities have frozen dividends or significantly lowered their growth rates" in order to
finance operations and pay down debt, "financially stronger companies may reconsider their
dividend policies.,,41
But in contrast to the assumptions Mr. Thornton builds into his DCF model, investors
focus logically shifts from dividends to earnings as a measure of long-term growth as payout
Avera, Di - Reb
A vista Corporation
ratios trend downward. As a result, growth in earnings, which ultimately supports future
dividends and the share price gains anticipated by investors, is likely to provide a more
meaningful guide to investors' long-term growth expectations. The fact that investment
advisory services, such as ffiES and First Call focus on growth in earnings indicates that the
investment community regards this as superior to dividends as an indicator of future long-
term growth. Indeed Financial Analysts Journal reported the results of a survey conducted to
determine what analYtical techniques investment analysts actually use.42 Respondents were
asked to rank the relative importance of earnings, dividends, cash flow, and book value in
analyzing securities. Of the 297 analysts that responded, only 3 ranked dividends first while
276 ranked it last. The article concluded:
Earnings and cash flow are considered far more important than book value and
dividends. 43
Did you err in not using a larger sample of utilities as claimed by Mr.
Thornton at page 34?
No. Mr. Thornton s claim that a larger sample results in "a more efficient
estimator" is contrary to common sense. My selection of these companies was guided by
Value Line s classification of utilities for investors. I chose a sample of western utilities
because there was evidence that investors believe that these utilities share risks that are
unique to the region. Throwing in more utilities from other parts of the country does not
improve information if these companies are not comparable in investors' eyes.
41 Standard & Poor s Corporation Industry Surveys: Electric Utilities (Aug. 7, 2003) at 8.42 Block, Stanley B.
, "
A Study of Financial Analysts: Practice and Theory,Financial Analysts Journal
(July/August 1999).
43
Id. at 88.
Avera, Di - Reb
A vista Corporation
Is there any validity to Mr. Thornton s claim at page 35 that your
dividend yield calculation mismatches price and dividends?
No. The price is observed at the same time as the dividend expectations.
There is no reason to believe that the publication of the Value Line each week causes prices
to move systematically because the information in Value Line Summary Index causes
investors' to alter their expectations, as suggested by Mr. Thornton. If this were the case, then
investors would certainly seek more timely and uniform distribution of Value Line, rather
than relying on weekly deliveries by U.S. mail.
Mr. Thornton argues at page 36 that you unreasonably assume that
companies will "suddenly and forever increase dividends by 6 percent per year" which
is "tremendously optimistic to the point of incredible." Do you make any incredible
assumptions?
No. I am attempting to replicate investor expectations, as reflected in mES
and First Call and other publications. First, as explained earlier and in detail in my direct
testimony, investors focus on earnings, not dividends in projecting future growth. This view
is confirmed in the writings of Professor Siegal referenced by Mr. Thornton:
It does not matter how much is paid as dividends and how much is reinvested
as long as the firm earns the same return on its retained earnings that
shareholders demand on its stock. The reason for this is that dividends not
paid today are reinvested by the firm and paid as even larger dividends in the
future.
Second, investors do not have an infinite horizon. Their projections of growth go out to the
foreseeable future. Few, if any, real world investors concern themselves with infinitely long
44 Exhibit JST -, p. 11 (emphasis original).
Avera, Di - Reb
A vista Corporation
horizons. As a practical matter, not only is it impossible to predict the distant future, it simply
doesn t matter. In terms of the DCF model, the present value of cash flows in far distant years
- beyond the foreseeable future - is so small as to have little effect on investment decisions
today.
Is Mr. Thornton correct to argue (p. 36) that "one cannot conclude that
investors reasonably expect a 6 percent dividend growth in the near future (through
2009) much less infinity"
No. Investors expect what they expect. If publications like mES and Value
Line reflect what investors expect, arid there is every indication they do, then it is reasonable
to conclude that what you see is what they expect. Mr. Thornton seems to think there is some
absolute benchmark for investor expectations other than what:we see revealed in the
marketplace. This view is contrary to that found in Professor Siegal's words on page 10 of
Mr. Thornton s Exhibit JST-
However, the risk and return on stocks and bonds are not physical constants
like the speed of light or gravitational force, waiting to be discovered in the
natural world. Historical values must be tempered with an appreciation of
how investors, attempting to take advantage of the returns from the past, can
alter those very returns in the future.
Please respond to Mr. Thornton s contention that the analysts' growth
projections you used to apply the DCF model are "overly optimistic" (p. 36).
First, in contrast to Mr. Thornton s allegations, a study reported in "Analyst
Forecasting Errors: Additional Evidence" found no optimistic bias in eamings projections for
large firms (market capitalization of $500-000 million), with data for the largest firms
Avera, Di - Reb
A vista Corporation
(market capitalization ~ $3 000 million) demonstrating a pessimistic bias.45 More
importantly, however, any bias in analysts' forecasts - whether pessimistic or optimistic - is
irrelevant if investors share analysts' views. The continued success of investment services
such as illES, and the fact that projected growth rates from such sources are widely
referenced, provides strong evidence that investors give considerable weight to analysts
earnings projections in forming their expectations for future growth. While the projections of
securities analysts may be proven optimistic or pessimistic in hindsight, this is irrelevant in
assessing the expected growth that investors have incorporated into current stock prices. As
an article in Journal of Applied Finance noted:
There is very little research on the properties of five-year growth forecasts, as
opposed to short-term predictions.
.. .
Analysts' optimism, if any, is not necessarily a problem for the analysis in
this paper. If investors share analysts' views, our procedures will still yield
unbiased estimates of required returns and risk premia.
Given the importance that investors place on estimates of earnings growth, there is no basis
to support Mr. Thornton s contention that securities analysts' earnings growth projections
should not be used in the DCF model.
Does Mr. Thornton use conventional inputs to apply the CAPM?
No. Mr. Thornton rejects the use of Value Line betas and creates his own
(lower) adjusted betas. Similarly, he follows his own views about the appropriate risk-free
rate and market risk premiums. Again, Mr. Thornton tells us why he has convinced himself
45 Brown, Lawrence D.Analyst Forecasting Errors: Additional Evidence Financial Analysts Journal
(November/December 1997).46 Harris, Robert S. and Marston, Felicia C.
, "
The Market Risk Premium: Expectational Estimates Using
Analysts' Forecasts Journal of Applied Finance 11 (2001) at 8.
Avera, Di - Reb
A vista Corporation
of the rightness of these inputs, but does not offer any evidence that real world investors
would apply the model his way.
Is there reason for the IPUC to be concerned about Mr. Thornton s low
betas?
Yes. His downward adjustment of the Value Line betas is a major driver of his
low CAPM estimates. In its recent decision in the Idaho Power case the IPUC noted the
concerns about the measurement and proper use ofbeta.47 Mr. Thornton puts great emphasis
on beta not only in his CAPM analysis but as a basis for arguing that utilities have much less
risk than the average stock. To the extent that investors use betas in assessing risk, they are
more likely to reference the published betas in a widely circulated and authoritative source
like Value Line, rather than Mr. Thornton s self-developed adaptations to Value Line. Most
surprising, however, is that buried in Mr. Thornton s discourse on betas is evidence that
validates the IPUC's healthy skepticism.48 The graph of betas presented by Mr. Thornton on
page 26 of his testimony reveals a sharp drop in "OLS betas" in the late 1970s and early
1980s. This was a period of turmoil in the electric utility industry as the second oil embargo
hit along with the Three Mile Island incident. To investors this was a time of great concern
about utilities with resulting dramatic drops in the prices of utility common stocks and
downgrades of utility bond ratings at a time when interest rates and inflation had been soaring
to new highs. As utilities were reeling in the aftermath of these changes, the stock market
generally was strong as inflation and interest rates began to fall and the economy shook off its
47 Idaho Public Utilities Commission, Order No. 29505 (May 25 2004) at 38.48 While using the CAPM as his sole risk premium method in the face of the IPUC's reservations about thismethod, Mr. Thornton disparages the comparable earnings method favored by this Commission on page 37
calling it "an inferior approach to estimate a cost of equity.
Avera, Di - Reb
A vista Corporation
malaise during the early years of President Reagan s administration. While investors almost
certainly regarded electric utilities to be increasing in relative risk, the unadjusted betas were
dropping because utility stock prices were going down while the market was rising. This
period was a statistical artifact that led most observers to understand that historical betas
should be interpreted with a prudent grain of salt.
Are Mr. Thornton s criticisms of your allowed ROE risk premium
approach correct?
, Mr. Thornton s criticisms of the allowed rates of return used in this
approach are without merit. First, he is incorrect to allege that the information regarding
average allowed rates of return in each year is unreliable simply because every item of
possible interest in each rate case is not also presented in my schedule. The allowed rates of
returns are taken from a recognized and widely-used publication from a firm with a long
history of accumulating and reporting the results of state regulatory commission decisions.
Mr. Thornton questions the potential for "upward bias " depending on the form of the DCF
model considered by regulators or whether they considered results of an "inferior approach
such as the comparable earnings method proposed by Ms. Carlock. But such criticisms miss
the point. Under this approach, it is not necessary to examine the actual tools and techniques
relied on by regulators to set allowed rates of return. Rather, what matters is that, after
reasoned consideration of the evidence presented by all participants to a rate proceeding,
regulators make an informed determination of investors' required rate of return at the time
they issue their decision. This determination is embodied in the authorized rates of return on
equity that I used to apply the risk premium approach.
Avera, Di - Reb
A vista Corporation
With respect to his theoretical arguments, Mr. Thornton is wrong about the risk
premium in the regression not being an independent variable.49 While the interest rate is
subtracted from the average allowed return each year, bond yields do not appear as an
independent variable in the analysis. Thus, if the risk premium had no association with the
level of interest rates, the regression equation would not show a statistically significant
relationship. In fact, the association found is highly significant using standard statistical
inference. Mr. Thornton also asserts that this study of authorized ROE's does not correct for
changes in industry risk. First, as explained in detail in my direct testimony, there is little
support for Mr. Thornton s contention that the risks associated with the electric power
industry have decreased over the period covered by my study. But irrespective of whether
risk was increasing or decreasing, this would be considered by regulators and captured in the
market data used to establish allowed rates of return. Mr. Thornton is also incorrect to claim
that declines in interest rates would lead to bias in the risk premiums. In fact, the average
interest rates used to apply this approach match the time period used to determine the average
allowed returns. Moreover, interest rates fluctuated considerable over the 29 years covered
by my study, which encompassed periods when interest rates were rising precipitously, as
well as times of moderating rates. And contrary to Mr. Thornton s allegation that my study is
out of step" by "mismatching" allowed ROEs and interest rates, my study specifically
adjusted for the impact of changes in bond yields on the equity risk premium. 50 Mr.
Thornton s suppositions are simply lacking in factual basis.
49 Thornton Direct at fn. 19.
50 Thornton Direct at 38.
Avera, Di - Reb
A vista Corporation
Is there any meaningful basis to Mr. Thornton s allegation that your risk
premium analysis based on realized rates of return is biased because it rewards
unsystematic risk" (pp. 39-40)?A. No. First, as I noted earlier in response to Dr. Peseau, the overall assumption
that investors care only about systemic risk and not company-specific risk is a substantial
simplification of reality. No investor is perfectly diversified and in the real world - as
distinct from Mr. Thornton s constructions - both macroeconomic risks and specific risks
affect investors' risk perceptions and return requirements.
Second, the assumption underlying the realized rate of return method is that historical
returns, measured over a sufficiently long time period, provide a surrogate for the forward-
looking rates of return required in the capital markets. This method does not depend on the
strict assumptions of the CAPM and avoids the controversy surrounding beta by looking
directly at returns for electric utilities. Nevertheless, these realized rates of return are a
function of actual prices in the capital markets, which are determined by real-world investors
that have the opportunity to "diversify into other industries.,,51 Thus, following Mr.
Thornton s logic, to the extent that these investors can eliminate risk through diversification
it would not be "priced in the market" or reflected in the values used to compute the realized
rates of return underlying my analysis. In other words, contrary to Mr. Thornton s assertions
the only compensation priced into realized returns would be for systematic risks.
51 Thornton Direct at 39.
52 This can be demonstrated by way of example. Subtracting my 5.2% risk-free rate from my 10.6% cost ofequity based on realized returns results in a risk premium for electric utilities of 5.4%. Dividing this premiumby the average beta of 0.77 for the fmns in my proxy group results in a market risk premium of 7.01 %, whichfalls squarely within the 6.1 to 7.8 percent range advocated by Mr. Thornton (p. 27).
Avera, Di - Reb
A vista Corporation
Third, Mr. Thornton again implies that declining risks may lead to an overstatement
of the cost of equity. Apart from the fact that his position is diametrically opposed to the
views of the investment community, as demonstrated in my direct testimony, it is also at odds
with the statistics he cites one paragraph previously, where he notes that the volatility of the
returns to electric utilities exceeded that for the S&P 500 over the 1994 to 2002 period.
Under Mr. Thornton s theoretical paradigm, higher volatility of returns relative to the market
is indicative of higher, not lower, investment risks.
Fourth, as I noted earlier in response to Dr. Peseau, there is no "mismatch" (p. 40) in
using triple-B bond yields to develop a cost of equity estimate for Avista. Adding the risk
premium to a triple-B bond yield, as I did, reflects the investment risks of a utility with the
lowest investment grade credit rating and is more likely to understate, rather than overstate
the returns required by equity investors.
Finally, the single academic study referenced by Mr. Thornton provides no meaningful
information to evaluate the realized rate of return approach or aid the IPUC in its
deliberations. As Mr. Thornton summarized, the final conclusion of this research was that
risk premiums for utilities "should be close to zero.,,54 Of course, no reasonable analyst
would contend that the current risk premium for electric utilities should approach zero and
such a nonsensical result is even inconsistent with the meager returns recommended by Mr.
Thornton himself.
53 Thornton Direct at 39.
54 Thornton Direct at 42.
Avera, Di - Reb
A vista Corporation
Q. Is there any reason to believe that the market risk premiums and
expected returns are declining, as Mr. Thornton (p. 40) and Dr. Peseau (p. 58) assert?A. No. Contrary to the assertions of these witnesses, a study reported in the
January/February 2003 edition of Financial Analysts Journal noted that the real risk premium
for U.S. stocks averaged 6.9 percent over the period 1889 through 2000 and concluded that:
Over the long term, the equity risk premium is likely to be similar to what it
has been in the past and returns to investment in equity will continue to
substantially dominate returns to investments in T-bills for investors with a
long planning horizon.
Combining this real risk premium with an inflation rate of 3 percent suggests a market equity
risk premium well above the 8.5 percent used in my CAPM analysis that Mr. Thornton
characterized as "unrealistically high. ,,56
Please respond to Mr. Thornton s criticism of the long-term debt cost you
used to apply the CAPM (p. 43-45).A. I agree with Mr. Thornton that:
The use of a long-term U.S. Treasury bond for the risk-free asset implies a
long-term holding period.
Common equity is a perpetuity and as a result, the return that investors require is predicated
on their expectations for the firm s long-term risks and prospects. This does not mean that
every investor will buy and hold a particular common stock into perpetuity, but even an
investor with a relatively short holding period will consider the long-term because of its
influence on the price that he or she ultimately receives from the stock when it is sold.
Similarly, Mr. Thornton recognized that in applying the DCF model, the analyst must
SS Mehra, Ranjnish
, "
The Equity Premium: Why Is It a Puzzle?Financial Analysts Journal (January/February
2003).
S6 Thornton Direct at 46.
S7 Thornton Direct at 43.
Avera, Di - Reb
A vista Corporation
consider "the present value of all future dividends expected to be received by a share of
stock ,S8 not just the
dividends to be paid during some shorter (e.two-year), intermediate-
term holding period. Indeed, as Mr. Thornton observed in his Appendix, under the DCF
model "we assume that dividends are paid infinitely (n~oo)."S9
In fact, credible sources unambiguously recognize that long-term Treasury bond yields
provide the preferred basis to compute a long-term cost of capital. Indeed, Roger Ibbotson
whose firm Ibbotson Associates provided data relied on in Mr. Thornton s CAPM
application, made the same conclusion over a decade ago, explaining that while the CAPM
can be applied using short-term bill rates, the appropriate basis for a long-term cost of equity,
especially in the context of rate setting, is the yield on long-term Treasury bonds:
Q. Should the CAPM be used to estimate the short-term or the long-term cost
of capital?
A. The CAPM was originally formulated to measure the short-term cost of
capital, but it may be adapted to measure the long-term cost of capital by using
the expected return on a long-term government bond, instead of the risk risk-
free rate of return, as the riskless rate. ...
Q. When is it appropriate to use the long-term cost of capital?
A. It is necessary to use a long-term cost of capital when discounting cash
flows projected over a long period. Also, regulated ratesetting processes often
specify or suggest that the rate of return should allow the firm to attract and
retain debt and equity capital over the long term. Thus, the long-term cost of
capital is typically the appropriate cost of capital to use in regulated
ratesetting.
58 Thornton Direct at 13 (emphasis added).
59 Thornton Direct at 53.60 Ibbotson, Roger G. and Sinquefield, Rex A.
, "
Stocks, Bonds, Bills, and Inflation: Historical Returns (126-
1987)," Research Foundation of The Institute of Chartered Financial Analysts (1989) at 122-25.
Avera, Di - Reb
A vista Corporation
More recently, Ibbotson Associates again emphasized the importance of using long-term bond
yields when applying the CAPM to estimate returns for long-term assets, such as common
stock:
The horizon of the chosen Treasury security should match the horizon of
whatever is being valued. ... Note that the horizon is a function of the
investment, not the investor. If an investor plans to hold a stock in a company
for only five years, the yield on a five-year Treasury note would not be
appropriate since the company will continue to exist beyond those five years.
In applying the CAPM, Ibbotson Associates recognized that the cost of equity is a long-term
cost of capital and the appropriate interest rate to use is a long-term bond yield. Mr.
Thornton s criticism of the long-term bond yields that I used is simply without basis and his
use of a shorter, intermediate term bond yield is similarly unfounded.
Did Mr. Thornton recognize that flotation costs are a necessary expense
that a utility must incur if it is to raise equity capital?
Yes. Mr. Thornton granted (p. 48) that "(f)lotation costs are a necessary cost
of business." Rather than recommend an upward adjustment to account for these costs
however, Mr. Thornton recommended that Avista be allowed to recover flotation costs "as an
expense item" through an accounting treatment.
Do you have any objection to the IPUC adopting an accounting treatment
for the recovery of flotation costs?
No. Allowing recovery of flotation costs as an expense item is certainly one
acceptable way to address this issue going forward. On the other hand, such a treatment
would ignore the costs already incurred in connection with past stock issuances. The only
practicable means available to ensure that Avista has the opportunity to earn investors' cost of
Avera, Di - Reb
A vista Corporation
capital is to include an allowance for past flotation costs in arriving at the fair rate of return
as Ms. Carlock and I have recognized. Choosing to ignore a "necessary cost of business" is
yet another reason explaining the extreme downward bias of Mr. Thornton s recommended
cost of equity.
Does financial theory preclude higher returns for higher risk, as Mr.
Thornton implies (p. 49-50)?
Of course not. Bond ratings are a widely recognized proxy for investment
risk. Mr. Thornton apparently is under the impression that investors would not necessarily
require a higher cost of equity from a
. "
D" rated company whose debt is in default because
investors can avoid risk by diversifying.,,63 This shows just how far Mr. Thornton s analysis
departs from common sense in order to justify a below-market return on equity. Lower bond
ratings, such as Avista s double-B corporate credit rating, evidence investors' understanding
that there is greater uncertainties surrounding the firm s ability to successfully meet its
financial obligations, especially during adverse market conditions. In fact, this potential for
greater variability translates into Mr. Thornton s CAPM paradigm, with Avista s beta
exceeding those of the utilities in the proxy groups referenced by Mr. Thornton and me by a
significant margin. Further, while I agree with Mr. Thornton that the interest of bondholders
and stockholders may not always be aligned, the risks of investing in common stocks clearly
exceed those associated with bonds. Thus, reference to yield spreads between bonds of
various ratings is far more likely to understate the risk differential perceive4 by common
stockholders.
61 Ibbotson Associates 2003 Yearbook (Valuation Edition) at 53.62 Thornton Direct at 48.
Avera, Di - Reb
A vista Corporation
Does this conclude your rebuttal testimony?
Yes, it does.
63 Thornton Direct at 49.
Avera, Di - Reb
A vista Corporation