BEFORE THE MARYLAND
PUBLIC SERVICE COMMISSION
)
In the Matter of the Investigation
)
Case No. 8879
into Rates for Unbundled Network )
Elements Pursuant to the
)
Telecommunications Act of 1996
)
)
SURREBUTTAL TESTIMONY OF JOHN I. HIRSHLEIFER
ON BEHALF OF AT&T and WORLDCOM
October 15, 2001
TABLE OF CONTENTS
I.
DR. VANDER WEIDE’S ESTIMATE OF THE COST OF EQUITY IS SYSTEMATICALLY BIASED
UPWARD. ............................................................................................................................................................ 1
A.
DR. VANDER WEIDE’S ASSUMPTION OF PERPETUAL GROWTH GUARANTEES AN EXCESSIVE RATE OF
RETURN. ................................................................................................................................................... 1
B.
THE RISKS OF SUPPLYING UNBUNDLED NETWORK ELEMENTS IN MARYLAND DO NOT JUSTIFY THE USE
OF DCF COMPARISON GROUPS THAT INCLUDE NON-TELEPHONE COMPANIES. ..................................... 17
1. TELRIC cost principles do not require the Commission to assume that Verizon Maryland faces
intense competition regardless of the facts. ............................................................................ 17
2. Dr. Vander Weide exaggerates the competitive risk of VZ-MD’s local telephone service generally
and fails to distinguish between the competitive risks of providing UNEs at wholesale and
providing local telephone services at retail. ........................................................................... 23
3. Dr. Vander Weide has not offered plausible or consistent reasons why telephone companies should
not form the appropriate DCF comparison group. ................................................................. 28
C.
THE S&P INDUSTRIAL COMPANIES SELECTED BY DR. VANDER WEIDE ARE NOT A VALID COMPARISON
GROUP FOR A DCF ANALYSIS OF THE COST OF EQUITY OF THE NETWORK ELEMENT BUSINESS. ........ 35
D.
DR. VANDER WEIDE’S MISCELLANEOUS CRITICISMS OF MY DCF ANALYSIS OF EQUITY COSTS ARE
WITHOUT MERIT .................................................................................................................................... 35
E.
DR. VANDER WEIDE’S CRITICISMS OF MY CAPM ANALYSIS OF EQUITY COSTS ARE ALSO UNFOUNDED
............................................................................................................................................................... 37
II.
DR. VANDER WEIDE HAS ASSUMED A CAPITAL STRUCTURE THAT IS INAPPROPRIATE FOR
THE WHOLESALE BUSINESS OF SUPPLYING UNBUNDLED NETWORK ELEMENTS................. 53
III.
DR. VANDER WEIDE’S COST OF CAPITAL ESTIMATES ARE GROSSLY AT ODDS WITH
ANALYSES BY INVESTMENT FIRMS AND OTHER INDEPENDENT ANALYSTS............................ 60
IV.
DR. VANDER WEIDE’S CONTRIVED “TESTS OF REASONABLENESS” UNDERSCORE THE
REASONABLENESS OF MY COST OF CAPITAL ESTIMATES, NOT HIS. ......................................... 74
Surrebuttal Testimony of John I. Hirshleifer
1
Q.
PLEASE STATE YOUR FULL NAME AND OCCUPATION.
2
A.
My name is John I. Hirshleifer and my business address is Charles River Associates, Inc.,
3
10877 Wilshire Blvd., Los Angeles, California 90024. I am a Vice President at Charles
4
River Associates, Inc. (CRA), an international financial and economic consulting firm.
5
Q.
ARE YOU THE SAME JOHN HIRSHLEIFER WHO PREVIOUSLY
6
SUBMITTED PREPARED DIRECT (5/25/01) AND REBUTTAL (9/5/01)
7
TESTIMONIES ON BEHALF OF AT&T AND WORLDCOM IN THIS
8
PROCEEDING?
9
A.
Yes, I am.
10
Q.
WHAT IS THE PURPOSE OF YOUR SURREBUTTAL TESTIMONY?
11
A.
The purpose of my surrebuttal testimony is to respond to the prepared rebuttal testimony
12
dated September 5, 2001 submitted in this proceeding by Dr. James H. Vander Weide
13
("Vander Weide Rebuttal") on behalf of Verizon Maryland ("VZ-MD") regarding the cost
14
of capital. In Sections I and II, I address Dr. Vander Weide’s criticisms with respect to
15
my cost of equity and capital structure estimates. In Section III, I refute Dr. Vander
16
Weide’s attempts to discard the cost of capital estimates used by investment banks and
17
industry analysts. In Section IV, I rebut Dr. Vander Weide’s “Tests of Reasonableness”.
18
19
I.
A.
20
21
22
23
DR. VANDER WEIDE’S ESTIMATE OF THE COST OF EQUITY IS
SYSTEMATICALLY BIASED UPWARD.
Q.
Dr. Vander Weide’s Assumption Of Perpetual Growth
Guarantees An Excessive Rate Of Return.
IN HIS REBUTTAL TESTIMONY DR. VANDER WEIDE ATTEMPTS TO
BOLSTER HIS INCORRECT USE OF THE SINGLE-STAGE DCF MODEL.
-1-
Surrebuttal Testimony of John I. Hirshleifer
1
(VANDER WEIDE REBUTTAL, PP. 55-60) IS HE ABLE TO CITE ANY
2
LEADING AUTHORITY SUPPORTING THE USE OF THE SINGLE-STAGE
3
MODEL WHEN THE GROWTH RATE SIGNIFICANTLY EXCEEDS THE
4
GROWTH RATE OF THE ECONOMY?
5
A.
No. While I cite numerous leading scholars and practitioners in my direct testimony that
6
clearly explain why the single-stage model is inappropriate for use in those
7
circumstances, it is striking that he can cite nothing that rebuts these authorities.
8
Obviously, these experts would not be advocating the superiority of multiple stage
9
models if any of Dr. Vander Weide’s arguments offered in regulatory proceedings
10
11
regarding the single-stage DCF model were true.
Q.
DR. VANDER WEIDE’S SINGLE-STAGE DCF MODEL ASSUMES THAT ALL
12
COMPANIES IN THE S&P INDUSTRIALS, INCLUDING ALL TELEPHONE
13
HOLDING COMPANIES, WILL GROW FOREVER AT RATES HIGHER THAN
14
THE GROWTH RATE OF THE ECONOMY. IN SUPPORT OF HIS MODEL
15
DR. VANDER WEIDE ARGUES THAT “IT IS COMMON FOR COMPANIES TO
16
GROW AT RATES SIGNIFICANTLY GREATER THAN THE RATE OF
17
GROWTH IN GNP FOR LONG PERIODS OF TIME.” (VANDER WEIDE
18
REBUTTAL, P. 57) HE CITES CERTAIN COMPANIES, SUCH AS WAL-MART,
19
INTEL, MERCK AND CENTURYTEL, AS EXAMPLES THAT HAVE GROWN
20
AT HIGH RATES FOR LONGER THAN FIVE YEARS. (VANDER WEIDE
21
REBUTTAL, P. 58) IS HIS REASONING CORRECT?
22
23
A.
No. No company, not even Wal-Mart, Intel, Merck and CenturyTel, will grow at those
rates perpetually. Studious analysts would be hard-pressed to agree that these particular
-2-
Surrebuttal Testimony of John I. Hirshleifer
1
companies will all grow at high rates for the next 20 years, let alone forever. Intel, for
2
example, has been the single most dominant microprocessor producer serving the
3
microcomputer industry, which grew from a base of close to zero in the early 1980s,
4
when microcomputers were unknown to consumers, to widespread use worldwide today.
5
Obviously, the entire S&P Industrials does not enjoy the incredible position that Intel was
6
in at the commencement of the 1980s. Indeed, Intel’s stock price dropped 70% from
7
September 1, 2000 until October 8, 2001, and the company has reduced prices on its
8
premier microchips.
9
It may be too early to know whether Intel’s decline in earnings is a short-term blip
10
or the harbinger of a longer-term slowdown in growth. One fact is clear, however: Intel is
11
unlikely to continue growing indefinitely at the pace it set in recent years as its markets
12
continue to saturate.
13
Q.
IN SELECTING A FEW COMPANIES THAT HAVE HAD SEVERAL YEARS OF
14
HIGH GROWTH, DID DR. VANDER WEIDE MENTION THE COMPANIES
15
THAT HAD AVERAGE, OR POOR, OR NEGATIVE GROWTH?
16
A.
No. And Dr. Vander Weide is unable to tell us which companies of his S&P Industrial
17
sample, or even of a sample of telephone companies, will grow at above-average rates,
18
and which will have average or below-average rates of growth. Just in the last three years
19
Laidlaw, Helmerich & Payne, Forster Wheeler, Fleetwood Enterprises, Pep Boys, Silicon
20
Graphics, IKON Office Solutions, Milacron and several others were dropped from the
21
S&P Industrial group. Laidlaw’s earnings growth rate averaged negative 1% over a 10-
22
year period; Pep Boys’ negative 8.5% over the past 5 years; IKON’s negative 7.0% over
-3-
Surrebuttal Testimony of John I. Hirshleifer
1
the past 10 years and negative 18.0% over the past 5 years.1 Had these and other
2
companies that were dropped from the S&P Industrials over the course of time remained
3
in the set, the expected growth rate for the aggregate sample would also likely be lower
4
than the rate currently forecasted.
5
Q.
IS THE S&P 500 GENERALLY DESIGNED TO BE AN INDEX OF LEADING
6
7
COMPANIES?
A.
Yes. The guiding principle for inclusion in the S&P 500 is that they are "leading
8
companies in leading US industries."2 In addition to rebutting Dr. Vander Weide’s
9
argument that all companies are expected to grow at high rates forever, this fact
10
highlights a selection bias that further taints his choice of comparables. By using an
11
index that is periodically repopulated by dropping selected poorly-performing companies
12
and adding better-performing companies, Dr. Vander Weide is assuring himself that he
13
will be using companies expected to have comparatively favorable growth expectations
14
on the whole.
15
Q.
DO OTHER SCHOLARS AND PRACTITIONERS AGREE THAT COMPANIES
16
OFTEN FAIL TO SUSTAIN ABOVE-AVERAGE RATES OF EARNINGS
17
GROWTH?
18
A.
Yes. They are well aware of the fact that not all of the companies that have grown for
19
many years at high growth rates will necessarily continue to exhibit such rapid growth.
20
Burton Malkiel, for example, has stated that,
1
Value Line Investment Survey, Stock Profile Reports, June 16, 2000, May 18, 2001, and April 20, 2001.
2
Standard & Poor's, General Criteria for S&P U.S. Index Membership, September 2000, p. 3.
-4-
Surrebuttal Testimony of John I. Hirshleifer
1
2
3
4
5
6
7
8
9
10
11
12
Corporations and industries have life cycles similar to most living
things. There is, for corporations in particular, a high mortality rate
at birth. Survivors can look forward to rapid growth, maturity, and
then a period of stability. Later in the life cycle, companies
eventually decline and either perish or undergo a substantial
metamorphosis. Consider the leading corporations in the United
States 100 years ago. Such names as Eastern Buggy Whip Company,
La Crosse and Minnesota Steam Packet Company, Lobdell Car
Wheel Company, Savanna and St. Paul Steamship Line, and Hazard
Power Company, the already mature enterprises of the time, would
have ranked high in a “Fortune Top 500” list of that era. All are
now deceased…
13
14
15
16
17
18
And even if the natural life cycle doesn’t get a company, there’s
always the fact that it gets harder and harder to grow at the same
percentage rate. A company earning $1 million need increase its
earnings by only $100,000 to achieve a 10 percent growth rate,
whereas a company starting from a base of $10 million in earnings
needs $1 million in additional earnings to produce the same record.
19
20
21
22
23
24
25
26
27
28
The nonsense of relying on very high long-term growth rates is
nicely illustrated by working with population projections for the
United States. If the populations of the nation and of California
continue to grow at their recent rates, 120 percent of the United
States population will live in California by the year 2035! Using
similar kinds of projections, it can be estimated that at the same time
240 percent of the people in the country with venereal disease will
live in California. As one Californian put it on hearing these
forecasts, ‘Only the former projections make the latter one seem at
all plausible.’3
29
Q.
DR. VANDER WEIDE ARGUES THAT THE SINGLE-STAGE DCF MODEL IS
30
VALID IF FIRMS CAN GROW AT A CONSTANT GROWTH RATE IN EXCESS
31
OF GNP FOR 40 OR 50 YEARS. (VANDER WEIDE REBUTTAL, P. 58) IS THIS
32
A HELPFUL DEFENSE OF THE SINGLE-STAGE MODEL?
3
Burton G. Malkiel, A Random Walk Down Wall Street, 1999, pp. 97-99 (emphasis added).
-5-
Surrebuttal Testimony of John I. Hirshleifer
1
A.
No. It appears to be a very strong argument against using the single-stage model. It is
2
impossible to predict which, if any, companies could grow at rates above the economy’s
3
growth for that length of time. It does not appear that many would grow at high rates for
4
that long. While some small companies with novel products can have many years of high
5
percentage growth, most do not.4 Sustained future periods of above-average growth are
6
less likely for the average company in the S&P 500 list: a company must already have a
7
relatively large capitalization to enter the list in the first place. And sustained future
8
periods of above-average growth are even less likely for Verizon, the largest telephone
9
holding company in the United States, and the other Bell holding companies. These are
10
already enormous companies, and their growth rates are likely to slow further as the high-
11
growth worldwide markets for wireless phones and data services saturate.
12
Investment bankers, for example, do not assume 40 to 50 years of high growth in
13
their valuation analyses. All use growth rate projections that assume that growth tapers
14
off over time in their DCF models. In valuing Sprint PCS, Morgan Stanley Dean Witter
15
(Morgan Stanley) projected an initially high free cash flow growth rate which consistently
16
declines every year: 76.9% (2003), 46.1% (2004), 25.3% (2005), 13.7% (2006), 8.7%
17
(2007) and 4.2% (2008). After 2008, Morgan Stanley specifically stated that it assumed a
18
6.0% growth of free cash flow in perpetuity.5
4
“While many investors recall the newly issued story stocks, such as Intel, Microsoft, and Wal-Mart, which
have made investors rich, most forget about the many such firms that fail to fulfill their promise when they
are issued. A study by Tim Loughran and Jay Ritter followed every operating company (almost 5,000) that
went public between 1970 and 1990. Those who bought at the market price on the first day of trading and
held the stock for five years, reaped an average annual return of only 5 percent.” Jeremy J. Siegel, Stocks
for the Long Run (1998), at p. 102 (citation omitted).
5
Morgan Stanley Dean Witter, “Sprint PCS Group,” March 13, 2000, p. 6.
-6-
Surrebuttal Testimony of John I. Hirshleifer
1
Similarly, Morgan Stanley projected declining free cash flow growth rates for
2
ALLTEL, one of the companies in my sample of comparables: 25.6% (2001), 18.4%
3
(2002), 12.8% (2003), 8.9% (2004), 8.4% (2005), 6.0% (2006), 4.5% (2007) and 1.2%
4
(2008) and assumed the perpetual growth after year 2008 to be 4%.6
5
If Morgan Stanley had assumed 40 years of growth for ALLTEL at the average
6
growth rate over the first five years of 14.8%, or even at 8.4% as of year 5, it would have
7
obtained a much higher valuation.
8
Q.
IN PERFORMING DCF VALUATIONS, HAVE YOU EVER SEEN AN
9
INVESTMENT BANK OR FINANCIAL ANALYST ASSUME THAT A
10
COMPANY’S EARNINGS OR CASH FLOWS WILL GROW AT EITHER A
11
HIGH RATE PERPETUALLY, OR FOR 40 OR 50 YEARS?
12
A.
No. I have examined numerous DCF valuations over the years and all have used forecast
13
methodologies similar to those used by investment banks as described in the prior answer.
14
This is because analysts are constrained by the reasonability of their valuation results. No
15
one reasonably expects that companies that are growing quickly now will grow at high
16
rates for long, long periods of time. If one were to make such assumptions, the resulting
17
valuations would be stratospheric, and it would be clearly evident to the analyst that the
18
assumptions made were simply wrong.
19
Q.
IS IT REASONABLE TO ASSUME THAT THE ENTIRE S&P INDUSTRIALS
20
WILL GROW AT RATES EXCEEDING THE GROWTH RATE OF THE
21
ECONOMY FOR 40 OR 50 YEARS?
6
Morgan Stanley Dean Witter, “Alltel Corporation,” March 13, 2000, p. 3.
-7-
Surrebuttal Testimony of John I. Hirshleifer
1
A.
No. It is fairly easy to predict that the companies which currently comprise the S&P
2
Industrials, used by Dr. Vander Weide as his comparable set, will not grow at rates
3
significantly above the economy’s growth rate for that length of time.
4
Q.
IF DR. VANDER WEIDE SAYS THAT ONLY 40 TO 50 YEARS OF HIGH
5
GROWTH ARE REQUIRED TO MAKE THE SINGLE-STAGE MODEL
6
ACCEPTABLE, IS HE REALLY SAYING THAT HE SHOULD BE USING A
7
TWO-STAGE MODEL WITH 40 YEARS OF SUPERNORMAL GROWTH AND
8
THEREAFTER GROWTH AT THE ECONOMY’S RATE?
9
A.
Of course. He is clearly validating the use of multiple stage models, although he
10
unrealistically assumes that virtually all companies will grow at high rates for long
11
periods of time.
12
Q.
IF HE WERE TO USE A TWO-STAGE DCF MODEL THAT ASSUMED 40
13
YEARS OF SUPERNORMAL GROWTH, WOULD HE ARRIVE AT THE SAME
14
COST OF EQUITY ESTIMATE THAT HE OBTAINS FROM A SINGLE-STAGE
15
DCF MODEL?
16
No, which entirely contradicts his assertion that 40 years of supernormal growth justifies
17
the use of a single-stage model. Even in comparison to a two-stage model with 40 years
18
of supernormal growth, the single-stage model yields a significantly higher cost of equity.
19
To illustrate, I calculated the costs of equity of a hypothetical company using a
20
one-stage DCF model and a 2-stage DCF model assuming that the high growth rate lasts
21
40 years. For this illustration I assumed that the company’s I/B/E/S-equivalent growth
22
rate equals the weighted average I/B/E/S growth rates of the telephone holding companies
23
in my comparables sample (12.15%), and that the company pays a dividend yield of
-8-
Surrebuttal Testimony of John I. Hirshleifer
1
2.53%, equal to the weighted average dividend yield of the sample. If it were assumed
2
that 40 years of growth were correct, the cost of equity estimated using the single-stage
3
DCF model overstates the cost of equity calculated using the 40-year two-stage DCF
4
model by at least 150 basis points. (See Attachment JH-1.)
5
Moreover, the present value of the projected dividend stream in year 41 and
6
beyond composes 41% of the total present value of the stock if the single-stage DCF
7
model is used. The present value of just the portion of the dividend stream projected for
8
year 100 and beyond accounts for 11% of the current stock value. Alternatively, using a
9
modified 3-stage DCF model which assumes the first stage to last 20 years and
10
convergence to the long-term growth rate of economy over the next 20 years, one arrives
11
at a cost of equity estimate of 12.38%, 230 basis points lower than the single-stage DCF
12
estimate.
13
Q.
HAS DR. VANDER WEIDE ALWAYS ARGUED THAT AN ASSUMPTION OF 40
14
TO 50 YEARS OF SUPERNORMAL GROWTH IS ALL THAT IS REQUIRED
15
TO USE THE SINGLE-STAGE MODEL?
16
A.
No. In his testimony in the 1997 Virginia UNE proceeding, his alternative, but similar,
17
argument was that the impact on present value of dividend growth rate assumptions
18
beyond 20 years was de minimis due to the effect of discounting.7 In fact, the present
19
value of constantly-growing dividends beyond year 20 accounts for more than 60% of the
20
company’s stock value using my hypothetical company.
7
Direct Examination of Dr. James H. Vander Weide, Before the State Corporation Commission of Virginia,
On Behalf of Bell-Atlantic-Virginia, Inc., Case No. PUC970005, pp. 204-5.
-9-
Surrebuttal Testimony of John I. Hirshleifer
1
Q.
WHAT COST OF CAPITAL WOULD RESULT IF A TWO-STAGE DCF MODEL
2
WITH A FIRST STAGE OF 20 YEARS OF SUPERNORMAL GROWTH IS
3
UTILIZED?
4
A.
If I were to calculate a cost of equity assuming 20 years of supernormal growth as Dr.
5
Vander Weide suggested, and then growth at the economy growth rate, the resulting cost
6
of equity would equal 11.48%, 320 basis points less than the single-stage DCF estimate.
7
Q.
HAS VERIZON ACKNOWLEDGED IN ANY OTHER CASE THAT THE
8
CHOICE BETWEEN DR. VANDER WEIDE’S SINGLE-STAGE DCF MODEL
9
AND YOUR THREE-STAGE MODEL HAS A SIGNIFICANT EFFECT ON THE
10
11
COST OF CAPITAL?
Yes. The cost of capital analysis submitted by Verizon’s predecessor NYNEX in the
A.
12
1996-97 UNE proceeding in Massachusetts illustrates this fact. Dr. Vander Weide was
13
NYNEX’s cost of capital witness in that case; and his methodology—including his
14
single-stage DCF—was essentially identical to the methodology he used here.8 In its
15
initial decision, the Massachusetts commission adopted Dr. Vander Weide’s methodology
16
in its entirety, with one exception: the commission directed NYNEX to file an alternative
17
cost of capital calculation that substituted AT&T witness Dr. R. Glenn Hubbard’s9 three-
18
stage DCF model for Dr. Vander Weide’s one-stage model.10 In response, NYNEX
8
In the Matter of the Board’s Review of Unbundled Network Elements Rates, Terms and Conditions of
Verizon-New Jersey, NJ BPU Docket No. TO00060356, 1 Tr. (Nov. 28, 2000), pp. 41-45 (Vander Weide).
9
Dr. Hubbard is currently Chairman of the President’s Council of Economic Advisors.
10
Id., pp. 47-50 (Vander Weide); D.P.U. 96-73/74, 96-75, 96-80/81, 96-83, 96-94 – Phase 4, Consolidated
Petitions of New England Telephone and Telegraph Company d/b/a NYNEX, Teleport Communications
Group, Inc., Brooks Fiber Communications, AT&T Communications of New England, Inc., MCI
Communications Company, and Sprint Communications Company, L.P., pursuant to Section 252(b) of the
- 10 -
Surrebuttal Testimony of John I. Hirshleifer
1
submitted a calculation showing that this one change would materially reduce the
2
weighted average cost of capital.11 When shown this document during cross-examination
3
in the New Jersey UNE cost proceeding, Dr. Vander Weide acknowledged that the choice
4
between his one-stage DCF and my three-stage DCF accounted for approximately 200
5
basis points of the difference between our cost of capital estimates.12
6
Q.
DR. VANDER WEIDE CLAIMS THAT VALUE LINE FORECAST DATA CAN
7
BE USED TO SUPPORT HIS ASSUMPTION THAT THE 5-YEAR I/B/E/S
8
GROWTH RATES FOR HIS GROUP OF “COMPARABLE” COMPANIES WILL
9
PERSIST INDEFINITELY IN THE FUTURE (VANDER WEIDE REBUTTAL P.
10
11
60). HOW DO YOU RESPOND TO THIS ASSERTION?
A.
I first note that in prior rebuttal testimonies, Dr. Vander Weide has claimed that Value
12
Line itself “publishes an estimate of each company’s long-run growth from internal
13
sources beyond the period beginning in 2003-2005”13 which according to him confirmed
14
that relatively high growth rates could be sustained for indefinitely long periods. This
15
assertion foundered because it was clear to any reader of the Value Line reports that
16
forecasts are provided for up to 5 years only. In fact, the Value Line reports cited by Dr.
17
Vander Weide provides no forecast beyond the year 2005. My staff additionally
Telecommunications Act of 1996, for arbitration of interconnection agreements between NYNEX and the
aforementioned companies (released Dec. 4, 1996) (“Phase 4 Decision”)
11
Letter dated December 18, 1996 from Bruce Beausejour, NYNEX counsel, to Mary L. Cottrell, Secretary to
MDTE.
12
NJ BPU Docket No. No. TO00060356, supra, 1 Tr. (Nov. 28, 2000), pp. 50-51 (Vander Weide).
13
See e.g., Responsive Testimony of Dr. Vander Weide on behalf of Bell Atlantic-New York, Case 98-C1357, filed June 26, 2000, p. 41.
- 11 -
Surrebuttal Testimony of John I. Hirshleifer
1
confirmed directly with Value Line that it does not make such long-term forecasts
2
asserted then by Dr. Vander Weide.
3
Q.
IF VALUE LINE DOES NOT MAKE FORECASTS BEYOND A FIVE-YEAR
4
HORIZON, HOW IS DR. VANDER WEIDE UTILIZING DATA OBTAINED
5
FROM VALUE LINE?
6
Dr. Vander Weide is saying that, by using the traditional book “b X r” method (where “b”
7
represents book earnings that are retained by the company, and “r” represents the book
8
return on book equity), he, not Value Line, is inferring a long-run growth rate by looking
9
at book retained earnings growth and assuming it will persist indefinitely into the future.
10
Dr. Vander Weide is simply taking data from Value Line reports and using a
11
method sometimes used in past traditional regulatory hearings for stable, regulated
12
industries which are not expected to experience significant variance from their historical
13
growth rates, and whose book value equities are approximately equal to the market value
14
of their equities. Similar to his assertions that 5-year analyst forecast growth rates are
15
expected by investors to persist forever, Dr. Vander Weide has not provided any evidence
16
that investors believe that supernormal growth rates obtained using this alternative
17
method will persist forever.
18
Q.
IS THE USE OF A METHOD WHICH IS BASED ON THE RETURN ON BOOK
19
EQUITY CONSISTENT WITH DR. VANDER WEIDE’S TESTIMONY
20
REGARDING THE APPROPRIATE CAPITAL STRUCTURE TO BE USED IN
21
ESTIMATING THE COST OF CAPITAL?
22
23
A.
No. Dr. Vander Weide has vociferously argued that a market-value capital structure is
the only one that can be used because a book capital structure is based on embedded costs
- 12 -
Surrebuttal Testimony of John I. Hirshleifer
1
and is backward looking. Consequently, his argument that a book value method is
2
appropriate for estimating growth, a critical input in estimating the forward-looking cost
3
of equity using the DCF method, is fatally discrepant.
4
Q.
DR. VANDER WEIDE CLAIMS THAT YOUR THREE-STAGE ASSUMPTIONS
5
ARE UNUSUAL AND ARBITRARY. (VANDER WEIDE REBUTTAL, P. 59).
6
WHAT ARE YOUR COMMENTS?
7
A.
This hardly constitutes a defense of the perpetual growth assumption for companies
8
currently experiencing high growth—an assumption that is not only arbitrary but
9
unequivocally incorrect. As cited in my direct testimony, for example, Professor William
10
Sharpe of Stanford and his co-authors indicated that sophisticated institutional investors
11
found the assumptions of single-stage and two-stage models overly simplistic, and that
12
they preferred three-stage models for providing the best combination of realism and ease
13
of use.14
14
Professor Aswath Damodoran of New York University illustrates many analytical
15
approaches for discounted cash flow modeling. Dr. Damodaran describes numerous
16
multiple-stage DCF models with varying formulations and characteristics. Dr.
17
Damodaran states that it is unrealistic to assume that a company with a high growth rate
18
would grow at this rate in perpetuity. Depending on how high the company’s current
19
growth rate is, Dr. Damodaran suggests different patterns and different lengths of time for
20
the high-growth period. After the period of high growth, Dr. Damodaran assumes that the
14
Sharpe, William F., Gordon J. Alexander and Jeffrey V. Bailey, Investments, Fifth Edition, Prentice Hall,
Englewood Cliffs, New Jersey, 1995, pp. 590-591.
- 13 -
Surrebuttal Testimony of John I. Hirshleifer
1
company will continue to grow at the stable growth rate of economy. He suggests the
2
following guidelines for defining the length of this first stage:15
3
If the Current Growth Rate Is:
Length of High Growth Period:
≤ 1 % higher than stable growth rate
No high growth
1-10 % higher than stable growth rate
5 years
> 10 % higher than stable growth rate
10 years
4
5
Notably, Dr. Damodaran never suggests that the single-stage DCF model should be used
6
for companies with growth rates significantly above the growth rate of the economy.
7
Dr. Damodaran indicates that for companies that have super normal growth rates
8
that are not particularly high, a two-stage or H Model could be used. Had I utilized either
9
Dr. Damodaran’s two-stage or H model with a 5-year initial stage for all of the individual
10
telephone holding companies, the cost of equity estimates would have been lower than
11
what I actually calculated.16 Consequently, my DCF model results are conservatively
12
high when compared to the results of these models.
13
Q.
UNDER WHAT CIRCUMSTANCES DOES DR. DAMODARAN SUGGEST
THAT THE “H MODEL” SHOULD BE USED?
14
15
A.
Dr. Damodaran states that:
15
Damodaran, Aswath, Applied Corporate Finance: A User’s Manual, John Wiley & Sons, 1999, p. 447.
16
In Dr. Damodaran’s two-stage model, the growth rate between years 5 and 20 equals the long-term growth
rate. In my model, however, the growth rates are higher than the long-term rate until year 20. Consequently,
the cost of equity resulting from my model will necessarily be higher than an estimate derived from Dr.
Damodaran’s 2-stage model.
- 14 -
Surrebuttal Testimony of John I. Hirshleifer
1
2
3
4
The H model is a two-stage model for growth, but unlike the
classical two-stage model, the growth rate in the initial growth phase
is not constant but declines linearly over time to reach the stablegrowth rate in steady stage.17
5
Dr. Damodaran indicates that the best use for this model is for firms that are growing
6
rapidly at the present, but for which the growth is expected to decline gradually over time
7
as their differential advantage over their competitors declines. Therefore, this model
8
appears suitable for use with telephone holding companies.
9
As shown in Attachment JH-4 of my direct testimony, the telephone holding
10
companies in the sample have five-year earnings growth rates between 11% and 14.8%
11
(4.7% and 8.5% above the stable growth rate of 6.29%). Had I applied Dr. Damodaran’s
12
H model to the set of comparables, the resulting costs of equity would have been lower
13
than those that I calculated. This is because in the H model the high initial growth rates
14
begin to decline immediately, while my DCF model assumes that the I/B/E/S five-year
15
growth rates do not decline over the first 5 years. After the initial growth phase, the
16
growth rate declines linearly to the long-term rate until year 20. Therefore, in every year
17
after the first my model utilizes higher growth rates than would be used in the H model.
18
Q.
WHAT DOES DR. DAMODARAN SAY ABOUT COMPANIES WHICH MIGHT
19
20
BE APPROPRIATE FOR THE CLASSICAL TWO-STAGE DCF MODEL?
A.
Damodaran suggests that one type of company for which this would be a suitable model
21
is a company:
… in an industry that is enjoying supernormal growth because
significant barriers to entry (either legal or as a consequence of
22
23
17
Damodaran, Aswath, Damodaran on Valuation: Security Analysis for Investment and Corporate Finance,
John Wiley & Sons, New York, 1994, p. 115.
- 15 -
Surrebuttal Testimony of John I. Hirshleifer
1
2
infrastructure requirements) can be expected to keep out new
entrants for several years.
3
4
5
6
7
8
9
The assumption that the growth rate drops precipitously from its
level in the initial phase to a stable rate also implies that this model is
more appropriate for firms with modest growth rates in the initial
phase. It is more reasonable, for instance, to assume that a firm
growing at 12% in the high-growth period will see its growth rate
drop to 6% after that than it is for a firm growing at 40% in the highgrowth period. 18 [emphasis added]
10
Q.
IF YOU ASSUMED THAT THIS WAS THE MOST APPROPRIATE MODEL TO
11
USE, WHAT IMPACT WOULD IT HAVE HAD ON YOUR DCF COST OF
12
EQUITY ESTIMATE?
13
A.
As I stated above, if I had instead used this model—which certainly appears applicable in
14
this case based on Dr. Damodaran’s analysis—it would have resulted in a lower cost of
15
equity than what I actually calculated.
16
Q.
DOES DR. DAMODARAN HIMSELF DESCRIBE A 3-STAGE DCF MODEL?
17
A.
Yes, although his 3-stage model is more complex model and differs in many ways from
18
the model I employ. Dr. Damodaran’s three-stage model allows for an initial period of
19
high growth, a transitional period in which growth declines, and a final stable-growth
20
phase; however, Dr. Damodaran states that his three-stage dividend discount model
21
requires year-specific payout ratios, growth rates and betas. The purpose for year-specific
22
betas is to compute distinct costs of equity for each phase of the model. This feature
23
allows an analyst to refine his valuation estimate by changing the expected cost of equity
24
in line with the analyst’s estimate of the changing risk characteristics of the firm being
18
Id., pp. 108-109.
- 16 -
Surrebuttal Testimony of John I. Hirshleifer
1
valued. My model does not assume changing payout ratios nor does it utilize betas.
2
Because it assumes that the cost of equity changes in each phase, Dr. Damodaran’s 3-
3
stage model cannot be used to solve for a cost of equity.
4
B.
5
6
7
1.
8
9
10
11
The Risks Of Supplying Unbundled Network Elements In
Maryland Do Not Justify The Use Of DCF Comparison Groups
That Include Non-Telephone Companies.
Q.
TELRIC cost principles do not require the Commission to
assume that Verizon Maryland faces intense competition
regardless of the facts.
DR. VANDER WEIDE CLAIMS THAT YOUR COST OF CAPITAL ESTIMATE
12
IS INCONSISTENT WITH THE FORWARD-LOOKING ECONOMIC COST
13
PRINCIPLES ESTABLISHED BY THE FCC (VANDER WEIDE REBUTTAL, P.
14
2). IS THIS CORRECT?
15
A.
No. Dr. Vander Weide overlooks numerous key provisions of the FCC August 8, 1996
16
Order which provide guidance for the determination of costs of capital associated with
17
UNEs.
18
Dr. Vander Weide states that:
19
20
21
22
23
[T]he forward-looking economic cost principle is economically
relevant only in a competitive market for telecommunications
services. Thus, the forward-looking economic cost principle, at its
heart, is based on the assumption that the market for local exchange
services is fully competitive. [Vander Weide direct testimony p. 35]
24
25
26
27
A proper definition of the cost of capital for use in Verizon MD’s
forward-looking cost studies is based on the assumption that the
market for local exchange services is competitive. [Vander Weide
direct testimony p. 49]
- 17 -
Surrebuttal Testimony of John I. Hirshleifer
1
2
3
[T]he FCC’s cost study principles require that cost studies
“replicate…the conditions of a competitive market” for unbundled
network elements. [Vander Weide direct testimony p. 44]
4
5
6
7
8
Mr. Hirshleifer’s cost of capital estimate is intended to be used as an
input to AT&T/WorldCom’s forward-looking economic cost studies,
which, according to the FCC, should produce rates that replicate the
results of a competitive telecommunications market. [Vander Weide
rebuttal testimony p. 7]
9
Dr. Vander Weide also argues in this testimony (Vander Weide rebuttal p. 7) that the
10
FCC requires the assumption of a competitive telecommunications market. This
11
assumption is contrary to the purpose of the 1996 Act and ¶¶ 688 and 702 of the FCC
12
August 8, 1996 Order. In particular, ¶ 702 of the FCC’s Local Competition Order of
13
August 8, 1996 makes explicit statements that entirely disprove Dr. Vander Weide’s
14
interpretation. I highlighted those statements at page 20 of my rebuttal testimony. In
15
contrast, the cost of capital that I have recommended is a risk-adjusted cost of capital
16
referenced in ¶ 702, and takes into account all relevant risks.
17
18
19
The FCC August 8, 1996 Order is replete with statements in other sections that
further contradict Dr. Vander Weide’s erroneous assumptions.
For example,
20
21
22
23
The 1996 Act encourages competition by removing barriers to entry
and providing an opportunity for potential new entrants to purchase
unbundled incumbent LEC network elements to compete efficiently
to provide local exchange services. [¶ 672] [emphasis added]
24
25
26
27
28
29
The incumbent LEC offerings to be priced using this methodology
generally will be “network elements,” rather than
“telecommunications services,” as defined by the 1996 Act. More
fundamentally, we believe that TELRIC-based pricing of discrete
network elements or facilities, such as local loops and switching, is
likely to be much more economically rational than TSLRIC-based
- 18 -
Surrebuttal Testimony of John I. Hirshleifer
1
2
3
pricing of conventional services, such as interstate access service
and local residential or business exchange service. [¶ 678]
[emphasis added]
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
Congress recognized in the 1996 Act that access to the incumbent
LEC’s bottleneck facilities is critical to making meaningful
competition possible. As a result of the availability to competitors of
the incumbent LEC’s unbundled elements at their economic cost,
consumers will be able to reap the benefits of the incumbent LEC’s
economies of scale and scope, as well as the benefits of competition.
Because a pricing methodology based on forward-looking costs
simulates the conditions in a competitive marketplace, it allows the
requesting carrier to produce efficiently and to compete effectively,
which should drive retail prices to their competitive levels. We
believe that our adoption of a forward-looking cost-based pricing
methodology should facilitate competition on a reasonable and
efficient basis by all firms in the industry by establishing prices for
interconnection and unbundled elements based on costs similar to
those incurred by the incumbents, …[¶ 679] [emphasis added]
19
20
21
22
…we find that incumbent LECs must prove to the state commission
the nature and magnitude of any forward-looking cost that it seeks to
recover in the prices of interconnection and unbundled network
elements. [¶ 680] [emphasis added]
23
It is evident from a reading of the FCC August 8, 1996 Order that the FCC does
24
not accept Dr. Vander Weide’s argument that cost should be based on
25
telecommunications services, nor that there should be a hypothetical assumption that the
26
risks of a fully competitive market exist for UNEs when in fact they do not. While Dr.
27
Vander Weide asserts in this proceeding that actual UNE competition justifies an
28
assumption of a fully competitive market, he has not met his burden of proof in this
29
regard. Dr. Vander Weide’s failure to provide convincing evidence means that his
30
already-questionable selection of the S&P Industrials as the basis for his cost of capital
31
calculation is not supportable.
- 19 -
Surrebuttal Testimony of John I. Hirshleifer
1
Q.
HAS ANY COURT AGREED WITH YOU ABOUT THE RISK ASSUMPTIONS
2
3
IMPLIED BY THE TELRIC STANDARD?
A.
Yes. In the 1997 UNE proceeding before the Delaware PSC, Dr. Vander Weide argued
4
for Bell Atlantic, as he now does in this Maryland proceeding, that the TELRIC standard
5
requires state commissions to assume that the supplier of unbundled network elements
6
faces intense competition. The Delaware Public Service Commission rejected this
7
argument for the same reasons I offer here. Bell Atlantic appealed to the United States
8
District Court in Delaware. The court upheld the Delaware Commission on this point,
9
again for the same reasons I have offered here:
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
Bell points to an apparent contradiction in assuming instantly
competitive prices for network elements (even though no such
competition now exists) but, in the context of determining cost of
capital, assuming little competition and, consequently, low costs of
capital. … The Telecommunications Act attempts to recreate the
prices that a hypothetical efficient company would charge for its
network elements and services in a competitive market. Indulging in
this fiction, however, does not change the fact that ILECs like Bell
do not face the same competitive risks as firms operating in a
competitive market. Indeed, ILECs have had no competition for
decades, and they will face little competition in the market for
network elements in the near future. See August 8, 1996 Order ¶
702, at 353. Therefore, in introducing competition in the local
telephone market, it makes perfect sense to recreate competitive
prices while acknowledging that the current lack of competition
warrants reduced costs of capital.19
Q.
HAS ANY OTHER ECONOMIC CONSULTANT TO VERIZON AGREED WITH
27
THE U.S. DISTRICT COURT ON THIS POINT?
19
Bell Atlantic-Delaware, Inc. v. McMahon, 80 F.Supp.2d 218 (D. Del. 2000) at 240 n. 19.
- 20 -
Surrebuttal Testimony of John I. Hirshleifer
1
Yes. National Economic Research Associates (“NERA”) is a consultant to Verizon. Dr.
A.
2
William Taylor, Verizon’s economic witness in multiple TELRIC proceedings, is a senior
3
vice president with NERA. In the UNE cost proceeding before the New York Public
4
Service Commission last year, an excerpt of a report authored by NERA was entered as
5
evidence. That excerpt states in part:
6
7
In terms of the more general concept of incremental costs, TELRIC
maintains the following specific assumptions.
8
9
10
11
12
First, the business decision being modeled is that of a hypothetical
local exchange carrier that offers unbundled elements to retail
providers (possibly itself) at undifferentiated prices. Hence the
increments in question are the total volume for the elements
demanded by the retail providers.
13
14
15
16
17
18
19
Second, the time horizon over which the ILEC offers the wholesale
elements is assumed to be the longest of the long-run. Implicit in
this definition are the assumptions that (1) the ILEC will effectively
be a monopolist in the provision of network elements for the
indefinite future and (2) competitors will need to obtain such
elements to compete over this time frame.20 [footnotes omitted;
emphasis added]
20
Dr. Taylor has himself acknowledged the distinction between attempting to
21
replicate the costs of a firm in a competitive market, and arbitrarily assuming that an
22
incumbent monopoly carrier faces the risks of a firm in a competitive market. Testifying
23
in the UNE proceeding in Virginia in 1997, Dr. Taylor agreed that it was not unheard of
24
for regulators to set prices in noncompetitive markets that replicate the prices that would
20
Excerpt from “An Economic Evaluation of Network Cost Models”, NERA, August 7, 2000, Exhibit 408,
State of New York Public Service Commission, Proceeding on Motion of the Commission to Examine New
York Telephone Company’s Rates for Unbundled Network Elements, Case 98-C-1357.
- 21 -
Surrebuttal Testimony of John I. Hirshleifer
1
result from a competitive market.21 Moreover, he conceded that it was possible for a
2
regulatory standard that sets rates at competitive levels to coexist with an environment in
3
which the regulated firm faces less competitive risks than a competitive firm would
4
face.22
5
DR. VANDER WEIDE CITES A FOOTNOTE TO THE FCC’S BRIEF TO THE
Q.
6
SUPREME COURT WHICH STATES THAT “AN APPROPRIATE COST OF
7
CAPITAL DETERMINATION TAKES INTO ACCOUNT NOT ONLY EXISTING
8
COMPETITIVE RISKS… BUT ALSO RISKS ASSOCIATED WITH THE
9
REGULATORY REGIME TO WHICH A FIRM IS SUBJECT.” (VANDER
10
WEIDE REBUTTAL, P. 21) IS THIS STATEMENT SUPPORTIVE OF YOUR
11
APPROACH, OR OF DR. VANDER WEIDE’S?
12
A.
It is helpful to mine. The quoted language makes clear that the cost of capital used in
13
determining UNE prices must take account of the particular competitive risks, existing
14
and anticipated by the market, entailed in supplying UNEs, as well as regulatory risks
15
anticipated by the market. By necessary implication, the analyst may not simply assume,
16
without analysis, that the relevant risks are those of a vigorously competitive market. My
17
cost of capital analysis considers both the competitive and regulatory risks that would
18
face a wholesale supplier of UNEs in Maryland. In contrast, Dr. Vander Weide’s use of
21
Ex Parte to Determine Prices Bell Atlantic—Virginia, Inc. Is Authorized to Charge Competing Local
Exchange Carriers in Accordance with the Telecommunications Act of 1996 and Applicable State Law,
Virginia State Corporation Commission, Case No. PUC970005, Tr. at 580 (statement of William E.
Taylor).
22
Id. at 581 (emphasis added).
- 22 -
Surrebuttal Testimony of John I. Hirshleifer
1
an incorrect comparison group based on S&P Industrial companies considers neither set
2
of risks.
3
IS DR. VANDER WEIDE’S POSITION REGARDING AN ASSUMPTION OF
Q.
COMPETITIVE RISKS CONSISTENT WITH VERIZON’S OWN POSITION?
4
5
A.
No. In its brief to the Supreme Court supporting its opposition to the use of TELRIC,
6
Verizon stated that:
7
8
9
10
…TELRIC similarly presumes that carriers in its fictional world of
constant network replacement would nonetheless continue to have
the same cost of capital established for incumbents in the stable,
low-risk monopoly system of the past…
11
12
13
14
15
The FCC’s order, moreover, requires the States to start with the
existing rate of return, and places the burden on incumbents to
demonstrate with specificity that the business risks—defined
exclusively in terms of facilities-based entry by competitors—justify
any change in the rate of return.23
16
Verizon cites ¶¶ 687, 688 and 702 of the FCC’s August 8, 1996 Order in support of these
17
statements.
2.
18
19
20
21
22
Dr. Vander Weide exaggerates the competitive risk of VZMD’s local telephone service generally and fails to
distinguish between the competitive risks of providing UNEs
at wholesale and providing local telephone services at retail.
HAS ANY COURT NOTED DR. VANDER WEIDE’S FAILURE TO
Q.
23
DISTINGUISH BETWEEN THE COMPETITIVE RISKS OF PROVIDING UNES
23
Brief for Petitioners, On Writ of Certiorari to the United States Court of Appeals for the Eighth Circuit, In
the Supreme Court of the United States, October Term, 2000, No. 00-511.
- 23 -
Surrebuttal Testimony of John I. Hirshleifer
1
AT WHOLESALE AND PROVIDING LOCAL TELEPHONE SERVICES AT
2
RETAIL?
3
4
A.
Yes. In Bell Atlantic-Delaware, Inc. v. McMahon, 80 F.Supp.2d 218, 240 (D.Del. 2000),
the court stated as follows (emphasis in original):
5
6
7
8
9
10
11
12
13
14
15
In assessing Bell’s case for an elevated cost of equity, the Hearing
Examiners criticized the testimony of Bell’s expert, Dr. James Vander
Weide. The Examiners noted that Vander Weide based his cost of equity
on the risk associated with Bell’s retail business instead of on the future
demand for Bell’s network elements that it will sell at wholesale. AT&T’s
expert, Bradford Cornell, also criticized Vander Weide’s analysis as
“ignor[ing] the critical fact that the business at hand in this proceeding is
not local retail phone service that already exists, but rather the new
business of leasing of network elements at wholesale for use in providing
competitive phone services to an existing retail market.” [citation
omitted] The distinction between wholesale and retail is crucial.
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
Retail competition is competition for the end user of telephone service.
That sort of competition is not at issue when determining the risks
associated with leasing unbundled network elements (e.g., loops and
switches) at wholesale. The risks associated with leasing “bottleneck”
network elements at wholesale is less than that associated with
competition for retail service. See August 8, 1996 Order ¶ 702, at 353
(noting that network elements “generally are bottleneck, monopoly
services that do not now face significant competition”). This is so because
Bell often is the only provider of these network elements, and it is to Bell
that new entrants must come to lease or purchase loops, switches, or other
network elements. Thus, even if retail competition intensifies, Bell’s
prominence as a wholesale provider of network elements will remain
largely unaffected—at least until new entrants build their own networks.
[footnote omitted] Accordingly, the Hearing Examiners correctly rejected
Vander Weide’s testimony as impermissibly attributing the risks of retail
competition to the competition in the sale of unbundled network elements.
See August 8, 1996 Order ¶ 691, at 348 (explaining that, “[o]nly those
costs that are incurred in the provision of network elements in the long run
shall be directly attributable to those elements”).
- 24 -
Surrebuttal Testimony of John I. Hirshleifer
1
Q.
IN HIS REBUTTAL TESTIMONY DR. VANDER WEIDE HAS ARGUED THAT
2
“SIGNIFICANT FACILITIES-BASED COMPETITION ALREADY EXISTS FOR
3
LOCAL EXCHANGE SERVICES IN MARYLAND, AND INVESTORS EXPECT
4
FUTURE COMPETITION TO INCREASE RAPIDLY.” (VANDER WEIDE
5
REBUTTAL, P. 24) WHAT IS YOUR RESPONSE TO THIS ARGUMENT?
6
A.
I have reviewed the evidence that Dr. Vander Weide suggests for competition in
7
Maryland. In general, his examples do not clearly distinguish between facilities offered
8
by network competitors, and end users that are retail customers of competitors but
9
ultimately use VZ-MD’s network elements. He thus ignores the crucial distinction
10
recognized by the U.S. District Court in Delaware. As stated by ABN Amro in its
11
January 20, 2000 report covering Bell Atlantic, for example, 1.1 million access lines that
12
Bell Atlantic supposedly lost to competition as of the close of the third quarter of 1999
13
were in fact being supplied by Bell Atlantic on a wholesale basis.24 In its August 2000
14
announcement of quarterly results, Verizon stated that its “wholesale business provid[es]
15
nearly 2.9 million switched access lines and 541,000 unbundled loops.”25 To the extent
16
that competitors are using Verizon’s network elements, Bell Atlantic retains those
17
facilities revenues and has lost nothing on the UNE level, which is the sole subject of this
18
proceeding.26 Without evidence that the network facilities business—as opposed to the
19
local exchange or telecommunications businesses, has become dramatically
24
ABN Amro also added that “[o]ffsetting these setbacks, Bell Atlantic added nearly 6 million lines from
year-end 1995 through the third quarter of 1999, a 3.2% CAGR.” (p. 24) [emphasis added]
25
Verizon Press Release, “Verizon Communications Announces Second Quarter Results,” August 8, 2000.
26
In its May 15, 2000 “Telecom -Wireline” report Morgan Stanley Dean Witter highlighted that in the first
quarter of 2000 it “saw some renewed strength in local and access revenues. Sales of value added services
and strong wholesale business more than offset local competition.”
- 25 -
Surrebuttal Testimony of John I. Hirshleifer
1
competitive—Verizon has failed to satisfy at least one element of its burden of proof
2
required by paragraphs 680 and 702 of the FCC August 8, 1996 Order.
3
Q.
DR. VANDER WEIDE ARGUES THAT THE MINIMAL CLEC PENETRATION
4
PORTENDS GREAT UNE COMPETITION IN THE FUTURE. (VANDER
5
WEIDE REBUTTAL, P. 24-25) HOW DO YOU RESPOND TO THIS
6
ARGUMENT?
7
Dr. Vander Weide’s predictions about future competitive developments are empty
A.
8
speculation; they ignore the high barriers to entry that appear likely to remain in the
9
foreseeable future. The recent bankruptcies of one-time potential competitors such as
10
Northpoint, Rhythms and Covad merely underscore the fanciful nature of Dr. Vander
11
Weide’s projections of competition.
12
Even on the level of retail local exchange, Verizon is experiencing a decline in
13
competitive entry:
14
15
16
17
18
19
20
21
22
23
24
Verizon is seeing a significant slowdown in the sequential growth in
its reported wholesale lines as CLECs are hit by the drop off in
funding. In Q2 00 533,000 wholesale lines were added, in Q3 00
364,000 lines were added, in Q4 00 335,000 lines were added, and in
Q1 01 only 84,000 wholesale lines were added. If this trend
continues, we could see far less share loss to competitors that we
currently have factored in to our projections. With most CLECs in
dire financial straits right now, any upside on the competitive front
could be a powerful offset to any potential lingering economic
weakness.27
Q.
IS IT POSSIBLE THAT FACILITIES-BASED COMPETITION, WHETHER
25
EXISTING TODAY OR EMERGING IN THE DISTANT FUTURE, WOULD
27
Dresdner, Kleinwort, Wasserstein Research, "Verizon Communications", June 29, 2001, p. 2.
- 26 -
Surrebuttal Testimony of John I. Hirshleifer
1
DEPRIVE VZ-MD OF REVENUES IN THE BUSINESS OF LEASING
2
UNBUNDLED NETWORK ELEMENTS?
3
A.
Given enough time, anything is possible. The market has already incorporated its
4
expectations of any such losses, however, in the price of Verizon’s stock. If such fears of
5
competition were significant to investors’ estimates of the required cost of capital, they
6
have already accounted for them in valuing Verizon’s stock.
7
Morgan Stanley states that,
8
9
10
11
12
13
14
15
16
17
18
There is no doubt that competitive pressures are significant in the
industry, with the recent opening up of New York to Bell Atlantic
long distance heralding the new era. Nevertheless, we continue to
firmly believe that the pie is growing, and those companies who
execute effectively can succeed despite competitive pressures. The
Bells have already absorbed significant local and toll competition
from CLECs as well as significant rate reductions over the past
several years. We also see the control of the customer, and the local
loop combining with scale advantages to create significant
competitive leverage for the local phone companies.28
Q.
FROM THE STANDPOINT OF FINANCE THEORY, IS COMPETITIVE RISK
19
GENERALLY ACCEPTED AS BEING RELEVANT TO THE DETERMINATION
20
OF THE COST OF CAPITAL?
21
A.
As I discussed in my direct testimony, capital market theory indicates that the market
22
would not increase the cost of capital for competitive risks which investors can diversify
23
away by purchasing a diversified portfolio of stocks.
24
Q.
IN HIS REBUTTAL TESTIMONY DR. VANDER WEIDE HAS TESTIFIED
25
THAT TELEPHONE HOLDING COMPANIES ARE ACTUALLY LESS RISKY
28
Morgan Stanley Dean Witter, “Telecom - Wireline”, January 21, 2000.
- 27 -
Surrebuttal Testimony of John I. Hirshleifer
1
THAN THE LEC’S BECAUSE THEY HAVE DIVERSIFIED (VANDER WEIDE
2
REBUTTAL, P. 41). CAN THIS BE TRUE?
3
A.
No. As I stated in my rebuttal testimony, engaging in businesses which are systematically
4
riskier than the wholesale network element business, such as wireless or international
5
ventures, will always make the risk of the telephone holding company greater than that of
6
the wholesale network element business. Overall risk can never fall because of the
7
acquisition of systematically riskier businesses.
8
In its last access charge rate represcription proceeding, the FCC stated that:
9
10
11
12
It seems counterintuitive to suggest, as Bell Atlantic does, that
diversification into riskier businesses could actually reduce the
business risk of an RHC so that it is lower than the business risk of
the regulated business.29
13
14
15
16
17
18
19
However, the record does show that the RHCs are also involved in
activities which are perceived as riskier than their regulated
telephone business. We therefore find that we should give some
weight in our decision to the possibility that a cost of equity estimate
for an RHC as a whole company might somewhat overstate the cost
of equity for interstate access service alone. [emphasis added] (Id., ¶
86, p. 7517).
3.
20
21
22
23
Q.
WHAT REASON DOES DR. VANDER WEIDE GIVE FOR USING AN S&P
24
25
Dr. Vander Weide has not offered plausible or consistent
reasons why telephone companies should not form the
appropriate DCF comparison group.
INDUSTRIAL COMPARABLE SET?
A.
He states that:
29
FCC Order 90-315, In the Matter of Represcribing the Authorized Rate of Return for Interstate Services of
Local Exchange Carriers, CC Docket No. 89-624, September 19, 1990, ¶ 84, p. 7517.
- 28 -
Surrebuttal Testimony of John I. Hirshleifer
1
2
3
4
Since the S&P Industrials are a group of competitive firms whose
composite risk is average, I have selected them as a reasonable proxy
for Verizon MD’s risk of providing unbundled network elements in a
competitive market. [Vander Weide rebuttal, pp. 50-51]
5
Notably, Dr. Vander Weide again offers no proof for this assertion.
6
Q. YOU PREVIOUSLY CITED DR. VANDER WEIDE’S BELIEF THAT THE UNE
7
MARKET IS ALREADY COMPETITIVE (VANDER WEIDE, P. 33). IF SO,
8
COULD HE HAVE JUST USED A SAMPLE OF TELEPHONE HOLDING
9
COMPANIES?
10
A.
Of course.
11
Q.
WHAT ARGUMENTS DOES DR. VANDER WEIDE CURRENTLY OFFER FOR
12
13
NOT USING THE CLOSEST COMPARABLE COMPANIES?
A.
14
15
16
17
18
19
20
21
The DCF and CAPM Models provide more uncertain estimates of
the cost of equity for companies such as the holding companies that
are experiencing radical restructuring and profound regulatory,
organizational and technological change. In addition, the four or five
telecommunications holding companies are simply too small a group
to obtain reliable cost of equity estimates. [Vander Weide rebuttal, p.
49]
Q.
22
23
Dr. Vander Weide’s reasons are as follows:
WAS DR. VANDER WEIDE OFFERING A DIFFERENT ARGUMENT IN THE
RECENT NEW YORK UNE COST PROCEEDING?
A.
Yes. In his New York UNE cost proceeding rebuttal testimony, Dr. Vander Weide
24
argued that telephone holding companies either involved in mergers or subject to merger
25
speculation could not be used as proxies for other telephone holding companies because
26
“the projected earnings growth associated with the mergers is not reflected in the
- 29 -
Surrebuttal Testimony of John I. Hirshleifer
1
analysts’ growth rates” used in DCF analyses. (Vander Weide New York responsive
2
testimony, p. 30) He assumed that the stock prices would immediately rise upon the
3
merger news and concluded that the DCF cost of equity would be biased downward. In
4
support of his claim, Dr. Vander Weide produced in his New York responsive testimony
5
an exhibit showing gradually rising I/B/E/S forecast growth rates for several merging
6
telecommunications companies.
7
I showed that these data did not support Dr. Vander Weide’s claim. First, Dr.
8
Vander Weide assumed that the increase of I/B/E/S forecast growth rates over the last
9
several years result solely from mergers. However, each of the companies selected by Dr.
10
Vander Weide has engaged in numerous high-growth endeavors during the period
11
illustrated in his exhibit. Obviously, a far greater proportion of growth rate increases
12
would derive from high-growth businesses than would arise from the cost cutting
13
measures which mergers make possible. Industry analysts have stated that “data and
14
wireless continue to expand [LECs’] piece of the revenue pie.”30 ABN AMRO reiterated
15
that it saw “three catalysts of Bell Atlantic growth: high-speed data, global wireless and
16
long-distance entry.”31 Verizon’s international business segment, as an example, grew by
17
18.6% in 2000 and 21.2% in 1999.32
18
Second, had Dr. Vander Weide investigated these mergers, he would have found
19
that the stock price of at least one of the companies declined after the announcement of
30
Morgan Stanley Dean Witter, “Telecom - Wireline”, May 15, 2000.
31
ABN AMRO, “Bell Atlantic Corporation”, January 20, 2000.
32
Verizon Communications Inc. SEC Form 10-K405 for the period ending 12/31/00.
- 30 -
Surrebuttal Testimony of John I. Hirshleifer
1
the merger.33 A decline in the stock price would result in a higher cost of equity if DCF
2
model calculations were performed keeping all other parameters unchanged. Moreover,
3
after the merger was announced, the aggregate market capitalization of the two merging
4
companies went down in 5 out of 6 cases.
5
Another example in the news was the failed WorldCom/Sprint merger. When
6
U.S. antitrust officials announced their intentions to investigate the planned merger,
7
Sprint’s stock price rose by 8.7% instead of declining. The day the Department of Justice
8
filed suit to block the merger, WorldCom’s stock price increased by 12.28%, while
9
Sprint’s stock price decreased by 9.56%. Contrary to Dr. Vander Weide’s assertions, in
10
the instances where a company’s stock price is depressed because of merger anticipation,
11
such as MCI WorldCom’s stock, a DCF calculation would have provided a higher, not
12
lower cost of equity estimate.
13
Q.
HAS DR. VANDER WEIDE ATTEMPTED TO DETERMINE WHICH
14
COMPANIES IN HIS S&P INDUSTRIAL SAMPLE ARE SUBJECT TO
15
MERGER OR ACQUISITION SPECULATION IN THIS PROCEEDING?
16
A.
No. And it is very ironic that Dr. Vander Weide does not cull out telephone holding
17
companies that he suggested would yield downwardly-biased cost of equity estimates in
18
the New York UNE cost proceeding. While Dr. Vander Weide has argued that mergers
19
and merger speculation are important to the selection of appropriate comparables, he has
20
not analyzed the companies in his sample to determine which ones are in industries that
21
have or are anticipating merger activity. Standard & Poor’s itself acknowledges that
33
SBC stock declined after the announcement of each of its three mergers: SBC/Pacific Telesis, SBC/SNET
and SBC/Ameritech. In three other mergers mentioned by Dr. Vander Weide, all stocks declined after the
announcements except for Bell Atlantic stock’s in connection with its merger with NYNEX.
- 31 -
Surrebuttal Testimony of John I. Hirshleifer
1
“[t]he S&P 500 is a great list of merger candidates—the companies are well known and
2
widely followed on Wall Street. The list is one of the first places an investment banker
3
turns when searching for a big target.”34
4
Other examples of industries represented in the S&P Industrials that have been
5
involved in substantial merger activity are the banking industry (Wells Fargo, Chase
6
Manhattan and US Bancorp merged with various smaller banks); the chemical industry
7
(Eastman Chemical, Great Lakes Chemical have been involved in mergers); the food and
8
beverages industry (Bestfoods, ConAgra, General Mills, HJ Heinz, Seagram, Kellogg
9
participated in mergers); the entertainment industry (Time Warner merger with AOL,
10
CBS with Viacom); the newspaper industry (Times Mirror merged with Tribune), etc.
11
Q.
HAS DR. VANDER WEIDE ATTEMPTED TO DETERMINE AND CULL OUT
12
COMPANIES IN HIS S&P INDUSTRIAL SAMPLE THAT ARE SUBJECT TO
13
RESTRUCTURING, OR TO REGULATORY, ORGANIZATIONAL OR
14
TECHNOLOGICAL CHANGE WHICH IN HIS VIEW WOULD MAKE DCF
15
COST OF EQUITY ESTIMATES MORE UNCERTAIN?
16
A.
No, he has not.
17
Q.
IN YOUR EXPERIENCE, HAS DR. VANDER WEIDE EVER
18
ACKNOWLEDGED THAT TELEPHONE HOLDING COMPANIES ARE
19
APPROPRIATE COMPARABLES FOR OTHER TELEPHONE HOLDING
20
COMPANIES?
34
www.spglobal.com/howmany.html.
- 32 -
Surrebuttal Testimony of John I. Hirshleifer
1
A.
Not that I can recall. For example, Dr. Vander Weide proposed the use of the S&P 500 to
2
verify the reasonableness of the USTA cost of equity estimate in the FCC’s access charge
3
rate represcription proceeding completed in 1990, well before the 1996 Act.35 The FCC
4
properly rejected the use of Dr. Vander Weide’s index approach in the 1990 proceeding.36
5
Dr. Vander Weide’s longstanding advocacy of S&P Index companies as a DCF proxy
6
group for local telephone companies clearly predates the recent regulatory and
7
competitive developments that ostensibly justify his approach.
8
Q.
DR. VANDER WEIDE OBJECTS TO YOUR USE OF A SAMPLE THAT ONLY
9
INCLUDES FIVE TELEPHONE COMPANIES. (VANDER WEIDE REBUTTAL,
10
PP. 48-50) IS IT PREFERABLE TO USE A LARGER SAMPLE OF
11
NONCOMPARABLE COMPANIES AS HE ADVOCATES?
12
A.
Obviously not. The purpose for using a larger sample, when there are enough comparable
13
companies that can be included in that sample, is to reduce measurement error in order to
14
arrive at averages that more closely represent the true mean for the comparable company
15
group. Even if one’s sample by necessity is small, however, the averaging process
16
reduces measurement error. In contrast, averaging over a group of noncomparable
17
companies yields a mean that in no way measures the parameter one is attempting to
18
estimate for the subject company or for its industry.
19
As Myers and Borucki put it:
35
“Bell Atlantic asserts that because the S&P 500 is a group of large industrial firms, it is an excellent
benchmark for determining the interstate access cost of equity and can be used to verify the reasonableness
of the results of the USTA cluster analysis. USTA argues that the S&P 400 is a proxy for the competitive
marketplace.” FCC Order 90-315, In the Matter of Represcribing the Authorized Rate of Return for
Interstate Services of Local Exchange Carriers, CC Docket No. 89-624, September 19, 1990, ¶144, p. 7524.
36
Id. at ¶162, p. 7526.
- 33 -
Surrebuttal Testimony of John I. Hirshleifer
In real life, errors in estimating investors’ forecasts of future growth
are inevitable. The errors will occur even if all the DCF method’s
assumptions are satisfied. This does not invalidate the method; all
approaches to measuring the cost of equity are liable to random
error. Responsible analysts attempt to average across similar
companies whenever possible.37
1
2
3
4
5
6
According to Dr. Vander Weide’s flawed reasoning, it would be preferable to use
7
8
a large sample of unrelated stocks trading on the Paris Bourse over a smaller sample of
9
comparable telephone holding companies.
10
Q.
DOES DR. VANDER WEIDE FOLLOW THIS RULE OF NOT USING SMALL
11
12
SAMPLES OF COMPARABLE COMPANY GROUPS?
A.
No. For example, in his New York responsive testimony, Dr. Vander Weide presented
13
alleged “tests of reasonableness” similar to those that he presents in his rebuttal testimony
14
for this proceeding. He calculated DCF results for groups of companies that he described
15
as “interexchange carriers” and “large industrials.” For the “interexchange carrier” group
16
he averaged over a sample of only two companies. For his “large industrial” group, he
17
averaged over a sample of just three companies. In both instances, he could have
18
included more companies that fell within those groups, but consciously chose not to do
19
so. In the case of “large industrials”, his own S&P Industrial sample group would have
20
allowed him to use a sample group of substantial size. In the present case he averages
21
over only three companies for his group of “natural gas distribution companies”, and
22
averages over a group of only four “telecommunications companies”, with Verizon itself
23
notably excluded from the sample.
37
Stewart C. Myers and Lynda S. Borucki, “Discounted Cash Flow Estimates of the Cost of Equity Capital—
A Case Study”, Financial Markets, Institutions & Instruments, vol. 3, no. 3, New York University Salomon
Center, 1994, p. 17. [emphasis added].
- 34 -
Surrebuttal Testimony of John I. Hirshleifer
C.
1
2
3
4
Q.
The S&P Industrial Companies Selected By Dr. Vander Weide
Are Not A Valid Comparison Group For A DCF Analysis Of The
Cost Of Equity Of The Network Element Business.
DR. VANDER WEIDE ALLEGES THAT THE GROUP OF S&P INDUSTRIALS
5
PROVIDES “A REASONABLE PROXY FOR VERIZON VA’S RISK OF
6
PROVIDING UNBUNDLED NETWORK ELEMENTS IN A COMPETITIVE
7
MARKET.” (VANDER WEIDE REBUTTAL, PP. 50-51) HAS ANY COURT
8
EVALUATED THE LEGITIMACY OF DR. VANDER WEIDE’S USE OF S&P
9
INDUSTRIAL COMPANIES AS COMPARABLES FOR TELEPHONE
10
11
COMPANIES?
A.
Yes. The United States District Court in Delaware, in upholding the decision of the
12
Delaware Public Service Commission in 1997 to approve a weighted average cost of
13
capital of 10.28% for UNE pricing, quoted with approval the following findings:
The [Delaware PSC Hearing] Examiners also discounted Vander Weide’s
analysis because he based his cost of equity calculation on the assumption
that Bell’s business was as risky as that of a Standard & Poor’s (“S&P”)
300 industrial firm. … Because these S&P firms employ a variety of
technologies and enjoy a wide array of market shares, the Hearing
Examiners concluded that the risks faced by these firms said little about
the risk Bell faced in the market for unbundled network elements. …
Instead, they accepted AT&T’s assessment of Bell’s risk, which it
premised upon the risk experienced by other telephone holding companies.
14
15
16
17
18
19
20
21
22
23
Bell Atlantic-Delaware, Inc. v. McMahon, 80 F.Supp.2d 218, 241 (D.Del. 2000)
24
(citations omitted).
25
26
27
D.
Dr. Vander Weide’s Miscellaneous Criticisms of My DCF Analysis
of Equity Costs Are Without Merit
- 35 -
Surrebuttal Testimony of John I. Hirshleifer
1
2
Q.
DR. VANDER WEIDE PROVIDES MANY ARGUMENTS TRYING TO
3
SUPPORT THE USE OF QUARTERLY COMPOUNDING. (VANDER WEIDE
4
REBUTTAL, PP. 53-55) DOES HE UNDERSTAND YOUR POINT REGARDING
5
WHY THE COMMISSION SHOULD NOT USE QUARTERLY
6
COMPOUNDING?
7
A.
No. Dr. Vander Weide forgets that UNE rates set by this Commission are amounts paid
8
to companies like VZ-MD, not to investors. Dr. Vander Weide’s method of calculation
9
would therefore give VZ-MD the benefit of quarterly compounding which it would not
10
otherwise get. As I noted in my direct testimony, this is best understood by comparing
11
VZ-MD to a company whose prices are completely unregulated. Times Mirror
12
Corporation, for example, (a newspaper publisher), received its cash flows from
13
subscribers approximately monthly. It then could reinvest those funds monthly to
14
increase its return on a monthly compounded basis. When Times Mirror paid dividends
15
to its investors, it did so quarterly. Therefore, Times Mirror received the benefit of
16
monthly compounding of its funds, while its investors also got the benefit of quarterly
17
compounding. It can be clearly seen, however, that Times Mirror never got the benefit of
18
quarterly compounding. If VZ-MD were allowed a quarterly compounded rate, its
19
investors would effectively get the benefit of quarterly compounding twice, first when
20
VZ-MD gets it, and second when investors reinvested their quarterly dividends received
21
from Verizon.
22
23
Q.
DR. VANDER WEIDE SAYS THAT YOUR FAILURE TO MAKE AN EQUITY
FLOTATION COST ALLOWANCE IS AKIN TO IGNORING ALL THE
- 36 -
Surrebuttal Testimony of John I. Hirshleifer
1
EXPENSES OF THE COMPANY (VANDER WEIDE REBUTTAL, P. 60). IS
2
THIS TRUE?
3
A.
Obviously not. For example, Dr. Vander Weide does not make a salary cost adjustment
4
to the cost of capital, nor does he adjust it for advertising costs, lobbying costs, (or even
5
for expert witness costs). Similar to flotation costs, these adjustments do not need to be
6
made to the cost of capital because the market anticipates such costs and incorporates
7
them in the cash flow expectations for the company. Adding a flotation cost adjustment
8
would in effect double count the cost of financing.
9
Q.
IN ADDITION TO THE ARGUMENTS THAT YOU PRESENTED IN YOUR
10
TESTIMONIES EXPLAINING WHY THE FLOTATION COST ADJUSTMENT
11
IS NOT APPROPRIATE FOR THE TELECOMMUNICATIONS COMPANIES,
12
ARE THERE ANY OTHER REASONS WHY SUCH AN ADJUSTMENT WOULD
13
NOT BE APPROPRIATE FOR VERIZON?
14
A.
Yes. Over the past five years Verizon has issued only minor amounts of common stock
15
and has in fact bought back stock. Given the high level of equity in its market capital
16
structure, there is no reason to expect large equity financings in the foreseeable future.
17
18
E.
19
Q.
Dr. Vander Weide’s Criticisms of My CAPM Analysis of Equity
Costs Are Also Unfounded
DR. VANDER WEIDE IMPLIES THAT YOU SHOULD HAVE ABANDONED
20
THE USE OF BARRA BETAS. (VANDER WEIDE REBUTTAL, P. 64) WHAT
21
DO YOU MAKE OF THIS CRITICISM?
22
23
A.
I find it incomprehensible. I indicated that I used the predicted BARRA betas because I
was not able to calculate a 5-year historical beta for the then newly-formed Verizon. In
- 37 -
Surrebuttal Testimony of John I. Hirshleifer
1
my prior testimonies over several years, I had used BARRA betas as a reasonableness
2
check on my historical betas. Dr. Vander Weide implies that BARRA could not supply a
3
predicted beta for Verizon because of data limitations. However, BARRA did in fact
4
provide a predicted beta for Verizon as of June 2000 as part of its beta service and did not
5
indicate any lack of confidence in it. It is further puzzling that Dr. Vander Weide appears
6
to suggest that all predicted BARRA betas are faulty—even if you accepted his incorrect
7
assertion.
8
Q.
9
DR. VANDER WEIDE CLAIMS THAT BARRA PREDICTED BETAS ARE
CALCULATED USING EXPLANATORY VARIABLES THAT ARE ALL
10
CALCULATED FROM HISTORICAL DATA. (VANDER WEIDE REBUTTAL,
11
P. 64) IS HE CORRECT?
12
A.
BARRA is the analysts’ mean growth forecast.
13
14
No. Dr. Vander Weide is simply misinformed. For example, one of the variables used by
Q.
BY CRITICISING BARRA BETAS DR. VANDER WEIDE IMPLIES THAT YOU
15
SHOULD HAVE INSTEAD USED VALUE LINE BETAS. (VANDER WEIDE
16
REBUTTAL P. 64.) ARE VALUE LINE BETAS CALCULATED USING
17
PURELY HISTORICAL FIVE-YEAR RETURNS?
18
A.
Verizon is precisely the beta I should not be using because of “data factors.”
19
20
21
Yes. Therefore, according to Dr. Vander Weide’s own logic, Value Line’s beta for
Q.
HAS VALUE LINE COMPUTED BETAS FOR VERIZON SINCE THE MERGER
OF BELL ATLANTIC AND GTE?
- 38 -
Surrebuttal Testimony of John I. Hirshleifer
1
A.
It does not appear so. For example, the April 6, 2001 Value Line report on Verizon
2
indicates that the beta is “NMF”, meaning that Value Line could not measure it according
3
to its techniques. This report was issued over nine months after the close of the merger.
4
WHAT IS YOUR RESPONSE TO DR. VANDER WEIDE’S CRITICISM
Q.
5
REGARDING ADJUSTMENTS FOR BETAS LESS THAN 1? (VANDER WEIDE
6
REBUTTAL, P. 72)
7
A.
First, Dr. Vander Weide fails to point out that there is no general agreement that betas
8
should be adjusted, and if so, how they should be adjusted. The rationale for adjusting
9
raw betas is to reduce measurement error. As I discuss extensively in my testimony, I
10
attempt to adjust for measurement error through the process of averaging, a technique
11
commonly employed. BARRA utilizes its own models for adjusting betas. According to
12
BARRA studies, BARRA predicted betas have more than 16 times the predictive power
13
of historical betas.38 Ibbotson Associates, as another example, uses 5-year regressions of
14
monthly returns against the S&P 500 and weighs the individual company’s beta with the
15
average beta for the corresponding industry.39 Compustat makes no adjustments to its
16
betas.
17
Q.
DR. VANDER WEIDE IMPLIES THAT ONE SHOULD USE 5-YEAR
18
HISTORICAL VALUE LINE BETAS. IS THIS CONSISTENT WITH HIS PRIOR
19
TESTIMONY?
38
Barr Rosenberg, “Prediction of Common Stock Betas”, Reprinted with permission from The Journal of
Portfolio Management, Winter, 1985, on www.Barra.com/ResearchPub/NonBarraPub/pocs/pocs-j.html.
39
Ibbotson Associates, SBBI: Valuation Edition 2000 Yearbook, pp. 96-97.
- 39 -
Surrebuttal Testimony of John I. Hirshleifer
1
A.
No. In rebuttal testimony filed in many other proceedings, Dr. Vander Weide has
2
vigorously objected to the use of historical betas computed over a 5-year time period
3
because in his opinion they were not sufficiently forward-looking proxies for risk. It is
4
therefore extraordinary that he now suggests that one can use 5-year Value Line betas.
5
In his 1994 testimony before the FCC, for example, Dr. Vander Weide
specifically criticized MCI witness Kahal’s use of Value Line betas for:
6
fail[ing] to recognize that some of Value Line’s risk indicators he
relies on … encompass a five-year time period that is too long to
reveal recent increases in the risk of investing in
telecommunications.40
7
8
9
10
11
To “more accurately measure the changed risk of investing in
12
telecommunications,” Dr. Vander Weide computed two-year weekly betas.
13
In New Jersey, Dr. Vander Weide testified:
14
15
Q. Did you also perform a capital asset pricing model (CAPM)
analysis of the cost of equity?
16
17
18
19
20
21
22
23
24
A. No. One of the major inputs to the CAPM is beta—a measure of
the relative risk of a security to that of the market as a whole. Betas
are estimated using historical security prices, usually over the past 60
month period. The use of a methodology which relies on historical
data over this lengthy period of time would be particularly
inappropriate in this case. The enormous changes that the
telecommunications industry has recently undergone would render
such historical measures of relative risk virtually useless in
estimating the forward-looking cost of equity.41
40
Affidavit of Dr. James H. Vander Weide In Support of Reply Comments of Bell Atlantic, Before the
Federal Communications Commission, CC Docket 94-1, June 29, 1994, p.19-20, ¶32.
41
Direct Testimony of Dr. James H. Vander Weide on Behalf of Bell Atlantic-New Jersey, Docket No.
TX95120631, November 4, 1996, p. 21, at line 10-20.
- 40 -
Surrebuttal Testimony of John I. Hirshleifer
Recall that Dr. Vander Weide’s primary argument in this proceeding for not using
1
2
telephone holding companies for his comparable sample is his belief that “the
3
telecommunications holding companies … are experiencing radical restructuring and
4
profound regulatory, organizational and technological change.” (Vander Weide Rebuttal,
5
p. 49)
6
In his 1996 rebuttal testimony in the same New Jersey proceeding, Dr. Vander
7
Weide suggested that one-year betas would be appropriate.42 In his 1997 rebuttal
8
testimony before the State Corporation Commission of Virginia, Dr. Vander Weide
9
calculated two-year weekly betas.43
10
Q.
HAS DR. VANDER WEIDE TESTIFIED REGARDING FORWARD-LOOKING
11
12
BETAS?
A.
Yes. Dr. Vander Weide stated in his direct testimony filed on behalf of Bell Atlantic-
13
New Jersey on November 4, 1996, that “if one is to use such a method [CAPM], one
14
should use a forward-looking beta which measures the future risk of the company.”44
15
Q.
HAVE YOU CONSIDERED FORWARD-LOOKING BETAS?
16
A.
Yes. As already noted, I used predicted betas provided by BARRA. These predicted
17
18
betas include changing fundamental and market data which are incorporated in the beta.
Q.
REGARDING THE EQUITY RISK PREMIUM TO BE USED IN THE CAPM,
DR. VANDER WEIDE HAS TESTIFIED THAT IBBOTSON ASSOCIATES’ 2001
19
42
Rebuttal Testimony of Dr. James H. Vander Weide on Behalf of Bell Atlantic-New Jersey, Docket No.
TX95120631, December 20, 1996, p.33, at lines 7-12.
43
Rebuttal Testimony of Dr. James H. Vander Weide on Behalf of Bell Atlantic-Virginia, Case No.
PUC970005, June 10, 1997, p. 95.
44
Direct Testimony of James H. Vander Weide on Behalf of Bell Atlantic-New Jersey, Docket No.
TX95120631, November 4, 1996, p. 21.
- 41 -
Surrebuttal Testimony of John I. Hirshleifer
1
YEARBOOK CONTINUES TO SPECIFICALLY RECOMMEND THAT A
2
HISTORICAL RISK PREMIUM BASED ON THE 1926-PRESENT PERIOD
3
SHOULD BE USED. (VANDER WEIDE REBUTTAL, P. 71) WHAT DOES IT IN
4
FACT SAY IN THE 2001 YEARBOOK?
5
It says specifically that “[a] proper estimate of the equity risk premium requires a data
A.
6
series long enough to give a reliable average without being unduly influenced by very
7
good and very poor short-term returns.”45 (emphasis added). It also says that the “period
8
starting with 1926 is representative of what can happen: it includes high and low returns,
9
volatile and quiet markets, war and peace, inflation and deflation, and prosperity and
10
depression.”46 (emphasis added) Ibbotson Associates also continues that “because
11
historical event-types (not specific events) tend to repeat themselves, long-run capital
12
market return studies can reveal a great deal about the future.”47
13
It is also worth noting that—while Ibbotson Associates disagrees with the
14
approach—it explicitly acknowledges that some analysts calculate expected risk premia
15
over shorter time periods.48
16
Q.
DOES ROGER IBBOTSON HIMSELF STATE THAT THE EQUITY RISK
17
PREMIUM ESTIMATE SHOULD ALSO CONSIDER FORWARD-LOOKING
18
APPROACHES?
19
A.
Yes. Professor Roger Ibbotson states that:
45
Ibbotson Associates, Yearbook 2000, Valuation Edition, p. 65.
46
Id., p. 66.
47
Ibid.
48
Ibid.
- 42 -
Surrebuttal Testimony of John I. Hirshleifer
1
2
3
4
The historical payoff for risk is a good guide to the future risk
premium, but it is not perfect. First, there is considerable estimation
error even assuming the 74 years returns were drawn from a
stationary distribution. …
5
6
7
8
Another way to estimate the ERP [equity risk premium] is to
recognize that the stock market is a part of the economy. … The
supply side estimate of the stock market [risk premium] is
substantially lower than the historical ERP. …
9
10
11
12
Overall, I think the best estimate of the ERP is to use some
combination of the historical ERP and the supply side estimate of the
ERP. 49 [emphasis added]
Q.
YOU MENTIONED EARLIER THAT DR. VANDER WEIDE SUGGESTS THAT
13
YOU SHOULD ONLY CONSIDER THE FULL 1926-PRESENT IBBOTSON
14
DATA PERIOD FOR EQUITY RETURNS WHEN TRYING TO EVALUATE A
15
RISK PREMIUM. (VANDER WEIDE REBUTTAL, P. 71) HAS DR. VANDER
16
WEIDE FOLLOWED HIS OWN RULE CONSISTENTLY?
17
A.
No. In his direct testimony on behalf of GTE South filed in Virginia on June 9, 1995, Dr.
18
Vander Weide chose the period starting in 1937 on the theory that it would be “most
19
meaningful” to use S&P 500 data after the passage and implementation of the Public
20
Utility Holding Company Act of 1935.
21
Q.
DO FINANCIAL PROFESSIONALS ALWAYS RELY ON HISTORICAL RISK
22
23
PREMIA?
A.
No. I have provided numerous citations of leading scholars and practitioners on this
24
subject in my direct testimony. Additionally, John Bogle, Chairman and Founder of The
25
Vanguard Group which runs mutual funds and has assets of $530 billion, stated at the
49
Research Roundtable: The Equity Premium, June 30, 2000. (http://ssrn.com/forum/).
- 43 -
Surrebuttal Testimony of John I. Hirshleifer
1
Financial Analysts Seminar Sponsored by the Association for Investment Management
2
and Research that:
3
4
5
6
Looking out over time, from the price levels in today’s market, a 2%
risk premium might be a reasonable guess for the coming decade.
Indeed, many respected investment advisers might place the probable
number at less than 2%.
7
8
9
10
11
12
13
14
15
Well, I’m often wrong (seldom in doubt), so first let’s explore what a
normal equity premium might be. I went to the acknowledged
authority on the subject, best-selling author (‘Stocks for the Long
Run’) and Wharton School Professor Jeremy J. Siegel. He
obligingly sent me a two-century history of equity premiums on U.S.
stocks over long-term U.S. Treasury bonds. The average equity
premium over this long, long period is 3.5%. I will leave it to you to
decide what is a fair number to use today, but, for the rest of my
analysis, I’m going to rely on this average.50
16
Another distinguished academic, Alfred Rappaport, states that:
17
18
19
20
21
22
23
24
25
26
The premium should be based on expected rates of return rather
than average historical rates. This approach is crucial because with
the increased volatility of interest rates over the past two decades the
relative risk of bonds has increased, thereby lowering risk premiums
to a range from 3 to 5 percent. Those who estimate the market risk
premium as the long-run average excess of stock returns over
government bond returns will typically obtain a figure in the 7 to 9
percent range. This historical approach ignores that market risk
premiums vary over time and at the present time can lead to
significant undervaluation.51 [emphasis added]
50
John C. Bogle, “The Riddle of Performance Attribution: Who’s In Charge Here--Asset Allocation or Cost?”
Remarks Before the Financial Analysts Seminar Sponsored by the Association for Investment Management
and Research, At Northwestern University, Evanston, Illinois, July 20, 1997. (Published at
www.vanguard.com).
51
Rappaport, Alfred, Creating Shareholder Value, The Free Press, New York, 1998, p. 39.
- 44 -
Surrebuttal Testimony of John I. Hirshleifer
1
Michael Mauboussin, Chief U.S. Investment Strategist at Credit Suisse First
2
Boston and Adjunct Professor at Columbia Business School, believes that the equity risk
3
premium used in the CAPM model should be estimated ex ante:
4
5
6
7
8
Ex-post definitions come with a lot of calculational baggage, most
notably choice of time period and data non-stationarity. …[U]se a
long-term discounted cash flow model to estimate expected return,
and then subtract a long-term Treasury yield to estimate the ex-ante
ERP.52
9
He believes that the risk premium has been in a range of 2-5% in recent years and
10
states that Credit Suisse First Boston uses about 4.0%.
11
Eugene Fama, Professor of Finance at the University of Chicago, estimates the
12
expected equity premium to be about 1-2%.53 John Cochrane, Professor of Finance at the
13
University of Chicago, believes that the risk premium is about or below 3-4%.54
14
Jay Ritter, Professor at the University of Florida, states:
15
16
17
18
19
20
21
22
In the 1980s, I followed the textbook mantra that the equity risk
premium should be based on extrapolating the historical average into
the future. By the late 1980s, I began to realize how wrong this was,
as the Japanese market soared. This approach predicted that in the
1990s there would be extremely high returns on Japanese stocks, just
as today it implies that there will be unrealistically high returns on
US stocks in the future.55
Q.
HAS DR. VANDER WEIDE STATED IN THE PAST HIS BELIEF THAT THE
23
MARKET RISK PREMIUM VARIES OVER TIME?
52
Research Roundtable: The Equity Premium, June 30, 2000. (http://ssrn.com/forum/).
53
Ibid.
54
Ibid.
55
Ibid.
- 45 -
Surrebuttal Testimony of John I. Hirshleifer
1
A.
Yes. In his testimony before the State Corporation Commission of Virginia, Dr. Vander
2
Weide stated that the equity risk premia over bonds "vary with the level of interest
3
rates."56
4
ISN’T THE IBBOTSON ASSOCIATES APPROACH TO ESTIMATING THE
Q.
5
EQUITY RISK PREMIUM FOUNDED ON THE THEORY THAT THE TRUE
6
RISK PREMIUM IS STABLE OVER TIME?
7
A.
Yes. Ibbotson Associates states that:
8
9
10
11
12
13
14
15
[T]he expected equity risk premium is unobservable in the market
and therefore must be estimated. … In using a historical measure of
the equity risk premium, one assumes that what has happened in the
past is representative of what might be expected in the future. In
other words, the assumption one makes when using historical data to
measure the expected equity risk premium is that the relationship
between the returns of the risky asset (equities) and the riskless asset
(Treasuries) is stable.57
16
Consequently, if Dr. Vander Weide believes that the risk premium varies with
17
18
interest rates, he cannot consistently advocate the Ibbotson approach.
Q.
DR. VANDER WEIDE ARGUES THAT PROFESSOR CORNELL STATED IN
19
HIS BOOK THAT THE IBBOTSON APPROACH TO ESTIMATING THE RISK
20
PREMIUM IS APPROPRIATE. (VANDER WEIDE REBUTTAL, PP. 68-69) IS
21
DR. VANDER WEIDE FAMILIAR WITH THE CURRENT THINKING ON THIS
22
SUBJECT?
56
Direct Testimony of Dr. James H. Vander Weide, Before the State Corporation Commission of Virginia,
On Behalf of Central Telephone Company of Virginia, The Chesapeake and Potomac Telephone Company
of Virginia, Contel of Virginia, Inc., GTE South Incorporated, United Telephone - Southeast, In., Case No.
PUC920029, p. 48, at 1-5.
57
Ibbotson Associates, SBBI: Valuation Edition 2000 Yearbook, p. 53.
- 46 -
Surrebuttal Testimony of John I. Hirshleifer
1
Apparently not. Professor Cornell’s book cited by Dr. Vander Weide was published in
A.
2
1993 and written some time before that date. Since 1993 a vast amount of literature has
3
been published regarding the equity risk premium: Ibbotson and Brinson58 and
4
Blanchard59 published their research findings in 1993; Siegel60 in 1994; Brown,
5
Goetzmann and Ross61 in 1995; Rappoport62 in 1998; Glassman and Hassett63 in 1999;
6
etc. Numerous articles have also been published noting the low equity risk premium. In
7
1999 Professor Cornell published an entire book devoted to subject of the equity risk
8
premium.64 Professor Cornell concluded that the equity risk premium at the time of the
9
writing of his book was in the range of 3.5% - 5.5%. My review of all of these sources
10
indicates that a 5.5% premium over long-term Treasury bonds appears to be conservative,
11
and may substantially overstate the actual current forward-looking expected risk
12
premium.
13
Q.
DR. VANDER WEIDE CLAIMS THAT HE HAS CALCULATED THE COST OF
14
EQUITY FOR THE S&P 500 USING THE SAME METHODOLOGY THAT YOU
15
USED FOR PRIOR TESTIMONIES BUT DID NOT DO FOR THIS
58
Ibbotson, Roger, and Gary P. Brinson, Global Investing: The Professional’s Guide to the World Capital
Markets, McGraw-Hill, 1993, at p. 45.
59
Blanchard, Oliver, “Movements in the Equity Premium”, Brookings Papers on Economic Activity, 75 (2)
1993.
60
Siegel, Jeremy, Stocks for the Long Run, Irwin, New York, 1994.
61
Brown, Stephen J., William N. Goetzmann and Stephen A. Ross, “Survival”, The Journal of Finance, Vol.
L, No. 3, July 1995.
62
Rappaport, Alfred, Id.
63
Glassman, James K., and Kevin A. Hassett, DOW 36,000: The New Strategy for Profiting from the Coming
Rise in the Stock Market, Times Books, 1999.
64
Cornell, Bradford, The Equity Risk Premium: The Long-Run Future of the Stock Market, John Wiley &
Sons, 1999.
- 47 -
Surrebuttal Testimony of John I. Hirshleifer
1
PROCEEDING, AND ARRIVED AT A COST OF EQUITY OF 10.93%, HIGHER
2
THAN MERRILL LYNCH’S COST OF EQUITY ESTIMATE OF 10.20%.
3
(VANDER WEIDE REBUTTAL TESTIMONY P. 66) ASSUMING THAT HE DID
4
THIS CORRECTLY, DOES THIS CAUSE ANY CONCERN TO YOU?
5
A.
Not at all. As explained in my direct testimony, this forward-looking cost of equity
6
estimate was utilized as one of several analysis tools for estimating the equity risk
7
premium. Assuming that Dr. Vander Weide’s calculations are correct, this 10.93%
8
estimate could also be used. Substituting 10.93% for 10.20% in Exhibit 6 to my direct
9
testimony yields forward-looking estimates of 6.00% over the long-run expected one-
10
month Treasury bill yield and 4.67% over the 20-year Treasury bond yield. However, in
11
my CAPM calculations I used risk premia estimates conservatively higher than these
12
estimates: 7.5% over long-run expected one-month Treasury bill yields and 5.5% over 20-
13
year treasury bond yields.
14
Q.
SHOULD THE FACT THAT MERRILL LYNCH ESTIMATED AN EXPECTED
15
RETURN (I.E., THE FORWARD-LOOKING COST OF EQUITY) FOR THE
16
MARKET OF 10.20% CAUSE DR. VANDER WEIDE TO QUESTION HIS HIGH
17
COST OF CAPITAL ESTIMATE OF 12.95%?
18
A.
Yes. Merrill Lynch is a sophisticated investment firm and also has been a financial
19
adviser to Bell Atlantic through at least two mergers with other giant telephone holding
20
companies. This is an obvious sanity check, similar to the costs of capital and discount
21
rates used by analysts, and by other investment banks in fairness opinions.
- 48 -
Surrebuttal Testimony of John I. Hirshleifer
1
Q.
DR. VANDER WEIDE CLAIMS THAT YOU HAVE MISSTATED THE
2
HISTORICAL EQUITY RISK PREMIUM CALCULATED OVER THE PERIOD
3
1926-1999 (VANDER WEIDE REBUTTAL PP. 68-69). HAVE YOU?
4
A.
No. He is again mistaken. Dr. Vander Weide assumes that I have simply taken the
5
arithmetic risk premium from the Ibbotson Associates Yearbook. Ibbotson Associates
6
calculates its arithmetic mean risk premium by taking the difference between the average
7
large company stock total returns (13.3%) and long-term government bond income
8
returns (5.2%).65 Contrary to Dr. Vander Weide’s assumption, I have calculated a range
9
of risk premia using geometric and arithmetic averages. My calculation66 of the
10
arithmetic average differs from that used by Ibbotson Associates because I take the
11
difference between the averages of large company stock total returns (13.3%) and long-
12
term government bond total returns (5.5%).67
13
Q.
DR. VANDER WEIDE IS CRITICAL THAT YOU CONSIDER GEOMETRIC
14
MEAN AVERAGES IN ADDITION TO ARITHMETIC AVERAGES WHEN
15
EVALUATING THE APPROPRIATE RISK PREMIA. HE SAYS THAT
16
IBBOTSON ASSOCIATES ARGUES THAT YOU SHOULD ONLY LOOK AT
17
THE ARITHMETIC MEAN WHEN ESTIMATING A HISTORICAL RISK
18
PREMIUM. (VANDER WEIDE REBUTTAL, P. 70) IS YOUR ANALYSIS
19
DEPENDENT ON WHAT IBBOTSON ASSOCIATES SAYS?
65
Ibbotson Associates, Stock Bonds Bills and Inflation 2000 Yearbook, pp. 124 and 185.
66
In my calculations, I utilized return data from both Ibbotson Associates and Dimensional Fund Advisers.
Ibbotson Associates and DFA returns differ only due to minor rounding.
67
Ibbotson Associates, Stock Bonds Bills and Inflation 2000 Yearbook, p. 124.
- 49 -
Surrebuttal Testimony of John I. Hirshleifer
1
A.
No. My analysis considers the arguments and data of Ibbotson Associates and also of
2
numerous other scholars and practitioners. Dr. Damodaran, for example, recommends
3
and utilizes geometric averages.
4
Q.
DOES DR. VANDER WEIDE RELY ON WHAT IBBOTSON ASSOCIATES SAY?
5
A.
Dr. Vander Weide’s reliance on the Ibbotson Associates approach is quite selective. As
6
noted above, in prior testimony he did not accept its foundational theory that the equity
7
risk premium is stable over time. He also ignores several other key propositions
8
embraced by Ibbotson Associates in the Valuation Edition 2001 Yearbook. These
9
propositions, if accepted, pull the linchpin from his entire analysis:
10
1.
The cost of capital is always an “expectational” or forward-looking concept (p. 9);
11
2.
The risk of the loss of business to competitors is unsystematic (i.e., investors can
12
diversify it away) so it is not entitled to a risk premium (p. 41);
13
3.
14
Multi-stage DCF models give better estimates of the cost of equity than does the
perpetual growth model which Dr. Vander Weide utilizes (p. 50);
15
4.
The terminal stage growth-rate in the DCF model should be sustainable. An
16
example of an indefinitely sustainable growth rate is the expected long-run growth
17
rate of the economy. (p. 50).
18
Q.
DR. VANDER WEIDE ARGUES THAT THE CONCEPT OF SURVIVORSHIP
19
BIAS IN MEASURING HISTORIC WORLD EQUITY RETURNS FOR
20
ESTIMATING AN EQUITY RISK PREMIUM DOES NOT APPLY TO STOCKS
21
TRADING IN THE U.S. MARKET. (VANDER WEIDE REBUTTAL, P. 71) IS
22
THIS A LOGICAL INTERPRETATION OF THE THEORY?
- 50 -
Surrebuttal Testimony of John I. Hirshleifer
1
No. His view, and in this instance, Ibbotson Associates’ view,68 is an extreme one. The
A.
2
theory postulates that historical U.S. stock returns overstate the returns you would have
3
obtained if you had been an international investor and had also invested in stock markets
4
that performed poorly relative to the U.S. stock market. In other words, using the
5
historical returns of a single, successful national stock market to estimate future returns
6
does not accurately reflect potential losses if a stock market were to perform poorly. As
7
of 1925 for example, you would not have known before the fact that the U.S. market was
8
going to be successful. Even if you had invested solely in U.S. stocks, there was a chance
9
that the U.S. market would have been one of the failures, and that you would have lost
10
much if not all of your money.
11
Of course, investors planning to hold an international portfolio of stocks will
12
estimate returns on the expectations for an international stock portfolio, not just on the
13
returns derived from stocks of companies in a single country. Dr. Vander Weide seems to
14
be saying with his argument that all investors in Verizon own, or will purchase only U.S.
15
stocks. This assumption is clearly not true. Verizon is one of the component companies
16
of the S&P 500, an index whose stocks are widely held by giant pension, mutual fund and
17
other managed portfolios, many of which are located and/or have investors outside of the
18
U.S., or themselves have diversified into various international holdings.
19
One need only look at how Verizon currently describes itself to understand its
20
global position:
68
Ibbotson Associates does state that the survivorship bias evidence is “compelling on a worldwide basis.”
The Valuation Edition 2001 Yearbook , Ibbotson Associates, p. 73.
- 51 -
Surrebuttal Testimony of John I. Hirshleifer
1
2
3
4
5
6
7
8
9
10
11
Verizon Communications is one of the world's leading providers of
communications services. Verizon companies are the largest
providers of wireline and wireless communications in the United
States, with 112 million access line equivalents and 27 million
wireless customers. Verizon International has investment interests in
telecommunications companies in 19 countries, with a global
presence that extends to 40 countries in the Americas, Europe, Asia
and the Pacific. Verizon has 3.2 million proportionate access lines
and 8.3 million proportionate wireless subscribers. It is a Fortune 10
company with approximately 260,000 employees and more than $65
billion in annual revenues.
12
13
14
15
16
17
18
19
20
Verizon is superbly positioned to capitalize on worldwide growth
trends that are transforming global telecommunications. Verizon
Global Solutions Inc. is building a global network to provide
seamless end-to-end communications by delivering data, voice, and
internet solutions to customers around the world. Verizon's global
network will link North America with major cities in Europe, Asia
and Latin America, and provide intra-regional communications.
Verizon's scale and scope make it the number one partner for anyone
wanting to access the U.S. market.69
Dr. Vander Weide’s view also poses a classic finance arbitrage. He is
21
22
fundamentally saying that an investor in only U.S. stocks would have one cost of capital
23
for Verizon, while an international stock investor would have a lower cost of capital—for
24
the same company. Therefore, one investor would apply the higher U.S. market-based
25
risk premium and value the multi-national company at a lower price, while another
26
investor would apply the lower world risk premium and value it at a higher price.
27
Because the international investor can pay more for Verizon, even in the U.S. stock
28
markets, it would bid up the price and arbitrage away price discrepancies caused by the
29
local investor’s parochial cost of capital.
69
http://www.verizon.com/international/.
- 52 -
Surrebuttal Testimony of John I. Hirshleifer
1
2
3
II.
DR. VANDER WEIDE HAS ASSUMED A CAPITAL STRUCTURE
THAT IS INAPPROPRIATE FOR THE WHOLESALE BUSINESS OF
SUPPLYING UNBUNDLED NETWORK ELEMENTS.
4
Q.
DR. VANDER WEIDE OFFERS AN ELABORATE ARGUMENT AGAINST THE
5
THEORETICAL SOUNDNESS OF USING A BUSINESS’S BOOK CAPITAL
6
STRUCTURE. (VANDER WEIDE REBUTTAL, PP. 34-36) HE CLAIMS THAT
7
YOU BASE YOUR COST OF CAPITAL ESTIMATE ON THE BOOK CAPITAL
8
STRUCTURE FOR VERIZON-VA. (VANDER WEIDE REBUTTAL, P. 34) IS
9
DR. VANDER WEIDE CORRECT?
10
A.
No. In fact, I clearly state in my testimony that the proper weights to be used for cost of
11
capital calculations should be the long-run target financing weights that a rational,
12
informed management team would employ for the network element leasing business. The
13
market value capital structure of the network element leasing business is not observable
14
because, as Dr. Vander Weide noted in his direct testimony, there are no publicly traded
15
companies which solely conduct the network element leasing business. (Vander Weide
16
Direct, p. 50) I also note that because the network element leasing business is less risky
17
than the aggregate business of the telephone holding companies, the market value debt
18
weights of the holding companies probably understate long-run target debt weights in the
19
capital structure of the network element leasing business.
20
Consequently, I calculate a cost of capital range using the telephone holding
21
companies’ market value capital structure to determine the high side of the range (which
22
provides the cost of capital for the telephone holding companies) and the book value
23
capital structure to determine the low side of the range, with the midpoint of the range
- 53 -
Surrebuttal Testimony of John I. Hirshleifer
1
considered to be the best estimate of the cost of capital for the business of network
2
element provision. This midpoint cost of capital estimate implies that a reasonable proxy
3
for the target market capital structure for the network element leasing business is also the
4
midpoint between the average market and book capital structures of the telephone holding
5
companies.
6
Q.
DOES DR. VANDER WEIDE HIMSELF RECOGNIZE THAT THE TARGET
7
MARKET CAPITAL STRUCTURE OF THE NETWORK ELEMENT
8
WHOLESALING BUSINESS IS NOT OBSERVABLE?
9
A.
Yes. At page 50 of his direct testimony he wrote that “… at the present time, there are no
10
publicly traded companies that have built telecommunications networks solely for the
11
purpose of providing local exchange service in a competitive market.” If there are no
12
publicly-traded local exchange companies, there are no publicly-traded network element
13
wholesaling businesses. Clearly, one cannot directly observe the capital structure of an
14
ILEC, let alone of a network element leasing business.
15
Q.
WHAT DO YOU MEAN BY THE TARGET MARKET CAPITAL STRUCTURE?
16
A.
As Copeland, Koller and Murrin note:
17
18
19
20
21
22
23
24
25
26
27
The theoretically correct approach to capital structure is to use a
different WACC for each year that reflects the capital structure for
the year. In practice, we usually use one WACC for the entire
forecast. We also think in terms of a target capital structure rather
than the current capital structure because at any point a company’s
capital structure may not reflect the capital structure expected to
prevail over the life of the business. Capital structure might be
affected by recent changes in the market value of the securities
outstanding and the “lumpiness” of financing activities, particularly
those involving securities offerings. Moreover, management may
have plans to change the capital mix as an active policy decision.
- 54 -
Surrebuttal Testimony of John I. Hirshleifer
1
2
3
All these factors mean that future financing levels could be different
from current or past levels.70
Q.
DOES DR. VANDER WEIDE PROVIDE ANY PROOF THAT HIS CAPITAL
4
5
STRUCTURE ESTIMATES ARE CORRECT?
No. The logical flaw in Dr. Vander Weide’s argument is obvious. If the true target
A.
6
capital structure of the network element leasing business is not observable, as he and I
7
both agree, it is not observable. No definitive proof can be provided by any party.
8
Therefore, analysts can only estimate the true capital structure based on sound judgment.
9
Using market value capital structures of holding companies with substantial high-growth
10
businesses that appear to be far riskier than the ILEC’s dominant network element leasing
11
businesses, as Dr. Vander Weide does, does not appear to be a sound approach.
12
Q.
DR. VANDER WEIDE EXTENSIVELY CITES YOUR COLLEAGUE DR.
13
CORNELL ON THIS SUBJECT. (VANDER WEIDE REBUTTAL, PP. 37-38)
14
DOES DR. VANDER WEIDE INTERPRET DR. CORNELL CORRECTLY?
15
While I am delighted at Dr. Vander Weide’s high regard for Dr. Cornell, he does not
A.
16
understand that Dr. Cornell entirely agrees with my view that the target market value of
17
the network element leasing business should be used, which can only be estimated, and
18
that market value capital structures of riskier holding companies should not be used.
19
Ironically, Dr. Vander Weide cites a specific passage from Dr. Cornell’s book which
20
states that “[i]f the comparable firms are publicly traded, their market value weights can
70
Copeland, Tom, Tim Koller, and Jack Murrin, Valuation: Measuring and Managing the Value of
Companies, 3rd Edition, McKinsey & Co., 2000, p.203-204.
- 55 -
Surrebuttal Testimony of John I. Hirshleifer
1
be calculated directly and averaged” (emphasis added), apparently forgetting his own
2
testimony that there are no publicly traded firms for the network element leasing business.
3
In fact, Dr. Cornell has offered cost of capital testimony in numerous state
4
TELRIC proceedings using substantially the same methodologies that I have used in this
5
and previous cases. Based on my recollection, those state commissions have generally
6
adopted capital structures recommended by Dr. Cornell or close thereto.
7
For example, the Ohio Commission found that:
8
9
10
11
12
13
14
15
16
Rather than adopting the cost of capital recommended by Ameritech,
we find that, on balance, the midpoint cost of capital
recommendation advanced by the AT&T/MCI witness Dr. Cornell
most accurately reflects the appropriate forward-looking cost of
capital for use in Ameritech’s TELRIC studies. In adopting the
AT&T/MCI recommendation, we note that Dr. Cornell provided the
most extensive support and analysis for his cost of capital
recommendation. Based on the record presented to us, we are most
comfortable with the analysis Dr. Cornell has undertaken.71
17
In that proceeding, the Ohio Commission adopted Prof. Cornell’s recommended
18
19
cost of capital of 9.74%.
Q.
DR. VANDER WEIDE QUOTES YOUR TESTIMONY IN AN OHIO UNE COST
20
PROCEEDING REGARDING THE FACT THAT THE STOCK MARKET
21
VALUES THE ASSETS OF AN ILEC AT MARKET VALUE. (VANDER WEIDE
22
REBUTTAL, P. 35) HE ALSO COMPLAINS THAT YOU LEVER AND
23
UNLEVER BETAS USING MARKET VALUE CAPITAL STRUCTURES OF
71
Opinion and Order, In the Matter of the Review of Ameritech Ohio’s Economic Costs for Interconnection,
Unbundled Network Elements, et al., The Public Utilities Commission of Ohio, Case no. 96-922-TP-UNC
(June 19, 1997).
- 56 -
Surrebuttal Testimony of John I. Hirshleifer
1
TELEPHONE HOLDING COMPANIES. DOES THIS CONTRADICT
2
ANYTHING YOU ARE SAYING IN THIS PROCEEDING?
3
A.
Not in the slightest. If Dr. Vander Weide had read my testimony more carefully, he
4
would have seen that my analysis starts with the estimation of the cost of equity for the
5
telephone holding companies. Consequently, I use market value capital structures and
6
unlever and relever betas using market value capital structures. Because the telephone
7
holding companies are riskier than the ILECs’ network element businesses, this cost of
8
capital estimate provides a ceiling to my range. The analytical step that Dr. Vander
9
Weide misses is that the cost of capital for the less risky network element business will be
10
less than the high side of the range, which incorporates the risks of all of the telephone
11
holding companies’ riskier businesses. The book value is used to estimate the low side of
12
the range.
13
Q.
IN THAT OHIO PROCEEDING, WHAT CAPITAL STRUCTURE AND COST
14
15
OF CAPITAL WAS ADOPTED BY THE OHIO COMMISSION?
The Ohio Commission adopted Staff’s recommendation to use the book capital structure
A.
16
for Cincinnati Bell Telephone, which contained 42.24% debt and 57.76% equity, and
17
specifically noted that this capital structure approximated the mid-point of the range that I
18
proposed.72 In that proceeding, the Ohio Commission adopted a cost of capital of 9.56%.
72
“We find that, under the facts and circumstances presented in this case, the staff’s book capital structure
should be adopted for purposes of determining the cost of capital. Staff witness Chaney recommends that a
capital structure of 42.24 percent long-term debt and 57.76 percent common equity be used for purposes of
this case. … The staff’s recommended capital structure approximates the mid-point of Mr. Hirshleifer’s
proposed range.” The Public Utilities Commission of Ohio, Supplemental Opinion and Order, In the
Matter of the Application of Cincinnati Bell Telephone Company for Approval of a Retail Pricing Plan
Which May Result in Future Rate Increases and For a New Alternative Regulation Plan, Case No. 96-899TP-ALT, p.13.
- 57 -
Surrebuttal Testimony of John I. Hirshleifer
1
IS DR. VANDER WEIDE’S TESTIMONY REGARDING CAPITAL STRUCTURE
Q.
2
3
CONSISTENT WITH HIS PRIOR TESTIMONY?
A.
Completely at odds with his current arguments, Dr. Vander Weide argued in his affidavit
4
in support of Bell Atlantic’s comments filed with the FCC on June 29, 1994 that it was
5
incorrect to use the capital structure of the regional holding companies (RHCs) in place of
6
the capital structure of the price cap LECs because some of the RHC’s have financial,
7
cellular and cable TV subsidiaries whose “capital structure does not reflect the actual
8
financing of the price cap LECs’ investments in telecommunications infrastructure.”73
9
Dr. Vander Weide also concluded that “[t]he capital structures of these subsidiaries
10
should be removed from the RHC’s consolidated capital structure to better reflect the
11
financing of the LEC’s telecommunications infrastructure.”74
Moreover, in his direct testimony regarding Virginia’s Experimental Plan (Case
12
13
No. PUC920029), Dr. Vander Weide stated that:
14
15
16
17
18
19
20
21
22
23
24
25
26
According to financial theory, the appropriate capital structure for
an enterprise is determined by its own business risk, the liquidity and
the market value of its own assets, and its own competitive strategy.
The proper capital structure for the LECs participating in the Plan
is related to their own business situation, not their parent
company’s. The parent companies of the LEC’s each have capital
structures that reflect their particular business situations. There is
evidence of parent company diversification into financial services,
real estate, cellular, interLATA services, cable television, and
overseas ventures. The LECs participating in the Plan have no
investment in their parent’s diversification efforts, and the risks of
these ventures are unrelated to the LECs business risks as local
telephone companies. [emphasis added]
73
Affidavit of Dr. James H. Vander Weide In Support of Reply Comments of Bell Atlantic, Before the
Federal Communications Commission, CC Docket 94-1 (June 29, 1994).
74
Ibid.
- 58 -
Surrebuttal Testimony of John I. Hirshleifer
1
Q.
DOES DR. VANDER WEIDE HIMSELF RELY ON BOOK VALUES FOR
2
3
OTHER ASPECTS OF HIS WORK IN THE CURRENT PROCEEDING?
Yes. For example, he estimates the cost of debt for VZ-MD by using yields on Moody’s
A.
4
A-rated bonds as a proxy. However, Moody’s uses book value capital structure ratios as
5
one of its analytical tools for assessing the riskiness of the subject companies. If Dr.
6
Vander Weide believes that book values can never be used for risk assessment, he should
7
not be relying on Moody’s.75 Dr. Vander Weide additionally uses book value weights for
8
estimating the market value of debt in his capital structure calculations. As I noted earlier
9
in this testimony, Dr. Vander Weide also relies on returns on book equity when he
imputes growth rates using the “b x r” method.
10
11
Q.
IN HIS SCHEDULE 2, DR. VANDER WEIDE PURPORTS TO ESTIMATE THE
12
IMPLIED MARKET CAPITAL STRUCTURE FOR A STAND-ALONE LOCAL
13
EXCHANGE COMPANY USING MARKET MULTIPLES. IS HIS ANALYSIS
14
CORRECT?
15
A.
No. Dr. Vander Weide is using multiples derived from the stock prices of telephone
16
holding companies. These stock prices reflect the valuation and risks of all the riskier
17
businesses operated by the telephone holding companies. Dr. Vander Weide is confused
18
between publicly-traded telephone holding companies and non-public local exchange
19
operations that are owned by telephone holding companies.
75
Standard & Poor’s also utilizes book value leverage ratios as one of it risk analysis tools.
- 59 -
Surrebuttal Testimony of John I. Hirshleifer
1
2
3
III.
DR. VANDER WEIDE’S COST OF CAPITAL ESTIMATES ARE
GROSSLY AT ODDS WITH ANALYSES BY INVESTMENT FIRMS
AND OTHER INDEPENDENT ANALYSTS.
4
Q.
WHAT HAS DR. VANDER WEIDE’S RESPONSE BEEN WHEN CONFRONTED
5
WITH THE EVIDENCE OF INVESTMENT BANKS AND FINANCIAL
6
ANALYSTS THAT THE COSTS OF CAPITAL FOR TELEPHONE HOLDING
7
COMPANIES ARE FAR LOWER THAN WHAT HE ESTIMATES?
8
9
A.
In his rebuttal testimony Dr. Vander Weide stated that Wall Street analysts “were hired to
provide an opinion regarding the fairness of the stock exchange ratio to be used in
10
proposed mergers, not to estimate the cost of capital input in UNE cost studies.” (Vander
11
Weide rebuttal, p. 82) In his prior testimonies in other states, Dr. Vander Weide has also
12
argued that “these analysts can not be used to justify any estimate of the
13
telecommunications companies’ weighted average costs of capital” since they “were
14
forced to assume a specific discount rate because they had not performed an independent
15
analysis of the weighted average cost of capital.” (see e.g., Vander Weide New York
16
responsive testimony p. 62).
17
As a former due diligence officer of a broker-dealer, I found Dr. Vander Weide’s
18
testimony nothing less than astonishing. It is incredible to even suggest that competent
19
fairness opinions in multi-billion dollar mergers could be rendered when one of the key
20
parameters of the valuation analysis is simply “assumed.” Merrill Lynch, Salomon Smith
21
Barney and Goldman Sachs are some of the world’s largest investment firms, with
22
expertise as financial advisors to giant multinational companies in mergers and
23
acquisitions.
- 60 -
Surrebuttal Testimony of John I. Hirshleifer
1
For example, to suggest that the weighted average cost of capital used in a merger
2
stock exchange ratio analysis was an “arbitrary assumption” (Vander Weide New York
3
responsive testimony, p. 65) is equivalent to saying that the whole analysis for which the
4
advisors were handsomely paid was just a charade.76 The investment advisors to the
5
transaction relied on those estimates as part of their valuation of the two companies,
6
which in turn was intended to ensure that shareholders received proper compensation as a
7
result of the proposed merger of the two companies. For investment advisors to rely on
8
“arbitrary assumptions” would have constituted a breach of fiduciary responsibility by the
9
investment advisors to management and by management to shareholders.
10
On appeal from the UNE decision of the Delaware PSC in 1997, the Federal
11
District Court in that state considered and rejected similar arguments by Bell Atlantic for
12
disregarding the discount rates used by Merrill Lynch in its exchange ratio analysis for the
13
Bell Atlantic-NYNEX merger. The discount rate for local service, 8% to 10%, supported
14
the 10.28% weighted cost of capital upheld by the Delaware PSC.77
15
DR. VANDER WEIDE’S SUGGESTION THAT THE DISCOUNT RATES USED
Q.
16
BY INVESTMENT BANKS WERE “SIMPLY ASSUMED” OR “ARBITRARY”
17
ALSO IMPLIES THAT THEY ARE ARBITRARILY WRONG. (VANDER
18
WEIDE REBUTTAL, P. 83) DO YOU BELIEVE THAT INVESTMENT BANKS
19
USE DISCOUNT RATES THAT THEY BELIEVE TO BE WRONG?
20
A.
Of course not. Investment banks use discount rates that they believe to be correct.
76
77
GTE and Bell Atlantic each paid $15 million to their financial advisors, Goldman Sachs, Salomon Smith
Barney, Chase Manhattan, Bear Stearns, Morgan Stanley Dean Witter, and Merrill Lynch, in connection
with this merger.
Bell Atlantic-Delaware, Inc. v. McMahon, 80 F.Supp.2d 218, 241 (D. Del. 2000).
- 61 -
Surrebuttal Testimony of John I. Hirshleifer
1
Q.
USE THESE DISCOUNT RATES. WHO COULD BE “FORCING” THEM?
2
3
DR. VANDER WEIDE SAID THAT INVESTMENT BANKS ARE “FORCED” TO
A.
Unless Dr. Vander Weide was suggesting that the telephone holding companies
4
themselves were somehow coercing their financial advisors into using discount rates that
5
were too low, no one was forcing them to use incorrect assumptions, or assumptions that
6
they did not believe in, in their valuation analyses.
7
Q.
8
9
WHAT REFINEMENT DID DR. VANDER WEIDE ADD TO THIS ARGUMENT
FOR THIS PROCEEDING?
A.
10
Dr. Vander Weide now argues that investment banks had to assume an incorrect cost of
capital because:
11
12
13
14
15
16
17
18
19
20
they simply could not simultaneously estimate the cost of capital at
the same time that they estimated the appropriate share price.
Indeed, these firms were faced with a situation where they had just
one equation to determine two unknowns—the value of the firm’s
stock, and the cost of equity. As a simple matter of mathematics,
there are many combinations of share values and cost of capital that
will solve a single equation; and hence, no unique solution exists for
either unknown. To resolve this dilemma, the investment bankers
chose not to estimate the cost of capital. Instead, they simply
assumed a discount rate. (Vander Weide rebuttal p. 83).
21
Q.
IS THERE ANY MERIT TO THESE ARGUMENTS?
22
A.
No. First, with respect to the DCF model, Dr. Vander Weide forgets one of the hotly-
23
debated topics in this very proceeding: the year-by-year earnings growth rate assumptions,
24
which are several other unknown parameters to the model. By necessity, all analysts
25
must estimate both the cost of capital and the growth rates if they are to use the DCF
26
model. This does not lead to Dr. Vander Weide’s faulty conclusion that investment
- 62 -
Surrebuttal Testimony of John I. Hirshleifer
1
bankers are using costs of capital and, by implication, growth rate assumptions that they
2
believe to be incorrect for estimating values in fairness opinions for the mergers of giant
3
corporations. Second, Dr. Weide inaccurately assumes that all investment bankers are
4
using the DCF model to estimate the cost of capital. Investment bankers are just as likely
5
to be using the CAPM, and potentially other models for cost of capital estimation. The
6
CAPM does not use stock prices or growth rates as inputs to the model.
7
Q.
WITH RESPECT TO THE FAIRNESS OPINION WORK PERFORMED BY
8
INVESTMENT BANKS, DID ANYTHING PREVENT THEM FROM USING DR.
9
VANDER WEIDE’S COST OF CAPITAL METHODOLOGY IF THEY
10
11
BELIEVED IN IT?
A.
No. Dr. Vander Weide’s simple application of the single-stage DCF model could easily
12
be calculated by investment banks within a very short period of time. Financial analysts
13
could have done the same for their analyst reports if they thought his approach was
14
applicable.
15
Q.
IN THIS PROCEEDING DR. VANDER WEIDE ARGUES THAT THE
16
DISCOUNT RATES USED BY INVESTMENT BANKS PROVIDE NO SUPPORT
17
FOR YOUR COST OF CAPITAL ESTIMATES BECAUSE THESE DISCOUNT
18
RATES WERE “SIMPLY ASSUMED” (VANDER WEIDE REBUTTAL, P. 83)
19
WHAT DIFFERENT ARGUMENT DID DR. VANDER WEIDE PRESENT WITH
20
RESPECT TO THIS SUBJECT PREVIOUSLY?
21
A
In the New York UNE cost proceeding Dr. Vander Weide alternatively argued that the
22
discount rates used by investment banks should not be considered because the banks' cost
23
of capital methodologies were not entered into evidence. [Vander Weide New York
- 63 -
Surrebuttal Testimony of John I. Hirshleifer
1
responsive testimony, p. 67] For this proceeding, however, he purports to know how the
2
investment bankers are calculating their costs of capital.
Dr. Vander Weide’s argument made no sense because he himself uses analyst data
3
4
which do not disclose underlying methodologies. For example, the I/B/E/S average
5
analyst earnings growth forecasts for hundreds of companies are critical inputs to his
6
single-stage DCF model. For each company in his sample, these forecasts are averages of
7
several forecasts made by different financial analysts. No explanation is provided as to
8
how each individual analyst made its forecast of future earnings growth.
9
I have also seen Dr. Vander Weide's testimony before several state commissions
10
stating that the 8% - 10% DCF discount rates disclosed in the Bell Atlantic/NYNEX
11
prospectus/proxy statement for valuing the telephone operations of the company were not
12
probative because, among other things, they were intended merely to provide relative
13
values of the companies for purposes of evaluating the fairness of the exchange ratio. He
14
testified that "when estimating the relative values, it doesn't really matter what the cost of
15
capital is, as long as the same cost of capital is used for both companies."78 This
16
argument was simply wrong. The choice of discount rates does affect the valuation of the
17
different segments of a subject company, which in turn would affect the exchange ratio.
18
Q.
WHAT NEW ARGUMENT DOES DR. VANDER WEIDE OFFER IN HIS
19
CONTINUING ATTEMPT TO CHALLENGE THE COST OF CAPITAL
20
ESTIMATES OF INVESTMENT BANKERS AND FINANCIAL ANALYSTS?
78
Transcript of Dr. Vander Weide's testimony on November 7, 1996 before the New York State Public
Service Commission, Case 95-C-0657, p. 3768, at 14-17.
- 64 -
Surrebuttal Testimony of John I. Hirshleifer
1
A.
In this proceeding, he argues that investment bankers and analysts use one cost of capital,
2
while investors inexplicably use a much higher one. He attempts to show this by
3
comparing the stock price valuations estimated by investment bankers to the actual stock
4
prices of the subject companies. (Vander Weide rebuttal p. 76-81) According to this
5
convoluted theory, everyone (i.e., investment bankers, analysts and investors) is using the
6
same earnings growth assumptions in their DCF valuation models. Thus, if investment
7
bankers and analysts estimate higher values than the actual market price for a stock using
8
their own cost of capital estimates, this “proves” that investors’ cost of capital is lower.
9
This reasoning is exceptionally misinformed. As I noted above, he appears to be unaware
10
or has simply forgotten that each investment bank is using its own unique forecasts of
11
earnings growth, which is one of the main professional activities engaged in by such
12
banks and analysts.
13
The average earnings growth expectation of the market for a company is an
14
unknown, yet Dr. Vander Weide blindly assumes that it is somehow known to all and
15
uniformly utilized. Therefore, his faulty logic leads him to the conclusion that only the
16
cost of capital estimate differs, whereas growth expectations are in fact differing. Of
17
course, no one knows what the true market growth expectations are (although it is fair to
18
assume that no rational observer expects above-market growth forever). Professional
19
analysts are using their skills to estimate several important, but unknown, parameters to
20
the DCF model: expectations regarding the future growth path of company earnings, and
21
the company’s cost or capital.
22
23
Q.
IF YOU HYPOTHETICALLY INDULGE DR. VANDER WEIDE’S ARGUMENT
THAT WALL STREET ANALYSTS ARE NOT ATTEMPTING TO ESTIMATE
- 65 -
Surrebuttal Testimony of John I. Hirshleifer
1
INVESTORS’ COST OF CAPITAL, WHOSE COST OF CAPITAL ARE THEY
2
ESTIMATING?
3
A.
This is where Dr. Vander Weide leaves us completely in the dark. He offers no reasoning
4
why investment bankers and analysts, sophisticated investment professionals, are not
5
trying to estimate the costs of capital of investors. Consider that their clients are
6
generally investors, or companies seeking investors to either buy, or maintain the value
7
of, their shares.
8
Q.
IN HIS NEW YORK REBUTTAL TESTIMONY, DID DR. VANDER WEIDE
ARGUE THAT ANY OF THE INVESTMENT BANKS’ COST OF CAPITAL
9
10
11
ESTIMATES SUPPORTED HIS POSITION?
A.
Ironically he did, although on a very selective basis. Dr. Vander Weide cited in his New
12
York responsive testimony the analyses performed by investment banks in connection
13
with proposed mergers of ALLTEL/Aliant and MCI Worldcom/ Sprint. (Vander Weide
14
New York responsive testimony p. 66) Dr. Vander Weide claimed that in these analyses
15
the assumed discount rates ranged from 10% - 12% for local operations. However, Dr.
16
Vander Weide did not cite these analyses correctly.
17
Warburg Dillon Read in fact used discount rates from 10.0% to 11.0%, not 12%,
18
to value Sprint FON group’s local telephone division (p. 59).79 It assumed for the long
19
distance telephone division discount rates ranging from 10.5% to 11.5%. In citing that
20
fairness opinion, however, Dr. Vander Weide failed to mention that in the same filing
21
Salomon Smith Barney, acting as MCI WorldCom’s financial advisor, used a weighted
79
WorldCom Inc.’s Amendment No. 3 To Form S-4 filed with the SEC on February 17, 2000.
- 66 -
Surrebuttal Testimony of John I. Hirshleifer
1
average cost of capital range of 8.75% to 9.75% to value Sprint’s local telephone division
2
(p. 74) and 9.25% to 10.25% for the long distance segment. It performed a DCF analysis
3
of Sprint PCS Group and MCI WorldCom, using discount rates reflecting a weighted
4
average cost of capital range from 10.5% to 11.5% for Sprint PCS Group and 11.5% to
5
12.5% for MCI WorldCom. Salomon Smith Barney used a higher cost of capital range for
6
Sprint’s PCS Group than for Sprint FON’s local segment. It also used a higher cost of
7
capital for the long distance segment than for the local telephone division.
8
When valuing the ALLTEL/Aliant merger, Merrill Lynch used discount rates
9
ranging from 10% to 12%, not for the local telephone business suggested by Dr. Vander
10
Weide, but for all the businesses of the companies.80 ALLTEL and Aliant have
11
substantial riskier businesses relative to their local exchange operations.
12
ISN’T DR. VANDER WEIDE NOW ARGUING THAT THESE SAME
Q.
13
WARBURG DILLON READ AND MERRILL LYNCH COST OF CAPITAL
14
ESTIMATES DO NOT REPRESENT INVESTORS’ COSTS OF CAPITAL?
15
A.
Yes. He is not at all consistent. And as noted above, he does not accept, or even consider
as a sanity check, Merrill Lynch’s 10.20% forward-looking cost of equity for the market.
16
17
Q.
WHAT ELSE DOES DR. VANDER WEIDE ARGUE IS A REASON THAT
18
19
INVESTMENT BANK ANALYSES SHOULD NOT BE USED?
Dr. Vander Weide also argues that investors “are not entitled to rely” on any single part of
A.
20
an analysis in a fairness opinion because of exculpatory language that the investment
21
bankers include in their opinion. (Vander Weide rebuttal, p. 83). This is a transparent and
80
Alltel Corp.’s Form S-4 which was filed to SEC on March 24, 1999.
- 67 -
Surrebuttal Testimony of John I. Hirshleifer
1
nonsensical argument. Investment banks routinely add this kind of exculpatory language
2
to mitigate their exposure to potential lawsuits from their issuance of fairness opinions.
3
By doing so, this does not lead to the conclusion that parties cannot consider the inputs
4
that the investment banks use in their models, nor to the conclusion that investment
5
bankers use assumptions that they believe to be wrong.
6
Q.
DO ANALYST REPORTS CONTAIN THIS EXCULPATORY LANGUAGE?
7
A.
Generally not, to my knowledge. In addition to the analyst reports that I have previously
8
cited, many other analyst reports also indicate that analysts consider the cost of capital for
9
various telecommunications businesses to be much lower than what Dr. Vander Weide
10
suggests. For example, Morgan Stanley Dean Witter used a 10% estimate of the cost of
11
capital in its DCF valuation of ALLTEL in a report dated March 13, 2000.81 In April
12
2000, BHF-Bank used an 8.2% cost of equity and 7.7% WACC for valuing Deutsche
13
Telekom.82 In January 2000 ABN Amro used a 10% discount rate to value Bell
14
Atlantic’s local, long-distance, directory assistance and wireless services.83 In its
15
September 29, 2000 DCF valuation of SBC Communications, A.G. Edwards employed a
16
8.70% cost of capital for the holding company in aggregate.84 Prudential Financial has
17
recently issued a report on regional Bell operating companies and integrated long-distance
18
companies and used a 9.5% discount rate for the RBOCs and a 10.5% discount rate for
19
the integrated long-distance companies.85
81
Morgan Stanley Dean Witter, “Alltel Corporation”, March 10, 2000, p. 4, and March 13, 2000, p. 3.
82
BHF-Bank, “Deutsche Telekom”, April 27, 2000, p. 6.
83
ABN Amro, “Bell Atlantic Corporation,” January 20, 2000, pp. 12, 19.
84
A.G. Edwards, “SBC Communications,” September 29, 2000, pp. 9-10.
85
Prudential Financial, "Wireline Telecommunications Services", May 29, 2001, p. 142.
- 68 -
Surrebuttal Testimony of John I. Hirshleifer
1
Q.
DR. VANDER WEIDE CITES IN HIS TESTIMONY THE COST OF CAPITAL
2
ESTIMATES FOR THE AGGREGATE TELECOMMUNICATIONS INDUSTRY
3
PROVIDED BY IBBOTSON ASSOCIATES’ COST OF CAPITAL QUARTERLY.
4
(VANDER WEIDE REBUTTAL, P. 83-84) HE STATES THAT THESE
5
ESTIMATES ARE HIGHER THAN YOUR ESTIMATES FOR THE UNE
6
LEASING BUSINESS. WHAT IS YOUR VIEW?
7
A.
It is interesting that Dr. Vander Weide considers this summary of cost of capital
8
calculations for telecommunications holding companies helpful to him, because when
9
looked at with a more analytical eye, I think that it is far more supportive of my cost of
10
capital estimate.
11
First, Dr. Vander Weide fails to note that this sample of 20 companies includes
12
companies engaged in all of the business activities riskier than the telephone network
13
leasing business. Therefore, all of his cost of capital calculations, even if one were to
14
assume that they were correct, overstate the cost of capital of the network element leasing
15
business. In my analysis, I include in my sample companies which have some
16
involvement in the local exchange business and to not include those which are much
17
more focused on riskier activities. Dr. Vander Weide focuses on the cost of capital
18
estimates for the “Large Composite” group, which in addition to RBHC’s and GTE,86
19
includes several companies much more heavily engaged in businesses riskier than the
20
local exchange business. For example, the ten large companies probably also include
86
The 2000 Cost of Capital Yearbook contains data as of March 2000. At that time GTE and Bell Atlantic
had not yet merged to form Verizon.
- 69 -
Surrebuttal Testimony of John I. Hirshleifer
1
AT&T, MCI Worldcom and Sprint FON Group.87 Three of these ten companies have
2
debt rated at BB or below according to the 2000 Cost of Capital Yearbook.
3
For this reason, and because the giant RBHC’s like Verizon are likely to have the
4
lowest cost of capital relative to riskier companies in this telecommunications group, Dr.
5
Vander Weide should be focusing his attention on the 25th percentile results for the costs
6
of capital.
7
Second, two of the cost of capital calculations use the CAPM model, which
8
employs Ibbotson Associates’ high estimate of the equity risk premium based solely on its
9
historical methodology. However, as I discuss extensively in my testimony, I consider
10
research performed by numerous scholars and practitioners that conclude that the
11
forward-looking equity risk premium is far lower than that determined using Ibbotson
12
Associates’ method. Consequently, Dr. Vander Weide is really saying once again that he
13
only believes Ibbotson Associates’ approach to the risk premium estimate, and that he
14
inexplicably ignores all other views, including the stated position of Roger Ibbotson
15
himself. As I explained above, however, Dr. Vander Weide has previously indicated that
16
the equity risk premium is not stable over time, so he does not believe the fundamental
17
premise of the Ibbotson Associates historical risk premium approach.
18
In contrast, I consider a far larger set of approaches regarding the equity risk
19
premium and utilize an estimate that falls somewhere between the high and low
20
approaches, even though the preponderance of the current research and opinions indicate
87
The 2000 Cost of Capital Yearbook does not identify the specific companies that make up its large
composite, but does indicate that AT&T is the largest company by sales and total capital.
- 70 -
Surrebuttal Testimony of John I. Hirshleifer
1
that the equity risk premium is currently very low. If a lower risk premium were used in
2
Ibbotson Associates’ CAPM model, a lower cost of equity would result.
3
Notably, Ibbotson Associates estimate of beta at the 25th percentile is lower than
4
the 0.77 beta that I use, and not at all close to the beta suggested by Dr. Vander Weide.
5
Therefore, if one were to use an equity risk premium in the range of 5.5% in their model
6
along with a beta of 0.65, it would result in cost of equity estimate of 9.84%. If even
7
lower equity risk premium estimates were used, lower CAPM cost of equity estimates
8
would result.
9
Another cost of equity calculation reflected in the Ibbotson Associates publication
10
uses the Fama/French model. Neither Dr. Vander Weide nor I have used this model, and
11
to my knowledge this model has not developed any serious following.
12
Ibbotson Associates calculates DCF-based costs of equity using two models, a
13
single-stage model and a three-stage model. To no one’s surprise, the single-stage model
14
that Dr. Vander Weide uses yields a higher cost of capital. In contrast, the three-stage
15
model yields a lower cost of equity estimate. Ibbotson’s three-stage model uses a higher
16
estimate of the long-run growth rate for the economy than does my model, which
17
averages the growth rate estimates of WEFA and Ibbotson Associates.
18
While he has dropped the argument for this proceeding, in his New York
19
responsive testimony Dr. Vander Weide then implied that Ibbotson Associates’ three-
20
stage model cost of equity estimate should be ignored. In making this suggestion, he
21
completely ignored Ibbotson Associates’ unequivocal statement in its 2000 Yearbook that
22
“[t]o produce a better estimate of the equity cost of capital, one can use a multi-stage
- 71 -
Surrebuttal Testimony of John I. Hirshleifer
1
discounted cash flow model.”88 In this instance, the 3-stage discounted cash flow method
2
cost of equity for industry SIC Code 4813 (Telephone Communications, Except
3
Radiotelephone) is 9.57% at the 25th percentile.
4
Q.
DR. VANDER WEIDE CLAIMS THAT A COST OF MONEY INPUT USED BY
5
AT&T IN 1997 IN ITS TOTAL INCREMENTAL COST MODEL ("TICM") FOR
6
LONG DISTANCE SERVICES SHOWS THAT YOUR FORWARD-LOOKING
7
COST OF CAPITAL ESTIMATE FOR VERIZON'S NETWORK ELEMENT
8
LEASING BUSINESS IS "UNJUSTIFIABLY LOW." (VANDER WEIDE
9
REBUTTAL, PP. 74-75) PLEASE RESPOND.
10
I find it ironic that Dr. Vander Weide purports to endorse AT&T’s undefined model input
A.
11
regarding the cost of money for certain long distance services in 1997, while in so many
12
other respects, in all TELRIC proceedings that I am aware of, he vehemently disagrees
13
with AT&T’s assumptions. More fundamentally, Dr. Vander Weide fails to explain, as
14
an initial matter, how this view regarding a long distance model relates to the forward-
15
looking cost of capital of an ILEC’s network element leasing business as of today.
16
To analyze this question, I return to some of the reality checks that I have cited in
17
my testimony. For example, Warburg Dillon Read, in its February 2000 assessment of
18
the WorldCom/Sprint merger, assumed discount rates ranging from 10.5% to 11.5% for
19
long distance telephone operations. In his New York responsive testimony, Dr. Vander
20
Weide specifically cited (although inaccurately) as evidence of the cost of capital for the
21
local telephone business the fairness opinions rendered for this merger, which opinions
88
Ibbotson Associates, SBBI: Valuation Edition 2000 Yearbook, p. 47.
- 72 -
Surrebuttal Testimony of John I. Hirshleifer
1
included Warburg Dillon Read’s range of 10% to 11% regarding the cost of capital for
2
Sprint FON’s local telephone operations (New York responsive testimony, p. 66). He
3
fails to explain why he then had faith in Warburg Dillon Read’s estimate for local
4
telephone operations, but is now somehow persuaded that the cost of capital for the long
5
distance business is dramatically higher than what Warburg suggests. In this same filing,
6
the fairness opinion of Salomon Smith Barney used a lower range of 9.25% to 10.25% to
7
analyze the long distance segment of Sprint FON Group and a range of 8.75% to 9.75%
8
for local operations. In May 2001 Prudential Financial has used a 10.5% discount rate for
9
the integrated long-distance companies and a 9.5% discount rate for the RBOCs.89
Consequently, Dr. Vander Weide’s reliance on this input to an AT&T model
10
11
appears to be inconsistent with other evidence. As the long distance business is generally
12
recognized as being riskier than local telephone operations, the investment banks’
13
estimates of cost of capital ranges for long distance operations appear to be far more
14
consistent on a relative basis with their own lower estimates of the cost of capital for local
15
exchange operations, and with my estimates and state commission decisions regarding the
16
cost of capital for the provision of network elements.
17
Q.
WHAT OTHER COMPANY-SPECIFIC INFORMATION FROM A SIMILAR
18
TIME PERIOD COULD DR. VANDER WEIDE HAVE CONSIDERED AS A
19
“SANITY CHECK”?
20
21
89
A.
In a 1997 Ameritech UNE cost proceeding, the Ohio Public Utility Commission
considered an internal Ameritech cost of capital estimate for the telephone holding
Prudential Financial, "Wireline Telecommunications Services", May 29, 2001, p. 142.
- 73 -
Surrebuttal Testimony of John I. Hirshleifer
1
company. This estimate was dramatically lower than what Ameritech was advocating as
2
its UNE cost of capital:
3
4
5
6
7
8
9
10
11
12
13
14
15
16
Ameritech has failed to convince this Commission that the riskadjusted cost of capital used either in its TELRIC studies (13.6
percent) or in its modified TELRIC studies (11.5 percent) reflects an
appropriate cost of capital. A number of concerns justify this
conclusion. At the outset, we note that both the 13.6 percent and
11.5 percent cost of capital figures recommended by the company for
use in TELRIC studies are higher than the weighted average cost of
capital of 10.5 percent that Mr. Domagola calculated for AIT (the
holding company) in June 1996 (MCI Ex. 2). This infers that the
risk associated with the provision of UNEs is greater than the risk
associated with the diversified operations of the holding company
parent. If anything, UNEs are the least risky element since they
represent bottleneck facilities which will be needed by competitors
for a considerable time to come.90
17
18
19
IV.
DR. VANDER WEIDE’S CONTRIVED “TESTS OF
REASONABLENESS” UNDERSCORE THE REASONABLENESS OF
MY COST OF CAPITAL ESTIMATES, NOT HIS.
20
Q.
IN HIS REBUTTAL TESTIMONY DR. VANDER WEIDE ALLEGES THAT
21
YOUR 3-STAGE DCF MODEL DID NOT CONFORM TO HIS SO-CALLED
22
“TESTS OF REASONABLENESS.” AS AN INITIAL MATTER, IF YOU
23
HYPOTHETICALLY ASSUMED THAT HIS CONCLUSIONS WERE TRUE, DO
24
THEY HELP OVERCOME THE PROBLEM NOTED BY ECONOMISTS OF
25
USING THE SINGLE-STAGE DCF MODEL WHEN FORECAST GROWTH
26
RATES EXCEED THAT OF THE ECONOMY?
90
Opinion and Order, In the Matter of the Review of Ameritech Ohio’s Economic Costs for Interconnection,
Unbundled Network Elements, et al., The Public Utilities Commission of Ohio, Case no. 96-922-TP-UNC,
June 19, 1997.
- 74 -
Surrebuttal Testimony of John I. Hirshleifer
1
A.
2
No they do not. The use of the single-stage model with the assumption of perpetual
supernormal growth results in cost of equity estimates that are too high.
3
Q.
WHAT FLAWS EXIST IN HIS TESTS?
4
A.
The flaws in his analysis are theoretical, logical and practical. As there are so many, I
5
only attempt to address some of them in this rebuttal to provide the Commission with a
6
sense of their ubiquity. First and foremost, Dr. Vander Weide is simply attacking a straw
7
man: he is saying that if he ignores the analytical procedures that I recommend for
8
estimating the cost of capital, he can perform what he describes as “tests” that gives him
9
results that he does not like. This premise is analogous to taking apart a car, throwing
10
away half of the parts, attempting to reassemble the parts without the benefit of the
11
blueprints, and then pronouncing that there are problems with the car.
12
Q.
WHEN DR. VANDER WEIDE ATTEMPTS TO COMPARE VALUE LINE
13
BETAS TO DCF RESULTS, HE ASSUMES THAT THERE IS A DIRECT
14
THEORETICAL LINKAGE BETWEEN THE DCF AND THE CAPM
15
APPROACHES. DOES THIS LINKAGE IN FACT EXIST?
16
A.
No. The two approaches are theoretically distinct, and as can be clearly seen in my direct
17
testimony, they do not have a single common parameter in their models. Dr. Vander
18
Weide is implicitly saying that both models are indisputably correct, and that if you knew
19
the “true” inputs (such as betas that had no measurement error), you would get the same
20
answers from each. There is no evidence in the academic literature that has yet proved
21
one or both theories to be unequivocally correct. Therefore, you cannot necessarily
22
assume that there is a direct relationship between the beta input to the CAPM, (even if
23
you could determine the “true” beta without any measurement error), and the cost of
- 75 -
Surrebuttal Testimony of John I. Hirshleifer
1
equity results from the unrelated DCF model. This theoretical independence is precisely
2
why I have used both models in my analysis and taken an average of the results of the two
3
to derive a balanced estimate of the cost of equity.
4
Q.
5
6
HAS DR. VANDER WEIDE USED THE CAPM TO ARRIVE AT HIS COST OF
CAPITAL ESTIMATE IN THIS PROCEEDING?
A.
No, and from my experience, he never has in prior TELRIC proceedings. As I discuss
7
above, Dr. Vander Weide has testified in the past that the CAPM is not suitable because
8
betas are not forward-looking in his opinion. He has specifically criticized Value Line
9
betas, which are computed over a 5-year historical period, as not being forward-looking.
10
So on the one hand, he posits that the CAPM should not be used at all, while on the other
11
he posits both that there is a direct relationship between the two distinct models, and that
12
Value Line betas are accurate measures of risk.
13
Q.
14
15
DOES DR. VANDER WEIDE’S SINGLE-STAGE DCF MODEL DEMONSTRATE
INCONSISTENCIES WITH HIS FORMULATION OF THE CAPM APPROACH?
A.
Yes. As a simple illustration, consider the estimate of Intel’s cost of equity using a
16
single-stage DCF model. As of June 2000, Intel was expected to grow at 19.55% for the
17
next 5 years and had a very low forward-looking dividend yield of 0.13%. Consequently,
18
the single-stage model estimates Intel’s cost of equity to be 19.68%. Now consider the
19
CAPM. Intel’s beta according to Value Line, as suggested by Dr. Vander Weide, was
20
1.1. Consequently, Intel’s cost of equity using the CAPM model is 6.26% + (1.1 * Risk
21
Premium). Whatever risk premium is selected, the cost of equity is substantially lower
22
than the 20% that the single-stage model calculates.
- 76 -
Surrebuttal Testimony of John I. Hirshleifer
1
Q.
ONE OF DR. VANDER WEIDE’S TESTS PURPORTS TO RANK AND
2
COMPARE THE RESULTS OF THE 3-STAGE DCF MODEL FOR VARIOUS
3
GROUPS OF COMPANIES. (VANDER WEIDE REBUTTAL, PP. 84-89). HOW
4
IS THIS FLAWED?
5
A.
As an example, Dr. Vander Weide inexplicably ignores my observation that companies
6
that pay low dividends may have cost of capital estimates which are biased downwards.
7
As he himself noted in his rebuttal testimony, when I estimated a cost of equity on the
8
market in prior testimonies, I specifically excluded companies with dividends under
9
1.5%. His own estimate of the cost of equity for the S&P 500 which, according to him,
10
used my methodology was 10.93%. In contrast, Dr. Vander Weide’s Table 1 now shows
11
9.00% as the cost of equity for the S&P 500, inconsistent with his own calculation. His
12
estimate of the S&P Industrials in Table 1 makes the same mistake, and he includes
13
Century Telephone in his small sample of local exchange carriers, which I specifically
14
excluded because the DCF approach may be less accurate for companies that pay small
15
dividends.
16
He is also not consistent regarding his position on sample size. He argues in his
17
rebuttal testimony that an average of four to five companies will not yield an accurate
18
estimate of a group’s cost of capital, presumably because of measurement error. Yet, for
19
purposes of this alleged ranking comparison, he is quite comfortable using an average of
20
only three “natural gas distribution companies”, even though there are many such
21
companies doing business in the United States.
22
23
Dr. Vander Weide also makes some rather broad assumptions about relative risk.
For example, Dr. Vander Weide’s “electric” group is composed of companies that are
- 77 -
Surrebuttal Testimony of John I. Hirshleifer
1
involved in electric, gas and nuclear energy, telecommunications, real estate, financial
2
services and international businesses. Over the past year there have been unanticipated
3
increases in natural gas prices that have had dramatic impact on certain electricity
4
markets, such as in California. As a result PG&E has entered bankruptcy and Edison
5
teeters on the brink. Therefore, it is not at all clear that these companies are as “low risk”
6
as Dr. Vander Weide tries to portray them.
7
Q.
DR. VANDER WEIDE TURNS TO STATISTICAL REGRESSIONS AS
8
ANOTHER ATTEMPT AT “TESTING REASONABLENESS.” DO THESE
9
TESTS SUFFER FROM THE SAME FLAWS?
10
A.
Yes. Dr. Vander Weide consistently ignores the analytical procedures that I use. He has
11
used cost of equity and beta inputs in his regressions that would not have resulted if I had
12
done the analysis. For example, my approach involves averaging betas and costs of
13
equity of comparable companies in order to reduce measurement error. In his regression
14
reflected at his Table 2, he does no such averaging and simply compares raw Value Line
15
betas against raw costs of equity. More significantly, he incorrectly assumes that there is
16
a direct linkage between the CAPM and DCF models. That incorrect assumption alone
17
renders the regression hypothesis meaningless. He also uses Value Line betas, which I do
18
not use—and which he himself has said are inappropriate. He does not explain why he
19
did not use other measures of beta, such as averaged betas, or BARRA predicted betas, or
20
Ibbotson Associates’ betas, or betas calculated over one or two-year time periods.
21
22
Q.
DR. VANDER WEIDE SIMILARLY ATTEMPTS TO CRITIQUE THE 3-STAGE
DCF MODEL BY RUNNING REGRESSIONS OF THE COST OF EQUITY
- 78 -
Surrebuttal Testimony of John I. Hirshleifer
1
RESULTS AGAINST THE I/B/E/S GROWTH RATES. DOES THIS SUFFER
2
FROM THE SAME PROBLEMS THAT YOU HAVE DISCUSSED?
3
A.
Yes. And significantly, Dr. Vander Weide does not test the correct theoretical
4
proposition. One question that could be tested is how cost of equity results from different
5
DCF models correlate with true growth rate expectations over the long-term. As I have
6
repeatedly cited from leading scholars and practitioners, analyst growth rates above a
7
sustainable long-run growth rate of the economy cannot persist forever. Moreover, I have
8
shown with actual telephone holding company examples that the analysts themselves do
9
not assume high-growth rates for particularly long periods of time.
10
Dr. Vander Weide’s regression is intentionally designed to test the relationship
11
between the 5-year I/B/E/S growth rates and the cost of equity results from the one-stage
12
DCF model, which assumes that the 5-year growth rates continue forever. Therefore, one
13
would expect a higher correlation for a model that uses the 5-year growth rate perpetually
14
as an input when compared with a model that uses the 5-year growth rates for only an
15
initial growth period, and then uses growth rates that change annually and linearly
16
converge to the growth rate of the economy in year 20.
17
Dr. Vander Weide has essentially created a self-fulfilling—but irrelevant—
18
proposition: it is no surprise that the 5-year growth rate correlates well with the results of
19
a one-stage model that uses the 5-year growth rate as the only growth input. This
20
obviously does not test a more important question: how do the results correlate with true
21
expected growth rates? As these rates can only be inferred with reasonable analytical
22
procedures, it would be difficult to properly specify a regression to test them. However, it
23
is already clear from simple logic, let alone the wealth of expert opinions, that
- 79 -
Surrebuttal Testimony of John I. Hirshleifer
1
supernormal growth rates cannot last indefinitely as Dr. Vander Weide unreasonably
2
suggests.
3
Q.
DR. VANDER WEIDE ALSO PURPORTS TO REGRESS DIVIDEND YIELDS
4
AGAINST COST OF EQUITY RESULTS. DOES THIS ANALYSIS SUFFER
5
FROM SIMILAR PROBLEMS?
6
A.
Yes. Dr. Vander Weide also makes an assumption that he fails to support: he states that
7
high dividend yield companies should have lower costs of equity. There has been a fair
8
amount of academic research regarding the relationship between dividend yields and
9
returns.
10
For example, Grinblatt and Titman state that:
11
12
13
14
15
Stocks with high dividend yields do, in fact, have higher returns, on
average, than stocks with low dividend yields…Stocks with zero
dividend yields have substantially higher expected returns than
stocks with low dividend yields, but for stocks that do pay dividends,
expected returns increase with dividend yields.91
16
Dr. Vander Weide himself has noted in his own prior testimony that increases in
the dividend yield “reflects a general increase in the cost of capital.”92
17
18
As higher dividend yields on average indicate higher expected returns, even if you
19
accepted Dr. Vander Weide’s approach, according to his own theory Dr. Vander Weide’s
20
regression appears to demonstrate the inferiority of the single-stage model.
91
Grinblatt, Mark and Sheridan Titman, Financial Markets and Corporate Strategy, McGraw-Hill, 1998, p.
532.
92
Affidavit of Dr. James H. Vander Weide In Support of Reply Comments of Bell Atlantic, Before the
Federal Communications Commission, CC Docket 94-1, June 29, 1994, p. 20, ¶34.
- 80 -
Surrebuttal Testimony of John I. Hirshleifer
1
Q.
2
3
DOES DR. VANDER WEIDE MISS THE MOST OBVIOUS TEST OF
REASONABLENESS?
A.
Yes. Dr. Vander Weide does not ask himself in any reasonable fashion why his cost of
4
equity results are so far away from those of analysts, which provide real-world views of
5
the cost of capital and are independent of this proceeding. Instead, Dr. Vander Weide
6
strains to offer several far-fetched arguments in an attempt to explain away this evidence.
7
Q.
DOES THAT CONCLUDE YOUR REBUTTAL TESTIMONY?
8
A.
Yes, it does.
- 81 -