-5...“ ask:
h a . . flunk.

3.5... t
:1 “a. mi».

.65..
«a. :71. .o.
0....

shun a L135...
A 9193.9...
: 19hr.

li'ri.

f.
. ' l I .O
xfl i: bt.
7.. no . 2.0.1..-
.

L531}. ,

It!!!

i

I ‘ . .
.1: ‘6:53:33 F

.1 a... .all,b.lau
:3..tnxihnai

.2! :

.2... 91-1..)

 

.. 1.1.1:“
i. .3125:
I I

‘ «a 5a).”..-
a

 

 

 

 

 

 

 

Ml IICH IANG ANTESTA

IIII I III

 

 

 

 

 

 

 

 

 

 

 

 

 

 

IIIIII I I I I IIIIII IIIIIIIII

410 1764

This is to certify that the

dissertation entitled

AN EMPIRICAL TEST AND COMPARATIVE ANALYSIS
OF REGULATED RATES OF RETURN

presented by
Richard Lee Patterson
has been accepted towards fulfillment

of the requirements for

Ph . D 0 degree in BUS 1118 SS Admin .

 

 

Major professor

Date 9/18/95

 

MSU is an Affirmative Action/Equal Opportunity Institution 0- 12771

 

LIBRARY
Michigan State
University

 

 

 

PLACE ll RETURN BOX to roman thin chock“ from your mood.
TO AVOID FINES return on or More data duo.

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

I .l -
——I____

 

 

 

 

 

 

 

 

__II:I__I
DI I-
II I-
I I

MSU I. An Affirm-live AdlaVEquni Opportunity Im
Willa-94

 

 

 

 

 

 

 

 

 

 

AN EMPIRICAL TEST AND COMPARATIVE ANALYSIS
OF REGULATED RATES OF RETURN

By

Richard Lee Patterson

A DISSERTATION

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

Department of Finance and Insurance

1995

ABSTRACT

AN EMPIRICAL TEST AND COMPARATIVE ANALYSIS
OF REGULATED RATES OF RETURN

By

Richard Lee Patterson

Historically, the financial literature has supported two primary methods used to
determine the allowed rate of return for wholly owned regulated subsidiaries, the
independent fu'm approach and the double leverage approach. Regulatory commission
reports show that rate regulation has included a third alternative, utilizing a consolidated tax
adjustment for holding companies owning regulated subsidiaries. This tax adjustment
reduces the tax benefits that arise from a holding company form of business, passing these
tax benefits back to the subsidiary utility's ratepayers as lower tax rate assessments. This
in turn lowers the affiliated subsidiary's allowed rate of return. The trend in subsidiary rate
of retum adjustment regulation has been toward the utilization of this third alternative
exclusively.

Similar to the implied criticism shown by the actions of public utility commissions,
Ezzell, Hsu and Miles (EHM) (1991) identify weaknesses in both the independent firm and
double leverage approaches. The independent firm approach recognizes differences in risk
between subsidiaries but ignores the debt held at the parent level, along with its interest tax
shield benefits. Conversely, the double leverage approach recognizes the debt held at the
parent level eliminating its interest tax shield benefits, but ignores any differences in risk
between subsidiaries. EHM (1991) through the use of a perpetual cash flow model develop
an alternative approach that recognizes both the debt held at the parent level and any
differences in risk between subsidiaries. The question arises, "Does EHM (1991) bridge
the apparent gap between financial theory and regulatory proceedings?"

EHM (1991) utilizes an unfamiliar discounting technique in determining the present
value of the parent level interest tax shield benefits. Due to the fact that the primary effect
of their rate of return adjustment approach depends on this value, the effectiveness and
understanding of their adjustment process is reduced. This dissertation places the EHM
adjustment procedure into the more familiar Modigliani and Miller (MM) perpetual cash

Richard Lee Patterson

flow environment. It does so using a completely different derivation from the one used by
EHM. I also simplify the EHM adjustment factor equation, placing it in more intuitive
terms without any loss of accuracy. Third, I show how the EHM methodology can be
modified to provide the exact same adjusted rate of return results to those found by my MM
perpetual cash flow model.

McGilsky (1986) performed an empirical study of different consolidated tax adjustment
methods being applied by public utility commissions. Her study defines these consolidated
tax adjustments in tax allowance equation form. Seven of these tax allowance equations
have been incorporated into my MM perpetual cash flow model. A comparative analysis is
performed for these seven consolidated tax adjustments. The results show that all of tax
adjustment procedures overcompensate for the parent level interest tax benefits and
therefore provide too low a return on the holding company subsidiary investment. The
same analysis is performed on all of the alternative procedures for determining the return
for a wholly owned regulated subsidiary defined by financial literature. These results show
that only the EHM approach consistently provides a zero net present value "fair" rate of
return for wholly owned regulated subsidiaries.

Finally, the empirical model of this dissertation fails to support the perpetual cash flow
model results the unfavorable consolidated tax adjustment affect The empirical evidence
implies that the consolidated tax adjustment affect is not a major factor in determining the
value of the holding companies included in this study. There is empirical evidence that
regulatory commissions have provided too low a return for subsidiaries within the local
telephone service industry.

In Ioving memory of my mother and father

iv

ACKNOWLEDGMENTS

I would like to thank the members of my dissertation committee. To Dr. Richard R.
Simonds, my chairman, for providing the timely guidance and encouragement. To
Dr. John L. O’Donnell for his inspirational comments and wisdom, along with introducing
me to Mr. Ronald G. Choura. To Dr. Michael A. Mazzeo for his insightful comments that
basically got the writing stage of this dissertation started. And to Dr. R. V. Ramamoorthi
for his support throughout my graduate studies at Michigan State University.

Though not a direct member of my committee, I would also like to extend my
appreciation to Mr. Ronald G. Choura for his assistance in obtaining the data information
necessary to complete the empirical portion of this dissertation. I shall always remember
his standard reply of “N 0 problem” to all of my data requests.

I would like to thank my family for their assistance and encouragement. Without their
help through the dark ages, I would not have been able to complete this degree.

Without reservation, I would also like to acknowledge my Lord and Savior for
answering my prayers for assistance. Without His help, I would not have been able to
meet such wonderful people like Jim Gerhardinger, Sheri Tice, and Tony Altobelli. Thank

you all for your help and encouragement.

TABLE OF CONTENTS

Page

LIST OF TABLES ix

LIST OF FIGURES xi
CHAPTER 1

INTRODUCTION .................................... 1
CHAPTER 2

LITERATURE REVIEW ................................ 5

Financial Theory .................................. 5

The Double Leverage Approach .................... 5

The Independent Firm Approach ................... 7

Other Alternatives ............................. 12

Empirical Evidence ................................ 15

Subsidiary Rate of Return Empirical Study ............ 15

Tobin's Q Empirical Studies ...................... 17
CHAPTER 3

PUBLIC UTILITY COMMISSION REPORTS REVIEW ......... 19

Double Leverage Rate Cases ......................... 19

Consolidated Tax Judgments ......................... 22
CHAPTER 4

TOBIN'S Q DERIVATION ............................. 27

Market Value ................................... 27

Common Stock .............................. 27

Preferred .................................. 27

Debt ..................................... 27

Replacement Cost ................................ 29

Plant and Equipment .......................... 29

The (Intermediate) Results .......................... 31

CHAPTER 5 Page
AN ALTERNATIVE DERIVATION OF THE EZZELL, HSU AND

MILES MODEL ................................ 33
The Perpetual Cash Flow Model ..................... 33
One Subsidiary, N 0 Debt at the Subsidiary Level .......... 35
Independent Market Value Balance Sheet ............ 36
Adjusted Market Value Balance Sheet .............. 37
Independent Cash Flows ...................... 38
Adjusted Cash Flows ......................... 40

One Subsidiary, Debt at Both the Subsidiary and Parent Levels . 40
Independent Market Value Balance Sheet ............ 41
Adjusted Market Value Balance Sheet .............. 42
Independent Cash Flows ...................... 43
Adjusted Cash Flows ......................... 44

Two Subsidiaries, No Debt at the Subsidiary Level .......... 45
Independent Market Value Balance Sheet ............ 45
Adjusted Market Value Balance Sheet .............. 46
Independent Cash Flows ...................... 47
Adjusted Cash Flows ......................... 48
Double Leverage Approach ..................... 49
Double Leverage Market Value Balance Sheet ..... 50

Double Leverage Adjusted Cash Flows ......... 50

Two Subsidiaries, Debt at Both the Subsidiary and Parent Levels 52
Independent Market Value Balance Sheet ............ 53
Adjusted Market Value Balance Sheet .............. 54
Independent Cash Flows ...................... 55
Adjusted Cash Flows ......................... 56
Other Allocation Pairings .......................... 57

Results Comparison Between This Derivation and EHM (1991) 60
Comparison Results, Discounting Using the MM Context 60
Comparison of the MM and EHM Discounting Methodology 62

CHAPTER 6
COMPARATIVE ANALYSIS OF THE ALTERNATIVE RETURN

ADJUSTMENT PROCEDURES ..................... 67
Tax Allowance Equations .......................... 68
Stand-Alone Method Using the Subsidiary's Interest Expense 68
The Narragansett Electric Company Method .......... 68
The Southwestern Bell Telephone Company Method . . . . 69
The New England Telephone & Telegraph Company Method 69
The Newton Water Company Method .............. 70
The Brockton Edison Company Method ............ 70

The Stand-Alone Method Using the Combined Parent/
Subsidiary Interest Expense ................ 71
The Actual Taxes Paid Method .................. 71
The EHM Method ........................... 71
The Beedles Method .......................... 72
The Double Leverage Approach .................. 73
The Perpetual Cash Flow Model ...................... 74
Scenario Results ................................. 78

vii

CHAPTER 6 (con'd.) Page
Initial Scenario .............................. 78
Second Scenario ............................. 79
Third Scenario .............................. 79
Fourth Scenario ............................. 80
Fifth Scenario .............................. 80
Sixth Scenario .............................. 82
Summary of Results .............................. 83
Performance Criteria .......................... 83
Evaluation of the Alternative Methods ............... 83
CHAPTER 7
EMPIRICAL STUDY ................................. 96
Data Selection ................................... 96
Hypotheses .................................... 97
The Statistical Model ............................... 98
Results ........................................ 101
Ordinary Least Squares Regression ................. 101
Positive Autocorrelation Corrective Action, Iterative
Approach .................................. 102
Positive Autocorrelation Corrective Action, First Difference
Approach .................................. 103
N on-Linear Regulated Sales Regression Results ........ 104
Natural Log Application ........................ 105
CHAPTER 8
CONCLUSIONS .................................... 1 13
APPENDD(
MARKET-TO—BOOK EMPIRICAL RESULTS ............... 116
Results ....................................... 1 16
Linear Regulated Sales Variable Results ............. 116
Non-Linear Regulated Sales Variable Results ......... 118
LIST OF REFERENCES ...................................... 125

LIST OF TABLES

Table Page
1 ADJUSTED WACCk SOLUTIONS (for k = 1,2) ............. 59
2 INTEREST TAX SHIELD ADJUSTMENT COMPARISONS,

INITIAL SCENARIO ........................... 85
3 INTEREST TAX SHIELD ADJUSTMENT COMPARISONS,
SECOND SCENARIO .......................... 86
4 INTEREST TAX SHIELD ADJUSTMENT COMPARISONS,
THIRD SCENARIO ............................ 87
5 INTEREST TAX SHIELD ADJUSTMENT COMPARISONS,
FOURTH SCENARIO .......................... 88
6 INTEREST TAX SHIELD ADJUSTMENT COMPARISONS,
FIFTH SCENARIO ............................ 89
7 INTEREST TAX SHIELD ADJUSTMENT COMPARISONS,
SIXTH SCENARIO ............................ 90
8 RATE OF CHANGE COMPARISON, TABLE 3 TO TABLE 2 . . . . 91
9 RATE OF CHANGE COMPARISON, TABLE 4 TO TABLE 2 . . . . 92
10 RATE OF CHANGE COMPARISON, TABLE 5 TO TABLE 3 . . . . 93
11 RATE OF CHANGE COMPARISON, TABLE 6 TO TABLE 5 . . . . 94
12 RATE OF CHANGE COMPARISON, TABLE 7 TO TABLE 6 . . . . 95
13 TOBIN'S Q REGRESSION RESULTS, ORDINARY LEAST
SQUARES REGRESSION MODEL, LINEAR REGULATED
SALES VARIABLE ............................ 107
14 TOBIN'S Q REGRESSION RESULTS, AUTOCORRELATION
CORRECTIVE ACTION, ITERATIVE APPROACH,
LINEAR REGULATED SALES VARIABLE ........... 108

Table
1 5

16

17

18

19

20

21

22

23

24

TOBIN'S Q REGRESSION RESULTS, AUTOCORRELATION
CORRECTIVE ACTION, FIRST DIFFERENCE
APPROACH, LINEAR REGULATED SALES VARIABLE .

TOBIN'S Q REGRESSION RESULTS, ORDINARY LEAST
SQUARES REGRESSION MODEL, NON -LINEAR
REGULATED SALES VARIABLE .................

TOBIN'S Q REGRESSION RESULTS, AUTOCORRELATION
CORRECTIVE ACTION, ITERATIVE APPROACH,
NON-LINEAR REGULATED SALES VARIABLE .......

TOBIN'S Q REGRESSION RESULTS, AUTOCORRELATION
CORRECTIVE ACTION, FIRST DIFFERENCE
APPROACH, N ON -LINEAR REGULATED SALES
VARIABLE ..................................

MARKET-TO—BOOK REGRESSION RESULTS, ORDINARY
LEAST SQUARES REGRESSION MODEL, LINEAR
REGULATED SALES VARIABLE ..................

MARKET-TO-BOOK REGRESSION RESULTS, AUTO-
CORRELATION CORRECTIVE ACTION, ITERATIVE
APPROACH, LINEAR REGULATED SALES VARIABLE .

MARKET-TO-BOOK REGRESSION RESULTS, AUTO-
CORRELATION CORRECTIVE ACTION, FIRST
DIFFERENCE APPROACH, LINEAR REGULATED
SALES VARIABLE ............................

MARKET-TO—BOOK REGRESSION RESULTS, ORDINARY
LEAST SQUARES REGRESSION MODEL, NON-LIN EAR
REGULATED SALES VARIABLE ..................

MARKET-TO-BOOK REGRESSION RESULTS, AUTO-
CORRELATION CORRECTIVE ACTION, ITERATIVE
APPROACH, NON -LINEAR REGULATED SALES
VARIABLE ..................................

MARKET-TO-BOOK REGRESSION RESULTS, AUTO-
CORRELATION CORRECTIVE ACTION, FIRST
DIFFERENCE APPROACH, NON-LINEAR
REGULATED SALES VARIABLE ..................

Page

109

110

111

112

119

120

121

122

123

124

LIST OF FIGURES

Figure Page
1 EXPANDED ADJUSTED WACC SOLUTIONS .............. 60

CHAPTER 1
INTRODUCTION

The purpose of this dissertation is to reconcile the techniques for determining the cost
of equity for wholly owned regulated utilities, between those defined by financial literature
and regulatory practice. Specifically, I consider whether or not regulatory practices are
based on financial theory when a company uses debt financing at both the parent and
subsidiary levels. Anecdotally, some believe that regulatory rulings in setting "fair" rates of
return have moved away from financial theory ."1 If this is true then the objective measure
of what is "fair" in theory is not being practically applied.

The utilization of leverage on two separate organizational levels is referred to as “double
leverage”. For example, assume Company S whose assets are financed with both debt and
equity, operates as a regulated, public utility. The common stock of Company S is in turn
owned by Company P, who acquired a portion of the funds used to obtain Company S's
common stock with debt. A single cash flow generating process has therefore been
leveraged twice. The question that arises is, "What should be Company S's allowable rate
of return for equity?" Since Company S's common stock is not publicly traded, the
normal security market directly determined rate cannot be utilized.

Double leverage has more significance in regulated than in unregulated industries. One
of the goals of public utilities regulators is to set revenue requirements that will permit the
utility to cover their operating expenses, while also providing a fair rate of return to
investors. These rates of return must be high enough to attract investment capital, while at
the same time not so high as to allow excessive profits at the expense of the utility's
customers. Achievement of this goal becomes more difficult when the utility is a wholly

 

1Measuring the cost of capital for unregulated firms with this type of debt financing is
equally important, but empirically difficult to test because unregulated firms are not
required to disclose rate of return information for each subsidiary. Thus the topic is more
easily pursued within the regulated industry context.

1

2
owned subsidiary, with its parent utilizing debt to support their equity investment in the
regulated utility.

Although a parent/subsidiary relationship is commonplace in the regulated utility
industries, double leverage primarily occurs in the water and telecommunications
industries. The Public Utility Holding Company Act of 1935 does not allow parent
companies to issue debt for the purpose of acquiring common stock of electric and gas
utilities. However, this restriction does not apply to the water and telecommunications
industries. An example of the usage of double leverage in the telecommunications industry
is the GTE Corporation. GTE has stated in Moody's Public Utilities Manual that they own
no plant, real property, franchises, or concession except indirectly through their
investments in subsidiaries, yet they show $4,833,443,000 of parent level debt, together
with $9,468,572,000 of debt at the subsidiary level. Presumably, some portion of GTE's
parent level debt has been utilized to support its equity holdings in the equity positions of
its regulated affiliates.

This issue has become more complicated in recent years because the parent companies
have moved into unregulated investment opportunities. If regulators allow too high a
return on the subsidiary's equity, the parent's unregulated business interests could be
subsidized by its regulated investments at the expense of the regulated subsidiary's rate-
payers. The opposite would occur if too low a return were set. The primary issue of the
double leverage approach being utilized to determine an allowable rate of return on equity
for regulated holding company affiliates, involves the affects of these cross subsidizations.

Historically, the literature has concentrated on two approaches to determine an
allowable rate of return for a wholly owned subsidiary, where parent debt has been used to
obtain part of the subsidiary's equity, -- the independent firm approach and the double
leverage approach.

The independent approach assumes the parent and subsidiary are independent of one
another. The parent's debt position, and, in turn its related interest tax shield are therefore
ignored. The rate of return on the subsidiary's equity is estimated using the market
determined costs of equity of other publicly traded firms of similar risk The independent
firm approach is supported by Brennan and Humphreys (1973), Lerner (1973), Brown

3

(1974), Jones and O'Donnell (1978), Pettway and Jordan (1983), Rozeff (1983)2, and
Beranek and Miles (1988).

The double leverage approach assumes the parent and subsidiary are not independent.
It uses the parent's weighted average cost of capital as the subsidiary's cost of equity. It
therefore recognizes the parent's debt position, along with its related interest tax shield.
The supporters of the double leverage approach include Bachman and Kristen (1972),
Copeland (1977), and Seeds (1978).

In addition to the independent firm and double leverage approaches, other approaches
have been included in studies by Fitzpatrick (1977) , Beedles (1984), and Ezzell, Hsu, and
Miles (1991). Fitzpatrick's adjusted capital cost method, (1) assumes that the subsidiary's
capital ratios to be equivalent to the consolidated capital ratios for the entire holding
company; and (2) computes the subsidiary's true cost of equity capital consistent with the
operating risk characteristics of the subsidiary. Given a parent, two subsidiary case,
Beedles' capital structure method allocates the parent's debt to each subsidiary, based the
operating risk of each subsidiary. Ezzell, Hsu, and Miles ( 1991) bring out some
interesting criticisms concerning both of these methods. Their main criticism of the double
leverage approach is that it ignores differences in risk among the subsidiaries. The
independent firm approach has a problem because it allows an equity wealth creation at the
parent level by ignoring the interest tax (shield) savings of debt canied at the parent level.
Ezzell, Hsu, and Miles build a model that uses this interest tax savings to lower the
subsidiary's cost of capital.

Regulatory public utility commissions do not currently utilize the double leverage
approach. In its place a majority of the state commissions, including most of those that had
previously used the double leverage approach, utilize the independent firm approach and
require firms to use a consolidated tax adjustment. The original intent of the this tax
adjustment imposed by regulatory authorities had been meant to deal with subsidiary
operating losses. Since the parent/subsidiary system was allowed immediate tax credits as
a result of these losses, the tax adjustment spreads the system's tax credits to all of the

 

2Rozeff‘s article proposes a modified double leverage approach that actually
determines the independent firm approach results for a parent, two subsidiary case. His
approach determines one subsidiary's equity return when the other subsidiary's equity
1rienturn, the parent's equity return, and the parent's weighted average cost of capital are
own.

4
system's profitable subsidiaries, not just to the parent and ultimately the parent's equity
shareholders.

State public utility commissions that utilized the double leverage approach in the early
1980's were concerned with the parent firm's shareholders receiving a return in excess of
their "fair" market return when debt was issued at both parent and subsidiary levels.
According to financial theory, the double leverage approach eliminates any wealth creation
at the parent level caused by the parent's interest tax shield. The regulatory commissions
have utilized the independent firm approach to adequately compensate the parent for
investment risk, but have also reduced the parent's interest tax shield affect through the use
of a consolidated tax adjustment. Independently, Ezzell, Hsu and Miles have derived a
methodology which again uses the independent fum approach to compensate for
investment risk, along with lowering the allowed return of each subsidiary to reduce the
parent's interest tax shield affect Based on the past practices of public utility commissions
and the recent developments in financial theory as discussed above, several questions
naturally arise. First, is the methodology developed through the regulatory proceedings
essentially equivalent to that presented by Ezzell, Hsu and Miles? Second, are either of
these two procedures consistent with the alternative approaches provided by Fitzpatrick and
Beedles? Third, can the effect of a consolidated tax adjustment utilized by a majority of the
public utility commissions be empirically tested?

In order to investigate these questions, this study is organized as follows. Chapter 2
provides a thorough literature review of the financial theory and empirical evidence that deal
with the concept of double leverage and its affect on the allowable cost of equity for wholly
owned regulated utilities. Chapter 3 reviews the W of rate case
decisions regarding the previously mentioned area of interest. Chapter 4 covers the
descriptive value to be utilized in the empirical test, Tobin's Q. Chapter 5 focuses on the
Ezzell, Hsu and Miles methodology, providing a better understanding of the Ezzell, Hsu
and Miles approach. Chapter 6 expands the perpetual cash flow model utilized in chapter
5, comparing the results of eleven different methodologies. Chapter 7 covers the empirical
test and its results. Chapter 8 provides concluding remarks.

CHAPTER 2
LITERATURE REVIEW

The research questions which I investigate encompass both theory and the empirical
testing of the theory. Thus this chapter reviews papers that provide normative models and
empirical results related to the regulation of allowed rates of return for a wholly owned
regulated subsidiary.

financial Ihegry

Theoretical literature concentrates on two approaches in addressing an allowable rate of
return for a wholly owned regulated subsidiary, the double leverage approach, and the
independent firm approach. There are three published articles supporting the double
leverage approach, seven articles favoring the independent firm approach, and three articles
that offer alternative approaches. The following three subsections discusses these papers
and their approaches.

WM

Backrnan and Kirsten (1972) is the earliest article found supporting the double leverage
approach. They state that when a holding company owns all of the common stock of an
operating company, it is not appropriate to determine the operating company's return on
equity by the usual comparatives with other utilities or industrial company common stock.
Because the operating company's common stock is not publicly traded, it does not have to
meet the capital attraction standard.3 The capital attraction standard must be met by holding
company securities, because they are the issues available to the public. So long as the

 

3The capital attraction standard applies to the firm's ability to attract additional capital
through the sale of its securities. Here the point in question relates to a wholly owned
regulated subsidiary being able to obtain additional equity capital from its parent company
as funding requirements arise.

5

6
holding company can attract the funds required by their operating companies, the capital
attraction standard has been satisfied. Because the offered securities often include debt
securities, these funds are usually obtained at a lower cost than if the operating company's
common stock is sold directly to the public. The ratepayer should benefit from these
savrngs.

Copeland (1977) asserts that the logical basis for the double leverage approach flows
from the context of the modern theory of capital budgeting. Competition and goal
maximization induces firms to increase investment as long as the internal rate of return on
the last investment project undertaken just equals the cost of capital. In competitive
equilibrium the net present value of the last investment project undertaken is zero. In turn,
for the rationale of utility regulation to emulate the competitive result, restricting the holding
company's return on investment, the equity of a wholly owned regulated subsidiary should
be the holding company's weighed average cost of capital.

Copeland addresses some common criticisms of the double leverage approach. He
provides that the double leverage approach does not discriminate against the holding
company arrangement, since it is obvious that subsidiaries of holding companies are
different than fums that are not subsidiaries. Because they are different, commissions must
be careful to determine whether or not the difference has implications for regulatory practice
and methodology. Copeland claims that included in the diversity between a subsidiary and
independently owned operating utility is a difference in the financial risk for capital invested
in the two utilities. He also refutes the argument that for the double leverage approach to be
valid it must be carried to its ultimate conclusion, applying the double leverage concept to
the individual investors who purchase the parent company's common stock. When
individual investors employ personal leverage so as to earn more than the return they
actually require on personal funds, disequilibrium exists which cannot last in a competitive
capital market Equilibrium is restored when the return on personal funds earned by the
marginal investor just equals the required return. In equilibrium the market yield on the
holding company's stock also just equals the marginal investor's weighted average cost of
capital. Because the competitive forces of the marketplace prevent the marginal investor
from earning any more than his weighted average cost of capital on his common stock
investments in the holding company, there is no need to worry about the problem of
"triple" leverage.4

 

4The immediate problem with Copeland's logic conceming "triple" leverage is that this
same argument could explain away any concern for worrying about "double" leverage.

7

Seeds (1978) is the final article supporting the double leverage approach. He poses that
determining a regulated utility's overall cost of capital for rate-making purposes is a very
complex issue. However, this complexity can be reduced by simply applying the double
leverage approach. This straightforward analytical analysis of the parent company's and
subsidiary's capital structures determine the overall rate of return to be applied to the
subsidiary to recover all capital costs of the subsidiary and the parent company.

In summary, four key points arise from these papers. First, the capital attraction
standard must be met by only the securities that are directly available to the public. Second,
the double leverage approach flows from the context of capital budgeting theory. Third,
double leverage's recognition of the differences in financial risk is the cause of any
differences when comparing its allowed rate of return for a wholly owned regulated
subsidiary to that of an independently owned operating utility. Finally, the double leverage
approach provides a simple, straightforward solution to the original complex issue
regarding a fair and reasonable rate of return on the equity of a wholly owned regulated
subsidiary.

W

The proponents of the independent firm approach are not only greater in number than
those of the double leverage approach, but their support of the independent firm approach
has continued over a longer period of time.

Brennan and Humpreys (1973) supported the independent firm approach by focusing
on the apparent weaknesses of the double leverage approach. They argue that the double
leverage approach does not make the wholly owned subsidiary's allowable cost of equity
funding dependent upon the cost of using the funds in its business, but upon its cost to the
purchaser or owner of the subsidiary's securities. This results in shifting the risk
governing the return allowed to the owner rather than the issuer of the securities. They cite
that the marketplace for capital does not concern itself with who owns the business, or the
source of the owner's capital, but with the earning power of the assets of the business and
the risk related to such ownership.

Lerner (1973) attacked the capital attraction standard criteria by Backrnan and Kirsten
cited earlier. Lerner argues that it would be contrary to sound management principles for a
holding company to retain or purchase additional common stock that did not meet or have

8
the potential to meet capital attraction standards. He also states that regulatory authorities
should not give consideration to the capital structure, sources of capital, or cost of capital of
a parent company in determining a subsidiary company's allowable rate of return for two
reasons: (1) the holding company form of organization may permit the entire entity to
borrow more funds than the individual operating companies, and (2) the debt of the parent
itself cannot be assigned legally, and should not be assigned hypothetically to any of the
operating utilities.

Brown (197 4) attacked the intent of treating the allowed return of an operating utility
differently, based on the ownership of its common stock. He states that when the double
leverage approach is utilized, the allowed cost of equity for wholly owned operating
utilities are lower than if these same operating utilities are allowed a higher return on equity
based on risk when owned by individual stockholders. As a result identical utility assets
employed in exactly the same use and under the jurisdiction of the same regulatory
authority can possess unequal return allowances. Brown is the first to point out that
imposing the parent's weighted average cost of capital for all regulated subsidiaries of a
holding company ignores any differences in risk between the subsidiaries.

Jones and O'Donnell (1978) challenge the regulatory legality of the double leverage
approach. They pose that the double leverage approach requires the assumption that each
subsidiary should earn identical rates of return on equity, so that the parent's equity return
does not exceed the allowance determined by the state commission. This constitutes
regulation of the holding company, not the operating subsidiary, which exceeds the state
commission's regulatory authority. They also point out that just as market evidence is not
available on the wholly owned subsidiary's common stock, valid market evidence
concerning the parent is likewise unavailable. Market evidence of the parent holding
company's publicly traded common stock is applicable to the consolidated operation, and
cannot be utilized to determine a cost of equity as if the parent is standing alone.

Pettway and Jordan (1983) follow a different approach to those previously taken. They
analyze the aspects of the independent firm and double leverage approaches utilizing the
following objective standards of rate of return regulation that they obtained from the
Bluefield and Hope cases5 and regulatory practices:

 

5Bluefield Water Works and Improvement Co. v. Public Service Commission of West
Virginia (1923) 262 US. 679, 692; Federal Power Commission v. Hope Natural Gas Co.
(1944) 320 US. 591, 603.

9
1. That the allowed return must be sufficiently low so as to eliminate "monopoly
rents" or "producer's surplus" in the operating company.6
2. That the allowed return must be sufficiently high so as to attract capital and to guide
the allocation of capital resources in a socially desirable fashion.
3. That the allowed return must exactly compensate investors of capital for the
riskiness of their investments in the public utility.

Pettway and Jordan find that double leverage provides a rate of return that would meet all
three criteria if and only if one of the following conditions are met:

1. The parent has only one subsidiary.

2. All the subsidiaries have equal systematic risk.

3. The systematic risk of the particular regulated subsidiary is equal to the weighted
average systematic risk of all the subsidiaries.

They find that the independent firm approach satisfies all three standards without the use of
any restrictions.

In a similar yet slightly different manner, Rozeff (1983) points out that the primary fault
of the double leverage approach is that it does not take into account the presence of other
subsidiaries. Rozeff then provides what he terms as "The Modified Double Leverage
Approach". Rozeff starts by establishing a fundamental financial equality :

In a competitive financial market, the weighted average required return on a set of
assets whose risks can differ equals the weighted average capital costs of the financial
securities used to finance the assets.

Utilizing the legal framework of the Hope Natural Gas case cited earlier by Pettway and
Jordan and his fundamental financial equality, he uses a parent, two subsidiary numerical
example to show that his modified double leverage approach properly arrives at a fair rate
of return. In his example, Rozeff shows that when the capital costs of debt for the parent
and both subsidiaries are known, along with the costs of equity for the parent and the first
subsidiary, the required rate of return on equity for the second subsidiary can be
determined. Although not stated, the results of Rozeffs modified double leverage

 

6The terms "monopoly rents" and "producer's surplus" apply generally to that portion
of the return on investments which is above the utility's cost of capital. (Taken from the
authors' footnote #2.)

10
approach are found to be consistent with those of the independent firm approach, when the
parent's cost of equity is a resultant of the costs of equity of its subsidiaries.

Beranek and Miles (1988) define a concept similar to the one posed by Rozeff. They
first state that the double leverage approach implicitly assumes that if the holding company
(not the utility) is permitted to earn its cost of capital, the objective standards of rate of
return regulation automatically be upheld for the operating regulated subsidiary. They
counter that allowing a holding company to earn its apparent cost of capital does not assure
its regulated subsidiaries are earning their costs of capital. Their basis for making this
statement is built around the concept that the parent company's cost of equity is determined,
at least in part, by the subsidiaries' costs of equity. Further, the parent's weighted average
cost of capital is itself a weighted average of the equity costs of all its subsidiaries.
Alternatively, the double leverage approach uses the parent's capital costs to determine the
subsidiary costs, which in most cases is inaccurate. They pose that if security prices are in
equilibrium, the total expected cash flow to the parent's claimants must equal the total
expected cash flow from the parent's interest in each of its subsidiaries depicted by the
following equation:

Epsp + £be = Eiksk (2.1)
where: Sk = value of equity in subsidiary k; k = 1, ..... , n
rk = expected rate of return on subsidiary k's equity; k = 1, ..... , n
Sp = value of the parent's equity
ip = expected rate of return on the parent's common stock
B1) = value of the parent's debt
rb = expected rate of return on the parent's bonds

If equation (2.1) is divided by Vp = Sp + Bp, the total value of the holding company, then

- S - B
b -
Jvg‘p'p + —Ev = Xxkrk (2.2)
p
Sk
where: xk = V—

1 1
The left side of equation (2.2) is the parent's weighted average cost of capital, while the
right side is a weighted average of all subsidiary equity costs. Since ZSk = Vp, the
parent's weighted average cost of capital is itself a weighted average of the equity costs of
all its subsidiaries.7 Given equations' (2. 1) and (2.2), Beranek and Miles point out that the
conditions under which the double leverage approach is valid are:

1. The parent has exactly one subsidiary.
2. Two or more regulated subsidiaries have identical costs of equity.

Beranek and Miles conclude their paper by stating that accepted financial theory leads to
the conclusion that the interests of ratepayers and capital suppliers are better served by
evaluating a subsidiary utility's cost of capital on the constitutionally valid independent firm
approach.

In summary these papers bring out four key points. First, the allowed cost of equity
for a wholly owned regulated subsidiary should be based upon the inherent risk of the
operating subsidiary, not who owns the operating utility or the source of the owner's
capital. Stated in a slightly different manner, the capital attraction standard should be
applied equally to all public utility companies regardless of ownership. Second, the parent
holding company and each of its wholly owned regulated subsidiaries are separate,
individual, economic entities, making the parent's capital structure, sources of capital, or
cost of capital irrelevant when determining a regulated subsidiary's allowable rate of return.
Third, employing the double leverage approach by imposing the parent holding company's
cost of capital as the allowed equity return on each of the parent's regulated subsidiaries
constitutes regulation of the parent holding company, not the operating utility. Last, the
independent firm approach meets all of the rate of return for regulated utilities criteria free
of any restrictions, while the double leverage approach can do so only if there exists only
one subsidiary, or if each regulated subsidiary within a holding company organization has
identical systematic risk.

 

u:iT‘his statement is noted to be similar in concept to Rozeff's fundamental financial
0‘1 1W-

12
MW

There have been other alternatives presented in the literature besides the double leverage
and the independent firm approaches. Fitzpatrick (1977) provides that there are basically
three different approaches to calculating the cost of capital for individual subsidiaries of
public utility holding companies:

1. Regulators employing the parent company's cost of equity capital and capital
structure in determining an individual subsidiary's weighted average cost of capital.

2. Determining subsidiaries' cost of capital based on their own merits regardless of
their parent's earnings history and capital structure (the independent firm approach).

3. The double leverage approach where the parent's weighted average cost of capital is
used as the allowed cost of equity for each subsidiary.

Fitzpatrick states the first concept stems from the argument that the different
subsidiaries of the holding company are ultimately financed from the same pool of financial
resources. He poses that this procedure ignores the premise that risk differentials between
subsidiaries should be compensated with different equity returns to the individual
subsidiaries. The second procedure ignores the possibility of the parent company electing
to create "artificial" capital structures for its subsidiaries that could not be supported in the
market without the financial support of the parent company. The third approach is
theoretically flawed since the parent's return is dependent on the financial characteristics of
the utility investments themselves. It is also unfair to corporate investors who may have
considerable leverage in their own capital structures.

Fitzpatrick next presents a two step alternative:

1. Determine if the subsidiary's capital structure is artificial, looking for variances
between the capital ratios within the same holding company system, or the average
capital ratios for the industry. If it appears that a subsidiary's capital ratios are
normal, its overall capital costs should be based on its reported capital structure.

If not, the consolidated capital ratios for the entire holding company should be
utilized.

2. Compute the subsidiary's true {cost of equity capital utilizing costs of equity of
publicly traded utilities that exhibit similar operating risk characteristics of the

13
operating subsidiary.

Reviewing Fitzpatrick's two step procedure, its results are consistent with the independent
firm approach if the capital structure of the subsidiary is not considered to be artificial. If
the subsidiary's capital structure is considered to be not normal, then the capital ratios for
the entire holding company should be used while still utilizing the subsidiary's costs of
financing (i.e. debt and equity).

Beedles (1984) follows a slightly different approach to that of Fitzpatrick. He asserts
that instead of altering return estimates to reflect subsidiary risk, one should adjust the
subsidiary's capital structure to reflect operating risk while employing the component costs
that are estimated for the consolidated holding company. In a two subsidiary scenario,
given equally risky subsidiaries, the parent's debt can be allocated in a manner so that the
resultant weighted average cost of capital for both subsidiaries are made equal. When one
subsidiary is considered to have greater operating risk, following financial economic
theory, a lower debt ratio is assigned to the riskier subsidiary, resulting in the weighted
average cost of capital of the riskier subsidiary being greater than that of the second
subsidiary.

Ezzell, Hsu and Miles (1991) focuses on the two primary approaches, finding
weaknesses in both the double leverage and the independent firm approaches. They state
that the independent firm approach has the ability to recognize differences risk between
subsidiaries in a holding company organization. However, by ignoring the utilization of
debt at the parent level allows the parent's interest tax shield benefits to be incorrectly
passed on to the parent company's common stockholders at the expense of the regulated
operating subsidiary's ratepayers. Conversely, the double leverage approach recognizes
the utilization of debt at the parent level, returning the parent's interest tax shield benefits
back to the ratepayers. The double leverage approach ignores any differences in risk
among a holding company's regulated subsidiaries. Their article portrays an alternative
approach that recognizes any differences in subsidiary risk, while at the same time returns
the parent's interest tax shield benefits to the subsidiary's ratepayers. This is accomplished
in the following manner:

1. Each subsidiary's unadjusted weighted average cost of capital is found utilizing the
independent firm approach, which accounts for any risk differences between
subsidiaries.

14
2. The present value of the parent holding company's interest tax shield benefits are
allocated back to the subsidiaries by reducing each subsidiary's unadjusted
weighted average cost of capital.

With the use of a perpetual cash flow model where the parent's interest tax shield benefits
are allocated back to the subsidiaries based on each subsidiary's investment, Ezzell, Hsu
and Miles (El-1M) define this reduction process in the form of an equation:

R
Zp SWLEW)

1-
S

where: WACCk“

subsidiary k's adjusted weighted average cost of capital

WACCk = subsidiary k's unadjusted weighted average cost of capital
zp = tkdp/(l+rf)
t = corporate tax rate
kdp = the cost of debt of the parent
rf = the risk-free interest rate
st = (1+ksw)/ksw
kSW = the unlevered cost of equity of the parent
Lp = the parent's market value debt ratio
11.11.
L“ = 31?
I1‘ = investment in subsidiary k
Lk = the market value debt ratio for subsidiary k

They define the parent's unlevered cost of equity, k as being equal to the summation of

SW’

the subsidiaries' "equity" weighted cost of equity.

 

Algebraically,
_ stk(l'l’k)va
8'” 2(1-Lk)VLk
where: ksk = the levered cost of equity for subsidiary k

the market value debt ratio for subsidiary k

Lk
VLk = the levered value of subsidiary k

15

Two of these variables, z.p and st, are a result of the manner in which EHM discount the

parent's interest tax shield. For the first period the parent's interest tax shield, tkdprVLp,

is discounted at the risk-free interest rate. For periods' two through infinity the parent's
interest tax shield is discounted at the parent's unlevered cost of equity, ksw. When placed

the more familiar Modigliani and Miller perpetual cash flow model where the parent's
interest tax shield is discounted every period at the parent's cost of debt, kdp, the

adjustment equation becomes:

WACCk“ = WACCk (71% )

where: WACCk“

subsidiary k's adjusted weighted average cost of capital

WACCk = subsidiary k's unadjusted weighted average cost of capital
t = corporate tax rate
Lp = the parent's market value debt ratio
IkLk
LW - 2:1k
Ik = the investment in subsidiary k
Lk = the market value debt ratio for subsidiary k

Summarizing, there was one common idea presented by each of the three "other"
alternatives: The debt carried at the parent level of a holding company organization cannot
be ignored. Of the three alternative procedures presented, the approach by EHM provides
the most objective alternative, providing a definitive equation for their adjustment
procedure. The alternatives by Fitzpatrick and Beedles will not be ignored, but the EHM
perpetual cash flow model will be the main focal point of chapter 5.

Empirical Evidencc

i' R fR mEmiri

To date, the empirical evidence in the literature regarding the issue of which approach
is more appropriate to utilize has been extremely sparse. O'Donnell and Walker (1987)
performed an event study which attempts to test market response to regulatory rulings
imposing the use of either the independent fum or double leverage approaches. They
circumvented the problem of not being able to directly observe the equity returns for

16
affiliated utilities by utilizing American Telephone and Telegraph (AT&T) affiliates that also
had existing minority equity positions. Twenty-two rate cases are tested over the 1968-
1980 period involving three AT&T affiliates having minority equity positions with publicly
traded stock. An 105 day estimation period (from t=-l35 to t=-3l) is used, and cumulative
average abnormal returns are obtained for a 61 day observation period (from t=-30 to
t=+30). The event date (t=0) is defined as the date the regulatory agency signed the rate
order as identified in the W. O'Donnell and Walker report that when
the regulatory agency used the independent firm approach investors earn normal returns,
while a double leverage adjustment results in a cumulative average abnormal return that
decreases 5.4% during the 61 day observation period.

While providing what would be considered logical results, three immediate concerns

arise:

1. Of the 22 rate cases, on only two occasions is there a reversal in the agency's
utilization of the two alternative procedures, reducing the effect of the agency's final
decision, unless the market had mistakenly expected a reversal that did not occur.

2. As brought out by O'Donnell and Walker, "Event studies are based on the
assumption that material information is publicly disclosed on one identifiable event
date. In contrast, rate cases are conducted in open hearings so that the outcome is
not necessarily an unanticipated event" There also can be an extended period
between the agency's decision and the actual signing of the rate order.

3. The time period used (1968-1980), which was used because of the AT&T affiliates'
having minority equity positions is now rather dated.

Given the results of this lone empirical study, I will attempt to study the affects of
public utility commission regulation upon the value of the firm. Tobin's Q will be utilized
to measure the profitability of holding companies that possess wholly owned regulated
subsidiaries. Instead of utilizing an event study approach, using calendar year-end data, a
Tobin's Q will be computed for each year-end, measuring the firm's profitability at that
point in time. Any regulatory rulings affecting the firm throughout the year will then be
reflected by a change in its Tobin's Q.

17

I!"DE"151'

The use of Tobin's Q as a measure of profitability is first introduced by Lindenberg and
Ross (1981). Subsequent studies have included such areas as market efficiency (Smirlock,
Gilligan and Marshall, 1984; Stevens, 1990), tender offers (Lang, Stulz and Walking,
1989 and 1991), takeovers (Chappell and Cheng, 1984; Servaes, 1991; Griffin and
Wiggins, 1992; Kin, Henderson and Garrison, 1993), Leveraged Buyouts (Opler and
Titrnan, 1993), and insider trading (John and Mishra, 1990; McConnell and Servaes,
1990). Also included in the literature concenring regulated utility regulation is an article by
Helrnuth (1990).

Helrnuth (1990) focuses on the affect of regulation on electrical utilities. Using 45
electric utilities, Helmuth determines an average yearly Tobin's Q for the years 1981-1984.
He then compares this electric utility average Tobin's Q to an average Tobin's Q for all
nonfmancial firms. He finds that a t-test for each yearly comparison can not reject the
hypothesis that an average electric utility Tobin's Q is equal to the average nonfinancial
firrns' Tobin's Q at the 1% confidence level. Helrnuth concludes that his results provide
evidence that the electric utilities earned a normal return from 1981 through 1984.

Smirlock, Gilligan and Marshall (1984) are interested in obtaining statistical support for
conflicting hypotheses concerning market concentration. An average Tobin's Q is
determined for each firm for the 1961- 1968 time frame. These averaged Q's are then
linearly regressed against the variables market share and concentration ratio. Control
variables, barriers to entry and market share growth, are also included. They find that the
coefficient for market share is positive and significant at the 1% level, while the
concentration ratio coefficient is not significant at any conventional level. These results did
not change when the control variables are included. The results provide support for the
efficient structure hypothesis.

Stevens (1990) took the work of Smirlock, Gilligan and Marshall, and expands the
regression to include the additional interactive variable of (market share x concentration
ratio). Including this additional term in the regression, Stevens obtains results that did not
support the conclusions of Smirlock, et al.

McConnell and Servaes (1990) are interested in statistically testing the relationship
between management ownership and firm value. Tobin's Q's are obtained for nonfinancial

18
firms for the years' 1976 and 1986. These then are regressed against insider ownership,
utilizing a quadratic regression model (Tobin's Q's are regressed against insider ownership
and insider ownership squared). This resulted in their finding strong evidence for the
existence of a curvilinear relation between insider ownership of equity and firm value.
Guarding against outliers biasing their results, the authors removed the observations for
firms had had Q's greater than 6.0. The following control variables are also included:

1. Financial leverage (debt market value / replacement cost)

2. Research and development expenditures (R&D expense / replacement cost)
3. Advertising intensity (advertising expense / replacement cost)

4. Replacement cost

With one exception, all of the control variables are found to be significant (the replacement
cost coefficient was not found to be significant for the 1976 sample).

Chen, Hexter and Hu (1993) tested the relationship between management ownership
and firm value, using the same dependent and independent variables utilized by McConnell
and Servaes (1990). They obtain Tobin's Q's from the National Bureau of Economic
Research Inc for 1976, 1980 and 1984, for large US corporations whose net annual sales
are greater than the smallest Fortune 500 firm in the year of analysis (they felt this selection
would better guarantee a higher degree of homogeneity in their sample than McConnell and
Servaes, who had included a number of small firms). Additionally, they also include a
cubic regression form (insider ownership, insider ownership squared, and insider
ownership cubed). Similar to McConnell and Servaes, the quadratic form performed the
best for the 1976 data. However, the cubic form outperformed the linear and quadratic
forms for the 1980 and 1984 sample years. They also include an additional control
variable, firm size (replacement cost/ total assets)

These last four studies did not deal directly with regulated utilities. However, their
statistical procedures are deemed to be more rigorous than Helmuth's methodology. More
importantly, the procedures utilized by these four can be adapted to the empirical concerns
of my study.

CHAPTER 3

PUBLIC UTILITY REGULATORY REPORTS REVIEW

This chapter discusses which of the approaches are supported by the public utility
commissions through their rate case rulings. While looking for rate case determinations
dealing with an adjusted allowed rate of return for a wholly owned subsidiary, I have
determined that two separate categories exist, those cases where "double leverage" is
specifically stated and cases where consolidated tax judgments are imposed.

Wm

Reviewing rate regulation cases from 1952 through 1992, twenty cases specifically
mention "double leverage" in their determinations. All of these cases involve companies
that are part of a holding company organization. The cases cover the period 1973 through
1984. They involve thirteen different state commissions. Twelve state commissions return
rulings that favor utilizing the double leverage approach, and are located in Alabama,
Alaska, Arkansas, Illinois, Iowa, Kansas, Maine, Montana, New York, Tennessee, Texas,
and Wisconsin. One commission, the Florida Public Service Commission ruled against
using the double leverage approach. The most active commission was the Arkansas Public
Service Commission, which had four rate case rulings favoring the use of the double
leverage approach. Of the twenty cases, eighteen involve telephone utilities, one a water
utility, and one case involves a holding company in the electric and gas transmission
industries.

The Florida Commission case that rejected the double leverage approach provided the
following rationale:

19

20
Re Florida Telephone Corporation, Docket No. 780912-TP, Order No. 9551,
September 17, 1980; 39 PUR 4th; 452-472:

The direct double leverage approach incorporates the debt and preferred stock costs
of Florida Telephone, but it assumes that the source of funds made available to the
operating company should be a major consideration in determining its rate of return.
We disagree. The important consideration should be the alternative investments
available to the stockholder(s). The rate of return analysis should be accomplished
within that context, without reliance on the source of a particular stockholder's funds.
Only in this manner can a capital structure truly reflect investment options in the
marketplace.

A "one stockholder" subsidiary is subject to capital attraction standards in the same
manner as other companies. If the parent company is not satisfied that the return on
investment of a subsidiary is commensurate with the return to be earned elsewhere on
investments with similar risks, it simply will decline to purchase additional equity from
the subsidiary. The subsidiary would either have to reduce service or sell shares in the
open market and is, therefore, met with market challenges similar to those of other
companies.

Of the cases ruling in favor of the utilization of the double leverage approach, the
following is a representative example of the most often cited reasoning of the parent's
stockholders eanring an unreasonably high return:

Re Hawkeye State Telephone Company, Docket Nos. U-270, U-392, September 7,
1973 (Iowa State Commission); 2 PUR 4th; 166-187:

Under our system of regulation a company is allowed to earn on the shareholders'
equity investment in utility plant a return equal to what it currently costs the company to
attract that equity in the stock markets. Where a company's equity is held directly by
shareholders this is accomplished by allowing a company a return on its equity
investment in utility plant equal to the market cost of equity.

However, when there is a parent-subsidiary relationship, where there is debt issued
by both parent and subsidiary, there exists a form of financial pyrarniding known as
"double leverage." If we were to ignore this double leverage and allow the subsidiary a
renrm on its "apparent" equity investment in utility plant equal to the market cost of
equity, this could result in the parent's shareholders earning more on their investment in
the company (represented by the book value of each share of the parent's common
stock) than the market cost of equity.

Another often cited rationale for using the double leverage approach involves the
importance of recognizing the debt held at the parent level:

21
Re RCA Alaska Communications Inc., U-78-4, Order No. 33, April 22, 1981 (Alaska
Public Service Commission); 42 PUR 4th; 683:

There are two circumstances in which the commission might reasonably disregard a
utility's actual capital structure and adopt a hypothetical capital structure for rate-making
purposes: (1) a utility's actual capitalization is determined to be inefficient and
unreasonable, thereby producing an inflated return rate and (2) a utility is part of a
holding company system in which the utility's book capital structure and capital costs
may not be a true reflection of the system's capital costs with respect to a particular
operating company.

As a matter of regulatory policy, the commission endorses the use of the double
leverage method to anive at a capital structure to the extent that a utility's parent
company has debt in its capital structure, since in that case the utility's booked equity
capitalization ratio actually underestimates the amount of debt which is being used to
finance operations.

Part of the regulatory process includes judicial review of state commission
determinations. The objectives of this review are to ensure that the commission has not
exceeded its authority, that the determinations are supported by the hearing's evidence, and
that the orders handed down will maintain the financial integrity of the utility. At the same
time the court is obligated to look out for the interests of the public. One double leverage
determination that was overturned by judicial review involves the following case:

Re New England Telephone and Telegraph Company, Docket Nos. 80-142 et a1,
March 30, 1981 (Maine Public Utility Commission); 42 PUR 4th; 182-252:

In Re New England Teleph. & Teleg. Co. F.D. No. 2213, June 10,1977, this
commission found that since the company was part of a financially integrated holding
company it was "essential to take into account the effect of the financial relationships
between the subsidiary and the parent in determining the proper cost of capital." That
same reasoning applies here. . . In Re New England Teleph. & Teleg. Co., supra, we
found that the recognition of double leverage was essential to determining a fair rate of
return to NET. At that time, however, there existed a 14 per cent minority interest of
publicly held shares. On appeal, the law court found, in New England Teleph. &
Teleg. Co. v Maine Pub. Utilities Commission (Me Sup Jud Ct 1979) 27 PUR 4th,
390 A2d 8, that based on the evidence in that case the existence of the minority interest
precluded the use of double leverage. In 1980, American Teleph. & Teleg. Co. moved
to buy out the minority shares. The record demonstrates no reason to believe that that
buy out was not accomplished. It is clear, therefore, that any legal or evidentiary
impediment of the application of double leveraging to this case no longer exists.

Currently, as reported in the National Association of Regulatory Utility Commissioners
Annual Report for 1990, no commission utilizes the double leverage approach in the
determination of a regulated utility firm's rate of return. It is not even given as a possible

22
selection. As an alternative, in Section 5, the Special Provisions of the Income Tax Laws,
I found that many of the state regulatory agencies require a consolidated tax adjustment,
assessing the utility's customers for the taxes paid, rather than what the taxes would be if
the firm were an independent utility paying the statutory tax rate amount. The following is
a summary of the column reporting "effective consolidated federal income tax:

categories it.

A - Actual tax (cost of service includes only those taxes to be paid) 17

N - Normalized (combination of actual taxes to be paid, and the tax laws provisions) 15

E - Either 7

O - Other 8

U - Undecided _4
Total 51 3

Given the lack of recent commission activity with regard to the double leverage approach, 1
also consider the rate cases dealing with consolidated tax judgments.

WW

Reviewing the rate cases that deal with consolidated tax judgments, sixty-nine cases
pertain to utilities operating in the telecommunications industry. As before, these cases
involve companies that were part of a holding company organization. They involve thirty-
three commissions, with sixty-two cases ruling for a consolidated tax adjustment, and
seven ruling for a stand-alone tax position. Identifying the commissions by state/district,
with the number of cases located in the parentheses, the following list is provided:

 

8The total number includes the fifty states, plus the District of Columbia.

23

Commission Rulings for the Commission Rulings for the
Consolidated Tax Adjustment Stand-alone Tax Position
Alabama (1) Kansas (3) North Carolina (1) Delaware (2)
Alaska (1) Louisiana (1) North Dakota (1) Louisiana (3)
Arkansas (2) Maine (3) Pennsylvania (3) Oklahoma (1)
Colorado (1) Maryland (4) Rhode Island (2) Texas (1)
Dist of Col (1) Massachusetts (1) South Carolina (1)
Florida (2) Michigan (3) South Dakota (1)
Hawaii (1) Minnesota (5) Tennessee (5)
Illinois (1) Montana (1) Washington (1)
Indiana (2) Nebraska (2) West Virginia (6)
Iowa (2) New Jersey (1) Wisconsin (3)

The Louisiana Public Service Commission is the only reversal listed.

An in-depth search of the W to obtain each state's (plus the District
of Columbia) mcst mm consolidated tax judgment provides the following results: 9

Stand-Alone Treatment: (An affiliated utility should be treated as if it were an
independent utility, calculating its tax allowance as though it were an independent firm.)

13 rulings:
1 general decision 3 telephone utility decisions
3 water utility decisions 3 electric utility decisions
2 gas utility decisions

Consolidate-Entity Treatment: (Consolidated tax benefits realized by an subsidiary
utility are allocated at least partially back to the utility's customers, thereby reducing the
rate that the customers must pay by the amount being allocated.)

32 rulings:

l7 telephone decisions 6 electric utility decisions
5 water utility decisions 4 gas utility decisions
No Ruling: 6 agencies

 

9The most recent decision is utilized rather than the most recent telecommunications'
judgments, based on the idea that the market would view future decisions concerning
telecommunications rate cases be consistent with the most recent one, changing
expectations and therefore the value of all regulated utilities operating in that state even
though the ruling only would be directly applied to the utility named in the decision.

24

The original basis for these findings are obtained from McGilsky (1986). She reports
that through 1984, the W identifies thirty-four states plus the District
of Columbia following the consolidated-entity tax adjustment, with eight states using the
stand-alone approach, and eight states having made no tax ruling. As shown above,
through 1992, utilizing each agency's most recent ruling, thirty-two agencies are utilizing a
consolidated-entity tax adjustment, thirteen agencies have directed that the stand-alone
approach be followed, and six have taken no position. In part the movement from a
consolidated tax adjustment to the stand-alone approach could be a reaction to three Internal
Revenue Service (IRS) private letter rulings. As an example of one of the changes in
position:

Re Iowa Public Service Co. (IPS), Docket No. RPU-87-3, Iowa Utilities Board,
June 17 1988; 94 PUR 4th; 239-289:

The Board finds IPS's reasons for basing tax expense on a "stand alone" tax rate to
be persuasive. The private letter rulings cited by [PS which support IPS's position are
dispositive of the situation. . . . The Iowa Supreme Court has further stated that a
private letter ruling, whatever its strength, could be recognized by the Board. Office of
Consumer Advocate of Iowa v. Iowa Utilities Board, 419 N.W.2d 373,375 (Iowa
1988). The Board cannot ignore the IRS interpretation, even if it should disagree with
the reasoning. The Board will defer to the private letter rulings.

Secondly, the ratepayers do not pay for losses of subsidiaries and should not reap
the tax benefits associated with the losses. The benefits from the losses should accrue
to the shareholders who hear the risk.

The following are examples of tax positions for rate case decisions concerning
companies operating in the telecommunications industry:

Stand-alone tax position:

Re GTE Southwest Inc., Docket No. 5610, 15 Tex PUC Bull 1, Texas Public Utility
Commission, February 23 1989; 106 PUR 4th; 194-342:

The court (See, e. g. City of Charlottesville, 774 F.2d at 1213, 1216) has
recognized that when a consolidated group uses the tax gains of one affiliate to offset
the tax loss of another, it is in effect cashing in a future tax reduction belonging to the
loss affiliate. The result, therefore, is not a reduction in tax liability, but only a deferral
of a tax liability. The consolidated group benefits only to the extent of the time value of
money. Therefore, reducing a utility's allowable income tax expense by the full
amount of its consolidated tax saving appropriates the current tax benefit from an
unregulated affiliate without a commitment to pay back the funds in the future when that
affiliate will face a higher income tax liability. If accelerated depreciation or investment

25

tax credits created the tax loss, the normalization rules are violated by the appropriation
of the current benefit.

Consolidated tax adjustment:

Re GTE South, Inc, Case No. 90-522-T-42T, West Virginia Public Service
Commission, May 31 1991; 123 PUR 4th; 257-282:

The Commission has addressed the issue of consolidated tax savings on a fairly
consistent basis for many years. In developing its proposed tax allowance in this case,
the Staff has reflected a savings in its federal income tax calculation relating to the filing
of a consolidated tax return. However, the Administrative Law Judge (ALJ) rejected
this calculation in large measure because of the proposed IRS' regulations which would
have apparently prohibited the Commission from reflecting the tax benefits associated
with filing a consolidated tax return.

On April 30 1991, however, the IRS withdrew its proposed utility normalization
rules. Since these proposed regulations have been withdrawn, and since the existence
of these regulations served as the basis upon which the ALJ rejected Staffs proposed
consolidated tax savings adjustment, the Commission must reverse the ALJ and
correspondingly reaffirm its long-standing policy regarding the treatment of the benefits
associated with the filing of a consolidated tax return when determining a utility's
allowance for federal income tax purposes.

The Commission has held in numerous cases that the filing of a consolidated tax
return allows the consolidated group to utilize certain parent company tax losses, that
could not otherwise be utilized. If the Commission were to disregard these losses,
which are offset against taxable income of the operating companies in a consolidated tax
return, the Commission would in effect he allowing a tax allowance in excess of the
actual tax requirement Such an allowance would be unreasonable. Thus, the
Commission will not depart from the utilization of a proper portion of the parent
company tax losses in a determination of the income tax allowance for GTE.

The following is a case that takes a position similar to Ezzell, Hsu and Miles (1991),
focusing directly on the interest tax shield benefits a parent would receive from its debt held
at the parent level when filing a consolidated return:

Re The Chesapeake and Potomac Telephone Company, Formal Case No. 595, Order
No. 5623, District of Columbia Public Service Commission, January 25, 1974;
4 PUR 4th; 1-82:

Another disputed issue in this case involves the company's allowance for taxes.
Although the Bell System files a consolidated return, the Bell parent company and each
of the subsidiary Bell operating companies pay a share of the total Bell System taxes
directly to the federal government. In reporting its federal income taxes, the Bell parent
receives the entire benefit of the interest deduction on debt which it issues. The

26

subsidiary operating companies take advantage of a deduction for interest only on debt
which the subsidiaries themselves issue.

The commission's staff has computed the company's tax allowance so that its
interest deduction is based upon the debt ratio and cost of debt for the entire Bell
System . . . The staff position is based upon the premise that it is essential to compute
the company's tax allowance on the basis of the debt structure and cost of debt for the
Bell System as a whole in order to equitably distribute the tax burden among the
operating companies and to avoid substantial tax bonuses to the Bell parent The staff
maintains that the company is not an independent entity and its capital structure is
artificial except as considered as a part of the Bell System. Consequently, it is urged

that the company's equity—which is held entirely by its Bell parent—is actually
financed, at least in part, through debt issued by the parent and that tax benefits flowing
from the issuance of that debt should be made available to reduce the allowable tax

liability in fixing rates for the company.

In summary there are two key points brought out by the public utility regulatory
reports. First, rate cases dealing with wholly owned adjusted subsidiary rate of return
regulation have focused on the same basic issue posed by the financial literature, "Should
the debt of the parent be recognized when determining a 'fair' rate of return for wholly
owned regulated subsidiaries?" Based on the number of cases imposing either the double
leverage approach (19 of 20), or a consolidated tax adjustment (62 of 69) public utility
commissions have ruled that the parent's debt position should be recognized. Second, state
commissions have used both double leverage and consolidated tax adjustments. The trend
of usage has been away from double leverage, to consolidated tax adjustments.

CHAPTER 4
TOBIN'S Q DERIVATION

Tobin's Q has been selected as the descriptive dependent variable for my empirical
study. Tobin's Q has been computed, using the Lindenberg and Ross (1981) algorithm.
In particular, the following procedures are performed to obtain the calculation:

Marketlalue
W

The recorded calendar year-end common stock market values are multiplied by the
calendar year-end common stock shares outstanding.

Wk

Because of the problem of obtaining complete price quotes for preferred stock, each
firm's aggregate preferred stock market value is calculated by dividing its total annual
preferred dividends by Standard & Poors’ December corporate preferred stock yield index
average for each year.

Debt

The market value of debt is the most difficult to obtain, since much of the publicly
traded debt is done off of the floor of the exchanges, and is therefore not reported. Any
transactions concerning privately held debt is also not reported. There is also the problem
of accounting for privately held debt. With this in mind, the question becomes, "Should an
imputed market value be determined, or should it be assumed that the debt's market value
equal its book value?" Klock, Thies, and Baum (1991) investigate different ways of
computing Tobin's Q, based on book, imputed market, and market values of debt Using
100 firms over the period 1977-1983, they find the imputed market value of debt Tobin's Q

27

28
is a marked improvement over the book value of debt Tobin's Q. Following their
methodology:

1. Compustat provides book values for total debt, TD, in addition to also reporting
those portions of TDt that will mature in years' one through five (Dlt, D2t, D3,,

D4t, and D50 for any year t. TDt, is broken down as
wt 5 Dlt + th + D3t + D41 + D5; + [314,

where Dlt = the debt that will mature in year t+l
D2t = the debt that will mature in year t+2
D3t = the debt that will mature in year t+3
D41 = the debt that will mature in year t+4
D5 = the debt that will mature in year t+5
DLt = the long-term debt maturing beyond year t+5

2. Dlt is assumed to have a market value equal to its book value

3. All of the remaining debt's market value is found in the following manner:

a. In any year t, all new debt is to be issued for 20 years, with a coupon rate
equal to that respective year's Moody's December corporate bond yield average.

b. For each of the Compustat reported debt figures of D2t - D5, (D1, is
assumed to be market value) are discounted back for periods 1 through 4
respectively, using the coupon rate set in step a, and Moody's December
corporate bond yield for year t as the discount rate.

c. For the remaining debt DLt (= TDt - 2Dit), DLt is assumed to be issued in

equal amounts over the remaining 15 year maturity [D6t through D20t =
(1/15)DLd. These book values are similarly discounted back for periods 5

through 19 respectively in the same manner set up in step b.

29

W

There are three basic categories within replacement cost:

a. plant and equipment
b. inventories
c. other assets, which includes such assets as cash, marketable securities, or land.

As was utilized by Lindenberg and Ross (1981), it is assumed other assets market value
equals book value. The following approach was utilized for plant and equipment and
inventories.

Wm

Lindenberg and Ross (1981) postulate that the replacement cost of net plant and
equipment (RN Pt) would can overtime as a result of the affects of four major components:

Price level changes - increases RN Pt.
Technological change - decreases RNPt.
Real economic depreciation - decreases RNPt.

#9310!—

Investrnent in new plant - increases RNPt.

Given the investment in new plant and equipment series (10, they develop the following

relationship:
1+¢
RNP =RNP- —— +1.120 4.1
' t1[(1+8)(1+t=)) ] ' ( )
where:

0 is the base year
4), = rate of growth of capital goods prices (for this firm) for year t
8, = rate of (real) depreciation for year L

St = rate of cost-reducing technical progress for year t.
" = implies "estimated"

30
Continuing with the recursion,

l

__1:¢__

Rm" = ETC [(1+5)(1_+_e) ]I’+HNP°n(1+5)(1+o)
5:0

where I-INPO = the book value of net plant in year 0.

The estimate of (pt used is the GNP deflator for nonresidential fixed investment.

The estimate for 5‘ used is

_1_)_EPt
5‘: HNP—'_t-. 1

where DEPt = the book depreciation for year t
HNPH = historical value of net plant in year H.

No good estimate was found for BI, so it was computed endogenously in one of two ways:

(1) By using formula (4.1) along with the reported replacement cost for years’ t and
t+1,

1+¢
t+1 = RNPt {+1
(1+8t+1)(1+9t+1)

 

RNP + It+1

If it is assumed the rate of technical progress is constant for a sufficient period prior to
t, the solution for 0t +1 can be used as an estimate for the years after t The difficulty with

this method is that it uses depreciation, inflation, and investment information for only one
year, t+1, in order to determine an estimate for many years.

(2) 0 can also be calculated by solving

1 l t

RNPt = 2 1T (1+9)r t i:¢——S L+ +HN1>o (1+0) tug-1+— W-l—S

68 53
1:0 s=t+r 3:01

using reported replacement costs for a specific year I. These measurements of technical
progress at the firm level could be aggregated to the industry level.

31
Using the variables exactly as defined by Lindenberg and Ross, an industry average
rate of technological change was attempted, using the second alternative to determine an
estimate for the rate of technological change for the telecommunications industry.
Replacement cost data for 27 firms in the 4800 SIC range was obtained from the tape
produced by Columbia University, which provided replacement cost data over the 1979-
1983 time period.

Lindenberg and Ross were not explicit in their explanation for new plant and equipment
investment, therefore capital expenditure is used. Compustat calendar year data is utilized
instead of fiscal year data, in an effort to neutralize any differences in reporting the data

WW

Probably the biggest surprise was that the yearly rates of technological change for the
included firms were predominantly negative. Summarizing these technological rate
estimates:

# negative 13 19 17 18 13
# positive 3 6 7 5 6
Range:
high .056 .047 .088 .222 2.077
low -.224 -.631 -.537 -.476 -.426
median -.079 -.057 -.029 -.026 -.036

Five year average: -.0455

A yearly rate is obtained by finding the median rate since using an arithmetic mean would
place too much weight on the outlying figures. It was also deemed inappropriate to
subjectively eliminate the outliers. Since the firm rates were both positive and negative, a
geometric mean could not be used.

Looking at equation (4.1), both the depreciation rate and the rate of technological
change are located in the denominator. I therefore hypothesized that Lindenberg and
Ross's estimate for the depreciation rate might be overstated, which if true would cause the

32
rate of technological change to be understated, or in this case negative. An alternative
estimate for the depreciation rate was used:

6 _ DEPt
‘ - GNPt-l

where DEPt = the book depreciation for year t.
GNPH = historical value of gross plant in year t-l.

This change provided a less negative estimate for the rate of technological change (-.0213).
However, the resulting estimate replacement costs for the 27 firms were not found to be as
accurate an indicator for the reported values that were obtained in the initial attempt

A third attempt was made, utilizing the change in gross plant for investment in new
plant and equipment in place of capital expenditures, feeling that this might more accurately
reflect any reduction of plant, property and equipment being done within the industry from
year to year. This caused a more negative average rate of technological change (-.0831),
with the initial result estimates again showing better accuracy.

A final attempt was made, combining the two changes stated above. This rate of
technological change estimate (-.0629) was found to provide the least accurate estimates.

It should be noted that the results do not necessarily question the logic of the different
alternatives, but that the original variables provide the best fit for the reported replacement
costs over the 1979-1983 time period.

CHAPTER 5

AN ALTERNATIVE DERIVATION OF THE EZZELL, HSU AND MILES MODEL

With the public utility commissions' favoring adjustments for parent held debt as a
consolidated tax adjustment, it would be valuable to take a detailed observation of the
approach found in Ezzell, Hsu and Miles (EHM) (1991). Instead of restating the points of
this paper, an equivalent model will be presented, utilizing the more familiar Modigliani and
Miller (MM) perpetual cash flow framework.

WW

Beginning assumptions:

1.

There are m subsidiaries, each of which is regulated and wholly owned by the
parent firm, p.

. Subsidiary k, for k=1,2, . . ,m, has an investment 1k in real assets, which generates

a level perpetual before-tax unlevered cash flow stream.

. N 0 operating synergy results from combining the subsidiaries into a parent

company.

. The parent has no investments in any operating assets other than those held by their

subsidiaries.

. Each subsidiary k and its parent company p maintain constant market value leverage

ratios of Lk = Dk/VLk and LI) = Dp/VLp, respectively, where D represents the
market value of debt and VL represents levered market value of the firm, inclusive

of any adjustments to each subsidiary's cost of capital.
The net tax shield on interest is equal to the corporate tax rate, t, with the present
value of the interest tax shield being thVLk for subsidiary k, and d‘pVLp for the

parent p.

. The allocation of any parent level interest tax shield benefits will be the same as

is utilized by EHM (qk = Ik/Elk).

33

34
8. To remain consistent with the results of the EHM (1991) perpetual cash flow
model, the tax allowance (TA) which in turn sets the allowed revenue uses the
flow-through method.

MM (1958) wrote a keystone paper concerning a firm's cost of capital, its valuation,
and its capital structure. A second article, MM (1963), extended their first paper, by
including corporate taxes. Among their results, they found the value of the levered firm to
be:

VL = Vu + tB (5.1)
where VL = value of the levered firm

Vu = value of the unlevered firm

t = corporate tax rate

B = market value of the firm's debt

In a manner similar to EHM, if the firm maintains a constant debt ratio, L, making
B = LVL, then equation (5.1) becomes:

vL = vu + tLVL (5.2)

This same format can be placed into a similar equation for a holding company system
containing a parent and one subsidiary. For this holding company system, the subsidiary
possesses all of the operating assets, while being wholly owned by the parent. To start,
for simplification, only the capital structure of the parent includes any debt The resulting

equation for the value of this system, Vsys, is:
Vsys = Vu(sub) + a‘pVLp (5 '3)
where Vsys = levered value of the system
Vu(sub) = unlevered value of the subsidiary
1..p = debt ratio of the parent
VLp = levered value of the parent

t = corporate tax rate

35
For the initial case V is also the levered value of the parent, VLp’ The unlevered

sys

value of the subsidiary is also the subsidiary's value of its equity, S1. Placing these
relationships into equation (5.3):

Solving for VLp’ equation (5.4) becomes:

81
va = 11E (5.5)

To illustrate the previous discussion:

Subsidiary's operating assets (11) = 1000

Subsidiary's debt ratio (L1) = .00 (100% equity)
Parent's debt ratio (Lp) = .30

Corporate tax rate (t) = 40%

If the subsidiary is allowed the risk-adjusted market rate of return required by investors:

Sl[ind] = I] = 100010

From equation (5.5):

51de 1000
VLDlindl = Tifpl = 1-(. )(.3) = “36-36

The difference between va[ind] and Sl[ind]’ 136.36, is the wealth created at the parent
level, the present value of the parent's interest tax shield, PVTSpfind].

 

10lhe " [ind] " denotes that the subsidiary and parent are being treated as independent
entities (or the independent firm approach).

36

WW

Placing these results into a market value balance sheet for the subsidiary, the parent,

and a consolidated statement provides:

 

S l .1. Recent :2 1'1 1
Assets 1000.00 1000.00 1000.00
PVTSpfind] . .13616 416,335
Total 1000.00 1136.36 1136.36
Debt 340.91 340.91
Equity 1000.00 12144 121—45
Total 1000.00 1136.36 1136.36

If there is a desire that there would be no wealth creation at the parent level,
VLpladjl = I], the resultant subsidiary equity value can be determined by an alternative

version of equation (5 .5):

The cash flows from the subsidiary's operating assets haven't changed, nor has the
subsidiary's operating or financial risk changed. The value of S1 is lowered by reducing

the "allowed" rate of return for the subsidiary, when the present value of the parents
interest tax shield has been taken into consideration. The market value of the operating
assets have 1191; changed, so there must be an additional amount recognized along with the
value of S1, for the market value balance sheet to "balance".

37

W

This revised balance sheet is now:

 

 

 

S l ‘1‘ E 2 1'1 1
Assets 1000.00 880.00 1000.00
Total 1000.00 1000.00 1000.00
Debt 300.00 300.00
Equity 880.00 700.00 700.00
PVTSpladJ-l .1201!)

Total 1000.00 1000.00 1000.00

Notice that this balancing amount is the "adjusted" present value of the parent's interest tax
shield,

The wealth creation at the parent level is no longer ignored, but is used to balance the
adjusted market value of the parent's investment in the subsidiary to the market value of the
subsidiary's operating assets.

The adjusted rate of return for the subsidiary can be determined using a cash flow
analysis. Given that the subsidiary's risk-adjusted market rate of return can be estimated
by the same measures used by the independent firm approach, one method for determining
the cash flows would be to start with the subsidiary's tax allowance as a foundation.
Equation (5.6) defines the subsidiary's stand-alone, independent firm approach tax
allowance.

38

. t
TAlfind] = [‘1 ' 111111] [TY I (56)
where TAI [ind] = tax allowance for the firm (independent firm approach)
r1 = capitalization rate for the firm [ledl + (1-L1)ksl]

weighted cost of debt for the firm [led1]
value of the operating assets for the firm

11
I1
t

corporate tax rate

Continuing with the previous example by providing the risk-adjusted market return for the
subsidiary's equity if viewed as independent of its parent, and the cost of the parent's debt,

ksl = its“, = 15% (L1: 0)

From equation (5 .6):
TAlfind] = [rl-i1][11][T-t_—t ] = [.15—0111000][%] = 100

W

The cash flows for both the subsidiary and the parent are: 11
S l . 1'

EBIT 250.00 0.00

-_hrt 4001) (21.02)
EBT 250.00 (34.09)
-_Ix (1.111.131) 13.63
NI 150.00 (20.45)
1.15111 1500)
Netp 129.55

 

11Depreciation for each period is used to equally offset periodic wear and tear on the
subsidiary's operating assets.

39
The independent firm approach rates of return are:

 

ksul = RI?” = 15.00% (=WACC1,sinceL1=0)
129.55 , _ 20.45+129.55 _
ksp = 795.75 = 16.29%, WACCp _ 1136.36 _ 13.20%

Using the same basic method to determine the cash flows presented above, an after
adjustment tax allowance equation can be derived from the following equations:

1. The tax base for the independent firm with no debt at the subsidiary level is
TB1 = al)(ksu1)’

with the tax allowance being

TAr = T3161}? ) = (Il)(ksul)(1+t )

2. If it is desired to acknowledge the present value of the parent's interest tax shield,
this can be done by removing this value from the original investment, 11.

TB1 = (11 ' tLpVLp[adj])(ksu1)

with the adjusted tax allowance being

TAlladjl = (11 ' [LDVLPladjlkalle-t ) (57)
From equation (5.7), the subsidiary's adjusted tax allowance is:

.4
TAM“ = (1000-120)(.15)(3 ) = 88

MW

40

The revised cash flows for both the subsidiary and parent are now:

5 l . 1.
EBIT 220.00
-_Int AIME)
EBT 220.00
-_Ix (8.8.110)
NI 132.00
:thll

Netp

12mm
0.00

(121211)

(30.00)
12,le

(18.00)
132.90

114.00

The adjusted rates of return for both the subsidiary and the parent are:

132

k * = m = 13.20 % (= WACC1*, since L1 =0)

sul

k, _114

Sp m = 16.29 % ;

18+114
WACCp* = 1000 = 13.20%

 

Notice that (l) the subsidiary's adjusted rate of retum equals the parent's WACCp, and
(2) the parent return rates, ksp* and WACC *, are not affected by the adjustment. This
results from the cash flows in the numerator for both ksp* and WACCp* being adjusted
lower, while the value of the respective denominators, S1) and VLp, have also been adjusted

downward with the removal of the wealth creation at the parent level.

Onc Subsidiag, Q1212; at Bcth mc Scbsigh'gy agd Parent Levels

Subsidiary's operating assets (11)
Subsidiary's debt ratio (L1)
Parent's debt ratio (Lp)

Corporate tax rate (t)

1000
= .50
.30

=40%

41
Using the risk-adjusted market rates of return for both subsidiaries:

From equation (5.5):

s .
. .._llml _ 500 _
VHarmer-"1411, ' 1—(. )(.3) ' 56818

The difference between VLp[ind] and Sl[ind]’ 68.18, is the value of the wealth created at the
parent level, the present value of the parent's interest tax shield, PVTSpfind].

WW

Placing these results into a market value balance sheet for both the subsidiary, the
parent, along with a consolidated statement provides:

 

Assets 1000.00 500.00 1000.00
PVTSplind] . .6818 _68._18
Total 1000.00 568.18 1068.18
Debtl 500.00 500.00
Debtp 170.45 170.45
Equity 518202 221.13 127.13
Total 1000.00 568.18 1068.18

If there is a desire that there would be no wealth creation at the parent level can be
accomplished using the following:

 

V .
_ Lplmdl
VLptadj] - [vafin (”14,111 I11 (5.8)

42

W

This would result in the following adjusted market value balance sheets:

 

S_ub1 Parent Consolidated

Assets 1000.00 468.09 1000.00
PVTSptadj] —— 453-32 ——

Total 1000.00 531.91 1000.00
Debtl 468.09 468.09
Debtp 159.57 159.57
Equity 468.09 372.34 372.34
PVTSptadj] 43-32 - ——

Total 1000.00 531.91 1000.00

The risk-adjusted market return for the subsidiary's equity required by individual investors,
as well as the cost of the parent's debt are,

 

WACC] = ksul [l-tLl] = .15 [1-(.4)(.5)] = 12%
WACCl-kdlLla-t)
k.“ = kdp = 10%

From equation (5.6):

TAmnd] = [r1 -il][11][lL_t ] = [.14- .05][1000][:—2— ] = 60

W

43

The cash flows for both the subsidiary and the parent are:

Subsidim Parent
EBIT 200.00 0.00
-_lnt (10.00) (11.01)
EBT 150.00 (17.05)
J}: (612.00) $.82
N I 90.00 (10.23)
1.1811 .2099
Netp 79.77

The rates of return for both the subsidiary and the parent are:

ks1 = 33,95% 18.00%; WACC] = 3001006?” = 12.00%

 

_ 79.77 _ , _ 10.23+79.77
ksp " T73 7.7 ‘ 20'0””, WACCp - 568.18

 

= 15.84%

If equation (5.7) is modified to incorporate the debt canied at the subsidiary level,

TA .- [(I- v -)WACC -LI (1-t)][t l (59)
1121411 - 1% Lptadu 1 11km (‘17) °

TAlladj] = [[1000-63.83][.12] - (.5)(1000)(.1)(.6)] [.4/.6] = 54.89

Adjustrdfiashflm
The cash flows for both the subsidiary and the parent are now:

Subsidim Parent

EBIT 187.23 3.19 12
-.m (£11m) (11.99)
EBT 137.23 (12.77)
-_Tx M32) .111
N1 82.34 (10.23)
11311 .2111!)
Netp 79.77

The resulting rates of return for both the subsidiary and the parent are:

 

 

ksl’“ = 832% = 16.47%; WACC1* = 30%;?)2'34 = 11.23 %
79.77 , _ 10.23+79.77 _
ksp* = W = 20.06 % , WACCp* — 568.18 — 15.84 %

As previously shown when the interest tax shield affect is recognized and the subsidiary
cost of equity is lowered, the parent's respective returns remain unaffected.

 

12Here EBIT represents the parent's interest income provided by the subsidiary's debt
portion of the present value of the parent's interest tax shield, PVTsz

EBITp=L1(PVTSp[adJ-])kd1 = (.5)(63.83)(.1) = 3.19

45

I 51.1.. 111:] 151.1. I 1
511121 311122

Subsidiary operating assets (1k) = 1000 2000

Subsidiary's debt ratio (Lk) = .00 .00

Parent's debt ratio (Lp) = .30

Corporate tax rate (t) = 40%

Using the risk-adjusted market rates of return for both subsidiaries:

Sklindl = 1k (Sl[ind] = 1000; 82[ind] = 2000)
A modified equation (5.5) is:

zslqind] _ 3000

VLprindl = 1—tr.p " 788— = 340909

Not surprisingly, once again the difference between VLpfind] and ZSkfindP 409.09, is both

the present value of the parent's interest tax shield and the value of the parent's wealth
creation.

 

Placing these results into a market value balance sheet for the subsidiaries, the parent,

and a consolidated statement provides:

 

 

5.11111 £11122 Balm mustered
Assets 1000.00 2000.00 3000.00 3000.00
PVTSpfind] . $22.92 A0292
Total 1000.00 2000.00 3409.09 3409.09
Debt 1022.73 1022.73
Equity 19181.99 2001200 2336.36 m

Total 1000.00 2000.00 3409.09 3409.09

Ifthe deSire IS for VLp[adj] = 3000,1henZSk = VLp[adj](l'tLp) = 3W(.88) = 2640

Each subsidiary's equity can then be determined, using the EHM allocation q's, qk =

Skladjl ‘

Slladjl

1k
[2}; ] VLpladjla'd‘p)

- [% ][2640] = 880;

WWW

The revised market value balance sheet is now:

Assets
PVTSpladjl

Total

Debt
Equity

$11121

1000.00

 

1000.00

880.00

Total

As was true for the original single subsidiary case, the sum of the "balancing" amounts

1000.00

5.0.122

2000.00

 

2000.00

1760.00
240,1!)

2000.00

Parent

2640.00
16.0.00

3000.00

900.00
2100.00

 

3000.00

(5.10)

2000

3000.00

3000.00

900.00
2100.00

 

3000.00

(PVTSp[adj]*) equals the present value of the adjusted parent's interest tax shield:

_1k_
21k

Conselidatrd

Each subsidiary's portion of PVTSp[adj] can be determined by the same allocation process

used above:

47

1k
PVTSpladmk)" = (fl: )(PVTSpladjl)

1000 ,
PVTSptadjlu)* = m )(360) = 120’

2000
Pvrspladmf = 3000 )(360) = 240

The cash flow analysis for the two subsidiary case, given the additional independent firm
approach market return information provides the following:

ksl = ksul = 15% (L1 = 0) kdp = 10%
1‘82 = kSllZ = 20 % (IQ = 0)

. t .4

. t .4
TA2[ind] = [r2-12][12][1—_t ] = [.20-0][2000][3 ] = 266.67

en 11 ashFl ws
The cash flows for the subsidiaries and the parent are:
5.11121 5.11122 Eaten].

EBIT 250.00 666.67 0.00
;I_nt .111!» .1020) um
EBT 250.00 666.67 (102.27)
ix (Ml) (266m) fllfll
NIk 150.00 400.00 (61.36)
121211,, 5.10.90
Netp 488.64

48
The rates of return for the subsidiaries and the parent are:

150 400
ksul = m = 15.00% ; ksu2 = m = 20.00%

(kSllk = WACCk, Since Lk = O, for k = 1,2)

488.64 , _ 61.36+488.64 _
ksp - 2386136 = 20.48%, WACCp _ 34®09 _ 16.13%

 

The same allocation process can be used to modify equation (5.7), the tax allowance after
adjustment, for the two subsidiary example:

11: t

TAuadj] = (31% )(3000-360)(.15)(% ) = 88.00

TA2[adj] = (§% )(3000-360)(.20)(% ) = 234.67
mm

The adjusted cash flows for the subsidiaries and the parent are now:

5.1.1121 5.1.1.122 Parent
EBIT 220.00 586.67 0.00
:10]. M) .101!» (20.00)
EBT 220.00 586.67 (90.00)
:18 W (2% 16m
NIk 132.00 352.00 (54.00)
1:2le 4.8.4.120

Netp 430.00

49
The resulting rates of return for the subsidiaries and the parent are:

132 352
ksul* = m = 13.20%; ksu2* = m = 17.60%
(ksuk* = WACCk“, since Lk = 0, for k = 1,2)

430 , _ 54+430 _
ks; _ m _ 20.48 %, WACCp* _ 70—00‘ _ 16.13 %

Again, like the single subsidiary case the parent rates of return are not affected by the
adjustments to the subsidiary return rates.

It should be recognized that a mathematical relationship does exist between the
subsidiary adjusted returns, ksul’“ and k * and the parent's weighted average cost of

su2 ’
capital, WACCP’“:

WACCp* = Z{ [g—‘I‘k ][ksk*]} = [-1- ][.132] + I} ][.176] = 16.13%

Th D lIevraeA roach

Utilizing the double leverage equations found in EHM (1991), the following results are
obtained when the following four equations are solved simultaneously:

 

 

 

WACC dkmw/xccp ‘31
ks = 2ksk[ind](l'Lk)VLk[DB]

WIDE] 2(1-Lk)va[DB]
WACCprDBl = kSWIDBl(l"LP)

WACCk B 11‘
VLleBl = Wiggl—

50
In this example,

Double Leverage Market Value Balance Sheet

 

 

51.11:, 811122 Parent classmates
Assets 1000.00 2000.00 2640.00 3000.00
PVTSpladjl 3.60.00 ___.
Total 1000.00 2000.00 3000.00 3000.00
Debt 900.00 900.00
Equity 1056.00 1584.00 2100.00 2100.00
Total 1056.00 1944.00 3000.00 3000.00
(Difference + 56.00 - 56.00)
Double Leverage Approach Adjusted Cash Flows
8.11121 $11122 Parent

EBIT 264.00 528.00 0.00

:1111 40.111) 4.0.181) (20.1KB

EBT 264.00 528.00 (90.00)

:12: (101.80) (2.11.20) 16.90

NIk 158.40 316.80 (54.00)

LEAH], 4112.0

N etp 421.20
The resulting rates of return for the subsidiaries and the parent are:

158.40 316.80
ksulfDB] = m- = 15.84%; 14mm] = m— = 15.84 %

51

421.20 , _ 54.00+421.20 _

 

As shown by the market value balance sheet no wealth creation occurs for the system
(VLp = 21k). However, cross subsidization does occur at the subsidiary level. Since sub
#1 is allowed a retum (15.84%) that is greater than its original risk adjusted market return
(15.00%), the result is a positive $56 value in excess of its true market value. Sub #2 is
allowed a return (15 .84%) that is less than its risk adjusted market return (20.00%),
resulting in a negative $56 value below its true market value. The net incomes shown were
obtained using the allowed equity return of 15.84 % times the book value of each
subsidiary. It follows that the subsidiary cash flows and returns are not affected by how
the parent's interest tax shield value is allocated to the subsidiaries. Additionally, the
parent's debt and equity values are obtained from

Sp z = [I'Lpnzslddb-adj] +Lptmlkn

Since neither of these values incorporates the allocation of the parent's interest tax shield
value, the parent cash flows and returns are also not affected. What is affected by the
allocation of PVTSp, is the market value of each subsidiary, and the amount of cross

subsidization that occurs between subsidiaries.

The following table shows the range and affect of different allocation pairings for the
preceding example:

q1 q2 NPVI NPV2 NPVSys
1.0 0.0 416 (416) 0
0.0 1.0 56 (56) 0

-.1556 1.1556 0 0 0

S41121 S11112
Subsidiary operating assets (1k) — 1000 2000
Subsidiary's debt ratio (Lk) = .5 0 40
Parent's debt ratio (LP) = .30
Corporate tax rate (t) = 40%

Given each subsidiary's constant debt ratio, Lk, then for the independent case:

Sk[ind] = “'1‘ka

The same modified equation (5.5) that was previously utilized provides:
= —— = 1931.82

Once again the difference between VLp[ind] and 28mm], 231.82, is the present value of the

parent's interest tax shield and the value of the parent's wealth creation.

53

W901

Placing these results into a market value balance sheet for the subsidiaries, the parent,

and a consolidated statement provides:

 

 

 

$11121 $11122 Eaters Conselidated

Assets 1000.00 2000.00 1700.00 3000.00
PVTSpfind] . 231182 231.82

Total 1000.00 2000.00 1931.82 3231.82
Deth 500.00 500.00
Debt2 800.00 800.00
Debtp 579.55 579.55
59th .3811!) 1.20000 13.5221 135221

Total 1000.00 2000.00 1931.82 3231.82
VLp[adj] is determined by modifying equation (5.8) for two subsidiaries:

vL ad. = [ VLPfin‘“ ][21k] (512)
pl 11 VLp[ind]+szIk

_ 1931.82 _
VLpradll - [ 9 1.8 +500+8 1 [300°] " 1793-25

Equation (5 . 10) is modified to include debt at the subsidiary level, obtaining Sk[adj]

(l-Lk)lk

2044911,] [Eskladfl] (5'13)

Skladjl =

500
1200
82M] = [W ] [1578.06] = 1113.92

Wheel

This results in the following adjusted market value balance sheets:

 

 

3.1.1121 £11122 Parent W

Assets 1000.00 2000.00 1578.06 3000.00
PVTSPIadjl - 45-192 ———-

Total 1000.00 2000.00 1793.25 3000.00
Debt1 464.13 464.13
Debt2 742.62 742.62
Debtp 537.99 537.99
Equity 464.13 1113.93 1255.26 1255.26
PVTSptadjl" 4114 44454 —— _—

Total 1000.00 2000.00 1793.25 3000.00
As before,

1000

PVTSpladj](2)* = (g-ggg )(21519) = 143.45

To assist in obtaining the cash flow analysis, the independent firm approach weighted
average cost of capital (W ACCk) for both subsidiaries can be determined by:

WACCk = ksukU-th)

WACCl .15(.80) = 12.00%; WACC2 = .20(.84) = 16.80%

55

Additionally, each subsidiary's capitalization rate, rk, and weighted cost of debt, ik, are:

r1 = .12+(.4)(.1)(.5)

.14;

.05 ;

11 = (.1)(.5)

Again, from equation (5.6):

TAmnd] = [.140-.05][1000][%]

Wm

The cash flows for the subsidiaries and the parent are:

31.1121
EBIT 200.00
-_Int (5.018))
EBT 150.00
-_Tx (60.00)
le 90.00
1.2le
Netp

The rates of return for the subsidiaries and the parent are:

ksl = fieliofi = 18.00%;

ik = kdkLk

r2 = .168 +(.4)(.1)(.4)

12 = (.1)(.4)

60

192

5.11.122
560.00
180.1!»

480.00
(122.00)

288.00

.184

Parent

0.00
(51.25)

(57.95)
23.15.

(34.77)

313m
343.23

30.00+90.00

1000 = 12.00 %

56
288 48.00+288.00

 

 

ks2 = m = 24.00%; WACC2 = 20% = 16.80%
343.23 , _ 34.77+343.23 _
ksp = 1357.27 = 25.38 %, WACCp _ 1931.82 _ 19.57 %

The adjusted tax allowance equation is obtained from modifying equation (5.9) for two
subsidiaries:

I
TAkIadj] = {Ink'd-pVLpradj]](§i: )(VVACClt)'LkItkdk(1'0} {Cit—[)1

TAlladj] = {[3000-(.4)<.3)(1793.25)] [19% ](.120)-(.5)(1000)(,1)(,6)} {% }
= 54.26

TA2[adJ-] = {[30°0'(-4)(-3X1793-25)][%&O% ](.168)-(.4)(2000)(.1)(.6)} {% }
= 175.93

Adjrlstedfiashflm

The adjusted cash flows for the subsidiaries and the parent are now:

$11121 $11112 Eaten
EBIT 185.65 519.83 9.33 13
J11: (512.112) 1801!» (5.129)
EBT 135.65 439.83 (44.47)
-_1‘x (5320) (111.23) 11.12
le 81.39 263.90 (26.68)
1281., 34.5.22
Netp 318.61

 

I
13 1313er =Zfi— Lk(PVTSp)kdk = (1/3)(.5)(215.19)(.1)+(2/3)(.4)(215.19)(.1)
k
= 9.33

57
The resulting rates of return for the subsidiaries and the parent are:

81.39 30.00+8 1.39

 

 

ksl* = W = 16.28%; WACC1* = 1000 = 11.14%

182* = %3-5%9 = 21.99 %; WACC2* = 4803358390 = 15.60%

ksp* = 13158816 = 25.38 %; WACCp* = 321278; 3.1361 = 19.57 %
:1 r :11 . E ..

I
If the allocation method is expanded beyond the EHM method (qk = fi- ), the
k

result is Figure 1. This figure is based on the two subsidiary case just covered, where debt
is held at both the subsidiary and parent levels. The range for qk has been expanded to

-.2 < ql < 1; 0 < q2 < 1.2. These allocation ranges include the double leverage solution
when NPV"‘k = NPVsys
displays the adjusted WACC k pairs for allocation pairs satisfying qu = 1. These adjusted
WACCk pairs are found by: (1) selecting a q] along the bottom horizontal axis, (2) moving
vertically to find WACC*1 on the NPV*1 = 0 solution line, (3) continuing vertically to find
WACC*2 on the NPV*2 = 0 solution line, and (4) finding the corresponding q2 value. For
example selecting q1 = 0, WACC*1 = .1200 (= unadjusted WACCI), with

WACC*2 = .1500 for q2 = 1. Also included in Figure 1 are the double leverage solutions.

= 0, which occurs when ql = -.1720; ‘12 = 1.1720. Figure l

The double leverage solutions illustrate that selecting an allocation pair affects the amount
of cross subsidization, not the double leverage adjusted WACC pairs.

If the allocation pair does not satisfy qu = 1, Figure 1 cannot be utilized. Not only
does this allocation pair affect NPVSys, but also the adjusted WACCk pairs.

Table 1 shows these affects when the first subsidiary allocation is held constant (q1 =
.3333), and the second subsidiary allocation is varied (0 < qz < 1). When q2 < .6667, it
causes NPVsys > 0, and a higher adjusted WACCZ. The q2 selection raises the market
equity value of the second subsidiary (higher allowed retum, greater market value). This in
turn increases the present value of the parent's interest tax shield (PVTSp). The higher
PVTSp causes a larger adjustment for the first subsidiary, resulting in a lower adjusted
WACCI. The opposite is true when ‘12 > .6667.

58

FIGURE 1

EXPANDED ADJUSTED WACC SOLUTIONS
(Two Subsidiary Case)

This figure displays the adjusted return rates for two subsidiaries, over the allocation
ranges of -.2 < ‘11 < 1.0 ; 0 < q2 < 1.2. To determine an adjusted WACC,C pair, select a q]

on the bottom horizontal axis, then move vertically up the figure to find the corresponding
adjusted WACCI, adjusted WACCZ, and q2. These pairs of adjusted WACCk and qk

satisfy the constraint S q = 1, and result in N PV1* = NPVZ“ = NPVsys = 0. (Subsidiary
parameters: Invstl = $1000, Invstz = $2000; WACCI = .120, WACC2 = .168; L1 = 50%,
L2 = 40%, Lp = 30%)

 

      

 

 

 

N '1 Q 02 02 N, to L0 <r co (\l .- o_
v-' v- v- o o o o' o' o' o' o' o' o
018 0.18
0.17 0.17
016 L 016
0.15 _ 0.15
8
< 0.14 0.14
3
Q 013 _ 0.13
2
< 0.12 0.12
0.11 0.11
0.10 \ 0.10
0.09 0.09
0.08 1 1 1 1 4 1 r 1 1 1 1 _ 0.08
N v- o. ". “1 0". V; In_ to ts m a: Q
0' o' o o o o o o o' o' o' o' v-

ALLmATKN DECMAL [ HJTI'W q1 ; TCPQ2 ]

 

O WACC(#1) I WAOC(#2)
. #1 DB SDLUIION #2 DB SDLU'HCN

 

9
5

£36.68
£98.63
834.80
843.0
885
83.:
33.8
33.3
3.6:. 5
cm. 5::
SNYRF
2:. .2

Aa>ov>az

Nphm0v§w
NON0.N0h.w
0_.hN.mhh.F
00Nh.005€
030005...
0Vh0.N00._.
0980—0;
00 5.000...
FOP0€v0€
0N¢T000€
0h00050...
0000.000._.

3.?

000—..0
00N_..0

00h0.00N
09;. P FN
0N00.0_.N
00v0érm
000 _..0 FN
000NOFN
hmvmfipm
«000.0FN
Nvmm. FNN
wkwodmm
hh00€NN
N000.0NN

3095

1 $00.35
t 80055

rehoéhoé
00N0.000._.
vvchfimo...
00V000P...
Ffimwdw P...
vumfimmw...
¢0F0.vmp._.
000NNVF€
000F.00_..F
mmwwdnwé
050008....
0000.00N...

Em

00N0.m0v
000560...
033460?
0mm~é0¢
000F600
0N0000v
«050.00%
0000.00V
05070004
_.\._.0.N0¢
0.00000?
NNNN.N0¢

Em

A$00.00
e$00.0v
$00.00

P m._m<._.

0000070
005 F0 _..0
v0000...0
N0000...0
0.00070
00:50.30
mvmmmwd
0vn0070
NO0N0...0
000v0...0
0h..00...0
0000090

_ .13ch
$.03;

33
a:
A S

8 n 2562 x w Ea w o €6.28 1 £2

a; u e. .20 mzo_5._0m $0055 8533....

0500070
0FONOP0
#00930
0500070
memmwd
00050.20
0.00070
0N000F.0
NNVNOT0
NON¢0w0
F200 v.0
0000070

5.09:5

000F:.0
«cm—.220
mhfizwd
vwfizpd
«amp—#0
30:70
Nmmwzd
vpmpzd
0.12.70
0h0_.:.0
000:...0
0000:.0

8.0055

0000.0

000.N
000...

0;
0.0
0.0
5.0
5000.0
0.0
0.0
V0
00
N0
_..0
0.0

3:6

Eu

$.62:
38>...
”2225630

 

EHM (1991) presented a formula for deriving a subsidiary's adjusted WACC. In the
MM discounting context, this formula is:

 

1-
WACCk’“ = WACCk (17$:— ) (5.14)
W
2111,,
where LW 2 21k

When Lk = 0 (no debt at the subsidiary level), LW = 0, which reduces equation (5.14)

down to,

WACCk* = WACCk(1-tLp) (5.14a)

For the single and two subsidiary examples with no debt at the subsidiary level, equation
(5. 14a) provides the same results shown in section In (with Lk = 0, WACCk“ = ksuk“):

WACC1* .15(.88) = 13.20%

WACC2* .20(.88) = 17.60%

The equivalent results using my alternative derivation, found on page 49, are the same.
For the two subsidiary example with debt at the subsidiary level (L1 = .5; L2 = .4),

= (.5)(1000)+(.4)(2000)

Lw 1000+2000 = '4333

Using equation (5.14):

 

1- .4 .3 .88
WACC1* = .120(1_. )E. 14) ) = .120 (.9—48 = 11.14%
_ 1-(.4)(.3) _ .88 _

61
The results of my derivation, found on page 57, are the same.

Patterson (1993) proved that the adjustment term in equation (4.15) can be defined in a
more intuitive manner,14

1- 21k
1'11wa ' 21k+PVTSpfindl

 

firm approach. This relationship can be shown here to be numerically correct for all of the

l is the present value of the parent's interest tax shield for the independent
preceding examples:

No debt at the subsidiary level, single subsidiary example:

 

1‘ = 88 , 21k = 1000 = 88

No debt at the subsidiary level, two subsidiary example:

 

 

1‘ 88 Elk 3000 88
l-tLp ' ’ Zlk+PVTSpfind1 5409-59 '
Debt at the subsidiary and parent levels, two subsidiary example:
1'53 £11. 3000
= .92827 ; = = .92827
4wa Elk+PVTSpfindl 3237782

Reviewing these equations, an even better interpretation of this adjustment term is:

1' = Vsys|adj|
"“pr Vsyslindl

where Vsys[adj] = value of the system that is desired (no parent level wealth creation)
Vsys[ind] = value of the system, independent case (with parent level wealth

creation)

 

1.4This same relationship is also true for the Ezzell, Hsu, and Miles (1991) adjustment,
provrded the approprrate PVTSpfind] (= ZpstLpVLp[ind]) value is utilized.

62

C . El 1111 lEllllE' . 111 11

The differences between the present value of the parent's interest tax shield, PVTSp,

using the MM discounting framework and the discounting technique used by EHM is the
following:

MM PVTSplindllMM] = “12012011410041”
EHM PVTSplindllEHM] = ZpstLpVLplindllEm

where: t = corporate tax rate
Lp = market debt ratio of the parent
VLp = levered value of the parent
V . = 2(1-Lk)lk
LplmdllMMl l-tL.p
2(1-Lk)lk
VLptindlrEHMl = 17pm
Ik = investment in subsidiary k
_ 9‘22
ZP — 1+l‘f
1tdp = parent's cost of debt
rf = risk-free interest rate
st = (1+ksw)/ksw
kSW = parent's unlevered cost of equity, based on the subsidiaries'

levered costs of equity:

_ stk( 1’11)va
SW 2(l-Lk)VLk

 

ksk = levered cost of equity for subsidiary k
Lk

‘01

market debt ratio for subsidiary k

levered value of subsidiary k

 

. 15The [ind] denotes that the values shown here do not recognize the parent's interest tax
shreld, and therefore represent the "independent-fum" values. Later the [adj] identifier depicts
that the values have been "adjusted" to remove any wealth creation at the parent level caused by
the parent's interest tax shield.

63

The EHM discounting methodology is more severe than that of MM. With EHM, the
interest tax shield is discounted at the risk-free rate, rf, for the first period, then at the

parent's unlevered cost of equity, ksw, from the second period on out to infinity. The MM

discounting technique discounts each period's interest tax shield at the parent's cost of
debt, kdp' This difference in discounting results in:

PVTSPNM] > PVTsplEHM]

The EHM discounting methodology also affects the impact of debt on each subsidiary's
weighted average cost of capital, WACCk:

MM: WACCkMfl = lama-111.)

where: kuk = unlevered cost of equity for subsidiary k
z .. .33
k - l+rf
kdk = cost of debt for subsidiary k
1+1:uk
a... = “1:...—
The net affect is:

WACCprMM] < WACCPEHM]

A third difference caused by the different discounting procedures is the affect on the

adjusted weighted average cost of capital, that eliminates any parent level wealth creation.

F P tte (1993) th ' ad' ' WACC‘: )
rom a I'SOl’l , 6 com aratrve [1811116111 113110118 _— are:
P J 6‘1 WACCk

64

W _ Ik‘qk‘LpVLpladllm

W ___ Ik‘quvstl‘pVLpladjrreml
WA km 11‘

where: qk - fractional share of PVTSp for subsidiary k

 

 

 

VL -
v . = mud] ] 21
Mad” [vafind]+zr., 1k [ 1‘]

Numerically, using the example for two subsidiaries with debt at both the subsidiary
and parent levels used in EHM (1991):

5.11121 3.11122
Subsidiary operating assets (1k) = 1000 2000
Subsidiary's market debt ratio (Lk) = .50 .40
Subsidiary's cost of debt (kdk) = 10% 10%
Subsidiary's unlevered cost of equity (kuk) = 15 % 20%
Parent's market debt ratio (Lp) = .30
Parent's cost of debt (kdp) = 10%
Corporate tax rate (t) = 40%
Risk-free rate (rf) = 10%
= 231.82
= 99.84
WACC/2mm = .20[1-(.4)(.4)] = .168000
WACCkaM] = kuk(1-ZkRskLk)
WACC 1[EHM] = .15[l-(.036363)(7.666667)(.5)] = .129091
WACszHM] = .20[1-(.036363)(6.000000)(.4)] = .182545

65

Eggmfi =‘k'qk‘1bVLptadl1m

 

 

 

Eggwfi _ Il'qld‘pvlarladjlflml _ 1000-(1/3)<.4>(.3><1793.25)
WA 1[MM] _ 11 - 1000
= .928270
( Ik WACC1* WACC2*
"he“ q, = 23’ we,” = m—

W = Ik'qkszswLpVLptadilrElnvn

 

w = Il'qlszswLpVLpladjlrEmvrr

= 1000-(ll3)(.036363)(5.084753)(.3)(1741.87)

 

1000
= .967793
The resulting WACCk* are:
WACC1[MM]* = .120(.928270) = .111392
WACC2[MM]* = .168(.928270) = .155949
WACC1[EI-M]* = .129091(.967793) = .124933
WACC2[EHM]* = .182545(.967793) = .176666

Focusing on the EHM methodology, as previously stated the parent's interest tax shield
is discounted at the risk-free rate, rf, for the first period, then at the parent's unlevered cost

of equity, ksw, from the second period on out to infinity. Given that the debt ratios are

markct value ratios, it would seem that the same discounting technique should be used
when finding the present value of the parent's debt. Since this discounting technique has
also been included when finding the weighted average cost of capital for both subsidiaries

66

and the parent, the market values for each subsidiary's debt should also be discounted in
the same manner. The point being made is that for the market value debt ratios Lk and LI)

to be accurate, along with remaining constant, the costs of debt for the subsidiaries and the

parent would have to equal:
k... k.
kdk = 0”“?ka- ); kclp = (1+‘r)(1+k‘:w )

WACC *
The modified EHM equations for PVTSp, WACCk, and W051: now mirror image the

MM equations since the EHM components of ZkRuk and zpst respectively within the
EHM equations reduce down to t when km and kdp are equal to the equations shown

immediately above. Not surprisingly, the numerical results for both the EHM and MM
methodologies for these variables are now equal.

Given the above results, what has been gained? One of the assumptions of the NM
perpetual cash flow model is that debt is riskless, and therefore is discounted at the risk-

free rate of interest The modified EHM perpetual cash flow model now recognizes the
business risk of the subsidiaries, setting kdk at a rate greater than the risk-free rate. Given

that the parent's unlevered cost of equity, ksw, is sensitive to changes in the subsidiaries'

financial risk, the business risk of the parent includes a component for the subsidiaries'
financial risk, with kdp being set at an appropriate rate that recognizes both the business and

financial risk of the subsidiaries.

CHAPTER 6

COMPARATIVE ANALYSIS OF THE ALTERNATIVE RETURN
ADJUSTMENT PROCEDURES

This chapter compares the various methods of return determination discussed in earlier
chapters. Chapter 2 defines five alternative procedures for determining a regulated wholly
owned subsidiary's rate of return. Chapter 3 discusses current public utility rate regulation
practices which consists of using consolidated tax adjustments when the commissions
deem an adjustment is necessary. This chapter places the five alternative methods
discussed in Chapter 2 into the perpetual cash flow model outlined in Chapter 5. In
addition, six consolidated tax adjustment methods are tested to determine if they result in a
net present value of zero for both the subsidiary and parent company levels.

In order to study the effect of consolidated tax adjustments, these adjustment
procedures have to be identified. McGilsky (1986) studies the aspects of the different
methods for handling an affiliated utility's federal income tax allowance being used by
public utility commissions. Included in her study are different consolidated tax adjustment
equations for obtaining a regulated affiliated subsidiary's tax allowance that public utility
commissions have utilized. From her study, I have identified seven alternative
consolidated tax adjustment methods along with her stand-alone (independent firm)
approach to be tested. Where necessary, McGilsky's notation has been modified to remain
consistent with the notation used in Chapter 5. In addition to these eight, I have included
three additional adjustment methods in the test, the double leverage approach, the EHM
approach, and the Beedles approach.

67

68

TBk = [11; ' ik] 1k

where: TBk = subsidiary k's tax base
rk = kadk + (I‘Lk)ksk (capitalization rate for subsidiary k)
Lk = market value debt ratio for subsidiary k
kdk = cost of debt for subsidiary k
ksk = risk related (independent firm) cost of equity for subsidiary k
ik = Llrkdk (weighted cost of debt for subsidiary k)
1k = invested amount in subsidiary k

TA=TBk[-li_t-]

where: TAk

tax allowance for subsidiary k
t = corporate statutory tax rate

IIIQ Narragansett EIQQIDE Company hdmhgdn

T31: = [’k'ik] It
TSs = [Imp-ind] t

where: TSs = tax savings of the corporate system
Intp_md = corporate parent's interest expense on a stand-alone basis

 

TBk

[mat-Ta] [1%. 1

16Re Cities Service Gas Company, Opinion No. 396, Docket No. G-l8799 (Federal
Power Commission 1963); 49 PUR 3d; 229

17Re The Narragansett Electric Company, Docket No. 1288 (Rhode Island Public
Utilities Commission 1978); 23 PUR 4th; 516

TAk

 

69
WW“
TBk = [rk ‘ is] 1k

where: iS = Lskds (weighted cost of debt for holding company system)

_ Z[Lk+(1-Lk)Lp] 1k
3 " 2:1k

 

(market value debt ratio for the system)

kds = cost of debt for the holding company system

TAlt = T131: [1% 1

Note: The Fitzpatrick method is mathematically the same as the Southwestern Bell
Telephone Company method, when an adjustment away from the stand-alone
approach is deemed to be necessary, and the cost of debt is the same for all of the
entities within the holding company system.

 

T31. = [rk-ik]Ik-[Ikip(1'Lk)P1

where: p = percentage of the parent's ownership in subsidiary k's equity

“‘1; = TBk[1—t-t- 1

 

18Re Southwestern Bell Telephone Company, Docket No. 60,800—U (Kansas State
Corporation Commission 1960); 34 PUR 3d; 257

19Re New England Telephone & Telegraph Company v. Maine Public Utilities
Commission, 390 A2d 8 (Me. Sup. Jud. Ct. 1978); 27 PUR 4th; 1

70

where: ta 2 effective tax rate

CTL = actual net consolidated federal income tax liability
PTL = total positive taxable income of the corporate system

te

WWW“

TB .
_ _ k|1nd|
TBk — TBlt[ind] zmflind] ] Imp-ind

where: TBkfind] 2 tax base of subsidiary k on a stand-alone basis

TAk = TBk[-1't—t 1

 

20Re Newton Water Company, Docket No. 790911 (Connecticut Public Utilities
Commission 1980)

21Re Brockton Edison Company v. Department of Public Utilities, Mass, 400 N.E.2d
838, (Sup. Jud. Ct. of Mass. 1980)

71

Il‘ .oH-éOI‘U‘JO' 1' _l' .IIDI'J.1' molten 1:_‘. E 1:22

TB]: = [rsk ' isk] Ik

 

where: rsk = Lsk kds+ (l-Lsk) ks-sk
[Lk+(1'Lk)Lp11k
LSk = I
k

 

_ ksk( I'll-p)‘kads( 1'0
l-L

s-slt —
ltds = cost of debt for the corporate system
islt = leckds
TAk = TBk [ft]
mainstream”

TB .
_ _ k|1nd|
TAlt - TAklind] [2113mm] ] [Inb-inc11'

W

I
TAk = { [Zlk'a’vap] (2)—11k )(WACCk)'Lkakdk(l't) } {Cit—()1

where: va = levered market value of the parent

 

22Re Continental Telephone Company of Maine, EC. No. 2183,
C. No. 440 (Maine Public Utilities Commission 1977); 18 PUR 4th; 636

23Re City Service Gas Company, Opinion No. 396, Docket No. G-l8799 (Federal

Power Commission 1963); 49 PUR 3d; 229

72
W

This method was developed in the following manner:

1. I first tested to see if the Beedles method of partitioning the parent's debt to the sub-
sidiaries would obtain the same returns as those found by the zero NPV returns
identified by the EHM methodology for the case when debt is canied at both the
subsidiary and the parent levels. The adjusted subsidiary debt ratio is then:

L at: ___ WACCk[EHM]'ks-sys[ind]
1‘ ktllt(l't)‘ks-sys[1ml]

where: WACCkaM] zero N PV weighted average cost of capital for

subsidiary k
risk adjusted market return for the equity of the

ks-sys[ind]
holding company system

I tested to see if the result accounted for all of the parent's debt by:

Dp-k = Lk*(11t)'Dk

where: Dp-k
Dk

the portion of the parent's debt carried by subsidiary k
market of debt for subsidiary k

This determined that not all of the parent's debt was accounted for (ZDp_k < Dp),
implying that a larger adjustment to WACCk was needed to adequately partition
all parent debt down to the subsidiaries.

2. An adjustment factor, d, was searched for that satisfied the following constraints:

(1) dis the same for both subsidiaries (the same property exhibited by the EHM

adjustment factor)
(2) EDp-k = Dp

73
3. The Beedles adjusted weighted average cost of capital for subsidiary k is then
found by:

4. Each subsidiary's adjusted cost of equity is then determined by:

WACCk[Beedles]*'kdkLk(l't)
ksk[Beedles] = l'Lk

4. The tax base (net income) was found by:
TBk = ksk[Beedles](l'Lk)Ik

5. The tax allowance was found by:
t
TAlt = T31; [17 1

It should be noted that the steps taken to obtain a testable method follows the general intent
of Beedles, but should not be viewed as the only way of representing his method. My
method does fall within these general guidelines outlined by Beedles:

1. Utilize the component costs for the consolidated holding company system.
2. The parent's debt is allocated to the subsidiaries.
3. A lower adjusted debt ratio is assigned to the subsidiary with greater operating risk.

I have taken additional measures to determine the adjusted cost of equity, ksk[Beedles]’ from
the adjusted weighted average cost of capital, WACCk[Beedles]*r which is not defined by

Beedles in his article.

Well

The adjustment process of this approach is based on the parent's weighted average cost
of capital, and as such, an explicit tax allowance equation has not been provided.

74

WW
Beginning assumptions:

1. There are m subsidiaries, each of which is regulated and wholly owned by the
parent firm, p.

2. Subsidiary k, for k=1,2, . . ,m, has an investment 1k in real assets, which generates
a level perpetual before-tax unlevered cash flow stream.

3. No operating synergy results from combining the subsidiaries into a parent
company.

4. The parent has no investments in any operating assets other than those held by their
subsidiaries.

5. Each subsidiary k and its parent company p maintain constant market value leverage
ratios of Lk = DkNLk and LD = Dp/VLp, respectively, where D represents the

market value of debt and VL represents levered market value of the firm, inclusive

of any adjustments to each subsidiary's cost of capital.

6. The net tax shield on interest is equal to the corporate tax rate, t, with the present
value of the interest tax shield being thVLk for subsidiary k, and tLpVLp for the
parent p.

7. The allocation of any parent level interest tax shield will be the same as was utilized
by EHM (qk = Ik/Zlk).

8. To remain consistent with the results of the EHM (1991) perpetual cash flow
model, the tax allowance (TA) which in turn sets the allowed revenue using the
flow-through method.

Using a two subsidiary simulation model, these eleven different methods are evaluated
for their affect on the allowed returns for the two subsidiaries, the parent, and the total
system. The initial parameters of the model are:

811121 811112
Subsidiary operating assets (1k) = 1000 1000
Subsidiary's debt ratio (Lk) = .00 .00
Subsidiary's unlevered cost of equity (ksuk) = . 15 . 15
Parent's debt ratio (Lp) = .30
Cost of debt (kd) = .10

Corporate tax rate (t) = 40%

75
Using the Narragansett method as an example, the following steps are used to obtain the
results for each method found in Table 2:

1. Determine the tax allowance for each subsidiary.
From the set-up parameters:
fk = kSUk = .15

Imp-ind

1.p[):Ik + PVTSp]kdp = Lp { 2:1k (1 + [113% D } kdp

.30 {2000(1+ [8—8 1)} .10 = 68.18

[.15-0]1000 = 150

TBk = [rk ‘ ik] Ik

[68.18] .4 = 27.27

TSS = [In[p_ind]t

TB
Tsk = TSS[XT1:1( ] = 27.27 [£78 ] = 13.64

TAk = [(TBkt)-Tsk][11—t ] = [(150).4 - 13.64][% ] = Lil

 

2. Find each subsidiary's net income.

y—a

NIk = TAk [% = 77.27 [33 1 = 115.9

3. Determine VLk , Sk, and Dk (subsidiary market values for itself, its equity, and its
debt)

Intk = 11.111911. = (1000)(0)(.10) = 0

(When Lk > 0, Int must be found simultaneously with the rest of the calculations in
step 3.)

76

V _ le+1ntk(1-t) _ 115.91+0 72 73
Lk _ WACCk_ind _ T :

 

ll
\1

8k = (l-Lk)VLk = (1)772.73 = 772.73

Dk = VLk" Sk = Q
. Determine ksk and WACCk_adj.
N11 _ 115.90

p—A
y—n
L11
\0

ksk=m-W=-

The consolidated tax adjustments each scale back the tax allowance, and therefore
net income. This in turn causes a reduction in VLk. The market value of the
subsidiary assets however do not lose value, since they can be "unbundl " from

the parent/subsidiary relationship and receive the unrestricted net income.
Therefore, the denominator for ksk is in terms of the unrestricted equity asset value.

This same logic is also utilized when obtaining WACCk-adj°

NIk+Imk(1't) 115.90

1k ‘1 ‘-

 

u—s
1...:
(I!
\O

. Determine va’ Sp, and DI) (market values for the parent).

281: 1545.45
va = m = T = 1756.20

 

 

(PVTSp = 1756.20-1545.45 = 210.73)

Sp = (l-Lp)VLp = (.7) 1756.20 = 1229.34

DD = VL -s = 526.86

77
. Determine ksp and WACCp.

Intp = Dp kdp = (526.86) .10 = 52.69

N1p = 23le + (2{ [212—1 ][kadk]}[1>vrsp] - Intp)(l—t)

= 115.91 + 115.91 + [0 - 52.69].6 = 200.21

NIp _ 200.21

ksp = [Elk-Dk]'Dp - 1473.14 = ___—-'1359

 

NIp+Intp(l-t) 200 21+3l 61
WACC = = ° ' = .
P zflk'Dk] 2000

ll;

. Determine ks_sys and WACCSyS.

 

 

= 1116+[ZDk+Dp][1-t] = 200.21+[52.69].6

 

 

WACCsys 21k 2000 = .1159
. Determine Tobin's sts
market value Z[D +S ]+D +S
. , _ sys _ k k p p _ 2634.30
Tob1ns QSYS " replacement valuesys _ 21k — m

1%:

. Determine N PVk, and NPVSys.
(the interest tax shield allocation being, qk = Ik/Zlk)

Nka = va+qk1>vrsp 1, = 772.73+(.5)210.73 -1000 = - 121.90

 

NPVsys = 23151ka = 243.80

-
___.—___—
___—

78

S . E l
I . . l S .
The initial parameters:
511121 5.0.122
Subsidiary operating assets (1k) = 1000 1000
Subsidiary's debt ratio (Lk) = .00 .00
Subsidiary's unlevered cost of equity (ksuk) = .15 .15
Parent's debt ratio (LP) = .30
Cost of debt (kd) = . 10
Corporate tax rate (t) = 40%

Table 2 provides the results of the first scenario for all of the methods. The results are
prioritized by descending N PVSys. The most surprising of the results is that only the

EHM, double leverage, and Beedles methods reduce the stand-alone (independent firm)
wealth creation down to the desired NPVsys = 0, and Tobin's Q = 1. All of the other

adjustment methods overcompensate for the stand-alone method's parent level wealth
creation (NPV = 272.72), driving all of their NPVsys < 0 and Tobin's Q < 1. The

stand-alone method using the combined parent/subsidiary interest expense (Std-alone, com)

sys

turns out to have a NPV that is equal and opposite in sign to the stand-alone, independent
method (Stand-alone). All of the adjustment methods provide a downward adjustment for
the SUbSidial'y realms (ksk-adj < ksk-ind; WACCk-adj < WACCk-ind)’ WhiCh iS meant to be
the result of eliminating the parent level wealth creation. The EHM, double leverage, and
Beedles methods make no downward adjustments to the parent's returns (ksp-adj = ksp-ind;
WACCp_adj = WACCp-ind)- This is the desired result for the following reasons:

1. As stated by Jones and O'Donnell (1978), the state commission's regulatory
authority should affect only the subsidiary, thereby not exceeding its authority by
imposing regulation on the parent.

2. The desired result of each the tax adjustments are to erase the parent level wealth
creation and leave the value of the holding company system equal to the
initial investment (Vsys = 21k). The payments to the security holders, discounted

by the market risk-adjusted returns determine this value. They should therefore be

reduced to provide Vsys = 21k.

79
The other seven methods not only cause lower subsidiary returns, they also reduce the

parent returns (ksp-adj < k WACCp-adj < WACC

sp-ind; p-ind)'

mm
Change in parameters:24 12 = 2000

Table 3 records the results for each procedure when the investment in subsidiary #2 is
increased. Equity values for the second subsidiary, Equity (2), for all of the methods
increase 100%, with parent values increasing by 50%. The N PVsys of the stand-alone
(independent firm) method shows an increase by 50%, and its Tobin's Q has increased.
The EHM, double leverage, and Beedles methods NPVsys and Tobin's Q all remain at zero

and one respectively. Of the remaining seven methods, their NPVsys all increase in

negative value by 50%, and their Tobin's Q have decreased from the values displayed by
Table 2. The subsidiary debt levels remained at zero, and are therefore labeled "not
applicable". None of the return data show any change from Table 2.

Table 8 is a rate of change comparison of Table 3 to Table 2. The expected result
would be that an increase in the investment aside from the appropriate increases in
subsidiary #2 values and parent values, should have no effect. This is true for the EHM,
double leverage and Beedles approaches. For the other approaches, based on the 50%
increase in investment, the stand-alone process shows a 50% increase in the parent level
wealth creation. The actual taxes paid, Southwestern Bell, New England, Newton,

Brockton, Narragansett, and stand-alone, combined interest all show 50% increases in the
negative value of N PVsys.

11.15 .

Change in parameters: 12
k

1000 (original value)

su2 = '20
Table 4 displays the affect of having equal investment in both subsidiaries at: = 1000),

with the second subsidiary having a higher unlevered cost of equity, ksuz = .20, with no

 

24Changes depict those changes between the current and preceding scenario.

80
other changes in the remaining parameters. Rather than interpreting these figures, Table 9
depicts the affects of an increase in business risk for subsidiary #2.

Table 9 provides the rate of change between Tables' 2 and 4. The desired affect is that
all of the subsidiary #2 and parent returns would reflect the increase in operating risk, with
all of the values remaining unchanged. This is the result for the stand-alone, EHM and
Beedles approaches. Knowing that the double leverage approach ignores differences in
subsidiary risk, although there is no wealth created at the parent level, there is cross
subsidization between subsidiaries. Of the other approaches, although it could not be
called cross subsidization, the actual taxes paid, Newton, Brockton, and Narragansett
methods all show increases in the equity values and returns of both subsidiaries. Their
negative NPVsys have become less negative, but are still negative, and their Tobin's Q's
have increased, but are still less than one. Southwestern Bell, New England, and stand-
alone, combined interest methods fare slightly better by not showing "sympathy" increases
in subsidiary #1 values. They also show less negative NPVsys, and Tobin's Q's less than
one.

mm
Change in parameters: 12 = 2000
Table 5 provides the raw data results.

Table 10 presents the relative change in data between Tables' 3 and 5. The stand-alone,
EHM, double leverage, and Beedles methods showed no change in performance than was
commented on for the third scenario. The relative movements of the remaining methods
outlined in the third scenario are the same. The fact remains that none of these remaining
methods are providing NPVsys = O or Tobin's Q's = 1, nor have they maintained any
consistency (relative change of all data remaining zero) in handling changes in the
parameters.

EN 5 .

Change in parameters: Market value debt ratiok = 50%, for k = 1,2

Table 6 provides the raw data results.

81

Table 11 presents the relative change in data between Tables' 5 and 6. The stand-alone
method displays a predictable 50% lower equity position of both subsidiaries, which
causes the parent level debt position to decrease by 50%. This in turn causes the parent
level wealth creation to decrease by 50%. The increase in financial risk at the subsidiary
level causes the stand-alone return data for all move in a predictable manner.

At first glance, it would appear that all of the methods have now deviated from the
stand-alone (independent firm) method rate change figures. The rate of change value
figures for the EHM method are slightly lower, for two reasons:

1. With the equity positions in both subsidiaries being reduced by the subsidiaries
now having debt positions, the parent's investment is now lower, causing the
parent's debt position to be lower, which lowers the value of the parent's interest
tax shield.

2. In Table 4, when no debt was held at the subsidiary level, all of the subsidiary's
portion of PVTSp_adj was balanced against its equity position in order to have the

investment side of the market value balance sheet equal the market value of each

subsidiary's assets. In Table 5, PVTSp_adj is being balanced against both the

subsidiary's debt and equity positions.

The EHM NPVsys remains at 0, and its Tobin's Q at 1.0.

With the addition of debt at the subsidiary level, the Beedles method no longer displays
any symmetry with either the stand-alone (independent firm) or the EHM methods. The
decrease in Equity (1) is greater than the decrease to Equity (2), with the decrease in the
parent values falling in between those of the subsidiary equities, diSplaying the problem of
cross subsidization of subsidiary #2 by subsidiary #1. The final result of the Beedles

method's adjustment process provides a loss of market value to the system relative to its
asset market values (Totalsys = 2,869.45 < 3,000 = Elk), resulting in NPVsys < 0.

The double leverage method shows a higher degree of cross subsidization of subsidiary
#2 by subsidiary #1. Here, the subsidiary with the lower operating risk is receiving too
high a return, with the opposite happening for the subsidiary with the higher operating risk

One of the characteristics of double leverage is that it does not allow a parent level wealth
creation to exist Appropriately, its NPVsys equals zero.

82

Of the remaining methods, all show a form of cross subsidization between
subsidiaries. The subsidiary with the higher operating risk is overcompensated, while the

subsidiary with the lower operating risk is being undercompensated. All of their respective
NPVsys remain less than zero.

5.15 .

Change in parameters: Market value debt ratioz = 40%

Table 7 provides the raw data results.

Table 12 displays the relative change between Tables' 6 and 7. All of the movement
shown by the stand-alone method is predictable, with its NPVsys and Tobin's Q both
increasing. This is caused by the increase of the parent level wealth creation.

The EHM method results differ somewhat from what is expected, but the unexpected
movement in the subsidiary #1 values can be explained. The increase in subsidiary #28
equity position in turn shows an increase in the parent's figures, caused by the parent's

increased investment position in subsidiary #2. Part of the increase is supported with debt,
which causes an increase in PVTSp, which decreases the part subsidiary #l's assets being

supported by its own debt and equity of 0.84%. The desired values for Totalsys, NPVsys
and Tobin's Q are retained.

Increasing the equity position of subsidiary #2 results in the Beedles method showing a
very slight increase in its cross subsidization of subsidiary #2 by subsidiary #1. The

higher parent level interest tax shield also results in a slight increase in Total but NPVsy

sys’ s

and Tobin's Q remain less than zero and one respectively.

The double leverage method continues to show the expected cross subsidization
problem. NPVsys and Tobin's Q remain at the desired values of zero and one respectively.

Of the remaining methods, the largest degree of cross subsidization is displayed by the
Southwestern Bell method, where an increase in subsidiary #2's equity position caused a

6.38% increase in the value of subsidiary #1. All of the other cross subsidization is very
slight (< 0.6%). All of these seven methods have NPVsys less than zero.

83

EE II"

All of the concerns cited in the literature and the public utility reports basically boils
down to two desired characteristics that a method should possess when determining a "fair"
rate of return for a public utility that is wholly owned by a parent company:

1. That the allowed rate of return compensate investors of capital for the riskiness of
their investments in a subsidiary public utility.

2. That any unfair business organizational advantages that are found to exist be
removed to the degree that this advantage no longer exists.

El'Elil '11]!

l. The Stand-alone (independent firm) method:

This method follows the guidelines of the first constraint, but has problems
with the second. When the parent has a debt position that supports its investment
position in a subsidiary, the interest tax shield that arises from this debt provides an
increase in value of the holding company system. Ignoring this interest tax shield
effect passes this value to the parent's shareholders at the expense of the subsidiary
public utility's customers.

2. The EHM method:

This method was developed to provide returns that fall within the requirements
of both constraints. It reduces the returns provided by the stand-alone (independent
firm) method so as to eliminate any increase in value caused by the interest tax
shield of debt held at the parent level.

3. The Double Leverage method:
This method has no problem with the second constraint, but has difficulties

with the first constraint By recognizing the parent's debt, it removes any interest
tax shield affect it might have. As long as all subsidiaries have equal systematic

84
risk, the double leverage method will provide the same return results as the EHM
method (see Tables' 2 and 3). If the subsidiaries have unequal systematic risk this
method violates the first constraint, allowing a cross subsidization condition to exist
(see Tables' 4 through 7).

. The Beedles method:

This method at times has difficulties with both constraints. As long as no debt
is carried at the subsidiary level, the Beedles method provides the same return
results as the EHM method (see Tables' 2 through 5). When debt is carried at the
subsidiary level the Beedles method violates both constraints, allowing cross
subsidization to occur, along with overcompensating for the parent level interest tax
shield affect which results in the value of the holding company system to fall below
its fair market value (see Tables' 6 and 7).

. The Fitzpatrick method:

When the capital structure of a subsidiary falls within what the author poses to
be "normal", this method follows the same procedures as the stand-alone
(independent firm) method, and therefore violates the second constraint for the
reason stated above. When the subsidiary's capital structure is considered to be
"artificial", and the consolidated capital ratios for the entire holding company are
used, it also allows cross subsidization and overcompensation for the interest tax
shield affect to occur, thereby violating both constraints (see SW Bell; Tables' 2
through 7).

. The "other" presented methods:

All of the other methods fail to fall within the guidelines of both constraints.
They show the affects of cross subsidization and overcompensation, violating both
constraints (see Act Tx Pd, SW Bell, New Eng, Newton, Brockton, Narragansett,
and Std-alone, com; Tables' 2 through 7)

85

88... $3... $3... $3... $3... $8.2 $82 $3... $3... 5868.868

8...... $3... $3... $3... $3... $3.2 $3.2 $3... $3... =888§z

5...... $3... $3... $3... $3... $3.2 $3.2 $3... $3... 8:85

88... $8. .. $8. .. $8. .. $8. .. $8.... $8.3. $8. .. $8. .. 8.82

58... $8.2 $8.2 $8.2 $8.2 $8.... $8.3. $8.2 $8.2 8m :62

.88... $8.2 $8.2 $8.2 $8.2 $8.... $8.3. $8.2 $82 .826

3.8... $8.2 $8.2 $8.2 $8.2 $8.2 $8.2 $8.2 $8.2 P. x. .3.

88. . $8.2 $8.2 $8.2 $8.2 $8.2 $8.... $8.2 $82 8.88

88.. $8.2 $8.2 $8.2 $82 $8.... $8.... $8.2 $8.2 >3 288

88.. $8.2 $8.2 $8.2 $8.2 $8.2 $8.... $8.2 $82 :5

38... $8.2 $8.2 $8.2 $8.2 $8.... $8.... $8.2 $8.2 88.88
.230 «.58. 88.00%. .3835 8.09;. .5005... .23... .3: .3... .3...
.22.... 8.3... 8.8%. a .8 .m .28 8...: 8... 888 8... 88.823...
.832 8.8... 8.88.. 8.83 32.8 2...... 8... 2.2. 8... .88282
.833 8.8... 8.8.»... 8.83 3...... 2.2. 8... 2.2.. 8... 828....
.888. 8.8... 3.38.. 2.33 5.... .m 3. .8 8... u... .8 8... 8.82
.8. .2. 2.2.... 2.2.4... 3.2.3 2.2a 8.8.. 8... 8.8.. 8... 8m :62
.8. .2. 2.2.... 2.2%. 3.23 2.2a 8.8.. 8... 8.8.. 8... .826
... .2... .38.. 3.2.... 8.8.. 3.8.». 8.8.. 8... 3.8.. 8... 8 x. .o<
8... 8.8...w 8.8.... 8.8.. 8.3m 8.8.. 8... 888 8... 886m
8... 8...8.~ 883.. 8.8.. 8.3.. 8.8.. 8... 8.8.. 8... >3 6.88
8... 8.8...m 8.8.... 8.8.. 8.3m 8.8.. 8... 8.8.. 8... :5
28B 2.23 .83.. 8. .8 2.4.5 8.8.... 8... 8.8... . 8... 82.6.8.8
.2352 .233... .3238 .358 .85.. .8235 .838 .5255 .558

.95.. __a .6. .. 8.2m :8. An .8...

$8.8 u 3...... 2... u .28... .88.. an a.

$8... u .38 2... u .8... 8.8.... u .8. 2 8m 9 .m.

$8... u $8 2... u .6... 8.8.... u E. a. .8 Au ...

60:9. BOD o:_u> use—Q2 ”9:30.. 8.20.333 “9:95.802... EEG... “9.0.5.52
”floaggwn.
Qm<zmow ..<Ez_

szwEfiSOO .rzwihwfigdx‘ odiw x<._. hmwmw...2_

N mams.

86

0000.0
P0h0.0
.0500
0000.0
.0000
.0000
000.0
0000..
0000..

.802:
$00.:
$00.:
$00.:
$000.
$000.
$00.0P
$0N.0F
$008;
$00.0—

«00¢...
$00.:
$00.:
$00.:
8000'
$00.0.
$00.0.
$00.0.
$005:
$00.0.

$00.2
$00.0.
$003:
$00€P
$00.3.
$006—
$00.0—
$00.0.
$00.0.
$00.0.

$02 3
$00. 3
$00. :
$00. pp
8000'
8000'
8000'
$00.09
$00.0.
$00.0.

0000. F
«.00...—

.3.30 8......

.88...
.28...
.28...
..... .8.
.223
.223
.28.
8...
8...
8...
8.8..

$238.2
o2.00.00

$00.0
$00.0

60:0. 500 o:_0> .00—.02

62.00.0—

.2383;

«0000.0
00600.0
00600.0
00000.0
20.2020
NNKNHN
.N.vv0.N
00000.0
00000.0
00000.0
00.00v.0

.233...
u .38

n .38
n 3....

$004:

.382.

v0.0w0.w
p0.¢v0.p
«0.30..
00.000;
00.000;
00.000...
00000.0
0000.0
00.00F.N
00.00..N
00000.0

.3238

020055200 0.220059. Dam—:0 x<h hwmmmhz

$3... $3... $8.2
$3... $3... $3.2
$3... $3... $3.2
$8... $8... $8.3.
$8.2 $8.2 $8.3.
$8.2 $82 $8....
$8.2 $8.2 $8.2
$8.2 $8.2 $8....
$8.2 $8.2 $8.2
$8.2 $8.2 $8.2
$8.2 $8.2 $8....
.3022. ...oo<2. .83..
3.2. 8...... 8.8....
8.8. 2.2.. 3.33..
8.8. 2...... 32......
8.8.. 8.8.. 8.83..
2.2.. R28 88....
2.2.. 3.28 8.8..
8.8.. .88 322..
8.8.. 8.8.. 8.8...
8.8.. 8.8.. 8.8...
8.8.. 8.8 882..
2.8.... 8.83 8.8...m
.358 .328 .3258
.2258... ..
2... u .23..
2... u .53..
2... u .6...
”2:22 0232.330
0.5.8.8 9288

0 w._m<._.

$00.0—

.8...

00.0
00.0
00.0
00.0
00.0
00.0
00.0
00.0
00.0
00.0
00.0

3.500

00000.0
00.000;

”gcgmoz: 520n—

300.0.

.3...

00.005
0N.Nhh
ohNhh
no. F02
00.000
00.000
v0.000
00.000
00.000
00.000

00.000...

.5258

u .3.
u 8.

$3. . . 58.8868
$0.... . . 23.30832
$3. . . 8:88
$8. .. 8.362
$8.2 .8 252
$8.2 8.5m
$8.2 P. x. ...<
$82 8.88
$82 >3 388
$82 20
$8.2 8......85m
..3..
8... 58.86365
8... .8888:
8... 82.85
8... 8.82
8... .8 262
8... .826
8... 8 x. .3.
8... 888
8... >3 288
8... 2m.
8... 8%83w
5.8..
523.0 .33 an .36.
2.28 An 3.
2 36 An .N.
.2 .8 Au E
”38.8...
”0.20520...

87

.88... $88. $88 . $88. $8.. . $88.... $888. $88. $8. .. 5888.838

8...... $8.3 $8.3 $3.... $88. $8.... $8.... $8.... $88. 288882

8.8... $8.3 $8.... $2.... $88. $8.... $8.... $3.... $88. 8828

...88... $.83 $8.3 $88.... $888. $8... $8... $88.... $888. 8.82

288... $8.3 $8.3 $8... $88. $88... $88... $8... $88. .8 82

8.88... $8.3 $8.3 $8... $88. $88... $88... $8... $88. .826

88.. $88. $88. $8... $888. $888. $888. $8... $888. .... ... .8

88.. $88. $88. $8... $888. $88. $8...... $8... $.88. 8.88

8...... $88. $88. $88. $88. $88. $8.8. $8.8. $8.8. >3 8.88

88.. $88. $88. $8... $.88. $88. $8...... $8... $88. 3m

88 ... $8.8. $8.8. $8.8. $88. $8...... $88. $888 $88. 2888.8
.330 8.2.8. 28.00.25 .3082. .3835 2.0925 .888... .8... .88... ..8...
.8888. 8. .2. . 88888.. 8888 88. . .8 8.8. 8... 8.8. 8... 58.88.28
.82.. 8888.. 8.88.. 8.88 88.8 8.88 8... 8.88 8... 288882
.82.. 8888.. 8.88.. 8.88 88.8 8.88 8... 8.88 8... 8.288
.88.... 8. .88.. 888.. 8888 8888 888.8 8... 888... 8... 8.82
.888: 8.28.. 88.8.. 888 8.888 8.8.. 8... 8.88 8... 88 82
.888: 8.28.. 88.8.. 888 8888 8.88 8... 8.88 8... .828
8.. 8.88 8.8.... 8.88 88.88 8.88 8... 8.88 8... .... ... .3.
8... 8.8.8 888.. 8.8.. 8.88 8.88 8... 8.88 8... 8.88
8... 8888 8.8.... 8.8.. 8.88 88.8. 8... 288.. 8... >3 88..
.8... 8888 888.. 8.8.. 8.88 8.88 8... 8.88 8... 28
28.8 28.88 8.88.. 88...... 28.8 8.8.... 8... 8.8.... 8... 88.88....
.3382 .2388. .3238 .338 .328 .3258 .358 .5238 ......8

.88. .... .... .. 8.88 .8. A" .8.

$8.88 u .38 ..8... u .88... .88.. A... .3

$8... u .38 8.... u .38... 8.8.... u .3. 8.. 88 Au .3

$8... u ...mo ...... u ...... 8.8.... u .... .. ....w A... ...

“00:0. 300 o=_0> .0352 ”9:30.. $20.9.-me ”0.00.5002: «Cohan. ”0.05232
”22050.0..—

n¥m<2mownxtlh
020953200 ng§< odfw x<._. hmmmm._.2_

v wAm<h

88

V0000
pmN00
pmN00
5000.0
0&v00
05v00
0000;
0000;
00004
0000.F

$00.vw
«wmméw
gmmép
$0F.mw
$00.0—
0000...:
$00.09
$0—.0F
$v0.mp
$0P0w

$00.:
$006?
$00€p
$070—
$00.0,
$00.09
{ommdp
$09.0F
«www.mp
$0p0p

$00Kw
«0000'
$m0dp
$0h.0w
$v00w
$30.0?
XENON
o\o0¢0m
$00.0N
$0v0m

$004”:
$00.0p
$00.09
$vm0p
$0089
$00K«
3558*
$00.59
$v0.mp
$00.5P

«009;

A830 8.588

$888.
A8880
88880
$888:
888m:
888m:
:8
88
88
88.8
888

Am>mv>mz
$00.00

$00.0
$000

8:2 58 3.8., .882

$0P0p

358$)

00000.0
hwdhfim
Dumhfiw
0005.0
5000.0
p00v0.~
vu.0N0.0
00000.0
00000.0
00000.0
00.00v.0

$5.82
u Ema

«6E8
u Eco

000—0.

@0055

00. «00;
N0.Nv0.p
Nodvm.’
v0.0n0é
V0000;
V0000;
NFONP.N
00009.0
00.00_..N
00.00w.N
00.000.N

3558.

020033500 ...Zwihmna< odim x<._. hmmmwhz

$8.2 $8.: $8.:
$8.8. $8.8. $88.
$8.8. $8.8 $88.
$8.8. $8.8. $8.8.
$8.2 $8.8. $88.
$8.2 $8.8. $8.8
$2.2 $8.2 $8.8
$8.: $8.8. $8.8
$88. $8.8. $8.8
$8.2 $8.2 $8.8
$8.8 $8.2 $88
@0055 8005; 3.6.8:.
8:8 885 888..
8.88 8.88 8.88;
8.88 888 8.88;
888 888 888;
8.88 888 8.85
8.88 588 8.8:
888 888 8.8:
888 888 888..
888 888 888..
888 888 888;
8.88.. 888 888.8
3.58 3va“. $2.88
$55.38. .0
88 u 68%.
8.8 u :8:
38 u .6:
”0520.. 8.0.2.033
0.558 :58”.

m ugh

$0v.0~

Em}

00.0
00.0
00.0
00.0
00.0
00.0
00.0
00.0
00.0
00.0
00.0

8:88

00. 000. N
00.000. —
ӣ29588. 293m

$00.0N

am:

00005
006;
000;
N00~0
00.000
00.000
0v000
00.000

00.000.—

00.000

00.000.—

3233

u a:
u 3.

$8. P F 58.82888
$ 8.9 288852
$88. 88.85
$88. 8382
$8.8. 8m .82
$8.8. 3min
$88. an. x» .2
$88. 8.88
$88. >3 8.88
$8.2 2m
$888. Ssaésm
:8:
88 58828.86
00.0 €239.82
88 88.85
88 8.382
88 8m 382
88 .826
88 E .c 8<
88 888
88 >3 8.88
88 2m
88 88.8.82m
5.80
5225 .38... An 35
.838 A» E
«8 26 Au Q
2 96 an E
U.....8525
“ESEQQ

89

838 $8... $8.8 $8.8 $8.8 $8.8 $8.8. $88 $8.... 58.88.33
8888 $8.8 $8.... $8.8 $8.8 $8.8 $8.8 $8.8 $3.8 ..88282
8888 $88 $8.8... $8.8 $8.8. $8.8 $8.8 $8.8 $3.8 88.83
3888 $3.8 $88 $5.... $88. $8.8 $8.8 $3.8 $88.8. 8.8:
3.88 $8.8 $8.8 $8.... $88. $8.8 $8.8 $8.8 $8.8 8m. >52
3.88 $8.8 $88 $8.... $88. $8.8 $8.8 $8.8 $8.8 38.5
8888 $8.8 $8.88 $88.... $8. .. $8.8 $8.8 $8.8 $8.8 P. x. .8...
888 $8.8 $88 $3.... $88. $8.8 $8.8 $8.8 $8.8 888m
8888.. $3.8 $8.88 $8.... $8.8 $8.8 $8.8 $8.8 $8.8 >3 8.88
8888.. $8.8 $88 $8.... $8. .. $8.8 $8.8 $8.8 $3.8 3m
8888.. $8.8 $8.8 $8.8 $8.8 $8.8 $8.8 $8.8 $8.8 88.8.88

.830 8.8.8.. .8388 .309; 8.09.2. .5082. .83... .83... .8... .3...

.838 8.88.8 8.83 88... 888 888 8.88 888 8.888 5888.928
.888. 888.8 8.38.. 8.8... 3.3. 8.38 8.38 888 888 888852
.888. 888.8 8.38.. 8.3.. 3.3. 8.38 8.38 8.8.. 888 88.85
.888. 8.88.8 8.38.. 88!. 8.8. 383 383 883 883 8.382
.388. 8. .88 8.8.. 8. .8. 888 888 888 8.8.. :88 8m 82
.388. 8. .88 8.8.. 8. .8 888 888 888 8.8.. 8.8.. .826
.88. 8.38.8 .88... 883 888 8.88 8.88 888. 888 P. x. .2
.888. 888.8 388.. 8.3.. 8.8. .888 .888 883 883. 8.83
88 888.8 8... ... 883 8... .8 883 883 8 .88 8.8... >3 8.88
.88. 8888 8.8 ... 883 8... .8 8.88 2.88 888. 883 .....m
888 8888 8.8 ... 8. . .m 888 888.. 888.. 888 888 9.8.8.8898

.8582 .3338. .3233. .358 .323. .8533 8:88 .5233 ....88

.85 ...... .8. .0 8.8.8 .8. An .88.
$88.88 u .38 88 u .83... .88 An ....
$8.8 u .858 88 u .58. 888.8 u .8: 8 ....w an .8.
$88.88 u .08 8.8 u ...... 888.. u .... ... ....m A... ...
“00:9 53 020) 3:32 ”9:22 €222.me ”flag-mo)... Egan. ”205.002

”20.9.3.8“.

OE<ZmOm IKE
02095300 §§< 0.910 x<._. hmmmm...2_

0 m._m<h

90

838 $.88 $8.8 $8.8 $.88 $3.8 $3.8 $8.8 $8... 58.2288
888 $88 $ 8... $ .... 8.8. $8. 8 $8. 8 $8.8 $8.8 ..888882
8888 $8.8 $8.8 $3.8 $8.8. $8. 8 $8.8 $8.8 $8.8 88.88
.388 $8.8 $8.8 $8... $.88. $8. 8 $8. 8 $8.8 $88 8.82
8.88 $8.8 $8.... $8... $3.8. $8.8 $88.88 $88. 8 $8.8 8cm 252
888 $8.8 $8... $8... $88. $8.8 $8.8 $8.8 $8.8 .826
888 $8.8 $8.8. $3.8 $8... $3.8 $3.8 $8.8 $.88 88 x. .3.
888 $8.8 $8.8 $88 $8.8. $8.8 $8.8 $8.8 $.88 888
88.. $.88 $3.8 $8.8 $...8 $8.8 $.58 $3.8 $3.8 >3 8388
88.. $8.8 $3.8 $8.8 $8... $8.8 $8.8 $8.8 $8.8 .28
83.. $ . .... $3.8 $8.8 $8.8 $8.8 $8.8 $8.8 $8.8 28.8.28
.8850 8.38 .8502? .8692. .38.? .5092. .8818. .88. .8: ...:
.. .885 88 8.8 888.. . .83 .88. .883 8.88 8.8. 8.8. 58.28.28
.88 85 8.8.8 8.8 ... .88. 888 8.88.. 888 8.8. 8.8. ..888882
.88 85 8.8.8 8.8 ... .88. 8.8. 8.88.. 888 3.8. 3.8. 88.88
.38 5 88 8.8 8.8 ... 8.8 8. .8 88.8.. 8.88 8.8. 8.8. 8.82
.885 8. .88 ...8... 388 888 888.. .888 3.3. 3.8. 88 >62
.83.. 8.88.8 8. .8... 88... 8.8 8.8.. 888 .88. .88. .826
.885 8.88.8 8.88.. 8.88 8.8 8.8... 8.8. 8.8. 8.8. 8 x. .2.
.8. . . 5 8.88.8 8.88.. 8 .8 8 8.38 8.3.. 38 .. 8.8.. 88.. 8.88
88 888.8 8.38.. 8.8 88 8 8.8.8.. 888 8.88 8.88 >3 8.888
.885 888.8 3.88.. 3.3.. 8.88 8.8 ... 8.8.. 8.8. 8.8. 28
8. .8 8. .88 3.88.. 883 8. .8 888.. 888 888 888 828.888
.8382 .8338. .3238 .8558 .8588 .8358 .8538 .5358 .558
.28.. ___. 8. .5 5288 .8. an .83
$88.88 u .858 88 u .88: .88 Au .3
$88. u .8va 88 u :8: 888.... u .85. 8 88 an .85
$8.8 u .58 8.8 u .6: 888.. u .5_ ... 88 an .5
”moan. 50b m2a> 5x55. “8:30.. “5.0.0.855 ”Ecgamo>c_ Egan. ”225:0“.-
”23ng
08288188

mZOQEmSOO hzwzwzd.‘ DAM-1m x<h Hmwmmhz

h wAmE.

91

888 888 888 888 888 888 888 888 888 28.28.38
88.8 888 888 888 888 888 888 888 888 ..88852
88.8 888 888 888 888 888 888 888 888 88.85
888 888 8888.8 888 888 8888.8 888 888 888 82382
888 888 888 888 888 888 888 888 888 8m 82
888 888 888 888 888 888 888 888 888 .826
888 888 888 8888.8 888 888 8888.8 888 888 on. 8p .2
888 888 888 888 888 888 888 888 888 888
888 888 888 888 888 888 888 888 888 >3 8.888
888 888 888 888 888 888 888 888 888 2m
8888 888 888 888 888 388.8 888.8 888 888 Sgagm
EEC «.58» 88809;) @092, @835 30055 3%.»: am: am: 2%.
888 888 888 888 888 88. P 82 888 «a 58.88.38
888 888 888 888 888 88.. 82 888 «a :38852
888 888 888 888 888 8888. F 82 888 s: 88.85
888 888 888 888 888 88. _ «a 888 a). 8.382
888 888 888 888 888 88.. «a 888 g. 8m 382
888 888 888 888 888 88.. 82 888 s: .8 am
888 888 888 888 888 88.. £8 888 «a on. .c 8<

s: 888 888 888 888 88.. 82 8888.8 82 8.88m

82 888 888 888 888 8888. , «a 888 82 >3 8.888

at 888 888 888 888 88.. «.2 8888.8 82 2m
8888 888 888 888 888 8888. _ 82 888 «a 88.8.88
38882 63:28 35.88 3:88 3858 @258 €588 €238 $58

as; __a 3. .. 5288 .88 A" .83

$8.88 135 38 u 68: .8288 Au 3.

888 n Qua 38 u :88: 888.8 888; u a: «8 88 A» a.

$88 u Eco 88 u .6: 8.88; 888; u E. x. 88 A» E

a 39; a 39:. ”285.88.
82.8 Eon 33> .952 8522 .8222..me Engage. Ewan.
“22ng

a $958 m._m<» . 8 mafia
8 39¢ oh 8 59:
28.38.28 @8256 no use

a mam/2.

92

3.88 88818 888.8 8888 888 8888 8888 8888 888 6888.888
888 888 $88 8888 888 888 888 .888 888 =3..w8:az
888 8.88 8.88 8888 888 .888 888 8888 888 88.88
8888 188 :88 8888 8888 888 888 8888 8888 8.382
81888 8888 8888 883.8 888 8188 8.88 33.8 888 88 382
81888 8888 8888 $38 888 8.88 8.88 83.8 888 .838
8888 888.8 888.8 88888 8888 8.88 8.88 88888 8888 8 x8 .2.
888 8818 88.8 8888.8 8888.8 8818 888.8 88888 888 888
888 818 818 818 818 88.8 88.8 818 818 >3 28:8
888 88.8 88.8 8888.8 8888.8 888.8 8818 8888.8 8888.8 .28
888 $8.8 88.8 8888.8 888 888.8 888.8 88888 888 28.8.8838
EEO 8.8.8. @8633 3.0935 $855 309;) 88-8: 38} am: :8:
38...: 3.88 838 888 3.88 8888 82 888 82 58.82868
£888. 888 888 888 8888 888 82 888 82 .8885:
£88.88 888 888 888 888 888 82 888 82 88.88
.8188. 8888 8888 8888 8888 8888 82 8888 a... 8.382
38.8.8 81888 8.888 81888 81888 888 82 888 82 8m 382
3818.8 8188 8.888 8.888 8888 888 82 888 82 .838
$81.: 8888 888 8888 888 8888 82 8888 82 P. .c 8<

8). 888 888 888 888 888 82 888 82 8.88

82 888 888 888 888 2818. 82 818 82 >3 8.88

82 888 888 888 888 888 82 888 8.8 28
888 8888.8 888 8888.8 888 888 82 888 8.8 88.8.8828
.8882 38:23 3:88 3:88 38858 $3.88 .8838 3258 358

as... =8 a. .8 £288 .93 Au Ea

$888 u 3va 88 818 u 888: .828 6 38

888 u Ema 818 8.8 u :88: 888.1 u a: 8.. 88 Au .8.

.888 u 3:0 818 8.8 n .6: 888.. u E. S 888 A" E

v 383. u 39: 8.88888_
“mesa.— EOD o:_a> .0332 ”ms—=9. 35.9..me nm~ccE~mc>S 2.2mm
”20.32am

AN wqm<§w w..m<._. - v m._m<._b
N w..m<._. O._. v w._m<.r
28328300 wGZ<I0 “.0 PEG

m mam/2.

93

8888 8888 8888 8888 888 8888 8888 8888 888 5888.868
8888 8888 8888 888 8888 «888 8888 888 8888 ..888882
8888 8888 8888 8888 8888 .888 8888 888 8888 88.88
81.88 88888 88888 88888 81.88 88888 8888 88888 8388 8.382
8 .88 88888 88888 88 $8 888 8888 8888 8838 888 8m >62
8.88 88888 88888 883.8 888 8888 8888 8838 888 .838
8888 8888 8888 8.888 8888 .888 3888 8.888 8888 8 x8 .2
888 888 888 88888 888 8888 8888 8888.8 888 888
888 888 888 888 888 8188 8 .88 888 888 >3 8.88
888 888 888 88888 888 8888 8888 88888 888 3m
8888 888 888 88888 888 8888 8888 8888.8 8888.8 28.8.8898
650 8.588 88880035 .3098; .8505) $0035 1888.8: 38: am: :8:
188.88 8888 8888 8888 8888 8888 82 8888.8 82 5888.928
888888 8888 8888 8888 8888 8888 82 8888 82 ..88882
888888 8888 8888 8888 8888 8888 82 8888 82 88.88
388.3 81.88 81.88 81.88 8388 8388 82 8388 82 8.382
.88 3.88 8 3.88 8 3.88 8 .88 8 $88 888 82 888 82 8m 382
888388 8:88 8:88 8:88 8.88 888 S: 888 82 .838
88. c 8888 8888 8888 8888 8888 82 8888 «a 8 x8 .2

82 888 888 888 888 888 82 888 82 88.88

82 888 888 888 888 881.8 82 888 82 >3 8.88

82 888 888 888 888 888 82 888 82 2m
8888 888 888 888 888 888 82 888 «a 88.8.28
8888882 8888.888 @258 3:80 3858 .8558 8:88 38:88 358

as... =8 a. .8 8.888 .88 An .8888

88888 u .58 88 8.8 u 888: .828 an .3

$88 .... 8E0 818 8.8 n :88: 888.8 a a: «a 88 9 Q

.8.88 n Eco 38 8.8 u .6: 888.1 u E. x. 88 A" E

8 398 8 385. ”8.8888.
”mass 58 Q:_d> «Std: ”9:20.. USEE¢EC ”9:98.502: Ewan.
uggggm

.8 8.858 885 - m 38;
8 885 08 m 388
208.8828 8.0280 “.0 3.5.

or m._m<._.

94

.88 .88.
8888
8888

.888.

.8888.

.8888.

.88 .88.

.8888.
888
888

.8888.

.8885 8.8.88

888 .8
.8 .888.
.8888.
88888
8 .888
8 .888
.888. 1.
82
8.8
8.:
.888.

88.8.2

$00.00
$00.00
$00.00
0 was».

.8818.
.8818.
.8818.
.. .8 .8.
.888.
.88....
.8888.
.8818.
.8818.
.8818.
.8818.

48.8.0055

.88 .88.
8888
8888

.888.

.8888.

.888.

.8858.

.8888.
888
888

.8888.

88.388

«000.00
«000.0
88.000

0 m._0<h
”80:2 58 3.8 .8882

000 ...0
..0000
8000.0
0000.0
0000.0
0000.0
00v0.0
00h 8.0
0hh0.0
h0h0.0
80h0.0

3.0055

2.8888.
.8888.
.8888.
.8888.
.8888.
.8888.
..8888.
.8 .888.
.8888.
..8888.
.888.

3.2300
n .300

u 8.88
n ...mo

.8888. 3.8.8.
.8818. .118.
.8818. .118.
.8818. .8818.
.88.... .8818.
.8818. .8818.
.88.... .88....
.8888. .8888.
.8818. .888 .8.
.8818. .8818.
.888. .888.
8.09.2. .5003...
.88888. .1888.
.8888. .8888.
.8888. .8888.
.8888. .8888.
.8888. .8888.
.8888. .8888.
..8888. .8888.
.8188. .8888.
..8888. ..8888.
..8888. .8888.
.888. .888.
.358 3.858.
.88. __8 3. ..
88 u .888...
8 .8 u .58...
8.8 u .6...

”8522 028.286....

.8 8.85.8 8.85 . 8 8385...
8 8.85 08 8 8.9.8

8008.0
0000.0
0000.0
0080.0
00.0.0
000.0
0000.0
0008.0
0080.0
0800.0
0800.0

.988...

.88888.
..8888.
28888.
.8888.
.8 .888.
.8 .888.
.88888.
.8888.
.8838.
..8888.
.888.

.8588

0008.0
0000.0
0000.0
0080.0
0080.0
0080.0
0000.0
0008.0
00—00
0800.0
0h00.0

.8...

a).
a?
a).
a).
a).
0E
a).
a).
a?
a?
0E

8.58

00000.0
00.000;

”2:85.882. 5288.

200E<m§00 0023.0 “.0 0.20

S 049..

0v00.0
0000.0
0000.0
8000.0
0000.0
0000.0
0000.0
8.0000
0080.0
0 800.0
0000.0

.88...

.8188.
.8888.
.8888.
.88888.
.88888.
.8888.
.88888.
.888.
.8888.
..8888.
.888.

. 1.8.88

.8:
....

0800.0
8000.0
0000.0
0000.0
0000.0
0000.0
0000.0
8008.0
0050.0
050.0
0000.0

. .8...

8.8
8.:
8.:
8.:
82
8.8
82
82
8.8
8.8
82

.058

68.82920
2850232

5.8.3.0
82382
0cm 382
30 .50
00 x» .o<
86000

>3 03:00

20

0:200:30

58.32820
:8 508.82

c8620
5.302
0:0 382
.00 .50
um .C. 8<
38090

>3 8.888

am

0:280:30

£0.90 .80» An .88.

228....
0.. 0:0
8* 0:0

9 a.
a .8.
a ...

“0.05.002

”228588..

95

8888
8.88.
.8 .88.
888
8888
8888
8888
8888
888
888
8888

.8585 8.583 .8>8.oo<3

.8 .88.
88.88
88 .88

.888.

.8888.

.888.8.

.8888.

.8838.

82
82
888 .8

.8>8.>n.z

$00.00
$00.0v
$00.00
5 0430.

0000.0
0000.0
0000.0
0800.0
K000
0000.0
0000.0
0v00.0
hh0o.0
0000.0
0h00.0

8888
8.88.
8.88.

888

8888

8888

8888

8888

888

888

88888

3.8.3.8.

6800.00
$00.00
6800.00
0 30¢...

”82.8. 53 33> 58:82

.8888.
...8888.
...8888.
.8888.
.8888.
.8888.
.188.
.88888.
.8888.
.....88.
......88.

EVOO<>>

NNNPO
000-.0
0008.0
000...0
0008.0
00070
V0080
0808.0
00:.0
000F.0
000—.0

3.2388
n 3.8

"8E8
u ...mo

88888 8888
.8888 .8888.
.8888 .8888.
88888 .8888.
88888 8888
88 .88 8888
88.88 .8588.
8888 .....88.
8888 .8888.
....88 ...888.
888 888
8.835 ...ooSs
E8 .8 28.8
888 ..8 888.8
888 .8 888.8
888.8 888.8
888 .8 888.8
88 .8 88 .8
..88 .8 888 .8
8 .8 .8 8 .8 .8
88. .8 88. .8
888 .8 888.8
888 .8 888 .8
3.588 3.858
.85. __8 .2 ..
888 u .888...
8.8 u .58...
8.8 u .6...

“8522 0222.30

.8888.
.8888.
.8888.
.8888.
.8888.
.88888.
.8888.
.88888.
.8888.
.8888.
.8888.

.288...

.0080
v00...0
V0070
0h0w.0
b00w.0
0008.0
0008.0
000.0
#0080
0009.0
0000.0

8.8588

.88 .88.
.8888.
.8888.
.8888.
.8888.
.8888.
.8888.
.88888.
.8888.
.8.88.
.8 .88.

.88...

.8888.
.8888.
.8888.
.8888.
.8 .88.
.88888.
..888.
.8 .888.
...888.
.8888.
.8888.

.8280

00000.0
00.000;

“8.88.5882: .8288.

.8 8.93.8 8:85 - 8 8.85..

0 040.; O._. n 049:

200.5828 0.02500 ".0 mks”.

0. 0490.

.8888.
.288.
.8888.
.8888.
.288.
.888 .8.
.8 .88.
0.8888.
.8888.
..888.
.8838.

.08...

8888
.8888.
.8888.
.8888.
8888
8888
.8 .88.
.8888.
.8888.
...888.
888

......888

.8._
....

888 88.82885
.8888. 2888232
...888. 88.8.8
.8888. 8.382

888 8:8 382

8888.8 .8838
.288. E x. .2
..888. 8888
.8888. >3 83:8
.8. .88. 2.0

888 28.8.8838

..8...

8888 58.2.2888
.8888. 28:82.82
.88.... 88.8.8
.8888. 8.382

8888 8:8 262

8888 .8838
.8888. P. x. :2
.8888. 88888
.88.... >3 83:8
.8888. 20

888 8:28.858

2.588
823.0 3.8. A" .88.

Ewan. Au 3.
88 888 An .8.
... 888 A" ...
“288882
”0.205201

CHAPTER 7

EMPIRICAL STUDY

This chapter discusses the design and results of an empirical study dealing with
regulated rates of return. The study is designed to investigate whether the value of holding
companies whose subsidiaries are regulated by commissions favoring the stand-alone tax
treatment are greater than those with subsidiaries regulated by commissions favoring
consolidated tax adjustments. Tobin's Q is used as a relative measure of the holding
company's market value. In the statistical model, I regress Tobin's Q against an
independent variable designed to be a proxy for the consolidated tax adjustment affect The
following sections, data selection, hypotheses, the statistical model, and results, discuss
the findings of this empirical test.

W911
To be included in the data, each participant must meet each of the following criteria:
1. Be included in the Compustat data tapes.
2. Provide local telephone service, being categorized as having a SIC of 4813.
3. Be included in the Energy and Regulatory Matters Information Service (ERMIS)

data tapes over the 1988-1992 period.

Fifteen firms meet these criteria:

Ameritech NYNEX

Bell Atlantic Pacific Telesis

BellSouth Telecommunications Rochester Telephone

Centel Southern New England Telephone
Cincinnati Bell Southwestern Bell

Contel (1988 & 1989) US West Communications

GTE United Telephone

Lincoln Telephone & Telegraph

96

97

The selection of this particular set of companies is based on the following explanation.
First, the area of interest are those regulated industries where a holding company
organization carries debt at both the parent and subsidiary levels. The Public Utility
Holding Company Act of 1935 restricts parent companies from issuing debt to purchase the
common stock of electric and gas utilities. To obtain holding company data with significant
levels of debt at the parent level, this reduces the selection process to the water and
telecommunications regulated industries. Second, Lerner (197 3) states that of all the
regulated industries, 95% of the telephone industry is organized under the holding
company form of corporate organization. Walker (1988) points out that parent holding
companies in the telephone industry are often permanent tax loss entities. This narrows the
selection to the telecommunications industry. Third, the telecommunications industry is
divided into categories by SIC code, 4812 - Radio Telephone Communications and
4813 - Telephone Communications, Except Radio Telephone. 4812 firms can be divided
into cellular telephone and paging companies. Both of these areas are not regulated,
removing the 4812 companies from the selection. 4813 firms can be divided into local
telephone service and long distance companies. Only local telephone service companies are
regulated by the state public utility commissions. Fourth, the hardest data to obtain for this
empirical study is the regulated sales revenue, broken down by the states in which it is
obtained. This "by-state" sales revenue data is included in the information reported to the
Federal Communications Commission for operating utilities that have a minimum annual
sales revenue of one hundred million dollars or greater. The ERMIS data tapes contain this
information for the fifteen holding companies listed above. Fifth, the included holding
companies make a major portion of the local telephone service industry. Also, being the
largest, if consolidated tax adjustments have any affect on the market value of the
companies operating in this industry, it should affect these firms. Last, similar to Chen,
Hexter and Hu (1993), selection of the largest holding companies in the industry provides a
higher degree of homogeneity.

Hiram

All of the data firms in this study are organized as parent/subsidiary systems. The
Tobin's Q's therefore have consolidated data included in their construction. As an
example, Ameritech has 14 subsidiaries, of which 7 are located within their Bell Group.
The Bell Group provides telephone service to 17 million customers in five midwestem
states, Illinois, Indiana, Michigan, Ohio, and Wisconsin. Ohio is categorized as being a
stand-alone tax state, with the other four being consolidated tax adjustment states.

98
Ameritech's telecommunications interests therefore operates within states which mostly
require a consolidated tax adjustment Conversely, NYNEX has telecommunications
subsidiaries that operate primarily in New York, a stand-alone tax state. Following the
logic that a stand-alone firm is allowed a higher return than a consolidated-entity firm, the
Tobin's Q for NYNEX is hypothesized to be greater than the Tobin's Q for Ameritech.

Most of the data firms also have business operations in unregulated areas, which again
would affect the holding company system's Tobin's Q. This introduces a second
underlying question: Are holding companies who own regulated subsidiaries receiving a
"fair" return on their regulated investments? The theoretical expected affect of the
regulation would be that moving from regulated to unregulated business operations, the
Tobin's Q would increase to "1.0". The logic behind this statement is based on the fact that
a majority of agencies, 32 of 51, utilize some form of a consolidated tax adjustment. Given
the simulation results of the included tax adjustment models in Chapter 6, those systems'
Tobin's Q's operating under these tax adjustment restrictions are hypothesized to be lower
than the systems operating in regulated areas utilizing the stand-alone (independent firm)
tax treatment. This then would result in systems whose revenues come predominantly
from regulated telecommunications would on average have a Tobin's Q less than 1.0.
Alternatively, systems operating predominantly in unregulated business concerns, because
of competitive forces, would theoretically have Tobin's Q's closer to 1.0. Maintaining our
focus on the primary question, this second question has been included as a control variable
in order to isolate any findings on the primary tax adjustment issue.

115”!!!”

The following percentage has been obtained for the data firms:

Regulated Adjusted Sales (REGSLS) = Consohdatgggffi $321531???“ Sales

The regression methodology utilized will be the procedures outlined in Chen, Hexter,
and Hu (1993). The regression equations will be the following:

99
Linear relationship:

Qi = a1 + b1(REGSLSi) + Eai

Quadratic relationship:

Qi = a1 + b1(REGSLSi) + b2(REGSLSi2) + g,

Cubic relationship:

Qi = a1 + b1(REGSLSi) + b2(REGSLSi2) + b3(REGSLSi3) + Q

If each of the public utility commissions are "allowing" a fair return to the
parent/subsidiary systems in the selected data, then none of the regression coefficients
should be significant. If any of the coefficients are significant, then it is expected that the
sign of any of the REGSLS variables will be negative. Consolidated tax adjustments lower
the allowed returns for the systems where this tax adjustment is imposed, which in turn
should cause the Tobin's Q's to be lower. Since the consolidated tax adjustment should
have a monotonic effect as more of a system's regulated sales are adjusted, both the linear
and higher order form variant coefficients should be negative, or at least have an combined
negative result

To guard against the results being based on some omitted variable, the following
control variables are systematically included:

Total regulated sales (TOTSLS = total regulated sales / total sales)
Financial leverage (FIN LEV = debt / replacement cost)

Firm size (FIRMSIZE = replacement cost/ total assets)

Advertising intensity (ADEXP = advertising expense / replacement cost)
Replacement cost (REPL COST, in $ billions)

Year group (YR = 1-5 for years' 1988-1992)

100

The expected results for the control variables, if significant, are the following:

Total regulated sales
(T OTSLS)

Financial leverage
(FIN LEV)

Firm size
(FIRMSIZE)

Advertising intensity
(ADEXP)

Replacement cost
(REPL COST)

Year

negative (based on the results of Chapter 6, if a consolidated
tax adjustment affect exists, with most state commissions
favoring a consolidated tax adjustment, as the proportion of
regulated sales increases the firm's Tobin's Q should
decrease)

negatixe (as financial leverage increases, the possibility of
financial distress increases, lowering the firm's Tobin's Q)

negatixe (as the size of the firm increases, firm grth
slows, lowering the firm's Tobin's Q; as replacement cost
increases above total assets implies an older, more mature
growth opportunity firm with a lower Tobin's Q)

nositixe (greater advertising intensity, supports greater firm
profitability, increasing the firm's Tobin's Q)

negative (denominator term for Tobin's Q, plus potentially
being an additional measure of firm size; higher replacement
cost, lower Tobin's Q)

unknown (This is included to test for marked changes in the
overall economic environment during the period 1988-1992.
With having only fifteen firms for 1988 and 1989, and
fourteen firms for 1990 through 1992, all of the data was
included for the entire period of 1988 through 1992.)

The desired expectation is that any displayed significance before the control variables are

included, will remain after their inclusion.

101

{11' l S E .

The results of the initial regressions are shown in Table 13. In regression #1
REGSLS and YR are regressed against Tobin's Q. REGSLS's coefficient shows no
degree of significance along with displaying a positive sign. These results infer that over
the period being tested, no significant consolidated tax adjustment affect has been found.
The control variable, YR, displays the same results, implying no significant changes in the
telephone industry's economic environment over the 1988 - 1992 period.25 The Durbin—
Watson statistic reflects there exists positive autocorrelation, implying the possibility of a
left-out key variable. Testing for our second underlying question if holding companies
within the local telephone service industry are receiving a "fair" return, regression #2
regresses TOTSLS against Tobin's Q shows no significance. Its coefficient does possess
the predicted negative sign. In this second regression YR again is not significant. In
regression #3 REGSLS and TOTSLS along with YR are regressed against Tobin's Q, with
no change in characteristics for both variables from the first two regressions.

Regression #4 tests the significance of the remaining four control variables FINLEV,
FIRMSIZE, ADEXP and REPL COST with YR. Similar to the reported 1980 results of
Chen, Hexter and flu (1993) regarding the financial leverage variable, FINLEV is negative
and insignificant. FIRMSIZE is negative and significant at the one percent level. ADEXP
is significant at the one percent level, and is the only variable whose coefficient depicts a
sign opposite from the expected. One possible explanation is that only the larger firms
show any advertising expenses, with a large-firm affect causing the sign to reverse from
the expected positive sign. YR has become significant at the five percent level for the first
time, with its sign being positive, implying an increasingly favorable economic
environment within the local telephone service industry over the 1988-1992 period.

Regression #5, incorporates REGSLS with the control variables run in regression #4..
REGSLS’s coefficient is again positive and insignificant. There is no change in the
regression #4 coefficient characteristics for FINLEV, FIRMSIZE, and ADEXP. Similar to

 

25This control variable has been included in every regression. Based on the reduced
number of observations for each year, each regression has been performed using the data
for the entire period. This variable has been included to determine the validity of including
all of the data in one regression.

102
the reported 1976 results both McConnell and Servaes (1990) and Chen, Hexter and Hu
(1993), the coefficient for REPL COST has become insignificant The coefficient for YR
is again positive, and is significant at the one percent level.

In regression #6, REGSLS is replaced with T OTSLS. The coefficient for TOTSLS is
still negative, and now significant at the one percent level. The only other change involving
the other control variables reported for regression #5 is the significance level for both
FINLEV and YR is now at the five percent level.

In regression #7, all seven variables are included. REGSLS shows no change in
characteristics, while TOT SLS’s coefficient has remained negative, and significant at the
one percent level. FINLEV, FIRMSIZE, ADEXP, REPL COST, and YR show no change

from regression #6.

As more variables have been added to the regression, the Durbin-Watson statistic has
declined further from two, implying more pronounced positive autocorrelation. It appears
that another key variable for the holding company business organization has been omitted.
Given that REGSLS has remained insignificant, even if this key variable were identified
and included, it is doubtful that there would be any change in REGSLS' performance.
Given these initial results it would appear our consolidated tax adjustment affect is not a
key factor in determining the value of a holding company system.

Positive autocorrelation can be dealt with by the use of transformed variables when as
in this case no other key independent variables can be found. Two forms of transformation
have been included in this study; the iteration approach, and the first difference approach.

 

The transformed dependent model:

Y't = Yt'th-l =(BO+B1XI+8t)'p(BO+let-l+8t-1)
= Bo(l-p)+Bl(Xt-pXt.1)+(at-p£t-1)
= [30(1'9)+131 (Xt-pXt_1)+ut

103
When more than one independent variable exists:

Y't = Bo“ 'PHZBi (Xit‘Pxit-1)+“t

The estimator for the autocorrelation parameter p, is r:

n
2 et-1(et)
r=2
n
2
2 e t—l
r=2

Table 14 provides the results of the iterative transformation. The independent variable
inclusions for regressions' #1 through #5 are the same those performed in Table 13. The
results for each iterative transformation regression are basically the same as those reported
in Table 13. FINLEV and REPL COST both display a higher degree of significance in
regressions' #4 through #7 than is reported in Table 13, while the significance of YR is
reduced slightly. The Durbin-Watson statistics in Table 14 all provide the desired values
close to two.

 

The transformation model:

p =1,then

Y't = Yt'Yt-l = (50+51Xt +8t)'(BO+let-l “it-1)
= 51(Xt'xr-1) + “t

When there is more than one independent variable:
Y't = 251 (Xit ' Xit—1)+ “r

Table 15 shows results of the first difference transformation model. For all
regressions REGSLS is again insignificant. Representative of this insignificance,

104
REGSLS changes in sign in regression #7. TOTSLS now shows significance at the five
percent level in regressions' #2 and #3, that increases to the one percent level in regressions
#6 and #7. FINLEV, FIRMSIZE, ADEXP and REPL COST all show significance at the
one percent level in regressions' #4 through #7. YR shows significance at the one percent
level for regressions' #1-#3, then becomes insignificant in regressions' #4 through #7.
For all of the regressions the Durbin-Watson statistic displays that the positive
autocorrelation has been significantly reduced.

ll-l'BllSlB '8]

Table 16 expands the independent variables to include REGSLS 2 and REGSLS 3.
Regression #1 regresses REGSLS, REGSLS 2, and YR against Tobin's Q. Both
REGSLS and REGSLS 2 are significant at the five percent level, with REGSLS showing
the predicted negative sign, and REGSLS 2 a positive sign. YR is again positive and
significant at the ten percent level. Regression #2 includes TOT SLS which has no affect on
the initial results for REGSLS and REGSLS 2, but YR has become insignificant TOTSLS
has a negative coefficient that is insignificant.

Regression #3 includes REGSLS, REGSLS 2, REGSLS 3, and YR as the
independent variables. REGSLS, REGSLS 2 and YR all show the same characteristics
shown in regression #1. REGSLS 3 is negative and significant at the ten percent level.
Regression #4 again includes TOTSLS, with no changes in performance of any of the
variables.

Regression #5 includes REGSLS, REGSLS 2, FINLEV, FIRMSIZE, ADEXP,
REPL cosr, and YR as the independent variables. REGSLS and REGSLS 2 maintain the
same sign but have become insignificant. FINLEV and REPL COST both are negative and
insignificant, while FIRMSIZE and ADEXP are negative and significant at the one percent
level. YR is positive and significant at the five percent level. When TOT SLS is added in
regression #6, TOTSLS is negative and significant at the one percent level.

REGSLS 3 is added in regression #8, with its coefficient being positive and
insignificant. REGSLS and REGSLS 2 have switched in sign, which is explained by their
being insignificant. The only other change is regression #6 results is that YR is less
significant at the ten percent level. The Durbin-Watson statistics again show positive
autocorrelation for all of the regressions in this table.

105

Table 17 shows the results of applying the iterative transformation. REGSLS and
REGSLS 2 both show increases in significance for the earlier regressions, but again as
more control variables are added become insignificant The Durbin-Watson statistics are
now more comfortably closer to two, but YR still shows significance at the ten percent
level.

Table 18 provides the first difference transformation results. Comparisons with
Table 17 shows that TOTSLS is now significant at the five percent level for the earlier
regressions, which again increases to the one percent level when the other control variables
are included. YR is significant at the one percent level for the earlier regressions, but
becomes insignificant with the inclusion of the other control variables. Similar to Table 17,
the Durbin-Watson statistics are closer to two than those shown in Table 16.

11 II 5 l. .
Two additional regression forms were attempted:
Q = a1 + ZBi(LN[Xi]) +£i
and
LN[Qi] = a1 + ZBi(LN[Xi]) +8i
where Qi Tobin's Q

LN[Qi] the natural log of Tobin's Q
LN[Xi] = the natural log of the independent variables presented above

The results of these regressions did not provide any additional substantive information,
suffering mostly from the loss of observations. This was caused by the abundance of
“zero” value figures for REGSLS and ADEXP. When any of the observations included a
“zero” value, since the natural log of zero cannot be found, that firm’s data for that year
were all dropped from the regression. Given the above, these results have therefore been
omitted.

106

Summarizing, this empirical study has not been able to directly uncover the existence
of a consolidated tax affect for holding companies operating in the local telephone service
industry. The regression results for REGSLS and its derivatives implies that the
consolidated tax affect is not a major factor in determining the value of the included holding
companies. Based on the results for TOTSLS there is support that commissions have
provided too low a return for subsidiaries within the local telephone service industry. This
same empirical process has been repeated using the market-to—book ratio as the dependent
variable. The market-to-book results which support these conclusions have been provided
in the Appendix.

107

v0
commie
omapé

L369:
38.0

339?.
88.?

.3391
$2.2.

£836-.
33.?

L83. 2
823.

29.8.9.
32...

"38.2
385

13.3.2
«and

b

no
Noam—to
ham—VF

Loewe:
2.85

389?.
88.?

cages;
23.2.

£336..
c .35.

L38. ...
2.2.?

£33.:
own P . T

338.3
33d

no
mmovd
mopné

1.338
$26

7:36..
88.?

£885-.
88.2.

2836;
536.

.38. 2
:85.

308.3
885

.526:
acme."

m

00
2.31.0
opomé

LESS
83.0

Louuo.l
28.?

238.3
3.63.

£336...
83.?

743.1
33.?

223.3
mnsm.r

v

we
vwmoé
anon.—

303.3
26.0

$393
33.?

3.25.2
owned

133$
2: I

m

3 2.
«39° 3.35
:3; 8;;
885.2 .88.:
38.0 366
$393
33.0-
$852
385

1:222 .5586:
an: 32.6

N F

mgm<> mug omhjnwmm 52..

dog gummuwm mg #84 #2290
2.3mm gmmfimm O mZBH

0F mqm<b

#2 r a ... 38%...» .
$0 an ... 3.2%.... r
a: r a s £5.29. .

so
«<1

”2.26% coaszgeao

3:23: a

222.38 E

3:25. ..

222:8". Boodwu

8:23... a

522.38 e09

8.3.33 ._

£22.38 manic

38....» a

322.38 Pfizm

3:233 a

222.38 999

8:23: a

822.38 98w.

3:233 a

222:8“. 2528
3.3.8053.

as + Ex an N + SE n E>

108

no
3.3.0
#059.
3.696

..83...
8.88

...888.
8.8.?

£8.58...
88....

£88....
88.?

£888..
..88-

£888..
88..-

.8288.
888

£838.
8....

N

10
o‘ood
mono;
0.806

..88...
8.88

«#000 F .N-_
5.00.0.

£888-.
82.....-

£883..-.
88.?

£8. .8.
88.?

£838..
8 .8. ..

£8 . .88.
8.8..

«0
~33...
mnooé
Nhnmd

L838...
3.88

£888-.
88.?

£383..
888..

£888..
88.?

L88... ...
8.8.?

.888.
888

£888.
8......

n

no
«:3...
Foooé
thnd

L88...
8888

£888-.
88.?

£888-.
888.-

£888..
88.?

L838.
88.?

£888.
8. . ..

v

no
owned
echo;
voopd

.888.
8.88

.8388.
88.?

.888.
3.88

£888.
838

m

8 8
888 8.88
88.. 83..
.88.... .88....

78:8. .88...
3.88 8.88
.88....
8.8.?
.8888.
888

£888. £838..
888 888

N ..

mdSm<> mm._<m DESDGmm $2...
I0<0¢am< m>F<mwt .20—#9 NZPOwGEOO 20:.(4wmm085<

Edam... Swmwmmumm O $7.8...

I m..m<._.

$8. _. a ... 88.....2» .
8 n a .8 ...-2.8.» ..
.8. a a .- 8258.» 8

c...
N<m

”2.2.8 88.838358
a

.2.»..».» ..
8.2.8»

.28....» ..
822:...»

.23....» ..
23.2.38

.23....» ..
522:8»

.28....» ..
822.8»

.2.»..».» ..
2.205000

.2.»..».» ..
.88.»...88

.2.»_.».» ..
20.2.38

k80 .EW.

gm

>m42E

mdhOh

aw.

..Z<...w28

. 5.89.08.

.....2. . ...u + ......_.x 8 . ....x. ...... w + ......8... « ...... 8 . ...> n ......

109

v0
named
oovmd

$2.31
82.?

.8899.
$85

:82.3
9552.

£88.:
2.2.?

£88.»;
.2 _ . 7

5.311
33.7

$813
86.?

n

no
vowed
Fhmmd

738.9.
3.3.?

.3899.
«wood.

133.3
83.8.

$28.2
confio.

$38.1
0% P . F -

.5332:
88.?

mo
hmomd
hand

.65.:
88.?

122.3
«mood.

£3221
9.3.2.

283.3
5.2.?

128d;
33.0.

380.2
385

m

we
smood
mowed

302.3
88.?

£885;
«mood.

£33.:
53.2.

132.1
33.?

28.3.2
8%.?

cm
award
oonwd

13292
2.8.0

..ESNE
33.?

393.3
866

m

3 on
335 3.96
38d 38d

.5992 233.2
88.0 836
..Soom.~-_
32.?
3.3.2
386
N _

#2 u a .a E§=c9o .
in la a 3328.. s
x; n u .- t§=co3 .

an
N<¢

6:236 assigning

53.3.. a
222.38

3.233 a
E20508

3.2...» s
220503

3.28.» a
522.38

3.235 ._
E20508

3:288 ..
E20508

3.223» a
E20508

.EOOEW.

EEK

>m..Z_m

wdho...

aw.

u 5.8200,".

2: + avax . .55 as w u .2; . :7 u E.>

m§> wm4<m awhSDGmm 32:
IO<8¢¢< mozwmwmmfi ...me .§0< Emmoo ZQDDMEOOOKZ
Wham—m gummvwm 0 $28...

9 w4m<h

110

No
m—mmd
town.—

.389:
$36

355?.
88.?

222.3
22.2.

.8231
$3.?

8536..
83.?

.._Eo~.m._
$2.?

:57:
SK...

:891
63.?

”3.3.2
38.0

:mmoha
o.om.~

no
mmhvd
003.“;

L8» v.3
38.0

823.2
386.

183.9.
22.1.

133.1
89.?

Cocoa;
88.0-

82::
$3.0

733.9.
33.9

$945.8
386

333.3
83...

x.

no
.366
name...

...32..:
23.0

33:...
38.?

£33.:
38.2-

138.3
:33.

83 .. ...
88.?

133.3
«2 ..7

$392
t: fio

£893
38.?

152......
«SEN

o

v0
«:16
Foam...

..Evowi
$2.0

.33..-.
«Sod.

£33.?“
35.1.

:2. _ , .1
33.?

”3&6;
88.0-

3.36.:
$8.0

“38.9.
33.?

£336;
seam."

m

M555; mw._<w owhfinowm m<wz_._.202

on
onw...o
500‘;

.839:
2.6.0

.33.:
83.0-

.3893
88.7

.gmaom.:
mmma.—

.._vm9.~._
33.?

...Smw.~_
83.0

v

be
vow—.0
sort."

.329:
88.0

.3891
{8.7

...:8.:
88d

.._mo\.~.~.H
$8.?

1326:
32.0

n

so
omood
oopvé

.32.:
335

$86.3
88.0.

L338
R36

L23. 2
38.?

15.3.3
58;

N

no
00090
(owe;

..35...
«29°

L388
236

...omno. 2
9.3.0-

22.8.3.
82.5

dog Zmewmwwm wm<30w ._.m(m._ >I<Zom0

or m4m<h

9...me zmewmwwm O 92.8...

32 u u s 28.2%.. .
.xh ..a ... 23:83 ..
4: a a ... 25.29. .

3:.
«(a

8:236 coasicsso

3.2:... a
522.38

3.2.2. ._
222.38

38:8» a
E22280

3.223» a
522.38

7.35. a
222.38

3.3.88 5
522:..8

3:23... ..
E20503

3:23» a
222.38

5.222» a
E20508

3:23.» a
222.38

.500 .Ewu

>327.

91.9.9.

ggwowu

magma

awn.

P255200

a c2829.;

Eu + Ex an n + 8a u E>

111

p0
ovvod
9:5;
nomad

.859:
$66

.33. ...
$86.

.3291
33.3.

:33».
38.?

..oamc. ...
83.?

333.9.
28...

.336.
336

.33.:
88.?

.386.
936

133.3
3%..

N0
ooomd
econ;
«050.0

.33.:
'93...

...aofi...
28.?

.8399.
2.3.8.

£83....
83....

.386.
3.0 . .o-

.335.
83.0

$899.
33.?

.32....
33.?

13.8.2
2.35

«0
00‘06
mvahé
Mum—v.0

.33.;
88.0

.33.. ...
N .85.

.326.
33.2.

2396-.
35.?

.83. ...
88.?

.8893
2.8.0.

.82.:
«or:

.33.. ...
82.?

$33.2
8;...

no
pvomd
homo;
wound

..NSo. :
«89o

..«cno. ...
.. .86.

£336..
82.3.

288.1
33.?

.82....
8&6-

Lack...
$5...

..No 8. ...
33.0.

233.3
885

wdSm<> mm._<w othDmuwm 5.023.202
IO<OmnE< m>z.<mm._._ .ZOFO< w>=bwmm00 20:.<.wmmOOOS<
$.5me zmemmwwm O $2.8...

mo
mom—.0
vpooé
ovomd

.33.:
«89°

.33.:
835.

.83....
etc...

...:oo.~_
83d

..83.~..
coco...

tomofiu.
3:6

00
000.6
pone.—
ooowd

..moon. z
885

.33. ...
38...

...Sfid.
3.2:

13.59.
memo...

.33.:
33.6

..-..a . ...u +

t mama?

co
abmpd
00pm;
omowd

.21.:
2:90

.2229.
83.?

...:w ..o.
335

.0893
2.3.9

.3093
52.6

no
Nomwd
mm .0;
«NON...

.EB...
83...

£38.“.
38...

£58.:
32.?

133...:
:35

8.: u u s 3823.. .
in 1.. .a 52.29. ..
x. u a .- t8=to3 .

.tu
N<m

5.3.35 cSEEEQB
a

.25.... a
222.38

.323. a
222.38

.23.... a
522.38

.2333 ._
222:8".

.25....» ..
E20508

.233. ..
E20508

.325.» a
222.38

.0325» ..
€22.33

.0388 a
522.38

.283» a
822.38

...80.EW.

gr.

>53“.

make.

9mg”.

«(q—wow”.

aw...

2.528

u 8.828”.

.....5x a . .ax. ...... u + 3-5.3.. u .....> a . .6. u a;

112

me
«0:20
nmm1.N

.38....
$8...

...»omo. ...
38.?

:33».
3.2.0..

u.» . o . ....-.
2.8.?

...»oo». ...
8...?

:88...
$3.?

88......
3o»...

.38....
SR?

.33....
88.?

no
:36...
owvmd

.33..-.
86.?

...».o».~..
88.?

£38.».
83:»?

£38.».
83.?

.33..-.
3.8.?

.88....
83...

.33....
«B»?

.38..-.
32.?

00
man
hpwvd

.39....
58...

:32».-.
38.?

zoom»...
83......

.89.».
33.?

...»8». .-.
32.?

22.86-.
«8...?

...».2...
one»...

...»3». .-.
88.?

to
wooed
oonmd

83......
38.?

...» .8...
88.?

:38».
38.2-

283?.
one»?

.32..-.
13.?

2&8...
one»...

...oowod.
38.?

w§m<> mm..<m thjbwwm ...—523.202

on
mean...
ooomd

£33.».
one»...

...:3. ...
83.?

....»m»...
8%.?

...»...od.
~»»»..

285.».
mono..-

so
amend
«Food

£33.».
83...

.88....
83.?

L323.
23..

138.».
u .8...

so
bpsnd
«cvod

.839».
88...

....»v». .-.
0.3.?

:33».
$8...

$38....
38.?

no
some...
«wood

£33.».
38...

12.3....
2b»...

:35...
$3.?

.8. u a ... .5225.» .
....» u a .- ...-2......» ..
2.. .. a .- 5222» »

...u
N<I

”2.236 58532350

.2.»..».» ..
222:8»

.22....» ..
222:8»

.22....» ..
«:305000

.22....» ..
£22.38

.23....» ..
222.38

.22....» ..
€22.33

.22....» ..
222.300

.22....» ..
50.2.33

.22....» ..
€22.38

...80 .EWu

>82...

“39.0..

ngwwu.

gammy.

am...

a 5.8203“.

...: + ........x - ....x. ...n. u n .....> . ...> u ....>

19.0..ng wOZwmwnEE ...me .75:b< wZHmemOO 20F<dmm000§

mp m...m<._.

”Baum Zmemmem O $2.8...

CHAPTER 8

CONCLUSIONS

The debate for determining the allowed rate of return for wholly owned regulated
subsidiaries within financial literature has focused on two key issues: subsidiary risk, and
the acknowledgment of debt by the parent to support its investment in the subsidiary. The
two primary approaches involved have been the independent firm and double leverage
approaches. Proponents of the independent firm approach state that the inherent risk of
the investment should dictate the return provided. They view the parent and each of its
subsidiaries as separate individual entities, making the parent's use of debt irrelevant in
subsidiary rate of return determination. In contrast, supporters of the double leverage
approach counter that the risk of only the publicly traded securities should be used to
determine a subsidiary's rate of return. They point out that recognizing the debt at the
parent level follows the context of capital budgeting theory. Ignoring parental debt
results in a regulated subsidiary earning more than its cost of capital, which produces a
windfall return at the expense of the subsidiary's ratepayers. Two alternative approaches
by Fitzpatrick and Beedles in the early 1980's presented methods that like the double
leverage approach recognizes parent level debt, while also identifying differences in
subsidiary risk.

The double leverage approach has also been supported by several public utility
regulatory commissions, reaching the height of its popularity in the early 1980's. The
decisions favoring the double leverage approach often based this decision on the parent
company shareholders receiving a return in excess of the market cost of funding used to
support their investment in the regulated subsidiary if the utilization of debt funding were
ignored. Of the rate cases involving the double leverage concept, 90 percent of these
rulings dealt with utilities operating in the local telephone service industry. I discovered
that the double leverage approach was not the only manner in which the regulatory
commissions recognized debt held at the parent level. What had originally been used by
commissions as a manner of spreading subsidiary tax credits across all subsidiaries of a
holding company, a consolidated tax adjustment was also being utilized as a means of

113

l 14
returning the parent level debt interest tax shield back to its subsidiaries. These
apportioned tax credits lowered the subsidiary revenue requirements, and in the process
also adjusted the subsidiary's allowed rate of return. Wide acceptance among
commissions for this alternative can be identified by the fact that the double leverage
approach is no longer being utilized by any of the fifty-one regulatory commissions.

Ezzell, Hsu, and Miles (1991) bridged the gap between financial theory and
regulatory decisions by offering an alternative that acknowledged both differences in
subsidiary risk and the interest tax shield benefits of the parent level debt. With the use
of a perpetual cash flow model they derived a rate of return adjustment equation.

This dissertation has presented an alternative derivation of the EHM approach, using
the more familiar Modigliani and Miller perpetual cash flow framework. The equations
presented by my alternative derivation simplify the makeup of the EHM equations.

Specifically:

EHM: WACCk*= WACCk (ll—[LL -tL pr ) 26

where: WACC k“

subsidiary k's adjusted weighted average cost of capital

WACCk = subsidiary k's unadjusted weighted average cost of capital
t = corporate tax rate
Lp = the parent's market value debt ratio
Ik‘L’k
L. = —
2:[k
Ik = the investment in subsidiary k
Lk = the market value debt ratio for subsidiary k

 

26This equation is presented in the MM perpetual cash flow framework. Because of
their manner of discounting the equation presented in their paper is slightly different. In
=1+k___:.)_.W
place of "t", are Zpst’ where zp = —E and st k

1+1'f sw

1 15
21k

 

Patterson: WACC k* = WACCk(

where: PVTSpfind] = the present value of the parent's interest tax
shield without an adjustment

Now it can be easily pointed out is that the EHM adjustment is an adjustment of the
subsidiary's rate base. It also reduces the adjustment components to those that are more
easily understood and can be more practically applied.

The second contribution of this dissertation is the systematic testing of alternative
approaches as to their ability to fairly compensate wholly owned regulated subsidiaries.
The five alternative rate of return adjustment approaches cited in financial literature,
together with seven alternative consolidated tax affect adjustment procedures previously
used by regulatory commissions were placed within the confines of the MM perpetual
cash flow model. For a parent with two wholly owned regulated subsidiaries, six
scenarios were tested. The results point out that only the EHM approach consistently
provides a zero net present value and Tobin’s Q equal to one in every scenario. The
stand-alone (independent firm) approach showed the expected wealth creation at the
parent level, returning a positive system net present value, and a Tobin's Q greater than
one. The seven alternative consolidated tax adjustment techniques taken from regulatory
commission decisions were found to overcompensate for the parental debt tax shield,
causing negative net present values and Tobin's Q's less than one. The Fitzpatrick and
Beedles approaches were found to also inadequately compensate for the debt held at the
parent level of a holding company.

Finally, the empirical study of this dissertation provides evidence that fails to support
the premise that the consolidated tax adjustment affect is a key factor in determining the
value of a holding company system operating in the local telephone service industry. The
study does provide support for the secondary conclusion that regulatory commissions
have allowed a less than "fair" return for regulated subsidiaries within this industry.

APPENDIX

APPENDIX
MARKET-TO—B 00K EMPIRICAL RESULTS

This appendix has been included to offer evidence that the empirical results found in
Chapter 7 are not the result of any unique qualities possessed by Tobin's Q. The
dependent variable used here is the market-to-book ratio. The market values for each
holding company’s financial securities are found in the same manner as outlined in
Chapter 4. The data selection, hypotheses, and statistical model are the same as those
defined in Chapter 7.

Results
l' E l lSlII'llE!

Regression #1 of Table 19 is the results when REGSLS and YR are regressed against
market-to-book. Here the coefficient for REGSLS shows the desired negative sign, but is
not significant. The coefficient for YR is positive and also not significant. When
TOTSLS and YR are regressed against market-to-book (regression #2), the coefficient for
TOT SLS is negative and significant at the one percent level. YR is again positive and
not significant. When REGSLS, TOTSLS, and YR are included as independent variables
(regression #3), TOTSLS remains negative and significant at the one percent level, with
REGSLS and YR showing no change in either sign or significance.

Regression #4 includes the remaining control variables FINLEV , ADEXP, and
TOTAL ASSETS along with YR as independent variables.27 When regressed against
market-to-book, FINLEV is negative, and significant at the ten percent level. ADEXP
and TOTAL ASSETS are both negative and significant at the one percent level. YR is
positive and for the first time significant at the one percent level. When REGSLS is
added to the regression (regression #5) its coefficient is positive, but still insignificant.

 

27FIRMSIZE (replacement cost/total assets) used in the Tobin's Q regressions was
excluded since a suitable replacement could not be found.

116

1 17
The other variable coefficients display no change from the characteristics reported for
regression #4.

Regression #6 replaces REGSLS with TOTSLS. The coefficient for TOTSLS is
again negative, and has retained its one percent level of significance. ADEXP and
TOTAL ASSETS show the same coefficient properties as in regression #5, while
FINLEV has maintained the same sign and increased in level of significance to the five
percent level. YR has also maintained its positive coefficient, but its level of significance
has fallen to five percent. When all of the independent variables are included (regression
#7), REGSLS displays a negative coefficient that is again insignificant. TOTSLS is
again negative and significant at the one percent level. FINLEV is negative and
significant at the five percent level. ADEXP remains unchanged in both sign and
significance. TOTAL ASSETS and YR both show the no change in sign, but their
significance is now at the five percent level. The Durbin-Watson statistic for all of the
regressions implies that a positive autocorrelation problem is present. When the results
of Table 19 are compared to those of Table 13, REGSLS has remained insignificant, but
in Table 19 now displays the desired negative sign. TOTSLS in Table 19 shows
significance at the one percent level for all of the regressions when it is included.

Table 20 provides the results when the iterative approach is used to correct for
autocorrelation. The only changes noted from Table 19 involve YR in the fourth through
seventh regressions, where its significance levels have dropped to the ten percent level
for regression #4, and are not significant in regressions’ #5 through #7. When Table 20 is
compared to Table 14, the changes stated for the first comparison are still accurate, and
the significance levels for YR have decreased.

Table 21 provides the results for the first difference approach transformation
application. REGSLS shows a positive sign in the first and third regressions, but remains
insignificant for all of its regressions. TOTSLS is again negative and significant at the
one percent level for all of its regressions, as is the case for FINLEV, ADEXP and
TOTAL ASSETS. YR shows a decrease is significance level as more independent
variables are added. When compared with Table 15, the results are basically the same,
with TOTSLS showing slightly higher levels of significance in the second and third
regressions.

 

118
II “I. E I 1511:1113 1

In Table 22, when regressed against market-to-book the coefficients for REGSLS and
REGSLS 2 are less consistent in sign than is displayed Table 16. Both terms switch sign
and lose significance when REGSLS 3 is included in the regressions. When all
independent variables are included, all REGSLS terms show five percent levels of
significance which is not present in Table 16. TOT SLS again shows a negative sign and
one percent level of significance for all regressions. The results for FINLEV, ADEXP,
and TOTAL ASSETS are consistent with those found in Table 16, as is true for YR. The
Durbin-Watson statistics again show positive autocorrelation.

When the iterative approach is applied, the results in Table 23 show that the drop in
significance for REGSLS and REGSLS 2 occurs more quickly than in Table 17.
REGSLS 3 shows the same positive sign and a ten percent level of significance when all
independent variables are included. TOTSLS once again is negative and significant at
the one percent level for all of its regressions. FINLEV, ADEXP, and TOTAL ASSETS
show basically the same characteristics as those of Table 17. YR is positive and slightly
less significant than in Table 17. The Durbin-Watson statistics again have moved closer
to two.

The first difference results in Table 24 when compared to Table 18 basically follow
the same pattern displayed in the preceding comparison.

Overall, the regressions utilizing market-to-book ratios support the evidence brought
out when Tobin's Q is used as the dependent variable. Most pronounced, during all of the
market-to-book regressions TOTSLS is consistently negative and significant at the one
percent level.

119

we
103.0
0000...

....»o». ..
88.»

...»o»».»..
3.8.?

..»3».»-.
».»».»..

CORD” FN.Nu_
whmhd.

..3.».»-.
.3». ..

.»3..?.
88.?

133.».
88.»

h

00
«03.0
1000..

...»»»». ..
88.»

..»»»..»..
38.?

...oo».»-.
83.»?

...8»».»..
»».».?

.88.».
»»»». ..

...»3.».
83.»

0

00
...—«00
000...

..»».».».
2.8.»

..»»»».»..
88.?

..».»».»..
»».»...-

..»»».... ...
3.»?

.38....
88...

.88.».
88..

n

no
00.0.0
N12...

...»»..».
38.»

..»»»».»..
88.?

..»».».»..
38....

... . 3....
28.?

..»»»».o..
38..

v

00
00070
owové

..»»..o.
88.»

..».»».»-.
3.8..-

..8».?.
18.?

..»»o».».
88.»

m

00
00070
oovvé

.»»»».o.
88.»

...».».»..
8o»..-

:83».
88.»

N

w..m<.m<> mug Duhsawm $2..

00
00.0.0
«~00;

.88...
83...

.83.».
33.?

1.63.»:
o»»o..

dog 6.8mm0mm mg .55.. 5.29.0

0F m.m<.r

Wham... gum—Gum $0.75;

.3. u a 3 38......» .
....» u.» 3 ...-2.32» ..
.2 a a 3 3.2.32» .

...»
N<m

23.3.» c8332.»...o

.23..»3 ..
322:8»

.23....» ..
322.38

.2.»..».» ..
322.38

.2.»..».» ..
322.38

.2.»..».» ..
322.38

.2.»..».» ..
E2233

.2.»..».» ..
€22.33

”Emmi ..(kO...

3Q“?

>53».

makah

am

#25028

a 3.8230

...» + ....x ...» w + ....» n ...>

120

 

#0
0000.0
1000.—
.0080

.8»...
88...

: 88.».
88...-

:28».
88......

:58».
88..-

:83...
88..-

.88....
2...».

:82».
»»o»..

h

00
0000.0
002.0.—
00020

:3»...
...»...o

: 32.».
38...-

:38».
28.»?

:8» . ....
o8... ..

:8»...
oov». ..

:38».
m»»»..

0

00
000v.0
0000.0
00080

..»»»»...
38...

:88».
$8....

:31...
88.»?

:22».
88...-

.23....
8......

:88»:
28...

0

00
0V0v.0
5000.0
.v «V0

......8. :
88...

:88».
88....

:32».
28.»...

:»..».»-.
8......

:28»:
88...

v

2.0
000.0

0.00..
0000.0

.85....
8......

:88».
28...

.2 . ..o.
28...-

:88».
»»»»..

0

00
000.0
0000.0
0000.0

.34»...
8......

:8. .».
08»...

:83».
8......

0

00
0000.0
0000.0
0000.0

.82....
2......

.»o».o.o.
38....

:88»:
3.2...

...... I ... 8 2.3.5.9»

....» u» ... 82......» ..
.... .. a .» ragga...» ..

...»
0<K

2.2....» c8.»;.c...so

.22....» ..
2.22:3»

.22.»... ..
€22.38

.223.» ..
E20503

.22....» ..
522:3»

.22.»... ..
£22.38

.22.»... ..
20.2.38

.22....» ..
€22.38

a

05% ..(kOh

é

>52...

0.5...0...

a”.

k2<kaOO

3 5.82001

......a - ...» + ......_.x a . ....x. ...». w + 3......» a .....> a . ...». u ......

w..m<_m<> ww4<w awhSDme ”202...
IO<Ommm< 0.5.5.0.: .ZO;0< m>POw¢mOO gar—0000:?

3.500”. SwmmEOMm 3.0%;

00 04mg:

 

121

00
#0000
0.00.0

.82.».
88.».

:88».
88.».

:88».
88.»?

:22».
88..-

:38».
8:..-

.88.».
88.».

h

00
0000.0
00V0.0

.82.».
88.»-

:88».
88.»-

:88».
88.»?

:28».
2» . . ..

:88».
2 .8. ..

00
000v.0
.0000

.88»...
83.».

:88».
88.»-

:88».
88......

:88».
.8 . . .-

.8. ..»..
88.»-

0

00
0 .080
0000.0

..88...
83.».

:88».
88.».

:88».
»»..».»..

:82».
2» . . ..

00
0.00.0
V0050

:38».
8»».»

:82».
88...

.83.».
28.»

a» 8
2.2.» 88.»
82» 88.»

:88». ...»v»...
88.» »2..»
:88».
8.»...
.88.».
88.»
» .

...». u a ... 28...»? .
....» nu .» 82...»... ..
.... u a ... 2.2.2...» .

....»
0<m

2.2.2» c2823...»

2.2.2» ..
522:8»

.22....» ..
222:8»

2.2.2» ..
2.22.38

.22.»... ..
822:8»

.22....» ..
€22.33

2.2.2» ..
222:8»

wkwww< ..(POF

809

>w.z_u

0.9.0..

am.

0 5.80.000

...» + ......x . ....x. ...... w n ...... . ...> u ......

04mim<> 00.2.0 thSDGwm «202..
IO<80¢< gamut-0 .5;“. .70....0< 94.50008 ZO:.<..mEOOOS<

.0 0.5.2.

mhgmm 75.000003 3.0.5.5.;

122

00
000v.0
0.5...

..»»»». ..
88.»

:88». .-.
8.8.».

:88».
.»»»»..

..88...
88.».

:88».
88...

...»8.».
8.».

....8.».
88».

...»»»»...
...8.»

..88...
88.8.

0

v0
00v0.0
00h...

...» .» .».
88.»

:88».
8.8a»-

:28».
88.1.

.88.».
88.».

..88...
8.».

....»8...
:8..-

...8»...
88.»

...8».»..
88...».

10
00.10
00.5..

.88....
88.»

....»8».
8-8.»-

:88».
88.»?

...88. ...
82.».

:88».
.8...

...8 . .»..
88.».

78.».
88.»

.... .8...
88.»...

00
.0000
00¢...

:88».
28.»

:88».
8.8.»-

:88».
2.8»?

.88.».
83.».

.88.».
.»8.»

.88.».
2.8.».

L88».
88.»...

0

wdSm<> 00.20 00.2.5000 04.023202

00
0v.0.0
«300..

.88.».
88.»

:88».
38...

...»..8».
»»»»..

....8». ...
.»8»-

.88.».
8....»

.28.».
88.»?

v

N0
000.0
000...

.. . .»...
8.».»

.:»»»..2
»»»»..

.88..-.
8o»..-

.88.».
.».....»

.»... . . ...
88.8.

no
0000.0
5000..

.82.».
86.»

:88».
38...

:88».
88.»

:38».
88.».

.88.».
88...».

N

00
001.0
~01...

... .8...
88.»

:88».
88.»

: .8 . .»..
88.».

.88....
88»...

.0002 290000000 00.300 ._.0<m.. >m<z_omo

0.53001 20.00me00 68.0....ng

mm 040‘...

2.». .. a .» ...-22:2» .
2.» u» ... ...-2.8.» ..
x. u a .- »..-2.8.» a

3...
«<1

2.2.2» c8.»z<c_»s»

.22....» ..
222:8»

.»..»..2» ..
2.22.38

.2.2.2» ..
2.22:...»

.»..»..2» ..
€22.38

.»..»..2» ..
£22.38

.»..»..2» ..
2.22..»8

.»..»..2» ..
E2228»

.22....» ..
E2228»

.»..»..2» ..
622:8»

0h000< 45.0...

>012».

01.0..Oh

9000*

90.00”.

am.

2.528

a 5.820»...

.... + ....x ...» u + .»... n ...>

123

2.5. n u .5 E8559. .
.85 us ... ...-2:5... :
x. n a .5 252%.... 5

«5 no 55 «5 no on 55 no 5..
5355 39.5 33.5 5535 5.35 5935 53.35 .355 «E
5.2.. 558.. .38.. .58.. 32... 5555.. 3.5.. 355.. ”2.2.2m 5252,2355
5535 5535 5535 5535 «35.5 5535 .855 9535 5

.35.... .353... .53.... .39.... .5355. .2555. .5555. .3555. 3.2.2» ..

«5555 .585 5:35 3555 2.55 2555 55.55 3555 £22.38 E
...585... 23.55-. ...Sao... 5.885.. 3.2.2. ..

Smut..- 8535. 8.355. 89285. 522.38 9.28559
5.5355. 5.385.. ....336. 2285.. 3.2.2. ..

2.35.. 585.1. 3555.. 555.3. 522.38 p.509
.2325. 2.5.5.... 5.55:5. ....52... 3.2.25 ..

$25. 2555. $555. 5355. 222.38 >32:
1135.. $3555-. 3.255-. 5.2.85. 3.2.25 ..

m. .5. .. o . .5. .. 38... «an. .. 222.38 999
..83... .853... ...553... .....E... 3.2.2» ..

235 2.55.. 83.. 255.. 222.38 938.“:

.83..-. .23.... .535... .5555... .33.... .335. 15555.... 5.2.55.5. 3.2.2....
$55... 255..- 55555 «535 9.85... Si... «585 8555 222.38 «(98%

852.5. .335. .38..-. .32.... .5585. .5585. 2555...... 5.35:... 3.2.2» ..

$85 «3&5 55555. 5855- «5.55. 555.5- 325. 325. 522.38 98w.

.38.... ....25. ... .355. ... ..omm... ... .885. .535. .5535. 85.35.. 3.2.2. ..
3255- 8.5.3.. $55.... 38.3 5855.. 55:35. «8555. 3.5.3. 222.38 22.5200
5 x. o m 5 5 m . 5 5.82525

..-..x . .... + .....2.x 5 . ....x. a... u + 3-...5... a .....> 5 . ...> n ....>

w..m<_m<> mm._<w owhSDGmm m<wz_._.ZOz
IO<O¢nE< m2...<mm._._ .20...0< MZFmeEOO 20F<dmm0005<
whimmm 20_mmwm@wm x080......m¥m(2

mm w..m<h

124

we
named
Fowvd

.55555.
5.55.5-

...5555. ...
55.5555.

5.553....
55555.-

5.55.5.5.
..555.

5.555....
5555...

.5555...
55555

.55555.
.5555.

.5535.
5555.5

55 ..5
5555.5 55.5.5
5555.5 5555.5

.5355. .5535.
5555.5. 5555.5.

5.3.55.5. ...5555. .-.

55-5555. 55.5555.
5.5.55.5. 5.5555,...
5.555.. 55555..

...5555. ... 5.55555.

.5555- 5.55.55-
5.355....
5 .55..-
..5555...
5555.5
..555... ..5555...
355.5. 55.55
.3555. ...5555...
5555.5 5555.5-
5 5

no
new...
omood

.3555.
555.5.

5.55555.
55.5555-

5.553....
52.5....

5.33.5.
55.5.5-

....5.5. ..
5555.5

.2555...
....55-

co
nonvd
Sofiw

5.. {.55.
5555.5

..5555.5.
5555..

...55555.
5555..

.555...
555....

.5535.
52.5.

50
0000.0
05nd

..5555...
.535

...5.55.5.
.55....

.5555...
555....

.8555.
5555.5-

.5 55
5555.5 5555.5
555.5 5.55.5

52555.5. ..5555...
5.555 53.5
5.5.55.5.
55.5..-

...... .55.... 5.55555.
5555.5 55555
5.55555. 5.55.5.5.
5555.5- 5.55.5-
5 .

555. .. a .5 28.5595 .
2.5 ..a .5 5.55.5525 ..
2.. u a .5 E8255... 5

...u
N<m

”2.2.25 552532.555

3.2.2. ..
522.38

3.2.25 ..
55.2.38

3.2.25 ..
5.22....8

3.2.2.... ..
5.22....8

3.2.25 ..
5.22....8

3.2.25 ..
£22.38

3.2.25 ..
2.0.5530

3.2.25 ..
€22.38

mmwm< ..(HO...

nag

>m..z.n.

0.950...

ggwm

3609...

am“.

5 50.55565...

...: + ......_.x - ....x. ...5. w u ..-..> . ...> n ....>

md<E<> wméw QwhSDomm 5.523.202
IO<Ommm< mOmemuED Emmi .ZO...Q< wZPOmmmOO ZQHEECOOOHQ‘
whswwm ZO_wmwm0wm Edy—Eva;

¢N m4m<h

LIST OF REFERENCES

LIST OF REFERENCES

Backman, J. and Kristen, J.B. (1972) "Double Leverage and the Cost of Equity Capital",
Public Utilities Fortnightly, (September 28), 32-38

Beedles, W.L. (1984) "A Proposal for the Treatment of Double Leverage", Public
Utilities Fortnightly, (July 5), 31-36

Beranek, W. and Miles, J .A. (1988) "The Excess Return Argument and Double Leverage",
The Financial Review , Vol 23, (May), 143-150

Brennan, J.F. and Humphreys, R.K. (1973) "The Double Leverage Concept Revisited",
Public Utilities Fortnightly, (April 26), 19-25

Brous, RA. and Kini, O. (1992) "Equity Issues and Tobin's Q: New Evidence Regarding
Alternative Information Release Hypothesis", The Journal of Financial Research, Vol
15, (Winter), 323-339

Brown, J .E. (1974) "Double Leverage: Indisputable Fact or Precarious Theory?", Public
Utilities Fortnightly, (May 9), 26-30

Chappell, H.W. and Cheng, DC. (1984) "Expectations, Tobin's Q, and Investment: A
Note", Journal of Finance, Vol 37, (March), 231-236

Chen, H., Hexter, J.L. and Hu, M.V. (1993) "Management Ownership and Corporate
Value", Managerial and Decision Economics, Vol 14, (July/August), 335-346

Copeland, BL. (1977) "Double Leverage One More Time", Public Utilities Fortnightly,
(August 18), 19-24

Ezzell, J.R., Hsu, BC. and Miles, J.A. (1991) "An Analysis of Regulated Rates of
Return for Wholly Owned Subsidiaries", The Journal of Financial Research, Vol 14,
(Summer), 167-180

Fitzpatrick, DB. (1977) "Subsidiaries' Capital Costs - A Compromise Approach", Public
Utilities Fortnightly, (June 23), 23-26

Griffin, 1M. and Wiggins, SN. (1992) "Takeovers: Managerial Incompetence or
Managerial Shirking?", Economic Inquiry, Vol 30, (April), 355-370

Helmuth, J.A. (1990) "Tobin's Q Ratio and Electric Utility Regulation", Review of
Business and Economic Research, Vol 25, (Spring), 1-10

125

126

John, K. and Mishra, B. (1990) "Information Content of Insider Trading Around
Corporate Announcements: The Case of Capital Expenditures", Journal of Finance,
Vol 45, (July), 835-855

Jones, J .R. and O'Donnell, J .L. (1978) "Double Leverage - Lawful and Based on Sound
Economics", Public Utilities Fortnightly, (June 8), 26-31

Kim, K., Henderson, G.V. Jr. and Garrison, S.H. (1993) "Examination of Tobin's Q for
Takeover Firms", Quarterly Journal of Business and Economics, Vol 32, (Winter),
3-26

Klock, M., Thies, CF. and Baum, CF. (1991) "Tobin's q and Measurement Error:
Caveat Investigator", Journal of Economics and Business, Vol 43, (August),
241-252

Lang, L.H.P., Stulz, RM. and Walkling, RA. (1989) "Managerial Performance, Tobin's
Q, and the Gains from Successful Tender Offers", Journal of Financial Economics, Vol
24, (September), 137-154

(1991) "A Test of the Free Cash Flow Hypothesis: The Case
of Bidder Returns", Journal of F mancral Economics, Vol 24, (October), 315-335

 

Lerner, EM. (1973) "What are the Real Double Leverage Problems?", Public Utilities
Fortnightly, (June 7), 18-23

Lindenberg, EB. and Ross, SA. (1981) "Tobin's Q Ratio and Industrial Organization",
Journal of Business, Vol 54, (January), 1-32

McConnell, M. and Servaes, H. (1990) "Additional Evidence on Equity Ownership and
Corporate Value", Journal of Financial Economics, V0127, (October), 595-612

McGilsky, D. (1986) "An Empirical Study of a Ratemaking Issue: Determining the Federal
Income Tax Allowance of an Affiliated Utility", unpublished dissertation

O'Donnell, J .L. and Walker, M.M. (1987) "A Comparative Market Test of Double
Leverage Techniques", Michigan State University working paper

Opler, T. and Titman, S. (1993) "The Determinants of Leveraged Buyout Activity: Free
Cash Flow vs. Financial Distress Costs", Journal of Finance, Vol 48, (December),
1985- 1999

Pettway, R. H. and Jordan, B. D. (1983) "Diversification, Double Leverage, and the Cost
of Capital", The Journal ofFinancial Research, (Winter), 289- 300

Rozeff, MS. (1983) "Modified Double Leverage - A New Approach", Public Utilities
Fortnightly, (March 31), 31-36

Seeds, K. J. (1978) "The Direct Approach to Subsidiaries' Capital Costs", Public Utilities
Fortnightly, (May 11), 18- 20

Servaes, H. (1991) "Tobin' s Q and the Gains from Takeovers", Journal of F mance
Vol 46, (March), 409- 419

127

Smirlock, M., Gilligan, T. and Marshall, W. (1984) "Tobin's Q and the Structure-
Performance Relationship", American Economic Review, Vol 74, (December),
105 l- 1060

Stevens, J .L. (1990) "Tobin's Q and the Structure-Performance Relationship: Comment",
American Economic Review, Vol 80, (June), 618-623

Walker, M.M. (1988) "The Double Leverage Adjustment and the Allowable Rate of Return
for a Regulated, Wholly Owned Subsidiary", unpublished dissertation

Public Utilities’ Cases

Bluefield Water Works and Improvement Co. v. Public Service Commission of West
Virginia ( 1923) 262 US. 679, 692

Federal Power Commission v. Hope Natural Gas Co. (1944) 320 US. 591, 603

Re Brockton Edison Company v. Department of Public Utilities, Mass., 400 N.E.2d 838
(Sup. Jud. Ct. of Mass. 1980)

Re The Chesapeake and Potomac Telephone Company, Formal Case No. 595, Order
N o. 5623, District of Columbia Public Service Commission, January 25, 1974;
4 PUR 4th; 1-82

Re City Service Gas Company, Opinion N o. 396, Docket N o. G- 18799 (Federal Power
Commission 1963); 49 PUR 3d; 229

Re Continental Telephone Company of Maine, EC. No. 2183, C. No. 440 (Maine
Public Utilities Commission 1977); 18 PUR 4th; 636

Re Florida Telephone Corporation, Docket N o. 7 80912-TP, Order No. 9551,
September 17, 1980; 39 PUR 4th; 452-472

Re GTE South, Inc, Case No. 90-522-T-42T, West Virginia Public Service
Commission, May 31 1991; 123 PUR 4th; 257-282

Re GTE Southwest Inc., Docket No. 5610, 15 Tex PUC Bull 1, Texas Public Utility
Commission, February 23 1989; 106 PUR 4th; 194-342

Re Hawkeye State Telephone Company, Docket Nos. U-270, U-392, September 7,
197 3 (Iowa State Commission); 2 PUR 4th; 166-187

Re Iowa Public Service Co. (IPS), Docket No. RPU-87-3, Iowa Utilities Board,
June 17 1988; 94 PUR 4th; 239-289

Re The Narragansett Electric Company, Docket No. 1288 (Rhode Island Public Utilities
Commission 1978); 23 PUR 4th; 516

Re New England Telephone & Telegraph Company v. Maine Public Utilities
Commission, 390 A2d 8 (Me. Sup. Jud. Ct. 1978); 27 PUR 4th; 1

Re New England Telephone and Telegraph Company, Docket Nos. 80-142 et a1,
March 30, 1981 (Maine Public Utility Commission); 42 PUR 4th; 182-252

128
Re Newton Water Company, Docket No. 790911 (Connecticut Public Utilities
Commission 1980)

Re RCA Alaska Communications Inc., U-78-4, Order No. 33, April 22, 1981 (Alaska
Public Service Commission); 42 PUR 4th; 683

Re Southwestern Bell Telephone Company, Docket No. 60,800-U (Kansas State
Corporation Commission 1960); 34 PUR 3d; 257

"71111111111lll'llllll