CGMMGRS STGCK PRICE PREMIUMS FOR VOTING
CONTROL '

"fliesis €01 the Degree of Ph. D.
MECHSGAN STATE UNIVERSITY
RAYMOND c. HELWIG ‘
.1973

 

LIBRARY

Midllb m State
University

 

This is to certify that the

thesis entitled

COMbDN STOCK PRICE PREMIUMS FOR VOTING CONTROL

presented by

Raymond C. Helwig

has been accepted towards fulfillment
of the requirements for

Ph.D _degree in Finance - Business

% Was

Major professor

Date 2/1/73

0-7639

 

 

 

 

 

ABSTRACT

COMMON STOCK PRICE PREMIUMS
FOR VOTING CONTROL

By

Raymond C. Helwig

This study examines the price premiums paid for
control stock, defined as additional amounts paid for types
or amounts of stock which facilitate voting control, as
compared with the price of otherwise comparable stock (of
the same company) which does not facilitate control. Three
methods of obtaining or retaining voting control which are
studied are (1) cash tender offers, (2) private purchases
of large blocks of stock, and (3) classified common stock.

Fifty-three cash tender offers, thirty-two private
purchases of large blocks of stock, and twenty-nine companies
with classified common stock are studied. Multiple regres-
sion analysis is used with cash tender offers and private
purchases of large blocks of stock. The classified common
stocks are analyzed both on a time series basis with the aid
of charts, and on a cross-section basis using two by two
tables and the chi-square test.

The dependent variable in each case is the premium

as measured by the price ratio, defined as the ratio of the

Raymond C. Helwig

price of control stock to the price of non-control stock.
For cash tender offers the price ratio is ratio of tender
offer price to market price (close or bid) two business
days before announcement of offer. The price ratio for
private purchases of large blocks of stock is the ratio of
block sale price to market price two business days before
date of agreement. The price ratio for classified common
stocks is the ratio of market price of voting common to
market price of non-voting (or reduced voting) common of
the same company. Several independent or explanatory vari-
ables are selected for the analyses, including some variables
suggested by other writers as possibly important.

The largest and most consistent premiums are found
for cash tender offers, with classified common stocks next.
Private purchases of large blocks of stock have the lowest
mean and widest range of premiums.

The regression equations explain a substantial por-
tion of the variation in stock price premiums, both for
cash tender offers and for private purchases of large blocks
of stock.

Support is found, in all three methods studied, for
the hypothesis that the percentage premium for control stock
is greater when management performance (as measured by
criteria such as rate of return on equity and per share
earnings growth) has been relatively poor. Dividend policy
is an important factor in the size of price premium on the

voting shares of classified common stock.

. Vfi-‘me—‘~.._—-l . A

Raymond C. Helwig

Other writers have suggested that high liquidity
makes a company attractive for takeover. The findings of
this study suggest an opposite effect-—low liquidity (as
measured by current ratio or quick ratio) serves as a warning
Sign which motivates investors to get control and to pay a
higher premium for control.

A book value per share which is high in relation to
market price is strongly associated with a relatively high
premium for private purchases of large blocks of stock, but
not for cash tender offers, and little if at all for classi-
fied common stock. Possible reasons for these differences
are suggested.

A wide prior year's price range is associated with
a larger premium for cash tender offers, but with a smaller
premium for private purchases of large blocks of stock.
Possible reasons for these opposite results are considered.

In private purchases of large blocks of stock,
corporate purchasers paid higher premiums than individual
purchasers, and corporate sellers received higher premiums

than individual sellers.

COMMON STOCK PRICE PREMIUMS

FOR VOTING CONTROL
By

Raymond CI Helwig

A THESIS

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

DOCTOR OF PHI LOSOPHY

Department of Accounting and
Financial Administration

1973

© Copyright by
RAYMOND CLARK HELWIG

1973

ACKNOWLEDGMENTS

I thank the members of my committee--Professors
Harry G. Brainard, Milton B. Dickerson, and Alan E.
Grunewald, Chairman--for their helpful guidance and con-

structive comments.

ii

TABLE OF CONTENTS

ACKNOWLEDGEMENTS . . . . . . . . . .

LIST OF TABLES O O O O O O O C O O 0

Chapter

I.

II.

III.

IV.

PRIVATE PURCHASES OF LARGE

INTRODUCTION 0 O O O O O O O 0

Purpose of the Study. . . . .
Need for the Study . . .

Methods of Obtaining or Retaining Voting

Control . . . . . . . .
Reasons for Control Having Value.
Hypotheses . . . . . . . .
Variables . . . . . . . .

CASH TENDER OFFERS. . . . . . .

Review of Literature. . . . .
Cash Tender Offers Studied. . .
Variables Employed . . . . .
Method of Analysis . . . . .
Findings. . . . . . . .
Discussion of Findings . . . .

Review of Literature. . . . .
Transactions Studied. . . . .
Variables Employed . . . . .
Method of Analysis . . . . .
Findings. . . . . . . -
Discussion of Findings . . . .

CLASSIFIED COMMON STOCK . . . . .

Companies Studied. . . . . .
Variables Employed . . . . .
Method of Analysis . . . . .
Findings. . . . . . . .
Discussion of Findings . . . .

iii

Page

ii

NH

coma-N

l7
I8

25
37

54

55
61
62
64
66
76

88

89
92
96
98
Ill

V. SUMMARY AND CONCLUSIONS.

Common Patterns Among
Inconsistencies Among
Other Findings

BIBLIOGRAPHY .
APPENDICIES .

Appendix
Appendix
Appendix
Appendix

C(3u1>

iv

Page,
0 O O O O O O 122

the Three Methods . 123
the Three Methods . .127
O I O O O O O 133

. . . . . . . 137
. . . . . . . 144
145
147

148
149

LIST OF TABLES

Cash Tender Offers: Variables Employed . .

Distribution of Actual Values of Price Ratio
Group A O O O O O I O I O I O 0

Regression Steps, Stepwise Addition of
Variables, Group A. . . . . . . . .

Regression Steps, Stepwise Addition of
Variables Group B. . . . . . . . .

Results of Regression Analysis, PREM is
Dependent Variable. . . . . . . . .

Unexplained Variations for Observations with
Predicted Value of Price Ratio Under
110.0%. C I O O I O O O O O 0 0

Observations with Unexplained Variations of
13.0% or More from Value of Price Ratio
Predicted by Regression Equation for Group

A C O O O O O I O O O O I O 0

Pairs of Variables Having Simple Correlations
of .40 or Higher with Each Other . . . .

Market Price One Week Before Announcement of
Offer as Percentage of Market Price Two
Days Before (Number of Offers). . . . .

Private Purchases: Variables Employed. . .
Actual Values of Price Ratio, Transactions
before November 3, 1967 and after March
16, 1969 (Number of Transactions). . . .

Comparison of Price Ratio with Date of
Transaction (Number of Transactions). . .

Regression Steps, Stepwise Deletion of
variables. a o o o o o o 0 o o o

Page

19

26

27

28

29

33

33

35

36

63

67

67

69

Table

3.7

Coefficients for Step No. 6, Stepwise
Deletion of Variables. . . . . . . .

Observations with Ratio of Book Value to
Market Price Greater than 1.00. Grouped by
Listing Status . . . . . . . . . .

Unexplained Variations for Observations with
Predicted Value of Price Ratio under 90.0%
or OV’er 119 09% O O O O O O O O I 0

Observations with Unexplained Variations of
13.0% or more from Predicted Value of
Price Ratio . . . . . . . . . . .

Pairs of Variables Having Simple Correlations
of .40 or Higher with Each Other . . . .

Classified Common Stocks: Variables Employed

Classified Common Stocks, Companies for which
Charts Suggest Possible Association of One
or More Independent Variables with the
Price Ratio . . . . . . . . . . .

Comparision of 1972 Price Ratio with Market
Price,25 Classified Common Stocks (Number
of Companies) . . . . . . . . . .

Comparison of 1972 Price Ratio with 1971 Rate
of Return, 22 Classified Common Stocks
(Number of Companies). . . . . . . .

Comparison of 1972 Price Ratio with 1971 Rate
of Return, 22 Classified Common Stocks.
Revised to Change Class Limit of Price
Ratio to 1.10 (Number of Companies) . . .

Comparison of 1972 Price Ratio with 1971
Quick Ratio, 18 Classified Common Stocks
(Number of Companies). . . . . . . .

Comparison of 1972 Price Ratio with 1971
Quick Ratio, 18 Classified Common Stocks.
Revised to Change Class Limit of Quick
Ratio to 1.00 . . . . . . . . . .

Comparison of 1972 Price Ratio with 1971 Debt
Ratio, 18 Classified Common Stocks (Number
of Companies) . . . . . . . .

vi

Page

70

71

75

99

102

103

104

106

106

108

Table 7 Page

4.9 Comparison of 1972 Price Ratio with Prior
Year's Dividends, 25 Classified Common
Stocks (Number of Companies) . . . . . 110

4.10 Comparison of 1972 Price Ratio with 1971
Shares Ratio, 23 Classified Common Stocks
(Number of Companies). . . . . . . . 110

4.11 Comparison of 1972 Price Ratio with 1972
Ratio of Book Value to Market Price, 23
Classified Common Stocks (Number of
Companies) . . . . . . . . . . . 112

4.12 Comparison of 1972 Price Ratio with 1972
Ratio of Book Value to Market Price, 23
Classified Common Stocks. Revised to
Change Class Limit of Price Ratio to
1.10 (Number of Companies) . . . . . . 112

vii

CHAPTER I

INTRODUCTION

Purpose of the Study

One purpose of this study is to determine the price
premium paid for control stock of various corporations
under the following three methods of obtaining or retaining
voting control:

1. cash tender offers

2. private purchases of large blocks of stock

3. classified common stock

The price premium paid for control stock is defined
as an additional amount paid for a type or amount of stock
which facilitates voting control, as compared with the
price of otherwise comparable stock (of the same company)
which does not facilitate control.

Another purpose is to explain the varying values
of the stock price premium for each of the above three
methods of obtaining or retaining control. A number of
additional variables, selected for this purpose, are

examined to determine the possible influence of each

variable upon the size of stock price premium.

A third purpose is to make a search of literature
relating to the value of control, and to relate the
empirical findings of this study to the findings and

analysis of others.

Need for the Study

Each of the three methods studied is an important
means of obtaining or retaining voting control. However,
very little empirical research has been published about
any of these methods. Further, despite the fact that the
methods are to some extent interchangeable, nothing has
been found which compares and contrasts them.

Information about these methods is needed by
investors, not only those seeking to obtain or retain
voting control of a corporation, but also those concerned
about the possible effects of others' desire for control
upon the price of their shares.

Investors interested in acquiring control need to
know not only the method which may be best to use in a
given situation, but also the premium likely to be required.
Other investors would want to know what premium they might
expect to receive on shares of a certain company if a
contest for control should develop.

Methods of Obtaining or
Retaining Voting Control
There are four main methods of obtaining or

retaining control of a corporation, of which one breaks

 

s

down into four sub-classifications. The four main methods
are:

1. Direct purchase of shares

a. privately-negotiated purchase of
large blocks of stock

b. purchase of shares on the open market

c. cash tender offer

d. exchange offer of securities of
offering company

2. Merger not preceded by purchase of shares

3. Proxy contest

4. Classified common stock.

Two of the above main methods are not included in
the study: merger not preceded by purchase of shares, and
proxy contest. Merger not preceded by purchase of shares
implies an acquiescence on the part of the management of
the acquired company. This acquiescence may be heavily
influenced by factors other than the exchange terms.
Further, because the securities issued in exchange often
include a newly-issued convertible preferred stock, it is
difficult to establish the value received per share of the
acquired company; trading in the new security may not
begin until considerably after the date of announcement
of the offer. A proxy contest does involve cost to those
attempting to capture control. However, complete data on
costs incurred by these insurgents are believed to be
unavailable.

Also, two of the four sub-classifications of direct
purchase of shares are excluded from the study: purchase
of shares on the open market, and exchange offer of

securities of offering company. The problem in studying

the purchase of shares on the open market would be obtaining

 
 
 

.ui

 

 
 
 

and interpreting information. The price paid for control
in this method is the average differential paid as the
large-scale buying drives the market price above the level
which would prevail in the absence of such buying. The
price paid in this method would be difficult to accurately
determine, because of the usual lack of availability of
information about the amounts acquired and prices paid
(particularly prices paid). Further, it would be impossible
to determine what prices would otherwise have prevailed.
The reason for the exclusion of exchange offer of securi-
ties is similar to one of the reasons for excluding merger:
the securities issued in exchange often include newly-
issued securities, such as debentures, convertible pre-
ferred stock, and warrants. It is difficult to accurately
determine the value of these new securities as of the date

of announcement of the offer.

Reasons for Control Having Value
The value of being able to control a corporation
varies with the circumstances of each corporation and the
circumstances and desires of those persons seeking control.
There appear to be three reasons that control of a corpora-
tion might have value:
1. There may be a possibility of increasing the
value of the stock by improving the company's
profitability or image, by merging it with

another company, or by liquidating it. This

is probably the most important reason for
seeking control of a corporation, and the
one which inspires this study.

Perquisites of control may be available.

For example, the person with control may

be able to have himself or a relative
installed as an officer at a level of status
and compensation (perhaps including stock
options) higher than the person could

obtain elsewhere. The status and fees of
being a director are usually available to
one having control. However, it is believed
that perquisites are not usually a major
motive; they are not examined in this study.
There may be a possibility of reducing com-
petition, though any such attempt is likely
to run afoul of the anti-trust laws. It
might be possible, for example, to secure a
source of supply of raw materials (perhaps
at a favorable price) by gaining control of
a supplier. A market for a company's product
might be assured by gaining control of a
customer or prospective customer. Control
need not be direct to be effective; it may be
indirect. For example, individuals who

already control one company may seek to gain

Hypotheses

 

control of a supplier to, or a customer of
that company. Even if reducing competition
were a motive, the motive would probably be
carefully concealed; it is not examined in

this study.

Following is a list of hypotheses:

1.

For a block of stock (or a type of stock)
which carries a large potential for voting
control (hereafter called control stock),
there tends to be a price premium relative
to the price of amounts (or types) of
shares which carry little potential for
voting control.

The percentage premium for control stock

is greater when management performance (as
measured by criteria such as rate of return
on equity and per share earnings growth)
has been relatively poor.

The percentage premium for control stock is
greater when there is a relatively great
possibility of profitable liquidation (as
roughly indicated by a relatively high ratio
of book value per share to market price).

The percentage premium for control stock is
greater when there is a relatively large
unused debt capacity (as roughly measured
by percent of capitalization comprised by
long-term debt).

The percentage premium for control stock is
greater when the corporation is relatively
liquid (as roughly measured by the quick
ratio or current ratio).

The percentage premium for control stock is
greater when the market price is relatively
low.

The percentage premium for control stock
will be greatest for unlisted companies and
lowest for NYSE—listed companies.

10.

11.

12.

13.

14.

15.

16.

The percentage premium for control stock
will be greater when the price of the
stock varied widely in the prior year (as
measured by the ratio of prior year's high
price to prior year's low price).

In a tender offer, the percentage premium
is greater when the market price is low in
relation to the previous year's high.

In a tender offer, the party making the
offer correctly forecasts management
endorsement of or opposition to the offer,
and the percentage premium is larger when
opposition is forecast (as revealed by sub-
sequent actual opposition).

In a tender offer, the percentage premium
is greater for offers that are relatively
successful (as measured by the ratio of
shares purchased through the offer to
shares offered to buy).

In a tender offer, the percentage premium
is greater when the length of the offer

is relatively short (as measured by number
of days from date of announcement of offer
through original expiration date of offer).

In a tender offer, the percentage premium
is greater when the proportion of shares
sought is relatively large (as measured by
shares offered to buy as percentage of
shares outstanding).

In a private purchase of a large block of
stock, the percentage premium is greater
when the block is relatively large (as
measured by the number of shares in the
block as a percentage of number of shares
outstanding).

In a private purchase of a large block of
stock, the percentage premium is greater
when the seller is a corporation than when
the seller is an individual.

In a private purchase of a large block of
stock, the percentage premium is greatest
when the purchaser is a corporation other
than the issuer, less when the purchaser
is an individual (or individuals), and
least when the purchaser is the issuer.

.e.

17. In a private purchase of a large block of
stock, the percentage premium is greatest
for those transactions followed by merger
with a corporation other than the purchaser,
less for transactions followed by merger
with the purchaser, even less for trans-
actions followed by a cash tender offer,
and least for transactions followed by no
such change in status of minority share-
holders.

18. With classified common stock, the percentage
premium of the voting shares is greater when
the ratio of the number of voting shares to
the number of non-voting (or reduced-voting)
shares is relatively low.

Variables

 

The main purpose of the study is to determine the
price premium paid for control stock and factors affecting
that premium. Therefore the dependent variable used in
studying each of the three methods of obtaining or retain—
ing voting control is derived from a price ratio: the
ratio of the price of control stock to the price of non—
control stock. The exact nature of this variable varies
with the method and is explained in the chapter discussing
each method.

Hypotheses numbers 2 through 18 listed in the
preceding section are designed to state factors which are
expected to influence the size of price ratio. In order
to test these hypotheses a number of independent or
explanatory variables are used. The independent variables
used vary for the study of each method and are listed and

discussed in the chapter describing each method.

CHAPTER II

CASH TENDER OFFERS

Review of Literature

 

Reasons for Making Offers.--In an essay on corporate
control, Oliver E. Williamson analyzes "the four incentives
for corporate control identified by Manne."

l. the possibility of realizing market power by
acquiring a rival firm

2. technical efficiency (economies of scale in
production, distribution, etc.)

3. the possibility that the management may be
incompetent or inefficient . . .

4. the incentive to capture the salaries and
perquisites of management which displacing
the incumbents would permit1
Williamson argues that incentives numbers 1 and 2 are
ordinarily obtained through merger, while there is a "wider
range" of possibilities for incentive number 3. One of
those possibilities is the cash tender offer.2

In his discussion of the cash tender offer, William-

son stresses detectable inefficiency of incumbent management

 

1Oliver E. Williamson, "Corporate Control and the
Theory of the Firm," Economic Policy and the Regglation of
Corporate Securities, ed. by Henry G. ManneTYWashington:
American Enterprise Institute for Public Policy Research,
1969). p. 308.

2Ibid., p. 309.

ll.

 

10

and introduces the concept of an inefficiency threshold.
He argues that where a desire to displace inefficient
management is the motivating factor for a tender offer, a
certain significant degree of management inefficiency must
exist before a tender offer will be made. There are a
number of possible reasons for the existence of this
threshold. There may be barriers to entry into the market
for corporate control, which, apart from size of the sub-
ject firm, have not yet been clearly defined. Secondly,
(and, Williamson feels, more importantly) there are diffi-
culties in detecting management inefficiency. Third, the
costs (other than the premium) may be substantial. These
costs include not only transaction costs, but also what he
terms transition costs. Transition costs involve the
adverse side effects of a change in management, such as
decreased performance by middle management as a consequence
of feeling threatened by the change. Further, the use of
a tender offer requires a premium that will be perceived
by existing stockholders as significantly large in order
to give a reasonable chance of having the minimum number of
shares tendered which the offerer desires.1

The management inefficiency required to make a
change of management attractive may not necessarily be due
to lack of management competence. It may also be due to
competent management that "pursues goals which are in

partial conflict with stockholder objectives."2

 

lIbid., pp. 311-316. 21bid., p. 309.

ll

Empirical Studies.--In 1967 a study of tender

 

offers was reported by Samuel L. Hayes, III and Russell
A. Taussig. The authors compiled a list of 280 external
cash tender offers made during the ten-year period from

July 1, 1956 to June 30, 1966.1

One purpose is to find characteristics which dis-
tinguish the companies which were the target of cash
tender offers from companies for which no cash tender
offer was made during the ten-year period. For this
part of the study they selected at random 50 cash tender
offers from the entire list of 280. One characteristic
of those companies is that earnings in the five-year
period before the offer had either been fluctuating
widely, trending downward, or included loss years.

Another characteristic is a relatively low rate of return
on stockholders' equity. Dividend cuts are another char-
acteristic. A relatively high quick ratio also is typical
of the target companies. The authors describe the pattern
of characteristics as representing "inept or at least
overly-conservative management." They imply that the cash
tender offers often are made because those making the
offer feel that new management should be found to try to

make more profitable use of the assets.2

 

l
S. L. Hayes, III and R. A. Taussig, "Tactics of

Cash Takeover Bids," Harvard Business Review, VL (March-
Aprilq 1967). PP. 135, 148.

2 . .
. R. A. Taussig and S. L. Hayes, "Are Cash Take-Over
Bids Uneth1cal?" Financial Analysis Journal, XXIII (January-
February, 1967), pp. 107-111.

 

12

Another objective of the Hayes and Taussig study
is to analyze factors associated with the amount of
premium--the amount by which the tender offer price
exceeded the price of the stock. In selecting the sample
for this part of the study they categorized 83 of the 280
offers as contested (identifiable opposition by incumbent
management of the target company) and selected 50 of the
contested offers for detailed study.1 Therefore, their
study of premiums excludes the large number of cash tender
offers where management of the target company either
endorsed the offer or announced no position either for or
against the offer.

Using this sample of 50 resisted cash tender
offers they note that the premiums range from zero to 44%,
with a median of 16%. They selected a number of variables
which they compare with the size of the percentage premium
through use of contingency tables. The variables include
offer success or failure, the market price of the stock,
and the length of the offer.2

On the basis of their limited analysis of size of
premium, the authors suggest the need for further research
in this area by saying "we suspect that differences in the
size of the bid premiums can be explained statistically by

a combination of variables . . . . Until someone is

 

1Hayes and Taussig, op. cit., pp. 141, 148.

2Ibid., pp. 140-41.

13

successful in isolating and precisely weighting the
critical variables, however, we conclude that the setting
of the premium, while crucial to the success of the bid,
is likely to remain an imprecise art."1

A 1971 book by George D. McCarthy and Robert E.
Healy includes a report of their study of 75 cash tender
offers for common stock of companies listed on the New
York Stock Exchange which were made during the period from
January 1, 1965 to April 28, 1967. (This period precedes
the one used for this study.) They examine characteristics
of the companies which seem to make them vulnerable to
tender offers; they also study the outcomes of the tender

offers. They calculate premiums of offer price above

market price and find a median of 21% with a range of 2%

to 53%.2

Lynn E. Dellenbarger, Jr. studied fifty mergers
occurring in the eight-year period between 1950-1957 which
involved the exchange of common stock only, and where all
constituent companies were listed on the New York Stock
Exchange prior to the merger.3 He focuses on the exchange
ratio as the dependent variable and explains the exchange

ratios through the use of multiple regression analysis.

 

lIbid., p. 148.

2George D. McCarthy and Robert E. Healy, Valuing a
Cknn an (New York, N. Y.: Ronald Press Company, 1971),
pp. -19, 322-25.

3Lynn E. Dellenbarger, Jr., Common Stock Valuation_
in Inmdustrial Mergers (Gainesville: University of Florida
Press, 1966), p. 7.

I‘
‘.q

.~.
..

14

Of the five independent variables which he uses, one is
by far the most dominant: "the ratio of the common stock
market prices for the quarter just prior to the initial
official action." Next in importance is "the ratio of the
average of earnings per share for the five years prior to
the merger." Of lesser importance are the other three
variables: "the ratio of the earnings per share in the
last full year prior to the merger," the ratio of book
values per share, and the ratio of dividends per share.
The ratio of market prices is quite highly correlated with
each of the other four independent variables, and the
author notes that this correlation may partially explain
the lack of greater dominance of these variables used in
the regression results.1

In addition to the five independent variables used
in the regression analysis, Dellenbarger also employs
five of what he terms "modifying factors": earnings
trend, earnings stability, financial leverage, cash
assets, and over-all financial strength. He uses these
factors to explain the unexplained variations resulting
from predictive use of his regression equations. He
concludes that earnings trend, and to a lesser extent
financial leverage and cash assets, do help to explain the
variations unexplained by the regression equations. How-

every these factors seem important only when there are

 

lIbid., pp. 70, 84-137.

15

extreme differences (with regard to one or more of these
factors) between the constituent corporations.

Dellenbarger does some further analysis which is of
particular relevance to this study. For his group B (a
subgroup of 45 mergers remaining after excluding 5 with a
constituent company having an earnings deficit within 5
years prior to the merger) he finds that in 36 of these 45
mergers the smaller constituent received a premium above
market price. That is, the exchange ratio was more favor-
able to the smaller company than it would have been if the
exchange ratio had been based solely on the ratio of
market prices. Further analysis of his appendix table
containing this data shows that the premiums range up to
43%, with a median of 13.5%. In 26 of these 36 cases, the
premiums are between 11% and 25%. Fifteen of the cases fall
within the relatively narrow range of 11% to 15%.2

There is a strong similarity between these premiums
found by Dellenbarger and the sizes of premiums on cash
tender offers reported by Hayes and Taussig. These pre-
miums might be regarded as a premium for control if it is
assumed that in these mergers the larger constituent com-
pany gained control of the smaller. This is probably a
realistic way of looking at the situation in most cases,
run: only from the vieWpoint of management but also from

tjue'viewpoint of stockholders. It is likely that in a

 

lIbid., pp. 95—137. 21bid., pp. 145, 154-155.

16

merger of companies of disparate size, the top management
of the smaller company, if it remains with the surviving
company at all, will find itself in a subordinate position.
The stockholders of the smaller company also might have
more at stake. In the merged surviving company the opera-
tions of the larger constituent will dominate, and the
proportionate interest in the smaller constituent will be
smaller than before the merger. Both the management and
controlling stockholders of the smaller company might
therefore demand a premium for surrendering control over
its assets and operations. Conversely, the management and
controlling stockholders of the larger company might be
willing to pay a premium to acquire control of the smaller

company.

Strategy.--In 1967 a study by Martin J. Whitman
entitled "The Strategy of Tender Solicitations" was pub-
1ished by Blair & Co., Inc., members of the New York Stock
Exchange. As a how-to-do-it manual on how to make a
tender offer succeed, it gives insight into factors other
than the size of the price premium which may affect the
outcome of the offer. One piece of advice is to avoid
active opposition by management, which may require a higher

premium, and may diminish the chances of obtaining control.l

 

1Martin J. Whitman, The Strategy of Tender Solici-
tations (New York: Blair & Co., Inc., 1967), p. 3.

 

 

17

We may not be able to tell what the offeror has in
mind by the terms of the offer; for example, although the
offeror would probably prefer not to be obligated to buy
any stock unless a certain minimum amount is tendered, he
is urged not to make such a minimum a condition of his
tender offer.1

Price premiums usually range from 10% to 30% above
the market price, but "since investors are dollar per share
conscious as well as percentage conscious, the lower the
price of the stock, the greater the percentage premium
that should be offered." Other factors to be considered
are management's attitude, stock ownership, the amount of
stock desired, the stock's price history, and the outlook

for earnings and dividends.2

Cash Tender Offers Studied

 

Fifty-three cash tender offers made for the common
stock of companies are statistically analyzed in order to
study the premium offered above market price and the
factors which help to explain the variations in the amount
of this premium.

All cash tender offers reported by Standard & Poors
Called Bond Record during all of 1968 and the first half of

1969 are examined. Of the 94 cash tender offers reported

 

11bido' pp. 4-12. 21bido ' p. 23.

18

during this eighteen months period, 53 are included in the
final analysis. A few of the 41 exclusions are made
because analysis of published information (such as state-
ments by management of the company making the offer)
suggests that the offer was made for a reason other than
an attempt to obtain or strengthen voting control. For
example, in the case of an offer of $18 per share for the
stock of Benrus watch, the offer was "a formality to comply
with the securities laws and no response is expected." It
followed an agreement by the family described as principals
of Benrus to sell all or most of their stock at $18 per
share.1 This was the only tender offer reported where a
negative premium existed: the market price was $20.
However, most of the 41 exclusions are made because
data is unobtainable for one or more of the variables used
in the study. Many are due to the lack of published daily
market price quotations for the stock; without daily price
quotations it is not possible to calculate the key variable
being studied--the ratio of the tender offer price to the

market price.

Variables Employed

 

The variables employed are listed and defined in
Table 2.1. The price ratio, which is analyzed as the
dependent variable, is the ratio of the tender offer price

to the market price (close or bid) two business days

 

1Value Line Investment Survey, 8-2-68, p. 456.

 

19

ABMEV

wofium poxsmz

 

ooflum pomeE on mnmnm mom mdam> x003 mo oaumm ou msHm> xoom Mz>m m
“can mHOMoQ How» Eoum
mmmmnowc Amy so .omcmno 0: Amy .mmmmuocfl AHV
"mumcm Mom mmcflcumm mummm Hofium mo coauownflo coflpomuflp mmcflcnmm mHQm m
Hoflum whom» m>flm can snow
mumB scans mumom 03» one How mmsflcsmw mmmum>m
on munch HOHHQ osu How mmcflcumm momsm>m mo oflumm campy mmcflcumm amem n
muflswm Rooum cOEEoo can .xooum pmuumwmsm .unmp
enmpnmcoH mo Hmpou on name enmuumcoH no capmm cappm pawn ammo w
mmeHHHQmHH pcmuuso ou manm>emoms measooom pom
mwflufiusomm waemwmxumfi EHmDIDHOSm .cmmo mo oflumm osumm xoflso UHDO m
Ammo» Hoflumv sumo: ems wo mmmucoonmm mm mEOOGH pmz eunuch mo mumm mom v
Hmwwo mo ucmsmocsocsm oHOMon
mmmp mmmcflmsn 03¢ Amen so mmoHov moflum poxnmz moflum umxmmz 9M2 m
ampmnsps Imv
.mmcmnoxm Hospo Amy .mmwz Adv "msumpm assumed mmcmnoxm moxm m
ABMSV ummmo mo paws
loocsoccm mHOMoQ mmmp mmmCHmsn 03D Amen so mmoaov
woflnm ymxnmfi mo omensmonmm mm moflsm Hmmmo Hmpcma oaumm mofium swam H
coflvflcfimmo mEmz mpoo HmQEdz
manmwum>

 

.pomoamsm moanmflum>

"mummmo mopeds cmMUII.H.N mqmde

20

 

>53 Op pmnmmwo mwumcm
Op nowwo emsoucp pmmmcousm moumnm mo ospmm

masocmwwuso
mmnmnm mo mmmwcmoumm mm >53 ou Umnmmmo mononm

Hmwwo mo dump soapmsflmxm Hmcflmflso nmsounu Hommo
mo pcoEoocsoccm mo memo Eouw mmmp mo sonfisz

coauflmommo Amy Ho hpaamuusmc Amy
ucoEmmH0pcm AHV "Hmmwo on coauomms m.ucmEmmmcmz

moflum 30H
m.ummm HOHHQ ou moflwm amen m.umo> uoflum mo oflpmm

Aessv
moflum umxums on moflsm 30H: m.umm> Hoflum mo osumm

pomoaousm mmumcm

unmsom mmHMSm

smmmo mo zumcmq
cofluomms
m.ucosmmmcmz
moans mOHHm

moaum
smegma on gene

BOO

8243

SEGA

mBUZ

Ude

MZHE

ma

vH

ma

NH

Ha

OH

 

A. a

5.

 

21

before announcement of the offer. This variable is a ratio
converted to percentage form. To get the actual premium
100% is deducted from the value of this variable. The two
day period is Chosen for two reasons. First, it is reported
as having been used in previous research on the size of
tender offer premiums.1 The use of the same period facili-
tates comparison of the size of premiums found with those
findings. Secondly, two days appears to be a reasonable
compromise in dealing with two conflicting risks: the risk
of a very short period is that word of the offer may have
gotten out prematurely and caused the market price to start
to reflect the higher tender offer price; the risk of a
longer period is that factors other than the tender offer
might have influenced the market price between the date
chosen and the date of announcement of the offer.

In order to attempt to explain the variation in the
size of the premium, 14 independent or predictor variables
are used. These variables were arrived at by considering
suggestions by other writers as to what might be important
and by examination of a number of other possibilities sug-
gested by financial theory. Selection was made on the
basis of a desire for wide coverage, plus a desire to avoid
duplication or overlap which might subsequently be reflected
in high simple correlation between two or more of the
independent variables. In order to avoid the possibility

of impaired reliability, the independent variables were

 

1Hayes and Taussig, op. cit., p. 140.

22

selected solely on the basis of logic; such factors as
degree of simple correlation with the dependent variable

were not utilized in the selection.1

Method of Analysis

 

A major goal of this study is to determine the
extent to which each independent variable explains the
variation in the dependent variable. In each case the
dependent variable has been influenced by many variables
acting in concert. Therefore, in order to study the
influence of any individual variable, it is necessary to
hold all of the other independent variables constant--thus
eliminating the effects of differences in the values of
each of these other variables. One method for conducting

this type of analysis is multiple linear regression.2

Regression Analysis.--The major output of multiple

 

linear regression is a prediction equation called a multi-
ple regression equation. The predicted value of the
dependent variable is on one side of the equation; on the
other side are a constant plus two or more other terms,
each of which consists of an independent variable multiplied
by the net regression coefficient for that variable.

For predictive purposes it is this multiple regres-
sion equation that would be used. In order to predict the

value of the dependent variable for a given observation, the

 

lM. Ezekiel and K. A. Fox, Methods of Correlation and
Regression Analysis (3rd ed.; New York: J6hn Wiley and Sons,
Inc., 1959). pp. 186, 195-196, 436.

 

 

2Ibid., pp. 151-52.

«it

Ԥ

\

 

23

value of the independent variables for that observation
would be plugged into the equation.

However, for measuring the relative contribution of
each independent variable in accounting for the variation
in the dependent variable, the equation is not directly
useful. The net regression coefficients cannot be mean-
ingfully compared with each other because the size of each
is influenced by the unit of measure employed for that
variable. For example, if height is an independent
variable, the use of inches instead of feet as the unit of
measure would greatly affect the size of the net regression
coefficient for that variable.

In order to make possible comparison of the rela-
tive contribution of each independent variable, beta
coefficients are calculated; the net regression coefficients
are normalized through adjustment by standard deviations
of the dependent variable and independent variable.1 The
resulting beta coefficients for each independent variable

may then be meaningfully compared.

Stepwise Procedure.--A number of methods are avail-

 

able for selecting the regression equation that best fits
the observations of the sample being studied. One of these
methods, stepwise linear regression is recommended by the

authors of one book on the subject as being the best.2

 

lIbid., p. 196.

2N. R. Draper and H. Smith, Applied Regression
Analysis (New York: Wiley, 1966): pp. 163-177.

 

24

It is a variation of the forward selection procedure. In
both of these methods variables are inserted one at a

time until the regression equation is satisfactory. The
order of insertion is determined by selecting for insertion
at each step the independent variable with the largest
partial correlation coefficient with the dependent variable;
this selection is made by trying all possibilities. When no
variable remains which meets the criterion established for
entry, the procedure terminates. The stepwise procedure
differs by re-examining at every step each variable entered
at an earlier step; it may no longer make an important
contribution because of its relationships with other
variables now in the regression; in that event, the vari-

able is removed from the regression equation.1

Two Groups.--The 53 cash tender offers used

 

consisted of 41 for the stock of non-financial corporations
and 12 for the stock of financial corporations. The latter
are insurance companies, with the exception of one finance
company and one mutual fund management company. The
presence of the financial companies poses two problems.

One problem is that two of the independent variables used,
QUIC (Quick Ratio) and DEBT (Debt Ratio) are considered

not applicable to the financial companies. The second
problem is that the presence of the financial companies

may be unacceptably destructive of the homogeneity of the

sample. These two problems are solved by using two groups

 

1
Ibid.

25

for the computer analysis. Group A includes all 53 obser-
vations. For this group the variables QUIC and DEBT are
omitted from the analysis. Group B is a sub-group of
Group A. It includes only the 41 observations of non-
financial corporations; in the analysis of this group all

variables are used, including QUIC and DEBT.

Findings

Price Ratio.--Table 2.2 shows the distribution of

 

the values of price ratio for group A, which includes all
53 cash tender offers studied. In terms of premiums
(price ratio minus 100%) the median is 16.7% and the mean
18.2%. There are three offers with premiums under 5% and
three of 40% or over. The range is from 1.9% to 46.7%.
Regression analysis is used to explain this variation in

the independent variable.

Regression Steps.--Table 2.3 shows certain informa-

 

tion regarding each of the nine regression steps for Group
A. The variable which is entered at each step is shown,
along with information on the cumulative effect of all
variables entered through that point. The nine steps
indicate that nine of the 12 independent variables used

for Group A made a significant contribution. The second
column shows the variables that were entered and the order
of their entry. The third column (R2) shows the cumulative

extent to which the variables in the equation at that step

26

TABLE 2.2.--Distribution of Actual Values of Price Ratio

 

Group A.
Actual Value of Number of
Price Ratio (percent) Observations

 

100.0 - 104.9
105.0 - 109.9

110.0 - 114.9 11
115.0 - 119.9 14
120.0 - 124.9 6
125.0 - 129.9 2
130.0 - 134.9 3
135.0 - 139.9 2
140.0 - 144.9 2
145.0 - 149.9 1

Total 53

 

Range = 101.9% to 146.7%
Median = 116.7%
Means = 118.2%

explain the variance in the dependent variable. As each
additional variable was entered, the value of R2 improved
from .08 (with only the independent variable GOT) to

.32 (with all nine variables). The next column shows how
the constant varies from one regression equation to
another; of course, the regression coefficients vary from
one step to another also, but they are not shown in this
table. The last column is calculated by the F-test for
significance of regression. The value, which varies

between 2.5% and 5.0% means that there is no more than a

27

TABLE 2.3.--Regression Steps,Stepwise Addition of Variables,

 

 

Group A.
Step Variable 2 Significance
Number Entered R Constant Level
1 GOT - .08 122.1 5.0%
2 LGTH .12 125.1 5.0%
3 RANG .19 113.5 2.5%
4 MKT .21 117.5 2.5%
5 BVMK .24 122.3 2.5%
6 EXCH .26 118.7 2.5%
7 EDIR .29 113.4 2.5%
8 HIMK .31 120.0 5.0%
9 MGTR .32 117.8 5.0%

 

5% probability that the regression results could have been
obtained by random sampling from the same pOpulation.l
Table 2.4 shows the same type of information for
Group B as Table 2.3 does for Group A. The Group B
regression also terminated with nine variables in the
regression equation, even though the number of independent
variables available for Group B was l4--two greater than
for Group A. The results in Group B are better than Group
A in two respects. First the value of R2 attained is
higher--.43 compared with .32. Second, Group B's final
regression equation has a higher level of significance--

2.5% versus 5.0%. Further, with Group B if one is willing

to settle for an R2 of .38 instead of the .43 obtained in

 

lEzekiel and Fox, 0p. cit., p. 397.

28

TABLE 2.4.--Regression Steps,Stepwise Addition of Variables,

 

 

Group B.
Step Variable‘ 2 Significance
Number Entered R Constant Level
1 GOT .14 122.6 2.5%
2 RANG .23 110.8 1.0%
3 LGTH .30 110.8 0.5%
4 MKT .36 116.9 0.5%
5 BVMK .38 121.6 0.5%
6 EXCH .39 118.4 1.0%
7 EDIR .41 114.2 1.0%
8 HIMK .42 118.2 2.5%
9 WANT .43 120.6 2.5%

 

Step No. 9, he can use the regression equation at Step No.

5 and obtain a significance level of 0.5%.

Regression Results.--Table 2.5 shows the regres-

 

sion equations achieved in the final step for both Group A
and Group B (step no. 9 in both cases). The regression
equation consists of the constant and each regression
coefficient of each variable multiplied by the value of

that variable. For Group A the equation takes the following

form:

PREM = 117.77 - 3.07 GOT + 1.81 EXCH - 0.06
MKT + 2.01 EDIR - 4.21 BVMK - 7.57
HIMK + 6.92 RANG + 1.11 MGTR - 0.28

LGTH

It \s-\

s \

h u

6

\>\;\\.

‘.

 

mm» o: mam>md wm.~
up unmoamsamam

no» mm> mam>wa mm
up ucmoauflcmflm
mm.m mm.m msam> m
uoUGMUAMHcmHm mo umou m
ma. mm. Emmy coaupcas

luouwp mamauasE
mo usmHOHmmmou

 

 

 

mm.oma 55.5HH ucmumcou

oz ma.o: mo.on Bzmz

mow mm.o| ma.ou vm.o| mm.on mama

mm». OH .0 :7” menu:

new mm.o Hone mm.o mm.m oz¢m

m oz NH.O- mm.a- sa.o- sm.s- msz

oz ma.ou mm.m| mH.o| Hm.v| Mz>m

mow mH.o Hm.H om.o Ho.m mHom

mom

mo» mm.on mo.on mH.ou mo.ou 9x2

mow ma.o mm.H ma.o Hm.H muxm

oz nm.ou mv.v| mH.o| no.m| 900

mmflmwnuomwm unmanamwooo unmaoflmmooo unmaoawwmou ucmfloammmoo manoeum>

sues comm coflmmmummm mumm coammmummm ucmpcmmmch
ucoumflmcoo

mmflcmmsoo whoaum>ummno Had "4 msouu

HMHUCMCHMICOZ "m msoum

 

.maemflum> ucoocomoa ma Ewan .mflmmamc< c0flmmoumom mo mpHSmomll.m.m mqmds

30

For Group B the regression equation is:

PREM = 120.58 - 4.45 GOT + 1.32 EXCH - 0.08 MKT

+ 1.31 EDIR - 5.58 BVMK - 4.28 HIMK

 

+ 7.01 RANG — 0.18 LGTH - 0.03 WANT 5

Because the regression coefficient for each variable
is influenced by the units in which that variable is mea-
sured, the beta coefficients are more useful for clues as to
the relative importance of each variable in the regression
equation. The association between each independent variable
and the dependent variable (PREM) is positive unless a
minus sign appears before the beta coefficient, in which
case it is negative. The last column shows whether or not
the direction of this association is consistent with the
hypothesis concerning that variable.

Of the ten independent variables which entered at
least one of the two regression equations, two show
moderate association, two weak association, and six very
weak association. The two variables showing moderate
association are length of offer (LGTH) and price range
(RANG). Length of offer shows a moderate negative associa-
tion with price ratio, which is in the direction expected.
One hypothesis is that the percentage premium is greater
when the length of offer is relatively short (as measured
by number of days from date of announcement of offer

through original expiration date of offer).

31

Price range (RANG), the other variable showing a
moderate association with the price ratio, has a positive
direction of association which is consistent with one of
the hypotheses: that the percentage premium for control
stock is greater when the price of the stock varied widely
in the prior year as measured by this ratio.

The two variables showing weak association are
shares purchased (GOT) and market price (MKT). Shares
purchased is negatively associated with price ratio, which
is inconsistent with one of the hypotheses: that the per-
centage premium is greater for offers that are relatively
successful (as measured by the ratio of shares purchased
through the offer to shares offered to buy). Market price
shows weak negative association with price ratio, Which is
consistent with the hypothesis that the percentage premium
for control stock is greater when the market price is
relatively low.

Of the six variables entering at least one of the
two regression equations and showing only very weak asso-
ciation with the price ratio, three show association in
the direction consistent with a hypothesis; these three
are earnings direction (EDIR), exchange (EXCH), and manage-
ment reaction (MGTR). The three independent variables
showing very weak association in the direction inconsistent
with a hypothesis are book value to market price (BVMK),

high to market price (HIMK), and shares sought (WANT).

32

In summary, both of the variables showing moderate
association do so in directions consistent with their
hypotheses; one of the two variables showing weak associa-
tion does; and three of the six variables showing very
weak association are consistent with their hypotheses.
Four variables do not show sufficient association with
price ratio to enter either regression equation: rate of
return, earnings trend, quick ratio, and debt ratio.
(Quick ratio and debt ratio have an opportunity to enter
only one of the two regression equations--that for group

A.)

Analysis of Residuals.--The next step is analysis

 

of the residuals, or unexplained variations. One method
is analyzing the out-on-a-limb predictions made by the
regression equation. The four predictions for Group A of
less than 110% (observations 13, 29, 30, and 83) and the
four predictions of 125% and higher (observations 15, 17,
23, and 24) are shown in Table 2.6. Of the four large
predictions, two (23 and 24) do not go far enough, while
the other two (15 and 17) go too far. Of the four small
predictions, one (30) does not go far enough, while three
(13, 29, and 83) go too far.

Another method of analyzing the residuals is to
look at the observations with the largest residuals and
try to see what may have been the cause. Table 2.7 shows
the four observations for Group A with unexplained varia-

tions of 13.0% or more. None of these cases involves the

33

TABLE 2.6.--Unexp1ained Variations for Observations with
Predicted Value of Price Ratio Under 110.0%
or Over 124.9%, Group A.

 

 

Observation Actual Predicted Unexplained
Number* Value Value Variation
% % %

13 114.8 109.1 5.7

15 ' 121.7 125.5 -3.8

17 119.3 126.0 ~6.7

23 140.4 131.8 8.6

24 139.1 135.3 3.8

29 116.7 109.7 7.0

30 106.7 107.9 -l.3

83 118.2 109.2 9.0
Median Unexplained Variation (disregarding sign) = 6.2%
Mean Unexplained Variation (disregarding sign) = 5.7%

 

*Companies are identified by observation number in Appendix
table.

TABLE 2.7.--Observations with Unexplained Variations of
13.0% or More from Value of Price Ratio
Predicted by Regression Equation for Group A.

 

 

Observation Actual Predicted Unexplained
Number* Value Value Variation
% % %
1 101.9 117.7 -15.8
21 107.5 121.9 -14.4
80 146.7 119.2 27.5
82 144.4 121.9 22.5

 

*Companies are identified by observation number in Appendix

table.

34

out-on-a-limb predictions shown in Table 2.6 and previously
discussed. On the contrary, these four cases have predicted
values grouped relatively closely around the mean actual
value of PREM (118.2%)--with a range of 117.7% to 121.9%.
Three of these four cases are due to extreme actual values
of PREM: under 105% or 140% or over; only six observations
from all of group A have such extreme values. In the other
case, the direction of prediction from the mean is wrong;

in observation number 21 the actual value of PREM is
relatively low (107.5%) but a moderately high prediction

is made (119.2%).

Simple Correlations.--Simple correlations between

 

pairs of variables are of interest largely for purposes

of considering why certain independent variables do not
enter a regression equation. Table 2.8 shows such corre-
lations of .40 or higher. Three of these variables do not
enter the regression equation for the group for which they
show simple correlation of .40 or higher with another
independent variable which does enter that regression
equation. Rate of return (ROR), which enters neither
equation, shows high correlation (-.58 and -.80) with

book value to market value (BVMK) which enters the equation
for both groups. Quick ratio (QUIC) shows a .50 correlation
with price range (RANG) for group B; quick ratio does not
enter the equation for this group (the only group for which,

it is used) and price range does. Shares sought (WANT)

35

TABLE 2.8.--Pairs of Variables Having Simple Correlations
of .40 or Higher with Each Other.

 

 

 

Correlation
Variables Group A Group B
ROR with BVMK -.58 -.80
EXCH with MKT -.45
QUIC with RANG NA .50
WANT with MGTR -.42
WANT with LGTH .43

 

shows a -.42 correlation with management reaction (MGTR)
for group B; management reaction does not enter that

equation, while shares sought does.

Price Action Prior to Announcement.--In the dis-

 

cussion of variables employed, one shortcoming mentioned

of using market price two business days before announcement
of the offer, is the risk that word of the offer may have
gotten out prematurely, causing the market price to start
to reflect the higher tender offer price. Table 2.9 shows
that during the five day period between one week before
announcement there were 38 instances of gains (percentages
under 100) compared with 15 of decline or no change (per-
centages of 100 or over). For 12 offers the rise was

6.5% or more (100 + 93.9 = 106.5%). In contrast only

three offers show a price drop of 4% or more. However this

greater frequency of gain seems to be largely attributable

\ .

.\.-

36

TABLE 2.9.--Market Price One Week Before Announcement of
Offer as Percentage of Market Price Two Days
Before (Number of Offers).

 

Market Direction During Five Day Period*

 

 

 

 

 

 

Percentage Total Rally Decline Mixed
104.0 and over 3 3

102.0 to 103.9 2 1

100.0 to 101.9 6 3

- 98.0 to 99.9 10 6 3 1
96.0 to 97.9 9 7 2
94.0 to 95.9 7 4 2 1
92.0 to 93.9 5 5
90.0 to 91.9 4 4

Under 90.0 3 2 1

Total 53 39 12 2

 

*"Trading Swings, Based on 425 Industrials," Standard &
Poor's Stock Guide, Year End, 1968, and Year End, 1969,
p. 7.

 

 

to the fact that 39 of these five day periods occurred
during periods of general market advance, as compared with
12 occurring during periods of market decline (as measured
by Standard & Poor's index of 425 industrials).1

The greater frequency of occurrence of offers dur-
ing rallies during the period studied is due to two
factors. First, there were 337 elapsed days of rallies

and only 209 elapsed days of declines.2 Second, the

 

l"Trading Swings, Based on 425 Industrials," Standard
& Poor's Stock Guide, Year End, 1968, and Year End, 1 6 ,
p. 7.

21bid.

37

relative frequency of announcements of offers during
rallies was 1.73 times that during declines. (There were
39 announcements during rallies, and 14 during declines).
In order for the predominance of price gains during
the five day periods to be attributed to the preponderance
of market rallies, there should be statistically signifi-
cant association between the percentages and whether they
occurred during rallies or during declines. Data abstracted
from Table 2.9 shows that for periods during rallies there
are 17 with percentages of 98 and over, and 22 under 98;
for periods during declines there are 7 with percentages
of 98 and over, and 5 under 98. The value of Chi square
calculated for a 2 by 2 table (not shown) including this
data is .801, indicating that the association is not
significant at the .10 level. However, this lack of
statistical significance may be due to the fact that the
classification rallies versus declines reflects a stock
market index, while the direction of movement of individual
stocks (even those in the index) is much more diverse;
during a rally many stocks reach their individual peaks

and start declining before the index starts to decline.

Discussion of Findings

 

Both regression equations explain a substantial
amount of the variation in the price ratios of the cash
tender offers and do so with very satisfactory significance

levels. The better of the two equations, that for Group B

38

(non-financial companies) has an R2 of .43 and is signifi-
cant at the 2.5% level.

Six hypotheses are supported in this part of the
study by association in the expected direction between
independent variables and the price ratio. Length of offer
and price range hypotheses are supported with moderate
association. Market price shows weak association; three
variables show very weak association: earnings direction,

exchange, and management reaction.

Length of Offer.--The moderate negative association

 

between length of offer and price ratio means that a
relatively long length of offer (number of days from date
of announcement of the offer through the original expira-
tion date) is associated with a relatively small premium.
The rationale for this hypothesis is that a short offer
length will require a relatively large premium to move
stockholders to act rapidly. A related possibility is that
a large premium and short offer length may sometimes be
used together to achieve a blitz effect and give the

Opposition little time to act.1

Price Range.--Price range shows a moderate positive
association with price ratio which means that when the
prior year's high price was high in relation to the prior

year's low price, the premium is relatively high. A

 

lWhitman, op. cit., pp. 23-24.

39

relatively-wide prioryear's price range could indicate a
relatively-high degree of general investor uncertainty
regarding the value of the shares. Also a wide price range
may make an individual stockholder more uncertain as to

what his stock is worth. Therefore, he will be more uncer-
tain as to whether a given tender offer premium is attrac-
tive. A likely response to this uncertainty is to do
nothing--which means not tendering his shares. A relatively-
large premium is therefore required to overcome this

uncertainty.

Market Price.--Market price shows a weak negative

 

association with price ratio. The negative sign indicates
that the higher the market price the lower is the premium
(which is expressed as a percentage of market value). This
finding is consistent with what Hayes and Taussig reported
having been told by investment bankers, and their findings
using a contingency table analysis.1 The reasoning in
support of a higher percentage premium when the market
price is relatively low is that the stockholder trying to
decide whether or not to tender his shares tends to think
of number of dollars premium per share as well as the per-
centage premium. If so, a 10% premium for a stock with a
market price of $8.00 would give a premium of $.80 and
this might appear less adequate than a 10% premium for a

stock with a market price of $100, which would be $10.00.

 

1Hayes and Taussig, op. cit., p. 140.

4O

Whitman also suggested this tendency in discussing
strategy.1

If many investors do behave in this manner it
would appear to be somewhat irrational. Indeed, a case
can be made for being willing to accept a smaller per-
centage premium on stocks with a lower market price.
This case is based on the fact that the tender price is
all net to the stockholder, while if he sells his stock
on the market he must pay a commission. Assuming that he
sells one or more round lots, his commission will be
higher as a percentage of the price when the price is
low. For example, neglecting the surcharge, which was
not in effect during the period under study, the commis-
sion on a 100 share round lot sale at $8.00 per share
would be $15.00, which is 1.875%. The commission on a
100 share round lot sale at $100 per share would be
$49.00, which is .49%. Thus one could argue that a
rational investor would be willing to accept a premium
1.385 percentage points lower on an $8.00 stock than on
a $100 stock.

McCarthy and Healy do not analyze the relation-
ship between market price and price ratio. However
their tabular presentation of data for the 75 offers they
studied makes possible c0ntingency table analysis similar

to that used by Hayes and Taussig for these variables. A

 

1Whitman, op. cit., p. 23.

41

contingency table (not shown) constructed from their
data shows that of the 32 offers with premiums under 20%,
16 had market prices of $30 and over, and 16 under $30.
Of the 43 offers with premiums of 20% and over, 19 had
market prices of $30 and over, and 24 under $30.1 The
value of Chi square calculated from this table is .249,

showing that association between the variables is not

statistically significant, even at the .50 level.

Earnings Direction.--Earnings direction shows a

 

very weak positive association with price ratio. This
variable has three possible values: a value of 1 indicates
an earnings per share increase from the previous year, a
value of 2 no change, and a value of 3 a decrease from

the previous year. Therefore the positive association

means that a relatively large premium is associated with

a decrease in earnings per share. This effect is consistent
with the hypothesis that the percentage premium for control
stock will be greater when management performance (as
measured by criteria such as rate of return on equity and
per share earnings growth) has been relatively poor. The
rationale of this hypothesis is that the ability to change
management (which voting control confers) is more sought
after when performance has been relatively poor. The
investor who can get control of the company is then impelled

to do so in order to change management policies and try to

 

1McCarthy and Healy, op. cit., pp. 322-25.

 

 

 

 

 

42

improve performance, and thereby increase the value of

his shares.

Exchange.--Exchange has a very weak positive asso-
ciation with price ratio. This variable has three possible
values, with the lowest for stocks listed on the New York
Stock Exchange, the intermediate value for those listed on
other exchanges, and the highest value for unlisted stocks.
The positive association of this variable with price ratio
therefore means that tender offers for unlisted stocks
tend to have higher premiums than offers for NYSE--listed
stocks.

One possible reason is that exchanges other than
NYSE are regarded as less-efficient markets than NYSE, and
in turn the over-the-counter market for unlisted securities
is regarded as less-efficient than any of the stock
exchanges. The market price at any given moment may be
regarded as a better measure of the actual value of the
stock to the extent that the market price has been deter-
mined in a more-efficient market. NYSE market prices would
be regarded as better indicators of value than market prices
on other exchanges, and prices on these other exchanges
would be regarded as better indicators of value than over-
the-counter prices. Since the purpose of the premium is
to convince stockholders that they are being offered more
for their shares than they are worth on the market, a lower
premium would be required above the more highly-regarded

NYSE market prices.

43

An additional possible factor in the association
is that for unlisted stocks the bid price was used, while
for listed stocks the actual closing transaction price
was used. To the extent that there is a spread between
the bid and asked prices of unlisted stocks, the use of
the bid price would tend to understate the market price

and thereby overstate the premium.

Management Reaction.--Management reaction has a

 

very weak positive association with price ratio. One
hypothesis (with which the direction of association is
consistent) is that the party making the tender offer

will correctly forecast management endorsement of or
opposition to the offer, and that the premium will be
larger when opposition is forecast. A possible reason
that the association is not stronger is that it is based
on announced opposition by management. In many cases it
may be known that management is opposed, and the premium
increased accordingly, but opposition is not announced.
McCarthy and Healy lend credence to this possibility by
suggesting that a management which does not control a sub-
stantial amount of stock in its company, and whose per-
formance has been poor, is necessarily restrained in any
resistance, because the chances of success of the offer are
high and "a belligerent attitude could lead to the speedy

unemployment of incumbent management."l

 

lIbid., p. 311.

44

Out—on-a-Limb Predictions.--The regression equations

 

make out-on-a limb predictions which fare quite well. Of
eight predicted values of price ratio under 110.0 or over
124.9 for Group A, only one is in the wrong direction from
the mean of actual values of price ratio. Further, the
largest residual, or unexplained value, of any of these

eight predictions is 9.0%.

Price Action Prior to Announcement.--The use of
market price two business days before the date of announce-
ment of offer appears to be satisfactory for calculating
the price ratio. For 12 of the 53 offers the rise in
price in the five day period, from one week before announce-
ment of offer to two days before, was 6.5% or more. These
price changes seem to be related to general market changes
during the same period.

Only two of the nine residuals of ten or more for
Group A (Table 2.7) seem likely to be attributable to word
of the offer getting out before the close of trading on
the second business day before the date of announcement.
Observation number 4 rose 17.8% in price in the five day
period despite the fact that it was a time of general
market decline. It would appear that word of the offer
may have gotten out in advance. Had it not been for this
price rise, the price ratio would have been 125.8% instead
of the 106.8% actually calculated. This would also explain

observation 4 having a residual of ~12.5% resulting from a

45

predicted value of 119.2% (Table 2.7). Observation number
72 is another offer on which advance word may have gotten
out, although the rise was not nearly as large as for
number 4. The rise during the five day period was 6.9%,
and this occurred during a period of mild market rally.
Had the price not risen the price ratio would have been
lll.7%--much closer to the 115.2% predicted than the
actually calculated 104.5% which resulted in a residual

of -10.8% (Table 2.7).

Book Value to Market Price.--Four independent
variables which entered one or both regression equations
(for Groups A and B) show association with price ratio
in the direction Opposite to that expected--inconsistent
with the hypothesis that each variable is intended to test.
Shares purchased shows weak association with price ratio;
very weak association is shown by book value to market
price, high to market price, and shares sought.

Book value to market price shows a very weak nega-
tive association with price ratio which means that a high
ratio of book value to market price is associated with a
relatively low premium. This result is opposite to that
expected. It has been said that a company interested in
acquiring control of another company looks for a situation
with a high ratio of book value to market price. McCarthy
and Healy point out that 29 of 70 companies subjected to

Offers had a book value per share higher than market

46

price.1 However, the rejection of this hypothesis is
consistent with the findings of Hayes and Taussig that
such companies seem no more subject to cash tender offers

than companies with a low ratio.2

High to Market Price.--High to market price has a

 

very weak association with price ratio, which means that
when the market price is low in relation to the prior
year's high price, the premium is relatively low. The
opposite had been expected. The rationale of the hypothesis
is that a given premium over the current market price will
look less attractive to the stockholder if the market price
is low in relation to the previous year's high. The stock-
holder may feel that "it will come back." On the other
hand, there is a possible explanation for the opposite
result obtained. If the market price is low in relation

to the prior year's high, the stockholder may be dis-
couraged with the stock and may welcome the opportunity

to dispose of it at even a small premium above the current

market price.

Shares Sought.--Shares sought has a very weak

 

negative association with price ratio, which means that
the higher the percentage of outstanding shares sought,

the lower is the premium. At first this statement appears

 

1McCarthy and Healy, op. cit., p. 315.

2Taussig and Hayes, op. cit., p. 140.

l,‘

47

to be illogical--instead we might expect that the more
shares we seek the higher the premium we would need to
offer. However, there is not necessarily a positive
association between the percentage of shares offered to

buy and the percentage desired to buy. On the contrary,
one source of strategy suggests that it can be a mistake

to offer to buy as many shares as you want; instead it

may be better to offer to buy a much smaller number and
reserve the right to buy more shares if more are tendered.l
Therefore, some offers for a low percentage of shares may

actually indicate great desire to have shares tendered

and may therefore involve a relatively high premium.

Shares Purchased.--Shares purchased shows a weak

 

negative association with price ratio, which means that

a relatively high premium is associated with a relatively
glow ratio of the number of shares purchased through the
offer to the number of shares offered to buy. The opposite
was expected: that a relatively high percentage premium
will be associated with offers that are relatively success-
ful. The explanation for the unexpected negative associa-
tion may lie in the denominator of the fraction used to
compute shares purchased: the number of shares offered to
buy. Here is the same problem already discussed in connec-
tion with shares sought: there is not necessarily a

positive association between the percentage of shares

 

lWhitman, o . cit., p. 11.

48

offered to buy and the percentage desired to buy. There
may actually be a negative association; some offers for
a low percentage of shares may actually indicate great

desire to have shares tendered.

Rate of Return.--Four independent variables do not

 

contribute enough to explaining the price ratio to enter
either regression equation. These four variables are rate
of return, earnings trend, quick ratio, and debt ratio.
(The last two are tested only with group B.)

Rate of return does not enter the regression
equation for either group A or group B. One hypothesis
is that rate of return on equity is one of the measures
of management performance, and that a poor performance,
as measured by a relatively low rate of return would be
associated with a relatively high premium. As indicated
in the discussion of earnings direction, the rationale
of this hypothesis is that the ability to change manage-
ment (which voting control confers) is more sought after
when performance has been relatively poor. Earnings dir-
ection gives very weak support for this hypothesis. Why
does rate of return fail to give any support? One possible
explanation is the high intercorrelation of rate of return
with another independent variable, book value to market
price. By far the highest correlation between any two
variables is between these two; the figure is -.58 for
group A and -.80 for group B (Table 2.8). Intercorrelation

between the variables may reduce the additional contribution

49

of one variable; the first of these two variables entered
may provide a significant contribution; then, when the
second variable is tried with the first already in the
equation, the second does not provide a sufficient addi—
tional contribution to be entered.1

The question remains as to why rate of return and
book value to market value are so highly negatively corre-
lated with each other. Considerable negative correlation
would be expected for the following reasons: (1) The
denominator in the fraction used to calculate rate of
return (net worth) is closely related to the numerator in
the fraction used to calculate book value to market price
(book value per share). Given the absence of preferred
stock, these two items will vary directly. (2) The
denominator of book value to market price (market price)
would be expected to vary (in the same direction, but not
necessarily the same degree) with the numerator of rate
of return (net income). Market price would be low in
relation to book value when the rate of return is rela-
tively poor; if net assets are earning a relatively low
rate of return, then a relatively low market evaluation

should be placed on those net assets.

 

1Draper and Smith, op. cit., pp. 169-72. Frederick
C. Mills, Statistical Methods (3rd ed.; New York: Henry
Holt & Co., 1955), pp. 648-52.

 

50

Earnings Trend.--Earnings trend enters neither the

 

group A nor group B regression equations. The purpose of
this variable is to measure earnings trend over a five

year period. One hypothesis is that the percentage premium
for control stock will be greater when management per-
formance, as measured by a variable such as earnings

trend, is relatively poor. As mentioned in the discussion
of earnings direction, the rationale of this hypothesis is
that the ability to change management (which voting control
confers) is more sought after when performance has been
relatively poor. Unlike earnings trend, earnings direction
did show a very weak positive association with price ratio.
Why did earnings direction enter in while earnings trend
did not? It is possible that earnings direction is a

more sensitive indicator of cause for stockholder dis-
appointment. Year-to-year earnings comparisons receive
great publicity and apparently considerable emphasis by
many investors. Consequently a decrease of earnings per
share from the previous year might be expected to have a
more measurable impact than the longer-term comparison

which is provided by earnings trend.

Quick Ratio.--Quick ratio is another variable which

 

does not enter the regression equation for group B. (It
is not used in analyzing group A for reasons explained
under "Method of Analysis.") One hypothesis is that the

premium for control stock will be relatively high when the

51

corporation is relatively liquid (as measured by the quick
ratio). The quick ratio was selected as the one variable
to measure liquidity because it seemed more satisfactory
than other measures for which data would be available for
all companies. Also, it is the measure used by Hayes and
Taussig in finding that takeover—prone companies had
"surplus liquid assets."1 In addition to the shortcomings
of the quick ratio as a measure of liquidity, another
possible explanation for its lack of contribution is its
intercorrelation with price range (Table 2.8). At .50
this intercorrelation is second in group B to the -.80
between rate of return and book value to market price,
which has been discussed. Another reason is that there
may be offsetting forces at work with regard to this
variable's association with price ratio. In addition to
the influence already discussed, it is possible that a
low quick ratio, perhaps below some level popularly con-
sidered normal, may be regarded as a warning sign of
possible failure, and that this warning sign will impel
stockholders to try to get control of the company to try
to save it and their investments; the result would be a
tendency to pay a higher premium for those companies with

a low quick ratio.

 

1Hayes and Taussig, op. cit., p. 142.

52

Debt Ratio.--Debt ratio does not enter the regres-

 

sion equation for group B. (Like quick ratio it was not
used for the analysis of group A.) One hypothesis is

that the percentage premium for control stock will be
greater when there is a relatively large unused debt,
capacity (as roughly measured by percent of capitalization
comprised by long-term debt). Hayes and Taussig used a
similar hypothesis with regard to vulnerability to takeover
attempts, but they also found no statistical support for
it.1

In addition to the limitations of this ratio as
a measure of unused debt capacity, there is also the
possible problem of offsetting forces, as in the case of
the quick ratio. A debt ratio above some level may be
considered a warning sign by investors who may thereby be
impelled to get control of the company to protect their
investments.

Another type of offsetting force with regard to
debt ratio is the possibility that investors will not
regard a high debt ratio company as unattractive for take-
over because it lacks excess debt capacity which could be
put to use to borrow money and attempt thereby to improve
earnings per share through financial leverage; instead
they will regard the company as attractive for takeover

because of the financial leverage which already exists;

 

lIbid.

53

this possibility is greatest when unfavorable financial
leverage exists, because the incentive for turning the
company around and changing the unfavorable leverage to
favorable leverage is greater than if no financial lever-
age exists. The result of this offsetting force would be

a tendency for positive association between debt ratio

and price ratio, that is a higher premium paid for companies

with a relatively high debt ratio.

Extreme Values.--Extreme values of the price ratio

 

pose problems for the predictions of the regression equa-
tion. Group A has six observations with price ratios

under 105% or 140% or over (Table 2.2). Three of these six
have residuals (unexplained variations) of 13% or more
(Table 2.7). This phenomenon would be expected in using
regression equations with a relatively low R2 (group A

is .32, group B .43). Since the equations can explain

only part of the variation in price ratios, they do poorest
at identifying the relatively few extreme values of this

variable.

CHAPTER III

PRIVATE PURCHASES OF LARGE BLOCKS OF STOCK

The private purchase of a large block of stock
from an individual stockholder appears on the surface to
be an ideal way to obtain voting control of a corporation,
particularly under certain circumstances.

One advantage is that it can usually be done with
much less publicity than the cash tender offer or proxy
contest, which require wide publicity. Negotiations for
a large block can be carried on privately, and if the
negotiations fail, may never be disclosed. If the sale is
completed, the transaction is likely to be required to be
reported to the Securities and Exchange commission for

publication in its monthly Official Summary of Securipy

 

Transactions and Holdings.. Such a report is required if

 

either the buyer or the seller of the block has been or
becomes (through purchase of the block) an insider, defined
as an officer or director or holder of 10% or more of any
class of the corporation's equity securities.

If such disclosure is required, the report must

include the number of shares bought and/or sold, whether

54

55

the sale (or purchase) was a private transaction, and the
number of shares owned after the sale (or purchase). The
price at which the transaction took place is not reported.
Unless one of the parties to the transaction chooses to

announce it, or unless the financial press publicizes the

transaction after seeing it in the Official Summary, it is

 

likely to largely escape notice.

Review of Literature

 

Possible legal and ethical problems involved in
this type of transaction have received considerable atten-
tion in legal periodicals. The question raised in these
articles is if, and if so under what circumstances, a
controlling stockholder should be permitted to accept a
price for his shares which is not made generally available

to other shareholders.

Perlman v. Feldman.--Considerable attention is given

 

to the case of Perlman v. Feldman, decided on appeal by the
United States Court of Appeals, Second Circuit, in 1955.
Henry G. Manne suggests that "there are few serious legal
problems with any of the direct purchase techniques" (for
obtaining control). However, he notes the influence of
Adolf A. Berle's contentions (in 1933 and 1958 writings)
that control is a corporate asset and any premium received
by an individual for the sale of control belongs in equity
to all shareholders. A number of legal writers have

continued to press for such a rule. He says, however,

56

that "the courts have refused to follow this thesis, and
there are numerous judicial statements to the effect that
one may claim a premium for control." Even in what he
regards as the difficult Perlman v. Feldmann case, he
maintains that "The court was explicit, however, that the
seller could retain that part of the premium received for
control not covering the power to allocate steel."l

In a 1971 article reconsidering the case of Perl-
man v. Feldmann and its implications for the sale of

control bocks, Leonard J. Conolly notes that neither this

case nor any other has upheld the corporate asset theory

of control, which suggests that control is an asset belong-

ing to the corporation—-not to selling stockholders. He
says that there are two types of situations that can make
sale of control illegal: (l) where the circumstances
suggest trouble for the corporation or (2) where a sale
of directorships is implied. Discussing this second type
of situation, he says, "Many agreements for the sale of a
controlling block of shares include provisions for the
immediate transfer of board control, usually by means of
seriatim resignations and appointments of nominees of the
buyer." In such a situation the key factor determining
whether or not such a transaction is legal is the propor-

tion of shares acquired:

 

1Henry G. Manne, "Mergers and the Market," Journal
of Political Economy, LXXIII (April, 1965), p. 116.

 

57

If a premium is paid for an absolute majority
of the outstanding shares under an agreement with
a typical seriatim-resignation clause, the courts
seem inclined to attribute the premium to the
voting control incident to the majority block.
Other factors remaining equal, this appears to be
lawful. But if a premium is paid for an insignifi-
cant fraction of the outstanding shares under a
contract containing a similar provision, the
premium cannot be attributable to control since a
small number of shares do not ordinarily represent
voting control. The courts will then attribute
the premium to the resignations of the directors.
This is an unlawful sale of directorships.

Other Cases.-—The greatest uncertainty as to

 

legality or illegality occurs when the proportion of out-
standing shares purchased is significant, but less than a
majority. In the case of Essex Universal Corporation v.
Yates, the proportion of outstanding shares sold was 28.3%,
and the three concurring opinions left doubt as to whether-
a sale of this magnitude could be held to constitute
working control. Conolly suggests that the block must be
large enough to carry control with it "as a practical
certainty." Even under these circumstances there remains
some uncertainty. The decision on Ferraioli v. Cantor
held that if a seller of a controlling block at a premium
invites any minority stockholder to participate, he must
invite all. Moreover, "certain judges have shown varying
degrees of sympathy" for the notion that a "buyer of con;

trolling shares should be required to make an offer to

 

1Leonard J. Conolly, "Perlman v. Feldmann and the
Sale of Control--A Brief Reconsideration," Business
Lagyer, XXVI (April, 1971), pp. 1261-63.

58

purchase the minority shares . . . at the same price . . ."
Conolly says that whether this approach will gain wide
acceptance remains to be seen.1
A 1968 case not cited by Conolly also pertains to
the sale of control. Christophides v. Porco was brought
under section 10(b) of the Securities Exchange Act of 1934
and SEC Rule lO-b. (The case of Ferraioli v. Cantor cited
by Conolly was also brought under these sections.) In his
opinion dismissing the suit, the judge held that:
Similarly, a purchaser is free to offer a
premium for a block of control stock. This is so,
even though control stock is purchased pursuant to
a plan to acquire the remainder of the shares at a
lower price, if, by private purchase or the normal
economics of the marketplace, this can be achieved.
It is only where faud, deceit or manipulation enter

that a violation of state law or of Rule lOb-S
occurs.2

Brown v. Halbert.-—The largest premium paid for

 

control which was uncovered in this part of the study is
138%, reported in the 1969 court case of Brown v. Halbert.
Halbert and his wife owned 53% of the outstanding common
stock of a savings and loan association; he was also
president, chairman of the board of directors,and manager.
Halbert was approached by the president of another savings
and loan association and was asked whether his association

was for sale. Halbert replied, "No, the Association is

 

11bid., pp. 1263-70.

2Christophides v. Porco, 289 F. Supp. 405 (D. C.
N. Y. 1968).

59

not for sale. However, my wife and I would entertain sell-
ing our stock." Halbert stated a price of two and one-half
times book value, which was eventually accepted, producing
a price of $1548.05 per share. Halbert agreed to permit
the purchasers to inspect the books, and to refrain from
paying dividends.l

Only after the sale had been negotiated were
minority stockholders informed. Halbert then assisted the
purchasers in buying the minority stock. (The court held
that at all times the purchasers had intended to acquire the
minority stock.) However, the offer to minority stock-
holders was $300 per share-~l9% of the amount paid Halbert
for his control block. Apparently few if any minority
stockholders accepted this offer. The purchasers subse-
quently took control of the association and acquired all
but a few of the minority shares at prices ranging from
$611 to $650 per share. These prices were 40% to 42% of
the price paid Halbert, and were close to the book value
per share.2 The implied premium above the highest price
paid to minority stockholders ($650) is 138%. (The $1548
paid is 238% of $650.)

Testimony by the expert valuation witnesses for
Brown established fair market values of the majority stock

at $1154 per share and the minority at $944 per share.

 

1Brown v. Halbert, 76 Cal. Rptr. 781 (Cal. App.
1969).

21bid.

60

This differential implies a fair premium of 22% for the
control stock. It is notable that not even the witness
for Brown contended that no premium should exist for the
control block. The court held that Halbert violated his
triple fiduciary duty by selling his stock at a price so
disproportionatetx3that available to minority stockholders
and by selling to purchasers who radically changed the
policy of the association (by announcing that there would
be no dividends for ten to twenty years) in a way that
tended to depress the price of the stock. The court held
that Halbert could not retain the advantage he derived
and ordered that the difference be computed between the
amount received by Halbert and that received by the
minority holders, and that this difference (plus interest)
be paid to minority stockholders in proportion to their

holdings.1

Conclusions from Literature.--It would appear from

 

this literature that, due to the uncertainty involved,
prospective buyers of large blocks of stock might be
reluctant to make such purchases at a premium unless they
wish to extend similar offers to all stockholders. It is
even more likely that prospective sellers of large blocks
would be reluctant to sell at a premium unless they can
convince the purchaser to extend the same offer to other

stockholders. Further, when purchases of large blocks are

 

lIbid.

 

61

made at a premium without a similar offer to other stock-
holders, it is very likely that both the purchaser and
seller will attempt to conceal the terms of the transaction.
Aside from the legal questions involved, it appears that

the ethics of such transactions are doubted by some judges
and other scholars. Therefore it is not surprising that
data on price is available for only a minority of those

stock transactions.

Transactions Studied

 

All transactions are studied which are listed in

1966, 1967 and 1971 issues of Official Summary of Security

 

Transactions and which meet the following additional

 

criteria:

1. transaction designated as private sale

2. transaction at least 10,000 shares of
common stock

3. transaction at least 5% of the common

shares outstanding

4. number of shares sold by seller greater
than his holdings after sale

5. no offsetting purchases during month
of sale

6. sale price available

The Official Summary is published monthly by the

 

United States Securities and Exchange Commission and is
based on reports required to be filed with the SEC by a
purchaser or seller of equity securities who falls into
one or more of the following categories:

1. officer

2. director

3. beneficial holder of 10% or more of any
registered equity security

62

Lack of information as to price at which the sale
took place requires elimination of two-thirds of the
transactions which meet the other criteria. Thirty-three
companies remain. One of these is eliminated because it
is the only financial company and would require that two
of the independent variables not be used. The result is

thirty-two companies studied.

Variables Emplgyed

 

Table 3.1 lists names of the variables with their
definitions. Variables 1 and 2, and 7 through 12 are
similar in nature and purpose to variables used in the
study of cash tender offers. Variable 1, the price ratio,
is the dependent variable whose variation in value is to
be explained. Variable 6, the current ratio, is used for
the same purpose as the quick ratio in the cash tender
offer study. Variable 5, block size, is used for a pur-
pose similar to that of shares sought in the cash tender
offer study. Variables 3, 4, and 13 are new (different
from any used in the cash tender offer study) and the
hypotheses in the introductory chapter which they are

designed to test deserve explanation at this point.

Purchaser.--The hypothesis that a corporate pur-

 

chaser will tend to pay a higher premium than an individual
purchaser is based on the expectation that a corporate pur-
chaser will be more likely to have merger in mind than an

individual purchaser, and will be in a better position to

TABLE 3.l.--Private Purchases:

63

Variables Employed.

 

 

No. Code Name Definition
1 PREM Price ratio Ratio of block sale price to
market price
2 EXCH Exchange Listing status: (1) NYSE,
(2) ASE, (3) Midwest SE,
(4) unlisted
3 SELL Seller Seller: (1) corporation,
(2) individual(s)
4 PURC Purchaser Purchaser: (1) corporation,
(2) individual(s), (3) issuer
5 SIZE Block size Number of shares sold as
percentage of shares out-
standing
6 CURR Current ratio Ratio of current assets to
current liabilities
7 DEBT Debt ratio Ratio of long-term debt to
total of long-term debt,
preferred stock, and common
stock equity
8 ROR Rate of return Net income as percentage of
net worth
9 EDIR Earnings Direction of prior year's
direction earnings per share: (1)
increase, (2) no change, or
(3) decrease from year before
that
10 BVMK Book value to Ratio of book value per share
market price to market price.
11 MKT Market price Market price (close or bid)
two business days before
date of agreement
12 HILO Price range Ratio of prior year's high
price to prior year's low
price
13 SUBS Subsequent Subsequent status of minority
status stockholders: (1) no change,

(2) cash tender offer, (3)
merger with purchaser, (4)
merger with corporation other
than purchaser

 

64

profit from merger. An issuer repurchasing a block of
stock is likely to pay even less, due to the possibility
that factors other than control are involved in the pur-

chase.

Seller.--The hypothesis that a corporate seller
will tend to get a higher premium than an individual seller
is based on analysis similar to that in the preceeding
paragraph concerning the purchaser. A corporate holder
may have more attractive alternatives to selling the stock
than an individual holder; also the individual holder mayv
be more likely to have other motives for selling than a

premium, such as need to diversify an investment portfolio.

Subsequent Status.--The hypothesis that transac-

 

tions followed by merger or cash tender offer are more
likely to involve high premiums than those followed by no
change in the status of minority shareholders, is based on
the belief that the subsequent occurrence of these events
will help identify transactions where they were most
envisioned by the purchaser in negotiations, and where they
therefore caused the purchaser to be willing to pay a

relatively high premium.

Method of Analysis
Price ratio, the ratio of block sale price to
market price, is the dependent variable. Stepwise deletion

of variables is used to determine the extent to which each

65

independent variable explains the variation in price ratio.
Under this method of regression analysis, all of the inde-
pendent variables are entered in the initial step, and a
regression equation obtained. For each subsequent step,
one independent variable is deleted and a new regression
equation calculated based only upon the remaining variables.
The order of deletion of variables is determined by the
relative contribution which each variable is able to make;
the variable deleted is the one whose deletion will reduce
R2 least, as determined by trying all possibilities. At
each step, the coefficient of determination (R2) and the
F-test for significance are calculated for the regression
equation. The procedure continues until all independent
variables have been deleted.

From the several resulting regression equations,
one is selected as being best, based on the values of R2
and the F-test for significance. That equation is examined
to see which of the independent variables are included,
and to determine the relative contribution of each. The
residuals (unexplained variation between observed value of
each of the observations and the value predicted by the
regression equation) are examined for this regression
equation, not only to get a better feel for the fit of the
equatiOn, but also to detect patterns in the observations
which may be responsible for the fit not being better than

it is. The raw data are also examined for possible clues

r-
or.

Y
.IIAI

v-o‘.‘

"v‘l.
"sLl

"A

‘2'.“

154.:
u; 5‘

(I:
”I

 

n
\
s

I I

66

to factors interfering with fit, and for other information

relevant to the study.

Findings

Price Ratios and Date of Transaction.--Table 3.2

 

shows the wide range of price ratios found for the private
purchases of large blocks of stock (31.4% to 137.5%).
Twelve of the transactions were at premiums of 10% or more
above market price. But five were more than 10% below
market price. Almost half were within 10% of the market
price (15 of the 32 transactions).

The differing distributions of the first two
columns of Table 3.2 show that the price ratios for trans-

actions selected from the 1971 issues of Official Summary

 

(second column) tend to be significantly lower than those
selected from the 1966 and 1967 issues (first column).

The median for the earlier group is 107.3%, compared with
100.0% for the later group. Table 3.3 summarizes this
same data in a 2 by 2 table. Taking a premium of .l% as
insignificant, 17 of 21 transactions in the earlier group
occurred at a premium, compared with only 3 of 11 companies
in the later group. Stated another way the observed fre-
quency in the upper left cell is 17, compared with an
expected frequency of 13.13. The Chi square test, after
application of Yates' correction for continuity, shows
that the relationship between price ratios and time period

of transaction is statistically significant at the .01 level.

67

TABLE 3.2.--Actual Values of Price Ratio, Transactions
before November 3, 1967 and after March 16,
1969 (Number of Transactions).

 

November 22, 1965 March 17, 1969

 

 

 

 

 

Price Ratio to to
% November 2, 1967 June 25, 1971 Total
Under 70.0% 1 1 2
.70.0 - 79.9 0 2 2
80.0 - 89.9 1 0 1
90.0 - 99.9 2 2 4
100.0 - 109.9 8a 3b 11
110.0 — 119.9 6 2
120.0 - 129.9 2 0
130.0 - 139.9 1 1
Total 21 ll 32
Mean 106.3 96.0 102.7
Median 107.3 100.0 106.6-
Range 31.4 - 137.5 62.0 - 133.3 31.4 -
137.5
aRange this cell is 101.9 to 107.4
bRange this cell is 100.0 to 100.1
TABLE 3.3.--Comparison of Price Ratio with Date of
Transaction (Number of Transactions).
November 22, 1965 March 17, 1969
Price Ratio to to
% November 2, 1967 June 25, 1971 Total
101.9 and over 17 3 20
100.1 and under 4 8 12
Total 21 ll 32
Chi square (with Yates' correction for continuity) = 6.73

'df = 1; significant at .01 level.

68

Regression Results.--Table 3.4 shows the regression
steps described in the discussion of method of analysis.
The first row shows that with all 12 independent variables
in the regression equation the R2 is .39, but the signi-
ficance level is an unsatisfactory 46.4%. In steps 2
through 6, as variables EDIR, EXCH, DEBT, SIZE, and SUBS
are deleted, R2 remains at .39 and the significance level
improves to 7.3%. Deleting these variables increases the
statistical significance without reducing the explanatory
power of the regression equation. This regression equation
in step number 6 appears to be the best equation for the
purposes of this study. Each of the next three deletion
steps causes the significance level to improve slightly
more, but at the cost of a decline in R2. Beyond step
number 9 the deletion of additional variables is clearly
disadvantageous, because not only does R2 continue to
drop (and even more rapidly) but also the significance
level starts to deteriorate.

Table 3.5 shows the coefficients for the independent
variables which are in this regression equation for step
number 6. This best regression equation consists of the
step 6 constant (141.9) plus the regression coefficients,

and takes the following form:

PREM = 141.9 - 10.31 SELL - 7.73 PURC - 9.53
CURR - 0.16 ROR + 27.81 BVMK + 0.21 MKT

- 5.84 HILO

69

TABLE 3.4.--Regression Steps,Stepwise Deletion of Variables.

 

 

Step Variable 2 Significance
No. Deleted R Constant Level
1 .39 148.6 46.4%

2 EDIR .39 148.8 36.0%
3 EXCH .39 149.7 26.4%
4 DEBT .39 149.9 18.2%
5 SIZE .39 148.3 11.8%
6 SUBS .39 141.9 7.3%
7 MKT .36 147.2 6.3%
8 SELL .32 128.6 6.1%
9 CURR .29 117.5 5.1%
10 PURC .23 110.2 6.4%
11 BVMK .14 121.1 11.5%
12 ROR .05 109.8 21.7%
13 HILO .00 102.7 100.0%

 

For comparing the relative contribution of each of these

variables, the Beta coefficients are useful. The two most
important variables, both showing strong association with
the price ratio, are the ratio of book value per share to

market price (BVMK) and the price range (HILO).

Book Value to Market Price.--The ratio of book

 

value to market price shows a strong positive association
with the price ratio, which is in the direction expected.
One hypothesis is that the greater the possibility of
profitable liquidation, as roughly measured by this ratio,
the greater will be the percentage premium for control

stock.

70

TABLE 3.5.--Coefficients for Step No. 6,Stepwise Deletion
of Variables.

 

 

Consistent
Independent Regression Beta with
Variable Coefficient Coefficient Hypothesis?
SELL -10.31 -.23 yes
PURC - 7.73 -.31 . yes
CURR - 9.53 -.37 no
ROR - 0.16 -.30 yes
BVMK 27.81 .53 yes
MKT 0.21 .21 no
HILO - 5.84 -.42 no

 

Table 3.6 shows some further analysis of the
variable book value to market price (which shows the
strongest association with price ratio in the best regres-
sion equation). All observations with a ratio of book
value to market price greater than one are listed, grouped
by listing status. It appears from this table that the
variables exchange and book value to market price have
a synergistic effect upon price ratio. Although the mean
book value to market price for unlisted observations (1.45)
is higher than for New York Stock Exchange listed observa-
tions (1.32), the price ratio for the unlisted group
(95.6%) is lower than for the NYSE group (122.8%). Further,
the lowest price ratio in the NYSE group (114.3%) is higher
than the highest (111.7%) in the unlisted group. It is

possible that the types of investors who purchase large

 

-‘B

x
lab—r

 

's‘.

71

TABLE 3.6.--Observations with Ratio of Book Value to Market
Price Greater than 1.00. Grouped by Listing
Status.

 

Observation Book Value to
Number* Market Price Price Ratio

 

New York Stock Exchange

29 1.38 133.3%
6 1.31 120.7
11 1.26 114.3
Group mean 1.32 122.8

Midwest Stock Exchange

31 1.17 117.7
Unlisted

30 1.61 100.1
24 1.54 111.7
27 1.21 75.1
Group mean 1.45 95.6

 

*Companies are identified by observation number in Appendix
table.

blocks of stock privately can more readily detect a value
criterion such as the ratio of book value to market value
when the company is listed; one factor in such a situation
could be the use of computerized data banks for screening
purposes; data banks are more likely to include listed

than unlisted securities.

Price Range.--The price range shows a strong nega—

 

tive association with the price ratio. This finding is
inconsistent with one of the hypotheses, which is that

the percentage premium for control stock will be greater

«
up NH

>4 .
‘uvtn

"4‘
-.r
'-..a
.

)

5::
(I)

 

"'v~~
‘1

514;,
t

 

 

‘ I

A“ “

Q.“ .

..
'IAX

 

'n

v.)
.~~-l

I
ll)
1))

 

v“

o.
:11
I

(I)
x

72

when the price of the stock varied widely in the prior

year as measured by this ratio.

Other Variables Showing Association with Price

 

Ratig.-—Three other independent variables show moderate
association with the price ratio: current ratio, purchaser,
and rate of return. Current ratio shows negative associa-
tion with the price ratio—-a direction opposite to that
expected; a hypothesis is that the percentage premium

will be greater when the corporation is relatively liquid.-
Purchaser shows a moderate negative association. This
direction is consistent with one of the hypotheses: that

in a private purchase of a large block of stock, the per-
centage premium is greatest when the purchaser is a cor-
poration other than the issuer, less when the purchaser'

is an individual (or individuals), and least when the
purchaser is the issuer. The third variable showing a
moderate association is rate of return--showing a negative.
association which is also consistent with a hypothesis:

that the percentage premium for control stock will be *
greater when management performance (as measured by criteria
such as this) has been relatively poor.

The other two variables which are in the regression
equation at step number 6 show weak association; they are
seller and market price. Seller shows a weak negative
association, consistent in direction with a hypothesis:

that the percentage premium for control will be greater

73

when the seller is a corporation than when the seller is

an individual (or individuals). Market price shows a

weak positive association; the direction of association of
this variable is opposite to that expected, as one of the
hypotheses is that the percentage premium for control stock
will be greater when the market price is relatively low.

In summary, three of the five independent variables
in this best regression equation showing either strong or
moderate association are consistent in direction with
hypotheses. One of the two independent variables showing
weak association does so in a direction inconsistent with
a hypothesis. In addition, five other independent variables
did not show sufficient association with price ratio to be
included in the best regression equation; these five,
appearing in the variable deleted column for steps 2
through 6 in Table 3.4 are the following:

Earnings direction

Exchange

Debt ratio

Block size
Subsequent status

Analysis of Residuals.--The next step is analysis

 

of the residuals, or unexplained variations. When the
residuals are totaled for observations 22 through 32,
which represent the later group of transactions, the sum
is -41.0 for the 11 observations or a mean residual for
these 11 of -3.73. Since the sum of residuals for the

entire group of 32 observations must be zero, the regression

74

equation is predicting a value which is, on the average,
3.73 percentage points too high for this later group of
transactions.

Another way of examining the residuals is to see
what happened to the out-on-a limb predictions made by
the regression equation. Table 3.7 shows the eight
observations for which a value of price ratio was predicted
of either under 90.0 or over 119.9. All but two had a
residual of at least 10.0%, and one of 58.0%. However all
eight of the predictions were in the correct direction from
the sample mean of 102.7; that is for every one of these
observations both the actual value and predicted value lie
on the same side of the mean.

Table 3.8 shows the six observations with the
largest residuals-~those of 13.0% or more. For three of
these observations the prediction was in the wrong direc-
tion; that is the actual value and predicted values lie
on Opposite sides of the mean; the three observations are
numbers 8, 20, and 30. The other three predictions were
in the right direction but did not go far enough away
from the mean. Three observations (numbers 5, 8, and 20
had residuals greater than 17.6%. Only two (numbers 51
and 22) of the six observations with the largest residuals
resulted from the out-on-a limb predictions listed in

Table 3.6.

75

TABLE 3.7.--Unexplained Variations for Observations with
Predicted Value of Price Ratio under 90.0%
or over 119.9%.

 

 

Actual Predicted Unexplained
Observation Value Value Variation
Number* (%) (%) (%)
5 31.4 89.4 -58.0
11 114.3 126.1 -11.8
15 84.2 76.6 7.6
22 74.6 88.5 -13.9
25 90.4 84.6 5.8
27 75.1 87.8 -12.7
29 133.3 120.5 12.8
32 62.0 72.0 -10.0
Median unexplained variation (disregarding sign) = 12.3%
Mean unexplained variation (disregarding sign) = 16.6%

 

*Companies are identified by observation number in Appendix‘
table.

TABLE 3.8.--Observations with Unexplained Variations of
13.0% or more from Predicted Value of Price

 

 

Ratio.
Actual Predicted Unexplained
Observation Value Value Variation
Number (%) (%) (%)
31.4 89.4 -58.0
120.7 107.5 13.2
137.5 97.3 40.2
20 127.9 100.8 27.1
22 74.6 88.5 -13.9
30 100.1 117.7 -17.6

 

*Companies are identified by observation number in Appendix

table.

76

Simple Correlations.--Simple correlations between
pairs of variables are of interest largely for purposes of
considering why certain independent variables may not have
entered the best regression equation. Table 3.9 shows

such correlations of .40 or higher.

TABLE 3.9.--Pairs of Variables Having Simple Correlations
of .40 or Higher with Each Other.

 

 

Variables Correlation
EXCH with MKT -.46
EXCH with HILO .56
PURC with SUBS -.50
DEBT with ROR -.44
DEBT with HILO .50
ROR with HILO -.77

 

Discussion of Findings

 

The best regression equation explains a substantial
amount of the variation in the price ratios of private
purchases of large blocks of stock and does so at a satis-
factory level of significance. The value of R2 for the
equation is .39 and the significance level is 7.3%.

Four hypotheses are supported in this part of the
study by association in the expected direction between
independent variables and the price ratio- Book value to

market price is supported with strong association. Rate

77

of return and purchaser are supported with moderate asso-

ciation. Seller is supported with weak association.

Book Value to Market Price.--The ratio of book

 

value to market price shows strong positive association
with the price ratio, which is in the direction consistent
with the hypothesis that the greater the possibility of
profitable liquidation (as roughly indicated by a rela-
tively high ratio of book value per share to market price),
the greater is the price ratio. Further, there seems to
be a synergistic effect between book value to market price
and exchange with regard to the effect on price ratio.

It appears from these findings that those who make
private purchases of large blocks of stock may tend to be
considering liquidation as an alternative to continuing
to operate the company after control is obtained. The
synergistic effect of exchange may indicate that such
prospective purchasers can more readily detect a value
criterion such as the ratio of book value to market value
when the company is listed on the New York Stock Exchange;
one factor may be the use of computers and computerized
data banks for screening purposes, since data banks are
more likely to include listed than unlisted companies.

Since it is bargaining between holder of a large
block of stock and the prospective purchaser which deter-
mines the price ratio, it is possible that sellers of

'large blocks are quite book value conscious and therefore

78

tend to demand a higher premium when the book value per
share is high in relation to the market price. Like
prospective purchasers, prospective sellers would tend
to view liquidation of the company as an alternative to

sale of their shares.

Rate of Return.--Rate of return (net income divided

 

by net worth) shows a moderate negative association with
the price ratio, which is in a direction consistent with
the hypothesis that the percentage premium for control is
greater when management performance (as measured by
criteria such as rate of return on equity and per share
earnings growth) has been relatively poor. Often a holder
of a control block would be in a position to bring about
a change in management. However, it is understandable
that a long-term major stockholder seeing the company
deteriorate, might prefer to sell control and let someone
else make management changes than to get involved in a

probably unpleasant task involving long-time acquaintances.

Purchaser.--Purchaser has a moderate negative

 

association with price ratio, which is consistent in dir-
ection with one of the hypotheses: that in a private
purchase of a large block of stock, the percentage premium
is greatest when the purchaser is a corporation other

than the issuer, less when the purchaser is an individual
(or individuals), and least when the purchaser is the

issuer. A corporation other than the issuer has at least

79

one more alternative than other purchasers: it can merge
the company it controls with itself; this additional
alternative is probably worth paying more of a premium than
other purchasers would be willing to pay. Another possi-
bility is that a corporation will find it easier to finance
the purchase of a large block of stock than will individuals

or the issuing corporation.

Seller.--Seller is weakly correlated in a negative
direction with price ratio, which is consistent with the
hypothesis that the percentage premium in private purchase
of a large block of stock is greater when seller is a
corporation than when the seller is an individual (or
individuals). The rationale is similar in one respect to
the foregoing discussion of purchaser; just as a corpora-
tion acquiring control has the option of merging, a
corporation considering sale of a large block of stock in
another corporation has an alternative--merging with the
controlled corporation; this option is not available to an
individual seller.

Another reason for smaller premiums in the case of
sales by individuals is that the individuals may have
reasons for being willing to part with control that would
be less applicable to a corporate seller; one example is
a desire to retire; another is a need for diversification

of investment portfolio.

.A

uq

V.

r1.

'4‘

80

Variables Not in Regression Equation.--Why did the

 

following five independent variables fail to show suffi-
cient association with the price ratio to be included in
the best regression equation?

Earnings direction

Exchange

Debt ratio

Block size

Subsequent status

One possibility is that some of these variables
have a relatively high simple correlation with one or more
variables which are in the best regression equation.

Table 3.9 shows that three deleted variables show such
correlation. Exchange is moderately correlated with two
variables in the best regression equation: -.46 with

market price, and .56 with price range. The second deleted
variable, debt ratio, is moderately correlated with rate

of return (-.44) and price range (.50), both of which are
in the best regression equation. The third deleted varia-
ble is subsequent status, which has a -.50 correlation with.
purchaser, which is in the best regression equation.

In addition to their possible influence on which
variables enter the best regression equation, the simple
correlations between pairs of variables are of interest
in themselves. The negative correlation between exchange
and market price is not surprising-~issues listed on the
New York Stock Exchange are expected to tend to have higher

market prices than those listed on the American Stock

Exchange, or unlisted. Similarly, the positive association

81

between exchange and price range indicates that issues
listed on the New York Stock Exchange tend to have a
narrower price range than other issues.

The negative association between purchaser and
subsequent status indicates that when the purchaser is
the issuer or an individual, there is less likelihood of
an event such as tender offer or merger, than when the
purchaser is a corporation. This relationship seems
logical.

The negative association of debt ratio with rate
of return suggests unfavorable financial leverage, with
those firms having a high debt ratio showing a lower rate
of return on net worth. Perhaps unfavorable financial
leverage tends to be characteristic of companies in which
private purchases of large blocks occur. The next (and
last) two relationships shown in Table 3.9 seem consistent
with this implication of unfavorable financial leverage
in the companies studied. The positive association of
debt ratio with price range seems logical because if a
high debt ratio increases risk, one would expect the extent
of stock price reaction to a given event to be magnified.
The negative relationship between rate of return and price
range follows logically from the previous two relationships;
since debt is negatively associated with rate of return,
and debt is also positively associated with price range,
rate of return is negatively associated with price range.

This last correlation is surprisingly strong (-.77);

82

however it seems to possess additional logic of its own.
A company with a low rate of return may have greater

variation in the price of its stock because of uncertainty

'. Lh-q-v-e-r-thi‘.’ .. .

and speculation as to the prospects for improvement;
investors might regard a low rate of return as a temporary
situation, while considering an average level of rate of
return to be a stable situation more likely to continue in
the future.

Another reason that certain variables are not in
the best regression equation may be offsetting forces at
work with regard to a given variable. For example, one
hypothesis is that the premium for control will be higher
when there is a relatively large unused debt capacity--as
measured by a low debt ratio. If investors feel that
opportunities for favorable financial leverage have been
neglected, they might be willing to pay a larger premium
to get control so they can have the corporation borrow
money and attempt to thereby improve reported earnings per
share.‘ On the other hand, debt ratio could have the
opposite effect on price ratio: a high debt ratio (perhaps
only above some level considered normal) might be regarded
as a danger sign which would impel investors to get control
of the company in an effort to protect their investments.
These offsetting effects could prevent the debt ratio from
showing sufficient association with the price ratio to be

included in the best regression equation.

83

Current Ratio.--A second aspect of the findings to

 

be examined is the fact that in the best regression equa-
tion, three of the seven independent variables have associa-
tion with the price ratio in a direction that is incon-
sistent with a hypothesis. One of these variables is the
current ratio, showing negative association while positive
association is expected. The reason for this situation

may be similar to a possible reason that debt ratio did

not remain in the regression equation-~there may be off-
setting forces at work. The hypothesis that a high current
ratio would be associated with a high price ratio is based
on the expectation that relatively high liquidity would
make takeover of control more attractive. However, there
is another possibility--that a low current ratio, at least
below some level considered normal, may be regarded as a
danger sign which will cause stockholders to want to get
control to try to protect their investments. If such
motivation exists, and is stronger than the motivation to
take over a liquid company, then the negative association

found between current ratio and price ratio is logical.

Market Price.--Another variable associated in a

 

direction inconsistent with a hypothesis is market price,
showing weak positive association while negative associa-
tion is expected. One possible reason for lack of a
negative association is that with private purchases of

large blocks, both buyer and seller are likely to be

84

relatively sophisticated and therefore more likely to
consider percentage premiums above market price than dollar
premiums. It is also possible that an irrational market
overvalues lower-priced stocks relative to higher-priced
stocks, and that sophisticated block purchasers therefore

will pay higher premiums for higher-priced stocks.

Price Range.--The third variable in this category

 

is price range, which shows strong negative association
with price ratio instead of the positive association
expected. The strength and direction of association of
the price range is surprising, especially in View of the
findings in the study of cash tender offers; in that study
price range is one of two variables showing strongest
association with price ratio, and the association is
positive, as hypothesized. The reasoning behind the hypo-
thesis in both studies is that if the price of the stock
varied over a wide range in the previous year, investors
will be more uncertain as to the value of the stock than
if the stock price stayed within a relatively narrow
range; therefore a larger premium above the present market
price will be required to overcome this uncertainty and
induce the holders to sell.

Perhaps uncertainty of this type characterizes the
typical investor at whom a cash tender offer is aimed, but
not the large block holder who is the potential seller in

a private purchase of a large block of stock; perhaps

85

these large block holders are more sophisticated than the
typical investor, and less uncertain as to the value of
their holdings. Also, there is the possibility that the
large block holder may possess inside information which

reduces his uncertainty as to the value of the stock.

Fit of Regression Equation.--Another aspect of the

 

findings needing consideration is why the fit of the best
regression is not better: why the value of R2 is not
higher than .39, and why there is so much unexplained
variation of observed values from predicted values. (One
factor is probably the existence of a few extreme values
of price ratio, particularly those less than 100%.) As
noted in Table 3.2, 9 of the 32 observations have price
ratios under 100%, 5 of which are under 90%. Further,
while there are only two observations under 70%, one of
these is 31.4%--and this is the only observation under
80% for the 21 transactions in the earlier time period.

The statistically significant difference between
price ratios in the earlier transactions and the later
transactions is probably an important factor preventing
better fit of the best regression equation. These differ-
ences are shown in Tables 3.2 and 3.3. One possible reason
for these differences between the two periods is increasing
concern with possible legal and ethical objections to
accepting a premium for a large block without making that
premium also available to minority shareholders. Two of

the court cases cited earlier in this chapter occurred

86

after the earlier group of transactions, but before most of
the later group. Another possible reason for the higher
premiums in the earlier period is that the conglomerate
phenomenon which was flourishing during the earlier period
had diminished considerably by the later period. The
higher level of conglomerate activity in the earlier

period may have increased the competition for control of
companies through large block transactions, heightened the
sellers' expectations as to premium for control blocks, and
increased purchasers' willingness to pay high premiums.

The individual nature of these transactions is
another likely hurdle to a better fit. Several independent
variables are used in an effort to explain as much of the
change in the value of price ratio as possible; however,
there are limitations on the number of variables which
can be used; some relevant considerations would be diffi-
cult to obtain information about, and some information
which is available is difficult to quantify sufficiently
to be useful in this type of analysis. Sellers may have
individual needs and desires, such as varying tax and
liquidity situations and differing plans (such as a desire
to retire). Purchasers are also likely to have varying
types and strengths of motivation, such as a desire for
power. ‘Finally buyers and sellers may have varying atti-
tudes toward a risk peculiar to this type of transaction--
the risk that it may be considered illegal or unethical

under some circumstances to accept a premium for control

87

stock which is not available to minority stockholders.
This legal-ethical problem is covered in detail in the
first part of this chapter.

Further study of the transactions having the
largest unexplained variations (listed in Table 3.7)
suggests possible reasons for the two observations which
had very large unexplained variations and also predictions
which were in the wrong direction from the mean: numbers
8 and 20. Observation number 8 was unusual in that it
represented a purchase by the parent company, which though
small, was sufficient to raise the parent company's owner-
ship to 80% and thereby permit consolidation for U. S.
income tax purposes. This situation probably was a major
factor in the price ratio of l37.5%--largest in this study—-
despite the prediction of 97.3%.

Observation number 20 was another case with a
large price ratio (127.9%) but a prediction much smaller
(100.8%). Here the competition for the company seems to
have been unusually great; the sellers were officers of
another company which then terminated merger discussions,
permitting merger with the purchaser. Situations of these
types are examples of the many individual influences which

can hamper the fit of the best regression equation.

CHAPTER IV

CLASSIFIED COMMON STOCK

Classified common stock is relatively rare, but
when utilized may facilitate obtaining and/or retaining
voting control of a corporation. Unlike the cash tender
offer or the private purchase of a large block of common
stock, it seems more likely to be used to retain control
than to obtain control.

As the name implies, classified common stock results
from dividing the common stock into two (or possibly more)
classes; each class can then be given distinctive charac-
teristics. However, the type of individual characteristics
which can be imparted to each class is sharply limited if
each class is to retain its status as common stock. A key
characteristic of common stock is that it is the residual
equity; it is entitled to no preference as to dividends or
distributions in liquidation. If a material preference is
given to one class, it is no longer common stock. There—
fore it is necessary to distinguish classified common stocks
from other types of classified stocks which may be preferred

stocks or hybrid types.

88

89

This part of the study deals with classified common
stocks having classes which do not differ materially except
as to voting rights. Voting rights are one usual charac-
teristic of common stock; however, the removal of voting
rights from one class appears to do less damage to its
status as common stock than would the changing of most
other characteristics. This opinion is based on the obser-
vation that voting rights are relatively unimportant to
many stockholders. In the absence of a proxy contest,
management routinely receives adequate proxies to carry out
its will. Even with a proxy contest (which is relatively
rare) management's ability to use corporate funds for
proxy solicitation gives it a tremendous advantage. Further,
it seems likely that many stockholders lack either the
experience, time, or interest to inform themselves of the

issues and to utilize their votes effectively.

Companies Studied

Three main sources are used for the selection of
companies studied: Moodyfs Manuals, Bank andguotation
Record and Financial Post.

The Moody's Manuals used are Moody's Industrial

 

 

Manual, 1968, and Moodyfs Bank and Finance Manual, 1969.

The blue page sections which give a listing and ten-year

price range of stocks and bonds are examined to find com-
panies having classified common stocks. The following

selection criteria are applied to the stocks found:

9O

1. Price information must be available on both
classes at least at monthly intervals for some years
since 1964.

2. Both classes must be residual equity with
voting rights the sole difference between classes.

In order to use an alternate source and to update

the sample, the entire June, 1972 issue of Bank and

 

Quotation Record is examined to find additional classified

 

common stocks; these stocks are subjected to the same two
selection criteria mentioned in the preceding paragraph.
Many Canadian securities (including those listed

on Canadian stock exchanges) are not included in Bank and

 

Quotation Record. Therefore to find additional Canadian

 

companies, all of the stock listings (including over-the-

counter issues) in the July 22, 1972 issue of Financial

 

Post are examined. A high proportion of the Canadian
classified stocks differed other than as to voting rights.
The non-voting shares would typically have a dividend
preference of approximately $.25 per year; after both
classes have received that amount they participate equally
in dividends. Classified stocks of this type are included
providing they have consistently paid at least the amount
of preference to both classes for the last several years.
For these companies it is felt that any distorting effect
of the dividend preference on the market prices would be

small.

 

3!-
'U

40-

v.»

HA

‘9

‘r'

n:
r.‘

J

 

(V)

'I

91

In some cases one class is made distinctly inferior
to the other class with regard to voting rights, while
still having such rights. For example, in the case of
Investors Diversified Services, Inc., both classes have
one vote per share, but the Class A Common participates
four times as much as the Class B Common in dividends,
other distributions, and upon liquidation.l From the stand-
point of voting power, the result is similar to giving the
Class A one vote per share, the Class B one-quarter vote
per share, and making them the same in all other respects.
This company is accepted for the sample and the price of
the Class B is compared with one—quarter of the price of
the Class A.

Another type of reduced-voting security which is
accepted for the sample is exemplified by Presidential
Realty Corp. Both the Class A Common and Class B Common
have one vote per share. However, they each vote as a
class, with the class A electing two-thirds of the direc-
tors, and the class B electing the other one-third. Since
there are approximately three times as many class B as
class A shares outstanding, the class B has clearly inferior
voting rights.2

The total sample of 29 consists of all companies

found which meet these criteria. This number is small in

 

lMoody's Bank and Finance Manual, 1969, p. 970.

 

21bid., p. 1366.

92

relation to the number of companies which have classified
stock. However, in a high proportion of cases there are
important differences between classes (other than voting
rights) which might be expected to influence their prices
and make them useless for studying the value attributed
to voting rights. Some are similar to preferred stocks.
Many companies are eliminated because of the lack
of price information on one or both classes. It is not
surprising that in many cases of classifiedcommon stocks
no price information is available on the voting common.
The purpose of such classification is usually to enable
one group of stockholders to retain control of the corpora-
tion. Therefore the voting common is often closely held
by such a group and no market exists for the stock. Even
for some companies which have both classes listed on
Canadian stock exchanges, the trades in the voting stock_
are so infrequent that it is not possible to calculate
price ratios (which require near-simultaneous prices of

both voting and non-voting stocks).

Variables Employed

 

Table 4.1 lists names of the variables employed,
with their definitions. The third through seventh variables
on the list (debt ratio through market price) are similar
in nature and purpose to variables used in the study of
cash tender offers. As with that study, debt ratio and
quick ratio are considered not applicable to financial

companies.

TABLE 4.l.--Classified Common

93

Stocks: Variables Employed.

 

Variable Name (and abbre-
viation, if any)

Definition

 

Price Ratio

Shares Ratio

Debt Ratio

Ratio of price of voting
common to market price of
non-voting (or reduced-
voting) common.

Ratio of number of voting
shares outstanding to number
of non-voting (or reduced-
voting) shares outstanding.

Ratio of long-term debt to
total of long-term debt, pre-
ferred stock, and common

stock equity.

Quick Ratio

Book Value to Market Price

Rate of Return

Market Price

Dividends per Share

_Net Working Capital to
Total Assets
(NWC/Total Assets)

Sales to Total Assets
(Sales/Total Assets)

Ratio of cash, short-term
marketable securities, and
accounts receivable, to
current liabilities.

Ratio of book value per share
to market price of non-voting
(or reduced-voting) common.

Net income as percentage of
net worth.

Market price of non-voting
(or reduced voting) common.

Dividends per common share in
calendar year (same for both
classes for all companies in
sample).

Ratio of net working capital
to total assets

Ratio of year's sales to
end-of—year total assets.

 

94

Price Ratio.--The price ratio is the dependent

 

variable for which it is desired to explain variation in
value. For each of the 29 companies forming the sample,
price quotations for each class of common stock are
obtained as of June 30 (or asnear that date as possible)
for each year 1965 through 1972 (to the extent available).
For each company a price ratio is computed for each year:
the ratio of the price of the voting common to the price

of the non-voting (or reduced—voting) common.

Shares Ratio.--The shares ratio is the ratio of the

 

number of voting shares to the number of non-voting (or
reduced voting) shares. The smaller this ratio is, the
greater is the total number of shares (voting and non-
voting combined) which can theoretically be controlled by'
ownership of a given number of voting shares. For example,
assume that a company has one million common shares out—
standing and that 50% of the voting shares are required to
control the company. If the ratio of voting to non-voting
shares is 4 (4 to 1), there would be 800,000 voting shares
and 200,000 non-voting shares. A total of 400,000 voting
shares would be required to control the company. On the
other hand, if the shares ratio is .25 (.25 to 1), there
would be 200,000 voting shares and 800,000 non-voting shares.
Only 100,000 voting shares would be required to control the
company.

One hypothesis is that the premium for voting

stock will tend to be greater when this shares ratio is

95

relatively low. In this example, if a person were willing
to pay a total premium (above the price of non-voting
shares) of $1,000,000 to get control of the corporation,
he would be willing to pay a premium of only $2.50 per
share if the ratio of voting to non-voting shares were 4,
but $10.00 per share if the ratio were .25. The per share
premium in both cases is calculated by dividing the total
premium he is willing to pay ($1,000,000) by the number of
voting shares required to control the company (400,000 and

100,000 respectively).

Dividends per Share.--Dividends per share is

 

included as an independent variable because although another
performance measure, rate of return, is included, dividends
are discretionary with those in control of a corporation,
and because dividends represent a tangible cash payment

to stockholders. Even if the rate of return is very good,
dividends might be kept low or omitted--the result could

be dissatisfaction on the part of some shareholders who
might thereby be motivated to attempt to get voting control

of the corporation to alter the dividend policy.

Other Variables.--The last two variables listed,

 

net working capital to total assets and sales to total

assets, are selected from variables used by Altman in his

study of prediction of corporate bankruptch' They are

 

1Edward I. Altman, "Financial Ratios, Discriminant
Analysis, and the Prediction of Corporate Bankruptcy,"Journal
of Finance, XXIII (September, 1968), pp. 589-609.

 

96

used for approximately half of the companies to determine
whether they provide additional help in explaining the
price ratio. NWC/total assets is used as a liquidity
measure supplemental to the quick ratio (like the quick
ratio it is not applicable to financial companies). Sales/
total assets is a performance measure to supplement rate

of return and dividends per share.

Method of Analysis

 

Time Series Analysis.--Of the 29 companies in the

 

sample, 24 have sufficient data to warrant charting the
variables over a period of years. The criteria used in
deciding which companies to chart are either 5 years or
more of data, or at least three years of data with substan-
tial variation in the price ratio. (All but one of the 24
companies charted meets the 5-year criterion.)

In charting variables for each year, the price
ratio used is as of June 30 of that year (or as near to
that date as obtainable). The independent variables include
income statement and/or balance sheet information--these
data are for the fiscal year ended prior to June 30. For
example, the June 30, 1972 price ratio is plotted at year
1972; for a company on a calendar year basis the period
used for other variables plotted at year 1972 is the year
ended December 31, 1971, for a company using a fiscal year

ending April 30, the 12 months ended April 30, 1972 would

97

be used; for a company using a fiscal year ending September
30, the 12 months ended September 30, 1971 would be used.
With the exception of Brown-Forman, absolute values
of the variables are charted. For Brown-Forman, relative
values of each variable are charted, with 1965 = 100,to
see whether relative values are advantageous. Relative
values do not seem to possess material advantages for this
study, so they are not used for the other companies charted.
After the companies are charted the problem remains
of analyzing each chart in a systematic way and summarizing
the results of that analysis. The method decided on is
comparing one independent variable at a time with the price
ratio. For each year plotted, the directions of change
(from the previous year) are noted, as well as whether the
changes are in the same or opposite directions. For the
entire period of the chart, the criterion is whether either
changes in the same direction or changes in the opposite
direction predominate. In order for possible positive
association to be inferred, the directions of change must
be the same at least 5 out of 7 times, or in the following
number of times when data for fewer than 7 periods are
available: 4 out of 6, 4 out of 5, 3 out of 4, 3 out of
3, or 2 out of 2. Inference of possible negative associa-
tion is based on similar requirements. Then, for each
variable, companies having positive associations are
enumerated, as are the companies showing negative associa—

tions between that variable and the price ratio. This

98

summarization makes it possible to determine the extent of
any support for the hypothesis which that independent

variable is used to test.

Cross-Section Analysis.--In order to supplement this
analysis, which is on a time series basis, a cross-section
analysis is used for each independent variable for which
there are at least 18 sample companies with adequate data
(available value for this variable for the latest year as
well as a 1972 price ratio). Each of these independent
variables is compared one at a time with the price ratio by
means of 2 by 2 contingency tables. Each of these tables
permits visual inspection of the association or lack of
association between one independent variable and the price
ratio. (The pairing of variables was done as explained in
connection with charting methods.)

For each table the cfld.square test is performed to
test the significance of the relationship between the two
variables. For tables where the cflfl.square test suggests
a significant relationship, cflfi.square is recomputed using

the Yates correction for continuity.1

Findings
Table 4.2 summarizes the analysis of the 24 charts

(which are located alphabetically in the Appendix) by the

 

1Samuel B. Richmond, Statistical Analysis (2d ed.;
New"York: Ronald Press, 1964), pp. 300-01.

 

99

TABLE 4.2.--C1assified Common Stocks,Companies for which
Charts Suggest Possible Association of One
or More Independent Variables with the Price

 

 

Ratio.
NWC/ Sales/

Market Rate of Quick Total Total Debt
Company* Price Return Ratio Assets Assets Ratio
Brown- .
Forman + NA NA +
Cannon - NA NA
H. Corby +
Cutter -
Duncan - - NA NA
Harding - - - -
Loblaw + -
Niagara + + NA
Plymouth + NA NA
Resorts + NA NA
Rolland —
Supertest + -
Traders - NA NA NA NA
Universal - NA NA
Versatile - - NA - +

 

+denotes possible positive association with price ratio
-denotes possible negative association with price ratio
NA variable not analyzed for this company

*Company names are listed in Appendix C.

100

method previously described. Only the 15 companies in the
table show possible association of one or more independent
variables with the price ratio. Only the six independent
variables listed in the table columns show association with

the price ratio for one or more of the 24 companies.

Market Price.+-One hypothesis is that the percentage

 

premium for control stock is greater when the market price

is relatively low, that is, that the association is negative.
Possible negative association is displayed in the charts

of six of the 24 companies. The six are denoted by minus
signs in the market price column of Table 4.2. On the

other hand, three of the companies (indicated by plus

signs in the same column) show positive association--opposite
to that hypothesized.

The strongest case for the hypothesis is illus-
trated by the chart of Harding. There are eight observa-
tions for each of the two variables, price ratio and market
price, giving Seven possible changes in each variable. In
all but the first period (1965-1966) there are changes in
the opposite direction. In addition, another look at the
chart reveals that the two times that the market price was
highest (around $22 in 1966 and 1972) the price ratio was
near its low--between 1.05 and 1.07; the year the price
was lowest ($9.00 in 1970) the price ratio was highest--
1.22.

Another company illustrating rather strong support

for this hypothesis is Traders. Of the seven periods,

101

five show changes in the opposite direction for the two
variables--all but the 1966-67 and 1971-72 periods. The
changes in both variables are quite large for the periods
1967-68, 1968-69, 1969-70, and 1970-71. This company had

a price ratio of less than 1 every year except 1972, which
means that every other year the voting stock sold for less
than the non-voting. A possible explanation for this
situation may be the ownership of 81% of the voting shares
of Traders by Canadian General Securities Ltd.1 One

result of this ownership could be strong control by Canadian
General which would reduce the incentive for Canadian General
to acquire additional voting shares, as well as lessen the
appeal to other investors. Another result of this owner-
ship could be a thinner market for the voting shares than
many investors would prefer.

Comparison of these two variables on a cross
section basis fails to lend any support for the hypothesis.
Table 4.3 shows that the 9 companies with price ratios of
.1.05 and over are quite evenly distributed between stocks
' priced under $15 and stocks priced $15 and over, as would
be expected for this sample if there is no association
between market price and price ratio. The observed fre-

quency in the upper left cell is five--close to the

expected frequency of 4.32 (%g x 9 = 4.32).

 

lMoody's Bank and Finance Manual, 1970, p. 1897.

 

_‘____._

102

TABLE 4.3.--Comparison of 1972 Price Ratio with Market Price,
25 Classified Common Stocks (Number of Companies).

 

Market Price

 

 

Price Ratio Under $15 $15 and over Total
1.05 and over 5 4 9
Under 1.05 7 9 16
Total 12 13 25

 

Chi square = .32; df = 1; not significant at .05 level.

Rate of Return.--One hypothesis is that the rate

 

of return is negatively associated with the price ratio-—
that a high price ratio is more likely to exist when the
rate of return is relatively low. Table 4.2 shows that
very little support for this hypothesis is found in the
analysis of the charts. Only two companies show possible
negative association, and in each case this is based on
relatively few time periods.

However, Table 4.4 shows that there is some possible
support for the hypothesis based on cross-section analysis.
Although not statistically significant at the .05 level,
possible negative association is visually apparent.
Although there are 11 each of lower return (under 10.0%)
and higher return (10.0% and over) companies, 5 out of 7
of the companies with price ratios of 1.05 and over are
lower return companies-~an observed frequency in the upper
left cell of 5 compared with an expected frequency of 3.5

11 _
(EX7_ 3.5).

 

103

TABLE 4.4.--Comparison of 1972 Price Ratio with 1971 Rate
of Return,22 Classified Common Stocks (Number
of Companies).

 

Rate of Return

 

 

Price Ratio Under 10.0% 10.0% and over Total
1.05 and over 5 2 7
Under 1.05 6 9 15
Total 11 ll 22

 

Chi square = 1.89; df = 1; not significant at .05 level.

Greater negative association between price ratio
and rate of return shows up when the same data are used to
construct another table with the class limit of the price
ratio changed from 1.05 to 1.10. The resulting table 4.5
shows that all four of the companies with a price ratio of
1.10 and over have a rate of return under 10.0%. This
observed frequency of four in the upper left cell compares
with an expected frequency of two (%% x 4 = 2). It appears
that the results shown in this table could be considered
statistically significant; the chi square test does not
show significance at the .05 level after applying the
Yates' correction for continuity. However, significance at

the .10 level is indicated by this test. Further, the

l O I l
Yates correction is said to over-correct.

 

lRichmond, op. cit., p. 301.

104

TABLE 4.5.--Comparison of 1972 Price Ratio with 1971 Rate
of Return,22 Classified Common Stocks. Revised
to Change Class Limit of Price Ratio to 1.10
(Number of Companies).

 

Rate of Return

 

 

 

 

Price Ratio Under 10.0% 10.0% and Over Total

1.10 and over 4 0 4 ;
4

Under 1.10 7 ll 18 j

Total 11 11 22 T;

Chi square (without Yates' correction) = 4.89; df = 1

Chi square (with Yates' correction for continuity) = 2.76;

df = 1

Significant at .10 level but not at .05 level.

Quick Ratio.--The hypothesis regarding this vari-

 

able is that the price ratio is greater when the company is
relatively liquid, as roughly measured by this ratio. Over-
all, based on both the time series and cross-section
analysis, this hypothesis does not seem to be supported;

in fact greater support seems to lie in the opposite
direction--that the price ratio is greater when the company
is relatively illiquid.

‘Table 4.2 reveals that based on the criteria estab-
lished for analyzing the charts, four of the companies seem
to lend support to the hypothesis. However, for one of
these four companies, Niagara, the chart shows that the
highest price ratios occurred in 1969 and 1970 when the

quick ratios were lowest.

105

Table 4.6 is the basis for the previous statement

that there seems to be greater support for an opposite

._ A3._...~..-—~ _..-—. _

hypothesis-~that the price ratio is greater when the

., -.~.-a——~ —~"

company is relatively illiquid. Although not statistically j

 

significant at the .05 level, possible negative association

between the price ratio and the quick ratio is apparent in ?

the table. Although an equal number of companies (nine) 1

have a quick ratio under 1.50 and 1.50 and over, six of

the nine companies with a price ratio 1.03 and over are

companies with a quick ratio under 1.50. Stated another way,

to observed frequency in the upper left cell is six, com-

pared with an expected frequency of 4.5 (I% x 9 = 4.5).
Possible association in this direction is even more

evident from Table 4.7, which is based on the same data

as Table 4.6, but with the class limit of the quick ratio

reduced from 1.50 to 1.00. Five of the six companies with

a quick ratio under 1.00 have price ratios of 1.03 or

over. The expected frequency in this upper left cell is

3 (lg x 9 = 3). The value of cflfl.square (4.00) is sub-

stantially higher than with the class limit of Table 4.6;

however, after recalculating cfifi_square with Yates' correc-

tion for continuity the value is 2.26, still not statisti-

cally significant at the .05 level.

Net Working Capital to Total Assets.--This ratio is

 

calculated for 12 of the companies in the sample and

analyzed on a time series basis. Only the chart for Harding

106

TABLE 4.6.--Comparison of 1972 Price Ratio with 1971 Quick

Ratio,l8 Classified Common Stocks (Number of
Companies).

 

Quick Ratio

 

 

 

Price Ratio Under 1.50 1.50 and over Total
1.03 and over 6 3 9
Under 1.03 3 6 9
Total 9 9 18
Chi square = 2.00; df = 1; not significant at .05 level.

TABLE 4.7.—-Comparison of 1972 Price Ratio with 1971 Quick
Ratio,18 Classified Common Stocks. Revised to
Change Class Limit of Quick Ratio to 1.00.

 

Quick Ratio

 

 

 

Price Ratio Under 1.00 1.00 and over Total
1.03 and over 5 4 9
Under 1.03 1 8 9
~Total 6 12 18
Chi square (without Yates' correction) = 4.00; df = 1
Chi square (with Yates' correction for continuity) = 2.26;
df = 1

Not significant at .05 level.

 

. . . .. .
A - A. h...) 1.. .1:

.' -. __..‘ _—_~

107

suggests possible association with the price ratio. The
Association is in a negative direction, as is the case with
Harding's quick ratio-~both opposite to the directions

hypothesized.

Sales to Total Assets.--This variable is calculated
for 10 of the companies in the sample. Table 4.2 reveals
that five of the ten companies show negative association
with the price ratio based on analysis of the charts.

Since this ratio is used as an additional measure of manage-
ment performance, the direction of association is consistent
with the hypothesis that the percentage premium for control

stock will be greater when management performance has been

relatively poor.

Debt Ratio.--One hypothesis is that the percentage
premium for control stock is greater when there is rela-
tively large unused debt capacity (as roughly measured by
a low debt ratio). This hypothesis means that price ratio
and debt ratio should be negatively associated. However,
no support for this hypothesis is found.

The time series analysis of charts, summarized in
Table 4.2,shows only two companies having association
between price ratio and debt ratio-~and the association is
positive instead of the negative association hypothesized.
For many of the other companies the debt ratio did not
change enough during the observed periods to permit analysis

of this variable.

 

1 Q: ...‘-"._~—" Lit .

108

Table 4.8 shows that the cross section analysis

reveals no association between debt ratio and price ratio.

TABLE 4.8.--Comparison of 1972 Price Ratio with 1971 Debt
Ratio,l8 Classified Common Stocks (Number of
Companies).

 

 

 

Debt Ratio
Price Ratio Under 15% 15% and over Total
1.03 and over 4 5 9
Under 1.03 5 4 9
Total 9 9 18

 

Chi square = .22; df = 1; not significant at .05 level.

. Dividends Per Share.--The criteria for determining
possible association in the analysis of the time series
charts are not appropriate for this variable; dividend
changes usually did not occur frequently enough to produce
enough changes in the value of this variable. Consequently

none of the sample companies meets these criteria. However,
further analysis of the charts does seem to suggest an
inverse relationship between dividends and price ratio.

For example, in the case of Chesapeake the two years of
price ratios greater than 1.00 (1970 and 1972) occurred
after the dividend had been eliminated. Similarly, the

only two years that Plymouth showed price ratios above 1.06

109

were the 1.16 and 1.13 of 1971 and 1972, after_the
dividend had been eliminated.

Rolland shows a similar possible association after
dividends were increased from $.30 in calendar years 1964
and 1965 to $.40 in calendar years 1966 through 1969. The
price ratio dropped from 1.21 in 1967 to 1.07 in 1968.
However, there is a time lag, as these dates indicate.

A much more persuasive case for the significance
of dividends is evident from the cross section analysis of
Table 4.9. Five of the six companies paying no dividends
in the prior year have price ratios of 1.03 and over. Only
7 of the 19 dividend payers fell in that price ratio cate-
gory. Stated in another way, the observed frequency in
the upper left cell is 5, while the expected frequency is
2.88 (5% x 12 = 2.88). The chi square test indicates

statistical significance at the .01 level.

Shares Ratio.--As in the case of the debt ratio,

 

the shares ratio for many companies did not change enough
during the observed periods to permit adequate analysis of
this variable on a time series basis. However, in the
case of Presidential the price ratio was generally higher
in years 1969 and later after the shares ratio had gradually
dropped from .59 to .33.

Table 4.10 shows that cross section analysis reveals

no association between shares ratio and price ratio.

110

TABLE 4.9.--Comparison of 1972 Price Ratio with Prior
Year's Dividends,25 Classified Common Stocks
(Number of Companies).

 

Prior Year's Dividends

 

 

Price Ratio No Dividends Dividends Total
1.03 and over 5 7 12
Under 1.03 l 12 13
Total 6 19 25

 

7.61

Chi square (with Yates' correction for continuity)
df = 1; significant at .01 level.

TABLE 4. lO.--Comparison of 1972 Price Ratio with 1971 Shares
Ratio,23 Classified Common Stocks (Number of
Companies.

 

Shares Ratio

 

 

Price Ratio 50% and Under Over 50% Total
1.03 and over 6 5 11
Under 1.03 6 6 12
Total 12 ll 23

 

Chi square = .05; df = 1; not significant at .05 level.

111

Book Value to Market Price.--This variable is

 

plotted on only a few of the charts but included are
Plymouth and Resorts which have wide variations in this
variable between substantially less than 1.00 and substan-
tially more than 1.00, as well as large fluctuations in
price ratio. However not even these two charts gave
evidence of association between these two variables.

Table 4.11 shows no evidence of association between
these two variables based on cross section analysis, with
the class limit of the price ratio set at 1.03. However,
when the class limit is raised to 1.10 and the same data
used, the result is Table 4.12. None of the eleven com-
panies with a ratio of book value to market price under
.90 has a price ratio of 1.10 and over. This observed
frequency of zero in the upper left cell compares with an
expected frequency of 1.91 (%% x 4 = 1.91). However, the
cflfi_square test with Yates' correction for continuity
indicates that the association is not statistically signi-

ficant at the .05 level.

Discussion of Findings

 

Rate of Return.--The hypothesis for which the study

 

of classified common stocks provides greatest support is
that the percentage premium for control stock is greater
when management performance (as measured by criteria such

as rate of return on equity and per share earnings growth)

112

TABLE 4.ll.--Comparison of 1972 Price Ratio with 1972 Ratio
of Book Value to Market Price,23 Classified
Common Stocks (Number of Companies).

 

Ratio of Book Value to
Market Price

 

 

Price Ratio Under .90 .90 and over Total
1.03 and over 5 6 11
Under 1.03 6 6 12

Total 11 12 23

 

Chi square = .05; df = 1; not significant at .05 level.

TABLE 4.12.--Comparison of 1972 Price Ratio with 1972 Ratio
of Book Value to Market Price,23 Classified
Common Stocks.Revised to Change Class Limit of
Price Ratio to 1.10 (Number of Companies).

 

Ratio of Book Value to
Market Price

 

 

 

Price Ratio Under .90 .90 and over Total
1.10 and over 0 4 4

Under 1.10 11 8 19

Total 11 12 23

Chi square (without Yates' correction) = 4.44; df = 1

Chi Squire (with Yates' correction for continuity) = 2.41;

Not significant at .05 level.

 

 

113

has been relatively poor. Rate of return shows statisti-
cally significant negative asSociation with price ratio,

which is in the direction consistent with the hypothesis.

Dividends Per Share.--Dividends per share is used

 

as an additional measure of management performance and
shows strong and statistically significant negative
association with price ratio, which is in the direction
consistent with the hypothesis. Companiespaying no
dividends in the prior year are much more likely to have
their voting shares selling at a premium of at least 3%
above their non-voting (or reduced-voting) shares than

are companies which paid dividends in the prior year.
Failure of a company to pay dividends may be regarded by
stockholders as constituting poor management performance
for two reasons. First, dividends constitute a tangible
return to stockholders, and omission of dividends may be
due to management's failure to earn enough on the company's
' assets, or to maintain sufficient liquidity, or both.
Secondly, even if earnings and liquidity are adequate, the
omission of dividends may be regarded (sometimes justi-
fiably) as intended to serve the interests of management
or controlling stockholders rather than other stockholders;
one cause of this situation may be management or controlling
stockholders who are in high income tax brackets and would
prefer to see funds retained for reinvestment in the hope
of lower-taxed capital gains, instead of paid out in

higher-taxed dividends.

 

114

Sales to Total Assets.--The ratio of sales to total

 

assets is used as a supplemental measure of management
performance on a time series basis for ten of the com-
panies in the sample. Five of the ten companies show
changes in this ratio which are predominantly in the
opposite direction from changes in the price ratio; in
other words, for these companies, when sales have declined
in the prior year in relation to assets, the price ratio
tends to increase from the previous year.‘ This association
is consistent with the hypothesis, because increases in
sales from the prior year are a popular measure of manage-
ment performance and a widely-publicized statistic, and
total assets often remains relatively constant from one

year to the next.

Book Value to Market Price.--The ratio of book

 

value to market price seems to show positive association
with price ratio based on cross-section analysis, although
the association is not statistically significant at the

.05 level. The direction of this association is consistent
with the hypothesis that the percentage premium for control
stock is greater when there is a relatively great possibil—
ity of profitable liquidation (as rOughly indicated by a
relatively high ratio of book value per share to market

price).

Methods Limitations.--One possible reason for lack

 

of greater support for other hypotheses may lie in the

*q :.-—;1 .
.. _-_.-.¢.— -gp—Ap-w» -- *- ‘-- "W, ' ' ‘1'”
1 . 11 A. s_ L L. ‘ A ‘ A ‘

 

115

limitations in the methods which were appropriate for the
type of data available in the study of classified common
stocks. Both the method used to analyze the time series
charts and the 2 by 2contingency tables used for cross
section analysis compare two variables at a time—-the
dependent variable and one independent variable. However,
many or all of the independent variables may be exerting
a simultaneous influence on the dependent variable; some
of these influences would be expected to occur in opposite
directions simultaneously. The result could be the off-
setting of, and therefore hiding of, the association of
any one independent variable with the price ratio.

It is also possible that certain of the independent
variables may be associated with the price ratio only
within certain ranges. For example, the greater associa-
tion shown by Table 4.5 as compared with Table 4.4 would
be consistent with a situation whereby a rate of return
of at least 10% tends to prevent a price ratio as high as
1.10, but not necessarily a price ratio between 1.05 and
1.10.

Similarly, the data analyzed in Tables 4.6 and
4.7 would be consistent with a situation where the quick
ratio is regarded as relatively unimportant unless it is
less than 1.00, and that a quick ratio at less than 1.00
would raise a caution flag that would impel shareholders
to attempt to get control of the corporation. A situation

such as this would be particularly likely to exist with a

116

variable such as the quick ratio where rules of thumb as to
"satisfactory" levels have wide circulation (though not

high regard).

Offsetting Forces.--A reason related to the one just

 

discussed is the possibility of offsetting forces at work
with regard to a given variable. For example, one hypothesis
is that high liquidity is a motivating factor for takeover.
Although this hypothesis is not supported by this part of
the study, it is possible that there is some such effect
but that it is hidden by the opposing effect of low
liquidity being regarded as a caution sign and therefore a
reason to attempt to get control. It is conceivable that
both relatively low and relatively high values of the quick
ratio encourage takeover, but for different reasons, and
that middle ranges of quick ratio provide no such incentive.
Another example of offsetting forces at Work is
the case of the debt ratio. One hypothesis is that the
premium for control is higher when there is a relatively
large unused debt capacity--as measured by a relatively low
debt ratio. Such an effect would logically occur when it
is felt by investors that opportunities for favorable
financial leverage have been neglected. On the other hand,
the existence of a relatively high debt ratio could also
impel stockholders to want to get control of the company.
In itself such a ratio might be regarded as a caution flag--

' as already discussed in the case of a relatively low quick

117

ratio. In addition, there is the financial leverage effect.
In this case financial leverage already exists--concern
about its effects would be the factor motivating takeover
attempts. Not only does the financial leverage caused by
debt increase the risk of loss, but it also increases the
degree to which earnings per share increases with a given
increase in earnings before interest and taxes. Therefore,
the existence of financial leverage might create a greater
incentive to get control of a company and try to improve

the earnings performance.

Nature of Securities Markets.--As mentioned early

 

in this chapter, the voting stock is more likely to be
closely held than the non-voting stock; one result could

be a poorer market for the voting stock than for the non-
.voting (or reduced-voting) stock. It is likely that many
investors would not be much concerned about voting power,
but would be greatly concerned with marketability. As a
result, the price ratio might be lower than would otherwise
prevail, and might become less than 1.00.

It seems necessary to consider the nature of the
securities markets, including the decision-making process
which investors use. One of the questions debated in the
financial journals in connection with the "random walk
hypothesis" has been the extent to which investors have
knowledge about factors relevant to their investment
decisions. The random walk hypothesis is based on the

premise that securities markets are efficient, with an

118

efficient market characterized by large numbers of compet-
ing, rational, profit-maximizing investors, and the wide
availability of important current information.1

If such efficient markets exist for many or most
securities, it is possible that classified common stocks
are an exception. The relative lack of marketability of
some classified common stocks (already discussed) could
be both a symptom and a cause of a market which is not
efficient--a cause in that once poor marketability exists
it tends to perpetuate itself because many additional
investors are dissuaded from entering the market because
it is poor.

What could be other causes of a market for classi-
fied common stocks which is not efficient? Lack of wide
availability of important current information could be one
factor. Relatively little has been published about classi-
fied common stocks, particularly about the price differ-
ential between voting and nonevoting (or reduced-voting)
shares.

' Attempts to generalize about the availability of
important current information in the securities markets
may be misleading because of the ways in which investment

information is generated and disseminated. Investment

 

lEugene F. Fama, "Random Walks in Stock Market
Prices," Financial Analysts Journal, XXI (September-October,
1965)! pp. 55-59.

 

 

119

information on a security, such as a brokerage house research
report, is often generated in response to already-existing,
widespread investor interest in that security. Therefore
more information is generated about those securities in

which there is large interest than in those in which there

is little interest. In many cases lack of interest in a
security may be due to lack of information about it; there-
fore the lack of information about a security tends to be
perpetuated.

Dissemination of investment information appears to
be similarly related to investor interest. Once information
is generated about a security, the extent to which the infor-
mation is disseminated appears to be related to existing
investor interest in that security. Investor interaction
appears to be a major factor in determining investor
interest. Brokers talk to one another and to investors
about a small proportion of securities, investors do like-
wise. This information seems to create a very high degree
of interest in a relatively small proportion of securities,
and the relative neglect of others. This varying interest
manifests itself in greatly varying trading volume in
relation to number of shares outstanding. Large volume of
trading in a security may result in further publicity such
as inclusion in a "most active" list carried by a large
number of newspapers.

It would appear that there is pronounced bimodality

in the availability of investment information about

 

‘1 '. .-
~. A. A) .L ‘; -n' .~. 1.
-
»
v . I u .

 

120

securities, due largely to the influences which have just
been discussed. If this is true, then the market for
securities may be segmented; one part may be relatively
efficient, the other part is probably not. Classified
common stocks would tend to be in the portion of the
market which is less efficient.

Such a situation might help to account for the
findings in the study which reveal that of the several
variables which seem to have a loqical basis for associa-
tion with the price ratio, the few which show some associa-
tion are those for which there is relatively great dissemi-
nation of statistical data. Even for relatively obscure
companies the Wall Street Journal publishes information
on sales, net income, and earnings per share. For most
companies this information is available on a quarterly
basis, and is published with comparative figures for the
year-earlier period. Major balance sheet figures are pub-
1ished at least annually. Dividend information is even
more publicized. Dividend declarations are announced in

a special daily feature of the Wall Street Journal; an

 

adjacent feature lists stocks which will go ex-dividend in
the next trading day or two. In addition, most daily stock
listings show the indicated annual dividend based on the
most recent quarterly dividend declaration. Not until

October, 1972, when the Wall Street Journal and some non-

 

financial newspapers began publishing a "PE ratio" column

for listed stocks, did earnings information begin to

1.4,. 1 1

 

 

approach the degree of daily prominence long held by
dividends.

As noted in the section on findings, dividends
show a far more significant association with the price
ratio than any other variable. The only other variable in
the cross section analysis reaching statistical signifi-
cance at even the .10 level is rate of return. The numer-
ator of the fraction used to calculate this variable is
net income-—one of the few widely-publicized pieces of
information previously mentioned. Net income is also the
key component for this variable because usually it is more
volatile than net worth which is the denominator and only
other component.

A study of small-bank stocks provides some addi—
tional support for this study's suggestion that when
information about certain usually-important factors is not
widely available, factors for which information is avail-
able become more important. It was found in that study
that dividends were very important in determining the
relative market price of the stock of a small bank and
that market price seemed to be more closely associated
with dividends than with earnings. It was suggested that
an important reason may have been greater availability of

dividend information than of earnings information.l

 

1James C. Van Horne and Raymond C. Helwig, The

Valuation of Small-Bank Stocks (East Lansing, Mich.:
Bureau of Business and Economic Research, Michigan State
University, 1966), p. 42.

-. _ .1. ". . _ _ ‘.
--—-— -'—- — - . 1 ' .
A ‘L 1.. x .23 “,“In; an: _ 1.

 

.‘l‘.

. ’4

+v‘“*‘; .1
.

\A . .
, . 1 H 4— .. o. a
1 .1 W. .,_,A v—V..—Va-—+—.~-l-._’_+
‘ , . . . ‘ . . v . ,

._..4_21‘-.’_.....__

CHAPTER V

1...1.'_fi__._._:.—.fl .

SUMMARY AND CONCLUSIONS

 

 

One purpose of this study is to explain the widely
varying stock price premiums found within each of the i
following three methods of obtaining or retaining voting 1*

control:

1. cash tender offers

2. private purchases of large blocks of stock )
3. classified common stock 1“

For the first two of these methods,regression

 

equations explain a substantial portion of the variation

in stock price premiums. Still, a considerable amount of
the variation remains unexplained. The number of indepen-
dent variables which may be used is limited. Further, some
relevant considerations would be difficult to obtain infor-

mation about, and some information which is available is

difficult to quantify sufficiently to be useful in this

. , n, . ‘ .
.1 1.; f. fl ‘, _. ..i ‘1‘. _ . A '1 . . A < A“ 4
n , q r . . 7 I ‘ V
- - ‘
- a .‘ . . 1 .

type of analysis. ‘ A
Although the combined ability of independent vari-

ables to explain the variation in price premiums is

important, the individual contribution of each variable is

also of interest. Some, such as liquidity measures and

122

123

measures of the extent of use of long term debt, are closely
related to decision topics in financial management which
receive considerable attention in courses from beginning
undergraduate to advanced graduate level. They also

receive attention in many investment courses.

Common Patterns Among the Three Methods

Price Ratio.--For all three methods of obtaining or

 

retaining voting control, premiums are generally found to
be paid for the type or amount of stock which facilitates
voting control, as compared with the price of otherwise
comparable stock (of the same company) which does not
facilitate control. The largest and most consistent
premiums are found for cash tender offers, varying from
1.9% to 46.7% with a mean of 18.2%. Next is classified
common stock, ranging (as of mid-year, 1972) from a dis-
count of 10% to a premium of 28% with a mean of 4.6%.
Private purchases of large blocks of stock show
the lowest mean premium (2.7%) but a very wide range--
from a discount of 68.6% to a premium of 37.5%. Consider-
able uncertainty and debate regarding the legality and
ethicality of accepting a premium for a large block without
making that premium also available to minority stockholders
probably accounts for some of the wide variation in this
last method. (The premium as measured by the variable
price ratio is the dependent variable in the study of each

of the three methods.)

 

124

Debt Ratio.--Debt ratio, the ratio of long—term debt

to total of long-term debt, preferred stock, and common
stock equity, shows a consistent pattern in the studies of
each of the three methods studied. In each case it does
not support the hypothesis that the percentage premium for
control stock is greater when there is a relatively large
unused debt capacity, as measured by a relatively low value
for this ratio. It does not enter a regression equation
for either cash tender offers or private purchases of large
blocks of stock, and shows no association with price ratio
for classified common stock.

Offsetting forces or alternate hypotheses may be
the explanation for lack of association of this variable.
Instead of a company being sought after (and a large
premium paid) because it has a low debt ratio, it may be
sought after because it has a high debt ratio-~representing

not the potential for financial leverage, but financial

leverage which already exists. This possibility is greatest

when unfavorable financial leverage exists, because the
incentive for turning the company around and changing the
unfavorable leverage to favorable leverage is greater than
if no financial leverage existed.

There may be a similar effect, which is more dir-
ectly related to this variable. A relatively high debt
ratio (perhaps only above some level which is popularly
considered normal) may serve as a warning signal to inves-

tors that their investment in the company is subject to

 

 

125

considerable risk. This signal motivates them to attempt
to get control of the company and either reduce the debt
ratio or reduce risk in other ways, such as by adopting

a more conservative policy with regard to investment in
assets. This last concept differs in that it is not
predicated on the existence of unfavorable financial

leverage at the time control is sought.

Liquidity Measures.--Measures of corporate liquidity

 

also show a consistent pattern among the three methods
studied, in that the hypothesis concerning them is not
supported. The hypothesis is that the percentage premium
for control stock is greater when the corporation is
relatively liquid (as roughly measured by the quick ratio
or current ratio). The quick ratio, used in the study of
cash tender offers, fails to enter a regression equation.
Current ratio, used in the study of private purchases of
large blocks of stock, is in the best regression equation
but shows moderate association in the direction-opposite
to that expected from this hypothesis.

In the study of classified common stock quick
ratio is also used; it shows no association on a time
series basis; on a cross section basis some association
exists in the direction opposite to that expected, but it
is not statistically significant. A supplemental measure
used for some companies in the study of classified common

stock is the ratio of net working capital to total assets;

_~_....
- .

 

. . . . .,
...._ A. ..<....., .

\. A. a
*1'“I. . - A
V ‘ . . -11.-

'A._.

- .,- ~..~——..—a. -.*—~..~. ~15. FHA: _; ’r_ ‘.
' ' I ‘- I. I .1 .1l_- 2 . . -
. . . ‘

\' ', ‘ .
A: A a) .1. _. ‘5'

 

126

it shows no association on the time series basis with
which it is used.

A factor in the failure of quick ratio to enter a
regression equation for cash tender offers is believed to
be its moderate positive correlation with another variable
which did enter: price range. Also, there seem to be
offsetting forces at work which are strong enough to cause
some association in the opposite direction from that
expected. A low quick ratio, especially below some level
popularly regarded as normal, is often regarded as a
warning sign to investors. Investors may be motivated by
this warning sign to try to get control of the company and

change its asset investment policies.

Correlation of Exchange with Market Price.-—In both

the non-financial companies studied for cash tender offers
and the companies subjected to private purchases of large
blocks of stock, there is a nearly-identical simple corre-

lation between the listing status and the market price:

-.45 and -.46 respectively. This correlation is consistent

with expectations that companies with higher-priced shares

are more likely to be listed on the New York Stock Exchange

than those companies with lower-priced shares, which are

more likely to be listed on another exchange or unlisted.

' ' x -~—-‘-7 .‘ .
..,..-.r+.~— ~,—._——fi.~ ~——ff 1 ,
. . 'I A... 1, I -, -v ..,_.‘,
_Y
n . , . A -

 

 

1‘ \ .
‘ -' .“‘fl .1. A a, 1.?
n
1 u

127

Inconsistencies Among the Three Methods

 

Rate of Return.--The independent variable rate of

 

return (net income as percentage of net worth) is used for

, _. _.
~u '7- ‘_

all three methods studied, but the results are inconsistent.

The hypothesis which this variable is used to test is that

 

the percentage premium for control stock is greater when

-. 1‘. L—J‘r—AV f
1. h .

management performance (as measured by criteria such as 1
rate of return on equity and per share earnings growth)
has been relatively poor. The variable does not enter a
regression equation for cash tender offers; one reason is
believed to be high simple correlation with another vari-
able (book value to market price) which does enter the
regression equations. However, another measure of manage-
ment performance, earnings direction, does enter the
regression equations for cash tender offers; this variable
is the direction of change of earnings per share from the
previous year; it may be a more sensitive indicator of
cause for stockholder dissappointment due to the wide
publicity given to year-to-year earnings comparisons.

Rate of return does remain in the regression equa-
tion for private purchases of large blocks of stock and
shows moderate association in the direction consistent with
the hypothesis. Although rate of return does not show
association with price ratio in the time series analysis
of classified common stocks, it does show statistically

significant association in the expected direction in the

128

cross section analysis of that part of the study. There-
fore, notwithstanding the foregoing inconsistencies, it
appears that there is general support for the hypothesis
that the percentage premium for control stock is greater
when management performance (as measured by criteria such
as rate of return on equity and per share earnings growth)

has been relatively poor.

Book Value to Market Price.--The ratio of book
value to market price proves to be an important explanatory
variable for the study of private purchases of large blocks
of stock. It remains in the regression equation and shows
strong association in the direction consistent with the
hypothesis that the percentage premium for control stock
is greater when there is a relatively great possibility of
profitable liquidation (as roughly indicated by a relatively
high ratio of book value per share to market price). Fur-
ther, New York Stock Exchange listing appears to have a
synergistic effect with this variable upon the price ratio.

This variable also enters the regression equations
for cash tender offers, but shows very weak association in
the opposite direction to that expected from the hypothesis.
In the study of classified common stock this variable seems
not to be important in the time series analysis; however,.
in the cross section analysis it seems to show association
in the expected positive direction, even though the associa-

tion is not statistically significant.

:1 u A: :3";

K. ._ ._ .

 

1 fizz-i: "é'.;{_11_...~ -. ..
. < \ l ‘

A..L_'.‘-

, '1 a
,. A a. J... 1—.‘..—..rw.....ar—n'——._ _...—_.'_. 1.... .—
‘ ' ' | > . a. o-) .
. , . . . \ .

129

There are two possible reasons for the greatly
differing influence of this variable between private pur-
chases of large blocks of stock and cash tender offers.
One reason is that the sellers of large blocks may be more
book value conscious than the typical investor who is the
potential seller in cash tender offers, and that they
therefore are more likely to take book value per share
into consideration in determining the price at which they
are willing to sell. Any such tendency wOuld be logical
because if the large block holders who are the potential
sellers have or can obtain control based on their holdings,
liquidation of the company is an alternative for them as
it is for the prospective purchaser.

Another possible reason for the differing results
is that private purchasers of large blocks may tend to be
more bent on liquidation than those who make cash tender
offers. A New York Stock Exchange listing may make com-
panies with a high ratio of book value to market price

more detectable.

Market Price.--One hypothesis is that the percentage

 

premium for control stock is greater when the market price
is relatively low. This hypothesis is supported by the
study of cash tender offers, although the association is
weak. However, this variable is also in the best regression
equation for private purchases of large blocks of stock but
with association (also weak) in the direction opposite to

that expected. With classified common stocks there seems

.._ ...,
I.-- «L— WW—vH-hy—‘y A A" -‘r L .
~ . . ._. . .J‘_.J- A_\.‘ ‘v

‘ .

 

‘1 .1, :1 _L'_11 -
‘_ ‘ 'r“

130

to be some possible association in the expected direction
based on time series analysis, but not with cross section
analysis.

A possible explanation for the differing results
between cash tender offers and private purchases of large
blocks of stock is greater sophistication of large block
sellers than typical investors who are the prospective
sellers in cash tender offers. The unsophisticated
investor would be susceptible to dollar illusion with
regard to price premium--he would look at the number of
dollars of premium instead of converting it into a per-
centage and emphasizing that. The large block sellers may
be more sophisticated and less susceptible to that illu-
sion. As shown by the example in the chapter on cash
tender offers, such an illusion results in illogical
behavior because of the effect of brokerage commissions
in open market sale--the alternative (for most investors)

to tendering.

Exchange.--The variable exchange is defined as the
listing status of the shares, whether New York Stock
Exchange, other exchange, or unlisted. It enters the
regression equations for cash tender offers and shows very
weak association with price ratio in the direction consistent
with the hypothesis that the percentage premium for control
stock will be greatest for unlisted companies and lowest

for NYSE-listed companies.

 

1 .'., LL ...._" -..-

131

This variable is not in the best regression equation
for private purchases of large blocks of stock. However,
New York Stock Exchange listing appears to have a syner-
gistic effect with book value to market price upon price
ratio. One reason may be the moderate negative simple
correlation of exchange with two other variables which are
in the best regression equation: price range and market
price.

There is another possible explanation based on the
rationale for this hypothesis which is that exchanges other
than the New York Stock Exchange (NYSE) are regarded as
less-efficient markets than NYSE,and the over-the-counter
market as even less efficient. The market price at any
given moment is regarded as a better measure of the actual
value of the stock to the extent that the market price has
been determined in a more-efficient market. Since the
purpose of the premium is to convince shareholders that
they are being offered more for their shares than they are
worth on the market, a larger premium will be required for
stocks trading in the less-efficient markets. Sellers of
large blocks of stock may be less uncertain of the value of
the shares they own then general investors, even when the
shares are unlisted. This lower uncertainty may be due to
greater soPhistication; it may also be due to inside infor-
mation. Since they have less uncertainty as to what their
non-NYSE shares are worth, a lower premium may be required

_to induce them to sell.

 

 

132

Price Range.--Price range, defined as the ratio of
prior year's high price to prior year's low price shows
inconsistencies similar to but more pronounced than those
of the variable exchange, probably for reasons which are
in part similar. The rationale for this hypothesis is that
a wide price range may make a stockholder more uncertain
as to what his shares are worth than does a narrow price
range. In order to overcome the greater uncertainty asso-
ciated with a wide price range, a larger premium is required.
For cash tender offers this variable enters the regression
equations and shows moderate association in the expected
direction. However, in the study of private purchases of
large blocks of stock the variable is in the best regres-
sion equation but shows strong association in the direction
opposite to that expected.

One reason for this difference of results may be
the greater sophistication of the private sellers of large
blocks, as compared with the typical investor who responds
to a cash tender offer. A wide prior year's price range
might be expected to create less uncertainty for a large
block holder as to what his shares are worth than it would
for a typical small investor. Inside information available
to the large block holder may play a part in this effect.

To the extent that a wide price range represents
greater risk in the view of the large block holder, and to
the extent that he does not want to hold a large block of

a risky security, he may be more anxious to sell the block

.. -..~ 4-d,—.¢---.'
v 4 l | I 4‘ A.
. , , 'n

 

 

133

. N. -
r.‘-,--(:',—) ‘ -
|

and diversify, even at a relatively low premium. Such an
effect could account for the strong negative association .3

found between price range and price ratio for private K3

a V... A A-“ V

l
r
t
l
A n.
purchases of large blocks of stock. v ,k
i
1

Other Findings

 

 

. . ._ . .‘
._ .~.4.., —-v-—-‘v . .
‘ ' ' u g»

‘4 ..A.~ .- . .
.. -...~_4_..__._,_..<

Cash Tender Offers.--In addition to those findings

previously discussed in this chapter, the study of cash

H- .._-_ ~_.___A.———-— -.
- I

tender offers provides others. Length of offer shows a 1
moderate negative association with price ratio, supporting i
the hypothesis that in a tender offer, the percentage pre- 1
mium is greater when the length of offer is relatively If
short (as measured by number of days from date of announce- :
ment of offer through original expiration date of offer).
'The rationale is that a short offer length will require a
large premium to move stockholders to act rapidly.
Management reaction shows very weak association
with the price ratio in the expected direction, meaning r
that the premium will be larger when opposition by manage- 9”
ment of the target company is (correctly) forecast. Lack
of a stronger association may be due to the reluctance of
some managements who oppose an offer to announce their
opposition, due to fear for their jobs if the offer succeeds.
The ratio of prior year's high price to market price
shows very weak association in the direction opposite to
that expected. One hypothesis is that the percentage

premium is greater when the market price is low in relation

 

134

to the previous year's high; the rationale is that inves-
tors will expect the price to "come back." The unexpected
result may be due to investors becoming discouraged because
the market price is low in relation to the prior year's
high, and welcoming the chance to dispose of their shares
at even a small premium above the current market price.
Shares sought shows a very weak association in a
direction opposite to that expected. The reason may be
that some offers for a low percentage of shares may actually
indicate great desire to have shares tendered and may there-
fore involve a relatively high premium. Shares purchased
shows unexpected negative association with price ratio,
which means that a relatively high premium is associated
with a relatively low ratio of the number of shares pur—
chased through the offer to the number of shares offered to
buy; as discussed in the chapter on cash tender offers, the

reason seems to be related to the one just discussed.

Private Purchases of Large Blocks of Stock.--In

 

addition to findings discussed in the earlier part of this
chapter, the study of private purchases of large blocks of
stocks provides some other findings. Purchaser is moder-
ately associated with price ratio in the direction consis-
tent with the hypotheSis that the percentage premium is

greatest when the purchaser is a corporation other than

the issuer, less when the purchaser is an individual, and

least when the purchaser is the issuer. The rationale is

. Ag ~7-~—~——:-u . . .

 

L. ZLL_L~_ -.
j, '4:

> A ’A , .
... a-.‘..x, A. .4, A) ., 7. .l a.
.
\
v n n y a . ~

-. ‘"-:—-*.#+.~—“’—~4» 7.. ;ffl
‘ ‘ A I I u‘ A.__4L_.L_. I‘. — v ‘ ,
. . . ‘\ , . .

 

135

that control of another corporation provides potential
benefits not available to an individual purchasing control,
for example, an alternative of merger with the purchaser.

Seller also shows association with the price ratio
in the direction expected, although the association is weak.
One hypothesis is that the percentage premium is greater
when the seller is a corporation than when the seller is
an individual. The rationale is similar to that just dis-
cussed for purchaser.

Block size is not in the best regression equation.
One hypothesis is that the percentage premium is greater
when the block is relatively large (as measured by the
number of shares in the block as a percentage of number of
shares outstanding).

The variable subsequent status also is not in the
best regression equation. The reason may be its relatively
high simple correlation with purchaser, a variable which is
in the regression equation. One hypothesis is that the
percentage premium is greatest for those transactions
followed by merger with a corporation other than the pur-
chaser, less for transactions followed by merger with the
purchaser, even less for transactions followed by a cash
tender offer, and least for transactions followed by no

such change in status of minority shareholders.

Classified Common Stock.--In addition to the findings
with regard to classified common stock discussed earlier in

this chapter, there are two additional findings for this

 

r L., L;

136

part of the study. Shares ratio shows no aSSociation with
price ratio either on the basis of time series analysis,
or cross section analysis. One hypothesis is that the
percentage premium of the voting shares is greater when
the ratio of the number of voting shares to the number of
non-voting (or reduced voting) shares is relatively low.

Dividends per share is found to be an important
factor in the size of price premium on voting shares. On
a time series basis changes in dividends usually did not
occur frequently enough to produce a sufficient number of
changes in the value of this variable for firm conclusions.
However, a much more persuasive case comes from the cross
section analysis, where it is found that companies paying
no dividends in the prior year are much more likely to
have their voting shares sell at a premium above the non-
voting (or reduced voting) shares than companies which
did pay a dividend in the prior year; the results are
statistically significant.

It is believed that the relatively great importance
of dividends in the case of classified common stock is
attributable to the nature of the market for classified
common stocks. Relative lack of marketability, and lack
of wide availability of important current information may
cause dividend information, which is both widely available
and broadly disseminated, to receive disproportionate
emphasis as a measure of management performance, as well

as emphasis as a tangible return to the stockholder.

 

h...

. n'y I
, 4 J ... .5 A‘
«
.
4 v 9 y i v

BIBLIOGRAPHY

137

 

' ' I} A
. ‘__4._

BIBLIOGRAPHY

Books

Dellenbarger, Lynn E. Common Stock Valuation_in Industrial
Mergers. Gainsville: UnIVersity of Florida Press,
1966.

Draper, N. R. and Smith, H. Applied Regression Analysis.
New York: Wiley, 1963.

 

Ezekiel, M. and Fox, K. A. Methods of Correlation_and
Regression Analysis. 3d ed. New York: John
Wiley & Sons, Inc., 1959.

 

McCarthy; George D. and Healy, Robert E. Valuing a Company.
New York: Ronald Press Company, 1971.

 

Manne, Henry G., ed. Economic Policy and the Regulation
of Corporate Securities. Washington: American
EnterprISe Institute for Public Policy Research,
1969.

 

Mills, Frederick C. Statistical Methods. 3d ed. New
York: Henry Holt & Co., 1955.

 

Richmond, Samuel B. Statistical Analysis. 2d ed. New
York: Ronald Press Company, 1964.

 

Van Horne, James C. and Helwig, Raymond C. The Valuation
of Small-Bank Stocks. East Lansing, Mich.: Bureau
of Business and Economic Research, Michigan State
University, 1966.

 

 

Articles

Altman, Edward I. "Financial Ratios, Discriminant Analysis,
and the Prediction of Corporate Bankruptcy."
Journal of Finance, XXIII (September, 1968), 589-609.

 

Conolly, Leonard J. "Perlman v. Feldman and the Sale of
Control-~A Brief Reconsideration." Business Lawyer,
XXVI (April, 1971), 1259-70.

 

138

 

 

139

Fama, Eugene E. "Random Walks in Stock Market Prices,"

Financial Analysts Journal, XXI (September-October,

Hayes, Samuel L. and Taussig, Russell A. "Tactics of Cash
Takeover Bids." Harvard Business Review, VL (March-
April, 1967), 135-48.

Manne, Henry G. "Mergers and the Market." Journal of
Political Economy, LXXIII (April, 1965), 110-20.

 

Taussig, Russel A. and Hayes, Samuel L. "Are Cash Take-Over
Bids Unethical?" Financial Analysts Journal, XXIII
(January-February, 1967), 107-11.

Investment Advisory Services

Moody's Bank and Finance Manual. New York: Moody's
Investors Service.

Moody's Industrial Manual. New York: Moody's Investors
Service.

Standard & Poor's Called Bond Record. New York: Standard
- & Poor’s Corporation.

Stock Guide. New York: Standard & Poor's Corporation.
Value Line Investment Survey. New York: Arnold Bernhard
& Co., Inc.
Other Publications
Bank and Quotation Record.
Brown v. Halbert, 76 Cal. Rptr. 781 (Cal. App. 1969).

Christophides v. Porco, 289 F. Supp. 405 (D. C. N. Y.
1968).

Financial Post.

U. 8. Securities and Exchange Commission. Official Summary
of Securit Transactions and Holdin 3. Vols. XXXII,
XXXIII, XXXVII (I936, I937, 1971). Washington,
D. C.: Government Printing Office.

Whitman, Martin J. The Strategy of Tender Solicitations.
New York: Blair & Co., Inc., 1967.

 

r—v—

\. A V
. , . A. w . A-
, A, . ,__,._,_.,. -_._,..,._,__.. -_l
. x' . v . ,

. . .
-r 5, .n‘, _~’ .. '7) .-Y
, V
.

140
General References

Books

Austin, Douglas V. and Fishman, Jay A. Corporations in

Conflict--The Tender Offer. Ann Arbor: Masterco
Press, Inc., 1970.

 

Baker, Ralph J. and Cary, William L. Cases and Materials on

Corporations. 3d ed. Abridged. Brooklyn: Founda-
tion Press Inc., 1959.

Baum, Daniel J. and Stiles, Ned B. The Silent Partners:
Institutional Investors and Corporate Control.
Syracuse: Syracuse University Press, 1965.

Bauman, W. Scott and Klein, Thomas A. Investment Profit
Correlation. Ann Arbor: The University of Michi-
gan, Bureau of Business Research, 1968.

Choka, Allen D. Buying, Selling, and Merging Businesses.
2d ed. Philadelphia: American Law Institute, 1965.

Harvey, John L. and Newgarden, Albert, eds. Management Guide

 

to Mergers & Acquisitions. New York: Wiley-
Interscience, 1969.

Mace, Myles L. and Montgomery, George C., Jr. Management
Problems of Corporate Acquisitions. Boston:
Harvard University, 1962.

 

Manne, Henry G. Insider Trading and the Stock Market.
New York: The Free Press, 1966.

Rabkin, Jacob and Johnson, Mark H. Federal Income, Gift
and Estate Taxation. Vol. IV. Mathew Bender, 1969.

Standard Research Consultants. Corporate Security Values

as Determined by the Tax Court. 2d ed. New York:
Standard Research Consultants, 1966.

Wyatt, Arthur R. A Critical Study of Accounting for Busi-
ness Combinations. Accounting Research Study No.

5. New York: American Institute of Certified
Public Accountants, 1963.

Articles
Andrews, William D. "The Stockholder's Right to Equal

Opportunity in the Sale of Shares." Harvard Law
Review, LXXVIII (January, 1965), 505-63.

 

i,£.x.. -,_.
, . . , . .~.

- n' «3*:

. k
\

' g - - AA .

, ._. A- b—-—;— A..--—---o~hdw—fr fr ‘fj'
. . . . a. skim—47A. ‘A ... -
. , . , . . :

 

'., L, 12““ —:V .
--. .., .

._.l . .___._.
. ' . . . .
_ A V . .M . A-“
. . , ‘ ‘ ‘

.— v—o I —-A

R‘,
, .— .
.. 4 .- *v“ —- 94 ”7’
K . - . . . l
A ‘ 5 ’ l . .

. “ ‘. ‘,
_ -,A _., .5 .4 .5 a,
. .
. \
4 t v . .

141

Berle, Adoph A. "Control in Corporate Law." Columbia Law
Review, LVIII (1958), 1212.

 

"The Price of Power: Sale of Corporate
Control." Cornell Law Quarterly, L (Summer, 1965),
628-40.

 

 

Berton, Lee. "Winning Losers." Wall Street Journal,
January 27, 1969.

 

Bridwell, Rodger W. "A and B Equities." Barron's,
November 11, 1957.

 

Byrne, Harlan 8. "Merger Come-Ons." Wall Street Journal,
June 6, 1969.

 

Hayes, Samuel L. and Taussig, Russell A. "Cash Takeovers
and Accounting Valuations." Accounting Review
(January, 1968), 68-74.

 

Hill, Alfred. "The Sale of Controlling Shares." Harvard
Law Review, LXX (1957), 986-1039.

 

Javaras, George B. "Equal Opportunity in the Sale of
Controlling Shares. A Reply to Professor Andrews."
University of Chicago Law Review, XXXII (Spring,
1965), 420-28.

 

Jennings, Richard W. "Trading in Corporate Control."
California Law Review, XLIV (1956), 1-39.

 

Manne, Henry G. "The SEC and Stocks." (Letters to the
Editor.) Wall Street Journal, November 5, 1969.

 

"Some Theoretical Aspects of Share Voting."
Columbia Law Review, LXIV (December, 1964), 1427-45.

 

 

 

Meyer, James E. "Determining the Optimum Tender Offer
Price." Financial Executive, XXXVII (August, 1969),
70-76.

Rukeyser, William S. "Getting Tough with Tenders."
Fortune, LXVI (August, 1967), 108-10.

Rustin, Richard E. "Big Board to Study Curb on Firms'
Efforts to Protect Themselves Against Take-Overs."
Wall Street Journal, February 20, 1969.

 

Sheehan, Robert. "Proprietors in the World of Big Business."

Fortune, June 15, 1967, pp. 178-83.

Soldofsky, Robert M. "Classified Common Stock." Business
Lawyer, April, 1968, pp. 899-902.

 

_ -‘.A-.._.. .__.. WA. ‘
«a -

 

142
Sommer, A. A., Jr. "Who's 'In Control'?--S.E.C." Business
Lawyer, XXI (April, 1966), 559-93.

Stabler, Charles N. "Takeover Strategy." Wall Street
Journal, January 26, 1970.

 

Vance, Jack 0. "Is Your Company a Take-Over Target?"
Harvard Business Review (May-June, 1969): PP. 93-98.

Wall Street Journal. January 7, 1969. "Corporation Mergers
Seen Continuing Rise from '68 High of 4,462."

. April 2, 1970. "Merger Activity Held Even in
First Quarter from '69, Study Shows."

 

. January 6, 1970. "Mergers Surged 37% in 1969,
Says a Study by W. T. Grimm."

 

 

. October 10, 1969. "Rule Adopted Barring Tender-
Bid Makers from Buying Same Stock Outside the Offer."

. March 24, 1970. "Top Executives and Staff
Men Tend to be Most Vulnerable in Mergers."

 

Other Publications

Michigan State University Agricultural Experiment Station.
Calculation of Least Squares (Regression) Problems
on the LS Routine. STAT Series Description No. 7.
East Lansing: Michigan State University Agricul-
tural Experiment Station, 1968.

Michigan State University Agricultural Experiment Station.
Stepwise Deletion of Variables from aLeast Squares
Equation. STAT Series Description No. 8. East

Lansing: Michigan State University Agricultural
Experiment Station, 1969.

Michigan State University Agricultural Experiment Station.
Stepwise Addition of Variables to Form a Least
Squares Equation. STAT Series Description No. 9.
East Lansing: Michigan State University Agricul-
tural Experiment Station, 1969.

U. S. Congress. Senate. Subcommittee on Securities of
Committee on Banking and Currency. Full Disclosure
of Corporate Equity Ownership and in Corporate
Takeover Bids. Hearings, 90th Cong., ISt Sess.
Washington: Government Printing Office, 1967.

 

. “14:; 'AI:
\ ‘.

.1:—-tl_ _ 1'

‘ __~_—-.~7-. ‘ ,
x. < .
-dr—-‘— -—-'—-.——A-o-—Ay— ‘7'“.1'14
A A— 'AH-_.. A ~, . ,, .
'- - - . - ‘

;._ '4—‘v

 

143

U. 8. Securities and Exchange Commission. Compilation of

 

Releases Dealing with Matters Frequently AriSing
Under the Securities Act of 1933.. Washington:
Government Printing Office, 1965.

. Compilation of Releases Dealingywith Matters

 

Arising Under the Securities Exchange Act of 1934
and Investment Advisers Act of 1946. Washington:
Government Printing Office, 1968.

. General Rules and Regulations Under the Securi-

 

ties Act of 1933. Washington: Government Printing
Office, 1966.

. Securities Act of 1933 as Amended to October

 

22, 1965. Washington: Government Printing Office,

1968.

. Securities Exchange Act of 1934 as Amended to

 

August 20, 1964. Washington: Government Printing
Office, 1965.

 

. General Rules and Regulations under the Securi-
ties Exchange Act of 1934. WaShington: Government
Printing Office, 1969.

A ’_¢ _ .
r’“:‘£__ 1 ‘ .. .
‘ , . - . ~ 1

,- 1' _ t;

_.L—_ I

.. ~‘v—‘A ._ _. ’
‘ l -9 ~.. ,

- ‘.- ~—~a- ¢--.'-- H- ~ wav- - v" 4.
< . . . . . .J‘. 4- : ~'- ‘
. , . . 5 . ,

 

‘.~—-—-'- ... -
5 .

v , , . -1
A A‘ ‘1...‘, __._Av
.. ‘4 '1 o- , " “I- h " —‘ " ‘.
. . ._,fi_o¢ A .--

n' A -
.. _+ —.-
‘ 1 .
y .

APPENDICES

144

145

 

o.vma m~.~m oo.ov .ou unsaasm cmoflumaa cum! mouamuoa on
~.-H mm.¢~ oo.am .ocH umuxmoummououmanmflz mmumanaa m~
H.mMH m~.na. oo.v~ .mm ..ocH mofluwcmmz museuaa «m
«.owa om.m~ oo.ov noses comna>mau>maumm mmumauaa mm
m.mma om.- oo.om .6ua 1.0.ov :coo mouhauoa -
m.hoa om.m¢ oo.om .wuq ..oo a 1.03 umzmum mmnoaua Hm
m.¢HH oo.va oo.ma .oo nouns smnuamz moumaum om
m.sHH oo.mm oo.ov mmanpmsccH mafia “mama monmwnm ma
m.mHH mA.HN oo.m~ .muoo romeo Hoflummsm «mum-» ma
m.aaa m~.¢a oo.na ago magmpmmm> osmaomm mmummam RH
m.maa m~.vm oo.mm .oo shamaomam oncouuomam mmumaum ma
A.H~H mm.ha oo.a~ .euq 1.6.03 caoo monoauoa ma
«.mHH mo.aa oo.o~ .ocH .ummzom mmumaun «a
m.¢aa om.om oo.mm .oo mcflmmwam :maummm mmuoum .ma
m.oaa mm.ms oo.om .06 a “scans mmuaum NH
m.mHH oo.hm oo.qv ammummama cogs: cumummz mmuwum as
m.oma mm.ms oo.mm .moo unmesupmcH Hoamma mmuaaus OH
H.bma oo.~m oo.mm Hmmpm caaamsmq a mmaon mmuoaum a
m.aoa oo.~m oo.~m .00 noon a ammm magnum muuaaum m
m.oHH mm.sm oo.oq uuoz amuse mousse A
A.moa oo.Hm oo.em .oo mmanaa>mo mounum m
m.vaa oo.on oo.om .muoo wua>mao moamum m
m.moa om.mm m~.vh mcflnmmocmoum amoflumaa mmumun v
m.NHH oo.m oo.m .oo ummumo musnxom mmumaum m
m.maa om.hm oo.om .oaH .cmusmsamzumODmmnsmm mmumum m
m.aoa mm.ov oo.as .oaH .mmacso a pummmouw mmumaum a
Oflumm m0 Hum mOHHm Gmwwfiohdm xOOpm quEwocfiogd cuwflgz
moflnm umxumz Hmmmo noflsz CH acmmEoo mo mEmz mo mama

 

QMHQDBm mammmo mmDZmB mmdu

4 xHQmemd

 

146

renumq mumnssz .Hvua mumnssc

F -A JI.1.! .4! 4.4L)

.pmms uos mum am can

.hh

mo mumwmcoo m msouw .mnmbss: m>oam Ham mo mpmwmsoo m msonwx

 

b.00d
N.0HH
v.vva
N.0HH
b.0va

0.~0d

0.0HH
0.mNH
H.0HH
H.50H
m.v0H
v.hHH
N.0HH
h.HNH
0.NHH
m.0NH
m.0HH
m.mHH
5.00H
m.m~H
0.0HH
0.00H
0.HNH
5.00H
5.0HH
m.wHH
v.0mH

00.mh

00.0HH

00.mv
mh.mm
00.mv
00.0m
mN.mN
00.0N
0m.vv
00.0w
mm.0m
mm.hm
mb.HH
mm.va
00.Hm
m0.mH
mh.mN
0m.mv
mm.mm
0m.mm
00.0w

ma.mHH

00.nm
mm.am

mm.vNH

00.m¢
00.mm

00.00
00.0MH
00.mm
00.50
00.00
00.mm
00.0m
00.mm
00.mm
00.00
0m.~m
00.Nm
mm.NH
0m.hH
00.mm
00.0w
00.0m
00.Nm
0m.mm
00.m0
00.0w
00.mNH
00.mv
00.vm
00.mvH
00.0m
00.0m

.osH comm a Haoopmz
.Hsmmd mqu ummzlummnw
.62 .uflmommo a mafiamcflm
.oo .mcH wmaA amusmfissoz
wmaA was wuflsfimpsH .mmmz
.oo .msH .smmz mocmofl>onm
mmaq mmumum sumpmm3I.MHHmo
.Ed .00 mocwHSmsH UmuHcD
.oo wocmusmsH um>ocmm
OUGMHSmsH madam macaw
mucousmcH maamm macaw
.QHOU mosmswm Hmumsmw
.pun mmflumzmum unflomsmu
.wuq mooom mpmamm

.muoo meoym mmaflm

.ou masmfisuumsH Had

.osH mmfluumspsH OED

.osH .cBopsusom

.ou manmo a wuflz xoameflm
.muoo ummuonflm woman

.00 w HsoEud

.muoo Esmaouumm mpmumad
.ocH mocflq Hwfi cuoummz
.30 a xmuom mmumum topaz:
.ou HHO Hflmaocflm

.oo HMOflusmsoumm swam
.mhoo umanm

mmlmlm
mmlomlm
mwlodlm

mmlmalaa
mwlmIH
0010MIm
mmINHIm
mmlwlm
mmlmalw

mmIMIm
mmlmmlm
wwlbalm

mmlmlm
mmlvalv
mmlmmlv

mmlmlw

mmlmlm
mwlwmlm
mmlmmla
mmlvNIH

A00IHMIH
mmlmalm
mmlmlmd
mmlmlma

mmINHINH
molmlaa
mmImHIHH

v0
mm
mm
H0
00
mm
05
mm
vb
Mb
mm
an
aw
0v
mm
mm
5m
mm
mm
vm
mm
mm
Hm
0m
mm
mm
hm

 

147

APPENDIX B

LIST OF TRANSACTIONS: PRIVATE PURCHASES
OF LARGE BLOCKS OF STOCK

 

 

Observa- Date of Name of Company in Market Price
tion No. Transaction which Stock Purchased Price Ratio
1 11-22-65 U.S. Consumer Products $ 9.88 114.9%
2 12-9-65 Elgin National Watch Co. 10.25 119.0
3 1-31-66 Sheller Mfg. Corp. 31.50 107.4
4 2-11-66 Flexible Tubing Corp. 9.50 110.5
5 3-7-66 Burnell & Co. Inc. 3.50 31.4
6 4-12-66 DWG Cigar Corp. 14.50 120.7
7 4-15-66 Paramount Pictures Corp. 71.38 116.3
8 5-15-66 Rimrock Tidelands Inc. 8.00, 137.5
9 12-1-66 Diners Club 19.25 101.9
10 12-8-66 Dayco Corp. 23.75 106.9
11 5-1-67 Shattuck (Frank G.) 14.00 114.3
12 5-11-67 Gorton Corporation 23.50 106.4
13 5-22-67 Glidden Company 29.75 104.2
14 6-1-67 McIntyre Porcupine Mines 39.00 105.0
15 7-5-67 Pathe Industries Inc. 2.38 84.2
16 7-20-67 Cudahy Corporation 9.25 99.0
17 8-2-67 Columbia Pictures Corp. 46.88 98.1
18 9-1-67 Metro-Goldwyn Mayer Inc. 55.00 107.3
19 9-7-67 Fargo Oils Ltd. 4.13 116.4
20 10-10-67 Universal American Corp. 21.50 127.9
21 11-2-67 Cott Corporation 11.00 102.3
22 3-17-69 Aurora Products Corp. 16.75 74.6
23 12-22-69 Computer Technology Inc. 17.25 91.3
24 8-1-70 All-Tech Industries Inc. 1.63 111.7
25 8-10-70 Williams Brothers Co. 24.63 90.4
26 11-6-70 Fotomat Corp. 1.75 100.0
27 ll-20-70 RCL Electronics Inc. 2.25 75.1
28 12-11-70 Ralston Purina Co. 24.38 100.0
29 3-18-71 Central Foundry Co. 11.25 133.3
30 6-1-71 Hexagon Laboratories Inc. 2.00 100.1
31 6-22-71 Nachman Corporation 8.50 117.7
32 6-25-71 Bally Mfg. Corp. 25.00 62.0

 

148

APPENDIX C

LIST OF COMPANIES: CLASSIFIED
COMMON STOCK

American Maize Products Co.*
Beneficial Standard Co.*
Bowser Inc.*

Brown-Forman Distillers Corp.
Cannon Mills Co.

Chatham Manufacturing Co.*
Chesapeake Life Insurance Co.
H. Corby Distillery Ltd.
Cutter Laboratories

Duncan Electric Co. Inc.
Harding Carpets Ltd.

Harvey Hubbell Inc.

Investors Diversified Services Inc.
Loblaw Companies Ltd.
Maclaren Power and Paper
Molson Industries Ltd.
Niagara Wire Weaving

A. C. Nielsen Co.

Northwest Engineering

Pembina Pipe Line Ltd.*
Peoples Credit Jewelers
Plymouth Rubber Co. Inc.
Presidential Realty Corp.
Resorts International

Rolland Paper Company, Limited
Supertest Petroleum Corp.
Traders Group Ltd.

Universal Telephone

Versatile Manufacturing Ltd.

*Company not charted.

.A A__.,_4’.~4~«‘A:—-vh _

 

15.4%.;11. v.

APPENDIX' D

149

I V.. -__~._._.._._.._—. I

 

b #- L—L —~—-" - .._ -

 

 

 

 

 

 

 

 

 

 

_ . . .
..... . . A

A.

. .1 :1 Ex 0. {A f 1.8.461» A sum. YVVRRQ ITTAlw! ... A III .. +1

[414‘].

\.

 

.. P 4* I? «h L . <- . . . . .9.
c . . . . . 4! . V

,
g
o
.

~'!
-4

1

r' 'T‘

.11.. 4.-

[4

. ”__f—1—
9 4'—4 Q -$-—
-‘ -.- .

w

- “'T_'—‘."'T“F—""!’

>b———¢-—‘—-—6—~4 — 1.7.4

.,.

9 4 *4-1—0-4 +4— .-4-_.

4

_ _
44.! 4 4 1.144
404 4+: 4.444 9 y
4.4

. m . . .
.4424 J4+414442PI 4+. 4.

. I 44
. -4 -44
.1. 4. -. 44.44.4444,. 4
44+44+4444 4.4.4.? 444444
4,- 4+- » 41.41.44 4444 4.

- 4 L. r- 4 14,444+ .+

.4, 4,444+- +4+ Ar 1 .A L, 44444
74 A .9. -94. 94.44.? I49

. . .
.
4,04 494. !.94. 4.. 4 494,4 .,
.4! 9 4. “9 n4 4.4 4 Lv 4

444 F. >4
_

40449.49 .4.

. _
!! 44%| 4.4.1..

1

494444 .9449.
In a 49. IO! .V V..+4.4! .0-

444..A1 LI. 4. +44 4

 

 

 

 

 

 

 

 

 

 

 

I? «I 4 4.4444 444-4 . 4 44 44 444% L -
.4 +34 -- 4 .

4+. r44... 492446: 64 .4 -4444
4V4~441 ., 4. _

4.. _ H _ -

.4

4 4 4 ‘Q 4
___.,__.._.,.4 1 1 #- —1-—q»

4'"! 4’4

4 a . ' ;
.->‘_ I ‘_
g

__'_ -.'
.

4..._ ‘

,.

4

1..

-1..4.l..L.-4.4 1 L4.

 

I 4.440 4} .I .
—
a 44494. 49.49444 .44
s
4 441414944) 494445444, 4
u v 4 I h
4.4. 4.4 J

9 4444f44449L .4 4.4 .44...
.

44449444494449 . 44

 

 

0494.94

4. 4. .- +444?
.‘ 44444444.‘-lo.4§ 544444.444 F 44-54. 44.1
.4 J 4. .494 4... ..5. .54qu .9. 9 94.4 4‘ .4 4. 444

-.¢. 0 + $. 44.94445 .9, 6.. o .
. -- .. .L. 4 4. . 9. 4.4.4.4; -
. ~ 4.. . + + _ 4 . +

~ 4...»... . . . 4 .

 

v 64., o. 9

 

 

t

4 - , ~
4 4
‘
.. .

. ~ ‘ . ..

....\ 4%.: xi.» .4

.( tfi {CK\

.: « ‘ +41...

“........:..‘..

 

L . _

1161161 61 4. 1 4- 61 4| .1 6|
. . . _

1 1 11.- .1 1 .- . L-

16L..-

H

6. 1.1
. 61 6 6
6 616.11.1410 1.1

.4!-.-1111 . - 1.

_
‘I- .914. ..

1. O. O O 4 1I|4

 

_ .
b. 6.416.6I 6 6 1| W1 .. +1

+11%-.. ..L1616-
. _
..1 6.11.116 6

- -.- -..|-1| s . , . ..- A 11.7.74 6-.

1+. .. 1.161141
4.1».17116..6.|+141 6-1

_. .
6_|16...6.1.5116.| 4. v 4. 61.? .1164.

.
. F.16I 6161--.4 4. fl .. 1+l-.fl.4161L16 619161
1.11? 1 o 6 o . .. . . . . . . - . . -.o. 6 61+1.6-L-.%1416.11..61
11111 6. . . .14.. -.6 6 .114. .1... 6.. . . .
.1 -.I.1-1i‘ |O -.6 . . . . | |.. . . _ - o . .- 4 I4. 6.- +1 1+. + .61. . Ll .

6.. ..... ..6. 41.416. 6.14. 4

I61!

 

«#L...ll...oll.L1-.4..-

01.611 .01.

1.11.11. H. L .. . - +1.

 

_ . . .
611.6116 $116.11 .61 .6 . I . . . bf .
.-6 4 6 I 616.161 ‘16. . . . 6 . . . .. .. . Io. . O .¢.|.4. 5311141

.1. 111+ .6 L...

41+ .. L . 1414....
.661. . k

. c .1..o1+| .o o... F

 

.‘

. \

~ . 6 ~ 4
6 L . . .
L

43“ lfikuk 6|. .! -8

 

|51_'1.|'|h 5.1.114

4 5 .6. 4. 1

.14-51.16.. 4

 

14-14Ii5I14II6 .515 .61.|||
5.1.1 5..
5 4. H11): .1144 5.

_
6.5.1+0115.|.6:.| 51

15||+|6_| :.-.+.14-.6I.

.4511149

[61 -1 51:4!

. . ..
16. 6115151516141 5.!

 

5| 1.411141414114151
. .. . . _
-11.- -r 111

.
. .

11111I 41.51611615- I 5- 615- . g .1. .

.
.1 16.1.6161|||6 6. . - 1. 5- 1.11.1

. .
1' .111...I.QI 1' l6 5! ' 1.161151511511415:
. .

. _ .
15.-14141514... . 141 . rIFV 61154111416.-
16 9| 5I||5i.k|1+15| 9 5| 415191|F16.1. T.+|14.11151 511.4115...“
_ . . . . H.
v6 .5-..’.5I...V.4| 6 5 4151616. . .15- 1.5151151
. . . _
141 5.. .16. 5 I 151-Ql 1| 414|||£ +l 5111.11 5. O b.11. (1.16- 115115

. ||1 1. 6 ¢ 161.6 1 6 5.6 o 6.15 .

. . .. v -
t 15 | n .6 6 6| 9 .6. .a 6. 4 .. ¢ 6 . . v . . . ~
. . . .

I 4
. 5 .
. . .. o

 

—‘ 5....H

 

> a.
M
.fi-

 

1-
'7
~11 ‘1
6.
3 H
‘4'
.
I
|
u
v
1 .
4
1‘.
'9
'v
1

 

 

 

 

flniummun

285

1

[M

T

111'

v”
mm
UIIIIO
Em

HHNMIUHW

30

1293