v .. IIIIII

WK rvMWWMNMflT
III“ II , .,... "I' .

. 'I'If‘m' I I .sz ’fltI‘I‘ l

- I III... III II. .. .III'III!‘ III... '9.
~ III. I'III'II III. I... .. I.
III IIIIIIIIIIIII (IIIII IIIIIII‘IIII “riff. III III? ’IIIII~I IIIIIIIIIJI IIIII )_
I‘ II IIIIIIII III III

II'IIIII 'III,
I IIIIIII'IIIIIIIIIIIIIIIIIIIIIIIIIIIII II IIIIIIIIIIII II. II I‘” II ”II IIIII
I IIbIII I'III'III'IIIII III III I IIIHI IIIIIIII IIIIII I IIIIII IIIIIIIvaI, ;
I I IIII‘IIIIIII IIIIIIII‘IIII II «I III III III I I III II II
I IIII IIIIIIII IIIIIII I IIIIIIII IIII IIIIIL

I ’II'II I
III III III III. II

III I:

. II III“, III . I .. . III.

I III I I. I III III

w IIII IIIIII‘IIII III II III III ' "
IIII IIIIIIIIII IIIIII. II "I III

ILIII

I I III
II III

I IIIIIII IIIIIIII I "II

II II IIIII IIIIIII III. III III .III IIIIIIIIIIII ’III' III“ II

III II' II II i-1 I II IIII IIII III III IIIII‘I'I

I III IIIIIIIII‘I III II III II I ,II IIIII III IIIIIIII‘IIIIII‘IIIIII IIIIIIIIII IIIIIIIII III ‘.

., I IIIIIIII III IIIII IIII IIIIIIIII IIIIIIIIIIIIIIII I IIIIIIII II I IIIII IIIIIIIIIII II IILIIII I

I i I III IIIIII IIIIII III .I-IIIIII IIIIII IIII'I.III.IIIIIIII .=1‘I.’-',.I'.;.I II": 'I
I II III IIIIIIII IIII IIII II IIIIIIII IIIIIIIIIII? III III III” .III 'III III III“ .‘ I -I "'

I IIIIIIIIIIIII ”I’IIIII IIII I III II III II. 'IIIII’I’IIIII‘IIIIIII IIII‘I“»"“ “ III.’ II

IIIII IIII IIIIII IIIIIII II IIIIIIII II IIIIII IIIIII IIIIIIIII IIIIIIIIII IIIIII II IIIIIIIIIII IIIII IIIII I

II: III " “"' II IIIIIIIIIIIIIIII IIII IIIIIIIIIIIIIII'II' II I

IIII- I' IIIIII'III III” I
I ‘ II I" IIII”IIIIIIII III IIIIII‘
I III III IIIII’II IIII I. III
I, III III III II III III III-III" III“
II IIIIIIIIII III.“ IIIII [III ml III II‘III:
I I IW‘IIIIII IIII IIIIIIIII'I I'III' I
l

IIIII'II II III: III II'IIIIIIIII .I. I I I
III. II'IIIIII. VIII I I I I
IIIII II” I {III .
.I IIIIIIIII" III I.I III III IIIIII IIIII IIkI IIII IIIIIIIII'II- IIIII
I'I'IIIII-IIKI

  II IIIII
I III

M
-—‘é——r—“‘~‘—‘—
-5;
__~:__‘
W
-———v=-

_‘ gag
M __

 

I'IIII .III"

I‘HESIS ‘\

Michigan State ‘
I

 

This is to certify that the

thesis entitled

A STUDY TO DETERMINE THE EFFECTS OF FLOATING
FOREIGN EXCHANGE RATES ON THE RISKINESS OF
MULTINATIONAL ENTERPRISES

presented by

Carl B. McGowan, Jr. P

has been accepted towards fulfillment
of the requirements for

Ph.D. Business - Finance

degree in

 

 

 

 

Mam/fl flaky

Major professor

Datefi/M /£ ”f0

0-7639

p, ‘ ovsnoue FINES:

-- . . ,. 25¢ per day per item

“mam g mumuc LIBRARY MATERIALS:
um-

i- ‘-_ 13,-“)?! Place in book return to remove
'l “2' charge from circulation records:

 

 

 

 

 

A STUDY TO DETERMINE THE EFFECTS OF FLOATING FOREIGN EXCHANGE

RATES ON THE RISKINESS OF MULTINATIONAL ENTERPRISES

BY

Carl B. McGowan, Jr.

A DISSERTATION

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

DOCTOR OF PHILOSOPHY

Department of Accounting and Financial Administration

1980

(:) Copyright by

Carl B. McGowan, Jr.
1980

ABSTRACT

A STUDY TO DETERMINE THE EFFECTS OF FLOATING FOREIGN EXCHANGE

RATES ON THE RISKINESS OF MULTINATIONAL ENTERPRISES

BY

Carl B. McGowan, Jr.

Research, performed prior to the collapse of the International
Monetary Fund in 1973, suggests that there are benefits to multinational
diversification. None of these studies have analyzed the impact that
the new international monetary system has had on this relationship.

This study develops a model for expected returns for a multi-
national firm. These returns can be divided into three parts:

(1) returns on domestic investments, (2) returns on foreign investments,
and (3) returns from changes in foreign exchange rates. Analysis of
covariance is used to test if the beta coefficients for a group of
multinational firms have changed and, if so, in which direction. Multi-
national firms are defined as those firms having real investments in

six or more countries. Firms in the Fortune 500 Directory of Manu-
facturing Firms were divided into three groups: (1) Firms with six

or more real investments overseas, (2) Those that had no overseas
investments, and (3) Those with one to five overseas investments.
Portfolio betas were then compared for the period from January, 1967

to December, 1970 to the period from January, 1973 to December, 1976

Carl B. McGowan, Jr.
to determine if the betas changed. The former period is one under the
fixed exchange rate system, and the latter period is one of floating
exchange rates. The domestic firms were used to determine if changes
in betas might have occurred because of changes in the overall market
risk-return relationship.

The results of this study indicate that multinational firms
have become less risky, as measured by beta,whi1e domestic firms have
become more risky. These results are consistent when other factors
such as industry, proportion of foreign sales, size of firms, and number
of countries in which investments are made, are examined.

Now that the international monetary system has changed and
foreign exchange rates are more volatile, the riskiness of multinational
firms might be expected to be higher causing their betas to rise. The
results of this study do not support this conclusion. Two reasons are
posited for this. First, it may be that multinational firms are less
risky because of other factors that have a greater impact on multi-
national firms than the change in the international monetary system.

If the domestic environment has become more risky, multinational firms
would be less influenced by these factors because their assets are
diversified across countries and not all subject to these adverse
effects. Domestic firms, which do not mitigate against the increased
domestic risk with foreign investment, must absorb the total increase
in risk. This could explain why multinational firms have become less
risky in absolute terms as well in relation to domestic firms.

A second possible explanation of the impact of the change in

the international monetary system is the method by which foreign

Carl B. McGowan, Jr.
exchange rate changes are determined. Under the fixed rate system, the
ultimate decision to change the foreign exchange rate was made in the
political arena. Thus, although all economic factors might indicate
that a foreign exchange rate parity change was necessary, for political
reasons, no parity change would be made until the situation became so
serious that a traumatic change would be required. Estimated foreign
exchange rate changes would have riskiness with respect to economic
predictions and uncertainty with respect to political changes. Under
the floating rate system, the political impact and the uncertainty of
that impact on exchange rates is reduced. Therefore, although foreign

exchange rate changes are now more volatile, they are less uncertain.

ACKNOWLEDGMENTS

I would like to acknowledge all of the assistance and guidance
for this dissertation from all of the faculty and doctoral students at
Michigan State. Dr. Jeffrey Towle was particularly helpful in the
early development of my dissertation. Dr. Richard Simonds was very
helpful in developing the model and research design. Both were helpful
in the writing and the presentation of the dissertation. I would, also,
like to acknowledge the help of Dr. Alden Olson, my dissertation chair-
man, who provided me with guidance and direction throughout the develop-
ment, execution and completion of my dissertation. I would like to
thank Mrs. Jo McKenzie who eliminated many problems for me by typing
my dissertation although I will retain credit for any remaining errors

or omissions.

iii

TABLE OF CONTENTS
Page
ACKNOWLEDGMENTS. . . . . . . . . . . . . . . . . . . . . . . . . . ii

Chapter
I 0 INTRODUCTION 0 O O C O O O I O O O O O O O O O O O O O O 1
The International Monetary System. . . . . . . . . . . 1
Fixed Exchange Rates - Advantages and
Disadvantages . . . . . . . . . . . . . . . . . . . . 5
Floating Exchange Rates - Advantages and
Disadvantages O O I O O O O O O O O I O O O I O O O O 6
Economic Impact of Foreign Exchange
Rate Changes. . . . . . . . . . . . . . . . . . . . . 8
The PrOblemo O O O O O O O O O O O O O O O O O O O O O 11
Research Design. . . . . . . . . . . . . . . . . . . . l3
Contribution to Finance. . . . . . . . . . . . . . . . 13
II. LITERATURE REVIEW. . . . . . . . . . . . . . . . . . . . 14
Initial Application. . . . . . . . . . . . . . . . . . 14
Less Developed Countries . . . . . . . . . . . . . . . l6
Holdings Period Effects . . . . . . . . . . . . . . . 17
Foreign Sales. . . . . . . . . . . . . . . . . . . . . 18
Size Effects 0 O O C O O O O O O O O O O O I O O O O O 21
Industry Effects 0 O O O I O O O I O O O O O O O O O O 23
The International Asset Pricing Model. . . . . . . . . 24
sumry O O O O O O O O O O O C O C I O O O O O O O O O 25
III. A RETURNS MODEL FOR MULTINATIONAL FIRMS. . . . . . . . . 27
The ShaIPE‘Lintner MOdEl o o o e e o o e o o o o e o o 27
Markowitz and Multinational Diversification. . . . . . 30
The Cash Flow Effects of Floating Exchange Rates . . . 33
sumary. I O O O O O O O O O O O O O O O O O O O O I I 36
IV. RESEARCH DESIGN. . . . . . . . . . . . . . . . . . . . . 38

Multinational Enterprise Defined . . . . . . . . . . . 39
Population . . . . . . . . . . . . . . . . . . . . . . 40
Data Sources . . . . . . . . . . . . . . . . . . . . . 41
Time Period. . . . . . . . . . . . . . . . . . . . . . 41
Analysis of Covariance . . . . . . . . . . . . . . . . 42
Summary. . . . . . . . . . . . . . . . . . . . . . . . 48

iv

Chapter Page

V. RESULTS. . . . . . . . . . . . . . . . . . . . . . . . 50
Introduction . . . . . . . . . . . . . . . . . . . . 50

Calculation of Statistics. . . . . . . . . . . . . . 50

Aggregate Test Results . . . . . . . . . . . . . . . 52

Firm Size Effects. . . . . . . . . . . . . . . . . . 53

Foreign Sales Level Effects. . . . . . . . . . . . . 54

Number of Countries Effects. . . . . . . . . . . . . 56

Industry Effects . . . . . . . . . . . . . . . . . . 57
Conclusions. . . . . . . . . . . . . . . . . . . . . 59

References . . . . . . . . . . . . . . . . . . . . . . . . . . . 61

CHAPTER I

Introduction

 

This chapter will describe the international monetary system in
the post-World War II period, how the system came about, and how it
collapsed. Both fixed and floating foreign exchange rate systems and
their advantages and disadvantages will be described. The impact of
foreign exchange rates on multinational firms will be detailed.
Finally, the problem that arises from this new international monetary
system and how to analyze the impact of this problem upon multinational

firms will be discussed.

The International Monetary System

 

The contemporary international monetary system has its origins
in the post-World War I period.1 At that time, each country tried to
revive its economy by increasing exports and limiting imports. Imports
were to be restricted by raising high tariff walls. Exports were to be
encouraged by devaluing the local currency which would make imports
relatively more expensive and exports relatively cheaper. However,
this outcome assumes that the trading partners of the country do not

react in the same manner. In fact, most countries created very high

 

1Thomas Willett, FloatinggForeign Exchange Rates and Inter-
national Monetary Reform, Washington: American Enterprise Institute,
1977, p. 2.

 

2
tariff barriers and devalued their currencies. The high tariff bar—
riers reduced world trade and the continual devaluations helped to
create higher inflation. These two actions helped to cause an almost
total disruption of trade during the interwar period.

Largely to preclude a similar situation after the Second World
War, representatives of the major trading nations met in August, 1944
at Bretton Woods, New Hampshire.2 The objectives of this meeting were:
(1) to promote European economic recovery, (2) to promote international
trade, and (3) to develop a viable international monetary system.

These objectives were to be achieved through three organizations - the
International Bank for Reconstruction and Development, the International
Trade Organization, and the International Monetary Fund - IMF.

The IMF had four goals. First, all major currencies should
again be convertible into gold. Second, exchange rate stability should
be attained. The third objective was to provide a mechanism by which
orderly foreign exchange rate changes could be made. The final objec-
tive was to create a system of international monetary reserves to be
used by countries having temporary foreign exchange outflows.

After all countries had returned to full convertibility, the
permanent mechanisms of the IMF were to come into full use. Foreign
exchange stability was to be enforced by intervention in the currency
market by the appropriate government at specified points. Originally,
when a currency moved more than plus or minus 1.0% from its stated par

value, the government was required to support the market, i.e., buy or

 

2Richard Hays, Christopher M. Korth, and Manucher Roudani,
International Business: An Introduction to the Multinational Firm,
Englewood Cliffs, NJ: Prentice-Hall, 1972, p. 149.

3
sell its currency to maintain or return to par. An intervention range
was used rather than a specific value. This recognizes the fact that

\\many factors influence foreign exchange rates, particularly in the short

 

m—Mg'gu—i-W ' ""'"""") H
- “in-.-

run, and that foreign exchange rates will normally fluctuate somewhat.

The IMF provides a stabilization fund for countries experiencing
temporary or seasonal pressures on their currencies. This fund is com-
posed of the international monetary reserves and potential reserves
maintained by members at the IMF. The primary monetary reserves are L/’///
gold, hard currency, and special drawing rights. The gold is part of
the original membership contribution each country makes in joining the
IMF. Hard currency is accumulated by countries through international
trade. Special drawing rights are fiat money created by the IMF to
increase international liquidity. Countries may draw on their reserves
in order to meet foreign exchange shortfalls or build reserves for
future needs.

Conditional reserves are composed of currency swaps and loans.
In a swap, Country A has too much currency B and Country B has a
shortfall of currency A. Country A trades its excess currency B until
a specified future date by which time the disequilibrium should be cor-
rected. The swap is a short term effort to allow a temporary disequil-
ibrium to be bridged without a permanent exchange rate adjustment.
Loans are temporary between the IMF, or a second country, and the
country with the currency deficit. Both swaps and loans may place
requirements, designed to alleviate the cause of the disequilibrium, on
the country receiving the loans.

The final source of reserves is special loans from the IMF or

lender countries. As a country needs additional loans, restrictive

4
covenants become more onerous. Thus, a country must voluntarily cor-
rect its disequilibrium or face externally devised solutions.

An alternative is a parity change. In theory, a country, must
first request permission to change its exchange rate if a permanent
disequilibrium exists. However, because market pressures react very
quickly, this request is usually a de facto devaluation. IMF approval
is a formality.

In the immediate post-war period, international trade was
hampered by the destruction of the European economies and by the fact
that the United States held virtually all of the available foreign ex-
change reserves of the world.3 Because of sales of our war material
to the allies for gold and on credit, most gold and hard currency was
in or owed to the United States. United States trade deficits were
needed to help rebuild the European and Japanese economies. The dollar
was overvalued in order to encourage the United States to import foreign

ods to stimulate foreign economies and to reduce exports by making U.S.
goods more expensive and less competitive.

After these objectives were met, the deficit became a burden
rather than a boon. Capital outflows occurred for several reasons:
U.S. multinational enterprise direct investment, foreign aid grants,
forgiveness of war debts, and military expenditures. In addition, the
relative inflation rate of the United States was higher than the rates
of her primary trading partners. The overvalued dollar caused a

weakening of the U.S. trade balance and gave European and Japanese

 

3Fred H. Klopstock, "The International Status of the Dollar,"
Essays in International Finance, Princeton, NJ: Princeton University
Press, 1957, p. l.

traders a competitive advantage.

The U.S. deficit could not go on indefinitely, and more and
more pressure was exerted on the U.S. to control the dollar. Yet, as
a key currency, the dollar could not be devalued, except by changing
its price in terms of gold. Since U.S. trade partners would not re-
value their currencies, raise their value vis-a-vis the dollar to

reflect their added strength, the dollar was revalued and made incon-
_____.

 

 

vertible by Nixon in August, 1971. Still, a good many disequilibria

M..-

 

.a-fl

-v...’--

occurred in the international economy and all currencies were eventu-

ally forced to float.4

Fixed Exchange Rates - Advantages and Disadvantages

 

The primary advantage of fixed rates is that rates are certain,

at least in the short term. Within the specified bands determined

through the IMF, rates are fixed. Reducing the likelihood and size of

v—mu-Ha in.“

 

 

 

 

-.

a foreign exchange loss facilitates and promotes international trade.5

“infinb )-
..— nil-4"
' "'hiVeu— _’ -m'M

,/’ ~

There are several disadvantages to a fixed rate system. The

f. i
If I,
I I

”’structure needed to maintain the fixed rate system is expensive. All
countries with convertible currencies must maintain large foreign ex-
change reserves to support temporary disequilibria, and a structure

such as the IMF must be maintained to coordinate and control the fixed

system.

 

4"International Payments Imbalances and the Need for
Strengthening Financial Arrangements," A Report to the Joint Economic
Committee, 87th Congress, lst Session, U.S. Government Printing Office,
Washington, 1961.

5Rita M. Rodriguez and Eugene E. Carter, International
Financial Management, Englewood Cliffs, NJ: Prentice-Hall, 1976,
p. 72.

 

6

A politically important disadvantage of fixed rates is that
to correct an incorrectly valued currency requires a dislocation of the
local economy. In terms of the quantity theory of money, the money
supply multiplied by the velocity of money is equal to the average
price in the economy times the number of transactions completed. Both
of these equal the total output of the economy. In order to cover a
trade deficit, a country must export goods or foreign exchange which
reduces the money supply. Assuming that the velocity of money and the
number of transactions are fixed, a decrease in the money supply causes
a decrease in the price level. Deflation causes a relative increase in
international competitiveness for the deflating country's goods which
increases exports and decreases imports. The net effect is to return
the balance of payments to an equilibrium level. fiIn the interim, the
local economy deflates and this increases unemployment. This latter
effect has political ramifications. Government is loathe to create
unemployment - votes for the opposition. Thus, governments are tempted
to support an artificially high foreign exchange rate in order to
buffer the local economy. However, in the long run, the foreign ex-

change rate must move to its real equilibrium value.

Floating Exchange Rates - Advantages and Disadvantages

 

 

Floating rates have two advantages. The rates are established
in a market place rather than in a political arena. Second, adjustments

for disequilibria occur in the foreign exchange rate and not in the

“Inn.

 

local economy. In a floating rate system, rates are determined by
currency dealers reacting to supply and demand conditions rather than

politicians reacting to political needs. Dealers in New York, London,

7
Zurich, Bonn, etc. adjust prices according to excesses of supply or
demand for a currency. If more people wish to sell a currency than
wish to buy it, the price will fall until supply equals demand. This
will have the same effect as deflating the local economy without
affecting the employment rate. The result will be that the price of
goods in the devaluing currency will cost less because the price of
local currency falls. It takes fewer units of nonlocal currency to
buy goods. Thus, exports are raised. Because it costs more of the
local currency to purchase foreign currency, the price of imports
rises and total imports fall.

The disadvantages of floating rates are uncertainty about

W‘flwem .. .,..

future foreign exchange rates, increased volatility of foreign exchange
rates, and the increased cost of hedging. Research indicates that
floating foreign exchange rates are more volatile than fixed rates

and that foreign exchange rates cannot berredicted in the short run

-WWW‘ u

u-—

because changes in the rates are a random walk.6 The effect of this

*—

 

‘v—r fit?
r—u

increased volatility on multinational firms is aggravated by the new
reporting requirement of FASB-8. Previously, firms could choose the
method of translation of subsidiary financial statements and offset

losses or gains with a contra account which would reduce the effects

of short term currency movements. EASE-8 requires that firms use the‘ffiy/
temporal method of translation and report gains and losses as theyg /: 3

occur, quarterly.

 

6Ian H. Giddy and Gunter Dufey, "The Random Behavior of Flex-
ible Exchange Rates: Implications for Forecasting," Journal of Inter-
national Business Studies 6 (Spring 1975):l-3l; William R. Folks, Jr.
and Stanley R. Stansell, "The Use of Discriminant Analysis in Fore-
casting Exchange Rate Movements," Journal of International Business
Studies 6 (Spring l975):33-50.

 

 

 

8
Economic Impact of Foreign Exchange Rate Changes

There are two types of gains and losses - transactions and

~—- '

translations. Transactions effects occur because currencies are ex-

“-1...“

 

changed at different rates. First, assume that a firm sends $1000 to
a subsidiary and exchanges the follars for 1000 local currency (LC).

The exchange rate is $1:1LC, initially, but changes to $1:2LC. When

the 1000 LC are repatriated, they will only purchase $500. This is a
transaction loss of $500.-

A translation loss occurs because succeeding financial state-
ments of a subsidiary are translated at different rates. The temporal
method of translation requires that accounts be translated at the
foreign exchange rate in effect when the event occurred and gave rise
to the account. Exhibit 1 shows the effect of a foreign exchange de-
valuation of local currency on the reported financial statements of a
subsidiary. If the local currency devalues from $1:1LC to $1:2LC, the
value of the firm's assets in dollars will decrease by $100. This
decrease will be offset by a decrease in liabilities of of $50 and a
decrease in owner's equity of $50. The income statement effect, when
translated at the new rates, will be a $50 loss rather than a $200
gain. If there are no changes in the foreign exchange rate, the parent
company reported income would rise by $200 from foreign operations.
With foreign exchange effects, parent reported income will be decreased
by $100. This loss would have to be reported immediately, even if the

W-__,_,‘ .‘h

\ decrease is temporary. If the foreign exchange rate returned to the
previous level and then rose 50%, there would be a $500 gain. This
gain and the previous loss would be offsetting and total reported in-

come from multinational operations for the two periods would be $400.

9
EXHIBIT 1

Multinational Operations

Balance Sheet

December 31, 19XX

 

Assets (1)
Cash 100
Inventory 100
Fixed Assets (net) 70

Total Assets 270

Liabilities and Owners' Equity
Current Liabilities 100
Owners' Equity 170

Total Liabilities 270

(1) value in local currency
(2) value in dollars if the exchange rate is $l:1 LC
(3) value in dollars if the exchange rate is $1:2 LC
(4) foreign exchange losses

Sales

Less:

LESS:

Multinational Operations

Income Statement

(2)
100
100

70
270

100
170
270

7

Year Ending December 31, l9XX

Q

4000

Cost of Goods Sold 600
Depreciation Expense 200
Other Operating Expenses 2800
EBIT 400
TAXES (50%) 200
Net Income 200

(1) value in local currency
(2) value in dollars if the exchange rate is $l:l LC
(3) translation rate if the exchange rate is $1.2 LC
(4) value at the new translation rate

 

e).
4000
600
200

2800

400

200

200

 

AA
U1UI
VV

(100)

( 50)

S 50)
(100)

(4)
3000
600
200
2100
100
150
( 50)

7David Norr, "Currency Translation and the Analyst," Financial'
Analysts Journal 32 (July-August l976):46-54.
Statement 8 Resolved Foreign Currency Accounting - Or Did It,"
Financial Analysts Journal 32 (July-August l976):55-61.

John K. Shank, "FASB

10
However, because the effects must be reported quarterly, reported
earnings are more volatile than they would have been if only permanent
changes were reported.

A multinational firm can, to some, nt, offset foreign ex-
change effects by_£sducing exposure and by hedging, i.e., purchasing’“ ”,5-
forward foreign exchange contracts. These may create additional real [3‘
costs if the firm uses either method to offset translation losses just
to reduce variability in reported earnings. Most firms do attempt to
minimize variability in reported earnings from foreign exchange effects,
i.e., have zero foreign exchange gains or losses.

If a firm anticipates that 200,000 LC dividend will be paid in
ninety days and if there are no foreign exchange parity changes, the
200,000 LC will purchase $200,000 at a foreign exchange rate of $1:1LC.
The firm, to protect its dividend, purchases a forward contract at a
rate of $1:l.1lLC. If the actual exchange rate is $1:2LC, the firm
would lose $100,000 without the contract - 200,000 LC times .5 $/LC.
With the contract, the firm loses $20,000 - 200,000 LC times .9 $/LC.

To protect an anticipated transaction, the $20,000 might be well spent.

However, if the firm purchased the same contract to offset a translation

'VMp-n: '

-nw-se-u-I “NW W "qr-4. u

loss, the firm would be losing $20,000 to offset a paper loss only.
This would be disadvantageous to the firm in cash flow terms.
If a firm increases local currency debt or local minority

interests, foreign exposure to foreign exchange effects is reduced

because decreases in asset values are offset by decreases in liability

 

8Michael Jillings and William R. Folks, Jr., "The Emergence

of the Foreign Exchange Risk Management Function," Working Paper 2,
Foreign Exchange Risk Management Project, Center for International
Business Studies, University of South Carolina.

11
values. However, local funds may be more expensive than U.S. funds,
even after foreign exchange effects. If the firm in Exhibit 1 had cur-
rent liabilities of 200 LC and 70 LC in equity, the decrease in current
asset value of $100 would be offset by a decrease in local currency
liabilities of $100. No foreign exchange loss would be reported. Thus,

the firm would be paying a higher interest charge to avoid a paper loss.

”W

M...‘ .
“Mm—nu. a. A

‘n‘h-A 44.. _, .

Raising local equity offsets foreign ékéhangé effects but, also, reduces
the firm's control and flexibility - another real cost to offset a paper
loss. Either of these actions would reduce the cash flow or control of

the subsidiary with no increase in real benefits.

The Problem

 

The question of whether or not multinational diversification is

9
useful in reducing risk has been dealt with extensively. This research

 

9Agmon, Tamir. "Country Risk - The Significance of the Country
Factor to Share Price Movements in the United Kingdom, Germany, and
Japan," Journal of Business (January, 1973), pp. 24-32.

Agmon, Tamir. "The Relations Among Equity Markets in the United
States, United Kingdom, Germany, and Japan," Journal of Finance
(September, 1972), pp. 839-856.

Grubel, Herbert G. "Internationally Diversified Portfolios:
Welfare Gains and Capital Flows," American Economic Review (December,
1968), PP. 1299-1314.

Grubel, Herbert G. and Kenneth Fadner. "The Interdependence of
International Equity Markets," Journal of Finance (March, 1971), pp. 89-
94.

 

 

Lang, Larry Ross. "The Effects of a Corporation's Multinational
Diversification on its Common Stock Holders' Return and on their Risk
Exposure," Unpublished Doctoral Dissertation, Michigan State University,
1973.

Lessard, Donald R. "WOrld National, and Industry Factors in
Equity Returns," Journal of Finance (May, 1974), pp. 379-91.

Lessard, Donald R. "International Portfolio Diversification:
A Multivariate Analysis for a Group of Latin American Countires,"
Journal of Finance (June, 1973), pp. 619-33.

12

indicates that, for a given rate of return, multinational firms have a
lower level of risk. However, there has been a structural change in
the international monetary system that has reduced the usefulness of
previous studies. The international monetary system has changed from
a fixed system under the Bretton Woods agreement to a floating system.
Since the floating system has created greater volatility of exchange
rates, this should impact on the riskiness of multinational firms.
However, most studies have been done with data only up to 1971 or

have used later data but ignored the change in the international mone-
tary system. This study is an attempt to determine if the change in
the international monetary system has impacted on multinational firm

riskiness as measured by beta.

— ",I n‘ —
_ .,-..-.n‘-a—r~.—--n-' "" -'-"

 

Levy, Haim and Marshall Sarnat. "International Diversification
of Investment Portfolios," The American Economic Review (September,
1970). pp. 668-75.

Rugman, Alan M. "Risk Reduction by International Diversifi-
cation," Journal of International Business Studies (Fall/Winter, 1976),
pp. 75-80.

Rugman, Alan M. "International Diversification by Financial
and Direct Investment," Journal of Economics and Business (Fall, 1977),
pp. 31-370

Severn, Alan. "Investor Evaluation of Foreign and Domestic
Risk," Journal of Finance (May, 1974), pp. 545-50.

Solnick, Bruno H. "International Pricing of Risk," Journal of
Finance (May, 1974), pp. 365-78.

Solnick, Bruno H. "Why not Diversify Internationally,"
Financial Analysts Journal (July-August, 1974), pp. 48-54.

Solnick, Bruno H. "An International Market Model of Security
Price Behavior," Journal of Quantitative and Financial Analysis
(September, 1974), pp. 537-54.

13

Research Design

 

This study will try to isolate the impact that floating foreign
exchange rates has had on multinational firms. Two groups of firms
were analyzed - a group of forty-nine firms that have zero multi-
national investments and a group of 193 firms with investment in six
or more foreign countries. Market risk measures were calculated for
each group for two time periods and compared. The first time period,
January 1967 to December 1970, was a period of fixed foreign exchange
rates under the Bretton Woods agreement. The second time period,
January 1973 to December 1976, was a period of floating foreign ex-

change rates. An analysis of covariance comparing the betas of thi::>

multinational and domestic firms before and after the collapse of

Bretton Woods was done to see if the betas changed. f”,x/)

-rr’
,r
.

Contribution to Finance

 

The objective of this study is to determine if the change in
the foreign exchange system from fixed to floating rates has had an
impact on the risk-return relationship for multinational firms. Its
contribution will be in assessing the value of multinational diversi-

fication since the change in the international monetary system."

CHAPTER II

Literature Review

 

To test whether multinational firms are now more risky because
the international monetary system has changed from a fixed to a floating
rate system, it is necessary to explain how common stock returns, and
the riskiness of those returns, are affected by multinational diversi-
fication. This chapter discusses studies that examined these factors.
Several possible alternative explanations for superior performance by
multinational finms will be examined, also. Finally, the international
asset pricing model will be examined. Proponents of the international
asset pricing model contend that investors are not benefited by multi-
national diversification by domestic firms. If investors can achieve
international diversification through portfolio investment, inter-
national diversificiation by domestic firms will not be rewarded in

the stock market. If the world's stock markets are segmented, then,

--‘,.__ 1,. _ _

multinational diversification will be beneficial.

Initial Application

To test whether multinational firms are now more risky, it is
necessary to develop a model to explain how common stock returns are
determined and how foreign exchange fluctuations ought to affect the
riskiness of those returns. The initial application of portfolio

theory to multinational investment sought to estimate the potential

14

15

gains from such international diversification.1 Correlations between
market indices of eleven different major countries were compared using
monthly rate of return observations for the period from January 1959

to December 1966. The monthly rate of return, r, was calculated as

follows:
P1 X1 12
ri = YO/IZ +F- x—' - 1.0
o o
where,
ri = the monthly rate of return
Y0 = the annual percentage dividend yield on
shares in the index
P = the value at the beginning of the period
0
of the common share price index
P1 = the value of the end of the period
of the common share price index
X0 = the dollar exchange rate at the beginning
of the period
X1 = the dollar exchange rate at the end

of the period
After the monthly returns were calculated for the period for
each index, a correlation matrix among all countries monthly returns
was calculated in order to derive variances and covariances for subse-
quent analysis. In addition, average rates of return, R, were calcu-
lated for the period using a geometric mean of the 95 monthly rates

of return.

 

1Herbert Grubel, "Internationally Diversified Portfolios:
Welfare Gains and Capital Flows," American Economic Review, 58
(December 1968):1299-1314.

16

Using the rates of return, variances, and covariances, portfolios of
Varying size and composition were constructed to create efficient port-
folios, i.e., portfolios with minimum risk for a given rate of return
or maximum rate of return for a given risk level. When the universe

of countries is expanded, the efficient frontier shifts upward, i.e.,
the rate of return for a given risk level rises. This implies that,

by expanding the investment universe, greater returns can be achieved

at the same level of risk.

Less Developed Countries

 

Using a somewhat different measure of return, the impact of
investing in less developed countries was tested for the period 1951
to 1967.2 Rates of return, ri, were calculated for the major stock

indices of twenty-eight countries.

ri — (Pl - Po)/Po
where,
r1 = the annual rate of return for each
country's index of common stock
P0 = the value of the index of common stock
at beginning of the period
P1 = the value of the index of common stock

at the end of the period
From these rates of return, arithmetic mean rates of return and a cor-
relation matrix among the annual rates of return were calculated for

use in further analysis.

 

2Haim Levy and Marshall Sarnat, "International Diversification
of Investment Portfolios," American Economic Review, 60 (September
l970):668-675.

l7

1 n
R1 = '—§_ 2 ri(t)
t=l
where,
Ri = the mean rate of return for each country
N = the number of countries in the study (28)

These data were used to calculate a locus of portfolios which minimize
the variance of the portfolio for a given expected rate of return. As
the number of countries was expanded, the efficient frontier shifted
upward, i.e., as the universe of potential investments increases, the

rate of return for a given risk level increases, also.

Holding Period Effects

 

An ex post variance and covariance matrix of returns for 51
Standard and Poor Weekly Index industry subindices, 28 Financial Times
Share Indices, and 28 West German Stock Market Indices were compared
to test the effects on the correlation between industry and aggregate
indices of increasing the holding period used.3 The time period of
the study was from January 1, 1965 to June 30, 1967. This time period
was roughly representative for all three markets. The weekly rate of

return, r, was calculated by solving the following equation for r.

PE = PE ert
1 o

 

3Herbert Grubel and Kenneth Fadner, "The Interdependence of
International Equity Markets," Journal of Finance, 26 (March 1971):
89-94 0

 

18
where,
PE = Po/Xo = the value of the index for the

beginning of week divided by the exchange
rate between dollars and the respective

currencies
PE1 = the value of the index at the end of the
week divided by the exchange rate at the
end of the week
t = the number of days in the observation
period (7) divided by 365
r = the annual rate of return with continuous

compounding
This time series of rates of return for the 108 assets was used to
calculate the variance/covariance matrix.

The longer length holding periods, one and three months, were
calculated as moving averages of different length rather than succes-
sive asset holding periods of different length. The results indicate
that as the holding period length increases, the correlation between

indices increases, also.

ForeiggfiSales

 

If multinational diversification is beneficial to investors in
a firm, earnings variability should be a decreasing function of the
proportion of total sales that are foreign. Using this as a hypothesis,
Rugman tested the effect that the proportion of foreign sales has on

the variability of firms' earnings.7

 

7Alan M. Rugman, "International Diversification by Financial
and Direct Investment," Journal of Economic and Business, 29 (Fall
l977):31-37.

 

l9

Var(E/K) = f([F/T], SIZE, IND)
Var(E/K) = the total variability of earnings for
a firm
F/T = the proportion of total sales that

are foreign

SIZE = a size variable for a firm (employees,
sales, or assets)
IND = a dummy variable for the industry of

the firm
Earnings variability was calculated for 492 firms for the period 1960
to 1969. In addition, regressions were run using industry and size as
independent variables to control for these factors.

The study concluded that variability is a decreasing function
of proportion of sales but with an R2 = .10. When size and industry
were included in the regression, the R2 rose to .12 and .13 respec-
tively. Although the results of this study are statistically signifi-
cant, they appear to be limited.

A second study tested size effects using a world market index:
The Capital International World Index, which is a market value weighted
international index of major securities on the 18 largest stock markets
and a United States index - the New York Stock Exchange Index.8
Returns were regressed for firms divided into deciles according to
their proportion of foreign sales for both of these indices at the same
time to see if the hypothesis would be confirmed that the higher the

proportion of non-U.S. activities, the lower the dependence on the U.S.

 

8Tamir Agmon and Donald R. Lessard, "Investor Recognition of
Corporate International Diversification," Journal of Finance, 32
(September l977):lO49-1055.

20

country factor. This study examined monthly returns for the period
from January 1959 to October 1972. Returns on firms in the U.S. were
calculated by dividing dividends during the period plus the ending value

of the securities by the beginning value of the securities.

4.
R = Li
P
o
where,
Ri = the rate of return for the security during
the period
D1 = dividends during the period
P0 = the value of the security at the beginning
of the period
P1 = the value of the security at the end

of the period
Firms were grouped in deciles according the proportion of non—U.S.
sales and the composite return series of all the firms was regressed

against the U.S. index and the world index.

R = a -+ 8 R + W’R + E
us w
where,
R = the return on a share of the corporation
Rus = the return on the NYSE index
w

R = the return on the rest of the world index

8 = the relationship between the return on the
share and the NYSE index

'Y = the relationship between the return on the
share and the rest of the world index

a = the intercept term

E = the residual term

21

This study concluded that as the proportion of foreign sales
increases, a increases also which implies a closer relationship to the
world index. When 8 and a were regressed against the proportion of
foreign sales, IS, to determine if they are related to IS, the results
were as expected t) was inversely related to IS and b' was directly

related to IS.

Bi = a + bIS +11
'1 = a' + b'IS + u'

where,
IS = the international sales ratio
b = the relationship between Bi and IS
b' = the relationship between Yi and IS
0 = the residual terms

Size Effects

 

The market model was used to test the effect that size has on
multinational firm riskiness for two groups of firms.9 The first group
of firms was composed of seventeen oil firms which are multinational
and a second group of eighteen firms of comparable size to the multi-
national firms which had a low percentage of foreign sales. Monthly
and annual rates of return were calculated for the period from 1951 to
1975.

(P - PC + D1)

P
o

l

 

Ri =

 

9Ibid., Rugman, Journal of Business and Economics.

22

where,

Ri = the monthly (annual) rate of return
for a firm

P0 = the value of the firm's share at the
beginning of the period

P1 = the value of the firm's shares at the
end of the period

D1 = the value of any dividends during the
period

The market model was used to calculate beta coefficients for

both portfolios of firms.

R. = a + BiRm + e.

1 1

where,
Bi = the relationship between Ri and the world
market factor

Rm = the return on the world market factor

a = the intercept term

ei = the residual term

The world market index, Rm, was constructed from the relative GNP
weighted annual share price indices taken from the IMF International
Financial Statistics. The results of the test indicate that, on aver-
age, beta coefficients for the multinational firms were less than beta
coefficients for domestic firms indicating that multinational firms

are less risky when size of firms is comparable.

23

Industry Effects

Because industry should affect the required rate of return for
a firm, this effect has been examined.4 Rugman tested the industry
effect of multinational diversification and concluded that the effect
was consistent across industries, i.e., in all industries, an increase
in the proportion of foreign sales for a firm caused a decrease in
riskiness, as measured by the variance of total earnings for the firm.
However, the explanatory power of this study was very low (average R2
= .10 without accounting for industry effects and .12 with industry
effects).5

Grubel and Fadner investigated industry effects, also.6 The
study concluded that as the proportion of foreign activity, [(exports
+ imports)/total sales], increases the correlation between returns of
industry stock market indices increases, i.e., firms in industries
with a large proportion of exports or imports to total sales tend to
be more highly correlated than industries with a lower proportion.
This results because "the greater the industry's involvement in

foreign trade, the more sensitive is its profitability to conditions

 

4Benjamin F. King, "Market and Industry Factors in Stock Price
Behavior," Journal of Business, 42 (January 1966):139-170; John D.
Martin and Robert C. Klemkosky, "The Effect of Market Risk on Portfolio
Diversification," Journal of Finance, 30 (March l975):l47-154.

5Alan M. Rugman, "Risk Reduction by International Diversifi-
cation," Journal of International Business Studies, 7 (Fall/Winter
l976):75-30; , "International Diversification by Financial
and Direct Investment," Journal of Economics and Business, 29 (Fall
l977):31-317.

6

 

 

 

Ibid., Grubel and Fadner, pp. 89-94.

24
throughout the world,"7 and, hence, the greater is its correlation with

other similar industries in the world.

The International Asset Pricing Model

Multinational diversification is beneficial if there are bar-
riers to investment, e.g., restrictions on currency flows, lack of ade-
quate information, or markets that are thin with respect to volume of
trading or number of traders. If barriers do exist, the international
market structure would be segmented and diversification would be bene-
ficial. Otherwise, markets would be homogeneous, and, therefore,
multinational diversification would not be beneficial.

Agmon tested the segmented versus homogeneous markets hypothe-
sis.10 The study compared market indices and a sample of individual
firms in the United States, the United Kingdom, West Germany, and Japan
from 1961 to 1966. It was hypothesized that if homogeneous world mar-
kets exist, the risk return relationship in all four countries could be
similar. If the markets are segmented, the risk/return relationship
would be different for each country.

Because of the small sample size for the non-U.S. group, the
results of the regression were not statistically significant. However,
when comparable samples of U.S. firms of equal size were compared these
results were not significant either. The author concluded that these

results neither reject nor support the segmented market hypothesis.

 

7Ibid., Grubel and Fadner, p. 92.

8Tamir Agmon, "The Relations Among Equity Markets in the United
States, United Kingdom, and Japan." Journal of Finance, 27 (September
l972):839-856.

 

25

When indices themselves were compared, though, the results indicated
that the relationship between the U.S. market and the other three mar-

kets was extremely low supporting the segmented market hypothesis.

Summagy

The initial application of portfolio theory to international
investment was developed in a two country, two asset model. This model
was tested by comparing the efficient frontier developed when invest-
ment was expanded to more than one country, in this case, eleven indus-
trial countries. The results indicated that a higher return could be
achieved, at the same risk level, if investment is made in multiple
countries. When the investment universe was expanded to include less
developed countries, the results were consistent with the earlier
results, i.e., as the investment expands, return increases, for a given
level of risk.

Further studies have been undertaken to determine if these
results are influenced by other factors - holding period, size, propor-
tion of international sales, or industry. As the holding period is
increased, the correlation between returns in various countries in-
creases. Size does not affect these conclusions. When a group of
multinational firms was compared to a group of domestic firms of com-
parable size and risk, the multinational firms had a higher return.
Likewise, these results are consistent across industries.

Not all research attempts to show the benefits of international
investment. An alternative approach is based on the international asset
pricing model. This approach attempts to show that multinational

diversification is not beneficial because there is an integrated world

26

securities market. This model posits that investors can achieve
international diversification through portfolio investment. If there
are no barriers to investment, multinational diversification is unnec-
essary and, therefore, not rewarded in the market. However, the re-
sults of this work indicate that the world asset market is segmented.
This result implies that multinational diversification could have

benefits to the investor.

CHAPTER III

A RETURNS MODEL FOR MULTINATIONAL FIRMS

The objective of this chapter is to develop a model of the
returns for multinational firms and to separate those returns into
three parts: (1) returns resulting from domestic investment, (2)
returns resulting from foreign investment, and (3) returns resulting
from foreign exchange changes. The Sharpe-Lintner model is used as
a general basis for explaining security returns. Next, a model of
returns is developed that shows the effect that multinational diver-
sification has upon expected returns. Finally, the foreign returns
for a multinational firm are divided into returns for investment and

returns resulting from foreign exchange rate fluctuations.

The Sharpe-Lintner Model

 

The mean variance model is developed from the work of Markowitz-
Sharpe-Lintner.l The model follows from the following assumptions sum-

marized in Jensen.2

 

1Harry M. Markowitz, "Portfolio Selection," Journal of Finance
(March 1952):77-91; William F. Sharpe, "Capital Asset Prices: A Theory
of Market Equilibrium Under Conditions of Risk." Journal of Finance,
19 (September 1964):425-442; John Lintner, "The Valuation of Risk
Assets and the Selection of Risky Investments in Stock Portfolios and
Capital Budgets," The Review of Economics and Statistics, 49 (February
1965):13-37.

 

2Michael C. Jensen, "Capital Markets: Theory and Evidence,"
The Bell Journal of Economics, 3 (Autumn l972):358-359.

 

27

28

All investors are single-period, expected utility
of terminal wealth maximizers who choose among
alternative portfolios on the basis of mean and
variance (or standard deviation) of return.

All investors can borrow or lend an unlimited
amount at an exogenously given risk-free rate of
interest, Rf, and there are no restrictions on
short sales of any asset.

All investors have identical subjective estimates
of the means, variances, and covariances of return
among all assets.

All assets are perfectly divisible and perfectly
liquid, i.e. all assets are marketable and there
are no transactions costs.

There are no taxes.

All investors are price takers.

The quantities of all assets are given.

With these assumptions, one can show that the expected return

for a given investment is determined by the riskless rate of return on

a market index, and the degree to which the return for the investment

varies with the index. The expected rate of return for investment, i,

 

would be:
E(R ) = R + [E(R ) - R ] °°V (R1: RI)
i f I f 2
0 (RI)
= Rf + [E(RI) - Rf] Bi
Alternatively,
E(Ri) = Rf + ri (piIoio.) = Rf + ri (Bi)
2
CII
where,
E(Ri) = the expected rate of return for invest-

ment i.

29

Rf = the rate of return for the riskless
asset.

E(RI) = the expected rate of return for index I.

cov(Ri,RI) = the covariance between the index and
investment i.

02(RI) = the variance of index I.

r1 = the risk premium for investment i.

Oil = the correlation coefficient of invest-
ment 1 and index I.

Ci = the standard deviation of the expected

return for investment i.

The beta of the investment, the degree to which the investment
responds to changes in the index, is of critical importance in deter-
mining the expected rate of return for an investment. If all other
things are held constant, a change in beta will cause a change in

expected return which will cause a change in the value of the invest-

 

ment.
0' C
B = °°V(R1’RI) = 011 i I
i 2 2
0' 0

Beta will change if the variance (standard deviation of investment i
changes and all other factors remain constant. The degree and direction
of this change is determined by the correlation coefficient between

investment i and index 1.3

 

3Ibid., Harry M. Markowtiz, pp. 77-91.

3O

Markowitz and Multinational Diversification

 

The Markowitz model has been applied to international diversi-
fication for a two country, two asset case by Grubel.4 Although Grubel
limits his model to financial asset investments, Rugman demonstrates
that the model is useful in testing the effects of direct investment.

In the context of this model, a multinational enterprise can be viewed
as a portfolio of assets in more than one economic and political environ-
ment: domestic and foreign. The return on this portfolio of assets

would be a linear combination of the returns on the various investments.

= + —
Rm XRD (1 X) Pm
where,
RMNE = the rate of return for firm 1.
RD = the rate of return for domestic investments

for firm 1.

RN = the rate of return for foreign investments
for firm i.

X = the proportion of firm i's assets invested
in domestic investments.

(l-X) the proportion of firm i's assets invested

in foreign investments.
The variance of the rate of return of a multinational enterprise,

0 is a function of the variability of returns in the separate econ-

MNE’

omies, the proportion of total investment in each economy, and the

covariance between return on investment in both economies.

 

4Herbert G. Grubel, "Internationally Diversified Portfolios:
Welfare Gains and Capital Flows," American Economic Review, 58 (December
1968):1299-1314.

5

 

Ibid., Alan M. Rugman, Journal of International Business Studies.

 

where,

(X)20§ + (l-X)Zo§ + 2 (X)(l-X)o

31

D,N

the variance of the rate of return on
investment for a multinational enterprise.

the variance of the rate of return on
investment in the domestic economy.

the variance of the rate of return on
investment in the foreign economy.

the proportion of total assets in the
domestic economy.

the proportion of total assets in the
foreign economy.

the covariance of returns in the domestic
and foreign economies.

Since the covariance of returns between domestic and foreign

investments can be written in terms of standard deviations and the cor-

relation of returns in each economy, the total variance for the firm can

be written as a function of these variables.

where,

D,N

= (x)20g + (l-x)202N + 2(x)(1-x)o

DONDD,N

the standard deviation of the rate of
return on investment in the domestic economy.

the standard deviation of the rate of
return on investment in the foreign economy.

the correlation coefficient between invest-
ments in the foreign and domestic economies.

If the variance of returns on investment in both economies is

assumed to be 12% and half of the firm's assets are in each country,

then,

32

2 _ 2 _ 2 _
OMNE _ (.5) (12) + (1 .5) (12) + 2(.5)(1 .5)(/i3)(/1—2)0D,N
= 3 + 3 + 6pD,N
= 6 + 60D N

Exhibit 4 shows the effect varying correlations between economies have
on the variance of returns for the multinational firm. If the return

)

of assets of both economies is perfectly positively correlated, (DD,N
= l, multinational diversification does not reduce the overall variance
of returns. Since the correlation coefficients between security mar-
kets are less than one, multinational diversification is beneficial to
investors. The degree to which investors benefit is inversely related

to the correlation coefficient, i.e., as the correlation coefficient

decreases, the benefit increases.

 

Exhibit 4

pD,N + l + .5 O - .5 -l
62 12 9 6 3 0
MNE
02 = 6 + 6( 1) = 12

MNE

2 = 6 + 6( 5) = 9
OMNE o
02 = 6 + 6( 0) = 6

‘MNE

2 = 6 + 6(- 5) = 3
OMNE '

2 = 6 + 6(- 1) = 0
UMNE

The example can be extended to show that the betas for firms

which are equal in all respects except that one firm is diversified

33

multinationally should differ. In all cases in which other factors are
held constant, the beta for the multinational firm will be lower because
the total variance for the firm will be lower as a result of multi-

national diversification. = .80 and if the variance of

If 00,1 = oN,1
the market index is 7.72%, the multinational firm will have a lower beta

than the domestic firm. The variance for the domestic firm is 12% and

the variance for the multinational firm is taken from Exhibit 4, assuming

 

 

that pN,I = .50
p 00 J— F—

8 = D,I D I = (.8)( 12)( 7.72) 1

D 02 7.72
I

p o o

8 = N,I N I = (.8)(/§)(/7.72) _

N --———-—- — .80
02 7.72
I

where,‘

SD = the beta coefficient of a firm with all
investment in the domestic economy.

8V = the beta coefficient of a firm with non-

L domestic investments.

pD I = the correlation coefficient between the

’ domestic firm's returns and the returns

on an index.

ON I = the correlation coefficient between the

’ multinational firm's returns and the

returns on an index.

0% = the variance of returns for an index.

The Cash Flow Effects of Floating Exchange Rates

If we view the rate of return of the multinational enterprise
in terms of capital gains and dividend yield, the rate of return, RN’

is the difference between the asset values at the beginning of the

34

period, V0, and the asset values at the end of the period, V1, plus

the dividends during the period, D1, all divided by the beginning asset

value.

 

This return can be further divided into returns on domestic investments

and returns on foreign investments.

Rn = (X)Rd + (l-X)Rn

The rate of return, in local currency, for multinational assets
is equal to the terminal value of assets less the initial value of the
assets plus dividends for the period, all divided by the initial asset

value. All of these terms are denominated in local currency.

 

~Lc LC ~L0
.. - +
REC = VN,1 VN,0 DN,1
LC
VN,O

If the exchange rate between dollars and local currency, (a),

is constant over that period, there will be no foreign exchange effect

as VN,O’ VN,1’ and DN,1 W111 all be converted into dollars at the same

rate. However, if (a) changes to cf), then VN 1 and DN 1 will be con-

verted at the (05 rate and will be converted into dollars at the

VN,0

the (a) rate. Thus, the rate of return on multinational assets, in

dollar terms, changes because the foreign exchange rate changes.

 

-3 a'(v§C1) - a(v§Co) +~a' (5301)
RN a(VLC )

N,O

35

 

 

~LC LC ~LC
t - + ' - _ !
RE, = (on O. a )Vle aVNJO + (0:. 0+0 )Dle
LC
aVN,0
“LC LC ”LC ~LC ~LC
~ - v + '.
p31 = 0“"151 N,0 D151) + (‘1 O”(Vim + DN,1)
L aVLC aVLC
N,0 N,0

 

RN = R11,1 + (1+thf1)( 6)

The dollar return on foreign assets is determined by the return on
assets in local currency plus a foreign exchange factor, FX. The return

on assets for the firm is affected by this.
RMNE = (X)RD + (l-X)RN + (l-X)FX

For example, assume that a firm invests $100 (2,000)LC) in a foreign
country at an exchange rate of $1:20LC. One year later, the value of
the assets is 2200 LC and a 200 LC dividend is paid. If the exchange
rate is still a, the return on investment is 20%. If, however, the
exchange rate rises (falls) by 5% the new return on investment will be
26% (14%). As the volatility of the exchange rate increases, swings

in return due to the foreign exchange factor will increase, also.

 

'-
Thus, if (0.00) is a random variable with an expected value of zero

and a standard deviation of o, the rate of return for a multinational

a'-a
a

 

will become more variable as the probability that ( ) equals zero

decreases. Since it has been demonstrated that changes in the foreign

36

exchange rate follow a random walk,6 the return for a multinational
firm should become more variable. From Exhibit 4, the variance of a
multinational firm with a correlation between the return on domestic
and foreign assets equal to .50 will be 9%. If the foreign exchange
rate becomes more variable, the return on multinational assets, in
dollar terms, becomes more variable and the variance of returns for

the multinational firm rises, also.

Cl
ll

(.5)2(12) + (.15)2(13) + (.5)(.5)(/12)(ff§)

= 3 + 3.25 + 3.12 = 9.37

Although the variance of returns for the multinational is still
less than the variance of returns for a domestic firm, 12%, the increase
in variability of foreign exchange rates raises the variance of the
multinational firm above the level of variability with fixed exchange
rates. Recent evidence indicates that floating foreign exchange rates,
even if partially managed, are more variable than fixed rates.7 Thus,
the riskiness of multinational enterprises should be higher with

floating foreign exchange rates.

Summary

The model tested in this study is developed from the work of

Markowitz. Markowitz demonstrated that the return on an investment

 

6Ian H. Giddy and Gunter Dufey, "The Random Behavior of Flexible
Exchange Rates: Implications for Forecasting," Journal of International
Business Studies, 6 (Spring 1975):l-31.

 

7Geoffrey Bell, "The New World of Floating Exchange Rates,"
The Journal of Portfolio Management, 4 (Spring l977):25-28; Thomas
Willett, Floating Exchange Rates and International Monetary Reform
(Washington: American Enterprise Institute), ch. 2.

 

 

37

is a weighted average of the returns on each asset in a portfolio. The
variance of the returns on the portfolio will be reduced to the degree
that the assets in the portfolio are less than perfectly positively cor-
related. Grubel reformulated this model to apply to multinational
enterprises which are viewed as portfolios of assets in more than one
country. To the extent that the correlation between economies of
countries is less than one, multinational diversification reduces the
total variance of returns for the multinational enterprise.

The return for a multinational enterprise is determined by the
return in both the domestic and foreign investments and the proportion
of assets invested in each. The returns on the foreign investment are
influenced by an additional foreign exchange factor which is determined
by the probability of a foreign exchange rate change. As the proba-
bility of a foreign exchange rate change increases, the rate of return
on foreign investment becomes more variable. Since floating foreign
exchange rates are more volatile than fixed exchange rates, the effect
should be to increase the required rate of return for a multinational

enterprise.

CHAPTER IV

RESEARCH DESIGN

This chapter describes the research design used to determine if
the change in the international monetary system has had an impact on
the risk-return relationship for multinational firms, and, if so, how
that relationship has been changed. A definition of a multinational
firm is developed based on the degree to which foreign operations affect
the firms earnings. A multinational firm is defined as one that has
manufacturing operations in six or more foreign countries. The Fortune
500 Directory of Firms for 1976 is used as the test group. Firms in
the Directory that had data available on the CRSP tapes for the test
period, 1967-1976, were divided into three groups: (1) Firms with no
overseas manufacturing, (2) Firms with manufacturing facilities in six
or more foreign countries, and (3) Firms with one to five foreign
operations. In order to get two distinct test groups, the third group
was eliminated in order to make the remaining groups as distinct as
possible. Analysis of covariance was used to determine if and how

beta coefficients for the two groups changed.

38

39

Multinational Enterprise Defined

 

A multinational enterprise can be viewed as a firm whose assets
or sales are in more than one country.1 This is a rather general defini-
tion and not of much operational value. It would include firms such as
Exxon with sales and assets in over 100 countries with small firms where
being multinational is limited to occasional exporting. Thus, the first
aspect of being multinational involves the significance of nondomestic
operations. They must be of material impact on the firm either by in-
volving a relatively large investment in assets or a large proportion
of sales.

The second aspect of being multinational involves the number of
countries in which a firm operates. Again, Exxon, with manufacturing
facilities in almost every Western country, is more multinational than
a Texas bottling company selling a portion of its output to a single
Mexican distributor. Therefore, a multinational enterprise should have
operations in several countries.2 A firm, by establishing manufacturing
facilities in nondomestic countries, is committing itself to being

multinational.

 

lJohn Fayerweather, ed., International Business - Government
Affairs. Cambridge, Mass.: Ballinger Publishing Company, 1973, p. 1;
Warren J. Keegan, Multinational Marketing Management, Englewood
Cliffs, N.J.: Prentice-Hall, Inc., 1969, p. 8; James W. Vaupel
and John P. Curhan, The World's Multinational Enterprises, Boston:
Harvard University, 1973, p. 10; Zenoff and Zwick, International
Financial Management, Englewood Cliffs, N.J.: Prentice-Hall, Inc.,
1969, p. 2.

 

 

 

 

 

2Larry Ross Lang, "The Effects of a Corporation's Multinational
Diversification on its Common Stock Holders' Return and on their Risk
Exposure," Unpublished Doctoral Dissertation, Michigan State University,
1973; Sidney M. Robbins and Robert G. Stobaugh, Money in the Inter-
national Enterprise, New York: Basic Books, 1973, p. 10.

 

 

40

For a firm to be a multinational, nondomestic resources and
commitments of the firm must be significant. A definition must involve
the proportion of total sales that derive from nondomestic sources and
the number of countries in which the firm has a manufacturing commit—
ment. In keeping with previous studies, the definition used in this
study involves two factors. Specifically, a multinational enterprise
is any firm that derived at least 20% of its sales from nondomestic
sources or has at least 20% of its investment overseas and operates

manufacturing subsidiaries in six or more countries.

Population

 

The population for this study was limited to New York Stock
Exchange firms in the Fortune 500 Directory of Firms. Firms had to
meet the previous definition of multinational for the entire period of
the study (1967-1976). This group of firms includes well over 75% of
all United States direct investment.3 Because of heavy regulation or
limited overseas exposure, the study did not deal with firms in the
following industries: banking, insurance, retailing, transportation,
and utilities.

The group of firms meeting this definition of a multinational
enterprise were subdivided into appropriate industries using the two
digit Standard Industrial Code. Each of these groups were used to

determine if the effects are different across industries. This was

 

3Ibid., Robbins and Stobaugh, p. 10.

41
done to test any bias induced by industry effects.4
The remaining firms in the sample were divided into two cate-
gories: firms with zero multinational content and firms with more than
zero, but less than required, foreign content. The former firms were
used for control purposes to compare to the group of multinational

firms.

Data Sources

 

Data for partitioning firms into the appropriate groups were
derived from annual reports and 10K reports filed with the Securities
and Exchange Commission for the years 1967 and 1976. Firms which meet
the definition of multinational for both of these years were classified
as multinational. Monthly rate of return data were taken from the CRSP

tapes. These data were used to calculate each firm's beta coefficients.

Time Period

 

The time period for this study will be from 1967 to 1976. This
time period will cover three important subperiods. From 1967 to 1971,
the foreign exchange system was a fixed system. From 1971 to 1973,
the foreign exchange system was both fixed and floating. During this
time, some currencies floated and some were fixed for varying periods
of time. The final period, 1973 to 1976, was a period of floating
rates. This study will examine the two forty-eight month periods

from 1967 to 1971 and from 1973 to 1976. Betas for industry groups

 

4Benjamin King, "Market and Industry Factors in Stock Price
Behavior," Journal of Business, 42 (January l966):139-l70; John D.
Martin and Robert C. Klemkosky, "The Effect of Market Risk on Port-
Folio Diversification," Journal of Finance, 30 (March l975):l47-l54-

 

 

42
and diversified portfolios were calculated for each period. This was
the primary analysis to test if the systematic risk for multinational
firms has changed because of the change in the foreign exchange

system.

Analysis of Covariance5

 

When comparing two groups of observations or a single group
before and after an event, a simple t-test may be sufficient. However,
as the number of groups rises, the probability of rejecting a true null
hypothesis of no differences between groups rises. Consequently, an
alternative procedure is required if several groups are to be compared.
An appropriate procedure is the analysis of variance.

A group of sample observations, Y divided into p classes

ij’
with m observations in each class is obtained.

Yij i=1,...,p j=l,...,m

The mean of each group is obtained by summing the observations in each

group and dividing by the group sample size.

_- 2 = - _- 2 - _- 2
E.(Yij Y) Em(Yi Y) + E(Yij Y)

13 1 J
The research question now is, are the groups similar or are the groups
different?
This question is answered by first separating the total variation
of the observations into two parts -- the variation between the separate

groups and the variation of observations within each group. This latter

 

5J. Johnston, Econometric Methods, 2nd, New York: McGraw-Hill,
1972, Chapter 6.

 

43

variation is the error term or individual, random variability.

- l m
Y. = —- Y.. i = l ... p
Z 9 9
1 m j=l 13
where,
ij (Yij-Y)2 = the total variation of the observations.
Em (Yi-Y)2 = the variation between groups.
2 - 2 . . . .
j (Yij-Y) = the variatlon w1th1n groups - error.
- 1
Y =-—— (2.. ) = the grand mean.
mp 13 13

The test statistic for this test is the F-ratio. This ratio is
the ratio of the mean squared differences between groups divided by the
mean squared differences within groups. As the groups become increas-
ingly dissimilar, the mean squared differences between groups becomes
relatively larger and the error term becomes relatively smaller causing
the test statistic, F, to rise. If the F that is calculated is larger
than the critical value for F, the groups are considered to be dis-
similar, i.e., at least one of the groups is unlike the others.

A shortcoming of ANOVA is that it assumes that the test obser-
vations, Yij’ are affected only by the variables included in the test.
If this is not true for a particular study, the error term may be over-
estimated, which will result in a reduced F value. Exhibit 76 demon-
strates this phenomenon. An analysis of variance would predict a mean
group difference of magnitude A. In fact, because the two groups began

at different mean levels, 1 versus 2, the mean group difference is

 

61bid., p. 193.

44

only B. Consequently, the F ratio calculated would be small and
possibly not reject the hypothesis that the group means are the same.
Analysis of covariance, ANCOVA, is a statistical method to correct the

effects of uncontrolled variables.

Exhibit 7

Dependent
Variable (Ri)

 

 

 

Independent
Variable (Rm)

The ANCOVA model begins with sample observations: Xrij and Yij.
There are "p" groups with "n" observations in each group. There are "r"

independent variables, Xrij, and one dependent variable, Yij. The

simplest model that can be used to explain this relationship is

Y = XB + u (1)

Where,

Y = an (nxl) vector with p subvectors.
= a vector of sample observations.

X = an (nxh) matrix of dependent variables.

8 = an (hxl) vector of coefficients for the
dependent variables.

h = the total number of variables, dependent
and independent.

n = the total number of observations.

45
This model implies that ”the grouping of the data into p classes has
no relevance or significance, the variation in Y is effectively
explained by the variation in the X variables."7 However, this model
is incorrect if there are group differences initially.
To test for group differences, the model must be expanded.

The expanded model has an additional element for the intercept vector.
Y=DOL+X8+u (2)

Where,

D
II

an (mp x [p-1]) matrix of dummy variables.

Q
II

a ([p-l] x 1) column vector of intercepts for
the independent variable groups.
If all group intercepts are equal, then, the ordinary least squares
residuals for this model are equal to the ordinary least squares
residuals for the simple model. If there are no group effects, the
alphas will all be zero. The F ratio is the ratio of the incremental
mean squared error explained by the expanded model divided by the mean
squared error term. If the F is larger than the expected F value,
then, the hypothesis that all of the a = O is rejected.

In order to test for varying slopes and intercepts, the model
is expanded again. This time the X matrix for independent variables
includes a column of ones which represent the intercept terms. The

vector includes a term that represents the intercept value.

 

71bid., p. 194.

Y1 = X181 + r1 1 = l,..., p (3)
Where,
Yi = a vector of independent variables.
Xi = a matrix of intercept and slope terms.
Bi = a vector of the coefficients for the
intercept slope terms.
ri = a vector of the residual values.

If a matrix, 2, is formed where the diagonal elements are Xi matrices
and all other values are zero, the model for testing differential inter-

cepts and slopes follows.

Y = 28 + r

Where,
Z a the block-diagonal matrix of Xi.
B = the (ph x 1) column vector of the i.

R = the column vector of R1.

Equation 1 assumes that the intercepts and slopes for each group
are equal. Equation 2 allows the intercepts to vary but holds the slopes
constant. The difference between the residual errors for Equation 1
and the residual error for Equation 2 is the incremental reduction
in error achieved by allowing the intercepts to vary. The difference
between the residual error for equation 2 and equation 3 is the
incremental benefit of allowing the slopes to vary. This is the rele-
vant test for this study. The test for the significance of this rela-
tionship is the ratio of the incremental residual reduction, 83,

divided by its degrees of freedom, (pk-p-k+l) to the residual error of

47
Equation 3, S4, divided by the appropriate degrees of freedom, p(m—k).

83/(pk - p - k + 1)

F = S4/p(m-k)

 

If this value is greater than the table value of F with (pk-p-k+l), p(mrk))
degrees of freedom, the null hypothesis of homogeneity of slope would be
rejected.

The sample points in this study are in the [Ri,Rm] space calcu-
lated for multinational firms before and after the foreign exchange
system changed. This results in a time series of sample points from
January, 1967 to December, 1976. This time period is divided into
three sub-periods: January, 1967 to December, 1970; January, 1971 to
December, 1972; and January, 1973 to December, 1976. These time periods
correspond to three eras. The first forty-eight month time period repre-
sents the fixed rate foreign exchange system of Bretton Woods. The
middle twenty-four month time period is transitional. The third forty-
eight month period represents the era after Bretton Woods collapsed.

ANCOVA was used to test the betas calculated for each of these
periods for portfolios of firms to determine if the beta relationships
changed as a result of the change in the foreign exchange system. The
primary hypothesis involves the slope coefficients of the portfolio
regressions. It was hypothesized that the slope coefficients have not
changed. If this hypothesis is rejected, the beta coefficients can be
further examined to see if they have become larger or smaller. If the
former case is true, investors are requiring a greater return from
multinationals and increasing their evaluation of the firms' riskiness.

In the latter case, the new foreign exchange system is viewed as making

48
a multinational firm less risky. In addition, a control group of non-
multinational firms was examined to determine if additional factors

have changed which could explain a change in the multinationals.

Summary

The objective of this study is to determine the affect that
floating foreign exchange rates have had on the riskiness of multi-
national enterprises. Mu1tinational enterprises are defined as those
companies that have investment in six or more foreign countries and
either 20% of sales derived from foreign sources or 20% of investment
overseas. Domestic firms are firms that have no overseas investment
or sales. The sample for taken from firms listed in the Fortune 500
Directory of Firms. Further data for segregating the firms into groups
were derived from annual reports and form lOK's for all firms that were
listed on the New York Stock Exchange for the study period and had a
full set of data available on the CRSP tapes. The firms with all data
were divided into three groups: (1) those which are multinational,

(2) those which are domestic, and (3) those which are in neither
category.

The monthly returns for each firm were regressed against the
returns on the Standard and Poor 500 Index, including distributions, to
calculate portfolio betas for both the domestic and multinational
groups for the fixed foreign exchange period, 1967-1970, and the
floating foreign exchange period, 1973-1976. Analysis of covariance,
using the market return as the covariate, was used to determine if the
slope coerficients for each group changed, from one period to the next.

In the case of the multinational firms, the betas decreased which

49
indicates a decrease in riskiness. The domestic firms' betas increased

indicating that they have become more risky.

CHAPTER V

Results

Introduction

 

The population for this study was the Fortune 500 Directory of
Manufacturing Firms. Of the 500 firms listed, 193 firms were found
that were multinational and had data available on the CRSP tapes for
the test period 1967-1976. Another 49 firms were found that had no
overseas investment for the period of the study. These domestic firms
were used as a control group to assure that any changes occurring in
the multinational group were not also occurring in the domestic group.
The data used to classify firms in these two groups were derived from
annual reports and Form 10K reports for the individual firms to deter-
mine the number of countries in which each firm had investments and

the proportion of multinational sales.

Calculation of Statistics

 

To test for homogeneous slope coefficients and to calculate
the regression coefficients for the various sample groups, two separate
regressions were run to obtain the F-ratio for homogeneity of the
slope coefficients over the period of the study. The first regression
assumes an additive model in which the slope coefficients are equal

for both time periods.

50

51

R = a + bD + CR

1 m
where,
Ri = the monthly return for security i
a = the intercept term for the regression
b = the regression coefficient for the dummy
variable D
D = the dummy variable to classify a return in

either the floating rate period (0) or the
fixed rate period (1)

c = the regression coefficient for the market
return variable Rm

R = the monthly rate of return on the market
index including dividends.

The results of this regression are compared to the results of a second
regression which includes an additive term to allow the slope coeffi-

cients for the two periods to vary.

R. = a + DD + CR + d(D'R )
1 m m

where,

D.
H

the slope coefficient for variable D'Rm

C1
2!
H

the variable obtained by multiplying D by Rm -
the interaction term.

The F-ratio is calculated by taking the mean difference in
explained variation of the two models and dividing by the mean residual

error term.

 

F (Ri - :§)/dfl
(1 - R2)/df2
where,
R2 = the explained variance in the full or inter-

active model

52

R3 = the explained variance in the additive model
(l-Ri) = the residual or unexplained variance

df2 = the degrees of freedom for the denominator

dfl = the degrees of freedom for the numerator

For this study, the degrees of freedom for the numerator is one and the
degrees of freedom for the denominator is assumed to be m for all cases

because the sample sizes are large, i.e., greater than 120.

Aggregate Test Results

 

The first test was to determine if there has been a change in
beta for both test groups from the fixed rate period, 1967-1970, to the
floating rate period, 1973-1976. The F-ratio for the multinational
group is 79 and for the domestic group is 4.73 (Table I). Both of these
ratios are significant at the .05 level. Consequently, the hypothesis
that the slope coefficients for each period - fixed rate vs. floating
rate - are equal is not supported by these data. Table I shows the

slope coefficients for both groups of firms. Multinational firms have
_____,__&‘__““M~___‘

‘Muc.__h .-““~
become less risky, as measured by beta, and domestic firms have become

_moreflfisky,"as measured by beta.
(”I TABLE I

AGGREGATE TEST RESULTS

 

 

 

Beta
Group Number F—Ratio 1967-70 1973-76
MN 193 79 * 1.16 .95
DOM 49 4.73** .88 1.26

 

*Significant at the .001 level.
**Significant at the .05 level.

53

Firm Size Effects

 

A major criticism of studies comparing multinational firms with
non-multinational firms is that multinationals are systematically
larger than non-multinationals. As can be seen in Table II, multinational
firms have larger sales, on average, than non-multinationals. This could

provide an alternative explanation for differences between the two

 

 

 

 

groups.
TABLE II
SIZE OF FIRMS
($ Millions Sales)
Multinational Firms - Quartiles
Number Mean Range

Highest 49 9302.2 3032.8 - 48630.8
Upper Middle 48 1987.4 1421.4 - 2967.6
Lower Middle 48 1054.7 747.0 - 1385.4
Lowest 48 510.0 327.7 - 745.9

Total 193 3245.1 327.7 - 48630.8

Domestic Firms - Halves

Higher 25 1504.5 689.9 - 3727.3
Lower 24 486.0 339.3 - 652.0

Total 49 1005.6 339.3 - 3727.3

 

To determine if size is a factor in comparing these two groups
of firms, the multinational firms were divided into quartiles. Each of

these sub-groups were tested to see if the slope coefficients changed

consistently within each quartile.

54

The domestic firms, because of the

smaller sample size, were divided in half, and each of these groups was

tested for homogeneous slopes.

that all six groups had slope coefficient changes.

TABLE III

SIZE EFFECTS

The R-ratios listed in Table 111 show

 

Multinational Quartiles

 

 

 

 

F-Ratio* Number 1967-70 1973-76
Highest 22 49 1.21 .98
Upper Middle 11 48 1.12 .97
Lower Middle 24 48 1.03 .84
Lowest 26 48 1.29 1.01

Domestic Halves

Higher 24 25 .94 1.24
Lower 16 24 .97 1.28

 

*All are significant at the .001 level.

Table III shows the slope coefficients for each of the six

groups before and after the foreign exchange system changed.

In all

cases, the multinational firm slope coefficients decreased, and the

domestic slope coefficients increased.

Foreign Sales Level Effects

It has been demonstrated that the proportion of foreign

sales to total sales impacts on the riskiness of the multinational

55

firm.1 To determine if the proportion of multinational sales affects

the results of this study, the multinational firms were divided into

four groups. One group of forty-eight firms was composed of firms that
did not report multinational sales. The remaining firms (145) were
divided into three groups from the highest proportion of multinational
sales to the lowest proportion of multinational sales. Each of these
groups was tested to determine if the slope coefficients were homogeneous
with respect to proportion of multinational sales from 1967-70 to 1973-76.
Table IV indicates the Reratios for each group. All four F-ratios are
significant at the .10 significance level, indicating that for each

group the slope coefficient shifted.

TABLE IV

FOREIGN SALES LEVEL

 

 

 

Beta
Group Number F-Ratio 1967-70 1973-76
Highest Proportion 48 5.93** 1.13 1.03
Middle Proportion 48 19 * 1.21 1.01
Lowest Proportion 48 22 * 1.17 .90
Non—Reporting 49 31 * 1.13 .86

 

*Significant at the .001 level.
**Significant at the .020 level.

 

1Rugman, Alan M., "Risk Reduction by International Diversifi-
cation," Journal of International Business Studies 7 (Fall/Winter
1976): 75-80; . "International Diversification by Financial and
Direct Investment," Journal of Economics and Business, 29 (Fall 1977):
31-37.

56
Table IV indicates the slope coefficients for each group for
the 1967-70 and 1973-76 periods. In all four cases, the slope coeffi-'

cients for each group declined.

Number of Countries Effects

 

Many studies classify multinational firms on the basis of the
number of countries in which the firm has investments.2 To test if
beta changes were consistent with respect to the number of countries
in which a multinational firm has investments, the multinational firm
group was divided into three groups. The test results indicate that in
all three groups, betas dropped significantly. Therefore, the number of
countries in which a firm has investments did not impact on the results

of this study.

TABLE V

NUMBER OF COUNTRIES

 

 

 

Beta
Group Number F-Ratio 1967-70 1973-76
(1) 6- 8 68 49* 1.27 .97
(2) 9-15 69 26* 1.15 .96
(3) 16-99 56 13* 1.04 .90

 

*Significant at the .001 level.

 

2Fayerweather, John, ed., International Business - Government
Affairs, Cambridge, Mass.: Ballinger Publishing Company, 1973, p. 1;
Warren J. Keegan, Multinational Marketing Management, Englewood Cliffs,
N.J.: Prentice-Hall, Inc., 1969, p. 8; James W. Vaupel and John P.
Curhan, The World's Multinational Enterprises, Boston: Harvard Univer-
sity, 1973, p. 10; Zenoff and Zwick, International Financial Manage-
ment, Englewood Cliffs, N.J.: Prentice-Hall, Inc., 1969, p. 2.

 

57

IndustrygEffects

 

A final question deals with industry effects.3 Multinational
firms were divided into industry groups and each group was tested for
beta shifts. Table VI describes each industry group - SIC, number of

firms in the group, the F-ratio, and the industry name.

 

 

TABLE VI
SIC Number F-Ratio Industry Name
10 5 4.05 Mining, crude oil production
20 19 0.00 Food
21 2 0.17 Tobacco
22 3 2.45 Textiles, vinyl flooring
25 l 3.01 Furniture
26 5 0.96 Paper, fiber, and wood products
28 27 22.00 Chemicals
29 14 13.00 Petroleum refining
30 5 2.19 Rubber, plastic products
32 5 4.26 Glass, concrete, abrasives, gypsum
33 9 8.97 Metal manufacturing
34 9 6.93 Metal products
36 15 5.52 Electronics, appliances
37 2 0.00 Shipbuilding, railroad and transporta-
tion equipment
38 9 0.43 Measuring, scientific, photographic
equipment
40 10 17.00 Motor vehicles
41 5 18.00 Aerospace
42 14 0.12 Pharmaceuticals
43 6 0.07 Soaps, cosmetics
44 7 4.22 Office equipment (includes computers)
45 16 4.68 Industrial and farm equipment
47 2 2.28 Musical instruments, toys, sporting goods
48 1 2.52 Broadcasting, motion-picture production
and distribution
49 2 6.18 Beverages

 

3Footnote 3 on following page.

58
Although not all industry groups had significant shifts in beta, of the
thirteen groups that did, twelve were decreases in beta and one was an
increase in beta. The eleven other industries groups appear to have
changed also, but to a lesser extent, on average. This coupled with
smaller average group sizes, probably accounts for the group differences

not being statistically significant.4

 

 

 

TABLE VII
INDUSTRIES
Beta
Industry Number F-Ratio 1967-70 1973-76

10 5 4.05+ 1.13 .75
25 l 3.01++ .99 .32
28 27 22 * 1.19 .91
29 14 13 * .95 .69
32 5 4.26+ 1.19 .91
33 9 8.97** 1.09 .77
34 9 6.93** 1.40 1.05
36 15 5.52+ 1.38 1.15
40 10 17 * 1.33 .91
41 5 18 * 1.45 .82
44 7 4.22+ 1.54 1.21
45 16 4.68+ .99
49 2 6.l8+ 1.11

 

*Significant at the .001 level.
**Significant at the .01 level.

+Significant at the .05 level.
++Significant at the .10 level.

 

3King, Benjamin, "Market and Industry Factors in Stock Price
Behavior," Journal of Business, 42 (January 1966):l39-l70; John D.
Martin and Robert C. Klemkosky, "The Effect of Market Risk on Portfolio
Diversification," Journal of Finance, 30 (March 1975):147—154.

 

 

4To test for sample size bias, power coefficients were calcu-
lated and compared for the average groups that had or did not have sig-
nificant beta changes. To detect a beta shift in the smaller sample
group, the shift would need to be 24% greater than the shift in the sig-
nificant groups. In fact, the shifts were smaller, and therefore not

detected.

59

Conclusions

 

Over the period when the international monetary system changed
from a fixed rate system to a floating rate system, multinational firms
have become less risky as measured by beta. In addition, multinational
firms have become relatively less risky than the domestic firms, which
have become more risky, as measured by beta. Betas for multinational
firms were shown to be lower after the change in the international
monetary firms regardless of several factors which might have affected
these results - the size of the firm as measured by sales, the propor-
tion of foreign sales to total sales, the number of countries invested
in, and the industry. In all sub-categories, except industry group 49
- beverages, betas for multinational firms decreased, and betas for
domestic firms increased, which supports the hypothesis that multi-
national firms are now less risky in spite of the change in the inter-
national monetary system.

Now that the international monetary system has changed and
foreign exchange rates are more volatile, the riskiness of multinational
firms might logically be expected to be higher, causing the firms' betas
to rise. This research does not support their conclusion. It may be
that multinational firms are less risky because other factors are more
important than floating foreign exchange rates on the riskiness of these
firms. If the domestic environment has become more risky because of
higher and less stable inflation rates, increased hostility toward
business, or increased government intervention, multinational firms
would be less influenced because their assets are diversified and not
all subject to these adverse effects. Domestic firms, which do not

mitigate against the increased domestic risk with foreign investments,

60
must absorb the total increase in risk. Therefore, multinational firms
have not only become less risky, but have become less risky when
domestic firms have become more risky.

A second possible explanation for these results deals with the
impact of the change in the international monetary system upon the way
foreign exchange rates are determined. Under the fixed foreign ex-
change rate system, the ultimate decision to change a foreign exchange
rate was made in the political arena. Thus, although all economic
factors might indicate that a foreign exchange change was needed, no
change would be made, for political reasons, until the situation became
so serious that only traumatic changes would do, e.g., the devaluation
of the Mexican peso by 50% in 1978. Any probability distribution of
expected foreign exchange rates for the future would have riskiness
with respect to economic predictions and uncertainty with respect to
political changes. Under the floating rate system, the political imr
pact and uncertainty on foreign exchange rates is reduced. Therefore,
although foreign exchange rates are now more volatile, they are less

uncertain.

REFERENCES

Agmon, Tamir. "Country Risk - the Significance of the Country Factor
to Share Price Movements in the United Kingdom, Germany, and
Japan," Journal of Business (January, 1973), pp. 24-32.

 

Agmon, Tamir. "The Relations Among Equity Markets in the United
States, United Kingdom, Germany, and Japan," Journal of Finance
(September, 1972), pp. 839-856.

 

Alexander, Sidney S. "The Effect of Size of Manufacturing Corporation
on the Distribution of the Rate of Return," The Review of
Economics and Statistics (August, 1949), pp. 229-35.

 

Arditti, Fred D. and Haim Levy. "Portfolio Efficiency Analysis in
Three Moments: The Multiperiod Case," Journal of Finance
(June, 1975), pp. 797-810.

 

Bell, Geoffrey. "The New World of Floating Exchange Rates," The
Journal of Portfolio Management (Spring, 1977), pp. 25-8.

 

Carbaugh, Robert J. and Liang-Shing Fan. The International Monetary
System. (The University Press of Kansas: Lawrence, Kansas),
1976.

 

Caves, Richard E. "International Corporations: The Industrial
Economics of Foreign Investment," Economica (February, 1971),
pp. 1-27 0

Cohn, Richard A. and John J. Pringle. "Imperfections in International
Financial Markets: Implications for Risk Premia and the Cost
of Capital to Firms," Journal of Finance (March, 1973), pp. 59-
66.

 

DeVries, Margaret G. "International Monetary Developments and Changes
in the World Economy," Finance and Development (September, 1974)
from Henley, Donald 8. International Business-1975, A Selection
of Current Readings. (MSU International Business and Economics
Studies: East Lansing, MI), 1975.

 

 

Einzig, Paul. The History of Foreign Exchange, Second Edition.
(MacMillan & Co. Ltd.: London), 1970.

Fayerweather, John (ed.). International Business-Government Affairs.
(Ballinger Publishing Company: Cambridge, Mass.), 1973.

61

62

Fieleke, Norman S. "Exchange Rate Flexibility and The Efficiency of
the Foreign Exchange Markets," Journal of Financial and Quanti-
tative Analysis (September, 1975), pp. 409-428.

 

Folks, William R. Jr. and Stanley R. Stansell. "The Use of Discriminant
Analysis in Forecasting Exchange Rate Mbvements," Journal of
International Business Studies (Spring, 1975), pp. 33-50.

 

 

Giddy, Ian H. and Gunter Dufey. "The Random Behavior of Flexible
Exchange Rates: Implications for Forecasting," Journal of
International Business Studies (Spring, 1975), pp. 1-31.

 

 

Glass, Gene V. and Julian C. Stanley. Statistical Methods in Education
and Psyghology. (Prentice-Hall, Inc.: Englewood Cliffs, NJ),
1970.

 

Gordon, Myron J. "The Savings Investment and Valuation of a Corporation,"

The Review of Economics and Statistics (February, 1962), pp.
37-51.

 

Grubel, Herbert G. "Internationally Diversified Portfolios: Welfare

Gains and Capital Flows," American Economic Review (December,
1968), pp. 1299-1314.

 

Grubel, Herbert G. and Kenneth Fadner. "The Interdependence of Inter-
national Equity Markets," Journal of Finance (March, 1971),
pp. 89-94.

 

Haley, Charles W. and Lawrence D. Schall. The Theogy of Financial
Decisions. (McGraw-Hill Book Company: New York), 1973.

Hays, Richard, Christopher M. Korth and Manucher Roudiani. Inter-
national Business: An Introduction to the World of the Multi-
national Firm. (Prentice-Hall: Englewood Cliffs, NJ), pp. 1972.

 

Hays, William L. and Robert L. Winkler. Statistics: Probability
Inference and Decision. (Holt, Rinehart, and Winston, Inc.:
New York), 1971.

 

"International Payments Imbalances and the Need for Strengthening Inter-
national Financial Arrangements," Report to the Joint Economic
Committee, 87th Congress, lst Session, U.S. Government
Printing Office, Washington, 1961 from Nehrt, Lee Charles.
International Finance for Multinational Business, Second Edition.
(Intext Educational Publishers: New York), 1972.

Jensen, Michael C. "Capital Markets: Theory and Evidence," The Bell
Journal of Economics (Autumn, 1972), pp. 357-98).

Jilling, Michael and William R. Folks, Jr. "The Emergence of the
Foreign Exchange Risk Management Function," Working Paper 2,
Foreign Exchange Risk Management Project, Center for Inter-
national Business Studies, University of South Carolina.

63

Johnston, J. Econometric Methods,_Second Edition. (McGraw-Hill:
New York), 1972.

 

Keegan, Warren J. Multinational Marketing Manggement. (Prentice-Hall,
Inc.: Englewood Cliffs, NJ), 1974.

 

Kindleberger, Charles P. American Business Abroad. (Yale University
Press: New Haven, CN), 1969.

King, Benjamin F. "Market and Industry Factors in Stock Price
Behavior," Journal of Business (January, 1966), pp. 139-70.

 

Klopstock, Fred H. "The International Status of the Dollar," Essays
in International Finance (Princeton University Press: Princeton,
NJ), 1957.

Knickerbocker, Frederick T. Oligopolistic Reaction and Multinational
Enterprise. (Harvard University: Boston), 1973.

 

 

Lang, Larry Ross. "The Effects of a Corporation's Multinational Diversi-
fication on its Common Stock Holders' Return and on their Risk
Exposure," Unpublished Doctoral Dissertation, Michigan State
University, 1973.

Lessard, Donald R. "World, National, and Industry Factors in Equity
Returns," Journal of Finance (May, 1974), pp. 379-91.

 

Lessard, Donald R. "International Portfolio Diversification: A Multi-
variate Analysis for a Group of Latin American Countries,"
Journal of Finance (June, 1973), pp. 619-33.

 

Levy, Haim and Marshall Sarnat. "International Diversification of
Investment Portfolios," The American Economic Review (September,
1970), pp. 668-75.

 

Lintner, John. "The Valuation of Risk Assets and the Selection of Risky
Investments in Stock Portfolios and Capital Budgets," The Review
of Economics and Statistics (February, 1965), pp. 13-37.

 

Markowitz, Harry M. Portfolio Selection: Efficient Diversification
of Investments. (John Wiley & Sons, Inc.: New York), 1959).

 

Markowitz, Harry M. "Portfolio Selection," Journal of Finance (March,
1952), pp. 77-91.

Martin, John D. and Robert C. Klemkosky. "The Effect of Market Risk
on Portfolio Diversification," Journal of Finance (March, 1975),
pp. 147-54.

McDonald, John G. "French Mutual Fund Performance: Evaluation of
Internationally Diversified Portfolios," Journal of Finance
(December, 1973), pp. 1161-80.

64

McElroy, Elam E. Applied Business Statistics. (Holden-Day: San
Francisco), 1971.

 

Miller, Norman C. and Marina V. N. Whitman. "A Mean-Variance Analysis
of United States Direct Foreign Investment in the European
Economic Community," Quarterly Journal of Economics (May, 1970),
pp. 175-92.

Modigliani, Franco and Hassein Askari. "The International Payments
System: Past, Present and Future," Sloan Management Review
(Spring, 1972), pp. 1-16.

 

Neter, John, William Wasserman, and G. A. Whitmore. Fundamental
Statistics for Business and Economics, Fourth Edition. (Allyn
and Bacon, Inc.: Boston), 1973.

 

 

Norr, David. "Currency Translation and the Analyst," Financial
Analysts Journal (July-August, 1976), pp. 46-54.

 

Pogue, Gerald A. and Bruno H. Solnik. "The Market Model Applied to
European Common Stocks: Some Empirical Results," Journal of
Financial and Quantitative Analysis (December, 1974), pp. 917-
44.

 

Robbins, Sidney M. and Robert G. Stobaugh. Money in the International
Entegprise. (Basic Books: New York), 1973.

 

 

Rodriguez, Rita M. and E. Eugene Carter. International Financial
Management. (Prentice-Hall: Englewood Cliffs, NJ), 1976.

 

 

Rugman, Alan M. "Risk Reduction by International Diversification,"
Journal of International Business Studies (Fall/Winter, 1976),
pp. 75-80.

 

Rugman, Alan M. "International Diversification by Financial and Direct
Investment," Journal of Economics and Business (Fall, 1977),
pp. 31-37.

 

Severn, Alan. "Investor Evaluation of Foreign and Domestic Risk,"
Journal of Finance (May, 1974), pp. 545-50.

 

Shank, John K. "FASB Statement 8 Resolved Foreign Currency Accounting-
Or Did It?" Financial Analysts Journal (July-August, 1976),
pp. 55-61.

 

Sharpe, William F. Portfolio Theory and Capital Markets. (McGraw-
Hill Book Company: New York), 1970.

 

Sharpe, William F. "Capital Asset Prices: A Theory of Market
Equilibrium Under Conditions of Risk," Journal of Finance
(September, 1964), pp. 425-42.

65

Solnick, Bruno H. "International Pricing of Risk," Journal of Finance
(May, 1974), pp. 365-78.

Solnick, Bruno H. "Why not Diversify Internationally," Financial
Analysts Journal (July-August, 1974), pp. 48-54.

Solnick, Bruno H. "An International Market Model of Security Price
Behavior," Journal of Quantitative and Financial Analysis
(September, 1974), pp. 537-54.

Solnick, Bruno H. European Capital Markets. (Lexington Books:
Lexington, Mass.), 1973.

"Statement of Financial Accounting Standards No. 8 - Accounting for
the Translation of Foreign Currency Transactions and Foreign
Currency Financial Statements," Financial Accounting Standards

Board (October, 1975).

Vaupel, James W. and John P. Curhan. The World's Mutinational Enter-
prises. (Harvard University: Boston), 1973.

Willett, Thomas D. Floating Exchange Rates and International Monetary
Reform. (American Enterprise Institute: Washington), 1977.

‘IIIIIIIIII