Economics of the Arms Race
Christopher- Green*
After reviewing briefly the economic con-
sequences of the arms race, especially in the
United States, the author analyses the dyna-
mics of arms races, using insights gained
from the study of economics. A distinction is
drawn between the arms race, which the
author defines as an upward adjustment,
either quantitative or qualitative, in military
stock and personnel, and military spending
per se; mere spending figures do not recog-
nize the importance of qualitative changes in
arms technology. There are powerful forces
propelling the arms race and its negative
macro-economic consequences are not se-
rious enough to halt it – military spending
need not be inflationary, it is not necessarily
inversely related to economic growth, it has
no decisive causal connection to employ-
ment, and it has not altered the industrial
structure of our economy in any significant
way. Arguments for disarmament must be
based, therefore, on a foundation firmer than
concerns about the economic conse-
quences of high levels of arms’spending. It
is here that the tools of economics can prove
most useful, by providing insights into the
behaviour of rivals, and a means of analyz-
ing the “stability” of the race. By drawing an
analogy to the economic experience with
cartels, the author concludes that a compre-
hensive disarmament agreement, even one
based initially on mutual trust, would not be
sufficient to effect a lasting end to the arms
race. After an initial arms limitation is con-
cluded, we will still require a means of legal
enforcement, or, as is inevitably the case
with cartels, the agreement will collapse and
rivalry will spring up once again.
L’auteur d6crit bri vement les consequences
dconomiques de la course aux armements,
surtout dans le contexte amrricain, pour en-
suite analyser la dynamique du phnnom~ne A
l’aide de constatations fournies par les
sciences 6conomiques. On distingue entre la
course aux armements, que l’auteur d~finit
comme un ajustement ascendant soit quanti-
tatifou qualitatif de ressources militaires ma-
trrielles et humaines, et les d6penses mili-
taires comme telles, l’ampleur desquelles ne
refl~teraient pas l’importance des develop-
pements technologiques. De puissants
agents alimentent la course aux armements et
ses effets macro-6conomiques ngatifs ne
sont pas assez s~rieux pour la faire cesser –
les drpenses militaires ne sont pas toujours
inflationnistes, ne sont pas ncessairement
fonction inverse de Ia croissance 6conomi-
que, n’ont aucun rapport direct avee le chO-
mage, et n’ont aucun impact particulier sur la
structure industrielle de notre 6conomie. Les
partisans du drsarmement doivent donc se
fonder sur d’autres motifs que des soucis
quant aux consequences economiques
qu’auraient les d~penses militaires 6le-
v~es. C’est k ce niveau que s’av~rent
utiles les apergus offerts par les sciences
6conomiques, permettant de mieux
comprendre le comportement des rivaux et
d’analyser la “stabilit6” de la course. En
tragant une analogie avec l’exprrience dco-
nomique des cartels, l’auteur conclut qu’une
entente sur le d6sarmement, meme si fondre
sur la confiance mutuelle, ne saurait apporter
une fin durable i la course. Apr~s un accord
prdliminaire sur Ia limitation des armes, il y
aurait toujours les probl~mes de mise en ap-
plication, ou, comme c’est inrvitablement le
cas avec les cartels, l’effritement de l’entente
suivi d’une reprise des rivalits.
*Of the Department of Economics, McGill University. I would like to thank Tom White and
Hanna Zowall for research assistance. Professors Ivan A. Vlasic, William G. Watson and
Richard Bird, as well as the editors of the McGill Law Journal, made many very useful
comments on an earlier draft. None of them are responsible for the views or any errors
contained in the paper. Cathy Duggan faithfully typed the various drafts of the manuscript.
REVUE DE DROIT DE McGILL
[Vol. 28
Synopsis
Introduction
I.
II.
Inflation
The Concept of the Race
The Pace of Military Spending
The Arms Race and Military Spending: Definitions and Measure-
ment
A.
B.
The Impact of the Arms Race on the Economy
A.
B. Economic Growth
C. Employment
D.
Impact on Industrial Structure
III. Economic Analysis of the Arms Race
A. Why Does the Arms Race Continue?
B.
Is the Arms Race “Stable”?
IV. Halting the Arms Race
Introduction
The arms race is a strange and depressing phenomenon. It is a race which
is unlikely to produce a winner, a fact which ironically sheds a ray of hope. A
“winner” may not avoid taking advantage of its position, while a prospective
“loser” may not accept its position graciously. The lack of a winner does not,
however, mean that the race ends in a “tie”. It could end in nuclear holocaust,
in which case all participants, and non-participants too, are losers. Or the race
could be “called off’, in which case there may be many winners and no losers.
Short of these possibilities, the arms race is a race without end, with no
well-defined finishing line.
A race without end not only seems pointless; it is tiring, if not exhaust-
ing. There would be an obvious gain if each participant were to decide that
enough was enough, drop out of the race, and direct its talents and energies
elsewhere. Unfortunately, the internal dynamics of an arms race do not
produce voluntary drop outs. Unilateral withdrawal by a major participant is
not likely to occur even in the unlikely event of “economic exhaustion”.
Multilateral withdrawal requires rational and collective action by all major
1983]
ECONOMICS OF THE ARMS RACE
participants in the race. But collective action requires overcoming powerful
forces which have so far combined to perpetuate the arms race.
I was asked initially to discuss the economic consequences of the arms
race. However, as my work proceeded, I found that much of what interested
me most went beyond the mere impact of military spending and instead
inclined toward an economic analysis of the dynamics of the arms race.
Another reason for not limiting myself to economic consequences is that it is
somewhat unsatisfying simply to render essentially negative judgments about
the relation between arms race “cause” and economic “effect”. Given the
“stagflationary” problems plaguing many nations, and the failure to convince
world political leaders of the moral-political degeneracy of the arms race, it is
perhaps not surprising that the past decade has produced a growing interest in
the economic consequences of the arms race and military spending.’ It is an
old story: if an ethical argument won’t suffice, perhaps a “pocket book” one
will. Nevertheless, the timing of that interest is curious. Although the decade
of the 1970s was characterized by rising inflation and unemployment, by
slowing growth and productivity, it was also a decade of decline in military
expenditures in relative terms, when taken as a percentage of world Gross
National Product [GNP]. This trend has recently, and unhappily, been re-
versed. On the face of it, the timing of events is not favourable to the case that
military spending is an important cause of current economic problems. As we
shall see, economic analysis is not more gentle to this hypothesis. However,
economic analysis does yield interesting insights into the forces propelling the
arms race, the stability of the arms race and the necessary conditions for a
lasting end to the arms race. These insights may be of more lasting value than
the simple analysis of economic consequences.
The paper is divided into four Parts. Part I is devoted to the definition of
concepts, in which the arms race is distinguished from military spending. It
also includes a brief overview of trends in military spending –
the chief
economic indicator of the course of the arms race. Part II contains an
examination of the likely impact of the arms race on such macro-economic
indicators as inflation, output and productivity growth, and employment, and
its impact, if any, on industrial structure. Because military spending is so
much more important in the United States than in Canada, the paper focuses
on American data and economic conditions. 2 In Part III, the focus is altered. I
‘The studies I have in mind have been the progeny of international organizations or
commissions. See, e.g., U.N. Doe A/32/88/Rev 1., Economic and Social Consequences of the
Arms Race and of Military Expenditures (1978); C. Sanger, Safe and Sound [:] Disarmament
and Development in the Eighties (1982); and Palme, et al., Military Spending: The Economic
and Social Consequences [1982] Challenge 4 (September-October).
2Because of my lack of knowledge, I will not attempt to discuss the economic consequences
of defence spending for the economy of the Soviet Union.
McGILL LAW JOURNAL
[Vol. 28
investigate the arms race from an economist’s perspective, searching for
insights into the behaviour of national military rivals, the factors sustaining
that rivalry, and its stability. The final Part suggests, from this economist’s
view, what are the necessary conditions if collective action is to result in
effective measures of arms limitation and disarmament.
I.
A.
The Arms Race and Military Spending: Definitions and Measure-
ment
The Concept of the Race
What is an arms race? Can it be measured? If so, how? Clear thinking on
these issues is essential if anything useful is to be said about the economic
consequences of arms races. I find it helpful to distinguish between “military
spending” and “arms race”, even though, for practical purposes, the former is
used as a rough measure of the latter. Military or defence expenditure is aflow
of spending per unit time, say a year. Thus, spending to procure weapons and
military supplies, to pay personnel, and to conduct operations, comprise
military or defence expenditure. An arms race is characterized by an upward
adjustment in the stock of destructive weapons and the necessary personnel to
employ them. A simple analogy with a washtub and its contents is helpful in
distinguishing the two concepts. In such an analogy, defence spending is
represented by the flow of water from the faucet into the tub. The arms race is
reflected in the rise or fall in the water level in the tub and sometimes by a
change in the temperature of the water. Changes in the water level are
analogous to quantitative changes in the stock of arms, while changes in water
temperature indicate qualitative changes in the stock of arms, measured in a
hypothetical unit of destructive power. In a similar manner, disarmament is
defined as a reduction in the destructive power of the weapons – quantitative
or qualitative. The washtub analogy to disarmament is a fall in the tempera-
ture and in the level of the water in the tub as it cools and flows down the drain.
Simply because arms races have a qualitative as well as a quantitative
side, dollars of expenditure on arms is not a particularly good measure of the
pace of the arms race. For example, Palme notes that, even among the super-
powers, less than twentyper cent of military expenditure is directed toward the
nuclear component of the arsenal. 3 Yet the destructive power of that twenty
per cent outweighs all else. As another example, the addition of a million
rifles or mortars to the United States or Soviet arsenals would involve a large
expenditure but would hardly be considered an intensification of the arms race.
In the same way, discarding some conventional and outmoded weapons should
‘Supra, note 1, 6.
1983]
ECONOMICS OF THE ARMS RACE
not be considered a step in the direction of disarmament by either of the nuclear
superpowers.
The confusion that can arise in attempting to define and measure an arms
race is indicated in an article by a ranking member of the U.S. military
establishment:
“Arms race”-
I use the quotation marks because the term is misleading. It is not true that
the respective numbers of nuclear weapons deployed by the United States and Soviet
Union have both been spiraling upward. America has over the last 15 years, repaired,
replaced and improved certain nuclear-weapons systems, but we have not increased the
size of our nuclear arsenal. In fact, it is smaller by several thousand nuclear warheads than
in 1967. The Soviet arsenal has grown since 1967 by some 6,000 nuclear warheads.’
If numbers of weapons were the sole basis for gauging an arms race, one
would conclude from the author’s remarks that the United States is disarming,
which it patently is not, while the Soviet Union is building up rapidly its arms
stockpile, which it clearly is. The truth is that during the decade of “detente”,
the nuclear arms race continued, and both superpowers now appear to be
stepping up the pace of the race.
Before turning to military expenditure trends – which can provide only
a crude economic measure of the pace of the arms race –
it is necessary to
clarify at least one more ambiguity about the term “arms race”. There are, in
reality, at least three arms races in progress. There is the growth in the
number, accuracy and destructive power of nuclear weapons and counter-
weapons deployed by the superpowers. Then there is the proliferation of
nuclear weapons to “third parties”. Finally, there is the stockpiling of conven-
tional weapons by both nuclear and non-nuclear states.’ Military spending
best approximates the pace of the last of these arms races.
B.
The Pace of Military Spending
World military spending has risen steadily during the past thirty years.
Measured in constant 1978 U.S. dollars, military expenditures rose from
4Perle, A Freeze Means Thin Ice, The New York Times (7 September 1982) A 23. At the
time, Mr Richard Perle was Assistant Secretary of Defense for International Security Policy.
The destructive and lethal power of some conventional weapons is beginning to match that
of lower-yield nuclear weapons. See Robinson, “Qualitative Trends in Conventional Muni-
tions: the Vietnam War and After” in Kaldor & Eide, eds, The WorldMilitary Order (1979) 64.
Note that I have not included the horrific race to “perfect” and stockpile the requisite
ingredients for biological war.
REVUE DE DROIT DE McGILL
[Vol. 28
$134 billion in 1950 to $446 billion in 1979.6 Currently between 4.5 and 5.5
per cent of world Gross National Product is absorbed by the military.7 More
than half of the world total is accounted for by the U.S.A. and the U.S.S.R.
However, military expenditures by developing countries are increasing at a
more rapid rate than in the developed nations,8 a fact accounted for in good
part by the very large increase in military spending by Middle Eastern states
during the last fifteen years.9
Despite the rise in world military spending, its relative “economic
importance”, as a per cent of GNP, has been declining, at least until very
recently. During the 1970s, world military expenditures fell steadily from 6.1
to 5.3 per cent of world GNP. 10 Almost all of the decline was accounted for by
the relative decline in military spending among developed countries. In this,
the United States led the way, with military expenditures declining, in real
6 See Stockholm International Peace Research Institute [SIPRI], Yearbook 1980 [:J World
Armaments andDisarmament (1980). In 1982, world military spending was expected to reach
$650 billion U.S. current dollars or roughly $500 billion constant 1978 dollars. See Palme,
supra, note 1, 5.
7The chief uncertainty revolves around estimates of military spending by the Soviet Union.
All major sources of information on world military expenditures reject outright the official
Soviet figures which put annual military expenditures at between $20 and $25 billion in current
U.S. dollars. However, there are important differences between the estimates which are
employed by the United States Arms Control and Disarmament Agency [ACDA] and the
International Institute for Strategic Studies [IISS] on the one hand, and the estimates made by
SIPRI on the other. R. Sivard, WorldMilitary andSocial Expenditures (1980) presents figures
more in line with the SIPRI estimates. In the late 1970s, ACDA and IISS figures were about 50
per cent higher than the SIPRI and Sivard estimates that Soviet military expenditures were in
the range of U.S. $100 billionperyear. One of the reasons why the ACDA figures, which are
based on CIA estimates, are so much higher than those of SIPRI and Sivard, is that the former
base their figures on estimates of what it would cost in the United States to produce the military
goods and services it is estimated that the U.S.S.R. has produced in a given year. For a good
discussion of the problems of comparing the CIA-ACDA estimates of Soviet military expendi-
ture with figures for NATO countries, or in making comparisons over time, see SIPRI,
Yearbook 1979 [:JWorld Armaments and Disarmament (1979) 28-32.
‘The income “elasticity” of military spending (which is equal to the percentage change in
military expenditure divided by the percentage change in GNP) is much higher for developing
than developed countries, but has been declining in both. Using figures from Sivard, ibid., the
elasticity was +.49
(1970-78) in developed countries compared to
+ 1.77 (1960-70) and + 1.19 (1970-78) for developing countries. For the 1970s the ACDA
figures for developed countries yield an elasticity of + .39 which is comparable to the Sivard
figure, but produces a lower figure for developing countries –
(1960-70) and +.37
9As aper cent of GNP, military expenditures by Middle Eastern countries rose from 10,3
per cent in 1970 to 15.1 per cent in 1975 and remained at the latter level for the remainder of the
1970s.
+ .89.
“In 1960, world military expenditures as a per cent of world GNP were about 6.5 per cent
(ACDA estimates).
1983]
ECONOMICS OF THE ARMS RACE
terms, by about 14per cent between 1970 Vietnam War levels and 1979.11 As
a result, American military expenditures fell from about 8.0 per cent to 5.2
per cent of GNP over the decade. 2 The decline in military expenditures [ME]
as a proportion of central government expenditures [CGE] was even more
dramatic. In 1970, the ratio of ME to CGE was roughly one third in both
developed and developing countries. By 1979, the ratio was just above a fifth
in both types of countries, with rapidly rising social welfare expenditures on
health, education and income maintenance filling much of the gap left by the
relative decline in military expenditures. For the United States, the ratio of
ME to CGE fell from 50 per cent in 1960 to under 25 per cent in 1980.
However, as the 1970s came to a close there were signs of a step-up in
military spending, particularly among the superpowers –
a rise confirmed in
the first two years of the 1980s. In the United States, the Reagan Administra-
tion’s stated objective is a nine per cent increase in real spending on the
military, which will raise the U.S. ME to GNP ratio from 5.5 in 1981 to 6.8
per cent in 1985. Both the U.S. and the U.S.S.R. blame each other for the
turn-around, but the real reasons are undoubtedly complex, being at least
partly embedded in new technology and in the perceived need to “modernize”
existing arms stockpiles. In any event, the sharp rise in the United States
defence budget and the heavy commitment of the Soviet Union to its military
establishment are beginning to weigh heavily on both economies. In the case
of the Soviet Union, the lack of economic slack and a declining rate of growth
make all the more obvious the economic drain on the civilian economy caused
by a large and growing defence establishment.’ 3 In the United States, the
sharp change in the level of defence expenditure and the method of financing
the current huge federal government budget deficits to which the rise in
military spending and Reagan tax cuts have contributed about equally, por-
“On this point, ACDA and SIPRI figures agree-with the latter actually indicating a larger
decline. During the same period, U.S.S.R. spending on the military is estimated to have risen
by 14.3 per cent (SIPRI) and 30.4 per cent (ACDA). The continued rise in real military
spending by the Soviets during “detente” is one of the stated reasons for the current United
States military build-up. It should be noted, however, that since 1976, the first year in which
the United States was no longer involved in Vietnam fighting, real military spending by the
U.S.A. has been rising. Thus much, if not all, of the decline in U.S. real military spending in
the 1970s is probably attributable to the phasing down and eventual conclusion of the Vietnam
War.
2 See Table 1 which also includes figures for Canada, the Soviet Union and NATO Europe.
13 The governments of both the U.S.S.R. and the U.S.A. are well aware of this fact and some
experts on both sides suggest, or at least hint at the possibility, that the American build-up is
aimed partly at producing economic exhaustion in the U.S.S.R. See United States Arms
Control and Disarmament Agency, World Military Expenditures and Arms Transfers 1970-
1979 (1982) 19-25; Konobeyev, The Capitalist Economy and the Arms Race [1980] Int’l
Affairs [Moscow] 192 (No. 8).
McGILL LAW JOURNAL
[Vol. 28
MILITARY EXPENDITURES [ME] AS A PERCENTAGE OF
GROSS NATIONAL PRODUCT [GNP]
TABLE 1
(4)
Soviet
Unionc
–
(5)
NATO
Europe’
(1)
(2)
(3)
–
–
–
–
Canadab
5.1
10.2
9.0
7.6
8.0
7.1
6.7
6.0
6.1
6.0
5.4
5.3
5.1
5.2
5.2
5.5
5.9
6.3c
United
Statesa
–
–
–
–
(7.9)
(7.0)
(6.6)
(6.0)
(6.1)
(5.9)
(5.3)
(5.3)
(5.1)
(5.2)
–
–
–
–..
Year
1950
1955
1960
1965
1970
1971
1972
1973
1974
1975
1976
1977
1978
1979
1980
1981
1982
1983
Figures from two sources are shown. The first column is taken from Stockholm International
Peace Research Institute [SIPRI], Yearbook (1980) and (1981). The second column in
parentheses is from the United States Arms Control and Disarmament Agency [ACDA],
World Military Expenditures and Arms Transfers 1970-1979 (1982) and is provided for fair
comparison with the figures in columns (4) and (5).
2.6
6.3
4.2
2.9
2.4
2.2
2.1
1.9
1.9
1.9
1.8
1.9
2.0
1.8
1.8
1.8
..
–
12.4
12.8
12.0
9.7
9.6
9.0
8.7
10.3
9.9
9.5
9.2
9.1
–
(14.7)
(14.3)
(14.4)
(14.0)
(14.0)
(14.1)
(14.3)
(14.0)
(14.1)
(14.4)
–
3.7
3.7
3.7
3.6
3.6
3.7
3.6
3.6
3.6
3.6
–
–
—
b SIPRI, ibid.
SIPRI, ibid., for the figures in the first column. United States ACDA, supra, note a, for the
figures in parentheses in the second column.
d United States ACDA, ibid.
I Estimates derived from Budget of the United States Government, Fiscal Year 1983.
tend a combination of cost-increasing bottlenecks in military-related indus-
tries and general deflation, in output as well as prices, in the economy as a
whole. This is a subject to which we shall return later.
1983]
ECONOMICS OF THE ARMS RACE
II.
The Impact of the Arms Race on the Economy
There are two major ways in which an arms race affects the economy.
First, where arms races involve high levels of defence spending, the way in
which that spending is financed can affect such economic aggregates as
inflation, output and employment. In addition, if there is a sharp change in the
level of defence spending, however financed, it can have dislocating effects
on the economy because the additional spending tends to be concentrated in a
few industries and sectors of the economy. Second, modem arms races are
highly technological in character. As a result, arms races often result in an
absorption of highly skilled and specialized resources such as engineers,
scientists, computer experts, and Research and Development facilities, which
might otherwise have been employed more constructively in civilian indus-
tries. While there are undoubted spillovers from military research and de-
velopment into civilian uses, as for example in aircraft, computer, satellite,
and communications technology, the employment in United States defence-
related work of 19, 22, 38, 54 and 59 per cent of all machinists, electrical
engineers, physicists, aircraft mechanics, and aeronautical engineers, respec-
tively, represents a real cost which could affect adversely productivity and
economic growth. 4 This said, what are the consequences of the arms race for
the economic indicators by which we generally measure the health of the
economy?
A.
Inflation
Military spending, per se, is not inherently inflationary. A dollar spent
on military goods and services is not economically unlike a dollar spent on
other goods and services such as education and health. While it is sometimes
argued that military expenditures are inflationary because they are contracted
on a “cost-plus” basis – with little control over rising costs –
such be-
haviour, if common, produces only once-and-for-all increases in the prices of
a subset of all goods and services. By contrast, inflation is a continuous
increase in the price level, affecting the prices of all goods and services. For
inflation to proceed the money supply must expand. This fact implies that it is
the method by which military expenditures are financed which may make
them inflationary. When military expenditures, or other government expendi-
tures, are financed fully by taxes, the withdrawal of spending power from the
private sector offsets the increase in spending by government. 5 However, if
,4 See Palme, supra, note 1, 10. See also Nelson, TheImpact ofArms Reduction on Research
and Development (1963) 53 Am. Econ. Rev. 435 (Papers and Proceedings).
“However, the balanced budget theorem in economics suggests that the increase in taxes
may have to exceed the increase in spending if the two are to be economically, as opposed to
budgetarily, offsetting.
REVUE DE DROIT DE McGILL
[Vol. 28
government expenditures are at least partially deficit financed, and the deficit
is “monetized”, that is, the government bonds issued by the Treasury in order
to cover the deficit find their way into the hands of the Central Bank, the
government spending will be inflationary. Only if military spending is more
likely to be money-financed than are other kinds of government expenditures
is it legitimate to say that military expenditures tend to be especially infla-
tionary.
Moreover, under certain circumstances, even deficit financing need not
be inflationary. If large deficits are financed by dumping billions of dollars of
government bonds into the capital markets, with little or no increase in the
money supply, real interest rates will be driven up, and both investment by
firms and interest-elastic spending by households and governments will fall as
a result. The net effect may be economic contraction of a sufficient magnitude
to cause the inflation rate to fall. The Reagan Administration’s policy of large
tax cuts, tight control over the money supply and increased military expendi-
tures is an example of this process in operation. This combination of policies
has produced large government budget deficits, unprecedentedly high real
rates of interest and severe economic contraction, but the United States
inflation rate has fallen from 12 per cent annually to under 5 per cent in little
over a year.’6
Historically, inflation and wars have often coincided. However, the
inflation of the 1970s occurred when government spending was rising, not
primarily as a result of military outlays, but because of rapidly rising expendi-
tures on social welfare including health, education and income maintenance
programmes. Table 2 shows that despite the relative decline in military
spending, the 1970s was –
a period of rising government budget deficits and inflation.
for the United States, as for many other nations
B.
Economic Growth
Sometimes it is suggested that high rates of military spending inhibit
growth. This view is based on three arguments. First, it is argued that military
adding nothing to an economy’s potential,
spending is “unproductive” –
while absorbing scarce resources. But this view overlooks the fact that, like
consumption spending, military spending may give “utility” in the form of
added security. In the absence of a supranational police force capable of
maintaining international peace, victims of aggression, as well as aggressors,
will spend large sums to protect their well-being which, in addition to tangible
forms of wealth, includes such intangibles as the nation’s “way of life” and
“honour”.
16The fear is that eventually large budget deficits will have to be monetized in order to
“stimulate” the economy, and that the result will be increased inflation.
1983]
ECONOMICS OF THE ARMS RACE
TABLE 2
UNITED STATES DEFENCE SPENDING
AND INDICES OF ECONOMIC PERFORMANCE
(1)
(3)
(2)
Defence
Outlays
as %
of GNPb
(5)
Deficit
(Surplus)
(6)
Productivity
(4)
Federal
Budget
Deficit
(Surplus)d
Inflation
Ratec
1.6
1.7
2.9
2.9
4.2
5.4
5.9
4.3
3.3
6.2
11.0
9.1
5.8
6.5
7.7
11.3
13.5
10.3
7.6g
6.0
Year
1960
1965
1966
1967
1968
1969
1970
1971
1972
1973
1974
1975
1976
1977
1978
1979
1980
1981
1982
1983
Sources: Economic Report of the President 1980; Economic Report of the President 1981;
and United States Congress, Congressional Quarterly, 13 February 1982, 228,230,234, and
444.
Fiscal year; Outlays = Expenditures + Net Lending.
(0.3)
1.6
3.8
8.7
25.2
(3.2)
2.8
23.0
23.4
14.8
4.7
45.2
66.4
44.9
48.8
27.7
59.6
57.9
110.0
180.0
as %
of GNP’
(0.6)
0.2
0.5
1.1
3.0
(0.4)
0.3
2.2
2.1
1.2
0.3
3.1
4.0
2.4
2.3
1.2
2.3
2.0
3.5g
5.39
Growth
Rater
1.6
3.8
3.2
2.0
3.3
0.2
0.7
3.3
3.5
1.9
(3.0)
2.1
3.5
1.9
0.5
(0.9)
(0.3)
9.3
7.6
7.9
9.1
9.7
9.0
8.4
7.7
7.0
6.1
5.8
5.8
5.5
5.4
5.1
5.1
5.2
5.5
5.89
6.3g
‘Percentage change in CPI year over year.
Fiscal year; in billions of dollars.
Fiscal year.
Rate of growth (decline) of output per man hour in private business sector.
Estimated.
A second, more convincing, argument is that high levels of government
military spending will reduce growth to the extent that resources are
absorbed, at least partially, at the expense of capital formation. The issue here
is the possible “crowding out” of investments which would otherwise have
McGILL LAW JOURNAL
[Vol. 28
been made. Note, however, that the potential for increasing government
expenditures to crowd out private sector investment is not limited t6 military
expenditures. In principle it applies to all kinds of government spending, with
the actual tendency to crowd out depending upon how close the economy is to
full employment and on how spending is financed.
A third way in which the level of military spending may affect economic
growth, even in cases where there is no “crowding out” of investment, is
through its potential’ impact upon the pace of technological change. If a large
portion of a nation’s best “brains” and research efforts is devoted to military
rather than civilian purposes and if these investments have little or no spillo-
ver into non-military uses, the long term rate of economic growth could be
reduced. But there is no necessary inverse relationship between military
spending and economic growth. The high rates of economic growth experi-
enced by many industrial nations in the 1950s and 1960s occurred in an era
when peacetime military spending as a proportion of GNP was at historic
heights. Moreover, there were undoubted spillovers from military research
and development into civilian uses in such fields as aeronautics, space,
telecommunications, and computers. Now that economic growth seems to
have slowed down, and there exists an unexplained decline in productivity,
some observers have suggested that one explanation may lie in a decline in the
spillover potential from current military-related research and development. 7
While the level of defence spending is not inherently growth-reducing or
inflationary, a large change in the defence budget can pose serious problems.
A $25 billion increase in defence and research and development procurement
represents only 1 per cent of United States GNP, but a very large increase in
the output of industries supplying defence, such as manufacturers of aircraft,
ordnance and communications equipment. Thus, a rapid rise in military
spending requires an abnormally large and rapid expansion of a relatively
small set of industries. The likely result is bottleneck cost increases as the
prices for particular materials and components and the wages and salaries of
persons with required labour skills are bid up. For example, whereas in the
Korean and Vietnam War build-ups, much of the additional spending was on
personnel as the armed forces expanded rapidly, the current American build-
up is concentrated on procurement and research and development. It is
estimated that over the five year period 1981-86, the real volume of military
procurement will rise by 80 per cent. This was expected to increase the
Pentagon’s non-personnel budget from 5.9 to 10.0 per cent of goods-
producing GNP –
defined as GNP minus services, even before taking
account of the depressing effect upon GNP of the recent recession in the
“See Palme, supra, note 1, 14-5. See also, supra, Table 2 for figures on the decline in
productivity.
1983]
ECONOMICS OF THE ARMS RACE
United States.'” By comparison, during the Vietnam build-up, non-personnel
outlays as aper cent of GNP minus services rose by 2.7 percentage points –
from 7.9 to 10.6 per cent. If it were not for the recent deep recession, with
substantial idle resources, the United States military expansion would be
especially prone to bottlenecks, with potential inflationary implications.
C.
Employment
In a period of high unemployment, there is a natural interest in how the
arms race affects employment. Does the arms race raise employment and
reduce the unemployment rate? Or does the arms race, in some complex
manner, actually reduce employment and raise the rate of unemployment? In
other words, were high post-war levels of military spending an important
factor in the generally low post-war unemployment picture? Is military
spending a curb on investment and long run economic growth and thereby a
contributor to rising unemployment rates in the past decade? One can find
proponents for each of these views, but most economists probably would not
be found in either camp.
After World War II brought an end to the Great Depression, some
skeptics of capitalism argued that high levels of military spending –
even war
itself – were necessary if capitalist countries were to maintain reasonably
full employment. The supposed analytical basis for this view was J. M.
Keynes’ concern that, without periodic large injections of government ex-
penditures, the inherent economic instability of modern economies, result-
ing, he thought, from volatile investment demand, would cause periods of
substantial unemployment. But even if one accepts the argument that capital-
ist economies are inherently unstable and most economists today would not,
we may ask, why are military expenditures necessary? Why not some other
form of government expenditure? The answer, which became popular in
some circles around the time of the Vietnam War, was that only military
expenditures could receive sufficient political support as an employment-
generating government policy.’ 9 But this argument lacks cogency because the
growth in government expenditures on health, welfare and education has far
outstripped the growth in government spending on the military.
An alternative view is that high levels of government spending absorb
real resources, that the taxes necessary to finance these expenditures have
been collected at the expense of investment as well as, consumption. More-
” See Schultze, Economic Effects of the Defense Budget (1981) 18 Brookings Bull. 1, 4.
19See Reich, Does the U.S. Economy Require Military Spending? (1942) 62 Am. Econ.
Rev. 296 (Papers and Proceedings).
REVUE DE DROIT DE McGILL
[Vol. 28
over, if spillovers from military research and development into civilian uses are
limited, productivity increase may be reduced, and so may a nation’s compar-
ative advantage in, and ability to export, high technology items.20 Accord-
ingly, if the arms race reduces investment and productivity growth, it will
tend to reduce the long term rate of economic growth. In turn, a decline in
economic growth will tend to raise unemployment, particularly if wages are
unable to adjust downward. The unemployment impact will be all the greater
during a period of rapid labour supply growth as has been the case in the last
decade or two, as the post war “baby boom” came of working age, and
married women moved in large numbers from the home into the labour
market.
While there is some evidence from cross-national comparisons that high
levels of military spending are associated with lower rates of investment and
productivity growth,2′ the plausibility of causal connection is weakened by
the inability to provide a coherent explanation as to why the same effects
would not be produced by other forms of government expenditure. One could
argue just as plausibly that it is the substantial growth in the relative as well as
the absolute size of government in the past two decades that has begun to have
an adverse impact on the growth of private sector investment, output and
employment. Because it is civilian –
particularly social welfare-related
expenditures –
rather than military expenditures which have accounted for
most of the growth in government spending, one might, with greater justifica-
tion, look to the former rather than the latter as a potential source of drag on
the economy.
While the argument that a high level of military spending is necessary to
the maintenance of full employment, or alternatively, that large military
expenditures are inherently employment-reducing, lacks plausibility, 22 it is
true nevertheless that employment in a given locale or region may be quite
dependent on defence expenditures 2 and that changes in defence spending
may produce cyclical changes in the unemployment rate. The tendency for
defence-related employment to be concentrated locally is an important phe-
nomenon which would create obvious adjustment problems if ever a serious
programme of disarmament were implemented 4.2 But the problems posed
‘See Rothschild, Military Expenditures, Exports and Growth (1979) 26 Kyklos 804.
21See, e.g., Smith, Military Expenditures and Capitalism [1977] 1 CambridgeJ. Econ. 61;
Palme, supra, note 1.
“The arguments are contradictory as well.
3E.g., Nova Scotia, Connecticut, Texas, California, and Massachusetts.
2′ The impact of disarmament on the Canadian economy has been examined in depth by G.
Rosenbluth, The Canadian Economy and Disarmament, rev’d ed. (1978) and by Bernard &
Truchon, Les retombdes iconomiques de l’effort militaire du Canada dans l’optique du
disarmement (1983) 28 McGill L.J. 684. For the impact of disarmament on the United States
1983]
ECONOMICS OF THE ARMS RACE
UNITED STATES DEFENCE EMPLOYMENT LEVELSa
TABLE 3
(1)
(2)
(3)
Defence-
Related
Civilian
Employment
(5)
(6)
(4)
Total
Defence-
Related
Total
Civilian
Employment
Employment,
Unemployment
Rate (%)
Department
of Defense
Personnelb
in Private
Sector’
2.3
3.2
2.9
2.4
2.0
1.9
1.7
1.7
3.7
4.8
4.8
4.3
3.8
3.3
3.3
3.2
Year
1965
1968
1969
1970
1971
1972
1973
1974
1 Source: Lynch, Regional Impact of the Vietnam War (1976) Q.J. Econ. and Bus. 37, 38.
b In millions. Includes military personnel and Department of Defense civilian personnel.
‘In millions. Includes persons employed by prime contractors and sub-contractors and persons
71.1
75.9
77.9
78.6
79.1
81.7
84A
85.9
6.0
8.0
7.7
6.7
5.8
5.2
5.0
4.8
4.5
3.6
3.5
4.9
5.9
5.6
4.9
5.6
providing supplies and services to the contractors.
In millions. Column (2) + Column (3).
In millions.
may be more political than economic. Adjustment to economic change is
often rapid. For example, the sudden 44 per cent increase in Vietnam War
procurement-related employment between 1965 and 1968 was followed just
as rapidly by a 40 per cent drop between 1968 and 1971, with most of the
increase and the subsequent contraction occurring in seven large midwestern
industrial states.’ The adjustment of labour supplies to changing demand
seems to have taken place with a minimum of dislocation.
The United States is itself cyclically sensitive to substantial changes in
defence spending. Table 3 shows the level of defence-related employment in
the United States during the Vietnam War. At the height of the conflict, in
1968, defence-related employment reached 8 million persons, or 9.5 per cent
of all employed persons. In Canada, the comparable figures for 1974 were
181,300 persons in defence-financed employment, or 2 per cent of the
economy, see W. Leontief, Input-Output Economics (1966). R. Bolton, Defense Purchases
and Regional Growth (1966) has estimated the impact of United States defence spending on
regional growth.
2See Lynch, Regional Impact of the Vietnam War (1976) 16 Q.J. Econ. and Bus. 37.
McGILL LAW JOURNAL
[Vol. 28
employed labour force.26 Table 3 also reveals that the Vietnam build-up
resulted in a rise in defence-related employment of 2 million persons between
1965 and 1968 and a subsequent decline of 2.2 million persons between 1968
and 1971. As defence-related employment rose, the unemployment rate fell
from 4.5 to 3.5 per cent during the period 1965-69, and as defence-related
employment fell, the unemployment rate rose by 2.4 percentage points.
While the defence expansion and contraction obviously contributed to the
cyclical movements in the unemployment rate, it would be an error to
overstate the direct military impact. The decline in the unemployment rate
from 1965 to 1969 was due at least partly to the well-known “Kennedy” tax
cut of 1964-65. Likewise, the rise in unemployment between 1969 and 1971
resulted, in good part, from restrictive monetary and fiscal policies put in
place after it became clear that the failure to tax-finance the Vietnam War
military build-up was producing inflationary strains on the United States,
indeed world, economy.
At times of high unemployment, the cyclical sensitivity of employment
to large changes in defence spending is sometimes used by proponents of
increased defence spending as a means to foster political support for such
spending. For example, with the American unemployment rate standing at
over 10per cent in the fall of 1982, some Reagan Administration officials saw
a “silver lining”, or perhaps a gold plated one, in the spiralling military
budget. In an attempt to head off Congressional cuts in the defence budget,
Caspar Weinberger, the United States Secretary of Defense, is reported to
have said that “[y]ou get 35,000 more jobs for every extra $1 billion you spend
on national defense”.27 Pentagon economists, tacking on the indirect employ-
ment impact of military spending, estimated the total employment effect of a
one billion dollar increase in defence spending to be between 50,000 and
60,000 jobs. If the Reagan Administration gets its way, it is estimated by the
developers of Pentagon economic models that, by 1987, military spending
will account for 3.4 million industrial jobs compared to 2.2 million in 1982.28
However, any attempt to justify a step-up of defence spending solely on
employment generating grounds is misguided. Unless the increased military
hardware is justified in its own right, it obviously would be preferable to
stimulate higher employment by producing goods and services that society
really demands.29 Moreover, as a number of analysts have pointed out,
6 See Rosenbluth, supra, note 24, xii-xiii. Input-output model figures for 1968 in Canada
are not available.
2Silk, Military Surge as Spur to Jobs, The New York Times (17 September 1982) D 2.
“‘ See ibid. The Pentagon-economists used the input-output analysis developed by Professor
Wassily Leontief, supra, note 24.
order to respond to the Soviet threat.
2 Of course, Mr Weinberger stressed that the increase in defence spending was needed in
1983]
ECONOMICS OF THE ARMS RACE
defence spending is not a particularly cost-effective method of job creation. It
is estimated that an extra $1 billion will create 76,000 jobs if spent on sewer
construction, 77,000 jobs if spent on teachers.3 Finally, if increasing defence
expenditures inflate the government’s budget deficit while the monetary
authorities sharply limit, money supply expansion, the likely result is to put
upward pressure on real interest rates, further reducing investment demand
and extinguishing jobs in the non-defence sector of the economy.
D.
Impact on Industrial Structure
Has the arms race affected the industrial structure of the economy? It
sometimes is argued that because military procurement and research and
development expenditures are concentrated on a relatively few large firms,
the arms race is contributing to an increase in industrial concentration. It is
argued further that the cost plus nature of defence procurement has dulled the
efficiency and altered the incentive structure of firms which depend heavily
on defence contract work.3 Moreover, the fact that most defence procurement
contracts are negotiated by a single buyer –
the United States Department of
Defense –
rather than awarded on a sealed bid basis, implies that the market
mechanism is bypassed and a “cozy” rather than “arms’s length” relationship
develops between purchaser and supplier. This, in turn, is said to create a
community of interest and to reinforce the idea of a military-industrial
complex.
The issues raised here are complex and easily subject to distortion. In the
short space available, it is difficult to do more than to suggest some general
directions of thinking. To begin with, there is no question that defence
spending is concentrated on relatively few firms. In the early and mid-1970s,
70 per cent of United States defence procurement centred on the 100 largest
defence contractors. Less than 20 per cent of defence procurement was
accounted for by small business.32 In addition, at the level of specific weapons
such as nuclear submarines, fighter aircraft, missile systems, and combat
10See Silk, Cost-Effective Job Creation, The New York Times (17 September 1982) D 2.
Unfortunately, none of these alternatives directly increases industrial jobs where unemploy-
ment is presently most critical.
31 Cost-plus pricing means that the price established between buyer and seller is equal to an
agreed markup over supplier’s costs, whatever they may be.
12 See Marfels, The Structure of the Military-Industrial Complex in the United States and its
impact on Industrial Concentration (1978) 31 Kyklos 409, 416. But Marfels says that only
three of the largest ten United States defence contractors are among the very largest industrial
corporations. He also says that “the firms which are heavily dependent on military contracting
are medium-sized firms only, although they are still in the group of the 50 largest industrial
corporations”.
REVUE DE DROIT DE McGILL
[Vol. 28
vehicles, the four top suppliers of each type of weapon usually accounted for a
high proportion of the contracts, often in excess of 90 per cent.33
It is true also that defence contracts are typically negotiated rather than
awarded on a sealed bid basis. Marfels reports that 90 per cent of all weapons
procurement is done without competitive bidding. When the penchant for
negotiated defence contracts is combined with (a) controls over profitper sale
rather than over cost; (b) the Department of Defense’s obsession with “per-
formance”; and (c) typically a high degree of excess capacity in U.S. defence
plants, it is not surprising that costs seem inflated, that large cost overruns are
common, and that the defence industry gives the appearance of inefficiency.
Adams has argued further that defence procurement policies have effectively
subsidized and protected large and often inefficient, firms, thereby inhibiting
turnover of leading firms and preventing competitive pressures from enforc-
ing the dictates of economic efficiency. 5
While performance in supplying the Department of Defense often would
not pass the more rigorous tests of efficiency confronting firms in the private
market place, the lasting effect on industrial structure is not clear. For
example, it is difficult to trace changes in the United States industrial structure
to defence procurement policies. Industry concentration ratios –
the share of
industry output accounted for by the largest 4 or 8 firms –
have not risen on
average. Those specific industries experiencing rising concentration have
tended to be consumer-, rather than producer-goods, industries. Moreover, a
recent study by Shepherd provides evidence that competition has increased in
the United States economy, and that the trend has been particularly marked
since 1958.36 While it is true that “aggregate concentration” (the share of all
manufacturing or industrial output accounted for by the 50, 100 or 200 largest
firms) has risen in the U.S., most of the increase took place in the first decade
after the end of World War II. In addition, the evidence suggests that at least
part of the increase in aggregate concentration is attributable to strong motives
toward diviersification by the largest firms –
sometimes by merger. In other
words, even absent the arms race, some increase in aggregate concentration
was likely.
33See J. Gansler, The Defense Industry (1980) 166.
‘See Marfels, supra, note 32. For a comprehensive study of the weapons acquisition
process in the United States, see M. Peck & F. Scherer, The Weapons Acquisition Process: an
Economic Analysis (1962).
SeeAdams, The Military-Industrial Complex and the New Industrial State (1968) 58 Am.
Econ. Rev. 652 (Papers and Proceedings); Adams & Adams, TheMilitary-Industrial Complex:
A Market Structure Analysis (1972) 62 Am. Econ. Rev. 279 (Papers and Proceedings).
ISee Shepherd, Causes of Increased Competition in the U.S. Economy, 1939-1980 (1982)
64 Rev. Econ. and Stat. 613.
1983]
ECONOMICS OF THE ARMS RACE
It is interesting that in recent years, leading military contractors, some of
which were noted in the 1960s for their inefficiency and inability to enter the
civilian market place, have now become diversified, often through merger, so
that they are no longer as dependent on military contracts. One factor operat-
ing here is the substantial number of prime contractors acquired by
conglomerates.37 Gansler reports that the percentage of defence-related busi-
ness of the 25 largest defence contractors fell from 40 per cent in 1958 to less
than 10 per cent in 1975.11 It would seem that, rather than altering the
industrial structure in any marked way, the arms race merely abets a degree of
inefficiency. In addition, there may be some tendency, conscious or uncon-
scious, to use defence expenditures as a kind of “industrial policy”, to keep
afloat large firms who, in the absence of defence contracts, would need some
form of market protection or financial bailout to survive. The obvious exam-
ples are Lockheed and Chrysler.
M. Economic Analysis of the Arms Race
We now turn away from the specific economic consequences of the arms
race to an analysis of arms race “inertia” and “stability”. We enter an area
which the reader will find more opaque, more subject to assumption and
qualification, than that through which he has just passed. It is hoped, howev-
er, that the topics are important enough and the insights sufficiently rewarding
that the reader will bear with me and the mode of analysis I have adopted in
approaching these difficult and complex issues.
A.
Why Does the Arms Race Continue?
Each year, enormous sums are spent by the superpowers to expand and
to modernize arsenals whose lethal power is nemany times that necessary to
obliterate life on earth. Each year, vast sums are spent on arms by poor
countries in what often appears to be little more than a boisterous display of
new found wealth. Alva Myrdal, a winner of the Nobel Prize for Peace, has
stated:
Every government defends its participation in the arms race as necessary to guard its
national security. But this is an illusion. What makes the arms race a global folly is that all
countries are now buying greater and greater insecurity at higher and higher costs. 9
3 See Marfels, supra, note 32.
38See Gansler, supra, note 33.
39A. Myrdal, The Game of Disarmament, rev’d ed. (1982) 7.
McGILL LAW JOURNAL
[Vol. 28
The key word is “global”. If nations prefer more security to less, it is
clearly globally nonoptimal to spend more for less. But the result is quite
predictable, even if each of the participants is perfectly rational. An example
will help illustrate the point. Suppose there are two nations, A and B, and that,
at the outset, the stockpile of arms of A and B are similar. Suppose further that
each nation values highly peace and security, but doesn’t trust its neighbour
a perceived rival. Finally, let us assume that security, which is measured
–
in index number form, is related negatively both to the ratio of rival to domes-
tic military capability and to the total stock of arms. Under these assumptions,
if both A and B increase their military spending, their security indices decline
from 1.0 to 0.8 as is indicated by a comparison of the upper left and lower
right hand boxes in Figure 1 below.
FIGURE 1
Nation B
No Increase
in Expenditure
on Arms
1.0 (A)
1.0 (B)
Increase
in Expenditure
on Arms
0.7 (A)
1.2 (B)
1.2 (A)
0.7 (B)
0.8 (A)
0.8 (B)
Nation A
No Increase
in Expenditure
on Arms
Increase
in Expenditure
on Arms
But if either A or B increases its expenditures on arms while its rival does not,
the nation which fails to increase its military capability suffers a perceived
decline in security, from 1.0 to 0.7, while its rival’s security is enhanced from
1.0 to 1.2. As the numbers in the example suggest, the insecurity produced by
a decline in relative military power exceeds the insecurity produced by an
increase in total world military power. If neither A nor B is certain what its
rival will do, and A and B are unwilling or unable to meet together and to
agree jointly to limit military spending, the safest strategy is to increase outlay
on arms. 40
1The assumption that the two nations begin with armaments which are quantitatively and
qualitatively similar is important in the context of deterrence theory. Deterrence theory
suggests that the real danger is “imbalance”. Thus, if A began with more armaments than B, an
increase in arms spending by B, but not by A, might make both nations more secure by reducing
the chances of a “first strike” for instance. In fact, claims to being behind in an arms race,
whether well-justified or not, are often used to promote a step-up in military spending. Of
course, one equally could argue that A and B would be even better off if A reduced its arms
level to that of B. But if A were willing to do so, would we, in fact, be confronted with an arms
race?
1983]
ECONOMICS OF THE ARMS RACE
We have here an example of the well-known “Prisoner’s Dilemma”.
Each agent acting on his own maximizes his gains or minimizes his losses by
following a strategy which is clearly inferior to the one he and his rivals would
choose if they acted collectively. So long as collective action is not possible,
or cannot be trusted because of sub rosa cheating, participants prefer
strategies which turn out to be globally inferior. The type of behaviour
reflected in the “Prisoner’s Dilemma” is not only rational but occurs in
numerous forms. For example, the “Prisoner’s Dilemma” explains why rival
firms in such consumer products industries as beer or cigarettes may push
advertising expenditures to levels which, at the margin, are not globally
profitable – where an extra dollar of industry advertising may add less than a
dollar to industry revenue. The dilemma faced by each individual firm is that
failure to match its rival’s advertising campaign threatens to prompt a sub-
stantial loss of market share and larger losses than does an imitative strategy.
Thus, an action which is collectively non-optimal is individually rational, as
long as cooperation, or collusion, is impossible or illegal. 4
The “Prisoner’s Dilemma” analogy is exceedingly apt in the context of
the arms race. It provides an explanation for arms race inertia which is not
dependent on conspiracy theories or the bogeyman of profiteering interest
groups. It does not even require actual aggressors. All that is essential is
incomplete information, distrust and the absence of an institution or an
organization, such as an international “police force”, capable of providing
and enforcing collective security.
As powerful as the “Prisoner’s Dilemma” analogy is, however, it does
not imply necessarily that the rational choice of an individual nation is to
increase its arms. Suppose that in the example above, the index of security
that would result if each nation were to increase its arms (the lower right hand
box) was 0.6 rather than 0.8. In that case, an individual nation’s security
index is actually lower when both rivals increase their armaments (0.6) than
when only the rival nation increases its arms (0.7).42 The rational action for a
nation is to halt any further arms expansion, even though security would be
increased if one’s own armaments increased while those of one’s rival did not.
The existence of a “Prisoner’s Dilemma” is not the only factor which can
explain arms race inertia making it difficult to stop an arms race short of war.
“None of the ideas are new. They underlie, at least partially, the important work of K.
Boulding, Conflict and Defense: A General Theory (1962); and T. Schelling, The Strategy of
Conflict (1960). See also M. Olson, The Logic of Collective Action; Public Goods and the
Theory of Groups (1971).
“Dumas discusses this possibility in detail and provides an analysis of how it might arise.
See Dumas, Armament, Disarmament and National Security: A Theoretical Duopoly Model of
the Arms Race (1979) 6 J. Econ. Stud. 1.
REVUE DE DROIT DE McGILL
(Vol. 28
Special interest groups and powerful incentives may play a role too. Modem
armed forces are led by well paid and privileged professionals who command
forces employing technologically advanced weapons developed by well-
rewarded scientists, engineers and skilled technicians, and supplied under
cost-plus contracts by firms whose market is guaranteed at taxpayer expense.
It would be surprising, almost unnatural, if these circumstances did not
generate powerful and varied voices warning sometimes with justification,
against “letting down our guard”, and advocating supposedly security-
enhancing improvements in weaponry. Compounding these forces is the fact
that the scientific knowledge on which modem weapons technology is based
cannot be destroyed, nor can it be easily turned off or kept secret.
In the United States, this complex of interest groups and incentives is
known as the military-industrial complex [MIC].43 Properly understood, the
term “military-industrial complex” is not an unreasonable characterization of
at least one of the forces in a modem arms race. Unfortunately, however,
there is a tendency in some quarters to embed the MIC concept in some form
of conspiracy theory. Doing so is not only simplistic, but overlooks the fact
that, at bottom, the problem is one of interest groups and incentives, not
unlike those which governments are constantly creating and attempting to
cope with in many other domains. The creation of large standing armies and
the demand for modem armaments to equip them, created the military-
industrial complex, not the other way around. Just as in the regulatory arena,
43 1t would perhaps be preferable to call it the military-industrial-technological complex.
That a similar complex set of interests exist in the U.S.S.R. as well, is evident from the paper
presented by Seymour Melman, The Conversion of Military Economy: The USSR, American
Association for the Advancement of Science Annual Meeting (January 1981). An explanation
for the arms race that rests on the concept of a military industrial complex is based on a very
different model of the government decision-making process than is the deterrence theory-
“Prisoner’s Dilemma” model. Graham Allison, Essence of Decision [:] Explaining the Cuban
Missile Crisis (1971) distinguishes between three models of government decision-making in
his analysis of the Cuban Missile Crisis: the “rational actor”; organizational process; and
governmental (bureaucratic) politics models. The “rational actor” model, in which it is
assumed that government makes constrained maximizing choices among a given set of
alternatives, is the model that underlies the discussion of the “Prisoner’s Dilemma”. In
contrast, the organizational process paradigm emphasizes the role of standard operating
procedures, organizational structure and routines, and parochial priorities and perceptions,
while the bureaucratic politics paradigm focuses upon the role of diverse interests, bargaining
and compromise among officials with unequal influence. Obviously, the military industrial
complex concept is much more closely related to the latter two models of government
decision-making behaviour than it is to the “rational actor” model. But in my view, and that of
Allison too, these decision-making models, although differing among themselves in fun-
damental ways, are not mutually exclusive. In fact, in an analysis of the arms race, they
complement one another nicely. The “rational actor” model is perhaps most useful in analyzing
high level strategic decisions while the latter models explain better the day-to-day decision-
making behaviour which contributes so much to arms race inertia.
1983]
ECONOMICS OF THE ARMS RACE
where regulations have begotten regulations,’ so the growth of large military
establishments has engendered its own momentum and self-protecting forces.
An example of the interaction between distrust and institutional incen-
tives that helps to propel the arms race is the apparent decision of both
superpowers to extend the competition into outer space.45 The “costs” in
resources and increased insecurity are obvious. Yet, on the other side of the
balance sheet, are the higher incomes, prestige and self-importance of the
various professionals, from military planners to scientists and engineers, who
stand to gain from widening the horizons to which the most advanced
technology and planning concepts are applied. Moreover, those who stand to
gain most will be able to argue, with some cogency, the historical inevitability
that “each time a new medium is opened to man it is exploited to gain a
military advantage”.46 In the absence of enforceable cooperation, the costs of
the arms race are the costs of avoiding military disadvantage.
Even attempts to limit the budgetary costs of the arms race can produce
perverse behaviour. 47 One of the ways in which the budgetary –
as opposed
costs of new weapons systems can be reduced is to lengthen
to opportunity –
production runs by selling abroad. Because of the heavy fixed costs associ-
ated with weapons development, the average cost of producing and develop-
ing a weapons system declines over a substantial range. As a result of foreign
sales, the costper unit of weapon or weapons system is reduced as is shown in
Figure 2, and so is the budgetary outlay for any given quantity, say X.,
purchased at home. If there are no sales abroad, budgetary outlays for X units
is OABX,. If production is X with X2 – X, units sold abroad, budgetary
outlays drop to ODCX, even though more resources are used up in the
production of the weapons system. Thus, sales abroad perpetuate the arms
race and use up additional resources while reducing the strain on the most
politically’ sensitive indicator, the supplier government’s own budget. Over
the decade of the 1970s, when the relative growth in world armament was
slowing, the annual value of arms transfers rose from 9 to 20 billion, in
“4For example, the continued regulation of railroads ultimately required the regulation of
trucking, as well as administrative control over technological innovations in transport which
threatened to alter existing competitive advantages.
” See, e.g., Vlasic, Disarmament Decade, Outer Space and International Law (1981) 26
McGill L.J. 135, 147 et seq.
46Attributed to Robert H. Stivers, Assistant Deputy Secretary of Defense for policy.
Reported in Wilford, Military’s Future in Space: A Matter of War or Peace, The New York
Times (19 October 1982) A 1.
47 Gansler, supra, note 33, found the real costs, i.e. after deducting for inflation, of military
equipment to be rising at a rate of fiveper centper annum. Thus, for a given budget-even one
the amount of weaponry that the military is able to acquire is
measured in constant dollars –
falling rapidly.
McGILL LAW JOURNAL
[Vol. 28
constant 1977 dollars, with over three quarters of the arms imports being
absorbed by developing nations. 8
$/unit
A
D
FIGURE 2
B
C
Average Cost
0
X1
X2
Quantity
B.
Is the Arms Race “Stable”?
Historically, most arms races have resulted in war. Is the present arms
race any different? If it is, will the arms stockpiles reach an “equilibrium”
level at which there is no tendency to increase further? These are difficult
questions about which it is not possible to make any definitive statements.
However, the existence of weak analogies between international and indus-
trial rivalry suggests, tentatively, that evidence on the latter may yield
insights into the former. First, it is necessary to delimit the analysis by making
certain distinctions.
Because the analysis will focus on the behaviour of the superpowers, it is
useful to distinguish “horizontal” from “vertical” proliferation of weaponry.
By “horizontal” proliferation I mean the spread of arms to “third” parties –
those not members of one of the two military alliances headed by a superpow-
4′ Note that the budgetary incentive to make sales of arms abroad is unrelated to the argument
that one goal of foreign arms sales is to enhance the trade balance (exports of traded items minus
imports of traded items). The balance of trade argument seems implausible given the small
fraction of total exports: 3.1, 0.3 and 0.7 per cent accounted for by arms sales made by the
U.S., Canada and NATO Europe respectively, in 1979. See United States Arms Control and
Disarmament Agency, supra, note 13. The ACDA figure of 16 per cent for the U.S.S.R. is
probably overstated due to the way in which Soviet arms are valued, while the U.S. figure of
3.1 per cent probably understates the real value of arms sales because some are given away and
others are sold at “below-market” prices.
1983]
ECONOMICS OF THE ARMS RACE
er. By “vertical” proliferation, I mean the arms expansion among the super-
powers and their close allies. The distinction is important because it seems
likely that even if the arms race between the superpowers does not end in war,
the horizontal proliferation of non-nuclear arms will almost certainly raise the
probability of non-nuclear war occurring somewhere in the world. 49 This fact,
plus the likelihood that armed third parties will play one superpower off
against the other, virtually assures a feedback from horizontal to vertical
proliferation and vice versa. Horizontal proliferation is especially destabiliz-
ing, while bringing stable third parties into one or the other of the alliances
headed by the superpowers may increase stability. 0
In order to investigate the stability issue, it is necessary to begin with a
relatively formal statement of the determinants of a nation’s decision to add to
its arms stockpile. This is no easy task. Nevertheless, there is some agreement
on where to start. The most thorough and influential attempt to present a
model of arms races was carried out over a thirty year period by the British
mathematician, Lewis Richardson.5 Richardson hypothesized that the time
rate of change of a nation’s armaments depends positively on its rival’s
armaments and negatively on the level of its own armaments, with a constant
term whose size depends on the nation’s aggression or “grievance”. Similar-
ly, the time rate of change in the rival’s armaments depends on the same set of
factors, so that mathematically we have, in the simplest case of two nations,
two differential equations which must be solved simultaneously. 52
4’I deal infra with the impact of the horizontal proliferation of nuclear weapons on the
probability of war.
-0See Siljak, Competitive Analysis of the Arms Race (1976) 5 Annals Econ. and Social
Measurement 283. As noted supra, the heavy costs associated with “vertical” proliferation
create a budgetary incentive to make foreign arms sales –
i.e., for the “horizontal” prolifera-
tion of arms.
11 See L. Richardson, Arms and Insecurity [:] A Mathematical Study of the Causes and
Origins of War (1960).
– 2Formally, Richardson’s theory is presented in terms of two linear differential equations
in the armaments of
expressing the factors that determine the time rate of change 4A. and IS
dt
two nations, A and B:
dt
dA
–
dt
= g + aS – bA
dS
–
dt
= h + cA-
eS
We can denote A and S as measures of American and Soviet military or strategic strength,
respectively; a, b, c, e, g, and h are all positive constants, where g and h reflect the degree of
“grievance” or “hostility” between the two sides. The negative sign in front of the bA and eS
terms is based on the assumption that the greater the size of a country’s own armaments stock
the less it will feel it necessary or desirable to increase that stock. As Richardson put it, ibid.,
15, the negative sign reflects the “fatigue and expense of keeping up defenses”. Thus, the bA
and eS terms have an economic interpretation: the resource limits upon increased arms
procurement.
REVUE DE DROIT DE McGILL
[Vol. 28
Richardson tested his model using data on the defence expenditures of
belligerents in the years preceding the outbreak of a number of European
wars, including the two World Wars. Richardson interpreted his results to
support, at least moderately, his basic hypothesis, and contended that a
science of arms races was not only possible, but necessary, if war were to be
averted in the future. Since Richardson’s work appeared, there have been
many attempts to refine his model, 3 and a few attempts to test it with data
from the current nuclear arms race.’
If arming nations behaved in the manner described by the Richardson
model, it would be possible to specify the conditions under which the arms
race would be “stable” in the sense that armament stockpiles would tend to an
equilibrium. 55 In fact, it turns out that the conditions required for stability have
significant economic content. Richardson-type arms races are potentially
stable because of the existence of what he termed the cost or “fatigue”
constraint. This is contained in Richardson’s specification that the rate of
change of a nation’s own armaments is related negatively to the level of its
own armaments, and that the larger one’s armies, the greater the resources
required to maintain them. Now, if the response coefficients of each of the
rivals is such that the drag posed by cost or fatigue outweighs the stimulus
from the rival’s level of armaments, then it can be shown mathematically that
a stable equilibrium exists. 6
There is a catch, however, and this brings us to another economically
meaningful factor affecting stability. The mathematically formulated
Richardson model assumes implicitly that each nation’s armaments are
homogeneous, that is, qualitatively similar. Changes in armaments levels are
51 See Caspary, Richardson’s Model ofArms Races: Description, Critique and an Alternate
Model (1967) 11 Int’l Stud. Q. 63; Brito, A Dynamic Model of an Armaments Race (1972) 13
Int’l Econ. Rev. 359; Intriligator, Strategic Considerations in the Richardson Model of Arms
Races (1975) 83 J. Pol. Econ. 339; Intriligator & Brito, FormalModels ofArmsRaces (1976) 2
J. Peace Sci. 77; Gillespie, et al., An optimal ControlModel ofArmsRaces (1977) 71 Am. Pol.
Sci. Rev. 226; Sorenson, Modeling the Nuclear Arms Race: A Search for Bounded Stability
(1980) 4 J. Peace Sci. 169.
‘See McGuire, A Quantitative Study of the Strategic Arms Race in the Missile Age (1977)
59 Rev. Econ. and Stat. 328.
–
55 Here equilibrium means a “position of rest”. Stability means that a small move away from
equilibrium will produce a move back to equilibrium. Incidentally, a stable arms race does not
mean that war will not break out, and an unstable arms race, in which armaments stockpiles
grow explosively, does not necessarily imply that the end result is war. Wars require some
“flash point” as well as the existence of armaments. It is perhaps unnecessary to remind the
reader that the economic consequence of arms races are nothing by comparison with the
economic consequences of a major war.
-‘See Sorenson, supra, note 53.
1983]
ECONOMICS OF THE ARMS RACE
changes in quantities, not qualities .57 Now the quantitative-qualitative distinc-
tion suggests a well-known economic parallel: the differential tendency to
equilibrium of homogeneous and differential oligopoly. The argument is as
follows. When two, or a very few, rival profit-maximizing firms producing a
homogeneous product face each other in a market, it is not only possible to
specify the equilibrium price and output levels, given demand and cost
conditions, but it is likely that the firms will adjust their prices and outputs to
these equilibrium levels, given normal specifications on their reaction func-
tions. However, if one introduces product differences, innovation or tech-
nological change into the picture, the stability of market equilibrium is
eroded, and it is even questionable whether an equilibrium exists. The
economic analogy implies that qualitative changes in the arms race, perhaps
reflected in technological breakthroughs, will tend to be destabilizing. Qual-
itative changes pose real threats to arms race stability by producing sudden
alterations in the rate of arms build-up, by increasing uncertainty about the
rival’s military capacity and by increasing the range of military choices. For
instance, “first strike” may suddenly modify a strategy of “massive retalia-
tion”. It is factors such as these which so disturb many knowledgeable
observers of the current arms race, and which probably account for the
worldwide increase in concern that the arms race is now entering a new and
more perilous phase. 8
There are other reasons to question Richardson’s model and to doubt
whether the arms race is headed toward stable equilibrium. 9 The Richardson
model is very mechanistic, ignoring learning behaviour, strategic considera-
tions, preferences for international power, and the use of weapons as a
51 In Richardson’s empirical work this assumption is reflected in the fact that he measured the
arms race in terms of defence expenditures. McGuire, supra, note 54, uses missiles and
wairheads as his common unit. Sorensen, ibid., shows that if differences between the accuracy
of American and Soviet missiles are taken into account (a qualitative factor), U.S. strategic
military strength exceeds that of the Soviets despite the fact that the U.S. has been outstripped
in missile numbers and nuclear megatonnage.
5 In response to the perceived strategic military power of the Soviet Union, the United States
is developing a number of new weapon systems such as the MX, Trident II and cruise missiles,
some of which are believed to give the U.S. a “first strike” capability. This poses the danger
that the Soviet Union, fearing such a first strike, would react by itself striking first. While these
conjectures are highly problematic, the mere thought should be enough to start a peace
movement.
59But see McGuire, supra, note 54, who not only found some modest empirical support for
the Richardson model, but who also concluded that the model did imply a stable equilibrium
albeit at substantially greater force levels than existed in the mid 1970s. However, Desai &
Blake, Modelling the Ultimate Absurdity: A Comment on “A Quantitative Study of the Strategic
Arms Race in the Missile Age” (1981) 63 Rev. Econ. and Stat. 619, show that McGuire’s data
and an alternative model specification only produce stability in the rate of growth of arma-
ments.
McGILL LAW JOURNAL
[Vol. 28
deterrent. 61 Moreover, the cost or fatigue factor may not be very important.
For example, Gillespie et al. tested empirically a modified form of the
Richardson model, 6’ and found evidence that, instead of viewing the arms
race in economically negative terms, the U.S.A. and the U.S.S.R. “perceive
each other to arm primarily for economic reasons”. Moreover, the Gillespie
study suggests that neither country seems concerned about “vertical” arms
proliferation. If true, we may infer that there is no chance that the “Prisoner’s
Dilemma” will be broken by an acute feeling of insecurity due to the rise in the
total quantity of armaments in existence. Gillespie et al concluded that “the
USA-USSR confrontation is one that lacks both stability and equilibrium and
is leading toward continued increments in arms”. However, on the brighter
side, when the U.S.A. and U.S.S.R. are included as members of their
respective alliances “the armaments race is both stable and equilibrium-
seeking”.62
Up to this point, we have focused on the stability of the arms race
between the superpowers and their allies – what was termed earlier “verti-
cal” proliferation. A gnawing problem is the “horizontal” proliferation of
nuclear weapons. Does the increase in the number of nations with nuclear
military capabilities increase the probability of nuclear war? One would think
so, by drawing again on an analogy from industrial behaviour. It is believed
widely by economists that the likelihood of price competition between firms
increases as the number of rival firms rises from 2 to 5 to 10 and so on. The
reason is that the larger the number of rivals, the more difficult it is for them to
to collude or
reach and to enforce an agreement among themselves –
cooperate tacitly or explicitly. It is, of course, unclear whether the collusion
analogy from economics is apt when, virtually by definition, arms race rivals
have no intention of colluding. In fact, if they did collude –
there would not be an arms race, certainly not one of such serious proportions
as that in which we are now enmeshed.
or cooperate –
In an interesting paper, Intriligator and Brito argue that, in certain
instances, proliferation may reduce rather than increase the probability of
60 See Intriligator & Brito, supra, note 53. Richardson, supra, note 51, 12, had an interesting
response however. After admitting that his “equations are merely a description of what people
would do if they did not stop to think”, he went on to say that nations “follow their traditions,
which are fixtures, and their instincts, which are mechanical … because they have not yet
made a sufficiently strenuous intellectual and moral effort to control the situation”. Although
Richardson’s view lacks the behavioural emphasis of modem social science, one wonders
whether he is not pretty close to the mark.
61 Modified to include an explicit goal or objective function: the achievement of a “balance of
power”. See Gillespie, supra, note 53.
21bid., 244.
1983]
ECONOMICS OF THE ARMS RACE
nuclear war.”3 Assuming that the new nuclear nations are responsible and that
they acquire nuclear capabilities sufficient to deter a strike by one of the
superpowers, then the probability of a deliberate initiation of war by one of
the first two powers is reduced. The reasoning is that the post-war distribution
of power is important to the initiating nation. ‘ With additional nuclear
nations, the initiator’s relative postwar position –
is
reduced. Its uncertainty about the reactions of the other nuclear nations is also
increased, the larger the number of nuclear nations. This is so whether or not,
in a conflict, the latter are belligerents or neutrals. But there is another side to
the coin. While nuclear proliferation may reduce the chances of a calculated
or “rational” attack, it also increases the probability 9f an accidental or
“irrational” war. And, as the number of nuclear nations increases,
the
accidental-irrational factor plays an increasingly important role.
if it “survives” –
According to Intriligator and Brito, the major policy implication flowing
from their analysis is that more attention should be paid to preventing
accidental or irrational war than to non-proliferation per se. But one could
argue the opposite, namely, that the best way to prevent accidents is to
prevent proliferation. Again, an analogy can be drawn from economics. The
best assurance against an outbreak of price competition in an industry with
two, or very few, rivals who are avoiding a price-competitive attack on each
other, is to promote high barriers to the entry of new firms into the industry.
Nevertheless, even blocked entry is not sufficient to prevent an “irrational”
outbreak of price competition between the established rivals. In addition to
non-proliferation – of firms or weapons –
cooperation seems necessary.0
However, another analogy from economics forces us to scrutinize the
side effects of cooperation carefully. The literature on cartels, and on indus-
tries whose price and output-entry conditions are publicly regulated, suggests
that limitation of price competition often stimulates non-price forms of
competition.’ Restricted by agreement or law from competing in one manner,
I Intriligator and Brito note that the only time nuclear weapons were used was when only one
nation had nuclear capabilities. See Nuclear Proliferation and the Probability of Nuclear War
(1981) 37(2) Public Choice 247. Now that there are two nuclear superpowers, the probability of
use is reduced because of a balance of nuclear terror caused by the ability to strike back.
‘ This reasoning assumes that there is something left to distribute after a war. An unfortunate
characteristic of much of the theoretical literature on nuclear war is its failure to integrate
explicitly what doctors and other scientists, knowledgeable about the effects of nuclear war,
have had to say about life – or what’s left of life- after a thermonuclear war. See Bates, The
Medical and Ecological Effects of Nuclear War (1983) 28 McGill L.J. 716 and Wolfe,
Chemical and Biological Warfare: Medical Effects and Consequences (1983) 28 McGill L.J.
732.
6’While as consumers we generally lose from cooperation (collusion) among firms, as
human beings we could gain from cooperation among the nuclear superpowers.
16A classic example was the United States airline industry prior to deregulation in 1978.
REVUE DE DROIT DE McGILL
[Vol. 28
rivals find other means –
to compete. This
analysis implies that even if participants in an arms race are prepared to agree
on some issues, it is important to consider the side effects.
and the necessary incentives –
Some economists have argued, for example, that the SALT I agreement
which produced limits on the quantities of certain strategic armaments and
reduced somewhat the level of military spending, may have accelerated the
qualitative arms race.67 Given the bureaucratic nature of the military estab-
lishments in both the U.S. and the U.S.S.R., it would not be surprising if
agreements to limit specific weapons resulted in a shift of effort and resources
toward weapons systems not covered by the agreement, for instance, cruise
missiles or increased accuracy of strategic nuclear weapons. Because the
qualitative arms race can be more destabilizing than a quantitative one, there
is a certain irony in the idea that the SALT I agreement, however well-
intentioned, may have reduced, rather than increased, long term security. In
any event, economic analysis suggests that regulating only one or some
dimensions of competition can be welfare- or security-reducing. Moreover,
defusing the arms race requires not only control over the “outputs” but
altering of the incentive structure of the “inputs”.
IV. Halting the Arms Race
The economic analysis employed in this paper suggests that there are
powerful forces propelling the arms race, but that the economic consequences
are not, on the whole, serious enough to do more than slow down somewhat
the growth of military spending. As I see it, an objective evaluation of the
economic issues suggests that if arguments for disarmament are to be persua-
sive, they will have to be built on something firmer than the economic effects
of the arms race, at least those short of war.6 However, economics does
67See, e.g., Brito & Intriligator, Strategic Arms Limitations Treaties and Innovations in
Weapons Technology (1981) 37(1) Public Choice 41; MacKay, Strategic Arms Limitation
Treaties and Innovations in Weapons Technology [:J A Comment (1981) 37(1) Public Choice
61.
,s Some readers will angrily, and perhaps justifiably, object that I have overlooked totally the
economic impact of the arms race on the less developed countries [LDCs]. It is true, of course,
that I have said almost nothing about the LDCs –
space would not permit. However, unless
one assumes that the resources that are currently being poured into the arms race would
otherwise be diverted to aid the “Third World”, it is not at all clear that the economic impact of
the arms race has been so disturbing to LDCs- unless of course it has embroiled them in war.
Incidentally, I suspect that the political consequences of the arms race are much more serious
for the LDCs because a large portion of the arms acquired are used to maintain internal “order”,
which usually means existing autocrats. It is interesting to note that E. Benoit, Defense and
Economic Growth in Developing Countries (1973) found –
a
unexpectedly, he claimed –
1983]
ECONOMICS OF THE ARMS RACE
provide some powerful tools for analyzing the arms race, its intransigence, its
lack of stability, and, as I will attempt to show in this final section of the
paper, what will be required ultimately to put an end to the race.
What will terminate the arms race between the superpowers? Trust?
Cooperation? I have already noted some problems with partial agreements.
But even a comprehensive arms agreement, covering all aspects, present and
future, of the arms race, although a highly desirable step in the right direction,
would not be sufficient to ensure enduring arms limitation and peace. Why
not? An economic analogy can explain why comprehensive agreements
between rivals are not sufficient, or perhaps even necessary, to ensure peace.
Again, we can call on the cartel literature for a simple explanation. Even the
most full-blown cartel agreement, covering all conceivable aspects of com-
petition between its members, can, and probably will, break down eventually
unless its provisions are enforceable in law. 9 Moreover, the likelihood of
breakdown increases with the number of members.
The implications of the cartel analogy for putting a lasting end to the
arms race are clear. Something more than agreement – or even trust which
can so easily turn to distrust –
is needed. Eventually, limits on armaments
and their use must be made legally enforceable. The legal imperative implies
some form of “constitutional” agreements establishing broad rules on the
development, deployment and use of arms. Enforceability implies a body – a
positive correlation between defence burden and growth in developing countries. However, the
work of at least one economist has called into question Benoit’s findings. See Palme, supra,
note 1, 21, fn. 15.
A quick scan of the data contained in Sivard, supra, note 7 seems to indicate a high rank
correlation between military spendingper capita and health and education spendingper capita,
in Third World countries. I suspect that as the real incomes of Third World countries rise, they
will continue to buy more of both security (armaments) and social welfare. Given that so many
of the LDCs are new nations whose borders often conflict with those delineated by ancient
tribal custom, and given the quite normal desire to increase protection of “property” as its value
rises, at least some of the demand by LDCs for armaments must be considered quite natural,
even if it is undesirable from the standpoint of world stability. Of course, these statements are
not intended to justify the transfer of arms to the Third World by the superpowers in an attempt
to win or keep allies, nor can they explain or justify the appalling arms race going on in the
Middle East.
6 Of course, because of the antimonopoly and antitrust traditions of most western nations,
they have not only refused to enforce cartel agreements, or other contracts in restraint of trade,
but have, in many cases prosecuted such arrangements as criminal conspiracies. But this, of
course, is beside the point. The real issue involves the behaviour of individual members of a
group. For a theory of group behaviour and an explanation of why, left to themselves,
self-interested individuals will not act to achieve common or group interests, see Olson, supra,
note 41.
McGILL LAW JOURNAL
[Vol. 28
to ensure compliance with those rules. 0 Reaching such a
“police force” –
“constitutional” agreement is, no doubt, a tall order. And, no doubt, some
readers will feel uncomfortable about any proposition that infringes on
national sovereignty. Let me attempt, then, to place these difficult issues in a
different light.
It is no coincidence, I believe, that military spending has grown enor-
mously during a period of unparalleled economic growth.7 As wealth in-
creases, there is a tendency to spend more to defend and secure that wealth. In
economic parlance, the income elasticity of military spending is greater than
zero, although not necessarily greater than one.7″ This statement applies to
individuals as well as to nations. However, in the case of individuals, certain
institutions have evolved which allow them to economize on the resources
required to protect themselves. These institutions include a police force and a
court system which not only prosecutes offenders but enforces contracts and
property rights. In fact, without a means of enforcing contracts and property
rights, and a degree of internal order, national income and welfare would be
much lower than it is today. In economic terms again, the roles played by the
police force and court system make them “public goods”. Once they are made
available to one person they are available to all. Take them away, and the
share of resources required to provide individual protection will rise
dramatically.73
Let us push the analogy with individuals a little further. Economists like
to think of individuals as “sovereign consumers”. Yet these same individuals
are expected to relinquish their “sovereignty” so far as appropriating property
and settling disputes is concerned –
at least those disputes for which mutual
agreement among the parties involved is not possible. The same reasoning, it
would seem, must apply to nations, if they are to economize on the resources
needed for security and are to minimize the probability that defensive or
protective weapons will be used in war. 74 It may not be easy to get national
7 Some economists, drawing on the “Coase theorem”, might argue that, in principle, the
rival parties should be able to reach through negotiations a mutually advantageous self-
enforcing arms control agreement; that no “police force” is needed. See Coase, The Problem of
Social Cost (1960) 3 J. Law and Econ. 1. However, in my view, the informational, transaction
cost and distributional problems which often make the “Coase theorem” difficult to apply in
practice, also plague its practical application to the arms race. Incidentally, a world police force
would provide at least one “market” for the “munitions makers” and their sidekicks, the
weapons technologists.
71 This is not to say that all of the growth in military spending is related to economic growth.
‘See supra, note 8.
73One only need think of feudal lords and their armies of vassals who in a period of national
disintegration provided virtually the only means of protection — if not order.
74The issue is even more clear when an individual or nation is hostile and aggressive.
1983]
ECONOMICS OF THE ARMS RACE
683
leaders to think as individuals, but it does seem that this will be necessary if a
future arms limitation or disarmament agreement is not to become a cruel
hoax on an unsuspecting humanity.
In concluding, it would be fair to say that in this paper more issues have
been raised than questions answered.. This result, however, is a common
malady in economics. Yet, economic “science” does have some useful
contribution to make to understanding the arms race. Its contribution may lie
less in providing economic arguments against the arms race than in offering
insights into why the race is run and in suggesting what institutional arrange-
ment are required to bring the arms race under control.
