The Impact of Defense Spending on GDP:

The Case of North America

by

Dale Bremmer
Professor of Economics
Department of Humanities and Social Sciences
Rose-Hulman Institute of Technology

and

Randy Kesselring
Professor of Economics
Department of Economics and Finance
Arkansas State University

April 2007

Presented During the “Topics in Economics” Session

at the 49th Annual Conference of the Western Social Science Association
Hyatt Regency Hotel, Calgary, Alberta, Canada
Friday, April 13, 2007, 8:00 a.m. – 9:30 a.m.
 The Impact of Defense Spending on Economic Growth: The Case of North America

I. Introduction

With the demise of the Soviet Union in the early 1990s, the fall of the Berlin Wall and

the end of the Cold War, some saw the need for decreased military spending by the United States

and they sought evidence of the peace dividend. But in 2007, six years after 9-11, the U.S. is

engaged in an ongoing five-year war in Iraq and a six-year conflict in Afghanistan. Given the

current strain on the deployment of the U.S. armed forces, concerns about the readiness of the

forces and the possible inability to engage in armed conflict on two separate fronts, some argue

that an increase in U.S. defense spending is needed. Debates about whether a military draft

should be reinstituted have begun. Given the increased fiscal pressure of defense spending on a

fragile federal government budget and a fragile economy, a more current analysis of the impact

of defense spending on GDP growth should be worthwhile.

This paper is a pilot study which examines whether increases in military spending have a

positive or negative impact on a country’s GDP. The study focuses on the three North American

economies: Canada, Mexico and the United States. While many others have written on this

issue, the approach used in this paper is somewhat unique in that a reduced-form “St. Louis” type

model is used to measure the impact of the central government’s economic policy tools on the

growth rate of GDP. Three policy tools are included in the model. Two different fiscal policy

tools are examined. First, the model includes central government expenditures (the standard

Keynesian policy tool). However, the model also includes a separate category of central

government expenditures—military spending. Finally, the role of monetary policy is captured by

including money supply changes in the model. The data set for each of the three countries

includes 43 annual observations between 1963 and 2005. The “St. Louis” model was estimated

Page 1
using current macroeconomic time-series techniques including VAR models, unit root tests, and

tests for cointegration. The impact of a change in policy is examined using impulse response

functions from the VAR models.

The results produced by the statistical models are asymmetric. Increased military

spending in Canada and Mexico leads to increases in GDP while increased military spending in

the United States decreases GDP. This result was confirmed with two different statistical tests:

pairwise Granger-causality tests and the behavior of impulse response functions generated by

each country’s VAR model. These results hint of a threshold effect where the impact of defense

spending on the economy depends on the size of the country’s military spending relative to total

central government expenditures and GDP. An increase in nonmilitary spending leads to rising

GDP in the United States and Mexico, but leads to falling GDP in Canada. The impulse

response functions indicate that innovations in the money supply are associated with rising GDP

in Canada and the U.S., but do not affect GDP in Mexico.

The rest of the paper is organized as follows. Following this introduction, the second

section of the paper provides a brief review of the past literature regarding the impact of defense

spending on a country’s economy. A description of the empirical model and the data is found in

the third section of the paper. This section also contains a discussion of the unit root tests and an

analysis of the pairwise Ganger-causality tests. The empirical results and the impulse response

functions are analyzed in the fourth section, while the final section of the paper offers a critique

of the results and some ideas for future research.

II. Literature Review

Previous papers on the impact of a nation’s defense spending on its economy have

concentrated on four areas. Some researchers have stressed the relationship between military

Page 2
spending and private investment. Others (especially development economists) have investigated

the effect that defense spending has on employment. Still others have investigated an economy’s

demand for defense expenditures. Finally, the most common approach has been to relate

military spending directly to the nation’s ability to produce output (GDP).1 The literature review

that follows has been divided into these four areas.

The impact of defense spending on private investment

Several studies have concluded that, in the case of developed nations, investment and

military expenditures are substitutes (Kennedy 1983, p. 198). Smith (1977, 1978, 1980, and

1983) argues that households are very resistant to cuts in private consumption and public

welfare. Given an exchange rate and a level of capacity utilization, the remainder of the nation’s

output is divided between defense spending and investment. The result is that a higher level of

military spending leads to a lower level of investment. In a study of 15 countries between 1960

and 1970, Smith found that investment spending and defense spending were negatively

correlated (-0.73). Lindgren (1984) points out that most studies are in agreement on the negative

correlation between defense spending and investment spending for industrialized, market

economies.

More recently, a study by Laopodis (2001) produced substantially different results.

Using Granger causality tests on a time-series of data (1960-1997) for Greece, Ireland, Portugal

and Spain no significant relationship (either positive or negative) was found between military

spending and gross private domestic investment. To further investigate this relationship,

Laopodis tested for cointegration of a number of variables relating to gross private domestic

1
For a recent study of the relationship between military expenditures and economic growth in
developing nations, see Chowdhury (1991).

Page 3
investment and estimated the associated error correction model. The results again indicated that

military spending had no significant effect on investment.

The impact of defense spending on employment

Regarding the impact of military expenditures on employment, the Marxist critique of the

1960s was that defense spending was a necessary, albeit wasteful, policy to stabilize and expand

capitalism (Baran and Sweezy 1968). They argued that by raising the level of defense spending

a country could solve the problems of under consumption and the unemployment associated with

it. Furthermore, capitalism would resort to such spending in an effort to reduce class conflict.

The hypothesis that increased military spending indirectly creates employment in the armaments

industry and directly creates more jobs in the armed forces does not necessarily need to rely on

the tenants of Marxist economics for support. Indeed, one of the primary concerns regarding

disarmament at the end of the Cold War was its hypothesized relationship to rising

unemployment.

In spite of this intuitive relationship, Chester (1978), Smith (1978), and deGrasse (1983)

could not find a statistically significant relationship between military expenditures and

unemployment. In his survey article of the literature, Lindgren concludes that “the relationship

between military expenditures and employment seems too complex to capture by correlation or

regression methods” (Lindgren, 1984, p.381). Recent studies confirm these earlier findings.

Dunne and Smith (1990) find no Granger causality between the share of defense spending and

the unemployment rate in nine of 11 countries in the OECD. Using data from 1962 to 1988, Paul

(1996) tested various economic hypotheses about the relationship between unemployment and

defense and non-defense spending in 18 OECD countries. However, Paul was unable to find a

uniform relationship between these variables across the various nations.

Page 4
 An interesting approach to this estimation problem was provided by Hooker and Knetter

(1997). Using a panel data model for the U.S, they found that defense procurement spending

which does vary from state to state had a significant impact on statewide employment. Both the

time series and the cross sections were accounted for with fixed effects. Additional evidence

was produced that the effects of defense spending on employment are decidedly nonlinear with

larger changes in defense spending produce significantly greater changes in employment.

Several other research efforts have been focused on specific countries. Wing (1991),

using input-output analysis, found that defense spending had a significant impact on employment

in Indonesia. Yildirim and Sezgin (2003) found that increases in defense spending tended to

affect employment negatively in Turkey. In a recent paper by Huang and Kao (2005), the

relationship between defense spending and employment in Taiwan was investigated. These two

used an unusual cointegration technique to demonstrate that defense spending damages

employment in the short run but benefits it in the long run.

The Demand for Defense Expenditures

Another interesting facet of research on military expenditures is the question of demand

determination. Early researchers in this area tended to estimate demand specifications for

defense spending that were closely related to standard household demand theory. Hartley and

Sandler (1990) provide a good review of these studies. The results seem to indicate that defense

spending is a normal good (positively related to GDP) and that, being a public good, there is

significant free riding.

Sandler and Murdoch (1991) add elements of game theory to the formulation of demand

models for defense expenditure. Murdoch et al (1991) test what is referred to as a median voter

demand model against an oligarchy model for members of NATO and get mixed results. Hanson

Page 5
et al. (1990), Hilton and Vu (1991) and Conybeare et al. (1994) all test empirical models of

defense demand concentrating on NATO countries. While all of these authors use slightly

different approaches they tend to get similar results regarding income or GDP (defense provision

is a normal good) and produce mixed results about free riding. They also tend to agree that

degree of external threat is an important determinant of demand.

A recent paper by Solomon (2005) uses a kind of cointegration technique (autoregressive

distributed lag approach) to investigate the demand for military spending in Canada. The results

indicated that the most important determinant of military spending in Canada was European

NATO spending. Interestingly, GDP was found to be insignificant.

The Impact of Defense Spending On GDP

Given the long-accepted, theoretical direct relationship between investment and

economic growth, if defense spending has a negative impact on investment, then it would seem

reasonable that defense spending would have an adverse impact on economic growth. This was

exactly the findings of two studies published in the seventies, Szymanski (1973) and Lee (1973).

Some studies attribute the negative effect of defense spending on economic growth to reduced

investment.2 Another study argues that defense spending restricts export growth and economic

growth because military expenditures compete for the same resources used in the production of

exports.3

However, other studies were unable to find any stable relationship between military

spending and economic growth.4 Chester (1978) found that military spending and economic

growth were positively related. A direct relationship between defense spending and economic
2
See Smith (1977 and 1978), Smith and Georgiou (1983) and Cappelen, et al (1984).
3
See Rotschild (1973).
4
See Nardinelli and Ackerman (1976), Faini et al. (1984), and de Grasse (1983).

Page 6
growth was found by Ahmed (1986) in a study of the UK. Weede (1983) found evidence that

supported his hypothesis that higher rates of participation in the armed services lead to more

economic growth. His argument was that service in the military leads to human capital

formation that is beneficial to economic growth. In his 1984 review essay that synthesizes past

articles analyzing of the impact of military spending on the economies of industrialized nations,

Lindgren (1984, p. 380) writes that the studies of the impact of defense spending on economic

growth are not as conclusive as those of investment. Nevertheless, the overwhelming conclusion

seems to be that military spending is not positively associated with economic growth but that

additional research is needed to clarify the issue. In terms of past modeling attempts, Lindgren

(1984, p. 376) notes that many studies used statistical techniques whose methods varied and

whose steps were not clearly described. These articles usually lacked the development of a

formal theoretical model on which to base econometric estimations. In examining the

relationship between defense spending and economic growth, Blackaby and Ohlson (1982, p.

291) noted that instead of “trying to provide a reasonable statistical structure” most of these past

attempts were “armchair theorists who conduct statistical exercises.”

Perhaps because of the above criticisms, models based on seemingly sounder theoretical

development have come into common use by researchers of this relationship. One of these

models is generally referred to as the Feder-Ram model because it was adapted by Biswas and

Ram (1986) from an export model developed by Feder (1983). Cuaresma and Reitschuler (2003)

used a version of this model for the U.S. economy. They found that at low levels of spending

(below an estimated threshold) defense spending has a positive impact on growth, but at high

levels the impact is either negative or insignificant. Yildirim et al. (2005) use a Feder-Ram

Page 7
derivation to find that military spending has a positive overall effect on the economies of the

Middle East.

A number of recent papers have attacked the problem through application of a type of

Granger causality estimation. Because they use Granger causality these papers generally limit

their analysis to examination of only two variables, defense spending and GDP. Dakurah et al.

(2001) examined 62 less developed countries (LDCs) with data from 1975 to 1995. A type of

two-way Granger causality was estimated for all of the countries. Only 23 countries exhibited

unidirectional causality, and in 16 of those 23 the relationship between defense spending and

GDP was positive. For the remaining 7, the relationship was negative. But, for most of the

countries, there was no relationship. Karagol and Palaz (2004) applied a similar technique to

Turkey for the years 1955 to 2000. However, they used impulse response functions to indicate

long-run causality. The result was that defense expenditures had a negative impact on GDP in

Turkey. Using a similar technique, Kollias et al. (2004) found that defense spending had a

positive impact on GDP in Cyprus. Their reasoning for this outcome was that the increase in

defense spending led to a greater feeling of security in Cyprus and, therefore, a better climate for

economic growth.

Atesoglu (2002) formulates a multivariable reduced-form Keynesian model patterned

after Romer (2000). Using cointegration estimation techniques for the years 1947 to 2000,

Atesoglu finds a positive long-run relationship between military spending and output for the U.S.

economy. Galvin (2003) also uses a multiequation, multivariable approach to investigate the

problem. Using a model similar to that of Deger (1986) three equations are postulated, one each

for growth, saving and defense expenditure. This model is somewhat unusual in that the data

used for estimation is strictly cross-sectional (64 developing countries). The author uses OLS,

Page 8
2SLS and 3SLS to derive overall estimates and separate estimates for middle and lower income

countries. The results indicate that military spending has a negative impact on growth for the

middle income countries but is insignificant for the low income countries. Klein (2004) also

used a Deger type model to estimate the effects of military spending in Peru. After some unit

root testing and difficulties with estimation results from the differenced system, adjustments

were made and the three equation model was estimated with OLS, 2SLS and 3SLS. The

resulting estimates reveal a negative relationship between defense spending and economic

growth in Peru. Finally, Dunne et al. (2005) provides a very good overview of the various

models used to estimate the relationship between defense spending and growth.

Obviously, the results from many different models estimated in several different ways

have been mixed. Consequently, the search is still on for a model and an estimation technique

that will provide consistent results. Our attempt begins by looking at the entire process in a

slightly different way. Economists are always concerned about policy making and, therefore,

policy variables. Governments have at their disposal several tools with which to affect the macro

economy. These tools are usually divided into two classes: monetary and fiscal policy tools.

Obviously, control over military spending is a “fiscal” policy tool. So, our approach is to use

military spending within a model that was originally set forth to distinguish the effectiveness of

these tools—the St. Louis model.

III. The Model, the Data, Unit Root Tests, and Granger-Causality Tests

To determine the relative impacts of defense spending, other central government

spending and monetary policy on a country’s output, a “St. Louis” or reduced-form expenditure

equal the annual percentage change in nominal GDP that was observed in
model is used. Let Yt

year t. To distinguish between the change in defense spending and other central government

Page 9

denotes the annual percentage change in nominal military expenditures in
expenditures, let D t

equals the percentage change in nominal nonmilitary expenditures in year t.

year t, while O t

is the annual
Finally, to control for the differences between fiscal and monetary policy, M t

percentage change in the nominal money supply in year t.

Specifying the VAR model

The vector autoregression (VAR) model is a seemingly unrelated system of four

regressions where each of the current dependent variables is a function of lagged dependent

variables. The VAR model below assumes that each regression has the same lag structure or
T T T T

=β +
Yt 0 ∑ βi Y
t - i +
i=1
∑ βT+i D
t - i +
i=1
∑ β 2T + i O
t - i +
i=1
∑β
i=1
3T + i


M t - i + ε1t (1)

T T T T

=α +

D t 0 ∑ α Y

i=1
i t-i
i=1

+
+ ∑ α T+i D t-i ∑α
i=1
2T + i O t - i + ∑α
i=1
3T + i


M t - i + ε 2t (2)

T T T T

=δ +

O t 0 ∑ δi Y
t - i +
i=1
∑δ
i=1


T+i D t - i + ∑δ
i=1
2T + i O t - i + ∑δ
i=1
3T + i


M t - i + ε 3t (3)

and
T T T T

=γ +
M t 0 ∑ γi Y
t - i + ∑ γT+i D
t - i +
i=1 i=1
∑ γ 2T + iO
t - i +
i=1
∑γ
i=1
3T + i


M t - i + ε 4t . (4)

On the right-hand side of each equation, T lags of each dependent variable are used as

explanatory variables. Also notice that the right-had sides of Equations (1)-(4) contain no other

exogenous explanatory variables.

The data

To initially test the impact of military spending on the growth rate of GDP, this paper

used a pilot case of the three North American countries: Canada, Mexico, and the United States.

This sample has two G7 countries, Canada and the United States, the only superpower. The

Page 10
results of these two developed countries can be compared with Mexico, the sample’s only

developing country. Annual data for all four variables were obtained for each of the three

countries for 43 years, from 1963 to 2005. The data can be found in Tables A1, A2, and A3 of

the appendix.5

Table 1 shows that the military expenditures in the United States, as a percentage of

central government expenditures, fell from a high of 49.15 percent in 1963 to a low of 15.70

percent in 1999. The U.S. wars in Afghanistan and Iraq increased this percentage to 19.84% in

2005. In the case of Canada, military expenditures as a percentage of central government

expenditures fell from 24.03 percent in 1963 to 5.86 percent in 2005. While this percentage

fluctuates over time, the general trend for Canada was for this percentage was to fall over the

data sample. Military spending as a percentage of central government expenditures was 10

percent in Mexico during 1965. While this percentage fell over time , it hit 1.5 percent in 1982,

increased to 4.7 percent in 1994, and was 2.41 percent in 2005.

Between 1963 and 2005, military expenditures as a percentage of GDP fell for each of

the three countries. As Table 1 shows, almost during the height of the Vietnam War, military

spending was 9.09 percent of GDP in the United States in 1967. After 9-11, military spending as

a percentage of GDP increased from 3.09 percent in 2001 to 4.07 percent in 2005. Military

spending in Mexico is always a smaller percentage of Mexican GDP. Military spending in

Mexico increased from 0.72 percent of GDP in 1963 to 1.18 percent of GDP in 1987 and then it

fell to 0.39 percent in 2005. Military spending as a percentage of GDP also fell in Canada. The

5
Data was obtained from various issues of World Military Expenditures and Arms Transfers, published by the U.S.
Arms Control and Disarmament Agency. This source has nominal and real military expenditures and central
government expenditures for the years between 1963 and 1999. Data on military expenditures between 2000 and
2005 was obtained from the military expenditure data base of the Stockholm International Peace Research Institute
(see www.sipri.org). GDP and money supply data was obtained from the International Monetary Fund as was data
on central government expenditures between 200 and 2005 (see www.imf.org).

Page 11
percentage fell from 3.56 percent in 1963. It flirted with being over two percent in the early

Seventies and the mid Eighties, but it was 1.12 percent in 2005.

The unit root tests

Regressions with time series data run the risk of obtaining spurious results. This occurs

when regressions estimated with one or more nonstationary data series result in statistically

significant results even though there may be no true underlying relationship. To avoid spurious

results, augmented Dickey-Fuller tests are performed on each data series to determine whether

the data exhibit unit roots and are nonstationary. The results of the augmented Dickey-Fuller

tests are reported in Table 2. The four data series are annual percentage changes in GDP,

military spending, other expenditures of the central government, and the money supply. Table 2

lists the results of the unit root tests on these series for Canada, Mexico and the United States

using both levels data and first differences of the data.

In the case of Canada, for each of the four variables, the augmented Dickey-Fuller tests

reject the null hypothesis of unit roots at either the one or five percent level. This result is not

surprising as the variables are percentage changes which are essentially functions of the first

differences. The Canadian series are stationary and the regressions can be safely estimated in

level form and cointegration test aren’t necessary.

According to the results in Table 2, the percentage change in Mexican GDP and the

percentage change in Mexican military spending are nonstationary in levels, but stationary in

first differences. However, in the case of the percentage change in other spending by the

Mexican central government, the null hypothesis of a zero root is rejected at the one percent

level while the null hypothesis that the percentage change in the Mexican money supply has a

zero root can be rejected at the five percent level. The unit root tests indicate that the first

Page 12
differences of the percentage changes in Mexican money supply and other central government

expenditures are stationary. Because statistical data indicate that level data of at least two of the

four series data exhibit nonstationarity, the Mexican data will have to be estimated in first

differences and cointegration test have to be performed.

Finally, level forms of the United States data appear to be stationary. The null hypothesis

of a zero root is rejected at the one percent level in three of the four series. However, the

percentage change in American military expenditures appears to be stationary only at the ten

percent level. As was the case with Canadian data, the United States data can be estimated in

levels form without checking for cointegration.

Pairwise Granger-causality tests

Pairwise Granger-causality test are reported in Table 3. These statistical tests investigate

whether changes in GDP causes changes in military spending or does changes in military

spending causes changes in GDP. For each of the three countries, these tests involved estimating

two simple linear regressions, one having the percentage change in GDP as the dependent

variable and the other having the percentage change in military spending as the dependent

variable. The explanatory variables of these regressions are lagged values of the percentage

changes in military spending and GDP. The number of lags was determined by picking that

model specification that minimized the Akaike Information Criterion (AIC).

In the case of Canada and Mexico, Granger-causality tests indicate unidirectional

causality in that changes in military spending affect changes in GDP but not the other way

around. The Canadian result was found at only the ten percent level while the Mexican result

was obtained at the one percent level.

Page 13
 However, the opposite result was found for the United States. In this case, Granger-

causality tests indicate that changes in GDP affect military spending but changes in military

spending do not affect GDP. The conclusion of unilateral causality from GDP to military

spending was found at the five percent level. Because a different result was found for the United

States other than Mexico or Canada, this suggest that military spending may have a threshold

effect in that the impact of military spending on a country’s economy may depend on the relative

size of that country’s defense spending.

IV. Estimation Results of the VAR Model and Impulse Response Functions

The VAR models are estimated for Canada, Mexico, and the United States. The

estimation results for Canada are reported in Table 4, the cointegration tests and the estimation

results for Mexico are reported in Tables 5 and 6, and Table 7 lists the estimation results for the

United States. The appropriate lag length for each country was determined using standard lag

exclusion tests.

Estimation of the Canadian, Mexican and United States VAR models

Tests indicate that only one lag was needed in the Canadian VAR model. The main

advantage of using the VAR framework is its a-theoretic, reduced-form framework that doesn’t

require careful specification of a structural model and its comparative advantage in forecasting.

Given the reduced-form structure of the model, interpreting the signs and the statistical

significance and the sign of an individual coefficient becomes more problematic. Since the

Canadian data was stationary, the results in Table 4 were obtained using levels data. The results

in Table 4 indicate that only one of the four regression models have a significant F-statistic.

Estimation results indicate that null hypothesis that all the slope coefficients are simultaneously

Page 14
zero is rejected at the five percent level for the regression model with the GDP growth rate as a

dependent variable.

Since the unit root tests indicated that Mexican data was nonstationary, cointegration

tests were performed and their results are reported in Table 5. Using the techniques developed

by Johansen, both the trace test and the maximum-eigenvalue test indicate the presence of at

most one cointegrating equation with an intercept present.6 Having determined the presence of

one cointegrating equation and its underlying structure, an error-corrected VAR model was

estimated and the results are listed in Table 6. All four of the regressions reported in Table 6

have significant F-tests implying that none of the models have slope coefficients that are

simultaneously zero.

The estimation results for the United States are reported in Table 7. Recall that the

United States data series were stationary; therefore, the estimation results in Table 7 were

obtained with data in level form. Three of the four regressions have statistically significant F-

tests. Only the regression where nonmilitary spending as the dependent variable fails to reject

the null hypothesis that all the slope coefficients are simultaneously equal to zero.

The impulse response functions

The impulse functions of the three VAR models are shown in Figures 1, 2, and 3. The

Canadian results are shown in Figure 1, the Mexican results are depicted in Figure 2 and Figure 3

shows the United States results. Each figure shows what will happen to the percentage change in

GDP given an innovation to either military spending, nonmilitary spending or the money supply.

The shock was always an initial, one standard deviation to the residual, orthogonalizing the

impulses by the Cholesky factor and adjusting for the degrees of freedom. Each figure shows

the accumulated response of the growth rate in GDP.

6
The specification of the cointegrating equation was determined by choosing that model that minimized the AIC.

Page 15
 The response of the growth rate of Canadian GDP to a shock in the other variables is

shown in Figure 1. The estimation results indicate that positive innovations to either the growth

rate in military spending and money supply lead to increases in the growth rate of GDP.

However, if the Canadian federal government increases the growth rate of other spending

Canadian GDP falls.

According to the information in Figure 2, positive innovations to government

expenditures, either military or nonmilitary, lead to increases in Mexican GDP. But, Mexican

monetary policy almost appears neutral as a positive innovation in the nominal money supply has

little long-run effect on the growth rate of Mexican GDP.

While Canadian and Mexican military spending stimulate GDP growth in their respective

countries, Figure 3 shows that an increase in the growth rate of military expenditures in the

United States leads to a fall in GDP. However, positive innovations to either nonmilitary U.S.

government spending or the U.S. money supply leads to increases in U.S. GDP. Like the

pairwise Granger-causality tests discussed in the previous section, the conflicting evidence from

the impulse response function may again show evidence of a threshold effect where the relatively

large military spending of the U.S. may have a different impact on the economy than a country

were military spending is a relatively smaller portion of the economy.

V. Conclusion

Both pairwise Granger-causality tests and impulse response functions from VAR models

show an asymmetric response of the growth rate in GDP of the North American countries to a

change in military spending. Increased military spending increases nominal GDP in Canada and

Mexico. But increased military spending in the United States reduces the growth in nominal

GDP. Evidence from Mexico and the United States indicates that nominal GDP would grow

Page 16
faster if the federal government increased the growth rate of nonmilitary spending. Increasing

the growth rate of nominal money supply lead to higher growth rates in nominal GDP in Canada

and the United States but had little long-run impact on Mexico GDP.

The results of the paper foreshadow a threshold effect in that military spending in the

United States had an opposite effect on GDP growth than in increases in military spending in

Canada and Mexico. Since the U.S. spends a relatively larger percentage of its government

expenditures and GDP on defense, the existence of a threshold effect will have to be examined

with a larger sample of countries.

The use of a “St. Louis” type model in this pilot study shows the promise of future work

with this methodology. The “St. Louis” type model avoids constructing a detailed, precise

structural model and the reduced-form model measures the impact of a country’s policy

decisions on the nation’s GDP.

Future work will extend this analysis to more countries. The countries will differ in

economic development and the size of military spending. Depending on data availability, a tax

variable should be added to control for another tool of fiscal policy. Data sets with quarterly data

will also be analyzed, but quarterly data on defense spending by developing countries is difficult

to obtain.

Economics is the “Dismal Science” and the forecast of the model in this paper is bleak.

Military spending in the United States will continue to rise with the fall of the Soviet Union and

the emergence of the U.S. as the only superpower with the military to police conflicts deemed

politically appropriate. The events of 9-11, the advent of the war on terror, and the ongoing

conflicts in Iraq and Afghanistan promise to put upward pressure on military spending. This

paper indicates that increased military spending occurs with a tradeoff of less growth in GDP.

Page 17
 References

Ahmed, S. “Temporary and Permanent Government Spending in an Open Economy,” Journal of

Monetary Economics, 17, 2, 1986, pp. 405-419.

Atesoglu, H. Sonmez, “Defense Spending Promotes Aggregate Output in the United States—
Evidence from Cointegration Analysis,” Defence and Peace Economics, 13, 1, 2002, pp. 55-
60.

Baran, Paul and Paul Sweezy. Monopoly Capital. Harmondsworth: Penquin, 1968.

Biswas, B and R. Ram, “Military Expenditure and Economic Growth in Less Developed
Countries: An Augmented Model and Further Evidence,” Economic Development and
Cultural Change, 34, 1986, pp. 361-372.

Blackaby, Frank and Thomas Ohlson. “Military Expenditure and Arms Trade: Problems of
Data,” Bulletin of Peace Proposals, 13, 4, 1982, pp. 291-308.

Bremmer, Dale and Randy Kesselring, “The Opportunity Cost of Super Power Status: The
Tradeoff Between Defense Spending and Economic Prosperity,” unpublished manuscript,
March 1998.

Cappelen, Adne, Nils Petter Gleditsch and Olav Bjerkholt. “Military Spending and Economic
Growth in the OECD Countries,” Journal of Peace Research, 21, 4, 1984, pp. 375 - 387.

Chester, Eric 1978. “Military Spending and Capitalist Stability,” Cambridge Journal of
Economics, 2, 3, 1978, pp. 293-298.

Chowdhury, Abdur. “A Causal Analysis of Defense Spending and Economic Growth,” Journal
of Conflict Resolution, 35, 1, 1991, pp. 80-97.

Conybeare, J.A.C., J.C. Murdoch and T. Sandler, “Alternative Collective-Goods Models of

Military Alliances—Theory and Empirics,” Economic Inquiry, 32, 4, 1994, pp. 525-542.

Cuaresma, Jesus Crespo and Gerhard Reitschuler, “A Non-Linear Defence-Growth Nexus?

Evidence from the U.S. Economy,” Defence and Peace Economics, 15, 1, 2003, pp. 71-82.

Dakurah, A. Henry, Stephen P. Davies and Rahan K. Sampath, “Defense Spending and
Economic Growth in Developing Countries: A Causality Analysis,” Journal of Policy
Modeling, 23, 2001, pp. 651-658.

De Grasse Jr., Robert W. Military Expansion Economic Decline: The Impact of Military
Spending on U.S. Economic Performance. New York: Council on Economic
Priorities/Sharpe, 1983.

Page 18
Deger, S., “Economic Development and Defence Expenditure,” Economic Development and
Cultural Change, 35, 1, 1986, pp. 179-196.

Dunne, P. and R. Smith. “Military Expenditure and Unemployment in the OECD,” Defense
Spending, 1, 1990, pp. 57-73.

Dunne, Paul J., Ron P. Smith and Dirk Willenbockel, “Models of Military Expenditure and
Growth: A Critical Review,” Defence and Peace Economics, 16, 6, 2005, pp. 449-461.

Faini, Riccardo, Patricia Annez, and Lance Taylor. “Defense Spending , Economic Structure
and Growth: Evidence Among Countries and Over Time,” Economic Development and
Cultural Change, 32, 3, 1984, pp. 487-498.

Feder, G., “On Exports and Economic Growth,” Journal of Development Economics, 12, 1983,
pp. 59-73.

Galvin, Hannah, “The Impact of Defence Spending on the Economic Growth of Developing
Countries: A Cross-Section Study,” Defence and Peace Economics, 14, 1, 2003, pp. 51-59.

Hanson, L., J.C. Murdoch and T. Sandler, “On Distinguishing the Behavior of Nuclear and Non-
nuclear Allies in NATO,” Defence Economics, 1, 1, 1990, pp. 37-55.

Hartley, K. and T. Sandler (eds.), The Economics of Defence Spending: An International Survey,
London: Routledge, 1990.

Hausman, J. "Specification Tests in Econometrics," Econometrica, 46, 1978, pp. 1251-1271.

Hilton, B., and A. Vu, “The McGuire Model and the Economics of the NATO Alliance,”
Defence Economics, 2, 1, 1991, pp. 105-121.

Hooker, Mark A. and Michael M. Knetter, “The Effects of Military Spending on Economic
Activity: Evidence from State Procurement Spending,” Journal of Money, Credit, and
Banking, 29, 3, 1997, pp. 400-421.

Huang, Jr-Tsung, and An-Pang Kao, “Does Defence Spending Matter to Employment in
Taiwan?” Defence and Peace Economics, 16, 2, 2005, pp. 101-115.

Judge, George G., R. Carter Hill, William E. Griffiths, Helmut Lütkepohl, and Tsoung-Chao
Lee. Introduction to the Theory and Practice of Econometrics. New York: John Wiley &
Sons, 1988.

Karagol, Erdal and Serap Palaz, “Does Defence Expenditure Deter Economic Growth in Turkey?
A Conintegration Analysis,” Defence and Peace Economics, 15, 3, 2004, pp. 289-298.

Kennedy, Gavin. Defense Economics. London: Duckworth, 1983.

Page 19
Klein, Thilo, “Military Expenditure and Economic Growth: Peru 1970-1996.

Kollias, Christos, Charis Naxakis and Leonidas Zarangas, “Defence Spending and Growth in
Cyprus: A Causal Analysis,” Defence and Peace Economics, 15, 3, 2004, pp. 299-307.

Laopodis, Nikiforos T., “Effects of Government Spending on Private Investment,” Applied

Economics, 33, 2001, pp. 1563-1577.

Lee, Jong Ryol. “Changing National Priorities of the United States,” Military Force and
American Society, Bruce M. Russett and Alfred Stepan, eds. New York: Harper and Row,
1973, pp. 61-105.

Lindgren, Göran. “Armaments and Economic Performance in Industrialized Market Economies,”

Journal of Peace Research, 21, 4, 1984, pp. 375-387.

Murdoch, J.C., T. Sandler and L. Hansen, “An Econometric Technique for Comparing Median
Voter and Oligarchy Choice Models of Collective Action—the Case of the NATO Alliance,”
Review of Economics and Statistics, 73, 4, 1991, pp. 624-631.

Nardinelli, Clark and Gary B. Ackerman. “Defense Expenditures and the Survival of American
Capitalism,” Armed Forces and Society, 3, 1, 1976, pp. 13-16.

Paul, Satya. “Defense Spending and Unemployment Rates: An Empirical Analysis for the
OECD,” Journal of Economic Studies, 23, 2, 1996, pp. 44-54.

Romer, David, “Keynesian Macroeconomics Without the LM Curve,” Journal of Economic

Perspectives, 14, 2, 2000, pp. 149-169.

Rotschild, Kurt W. “Military Expenditure, Exports, and Growth,” Kyklos, 26, 4, 1973, pp. 804-
813.

Sandler, T., and J.C. Murdoch, “Nash-Cournot or Lindahl Behavior? An Empirical Test for the
NATO Allies,” Quarterly Journal of Economics, 105, 4, 1991, pp. 875-894.

Smith, Ron P. “Military Expenditure and Capitalism,” Cambridge Journal of Economics, 1, 1,

1977, pp. 61-76.

Smith, Ron P. “Military Expenditure and Capitalism: A Reply,” Cambridge Journal of

Economics, 2, 3, 1978, pp. 299-304.

Smith, Ron P. “Military Expenditure and Investment in OECD Countries, 1954-1973,” Journal
of Comparative Economics, 4, 1980, pp. 19-32.

Smith Ron P. and George Georgiou. “Assessing the Effect of Military Expenditure on OECD
Economies: A Survey,” Arms Control, 14, 1, 1983, pp. 3-15.

Page 20
Solomon, Binyam, “The Demand for Canadian Defence Expenditures,” Defence and Peace
Economics, 16, 3, 2005, pp. 171-189.

Szymanski, Albert. “Military Spending and Economic Stagflation,” American Journal of

Sociology, 79, 1, 1973, pp. 1-14.

Weede, Erich 1983. “Military Participation rations, Human Capital Formation, and Economic
Growth: A Cross-national Analysis,” Journal of Political and Military Sociology, 11(Spring),
11-19.

Wing, M.M., “Defense Spending and Employment in Indonesia,” Defence Economics, 3, 1,

1991, pp 83-92.

Yildirim, J. and S. Sezgin, “A System Estimation of the Defence-Growth Relation in Turkey,” in

Arming the South: The Economics of Military Expenditure, Arms Production, and Arms
Trade in Developing Countries, J. Brauer and P.J. Dunne, eds., London: Palgrave, 2002.

Yildirim, J., Selami Sezgin and Nadir Ocal, “Military Expenditure and Economic Growth in
Middle Eastern Countries: A Dynamic Panel Data Analysis,” Defence and Peace Economics,
16, 4, 2005, pp. 283-295.

Page 21
Table 1: Military Expenditures as a Percent of Central Government Expenditures and GDP:
Canada, Mexico, and the United States, 1963 - 2005
Military Expenditures as a Percent Military Expenditures as a Percent
of Central Government Expenditures of GDP
Year Canada Mexico United States Year Canada Mexico United States
1963 24.03 9.86 49.15 1963 3.56 0.72 8.47
1964 22.90 9.64 46.22 1964 3.44 0.72 7.72
1965 22.60 10.00 44.05 1965 3.16 0.71 7.20
1966 22.11 9.93 46.87 1966 3.11 0.72 8.07
1967 18.70 8.10 49.40 1967 3.11 0.68 9.09
1968 16.30 8.10 45.20 1968 3.10 0.69 8.87
1969 11.60 6.60 44.10 1969 2.62 0.70 8.27
1970 12.90 6.80 39.80 1970 2.39 0.70 7.50
1971 11.60 7.60 35.20 1971 2.23 0.73 6.64
1972 10.50 4.90 33.40 1972 2.03 0.99 6.27
1973 9.70 4.40 30.10 1973 1.83 0.88 5.67
1974 8.50 4.30 30.30 1974 1.68 0.43 5.73
1975 7.70 4.40 26.20 1975 1.57 0.56 5.55
1976 8.20 3.90 23.60 1976 1.34 0.54 4.99
1977 8.70 3.90 23.80 1977 1.62 0.44 4.97
1978 8.70 2.80 23.00 1978 1.92 0.36 4.76
1979 8.20 2.70 23.30 1979 1.83 0.33 4.77
1980 8.50 2.10 23.10 1980 1.92 0.23 5.16
1981 7.80 2.30 23.60 1981 1.84 0.32 5.43
1982 8.10 1.50 25.00 1982 2.14 0.48 6.03
1983 7.90 2.00 25.50 1983 1.78 0.69 6.16
1984 8.20 2.70 26.40 1984 1.87 0.79 7.14
1985 8.60 2.60 25.70 1985 2.02 0.58 6.12
1986 9.20 2.30 27.10 1986 2.07 0.86 6.29
1987 9.30 2.30 27.20 1987 1.93 1.18 6.08
1988 9.30 2.10 26.20 1988 1.70 0.69 5.74
1989 7.10 2.50 25.50 1989 1.31 0.65 5.54
1990 7.00 2.60 23.50 1990 1.33 0.54 5.27
1991 6.30 4.30 19.60 1991 1.26 0.52 4.67
1992 6.20 4.60 21.10 1992 1.33 0.51 4.81
1993 6.90 4.00 19.90 1993 1.53 0.52 4.48
1994 6.70 4.70 18.80 1994 1.51 0.63 4.07
1995 6.40 3.90 17.40 1995 1.38 0.81 3.77
1996 6.00 3.60 16.50 1996 1.23 0.67 3.47
1997 5.80 3.60 16.30 1997 1.12 0.62 3.32
1998 5.90 3.80 15.80 1998 1.28 0.60 3.13
1999 5.90 3.74 15.70 1999 1.26 0.56 3.03
2000 5.77 3.34 16.18 2000 1.15 0.52 3.07
2001 6.10 3.49 15.88 2001 1.17 0.54 3.09
2002 6.09 3.09 16.98 2002 1.15 0.50 3.41
2003 6.04 2.84 18.44 2003 1.15 0.46 3.79
2004 5.93 2.55 19.50 2004 1.14 0.41 3.97
2005 5.86 2.41 19.84 2005 1.12 0.39 4.07

Page 22
 Table 2
Augmented Dickey-Fuller (ADF) Unit Root Test
Regression Assumptions Statistic Critical Values
*
Data Type Data Lags Constant Trend τ τ99% τ*95% τ*90%
Canada
Growth Rate in Nominal GDP Levels ′65-′05 0 Yes Yes 4.18** 4.20 3.52 3.19
Growth Rate in Military Spending Levels ′65-′05 0 Yes No 4.98* 3.60 2.94 2.61
Growth Rate in Non-Military Spending Levels ′66-′05 1 Yes Yes 5.60* 4.21 3.53 3.19
Growth Rate in Money Supply Levels ′65-′05 0 Yes No 5.60* 3.60 2.94 2.61

Mexico
Growth Rate in Nominal GDP Levels ′67-′05 2 No No 0.98 2.63 1.95 1.61
st
Growth Rate in Nominal GDP 1 Differences ′67-′05 1 No No 6.89* 2.63 1.95 1.61
Growth Rate in Military Spending Levels ′67-′05 2 No No 1.58 2.63 1.95 1.61
Growth Rate in Military Spending 1st Differences ′65-′05 1 No No 7.88* 2.63 1.95 1.61
Growth Rate in Non-Military Spending Levels ′66-′05 1 Yes No 4.48* 4.20 3.52 3.19
st
Growth Rate in Non-Military Spending 1 Differences ′68-′05 2 Yes No 6.17* 2.63 1.95 1.61
Growth Rate in Money Supply Levels ′65-′05 0 Yes No 3.42** 3.60 2.94 2.61
st
Growth Rate in Money Supply 1 Differences ′65-′05 1 No No 7.32* 2.63 1.95 1.61

United States
Growth Rate in Nominal GDP Levels ′65-′05 0 Yes Yes 4.48* 4.20 3.52 3.19
Growth Rate in Military Spending Levels ′66-′05 1 Yes No 2.89*** 3.61 2.94 2.61
Growth Rate in Non-Military Spending Levels ′65-′05 0 Yes Yes 6.92* 4.20 3.52 3.19
Growth Rate in Money Supply Levels ′65-′05 0 Yes No 4.09* 3.60 2.94 2.61

*, **, and *** indicate the null hypothesis of a unit root is rejected at 1%, 5% and 10% level, respectively. The lag lengths in the ADF regressions were
determined by choosing the model with the largest AIC.

Page 23
 Table 3
Granger Causality Tests: Growth Rates in Nominal Military Spending and GDP
F-Statistic
Canada Mexico United States
(1966-2005) (1966-2005) (1967-2005)
Null Hypothesis 2 lags 2 lags 3 lags
Growth Rate in Military Spending Does Not
2.857*** 12.574* 0.297
Granger Cause Growth Rate in Nominal GDP
Growth Rate in Nominal GDP Does Not
0.020 0.417 3.310**
Granger Cause Growth Rate in Military Spending

*, **, and *** indicate the null hypothesis of no Granger causality is rejected at 1%, 5% and 10% level, respectively.
The lag length used in each VAR was determined by choosing the model with the largest AIC.

Table 4
VAR Model for Canada: Annual Data, 1965 – 2005
Dependent Variables
GDP Military Non-Military Money
Growth Spending Spending Supply
Independent Variables Rate Growth Rate Growth Rate Growth Rate
***
Intercept 11.264 6.227 14.776* 5.711***
(5.955) (4.235) (4.436) (3.505)
*
GDP Growth Rate-1 0.703 -0.025 -0.150 0.142
(0.122) (0.087) (0.091) (0.072)

Military Spending Growth Rate-1 0.576** 0.270 0.262 0.106

(0.265) (0.188) (0.197) (0.156)
Non-Military Spending Growth Rate-1 -0.383 -0.092 0.012 -0.078
(0.247) (0.176) (0.184) (0.146)

Money Supply Growth Rate-1 0.004 -0.027 -0.179 0.063

(0.269) (0.192) (0.201) (0.159)

R-squared 0.570 0.058 0.159 0.142

F-statistic 11.916† 0.551 1.704 1.485

Standard errors are in parentheses. *, **, and *** indicate the null hypothesis that the coefficient is equal to zero is
rejected at the 1, 5, or 10 percent level, respectively, using a 1-tail t test. † indicates the null hypothesis that all the
slope coefficients are simultaneous equal to zero is rejected using an F test at the 5 percent level.

Table 5
Johansen Cointegration Test - - Mexico
Null Hypothesis Eigenvalue Trace Test Max-Eigenvalue Test Lags
H0: At Most 1 Cointegtrating Equation 0.376 39.300* 18.313* 1
*
indicates one cointegrating vector at the 1% level.

Page 24
 Table 6
Vector Error Correction Model for Mexico: Annual Data, 1968 – 2005
Estimates for the One Cointegrating Equation
GDP Growth Military Spending Non-Military Spending Money Supply
Rate-1 Growth Rate-1 Growth Rate-1 Growth Rate-1 Intercept
** * *
1.000 -0.238 0.447 -1.338 -0.566
(0.118) (0.110) (0.089) (2.470)

Error-Corrected VAR Models

Dependent Variables
(First Differences)
GDP Military Non-Military Money
Independent Variables Growth Spending Spending Supply
(First Differences) Rate Growth Rate Growth Rate Growth Rate
The Cointegrating Equation 0.076 0.336 -0.640 1.271*
(0.209) (0.614) (0.800) (0.162)

GDP Growth Rate-1 -0.227 0.280 1.247 -0.812*

(0.314) (0.920) (1.199) (0.242)
***
GDP Growth Rate-2 -0.051 0.659 2.232 -0.734*
(0.298) (0.875) (1.140) (0.230)

GDP Growth Rate-3 -0.057 -0.977 -0.194 -0.762*

(0.264) (0.774) (1.009) (0.204)
Military Spending Growth Rate-1 0.076 -0.759** 0.188 0.276*
(0.118) (0.345) (0.450) (0.091)
**
Military Spending Growth Rate-2 -0.211 -0.903 -0.486 -0.010
(0.148) (0.434) (0.565) (0.114)

Military Spending Growth Rate-3 -0.113 -0.137 0.101 -0.138

(0.122) (0.357) (0.465) (0.094)

Non-Military Spending Growth Rate-1 0.126 0.333 -0.304 -0.405*

(0.097) (0.285) (0.372) (0.075)

Non-Military Spending Growth Rate-2 0.108 0.031 -0.539 -0.253*

(0.108) (0.318) (0.414) (0.084)

Non-Military Spending Growth Rate-3 0.070 -0.102 -0.451 -0.090

(0.078) (0.228) (0.297) (0.060)
***
Money Supply Growth Rate-1 0.095 -0.318 -1.296 0.212
(0.180) (0.526) (0.685) (0.138)

Money Supply Growth Rate-2 -0.008 -0.209 -0.954 0.198

(0.167) (0.491) (0.640) (0.129)

Money Supply Growth Rate-3 -0.032 -0.305 -1.049*** 0.148

(0.139) (0.407) (0.530) (0.107)

R2 0.622 0.507 0.553 0.863

F-Statistic 3.425† 2.140† 2.572† 13.135†

Standard errors are in parentheses. *, **, and *** indicate the null hypothesis that the coefficient is equal to zero is
rejected at the 1, 5, or 10 percent level, respectively, using a 1-tail t test. † indicates the null hypothesis that all the
slope coefficients are simultaneous equal to zero is rejected using an F test at the 5 percent level

Page 25
 Table 7
VAR Model for the United States: Annual Data, 1967 – 2005

Dependent Variables
GDP Military Non-Military Money
Growth Spending Spending Supply
Independent Variables Rate Growth Rate Growth Rate Growth Rate
Intercept 2.1066 -4.3115 -1.6134 0.6017
(1.2960) (3.9598) (4.1165) (2.3595)
GDP Growth Rate-1 0.2321 0.5636 -0.1963 -0.4158
(0.2132) (0.6513) (0.6771) (0.3881)
GDP Growth Rate-2 0.0098 -0.9731 -0.0636 -0.0069
(0.2187) (0.6681) (0.6946) (0.3981)
* **
GDP Growth Rate-3 0.1877 2.2460 1.3036 0.2156
(0.1901) (0.5807) (0.6037) (0.3460)

Military Spending Growth Rate-1 -0.0003 0.3065*** 0.1225 0.0934

(0.0494) (0.1509) (0.1569) (0.0899)
Military Spending Growth Rate-2 0.0203 0.2469 0.0243 0.0760
(0.0494) (0.1509) (0.1569) (0.0899)

Military Spending Growth Rate-3 -0.0670 -0.1129 -0.1781 -0.1165

(0.0507) (0.1550) (0.1611) (0.0923)

Non-Military Spending Growth Rate-1 0.0725 -0.4641** -0.0721 0.0585

(0.0618) (0.1887) (0.1962) (0.1125)

Non-Military Spending Growth Rate-2 0.0165 0.1397 0.2587 0.4182*

(0.0657) (0.2009) (0.2088) (0.1197)

Non-Military Spending Growth Rate-3 0.1092 -0.0053 0.2503 -0.0539

(0.0771) (0.2354) (0.2448) (0.1403)

Money Supply Growth Rate-1 0.0504 -0.1557 -0.3919 0.3690***

(0.1097) (0.3352) (0.3485) (0.1997)

Money Supply Growth Rate-2 0.0405 -0.2768 -0.0484 0.0294

(0.0930) (0.2843) (0.2955) (0.1694)
Money Supply Growth Rate-3 -0.0132 -0.2232 0.2174 -0.0547
(0.0896) (0.2739) (0.2847) (0.1632)

R2 0.5480 0.5564 0.3883 0.5314

F-statistic 2.6266† 2.7172† 1.3754 2.4571†

Standard errors are in parentheses. *, **, and *** indicate the null hypothesis that the coefficient is equal to zero is
rejected at the 1, 5, or 10 percent level, respectively, using a 1-tail t test. † indicates the null hypothesis that all the
slope coefficients are simultaneous equal to zero is rejected using an F test at the 5 percent level.

Page 26
Figure 1: Canadian Impulse Response Functions

Page 27
Figure 2: Mexican Impulse Response Function

Page 28
Figure 3: United States Impulse Response Function

Page 29
 Appendix
Table A1: Canada Data
(in billions of Canadian dollars)
Military NonMilitary Money Military NonMilitary Money
Year Spending Spending Supply GDP Year Spending Spending Supply GDP
1963 1.708 7.110 7.392 47.961 1985 9.789 113.827 76.380 485.714
1964 1.809 7.900 7.872 52.549 1986 10.610 115.329 87.284 512.541
1965 1.833 8.110 8.395 57.930 1987 10.780 115.917 92.557 558.949
1966 2.015 9.110 8.933 64.818 1988 10.424 112.087 98.433 613.094
1967 2.168 11.595 9.497 69.698 1989 8.584 120.900 104.969 657.728
1968 2.361 14.483 10.632 76.131 1990 9.042 129.178 106.000 679.921
1969 2.197 18.936 10.852 83.825 1991 8.650 137.305 110.424 685.367
1970 2.158 16.727 11.096 90.179 1992 9.295 149.922 117.408 700.480
1971 2.198 18.951 12.862 98.429 1993 11.146 161.539 124.448 727.184
1972 2.235 21.288 14.760 109.913 1994 11.649 173.864 131.553 770.873
1973 2.356 24.291 16.491 128.956 1995 11.185 174.772 143.629 810.426
1974 2.582 30.377 17.655 154.038 1996 10.294 171.570 160.836 836.864
1975 2.732 35.482 21.572 173.621 1997 9.900 170.690 174.282 882.733
1976 2.678 32.659 21.455 199.994 1998 11.734 198.884 183.995 914.973
1977 3.586 41.218 23.901 220.973 1999 12.361 209.513 203.488 982.441
1978 4.708 54.110 25.953 244.877 2000 12.326 213.480 221.319 1075.570
1979 5.127 62.528 26.842 279.577 2001 12.972 212.610 326.757 1107.460
1980 6.052 71.199 30.683 314.390 2002 13.332 218.895 347.829 1154.950
1981 6.630 85.000 37.669 360.471 2003 13.952 231.003 365.912 1214.600
1982 8.115 100.191 41.275 379.859 2004 14.749 248.534 395.706 1290.180
1983 7.313 92.572 46.411 411.386 2005 15.379 262.369 416.351 1368.730
1984 8.399 102.421 56.727 449.582

Page 30
 Table A2: Mexico Data
(in billions of nominal pesos)
Military NonMilitary Money Military NonMilitary Money
Year Spending Spending Supply GDP Year Spending Spending Supply GDP
1963 0.0014 0.0142 0.0235 0.1948 1985 0.2731 10.5021 3.4270 47.1675
1964 0.0016 0.0166 0.0275 0.2214 1986 0.6748 29.3385 5.6220 78.7870
1965 0.0018 0.0180 0.0291 0.2520 1987 2.2878 99.4686 12.4360 193.1620
1966 0.0020 0.0204 0.0323 0.2828 1988 2.8868 137.4690 20.5860 416.3050
1967 0.0021 0.0258 0.0348 0.3063 1989 3.5445 141.7807 28.6570 548.8580
1968 0.0024 0.0290 0.0404 0.3391 1990 4.0220 154.6930 46.9730 738.8980
1969 0.0026 0.0400 0.0459 0.3749 1991 4.9200 114.4196 106.3730 949.1480
1970 0.0031 0.0458 0.0506 0.4443 1992 5.7565 125.1416 122.0800 1125.3300
1971 0.0036 0.0469 0.0545 0.4901 1993 6.5116 162.7911 143.9000 1256.2000
1972 0.0056 0.1143 0.0648 0.5647 1994 8.8766 188.8631 145.1150 1420.1600
1973 0.0061 0.1386 0.0800 0.6909 1995 14.8931 381.8732 148.5340 1837.0200
1974 0.0039 0.0910 0.0992 0.8997 1996 17.0228 472.8547 206.0700 2525.5800
1975 0.0061 0.1395 0.1211 1.1001 1997 19.7962 549.8931 267.2110 3174.2800
1976 0.0074 0.1899 0.1571 1.3710 1998 22.9315 603.4595 323.9120 3846.3500
1977 0.0081 0.2072 0.1984 1.8493 1999 25.8131 689.9210 407.6010 4594.7200
1978 0.0084 0.3017 0.2629 2.3374 2000 28.3345 848.5020 465.4850 5491.7100
1979 0.0103 0.3801 0.3539 3.0675 2001 31.2980 897.0490 526.3460 5809.6900
1980 0.0105 0.4984 0.4633 4.4700 2002 31.2239 1009.7300 600.8390 6263.1400
1981 0.0198 0.8623 0.6140 6.1368 2003 31.7300 1118.0300 684.6900 6891.9900
1982 0.0466 3.1059 0.9810 9.7695 2004 31.8210 1248.1100 743.2210 7709.1000
1983 0.1230 6.1488 1.3850 17.8823 2005 32.6380 1356.4200 865.8910 8374.3500
1984 0.2326 8.6152 2.2580 29.4020

Page 31
 Table A3: United States Data
(in billions of nominal dollars)
Military NonMilitary Money Military NonMilitary Money
Year Spending Spending Supply GDP Year Spending Spending Supply GDP
1963 52.30 106.40 168.740 617.750 1985 258.20 1004.67 691.352 4220.250
1964 51.21 110.80 177.539 663.625 1986 280.90 1036.53 814.087 4462.820
1965 51.80 117.60 184.447 719.125 1987 288.00 1058.82 815.898 4739.470
1966 63.60 135.70 187.931 787.800 1988 293.00 1118.32 855.445 5103.750
1967 75.70 153.24 202.161 832.575 1989 304.00 1192.16 863.436 5484.350
1968 80.73 178.61 219.015 909.950 1990 306.00 1302.13 907.119 5803.070
1969 81.44 184.68 226.926 984.600 1991 280.00 1428.57 985.804 5995.920
1970 77.85 195.61 241.016 1038.520 1992 305.00 1445.50 1104.030 6337.750
1971 74.86 212.68 257.462 1127.100 1993 298.00 1497.49 1210.250 6657.400
1972 77.64 232.45 280.832 1238.770 1994 288.00 1531.91 1206.610 7072.230
1973 78.36 260.33 297.690 1382.730 1995 279.00 1603.45 1179.210 7397.650
1974 85.91 283.53 302.956 1499.980 1996 271.00 1642.42 1172.300 7816.820
1975 90.95 347.14 323.747 1638.330 1997 276.00 1693.25 1178.600 8304.330
1976 91.01 385.64 335.639 1825.280 1998 274.00 1734.18 1187.170 8746.980
1977 100.90 423.95 369.722 2030.920 1999 281.00 1789.81 1281.510 9268.430
1978 109.20 474.78 408.052 2294.700 2000 301.70 1864.40 1211.510 9816.970
1979 122.30 524.89 446.184 2563.300 2001 312.74 1969.50 1204.600 10127.900
1980 144.00 623.38 464.535 2789.520 2002 356.72 2101.10 1215.960 10469.600
1981 169.90 719.92 484.822 3128.430 2003 415.22 2252.10 1283.750 10960.700
1982 196.40 785.60 522.694 3255.020 2004 464.68 2383.00 1356.000 11712.500
1983 218.00 854.90 562.481 3536.670 2005 507.09 2555.90 1343.800 12455.800
1984 280.90 1064.02 609.688 3933.170

Page 32