View original document

The full text on this page is automatically extracted from the file linked above and may contain errors and inconsistencies.

February 15, 1991

Federal Reserve Bank of Cleveland

The Effect of War
Expenditures on U.S. Output
by Charles Carlstrom,
Jagadeesh Gokhale, and
Sharon Parrott

A he war with Iraq has become the
most extensive U.S. military involvement
since Vietnam. Although the duration and
the political ramifications of the operation
are still unclear, it is likely that the United
States will bear a substantial share of the
war's financial cost, placing an additional
burden on the federal budget. Aside from
the obvious concerns about war's devastation and the tragic cost in human lives,
policymakers have also expressed concern about the impact that the war will
have on the U.S. economy.
Many analysts believe that the economy
will be adversely affected because the
cost of the war will reduce consumption
either by raising taxes or by increasing
the federal budget deficit. Some cite the
oil price shock that followed Iraq's invasion of Kuwait and the subsequent drop
in consumer spending as additional factors working against the economy. However, others insist that the stimulative
effect of increased military expenditures
could help to pull the economy out of
its current downturn, noting that wars
have typically been associated with
high output.
This Economic Commentary examines
the effects of temporary war-related increases in government spending on output and real interest rates in the United
States. The model that we use differs
significantly in its approach and structure from more frequently used largescale econometric models. Our simulaISSN 0428-1276

tions indicate that the war with Iraq will
have only a minimal impact on the
• Costs of the Current
Military Buildup
In assessing how the cost of Operation
Desert Storm will affect the U.S. economy, it is necessary to separate the
additional cost of war from the cost of
ongoing military operations that the
United States would have incurred even
if troops had not been called to the Persian Gulf. Because the length of the
war is uncertain and the financial support of our allies is unclear, current cost
estimates are tenuous at best.
The cost of the war can be broken down
into two components: baseline costs and
incremental costs. Baseline costs are mainly those that would be incurred regardless
of where our troops are stationed. For the
U.S. military units deployed under Operation Desert Storm, this amount has been
estimated at $ 105 billion for 1991.' Because these costs would have been realized even in the absence of the Persian
Gulf crisis, they are not fully attributable
to Operation Desert Storm and therefore
are not included in our estimates of the
total cost of the war. However, given the
political developments in Eastern Europe and the pressures to reduce the federal budget deficit, defense allocations

Will Persian Gulf War expenditures
stimulate the U.S. economy? This
Economic Commentary looks at how
war-related temporary increases in
government spending affect real interest rates and output. The authors
find that even under the most pessimistic scenario, the Persian Gulf
War's impact on the economy should
be minimal.

for fiscal year 1991 might have been cut
in the absence of the crisis.
Under Operation Desert Shield (the first
phase of the military buildup), incremental costs included those associated
with transporting troops and equipment,
maintaining equipment in the desert environment, replacing spare parts, and
providing additional training for combat
under foreign conditions. The Comptroller General of the United States
recently placed the incremental cost of
maintaining 450,000 troops in a combatready position at $34 billion per year.
Now that hostilities have started, the incremental costs related to actual combat
are undoubtedly higher, but also more
difficult to estimate. Additional costs
stem from increased medical and logistical support services and from the replacement of destroyed ammunition and
equipment. The incremental costs of

active war could be six times the cost of
simply maintaining combat-ready troops.
This figure is consistent with daily cost
estimates of between $500 million and
$ 1 billion provided in Congressional testimony. If the war is over within a
year, the cost of replacing equipment can
be extended well into the future. However, if the war lasts significantly longer,
equipment will have to be replaced
much sooner and at a higher unit cost.
Assuming that Operation Desert Storm
lasts one year, we estimate a total cost
of $238 billion, which includes $34 billion of incremental costs for troop maintenance plus six times that amount for
combat-related costs. These estimates
do not include the cost to the United
States of maintaining a military force in
the region for an indefinite period once
a cease-fire occurs. Should this be
necessary, we assume that maintaining
a military presence for 10 years will
cost about $12 billion per year.
For simulation purposes, we assume that
half the incremental cost of the Persian
Gulf War ($119 billion) will be spent in
the current year, which translates into a
9.8 percent increase in total government
expenditures in the first year. The
other half, which primarily represents
the cost of replacing destroyed equipment, is assumed to be amortized over
the subsequent 10 years at an annual rate
of $14.6 billion. This amount, combined
with the additional cost of maintaining
a military presence once the war ends,
yields a 2.5 percent annual increase in
government spending for the next 10
years. We acknowledge that these figures are likely to be overestimates of
the war's actual cost, because 1) we
have not taken into account foreign contributions, 2) we assume that all equipment lost will be replaced, and 3) we
believe that the duration of the war will
be considerably less than one year.
To compare the costs of the current conflict with those of past wars, it is useful
to examine U.S. government and military
expenditures since 1941 (see figure 1).
One study estimates that temporary increases in government spending during
the four years of the Korean War (1951-

Billions of 1989 dollars

1941 1946 1951 1956 1961 1966 1971 1976 1981 1986
NOTE: Shaded areas represent war years in which government expenditures temporarily increased.
SOURCES: Economic Report of the President, February 1990, and Data Resources, Inc.










NOTE: Temporary changes in military expenditures are calculated as a percentage of government expenditures.
SOURCES: Authors' calculations and Robert J. Barro, "Output Effects of Government Purchases" (footnote 4).

1954) were 12.9, 12.7, 6.8, and 4.6 percent of total government expenditures,
respectively. During the three middle
years of the Vietnam War (1967-1969),
the respective increases were 5.1, 6.4,
and 3.5 percent. As seen in figure 1,
these increases are several orders of
magnitude smaller than those associated
with World War II.
• War and Output
Historically, temporary increases in government expenditures (most often precipitated by wars) have been accompanied by higher output. As figure 2
shows, World War II provides the most

graphic example of this phenomenon.
In this section, we explain why temporary increases in government spending should lead to higher output in the
short run.
A war-induced temporary increase in
government expenditures can be
financed either by a simultaneous increase in taxes or by an expansion of
the federal budget deficit. Both methods
of financing should result in higher interest rates and output. However, the
magnitude of these increases may differ.
First, suppose that the temporary increase in government spending is

Change in:
(percentage points)

Korean (1951-1954)
Vietnam (1967-1969)
Persian Gulf

First Year

Long Run



Real interest rate
(basis points)
First Year

Long Run


a. The real interest rate equals the model's predicted, marginal product of capital and should not be viewed as
a riskless rate. Its baseline value for the no-war scenario is 7.1 percent.
SOURCE: Simulations based on Auerbach and Kotlikofrs overlapping generations model (footnote 8).

financed by a concurrent increase in
taxes.5 Higher taxes during war years
(as compared to non-war years) lower
current after-tax income. Because individuals desire to smooth consumption
over time, they opt to increase borrowing in order to finance current consumption. As borrowing increases relative to
saving, the real interest rate rises, causing individuals to increasingly smooth
consumption by working longer hours.
It is this increase in hours worked that
causes output to rise. One study shows
that a temporary increase in government
expenditures equal to 1 percent of GNP
yields a 0.6 percent temporary increase
in output. As these numbers suggest,
when government expenditures rise, private consumption and investment fall.
So far our analysis has assumed that
wars are financed by higher taxes.
However, war efforts are typically
financed by higher government
deficits. Whether the use of deficit
rather than tax finance will have different effects on interest rates and output depends on the response of private
saving to higher government deficits.
Individuals tend to compensate for the
higher future taxes that will be needed
to pay for war-induced deficits by increasing their saving rate. However, the
extent of this increase is unclear. One
theory, known as the Ricardian Equivalence Theorem (RET), holds that the increase in personal saving will equal the
increase in government borrowing. If
this is true, the effect of a temporary increase in government spending on output and real interest rates will be iden-

tical under either arrangement. If RET
does not hold, however, the increase in
saving will fall short of the increase in
government borrowing. Consequently,
the effect of a temporary upturn in government expenditures on interest rates
and output would be somewhat larger
under deficit financing than under tax
Interest rates are also affected by international investment. The cost of the
Gulf War expected to be borne by many
U.S. trading partners is small relative
to the size of their economies. Therefore, these countries will not experience
the same upward pressure on interest
rates faced by the United States. Higher
U.S. interest rates relative to those
abroad will attract international investment, which will relieve the upward
pressure on interest rates and mitigate
increases in output.
• Estimates of Output
and Interest-Rate Effects
Many models have been developed to
study the output effects induced by increases in government expenditures.
These models may generate varying
results because of the different assumptions that they embody. For example,
large-scale macroeconomic models,
which are frequently used for policy
analysis purposes, generally yield larger
output effects than models that assume
flexible prices. The larger results generated by such models stem from assumptions about "sticky" prices. Because
the United States has been involved in
only a few wars over the last 40 years, it
is difficult to test which type of model
yields the most reliable predictions.

We have chosen to generate the output
and interest-rate effects of deficitfinanced war expenditures through a
dynamic, flexible-price model developed by Auerbach and Kotlikoff, because a previous study based on a
flexible-price model was able to successfully mimic the output effects of
World War II. In our simulations, the
war-related debt is serviced by higher
taxes on current and future generations.
Individuals increase their saving in
response to higher government deficits
in order to compensate for their own
higher future tax burdens. They do not,
however, increase their saving to compensate for the higher taxes faced by
their offspring. As a result of this financing arrangement, higher taxes for the indefinite future decrease labor supply and
saving, the latter of which eventually
leads to a reduced capital stock. Decreases in labor supply and capital stock
both reduce future output.
We use this model to highlight the qualitative importance of consumption and
savings decisions, labor supply, and fiscal
constraints. We do not expect our simulations to capture the precise historical values of interest rates and output, because
the model does not account for many realworld factors (the mitigating effects of international capital flows, for example).
To gain perspective on how the Persian
Gulf War will affect interest rates and
output, we analyze the economic impact
of the Korean and Vietnam Wars as well.
Table 1 gives the simulated values of
output and real interest rates for all three
wars. Note that the model predicts a
negligible change in the interest rate during the first year of both the Korean and
Vietnam Wars. For the current war, the
predicted first-year rise in the real rate of
interest is also negligible, so it is not surprising that our model predicts an increase in output of only 0.10 percent.
Due to higher future taxes, output begins
to decline within five years, eventually
reaching a long-run value that is roughly
0.20 percent lower than it would have
been in the no-war scenario. Because
future taxes are higher, interest rates

must rise to maintain the same after-tax

Nonetheless, it appears that the impact

rate of return on savings. For the Gulf

of a temporary increase in government

War, this increase is less than five basis


spending of the size expected to result
from the Persian Gulf War will not be
sufficient to cause the economy to devi-

The model shows that temporary increases
in government expenditures initially lead
to higher output. Our predictions are qualitatively consistent with the increase in output that occurred during past wars, as
shown in figure 2; however, the actual increases appear to be larger than expected.
These differences result from factors not
accounted for in the model. Actual output
may rise above or fall below trend because of business-cycle dynamics unrelated to war finance. This appears to have
happened during the Vietnam years, when
an economic expansion that began prior
to the 1967 escalation of the war continued throughout the war years.

• Conclusion
According to our analysis, the fiscal
effects of the Persian Gulf War on the
U.S. economy will be minimal. The projected short-run increase in output is not
expected to be large enough to pull the
economy out of a recession, and over
time, output should be somewhat lower
as a result of the higher future taxes
needed to pay for the war. We recognize,
of course, that aspects of the war not
covered in our analysis, such as oil price
shocks and other business-cycle factors,
will also influence the economy.

ate significantly from its present course.

• Footnotes
1. The estimates used in this analysis are
taken from testimony given before the House
Budget Committee by Charles A. Bowsher,
Comptroller General of the United States,
Report No. T-NSIAD-91-03 of the General
Accounting Office, January 4, 1991.
2. Briefing by Federal Reserve Board Chairman Alan Greenspan before the U.S. House
of Representatives, Committee on the Budget,
January 22, 1991. The estimate was provided
by Committee Chairman Leon E. Panetta.
3. Total government expenditures include
federal, state, and local government spending, but exclude transfer payments, since
they do not affect government absorption of
4. These percentage increases were derived
using results from Robert J. Barro, "Output
Effects of Government Purchases," Journal
of Political Economy, vol. 89, no. 6 (December 1981), pp. 1086-1121.
5. We assume here that taxes are lump sum,
whereas simulations presented later in this
paper incorporate income taxes. Both forms
of taxation yield the same qualitative effects
in the short run.
6. See Marianne'Baxter and Robert G. King,
"Fiscal Policy in General Equilibrium," University of Rochester Working Paper No. 244,
September 1990.

7. Prices are said to be sticky if they
respond sluggishly to changes in market
demand/supply conditions.
8. See Alan J. Auerbach and Laurence J.
Kotlikoff, Dynamic Fiscal Policy. Cambridge, England: Cambridge University
Press, 1987; and Mark A. Wynne, "The Aggregate Effects of Temporary Government
Purchases," Research Paper No. 9007, Federal Reserve Bank of Dallas, April 1990.
9. Details regarding the simulations are
available from the authors upon request.
10. The long-run output and interest-rate
effects depend entirely on the financing
arrangement assumed in the simulations. If
the debt due to war expenditures were paid
off after a finite period of time, the output
and interest-rate effects would be larger until
the debt was retired. However, interest rates
and output would eventually return to the
same levels as would have been realized in
the absence of the war.

Charles Carlstrom andJagadeesh Gokhale
are economists and Sharon Parrott is a research assistant at the Federal Reserve Bank
of Cleveland. The authors would like to thank
Alan Auerbach and Laurence Kotlikoff for
the use of their fiscal policy simulation
model, and Randall Eberts, June Gates, William Gavin, David Reifschneider, and Mark
Sniderman for helpful comments.
The views stated herein are those of the
authors and not necessarily those of the
Federal Reserve Bank of Cleveland or of the
Board of Governors of the Federal Reserve

Federal Reserve Bank of Cleveland
Research Department
P.O. Box 6387
Cleveland, OH 44101

Address Correction Requested:
Please send corrected mailing label to
the above address.

Material may be reprinted provided that
the source is credited. Please send copies
of reprinted materials to the editor.

U.S. Postage Paid
Cleveland, OH
Permit No. 385