Full text of Review (Federal Reserve Bank of St. Louis) : November/December 1990

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

3 T h e Internationalization o f the Beer
B rew in g Industry
20 Alternative Measures o f M on ey
as Indicators o f Inflation: A
Survey and Some N e w Evidence
34 T h e Trade-Related Aspects o f Intel­
lectual P ro p e rty Rights: W h at Is
At Stake?
47 T h e Economic Consequences o f
Reducing M ilitary Spending

THE
FEDERAL
A RFSEim

J&HXSWoi

A r S T .m r is

1

F ederal R eserve Bank o f St. Louis
R eview
November/December 1990

In This Issue . . .




The brew in g industry has evolved over the past 25 years from an in­
dustry that concentrated on domestic markets to one that view s itself as
part o f a global market. In the first article o f this Review, Jeffrey D.
Karrenbrock examines "Th e Internationalization o f the Beer Brewing In­
dustry.” Th e author provides a case study o f h ow and w hy brew ers
have used merchandise trade, licensing agreements and foreign direct
investment to penetrate foreign markets. In addition, Karrenbrock
demonstrates how certain economic factors, such as economies o f scale
and trade barriers, can affect which type o f international transaction is
used in different markets. A fe w specific countries’ barriers to beer
trade are also discussed.
* * *
On October 6, 1979, the Federal Reserve announced its landmark deci­
sion to implement m onetary policy by targeting the grow th rate o f M l
rather than some level o f the federal funds rate. Th e Federal Reserve
abandoned this practice less than three years later prim arily because
empirical regularities in the behavior o f M l velocity that had made M l
targeting an attractive strategy seemed to disappear. Although no one
has yet offered a definitive explanation fo r the observed changes in M l
velocity, these changes nonetheless have led to a general discrediting o f
m onetary aggregate targeting as an approach to m onetary policy.
In the second article o f this issue, "Alternative Measures o f M oney As
Indicators o f Inflation: A Survey and Some N ew Evidence,” Michael T.
Belongia and James A. Chalfant review the arguments about w h y M l
velocity may have changed, then discuss m ore general issues about the
construction o f a measure o f the money stock and the properties one
w ould expect it to exhibit. Noting that the composition and weighting o f
components in a m onetary aggregate each may contribute to measure­
ment erro r and unexplained shifts in behavioral relationships, the
authors then review other recent efforts to construct new m onetary ag­
gregates from first principles o f economic and statistical theory. They
compare the empirical perform ance o f the alternatives against various
final use criteria and find that, although some measures o ffe r im ­
provem ents over the current official aggregates, none can be judged at
this time to be the definitive measure o f the m oney stock.
* * *
In the third article in this issue, "Th e Trade-Related Aspects o f In­
tellectual Property Rights: W hat Is At Stake?” Alison Butler examines
one o f the major n ew issues being negotiated in the current round o f
multilateral negotiations. W ith increased trade in high-technology pro­
ducts, intellectual property rights have becom e an increasingly im por­
tant and contentious issue in international trade. The protection o f in-

NOVEMBER/DECEMBER 1990


FEDERAL RESERVE


2
tellectual property, w hich includes such things as patents, copyrights
and trademarks, is justified by most economists as being necessary fo r
invention, w hile others, prim arily policymakers, argue that these rights
only serve to protect the m onopoly pow er o f innovating firms.
The author analyzes the questions surrounding the protection o f in­
tellectual property, how innovation is affected by trade, and the benefits
o f protecting intellectual property internationally. In addition, she ex­
amines w h y the incentive to protect intellectual property differs betw een
industrialized and developing countries and what likely needs to happen
b efore an agreem ent can be reached.
* * *
Before the crisis in the Middle East this summer, the easing o f interna­
tional tensions had reduced, fo r many, the urgency fo r the United
States to continue building or even maintain its m ilitary strength. As
support fo r allocating the nation’s resources to defense weakened, peo­
ple began to argue about the potential fo r a significant “ peace dividend”
available to the U.S. economy.
In the fourth article o f this issue, "The Economic Consequences o f
Reducing M ilitary Spending,” Michelle R. Garfinkel examines some issues
that have been overlooked by those searching fo r ways to use the peace
dividend. Garfinkel attempts to answer tw o important economic ques­
tions: How can the resources from reduced m ilitary spending be real­
located efficiently to nonmilitary uses, and h ow can w e identify what
these uses should be? W hile the recent course o f events in the Middle
East might appear to detract from the prospects o f a large dividend in
the near future, it does not detract, the author says, from the relevance
o f these questions. Assuming that conflict is not a permanent obstacle to
realizing a large dividend later, it provides us with m ore time now to
evaluate the various options associated w ith future defense cuts.
* * *

BANK OF ST. LOUIS

3

Jeffrey D. Karrenbrock

Jeffrey D. Karrenbrock is an economist at the Federal Reserve
Bank of St. Louis. David H. Kelly provided research assistance.

The Internationalization of
the Beer Brew ing Industry

T

HE BEER BREWING INDUSTRY has been
undergoing a process o f internationalization fo r
the past 25 years. This article examines the roles
that three types o f international transactions —
merchandise trade, licensing agreements and
foreign direct investment — have played in this
internationalization. As in other industries, a
fe w general economic factors explain much o f
the increase in international brew in g activity.
W hat makes beer brew in g a particularly inter­
esting case study is that it provides an oppor­
tunity to demonstrate h ow certain economic
factors, such as economies o f scale and trade
barriers, can affect the internationalization o f
an industry.

THE INTERNATIONALIZATION OF
THE BREWING INDUSTRY
Merchandise Trade
As w ith most other goods, w orld merchandise
trade in b eer has expanded rapidly over the
past 25 years (see figure 1). Much o f the in­
crease in w orld beer trade — and in w orld
trade in general — can be attributed to such
factors as low er trade barriers, m ore efficient
communication and transportation technology,
and grow th in real personal incomes. Th e value
o f w orld beer trade increased from $149 million
in 1965 to $2.08 billion in 1987, a 14-fold in­
crease; at the same time, w orld trade in all

goods increased to m ore than 12.5 times its
1965 value. In m ore recent years, betw een 1980
and 1987, w orld trade in beer expanded 83.8
percent, w hile total w orld trade g rew only 23.4
percent. Despite its rapid growth, trade in beer
in 1987 accounted fo r less than one-tenth o f
one percent o f total w orld merchandise trade.
On a volume basis, w orld trade in beer has
nearly tripled since 1965, grow in g at an average
annual rate o f 6.5 percent betw een 1965 and
1987. The largest exporters o f beer in this
grow in g market, ranked by volume, are the
Netherlands, W est Germany, Czechoslovakia,
Belgium and Canada (see table l ) . 1 The largest
im porters are the United States, the United
Kingdom, France, Italy and W est Germany.
Beer imports as a percent o f total consumption
(IPC) and exports as a percent o f production
(EPP) are larger fo r some o f the smaller exporters
and im porters than they are fo r some o f the
larger exporting and im porting countries. As
table 1 shows, among 25 importers, beer IPC
ranges from a low o f 0.2 percent in N orw ay to
16.4 percent in Italy. The percent o f beer con­
sumption accounted fo r by imports in the largest
beer im porting country, the United States, is
about 5 percent.
Similarly, among exporters, figures fo r EPP
range from 0.4 percent in the United States to

'The FAO Trade Yearbook indicates that Mexico was the
world’s third-largest exporter of beer in 1987.



NOVEMBER/DECEMBER 1990

4

Figure 1
World Merchandise Trade
Billions of dollars

SOURCES:

Millions of dollars

FAO Trade Yearbook and International Financial Statistics Yearbook.

41.6 percent in Ireland. The export numbers as
a percent o f production fo r such countries as
the Netherlands and Luxembourg, how ever, are
questionable as these countries do a significant
amount o f re-exporting to other countries (that
is, much o f the beer reported as exports may
simply be im ported and then re-exported fo r
consumption elsew here).2
Few o f the countries listed in table 1 are
strictly im porters or exporters o f beer. Most o f
the countries that export beer also im port some
beer and vice-versa. This pattern o f trade is
known as intra-industry trade. An examination
o f the IPC and EPP statistics in table 1 show

2Ott (1988) notes that the Netherlands has a long history of
re-exporting imported goods.
3Ott (1988) notes that re-exported goods from the
Netherlands are not included in the country’s import
figures.

FEDERAL RESERVE BANK OF ST. LOUIS


that intra-industry trade in beer is m ore im por­
tant to some countries than others. The largest
exporter o f beer, the Netherlands, im ported on­
ly 4.3 percent o f the beer it consumed in 1987.3
Similarly, the tw o largest im porters o f beer, the
United States and the United Kingdom, exported
only 0.4 percent and 1.9 percent o f their beer
production in 1987. Ireland, on the other hand,
exported nearly 42 percent o f its production,
w hile im porting m ore than 12 percent o f its
beer consumption.
The degree o f intra-industry trade fo r a coun­
try can be measured using a simple index, cal­
culated fo r a given product as the absolute value

Selected 1987 Brewing Industry Statistics__________________________________________________________

Average Annual Growth Rate, 1975 to 1987

Australia
Austria
Belgium
Canada
Czechoslovakia
Denmark
W. Germany
Finland
France
E. Germany
Hungary
Ireland
Italy
Japan
Luxembourg
Netherlands
New Zealand
Norway
Poland
Portugal
Spain
Sweden
Switzerland
USSR
United Kingdom
USA

Exports
1000 HL

Imports
1000 HL

728.3
361.0
2537.0
2415.8
2698.0
1934.0
5706.0
22.8
672.0
N/A
0.0
2066.0
73.1
293.3
270.3
5725.2
83.2
19.2
327.1
74.5
121.0
44.9
36.7
N/A
1140.0
919.7

70.0
285.0
565.5
448.0
0.0
20.8
1301.6
13.9
2445.9
N/A
1290.0
415.0
2162.8
224.0
42.2
534.9
70.5
405.0
106.2
34.3
735.0
344.4
485.9
900.0
4093.1
10991.1

Exports
Imports
As Percent of As Percent of
Production Consumption
3.9%
4.0
18.1
10.5
12.1
22.8
6.2
0.7
3.4
N/A
0.0
41.6
0.7
0.5
40.9
32.6
2.0
0.9
2.8
1.5
0.5
1.1
0.9
N/A
1.9
0.4

0.4%
3.2
4.7
2.1
0.0
0.3
1.5
0.4
11.3
N/A
12.6
12.5
16.4
0.4
9.8
4.3
1.7
0.2
0.9
0.7
2.8
7.9
10.5
1.7
6.5
5.0

Intra-Industry
Trade Index
1975
1987
.83
.33
.23
.63
1.00
.98
.54
.60
.56
.25
N/A
.98
.95
.73
.79
.69
.68
.89
.74
.97
.58
.91
.75
N/A
.54
.80

.82
.12
.64
.69
1.00
.98
.63
.24
.57
N/A
1.00
.67
.93
.13
.73
.83
.08
.62
.51
.37
.72
.77
.86
N/A
.56
.85

Exports
Imports
Per
As Percent of As Percent of
Capita
Production Consumption Consumption
15.0%
7.4
2.4
13.8
5.9
-0.3
8.2
53.7
2.4
N/A
0.0
2.1
20.5
1.8
-0.7
6.2
7.2
-12.2
2.7
10.9
14.5
72.0
2.8
0.0
3.1
22.1

17.0%
-0.7
-5.3
52.0
0.0
*
6.3
171.1
2.4
N/A
-2.86**
39.0
5.7
19.6
0.5
2.4
36.7
22.6
12.3
N/A
19.3
8.4
6.4
8.8
3.4
13.7

- 1.7%
1.0
-0 .5
-0 .4
-0 .7
-0 .2
-0 .2
1.9
- 1.1
1.6
2.5
-2 .0
6.9
2.0
-0 .8
0.6
-0 .7
1.1
-1 .5
3.7
2.7
-1 .2
-0 .3
- 1.4
-0 .4
0.8

Imports
Per
Capita
14.7%
0.2
-5 .7
50.7
0.0
*
6.1
184.6
1.3
N/A
-0.83*
35.9
12.9
22.6
-0 .3
2.8
33.9
26.6
10.7
N/A
22.4
7.0
6.1
7.0
2.9
14.7

SOURCE: Derived from information in the Brewers Association of Canada’s International Survey: Alcoholic Beverage Taxation and Control Policies.
1 HL = 100 liters = 26.4 gallons
* Denmark’s beer imports were generally small and declining throughout the period. However, a few relatively large increases in imports in later years caused imports as a
percent of consumption and imports per capita to grow at average annual rates of 2,218 and 2,146.4, respectively.
** Average annual rate of growth for 1978-1986.



NOVEMBER/DECEMBER 1990

Table 1

6

o f the difference betw een exports and imports
divided by the sum o f exports and imports.4 If
the index is close to zero, the degree o f intra­
industry trade is substantial. An index value o f
one indicates that there is no intra-industry
trade. This index, labeled the "Intra-Industry
Trade Index” and shown in table 1, was calcu­
lated fo r beer trade in 1975 and 1987.
O f the 23 countries fo r which the index could
be calculated fo r both years, nine countries’ in­
dexes rose over the period, indicating less intra­
industry beer trade. In 12 countries, the indexes
declined, indicating that intra-industry beer trade
had increased. Only five countries in table 1 had
an intra-industry index value o f less than 0.5 in
1987. The majority (70 percent) had an index
value o f m ore than 0.5 in 1987, which indicates
that intra-industry trade plays a m inor role in
the brew in g industry in general, although it has
becom e m ore prevalent during the past 15 years.
Th e grow th rates o f IPC and EPP provide fu r­
ther evidence o f the increasing importance o f
intra-industry trade to the brew in g industry. O f
the 20 countries in table 1 reporting increased
EPP betw een 1975 and 1987, 16 also reported
increased IPC. Similarly, o f the 19 countries
reporting increased IPC, 16 reported increased
EPP. In sum, merchandise trade in beer has ex­
panded rapidly during the past 25 years, with
intra-industry trade playing a small, but growing,
role in beer trade.

Licensing Agreements
Brewers also use licensing agreements to make
their products available to foreign consumers. A
typical license agreem ent allows a b re w e r in
one country to b rew and market the beer o f a
foreign brew er. One example is AnheuserBusch’s (A-B) licensing agreem ent w ith the Cana­
dian b re w e ry John Labatt Ltd. This agreement
allows Labatt to b rew and market some A-B
beers, such as Budweiser and Michelob, in Cana­
da. In return, Labatt pays a royalty fe e to A-B.
O f course, the licensing brew ers insist that the
consistent quality o f their products be main­
tained. In essence, the licenser is selling its

“See Gray (1987), pp. 243-49.
5The U.S.-Canada Free Trade Agreement eliminates the
federal tariffs on beer between these countries, but does
not alter the pricing practices of the provincial liquor
outlets. See Carter, et al (1989) for a more detailed
description of Canadian barriers to beer trade.

FEDERAL RESERVE BANK OF ST. LOUIS


know-how in brew in g a specific beer, the right
to use a trademark and the name recognition it
has built fo r that tradem ark in exchange fo r a
royalty payment from the licensee. There w ere
at least 30 licensing agreements among various
brew ers around the w orld in 1987 (see table 2).
Several factors that are not mutually exclusive
prom ote the use o f licensing agreements. First,
some firm s use licensing agreements to circum ­
vent trade barriers. For example, U.S. beers
that are b rew ed in Canada under license agree­
ment are not subject to either the Canadian
federal ta riff or the discriminatory mark-ups
that other im ported beers face at the provincial
governm ent outlets.5 Second, the physical quali­
ties o f beer prom ote the use o f licensing agree­
ments. Beer is about 90 percent water, so trans­
portation costs can be reduced through local
production. In addition, beer has a shelf-life o f
about three to fou r months, o f which tw o to
three weeks could be taken up by overseas
shipment. Also, w hen companies enter new
markets, they often find it m ore profitable to
license existing plants and distribution systems
to handle their products rather than build their
ow n plants and establish their ow n distribution
systems.
The im port and export figures discussed p revi­
ously did not include consumption o f foreignheld brand names that are b rew ed domestically
under a licensing agreement. Thus, the degree
o f internationalization is understated w hen only
merchandise trade is analyzed. Inform ation on
the amount o f beer b rew ed under licensing
agreements is usually closely held b y the com­
panies involved, and not much data are publicly
available. The Conference Board o f Canada,
how ever, has estimated the impact o f licensed
b rew ed beer in the Canadian beer market and
its findings serve to demonstrate how important
licensed production can be.
In the Canadian market, three U.S. brewers,
Anheuser-Busch, Coors and Miller, had licensing
agreements w ith the three largest Canadian
b rew ers— John Labatt, Molson and Carling
O’K eefe— respectively, in 1986.6 Th e Conference

6Molson and Carling O’Keefe agreed to merge their
breweries in 1989. The new company is called Molson
Breweries and will continue to brew for Coors and Miller
under license.

7

Table 2
Licensing Accords Between Major Brewers
Anheuser-Busch
Heineken
Kirin
Guinness
Kirin brews Heineken in Japan
Guinness brews Bud in Ireland
Labatt
Whitbread
Labatt brews Bud, Michelob & Bud Light in
Whitbread brews Heineken in U.K.
Canada
Kirin
National
National brews Bud in Israel
Molson
Molson test brewing Kirin in Canada**
Oriental
Oriental brews Bud in Korea
Lowenbrau
Suntory
Allied-Lyons
Suntory brews Bud in Japan
Allied-Lyons brews Lowenbrau in U.K.
United
Asahi
United brews Bud in Denmark
Asahi brews Lowenbrau in Japan
Watney Mann
Miller
Watney brews Bud in U.K.
Miller brews Lowenbrau in U.S.
Artois
Molson
Whitbread
Molson brews Lowenbrau in Canada
Whitbread brews Stella in U.K.
Miller
Carlton & United*
Carling O’Keefe'
Carling O’Keefe*
Carling brews High Life, Lite in Canada
Carling brews Foster's in Canada
Courage *
Courage *
Courage brews Lite in U.K.
Courage brews Foster’s in U.K.
United
Watney Mann
Watney brews Foster’s in U.K.
Beamish & Crawford
Beamish & Crawford brews Carlsberg in
Castlemaine Parkins
Ireland
Allied-Lyons
Carling O’Keefe *
Allied brews XXX in U.K.
Carling brews Carlsberg in Canada1
Coors
Falken
Falken brews Carlsberg in Sweden
Asahi
Asahi brews Coors in Japan
Heileman
Molson
Heileman brews Tuborg in U.S.
Molson brews Coors in Canada
Suntory
Suntory
brews
Carlsberg
in
Japan
Guinness
Labatt
Watney Mann
Labatt brews Guinness in Canada
Watney brews Carlsberg in U.K.

’ Parent company, Elders IXL.
**No licensing agreement in effect.
SOURCE: Modem Brewery Age, July 13, 1987.
'John Labatt Ltd. acquired the right to brew Carlsberg in Canada, effective July 1, 1988.




NOVEMBER/DECEMBER 1990

8

Board estimates that brands produced in Canada
under license w ith U.S. brew ers in 1986 may
have accounted fo r as much as 15 percent o f
beer sales in Canada.7 This amounts to approx­
imately 2.7 million barrels o f U.S. brands pro­
duced and sold under license agreements in
Canada in 1986. I f these estimates are correct,
the volume o f licensed production o f U.S. beers
in Canada was m ore than 17 times the amount
o f beer exported directly to Canada in 1986 and
m ore than four times the amount o f total U.S.
beer exports to all countries (exclusive o f ship­
ments to U.S. m ilitary bases and Puerto Rico).8
In terms o f Canadian consumption, the licensed
b rew ed beer might have accounted fo r 15 per­
cent o f Canadian beer consumption compared
with the 2.1 percent o f domestic consumption
accounted fo r by imports.
The numerous licensing agreements with
brew eries in Japan and the United Kingdom
might indicate that beer produced under license
represents a significant part o f the foreign beer
consumed in these countries. At least in some
countries, beer produced under license clearly
accounts fo r a much larger portion o f foreign
beer consumption than does im ported beer.

Foreign Direct Investment
In addition to merchandise trade and licensing
agreements, the internationalization o f the b re w ­
ing industry has been characterized b y the in­
creasing production o f beer by "foreign-ow ned”
firms. This production reflects the increasing
frequency o f foreign direct investment (FDI), in
which one b re w e r purchases an existing firm or
invests in a new or existing facility in a foreign
country. Like licensing agreements, FDI is a
substitute fo r merchandise trade. Firms may be
prom pted to use FDI fo r the same reasons they
use licensing agreements. In addition, such fac­
tors as low er labor and energy costs and less
governm ent regulation may also encourage the
use o f FDI.
Several b rew ers have invested capital in
brew eries outside their home countries. T w o
Australian brew ers, Elders IXL and Bond, have

Conference Board of Canada (1989), p. 9.
8Data for U.S. beer exports to Canada and total U.S. beer
exports, exclusive of shipments to military bases, Puerto
Rico and the territories, were provided courtesy of R.S.
Weinberg & Associates.
9Carrington (1989).

FEDERAL RESERVE


BANK OF ST. LOUIS

used this method o f globalization extensively.
Elders purchased Courage Ltd. o f England in
1986, then purchased Carling O ’K eefe breweries
o f Canada in 1987. Early in 1989, Britain's De­
partment o f Trade and Industry blocked the
proposed takeover o f Scottish & Newcastle
Breweries PLC by Elders IXL Ltd.9 In 1990, Elders
IXL announced that its United Kingdom Courage
brew eries w ould purchase Grand M etropolitan’s
U.K. brew ing and brands interests and its beer
distribution and wholesaling activities. Further­
more, Courage’s 4,900 pubs w ill m erge w ith
Grand M et’s 3,570 pubs in a joint venture under
the name Inntrepreneur Estates.10 Bond has
purchased tw o U.S. breweries, Pittsburgh B rew ­
ing and G. Heileman.
Japanese b rew ers also have actively invested
in foreign b re w e ry operations. In 1989, Tatsuuma-Honke Brewing Co. announced plans to
build a sake b re w e ry on the grounds o f the
Coors b rew ery in Colorado. In 1990, Asahi
Breweries Ltd. announced plans to invest $70
million to open a b re w e ry near Denver, Colorado,
w h ere it w ill produce a dry b eer.11 Finally, the
Canadian brew ing company John Labatt Ltd.
purchased Latrobe Brewing Co. o f the United
States in 1987.

THE ECONOMICS OF INTERNA­
TIONALIZATION
Underlying the preceding description o f the
internationalization o f the brew ing industry are
some economic factors. The next section outlines
the reasons w h y demand fo r foreign beer can
exist in a country that already produces some
domestic brands and discusses how changing
relative prices and rising income can expand
the demand fo r foreign beer. The second sec­
tion analyzes the basic economic factors that
determine the type o f international transaction
a b rew ery w ill use to put its products in the
hands o f foreign consumers. A m ore technical
presentation o f the economics discussed in these
tw o sections is provided in appendixes to this
article.

10See Thornhill and Harris (1990), Harris (1990) and
Sherwell (1990).
" “ Japan’s Asahi Plans Brewery in U.S.” (1990).

9

Foreign Demand: The Attributes
Approach

the low-calorie beer the rest o f the time.

Why Demand f o r Foreign B eer Can Exist—

The demand o f foreign beer can expand if its
price falls relative to the price o f domestic beer.
I f a consumer had been purchasing domestic
beer, the relative fall in the price o f the foreign
beer may be enough to compensate him fo r any
perceived loss in services due to switching from
the domestic to the foreign brand. In this case,
the quantity demanded o f the foreign beer w ill
increase. Th e decline in the relative price o f the
foreign beer may also encourage people w ho
already consume it to purchase more. U n for­
tunately, data on im ported beer prices are
scarce, and thus the role that changing relative
foreign beer prices has played in the globaliza­
tion o f the industry is uncertain.

One reason w hy people consume foreign beer is
that they can buy it at a price at w hich they
w ant m ore beer than domestic brew ers want to
produce. That is, the quantity o f beer demand­
ed is larger than the quantity o f beer supplied
domestically at the price o f foreign beer, and
therefore, some foreign beer is im ported to
meet the excess demand. Another reason w hy
people consume foreign beer is that at least
some consumers p refer the attributes, or char­
acteristics, o f the foreign beer over domestic
brands. This second possibility is discussed in
this section.
In general, consumers purchase beer fo r the
"services” that they feel it can provide.12 Con­
sumers have a w ide variety o f beer brands to
choose from and, subject to price and income
limitations, w ill choose those brands that have
the attributes that most closely match their de­
sired services from drinking beer. Many attri­
butes, such as taste, caloric content, alcohol
content and packaging, distinguish one brand
from another, and each combination o f charac­
teristics offers a distinctly different package of
services.
Brewers differentiate their products on the
basis o f attributes and price. Consumers com­
pare the package o f services provided by a par­
ticular beer and its price to the services and
prices o f other brands. If consumers p refer the
services o f foreign beer over domestic beer, at
given market prices, then demand fo r foreign
beer w ill exist in a country.
O f course consumers do not necessarily con­
sume only one domestic beer or foreign beer.
Consumer satisfaction may be maximized by
purchasing a combination o f domestic and fo r ­
eign beers. Suppose a consumer prefers the
taste o f a high-calorie foreign beer over all other
domestic brands, but needs to watch his caloric
intake and finds the taste o f a particular brand
o f light beer to be acceptable. This consumer
might purchase both the foreign and domestic
beers, drinking the foreign beer in limited
amounts, say, on special occasions, and drinking

12Much of the information in this section on the attributes
model is taken from Douglas (1987).
13For a review of beer demand estimates, see Ornstein
(1980). Also see Heien and Pompelli (1988). The estimated



Growth in the Demand f o r Foreign B eer—

Increases in consumer incomes can also spur
the demand fo r foreign products. W hen con­
sumers’ incomes increase, they are able to pur­
chase m ore o f all o f the products they desire. In
general, how ever, the quantities purchased o f
some goods, like flour, decline as incomes rise,
w hile quantities o f other goods purchased, like
furniture, increase as incomes rise. The statistical
evidence relating beer consumption to income
grow th is m ixed.13 Some studies have shown
that the quantity o f beer consumed increases as
income increases, w hile others have shown the
opposite.
Although little w ork has been done to estimate
the relationship betw een foreign beer consump­
tion and income in general, there is some data
to suggest that the demand fo r foreign beer
might be positively influenced by increases in
income. All 21 OECD countries in table 1 that
provided im port data on beer had positive per
capita gross domestic product grow th between
1975 and 1986; 16 o f these reported a positive
average annual rate o f grow th o f beer imports
per capita and imports as a percent o f consump­
tion. In addition, the market fo r im ported beers
g rew much m ore rapidly than most domestic
beer markets during the late 1970s and early
1980s, a period o f income grow th fo r most
countries. These figures roughly suggest that
per capita income grow th has contributed to
the internationalization o f the brew ing industry.

income elasticity of beer in these studies ranged from
-0.46 to 0.79.

NOVEMBER/DECEMBER 1990

10

Where to P rodu ce Beer: Some
Econom ics
Once a firm determines that foreign demand
fo r its products exists, it must determine the
lowest-cost method o f supplying these products
to the foreign market. Should the firm use direct
exports, a licensing agreem ent or direct foreign
investment to enter the target market? The
answer is relatively simple, in theory, and is
based on the principle o f profit maximization.
A brew ery's total cost o f supplying a foreign
m arket is equal to the b eer’s cost o f production
plus transportation and distribution costs, m ar­
keting costs and overhead. A b re w e ry ’s cost o f
producing beer is a function o f its production
technology and the cost o f its inputs, such as
labor, agricultural ingredients and packaging
materials. Research has shown that the average
cost o f producing beer declines as production
expands.14 That is, economies o f scale exist in
the brew ing industry. Economies o f scale en­
courage direct exporting w hen the quantity
demanded o f foreign beer is relatively small and
encourage foreign production either through
licensing or foreign direct investment w hen the
quantity demanded is relatively large.
Suppose a U.S. b re w e r and a Japanese b rew er
have identical production functions exhibiting
economies o f scale and that the firms pay the
same price fo r their inputs. That is, their average
cost o f production curves are equal and are
shaped as shown in figure 2. (For simplicity,
assume that the U.S.-Japan exchange rate is fixed
throughout and, given this exchange rate, Japa­
nese prices are stated in U.S. dollars.) As the
brew ers expand production, the average cost o f
producing a unit o f the product falls up to a
point, after which average costs no longer de­
cline but stabilize. Assume that Japanese de­
mand exists fo r a beer—called Colony—produced
b y the U.S. firm. The U.S. firm must determine
w hether it can supply the Japanese market
cheaper b y producing Colony domestically and
exporting or by producing it in Japan, either
under license or by FDI.
N ow suppose that supply and demand condi­
tions and price in the U.S. are such that

14Elzinga (1973), Fuss and Gupta (1981), Keithman (1978)
and Scherer (1973) all provide evidence that economies of
scale exist in the brewing industry. See Thompson (1985)
or any micro-economic text for a discussion of the reasons
why economies or diseconomies of scale can exist at the
plant level.

FEDERAL RESERVE BANK OF ST. LOUIS


Am erican consumers consume Q us units o f Col­
ony, as shown in figure 2. (Note that the supply
and demand curves are not shown in figure 2
and the quantity Q os is simply given.) This sub­
stantial amount o f consumption allows the U.S.
firm to achieve significant economies o f scale,
producing Q us units o f Colony at an average
cost o f C2 per unit. Also assume that market
conditions are such that a relatively small quan­
tity o f Colony, Q j, , is demanded in Japan. Since
the U.S. b re w e r is already producing some Col­
ony fo r domestic consumption, expanding p ro­
duction to meet the extra demand o f Colony in
Japan w ould allow the U.S. b re w e r to m ove
dow n its average cost-of-production curve from
point A to point B, w h ere it could produce Col­
ony fo r C,.15
The alternative to producing the beer in the
United States and exporting it is to produce Col­
ony in Japan. Since Colony is currently not be­
ing produced in Japan, the Japanese firm or
branch b re w e ry built by the U.S. firm would
have to b rew the relatively small amount o f Col­
ony, Q ,, , at a high average cost o f production,
C3. In other words, the relatively small quantity
o f production w ill not allow the Japanese plants
to achieve significant economies o f scale. Thus,
producing Colony in the United States fo r ex­
port would save the b re w e r C3- C , per unit of
Colony. If the cost o f transporting Colony to
Japan and distributing Colony in Japan is less
than the difference betw een C3 and C,, then the
U.S. b rew er w ould maximize profits by exporting
Colony to Japan. If the quantity demanded o f
Colony in Japan w ere larger, it might be m ore
profitable fo r the b re w e r to use a licensing
agreement or foreign direct investment.
Suppose that the quantity o f Colony sold in
Japan grow s to Q ,2 as shown in figure 3, while
sales o f Colony in the United States remain at
Q us. Since the U.S. b re w e r has exhausted its
economies o f scale, it cannot produce Q us+ Q ,2
at a low er per unit cost than that fo r Qus+Qj,.
The Japanese brew ery, how ever, by increasing
production from Q n to Q J2 could n ow match
the U.S. b re w e r’s cost o f production because it
has also achieved the low est possible average
cost o f production. Thus, given equal average

,5For simplicity, we have ignored any quantity response,
stemming from lower prices, that might occur in the U.S.
market as a result of the expanded output.

11

Figure 2
Average Cost of Production: U.S. Plant Achieves
Economies of Scale

Figure 3

of output

Average Cost of Production: U.S. and Japanese
Plants Achieve Economies of Scale




of output

NOVEMBER/DECEMBER 1990

12

production costs, the U.S. firm w ill now prefer
to either negotiate a licensing agreem ent with
the Japanese b rew er or use FDI, thereby saving
the additional export-related expenses o f ship­
ping Colony overseas and distributing it within
Japan.
Like transportation costs, trade barriers also
offset production cost advantages. I f a target
country has high tariffs or distribution systems
fo r im ported goods that are relatively costly,
production cost advantages in the home country
may be offset and licensing and foreign direct
investment becom e the only feasible methods o f
entering the target market. As shown below,
trade barriers have had a significant effect on
the choice o f licensing agreements and foreign
direct investment in the internationalization o f
the brew ing industry.

BARRIERS TO BEER TRADE
Japan
In Japan, tw o types o f barriers inhibit foreign
beer from entering the country. The most
significant o f these is the Japanese distribution
system. Th e Japanese have a complex multi­
tiered system, comparable to the U.S. beer
distribution system, in which beer moves from
producer to w holesaler to consumer.16 In addi­
tion, Japan has little warehouse space, which
means shipments are smaller and m ore frequent
than in the United States. Both aspects o f the
Japanese distribution system raise the cost o f
distributing beer in Japan, relative to less com ­
plex systems. Japan also charges a small customs
duty on im ported beer. These factors raise the
cost o f exporting beer to Japan and make licens­
ing agreements or foreign direct investment
relatively m ore attractive methods o f selling
beer in Japan.

Canada
A G A TT panel ruled in 1988 that specific
practices o f the Canadian provincial govern ­
ments discriminated against im ported b eer.17
Canadian trade barriers include discriminatory
mark-ups at provincial liquor outlets and dif­
ferent marketing techniques fo r foreign beer,

16See VandeWater and Curley (1990).
17G. Heileman Brewing Co. filed a Section 301 trade action
against Canada over unfair pricing and distribution prac
http://fraser.stlouisfed.org/
FEDERAL RESERVE
Federal Reserve Bank of St. Louis

BANK OF ST. LOUIS

such as smaller packages and w arm foreign
beer sales at the governm ental outlets. These
non-tariff barriers have prom pted U.S. brew ers
to use licensing agreements in Canada even
though several b rew ers have U.S. plants that
are located quite close to the Canadian border.
The Australian b re w e r Elders IXL has chosen to
use foreign direct investment to enter the Cana­
dian market. This creates an interesting situa­
tion in w hich a U.S. beer is being made under
license in a Canadian b rew ery that is partially
ow ned by an Australian brew er.

United Kingdom
As in Japan, distribution practices are the
main barriers to trade in the United Kingdom.
Most beer consumed in the United Kingdom is
draft beer, and most o f this is sold in pubs.
Many pubs are ow ned outright by breweries,
managed b y the brew eries or leased to in­
dividuals w ho enter into exclusive supply agree­
ments w ith the breweries. This system was the
subject o f eight investigations betw een 1966 and
1986, that focused chiefly on pricing and supply
competition.18 Given the relationship betw een
the pubs and the domestic brew eries, foreign
label brew ers have problems getting local
brew ers to carry their products in British pubs.
Thus, many foreign brew ers have chosen to use
licensing agreements with domestic firms to
penetrate the U.K. beer market. Foreign direct
investment has also been used to enter this
market.

SUMMARY
The brew ing industry has evolved from an in­
dustry that concentrated on domestic markets
to one that views itself as part o f a global m ar­
ket. This internationalization has occurred via
the use o f merchandise trade, licensing agree­
ments and foreign direct investment. M erchan­
dise trade in beer has developed in an intra­
industry pattern, whereas international transac­
tions in licensing agreements and foreign direct
investment have not developed, in general, in a
bilateral pattern. Licensed production and p ro­
duction at foreign-ow ned brew eries likely ac­
count fo r an unknown, but probably large, part
o f foreign beer consumption in some countries.

tices of provincial governments. See Daily Report for Ex­
ecutives (1990).
18Brewers Association of Canada (1989), pp. 387-88.

13

T w o conditions must hold fo r trade in similar
goods, such as beer, to occur. First, the foreign
product must o ffer a combination o f desired at­
tributes that are not available to domestic con­
sumers from domestic products. Second, it must
be profitable to produce the product fo r the
foreign market. W hich type o f international
transaction will be used to supply a foreign m ar­
ket with beer is related to the existence o f eco­
nomies o f scale, distribution costs and trade
barriers.

REFERENCES

Brewers Association of Canada. International Survey:
Alcoholic Beverage Taxation and Control Policies, 7th ed.
(Ottawa, Canada, June 1989).
Carrington, Tim. “ Britain Blocks Elders Takeover of Big
Brewer; Government Urges Reforms in Reflecting Firm’s
Offer for Scottish & Newcastle,” The Wall Street Journal,
March 22, 1989.
Carter, Colin A., Jeffrey Karrenbrock and William W. Wilson.
“ Freer Trade In the North American Beer and Flour
Markets,” in Andrew Schmitz, ed., Free Trade and
Agricultural Diversification: Canada and the United States,
(Westview Press, 1989).
The Conference Board of Canada. The Canadian Brewing In­
dustry: Historical Evolution and Competitive Structure, (Ot­
tawa, Canada, February 1989).
Daily Report for Executives. (The Bureau of National Affairs,
Inc., Washington D.C., May 18, 1990), p. A-3.
Douglas, Evan J. Managerial Economics: Analysis and
Strategy, 3rd ed. (Prentice-Hall, Inc., 1987), pp. 86-103.
Elzinga, Kenneth G. “ The Restructuring of the U.S. Brewing
Industry,” Industrial Organization Review (Vol. 1, 1973), pp.
101-14.
Finnegan, Terri. “ International Licensing Pacts on the Rise,”
Modem Brewery Age (July 13, 1987).
Food and Agricultural Organization of the United Nations.
FAO Trade Yearbook, various issues.
Fuss, Melvyn A., and Vinod K. Gupta. “A Cost Function Ap­
proach to the Estimation of Minimum Efficient Scale,
Returns to Scale, and Suboptimal Capacity,” European
Economic Review (February 1981), pp. 123-35.




Gray, H. Peter. International Economic Problems and
Policies (St. Martin’s Press, Inc., 1987).
Harris, Clay. “ Time for Takeovers, Gentlemen Please,”
Financial Times, March 14, 1990, London, England.
Heien, Dale, and Greg Pompelli. “The Demand for Alcoholic
Beverages: Economic and Demographic Effects,” Southern
Economic Journal (January 1989), pp. 759-70.
Horstmann, Ignatius, and James R. Markusen. “ Licensing
Versus Direct Investment: A Model of Internationalization by
the Multinational Enterprise,” Canadian Journal of
Economics (August 1987), pp. 464-81.
International Monetary Fund. International Financial
Statistics Yearbook (1989), pp. 122-23.
“Japan’s Asahi Plans Brewery in U.S.” St. Louis PostDispatch, January 3, 1990.
Keithman, Charles. The Brewing Industry, Staff Report of the
Bureau of Economics, U.S. Federal Trade Commission
(1978).
Lancaster, K. "A New Approach to Consumer Theory,” Jour­
nal of Political Economy (April 1966), pp. 132-57.
_______ . Consumer Demand: A New Approach (Columbia
University Press, 1971).
Ornstein, Stanley I. “Control of Alcohol Consumption
Through Price Increases,” Journal of Studies on Alcohol
(September 1980), pp. 807-18.
Ott, Mack. “ Have U.S. Exports Been Larger Than
Reported?” this Review (September/October 1988), pp.
3-23.
Scherer, F.W. “ The Determinants of Industrial Plant Sizes in
Six Nations,” Review of Economics and Statistics (May
1973), pp. 135-45.
Sherwell, Chris. “ Foster’s Brewing Sees Global Sales of
A$9bn,” Financial Times, March 14, 1990, London, England.
Thompson, Arthur A. Economics of the Firm: Theory and
Practice, 4th ed. (Prentice-Hall, Inc. 1985).
Thornhill, John, and Clay Harris. “GrandMet and Elders Un­
veil Swap Details,” Financial Times, March 14, 1990, Lon­
don, England.
VandeWater, Judith, and John Curley. “ Beer Barrels
Eastward,” St. Louis Post-Dispatch, June 11, 1990.
Weinberg, R.S. & Associates. “Tax Free Removals, Destina­
tion of Exports and Consumed on Brewery Premises,
1947-1988” (Brewing Industry Research Program, St. Louis,
Missouri).

NOVEMBER/DECEMBER 1990

14

Appendix A
Foreign Demand fo r Beer: The Attributes Model
The attributes model, introduced by Lancaster
(1966, 1971), can be used to show how demand
fo r foreign beer can exist in a country in which
domestic brands are already produced.1 Suppose
a consumer chooses betw een tw o brands o f
beer so that, subject to income limitations, his
satisfaction from the services provided by the
beer is maximized. For simplicity, assume the
consumer values only tw o attributes o f beer:
taste and low calories. The tw o types o f beer
provide these attributes in differing proportions
and at different prices.

Figure A1
Maximization of Utility by Consuming
Only Domestic Beer

A fter sampling both products, the consumer
rates each brand on a scale o f 1 to 3, 3 being
best, fo r both taste and calorie content, as shown
in table A l. Brand F is a foreign beer that tastes
great, but is high in calories (thus receiving a
low rating on caloric content) giving it a rela­
tively high ratio o f taste-to-calorie appeal. Brand
D, a domestic beer, does not taste quite as
good, but is very low in calories. This beer then
has a relatively low taste-to-calorie appeal ratio.

Table A1
Attribute Ratings and Prices of Three
Beers____________________________

Attribute Rating
Brand Calories Taste

Ratio of
taste to Bar price Bottles
calories per bottle per $12

D

3

1

1/3

$2

6

F

1

3

3

$3

4

The amount o f each beer the consumer can
purchase is determined by his income and
the price o f the products. Assume that the
consumer has decided to spend $12 on beer
during a visit to a local bar and the cost of
each kind o f beer is as shown in table A l. If
he spends the entire $12 on only one p ro­
duct, he could buy at most six bottles o f D or
four bottles o f F. Four bottles o f F would
provide 4 units (4 bottles X 1 unit o f caloric

1Much of the information in this appendix follows the
discussion of the attributes model as presented by
Douglas (1987), pp. 86-102.

FEDERAL RESERVE


BANK OF ST. LOUIS

appeal per bottle) o f the caloric attribute and 12
units (4 X 3) o f the taste attribute.
The tw o products are depicted in figure A l in
an attribute space as rays from the origin. The
slope o f each ray is determined by the ratio of
taste to calorie appeal. If the consumer drinks
brand F, then he moves out along ray F, absorb­
ing the tw o attributes in a ratio o f 3:1. Points A
and B represent the maximum amount o f the
tw o attributes that can be obtained by consum­
ing beers F and D, respectively, given the spend­
ing constraint o f $12.
Joining points A and B provides the consumer's
efficiency frontier. The efficiency fron tier is the
outer boundary o f the attainable combination of
the tw o attributes, given the budget constraint
o f $12. It is called efficient because a utilitymaximizing consumer w ill get m ore utility by
being on the fron tier rather than within the
frontier, even though these interior points are
attainable.

15

Figure A2

Figure A3

Maximization of Utility by Consuming
Only Foreign Beer

Maximization of Utility by Consuming
Both Domestic and Foreign Beer

H ow do w e know which beer, or combination
o f beers, the consumer w ill choose? In the at­
tributes model, consumer preferences betw een
attributes can be expressed using indifference
curves. Like indifference curves used to express
the marginal rate o f substitution (MRS) betw een
tw o products, the attribute indifference curves
express the MRS betw een attributes, and higher
indifference curves represent higher levels o f
utility. The beer consumer's assumed attribute
indifference curves I, and I2 have been superim­
posed on the attribute space in figure A l. As­
suming that a consumer wishes to maximize his
satisfaction from taste and caloric content, he
w ould choose to be on the highest attainable in­
difference curve, which occurs at point B in
figure A l.
The position and slope o f the indifference
curves w ill determine the brand or brands o f
beer chosen. This particular consumer has an
indifference curve that is relatively steep indi­
cating that, compared with a consumer with a
flat indifference curve, he is willing to give up a
lot o f taste to get a few less calories.

the foreign brand F, at point A. Thus, in a socie­
ty w h ere some consumers p refer the attributes
o f foreign beers over domestic beers, a demand
fo r foreign beer w ill exist.

Figure A3 shows an example o f a consumer
w ho would purchase both foreign and some
domestic beer. Neither o f the beers provides the
attributes exactly in the ratio represented by
point N. The consumer could reach this point,
how ever, by consuming some o f both products.
Bv consuming L units o f the domestic brand,
the consumer would obtain Y, units o f taste and
X ( units o f caloric appeal. By spending the rest
o f his budget on brand F, the consumer would
travel along the path LN, which has the same
slope as ray F, to obtain the X2-X , units o f calo­
rie appeal and the Y 2- Y , units o f taste needed
to reach his maximum level o f utility at point N.
Alternatively, the consumer could have started
by consuming M units o f brand F and then con­
sumed L units o f brand D to reach point N.

Growth in Foreign Demand
N ow suppose that a different consumer, w ho
gives the beers the same attribute ratings, is
willing to consume a lot m ore calories to gain a
bit better taste. The shape o f this consumer’s in­
difference curve w ould be m ore flat, and as
shown in figure A2, this person would choose



The demand fo r foreign beer can increase if
the relative price o f the foreign beer falls. As
shown in figure A4, w hen the price o f the fo r­
eign beer falls, the maximum amount o f the
foreign brand that can be purchased increases,

NOVEMBER/DECEMBER 1990

16

Figure A4

Figure A5

The Price Effect Shown by the Attribute
Approach

The Income Effect Shown by the Attribute
Approach

shifting out the efficiency fron tier from AB to
BC. In this example, the consumer goes from
buying some o f both brands at point X to only
buying the foreign brand at point C.

consumption o f each beer. T o maximize utility,
the consumer shifts to point C on the higher in­
difference curve I2, w h ere he consumes some o f
both brands o f beer. Thus, a higher level o f in­
come has induced the consumer to purchase
m ore o f the foreign brand o f beer.

Expanding incomes can also explain increased
consumer demand fo r foreign beer. In figure
A5, the consum er’s increase in income has led
to a shift from buying only the domestic beer to
buying some o f both beers. Initially the con­
sumer’s efficiency fron tier is AB, the highest in­
difference curve attainable is I t and the con­
sumer purchases only the domestic product D.
W hen the consumer’s income increases, the e ffi­
ciency frontier shifts out in a parallel fashion to
A'B', because prices and attributes are fixed and
only income is changing. For the given prices
and attributes, the increased income allows m ore


FEDERAL RESERVE


BANK OF ST. LOUIS

O f course, the example could have been con­
structed to show how a shift in the consumer’s
income could have led to a reduction in the
amount o f foreign beer purchased. Several re ­
searchers have estimated the demand fo r beer,
but no conclusive evidence has been provided
as to w hether b eer consumption expands as a
consumer’s income expands. Some data, as dis­
cussed in the text, how ever, suggest that the de­
mand fo r foreign beer might be positively in­
fluenced by increases in income.

17

Appendix B
Whether to Produce Domestically or Abroad: Some
Economics

Assume that the long-run average cost (LRAC)
curves o f any plants w h ere Colony could be
b rew ed are identical. These potential plants in­
clude the U.S. parent firm and any branch plant
it may establish in Japan, as w ell as any existing
b rew ery in Japan that could b rew Colony under
license. In addition, assume that the U.S.-Japan
exchange rate is fixed throughout and, given
this exchange rate, Japanese prices are stated in
U.S. dollars.

in figure Bl, part o f the demand curve is above
the LRAC curve, indicating that consumers are
willing to pay a price fo r Colony that is above
the U.S. firm 's average cost fo r some levels o f
production. In Japan, how ever, if the demand
fo r Colony is relatively small, say D ,, as shown
in figure B2, then Colony cannot be b rew ed
profitably there. At each quantity along D „ , the
price consumers are willing to pay is below the
Japanese plant's average cost o f producing Col­
ony. Thus, when demand is D,,, direct exports
from the U.S. b re w e ry will be the only w ay the
Japanese market might be supplied with Colony.
However, if the Japanese demand schedule for
Colony w ere much larger, say, at D J2, then the
Japanese b rew ery could at least cover its p ro­
duction costs fo r some level o f production.
Thus, at this larger level o f demand, both p ro­
duction in the United States and in Japan are
potential routes o f supplying Japan with Colony.

Consider the level o f demand fo r Colony in
the tw o countries. In the United States, as shown

Consider the cost o f producing different levels
o f Colony fo r foreign consumption m ore closely.

Suppose that a U.S. b rew er wants to sell its
beer, Colony, in Japan. The U.S. b rew er has
three methods o f supplying Japan with Colony:
1) produce Colony domestically and export to
Japan, 2) negotiate a licensing agreement w ith a
Japanese b rew er w h o w ould b rew and distribute
Colony in Japan, or 3) produce Colony in Japan
either by purchasing an existing Japanese b re w ­
ery or building a new brew ery. W hich is the
cheapest?

Figure B1

Figure B2

U.S. Production Costs and Demand
for Beer

Japanese Production Costs and Two
Demand Curves for Beer




NOVEMBER/DECEMBER 1990

18

Figure B3

Figure B4

U.S. Plant’s Production Costs

Japanese Plant’s Production Costs
Compared to U.S. Plant’s Cost of
Exporting

The goal is to determ ine which type o f interna­
tional transaction allows the firm to provide
Colony to the Japanese market at the lowest
cost. First, consider the U.S. b re w e r’s cost of
producing and exporting Colony to Japan. As­
sume fo r a given market price, the quantity o f
Colony demanded in the United States is Q us, as
shown in figure B3. The U.S. b rew er produces
this amount at an average cost o f C4 per unit.

U.S. brew er's cost o f producing different
amounts o f Colony fo r export to Japan, shown
by the line LR A C *S in figure B4.

N ow suppose that fo r a given price, the quan­
tity demanded o f Colony in Japan is Q,,, as
shown in Figures B3 and B4. Since the U.S.
b rew er already produces some Colony fo r do­
mestic consumption, by expanding production
to Q us + Q ji to meet the export demand, the U.S.
b re w e ry could m ove down its LRAC curve from
point A to B in figure B3, low ering its average
cost o f producing Colony from C4 to C^.1 If the
quantity demanded o f Colony in Japan w as
larger, at Q J2, the U.S. b re w e ry ’s average cost
o f producing it w ould fall even further to C, at
point C. By doing a similar analysis fo r other
quantities o f U.S. exports, w e can develop the

Figure B4 allows a straightforw ard com­
parison o f the production costs o f exports to
Japan, LRAC£,S, with the production o f these
quantities in Japan, LRACj. It shows that the
average cost faced by the U.S. b rew er produc­
ing a given amount o f Colony fo r export is low er
than the Japanese b re w e r’s average cost, LRAC „
up to the quantity Q J3, but higher fo r all subse­
quent levels. This is possible, even when cost
curves are identical across countries, because
the U.S. plant was already producing Colony fo r
domestic consumption and that by expanding
production to meet export demand, the average
cost o f producing Colony fell. The Japanese
plant currently is not producing any Colony; if
it w ere to start brew ing Colony fo r Japanese
consumption, it w ould have to start at a higher
cost on its long-run average cost curve.
Economies o f scale, how ever, do not continue
indefinitely. Consequently, the cost o f producing

1For simplicity, the analysis ignores any potential sales
price decline in the United States that may stem from the
brewer achieving greater economies of scale. Thus, the

quantity demanded in the United States, after production
is increased to meet export demand, is assumed to remain
at the same level as before export production occurred.


FEDERAL RESERVE BANK OF ST. LOUIS


19

Colony eventually starts to rise and the Japanese
plant can produce Colony cheaper than the U.S.
plant after point Q J3.
Thus far, the discussion has focused solely on
the cost o f producing Colony. Transportation
and distribution costs are likely to influence
w h ere production is located. Assume that it
costs the same per unit to ship and distribute a
small amount o f Colony as it does a large amount
o f Colony. Since the U.S. plant has to ship Colony
overseas, it is reasonable to assume that its
transportation and distribution costs w ill be sig­
nificantly higher than a Japanese plant’s w ould
be if Colony w ere produced there. Adding these
average per unit transportation and distribution
costs to the plant's respective long-run average
cost o f production curves gives the tw o dashed
lines, LR A C *s+ t us and LRACj + t,, shown in
figure B4. The U.S. plant can produce, transport
and distribute Colony to the Japanese market at
a low er cost than the Japanese b rew ery can up
to the quantity Q*. Notice that the additional
costs o f transportation and distribution have
low ered the quantity at which the U.S. b rew ery
can compete from Q J3 to Q*. At quantities
beyond Q*, the Japanese firm can produce and
distribute Colony fo r less than C3, giving it a
cost advantage over the U.S. b rew er.2
Because o f the cost advantage, fo r any given
quantity o f Colony demanded in Japan up to Q*,
the U.S. firm would p refer to produce Colony
domestically and export the product to Japan. If
the quantity o f Colony demanded in Japan w ere
greater than Q*, the U.S. firm w ould either at­
tempt to negotiate a licensing agreement with
the Japanese b rew er or purchase or build a
Japanese b rew ery fo r production o f Colony.
W hether the U.S. b rew er w ould choose to
license production or open a branch b rew ery in
Japan w ould depend on several factors. Horstmann and Markusen (1987) note that if the li­

censee and the plant to be built or purchased
are equally efficient, then the need to give the
licensee the incentive to maintain the reputation
o f the licenser’s product w ill result in FDI
always dominating the use o f licensing. They
also conclude, how ever, that if the licensee and
branch plant are not equally efficient—that is, if
their LRAC curves are not identical—then other
factors such as the size o f the market, the ex­
istence o f close substitutes in the target market
and the level o f interest rates in the tw o coun­
tries will determine w hether licensing agree­
ments or FDI w ill be used.
In addition to production and distribution costs,
brew ers also face ta riff and non-tariff trade bar­
riers, which raise the cost o f supplying a coun­
try with beer. In terms o f figure B4, it is con­
ceivable that the U.S. b rew ery could have an
average cost o f production considerably below
brew eries in Japan, but that trade barriers in
Japan are so high that licensing agreements or
foreign direct investment becom e the preferred
method o f supplying the foreign country at all
levels o f demand. Here, the LR A C *s+ 1us curve
can be used to incorporate this idea. Let the t us
variable now stand fo r transportation and
distribution as w ell as costs associated with
trade barriers, such as tariffs. The existence o f
trade barriers simply shifts the U.S. b re w e r’s
export cost curve upward, pushing Q* closer to
the origin.
Other realistic problems associated with inter­
national transactions have been ignored in this
example. Some o f the other factors that would
affect how a firm supplies a foreign market in­
clude differences in production technology and
input costs, governm ent restrictions on foreign
investment, costs o f negotiating and monitoring
licensing agreements, exchange-rate movements
and the role other products being produced at
the brew eries might have on the plant’s cost o f
production.

2Of course, the Japanese firm will eventually reach its
points of diseconomies of scale and its average cost of
production will rise above Cr



NOVEMBER/DECEMBER 1990

20

Michael T. Belongia and
James A. Chalfant

Michael T. Belongia is an assistant vice president at the
Federal Reserve Bank of St. Louis. James A. Chalfant is an
associate professor of agricultural and resource economics at
the University of California-Berkeley. Scott Leitz, Lora Holman
and Lynn Dietrich provided research assistance.

Alternative Measures of
Money as Indicators of Infla­
tion: A Survey and Some New
Evidence

I n OCTOBER 1982, the Federal Open Market
Committee (FOMC) officially de-emphasized M l
as an interm ediate target variable in conducting
m onetary policy. Th e FOMC further reduced
the emphasis given to M l in 1987 by not setting
target ranges fo r its rate o f growth. The typical
explanations given fo r this de-emphasis have
stressed a breakdow n in the 1980s in the rela­
tionships betw een M l grow th and both the rate
o f inflation and the grow th rate o f nominal GNP
that had characterized the post-war period.1
The de-emphasis o f M l as a policy guide has
prom pted many observers to seek a replace­
ment. Some analysts have argued that financial

1Specifically, in the postwar period through 1981, M1
growth was related closely to nominal GNP growth over a
period of one year and to the rate of inflation over a period
of three to five years. The breakdown in these relation­
ships also has been discussed as a decline in the growth
rate of M1 velocity. See Stone and Thornton (1987), for ex­
ample, for a survey of explanations that have been offered
for this abrupt shift in what had been quite stable relation­
ships. Tatom (1988) and others, in contrast, have disputed

http://fraser.stlouisfed.org/
FEDERAL RESERVE BANK OF ST. LOUIS
Federal Reserve Bank of St. Louis

innovations and deregulation have altered the
relationship betw een interest-bearing and noninterest-bearing assets. One group o f individuals,
fo r example, has argued that the interest-bearing
components o f M2 are good substitutes fo r the
narrow er set o f assets in M l and, on this basis,
concluded that M2 now is the best m onetary ag­
gregate fo r policy actions. A nother group, rea­
soning that interest-bearing accounts act m ore
as savings balances than transaction accounts,
has advocated the narrow measure o f M IA (cur­
rency plus non-interest-bearing checkable depos­
its).2 Still others have advocated the use o f
weighted monetary aggregates, arguing that the
individual assets included in M l, M2 and the

whether historical relationships between M1 and nominal
variables were fundamentally different in the 1980s.
2For the arguments favoring M2, see Hetzel (1988). For the
evidence favoring M1A, see Hafer (1984) and Darby, et al.
(1989).

21
broader aggregates possess different degrees o f
"moneyness” according to the interest they pay.
If these assets have different degrees o f m oney­
ness, they should be weighted differently when
added together.3
This article focuses on research that addresses
both the composition and w eighting issues in
deriving a new monetary aggregate from the
first principles o f demand analysis.4 It surveys
this w ork and evaluates the perform ance o f tw o
new money stock measures as indicators o f
future inflation.

W H Y IS THE DEFINITION OF
“MONEY” IMPORTANT?
W hat truly constitutes "m oney” in the real
w orld is an interesting and important issue
because, as economic theory predicts, money
has pow erfu l effects on economic activity. Un­
fortunately, theory alone does not o ffe r explicit
guidance in choosing a unique and unambig­
uous real w orld counterpart to the abstract
theoretical concept o f "m oney.” W ithout clear
theoretical guidance, many economists have
adopted the conclusion that the appropriate
definition o f money is the one that serves best
in the chosen empirical application.5
One theoretical fram ew ork that attributes an
important causal role to m oney in determining
fluctuations in economic activity is the Quantity
Th eory o f Money. This theory is based on the
Equation o f Exchange, which typically is w rit­
ten as:
(1) M V = PT,
w h ere M is the quantity o f money, V is its
velocity o f circulation, P is the average price
level and T is the volume o f transactions. The
Quantity Th eory derived from this equation
predicts that the price level varies directly with
the quantity o f money. In its strictest form,
which has been subjected to empirical testing,
the Quantity Th eory predicts that the causation
runs from changes in m oney to changes in
prices (but not the reverse) and that, accounting

3See, for example, Barnett (1983), Barnett et al. (1984) or
Spindt (1985).
4See Belongia and Chalfant (1989) and Swofford and
Whitney (1990a,b).
5See Friedman and Schwartz (1970), p. 91.



fo r the short-run influences o f other variables,
changes in m oney and prices are proportional.6
Obviously, the definition o f m oney is im por­
tant in testing such propositions. The definition
o f m oney is also important fo r m onetary policy
purposes. Assume, fo r example, that a central
bank believes the implications o f the Quantity
Theory. It might then direct its policies to
achieve a grow th rate o f m oney consistent with
zero inflation. This raises the issue o f how to
measure m oney in the real w orld. As Laidler
(1989) points out, both critics and advocates o f
the Quantity Th eory frequently disagree about
the appropriate definition o f m oney to test its
implications. Thus, a central bank committed to
achieving a policy goal fo r "P ” still faces con­
siderable uncertainty about which "M ” should
be targeted and how to do so.

FINANCIAL INNOVATIONS AND
THE CONDUCT OF MONETARY
POLICY
For m ore than 30 years after W orld W a r II,
the trend grow th rate o f M l velocity was nearly
constant at 3 percent. For many economists,
this stability made M l grow th a reliable leading
indicator o f the effects o f current m onetary ac­
tions on future econom ic activity. As figure 1
shows, how ever, this seemingly stable trend
was disrupted sharply in 1981, perhaps because
o f the nationwide introduction o f interestbearing checkable deposits in January o f that
year. Because these deposits not only paid rates
o f interest that w e re close to those on savings
deposits but also allowed depositors to w rite
checks directly on these deposits, the hypothesis
was that at least some portion o f NO W account
balances w ere savings. If so, these new ac­
counts combined “ idle” savings balances with
"active” transaction balances in traditional de­
mand deposits. The larger quantity o f balances
in M l accounts, fo r the same level o f GNP,
w ould be reflected in a fall in the grow th rate
o f M l velocity.7 M ore important, uncertainty
about the effects o f financial innovations on M l,

6See Laidler (1989) for a thorough review of the Quantity
Theory’s history, implications and controversies.
7See Gordon (1984), for example, for more discussion of
this argument.

NOVEMBER/DECEMBER 1990

22

Figure 1
Log of M1 Velocity1

'Velocity data are indexed to a base value of 100 in 1/1963.
in addition to velocity growth, weakened M l ’s
informational content in predicting future price
movements.8
This uncertainty has been an important ele­
ment o f m onetary policy discussions. The record
o f the FOMC, fo r example, shows that, at the
meeting o f February 8-9, 1983:
Committee members’ views varied considerably
on the weight to attach to M l ... the perfor­
mance of that aggregate had been subject over
the past year or more to substantial uncertain­
ties related to the growing role of NOW ac­
counts and an apparent shift in the behavior of
its income velocity ... Only modest allowance
was made for the possibility that the new Super
NOW accounts would draw funds into M l from
other sources.9

8Monetary control also will be impaired if shifts of savingstype balances into interest-bearing checkables weakens
the link between the monetary base and M1.

http://fraser.stlouisfed.org/
FEDERAL RESERVE BANK OF ST. LOUIS
Federal Reserve Bank of St. Louis

T w o months later, in a statement to Congress,
Federal Reserve Chairman Paul Volcker reported:
To some extent — but it cannot be measured
with any degree of certainty — the decreases in
“velocity” may reflect the changing nature of
M l; with interest-bearing NOW and Super
NOW accounts making up an increasingly large
proportion of Ml, this aggregate may be in­
fluenced by "savings" behavior as well as "tran­
sactions” motives. That is a longer-term factor,
and the growth in M l over the shorter-run
may have been affected by the reduced level of
market rates — particularly relative to interestbearing NOW accounts — and slowing inflation
as well. The range o f uncertainty on these points
is substantial (emphasis added) and has led the
Federal Open Market Committee to place less
emphasis on M l in the implementation of
policy over the short term.10

9Federal Reserve Bulletin, April 1983, pp. 288-289
10Federal Reserve Bulletin, May 1983, p. 338.

23

Nonetheless, after being “ m onitored” but not
targeted in 1983, M l was chosen as an interm e­
diate target o f m onetary policy from 1984 to
1986. In 1987, the FOMC again declined to set
target ranges fo r M l growth, instead choosing
to emphasize M2 and M3 in policy discussions.11
These official statements point to considerable
uncertainty about how interest-bearing checking
accounts w ere being used and how their use
might affect the behavior o f the m onetary ag­
gregates generally, the behavior o f M l specifi­
cally, and the relationship o f M l to important
economic variables, such as spending and prices.

APPROACHES TO THE
DERIVATION OF ALTERNATIVE
MONETARY AGGREGATES
Choosing a m onetary aggregate requires at
least four steps.12 First, the group o f assets to
be included in the aggregate must be selected.
Second, individual asset categories must be ag­
gregated appropriately. Third, the aggregate
must be analyzed in terms o f the central bank’s
ability to control it. Finally, its perform ance in
some ultimate use must som ehow be evaluated.
Before the 1980s, those w ho w ere skeptical o f
using the M l aggregate fo r policy analysis gen­
erally focused on the fourth point, the m ethod­
ology used to link M l either to nominal GNP
grow th or the inflation rate. Many critics, fo r
example, raised issues about the exogeneity o f
changes in the m oney stock, reverse causation
running from income or prices to m oney or
technical problems associated w ith estimation o f
reduced-form equations.13 W h ile such m ethodo­
logical criticisms continued throughout the
1980s, much m ore attention in recent years has
been focused on the aggregates themselves
rather than their final uses. The official money
stock measures particularly have been criticized
both fo r the w eighting scheme used to ag-

11For more detail on the uncertainty about the behavior of
different aggregates, see Hafer and Haslag (1988) or Gar­
finkel (1990).
12Barnett, et al. (forthcoming) argue that this process also
involves addressing a fifth step—monitoring how closely
the index number tracks the economic aggregator
function—that has been overlooked by previous research
in this area.
13See, for example, de Leeuw and Kalchbrenner (1969),
Davis (1969), Gramlich (1971), Goldfeld and Blinder (1972)
and B. Friedman (1977), among others.
14The full quote is, “ Money never bears interest except in
the sense of creating convenience in the process of ex­
change.” See Fisher (1963), p. 9.



gregate individual asset categories and fo r the
specific assets used to construct an aggregate.
These criticisms are discussed in m ore detail
below.

Composition o f Monetary
Aggregates
Historically, debates about w hat is m oney
have focused on the things—coins, bank notes,
etc.—that should be included in a m onetary ag­
gregate. Before the em ergence o f interestbearing checkable deposits, the logic behind the
assets included in M l coincided w ith Irving
Fisher’s dictum, “M oney never bears interest... ,"14
For economists o f this school, the m onetary ag­
gregate also was view ed as containing only
those assets that could be readily exchanged fo r
goods.15
Since the birth o f m oney market mutual
funds (MMMFs) and, even m ore so, the nation­
w ide introduction o f N O W accounts in January
1981, how ever, the question o f which assets are
correctly identified as m oney has becom e com ­
plicated . Many argue that the historical justifi­
cation fo r the group o f assets in M l is no longer
clear. What, fo r example, distinguishes the NOW
accounts in M l from MMDAs or MMMFs, which
appear to have characteristics similar to NOWs
but are excluded from M l? Conversely, because
N O W accounts pay interest and can be used as
savings accounts, should they be excluded from
a m onetary aggregate designed to include those
assets that can be readily exchanged fo r goods?

Tests f o r Weak Separability and
Asset Selection
One approach to deciding which financial
assets should be considered as a group fo r ag-

15lbid., p. 8.

NOVEMBER/DECEMBER 1990

24

gregation treats financial assets as commodities
that are held fo r the services they provide. In
this approach, individuals are assumed to
allocate their financial wealth across the spec­
trum o f available assets, according to their
preferences fo r the characteristics o f each asset
and the relative returns each asset yields.
Treating financial assets in this manner leads to
inserting them as arguments in the utility func­
tion o f the representative individual and analyz­
ing the demand fo r financial assets in much the
same w ay that the demand fo r other com ­
modities can be studied.16
This approach leads to a search fo r “ separa­
ble” com m odity groups. Separability is related to
the notion o f multi-stage budgeting in which
consumers first allocate their budget across
broad expenditure categories such as “ shelter,”
"transportation,” “ clothing” and “food.” Then
after allocating, say, 20 percent o f their total ex­
penditures to food, further allocations are made
within the food group as expenditures are bud­
geted fo r "m eat,” "dairy products,” “ fruits and
vegetables” and so on. The process continues
w ith the gross expenditures fo r meat allocated
across beef, pork and chicken. The result o f this
process can be pictured as a branching tree
diagram in w hich broad commodity groups are
continually divided into smaller, m ore specific
com m odity groups.
T o decide w hich financial assets may be
grouped under the heading "m oney,” the re ­
search task is to find what goods are in the
“ m oney" group and what goods are unrelated to
utility from the services o f m onetary assets.
This identification can be done form ally by
testing w hether an asset collection is w eakly
separable from other assets and commodities.
Indeed, Barnett (1982, pp. 695-96), has shown
that a necessary condition fo r the existence o f a
m onetary aggregate is that the subset o f m one­
tary goods is w eakly separable from nonmone­
tary goods.17

16There has been a steady controversy about whether
money can be modeled as an argument in the utility (or
production) function; however, Feenstra (1986) has shown
that this approach is equivalent to other standard ap­
proaches, including the Clower (transactions) constraint;
see Clower (1970).
17For more detail on weak separability and other points
related to finding an admissible monetary aggregate, see
Varian (1982, 1983), Barnett (1987), Ishida (1984), Serletis
(1987), Fayyad (1986), Hancock (1987) and Belongia and
Chalfant (1989).

FEDERAL RESERVE


BANK OF ST. LOUIS

W eak separability, a result from demand
theory, implies that the consum er’s marginal
rate o f substitution betw een tw o assets in a
group is unaffected b y the quantities o f any
good not in the group. This ensures that changes
in the quantities consumed o f goods not in the
group w ill not affect the marginal utilities de­
rived from goods within the group. The practi­
cal implication o f w eak separability is that it
permits demand analysis that is concerned only
w ith a specific collection o f assets; it does not re ­
quire that one study all goods in the economy.
T o view the problem differently, let some
group o f assets, A, consist o f goods a l, a2 and
a3; that is, A = (ai, a2, a3). Assume, how ever,
that this group is not w eakly separable from
other assets; instead, the separable group also
should include asset a4. By arbitrarily cutting
o ff the group to include only al-a3, changes in
the quantity o f a4 consumed w ould affect the
marginal rates o f substitution w ithin the group
o f al-a3 and give rise to seemingly unstable
preferences fo r asset holdings or, from a p ro­
duction function point o f view, an unstable
transactions technology. In a m ore concrete
sense related to traditional empirical w ork, the
omitted variable a4 w ould affect an estimated
demand function fo r an aggregate variable con­
structed from assets al-a3 b y acting as an om it­
ted shift variable. For example, estimating a
m oney demand function based on M l balances
when MM DAs should be part o f this group
could be the source o f instability in such a de­
mand function.18
Recent w ork by S w offord and W hitney (1987,
1988) and by Belongia and Chalfant (1989) has
examined the conditions o f w eak separability
fo r the asset groups that correspond to M l and
M IA (M l minus interest-bearing checkable de­
posits), as w ell as broader collections o f assets
that include m ore components o f M2. This re ­
search checked fo r the existence o f a well-behaved utility function that was consistent with

18ln fact, studies of money demand functions often include a
dummy variable around 1981-82 to capture the effects of
financial innovation. Aggregates based on weakly sep­
arable groups, however, should be stable and eliminate
the need to account for shifts in functional relationships
with dummy variables.

25

the financial balances held in various form s at
existing prices and levels o f income. Upon finding
that the data w ere consistent w ith demand
theory, the separability tests w ere then used to
identify the specific financial assets, in combin­
ation w ith currency, that w ere w eakly separable
from other financial assets in the utility func­
tion. These w eakly separable asset groups w ere
then candidates fo r aggregation.

Weighting o f Asset Components
Another criticism o f M l as an indicator o f
m onetary actions—specifically w hen examining
w h y the financial innovations o f the early 1980s
might be associated with the sharp slowing in
M l velocity g row th —focuses on the potential e r­
rors associated w ith the simple-sum weighting
scheme used to derive M l and other official
m oney stock measures. The simple-sum aggrega­
tion scheme, which gives the same w eight to
currency as it does to dollars deposited in ac­
counts with quite different characteristics,
violates fundamental aggregation principles, so
the argument goes.
For example, a dollar o f currency (which is
term ed pure m oney because it pays no explicit
rate o f return) is weighted equally in M l w ith a
dollar in a N O W account deposit. This weighting
occurs even though the N O W account pays a
market rate o f return that, in recent years, has
approximated the return on savings deposits.
Thus, even though some N O W account balances
may be held as savings, they are included dollar
fo r dollar in M l w ith the other assets that tradi­
tionally have been considered to be transactions
balances.19 Such equal weights are justified only
if the assets are perfect substitutes fo r each
other; both the raw data as w ell as much em­
pirical w ork demonstrate that these conditions
do not hold.20

19Some analysts, in fact, have attempted to infer the proportion of NOW balances held for that asset’s savings
characteristics (its interest rate and deposit insurance) on
the basis of turnover data relative to demand deposit turn­
over rates. See Spindt (1985).
20See Barnett (1982) for the technical arguments about the
shortcomings of simple-sum weighting. Ewis and Fisher
(1984, 1985) and references they cite are sources that find
low elasticities of substitution between M1 components
and near-monies.
21See, for example, Friedman and Schwartz (1970),
pp. 151-52.



The fundamental point here is that each
dollar in a N O W account should receive less
w eight than a dollar o f currency, which p re­
sumably is held purely fo r the conduct o f trans­
actions. Similarly, if the assets collected in M2
are found to be a w eakly separable group, each
o f its components should be w eighted different­
ly w ith the weights declining as own-rates o f in­
terest rise. A t this point, it should be noted
that, w hile simple-sum weighting is clearly
w ron g in principle, it may not in the past have
m attered empirically; that is, it may not have
produced major distortions in the estimated
relationships betw een "m oney” and other vari­
ables. It also should be noted that the interest
in w eighted m onetary aggregates is not new;
the deficiencies o f simple-sum aggregates w ere
recognized m ore than 20 years ago.21
By form alizing this intuitive argument, Barnett
and others constructed a series o f Divisia m one­
tary aggregates.22 A Divisia index, w hich is
m erely one o f many statistical index number fo r ­
mulas that have desirable aggregation prop er­
ties, has an advantage over simple-sum aggrega­
tion in that it internalizes pure substitution e f­
fects. This means that the measured value o f
the index w ill not change unless the utility or
production functions underlying the index actu­
ally produce a different level o f utility or output.
In contrast, the failure o f simple-sum weighting
to internalize substitution effects means it can
change even w hen there has been no change in
the flo w o f transaction services from the group
o f assets.
Unlike the simple-sum weighting scheme cur­
rently used, any good index gives different
weights to balances held in different deposit
categories. In the case o f the Divisia index ap­
plied in m onetary studies, the weights are based
essentially on the difference betw een the rate o f
return on a pure store o f value (typically, a
long-term bond) and each asset’s ow n rate o f

22See Barnett (1983, 1984) and Barnett, Offenbacher and
Spindt (1981, 1984).

NOVEMBER/DECEMBER 1990

26

return.23 The larger this interest rate differen ­
tial, the higher the opportunity cost o f holding a
particular financial asset and the greater its
"moneyness.” The largest interest differential
and, therefore, the greatest "moneyness” occurs
fo r currency, whose rate o f return is zero.
The research cited above has put the defini­
tion and measurement o f m onetary aggregates
on firm er economic and statistical footing; the
specific aggregates it has produced, how ever,
have been subject to criticism. One criticism has
been that w eighting may be, or at least may
have been, unimportant empirically. Specifically,
the relative perform ance o f simple-sum vs.
Divisia measures against various final uses cri­
teria has been disputed.24 Perhaps the most im ­
portant criticism, how ever, is that this research
has failed to analyze w hether the re-weighted
versions o f the official m onetary aggregates,
M l, M2, M3 and L, represent asset groupings
that can or should be aggregated. The shaded
insert at right illustrates how these issues can
affect the behavior o f a m onetary aggregate.

23A Divisia monetary aggregate is constructed in the follow­
ing manner: Let q„ and plt represent the quantities and
user costs of each asset to be included in the aggregate
at time t. The expenditure share on the services of
monetary asset i in period t is:
8 _

p,q,

The user cost of each asset is measured as:
P„ =

P ;(R ,-r„)(1 -M )
1 + R,(1 - M)

where P*is the geometric mean of the CPI and GNP
deflator, R, is the maximum available expected holding
period yield, rit is the observed or imputed nominal own
rate of return on asset i and M is the average marginal tax
rate [see Barnett (1978)]. The growth rate of a Divisia ag­
gregate then can be written as
G(Q,) = j I 1 s* G(qJ,
where s* = 1/2(s„ + slit.,), and n is the number of assets
in the aggregate. For a more detailed explanation of this
weighting procedure and its application to the Divisia ag­
gregates, see Barnett (1978) and Barnett et al. (1984).
24Barnett (1984), Serletis (1986), Hancock (1987) and Ishida
(1984) find the Divisia aggregates to be superior to simplesum measures in a variety of final uses (e.g., stability of
velocity, relationship with nominal spending or prices). Bat­
ten and Thornton (1985), against the standard of perfor­
mance in nominal spending equations, find little difference
between the two.

FEDERAL RESERVE BANK OF ST. LOUIS


DIVISIA M IA AND THE MONEY
METRIC INDEX
Previous research that has tested fo r weakly
separable asset groups has suggested tw o
w eighted m onetary aggregates as possible im­
provem ents over traditional simple-sum mea­
sures. These new aggregates are Divisia M IA
and the M oney M etric Index (M M I).25 The fo r­
m er includes only currency and traditional de­
mand deposit balances, which are w eighted by
their user costs to construct a Divisia aggregate.26
The latter includes a w id er range o f assets: cur­
rency, demand deposits, other (interest-bearing)
checkable deposits, savings deposits and o v er­
night repurchase agreements. The interested
reader is directed to the original sources fo r
detail on the testing strategies that led to the
choices o f these asset groups.
The velocities fo r these tw o measures and M l
are shown in figure 2. Perhaps the most in­
teresting aspect o f figure 2 is that, in the early
1980s, w hen M l velocity began its heralded

25The more narrow measure, Divisia M1A, is derived in
Belongia and Chalfant (1989). The broader measure, the
Money Metric Index (MMI), is discussed in Swofford and
Whitney (1990b). Both groupings were determined by nonparametric tests for weak separability described in Varian
(1982). While parametric tests for weak separability have
been reported by Serletis (1986) and Hancock (1987),
Swofford and Whitney (1986) have noted that such tests
are sensitive to the choice of the functional form for the
utility or production function.
26The user cost formula is derived in Barnett (1978). We use
a simplified formula:
user cost =-p :(1r ,+ -R.r„)
Although the return on something completely illiquid, such
as the return on human capital, would be the preferred
choice to represent Rt, practical measurement issues have
led to the use of long bond rates in empirical studies.
The approach used by Swofford and Whitney (1990a)
actually calculates upper and lower bounds for the index
rather than a unique value for each time period. To avoid
any problems associated with choosing an incorrect value
of their index, we simply used the Divisia weighting
scheme to aggregate the asset collection they identified.

27

Calculating A Divisia Monetary Aggregate
Th e example below illustrates the conse­
quences o f using a m onetary aggregate that
is composed o f the "incorrect” assets or that
employs simple-sum weighting o f its constitu­
ent assets even if they are the proper ones to
use. For simplicity, w e assume that only four
asset categories exist: currency, traditional
(non-interest-bearing) demand deposits, NO W
accounts (interest-bearing checkable deposits)
and savings deposits. W e then define M l as
the sum o f C + DD + NOWs and M2 = M l
+ Savings. From these definitions, w e con­
struct four aggregates: simple-sum M l and
M2 and w eighted versions o f M l and M2.
For purposes o f these examples, w e start
w ith the follow ing initial conditions:
Dollar Balances in Each Asset Category
C = 100
DD = 50
NOW s = 50;
Savings = 50;
M l = 200
M2 = 200 + 50 = 250
Interest Rates on Each Asset
Benchmark rate (R) = 0.1; rc = 0; rDD = 0.03;
^now = 0.05, rSAV = 0.05.

A simplified expression fo r the user cost
formula, cited in footnote 26, is:
D _p
plt = _ i___ ii. Thus, the user cost fo r each
1 + R,
asset is:
1 -0
1
C: — -----— = — = .0909
1 + .1
1.1

1.1

0.0636

NOWs: 1.~ ....... = 0.04545
1.1
SAV:

1 ~

DD: .0636(50) + 14.55 = 21.9
NOWs: .04545(50) + 14.55 = 15.6

Similarly, fo r M2 assets:
C: .0909(100) + 16.82 = 54.1
DD: .0636(50) + 16.82 = 18.9
NOWs: .04545(50) + 16.82 = 13.5
SAV: .04545(50) - 16.82 = 13.5
The divisors in each column, 14.55 and 16.82,
are the sums o f expenditures on all assets in
that group. W ith this information, w e can il­
lustrate how the tw o issues o f c o n c e r n w eighting and composition—affect the behav­
ior o f each aggregate.
An Illustration: Assume, fo r quarterly data,
$10 is shifted from savings to demand de­
posits. For the simple-sum aggregates, the
dollar changes in levels and corresponding
grow th rates are:

M l:

User Costs

1 — ns
DD: —----- ^ =

C: .0909(100) -*- 14.55 = 62.5

= 0.04545

1.1

100 + (50 + 10) + 50
= 210

Change

Grow th
rate

+10

21.55%

M2:
100 + (50+10) + 50
+ (50-10) = 250

0

0%

Weighted M l : The grow th rate o f a Divisia in­
dex is constructed as the grow th rate o f each
asset category w eighted by the average shares
o f the tw o periods.1 N ew budget shares at
quarter t + 1 are 59.9 (C), 25.1 (DD) and 15.0
(NOWs). A verage shares (S*) fo r the tw o per­
iods are 61.2 (C), 23.5 (DD) and 15.3 (NOWs).
This yields:
0.612»(0) + 0.235»(grow th rate o f DDs)
+ 0.153*(0)
= 0.235 [(| Q )4 - 1]* 100

Shares
Expenditure shares are calculated as
s.t = _Pit9h_
SPi.qn

xhus, fo r M l assets:

= 0.235*(1.0736)* 100 = 25.23%
Weighted M2: For this aggregate, the budget
share o f DDs rises while that o f savings de­

i




NOVEMBER/DECEMBER 1990

28

clines. Using the average budget shares across
the tw o periods and recognizing that the
changes in C and NOWs both equal zero, w e
have:

0 + .2069 » (107.36) + 0 + .1211* (-59.04)
= 22.22 + (-7.15) = 15.07%

Thus, a simple $10 shift from one category to
another produces growth rates ranging from

'The weighted measures are indexes set equal to 100 in
the same base year as opposed to M1 and M2, which
are measured in dollars. Thus, comparisons of changes
in levels are uninformative.
2ldeally, we would like to isolate the effect of an interest
rate change alone, but this is not done for two reasons.

zero (simple-sum M2) to 25.23 percent (Divisia
M l).
The intuition behind these results is that $10
has been shifted from one deposit category
(savings) to another with greater "moneyness”
because demand deposits have a higher user
cost, i.e. their ow n rate o f return is lower.
Thus, fo r example, Divisia M2 will show growth
while simple-sum M2 will not because $10,
while still in M2, has been shifted into a cate­
gory that receives a larger weight when its
growth rate is calculated.2

First, economic theory predicts changes in quantities
demanded or supplied in response to changes in
relative prices. Second, as can be seen from the
calculations for the growth rate of a Divisia index, the
index will not change unless quantities change.

Figure 2
Logs of Velocity of M1, Divisia M1A, and MMM

’ Velocity data are indexed to a base value of 100 in 1/1963.


http://fraser.stlouisfed.org/
FEDERAL RESERVE BANK OF ST. LOUIS
Federal Reserve Bank of St. Louis

29

Figure 3
Growth Rates of M1, Divisia M1A, and MMI
Percent

Percent

-2 0
1964

66

68

70

72

74

76

78

80

82

84

86

1988

decline, the velocities o f both weighted aggre­
gates actually increased. Also interesting is the
sharp drop in MM I velocity around 1986, at a
time w hen M l velocity also fell sharply but
Divisia M IA velocity fell only moderately. Since
1986, all three appear to be on similar paths.

figure 3 indicates that the behavior o f these ag­
gregates was much m ore divergent throughout
the 1980s than in previous decades.

These trends are exhibited in different form
in figure 3, which depicts the grow th rate o f
each aggregate. Although the three series often
track closely, there are notable exceptions. For
example, the grow th o f M M I slows sharply in
1978-79 whereas the other measures continue
to g ro w at rates near 5 percent. In 1980-81,
both w eighted measures show sharply neg­
ative grow th rates w hile M l grow th continues
to be positive; this is particularly interesting
because the decline in M M I over the same per­
iod w ould refute the idea that this episode
m erely reflects shifts out o f narrow M IA bal­
ances into interest-bearing accounts. In general,

Long-Run Relationships betw een
M onetary Measures and the Price
Level




Th e empirical perform ance o f alternative
m onetary aggregates and their usefulness as
targets fo r m onetary policy might be evaluated
in a variety o f ways, fo r example, the use o f
St. Louis-type spending equations or short-run
m oney demand functions w ith tests o f co effi­
cient stability over time. Either approach is quite
sensitive to specification; fo r example, the
results appear to depend crucially on both the
sample period chosen and the lag lengths used

NOVEMBER/DECEMBER 1990

30

fo r all explanatory variables.27 M oney demand
equations also are sensitive to the specification
o f such things as the scale and interest rate
variables and the equations’ short-run dynamics.
Such short-run empirical controversies, how ever,
are irrelevant fo r our purposes in this study. As
Laidler (1990) has pointed out, many key discus­
sions o f m oney’s effect on economic activity
specifically emphasize long-run relationships,
testing hypotheses w ith annual or cycle average
data, and never argue that economists should
be able to m odel the short-run dynamics in­
herent in quarterly data.
T h erefore, w e w ant to focus on one desirable
characteristic o f any aggregate monetary vari­
able: stability in its long-run relationship w ith a
variable o f prim ary importance to m onetary
policy. If w e return to the equation o f exchange,
M V = PT, w e can see that long-run inferences
about future movements in the price level (P)
can be made from the behavior o f the appropri­
ate m oney stock (M) if w e have some notion
about the long-run behavior o f output (expres­
sed as the volum e o f transactions, T). For this
analytical fra m ew ork to be useful, the velocity
o f circulation (V), given some long-run forecast
o f output, must be stable. Indeed, this simple
fram ew ork fo r predicting the long-run behavior
o f inflation is the basis o f the w idely discussed
“ P *" analysis o f Hallman, Porter and Small (1989).
One w ay to investigate the stability o f velocity
is to test w hether its time series has a unit root.
I f a time series has a unit root, it is said to be
nonstationary; this means that “ shocks” to its
path may m ove the series permanently away
from its previous pattern. In this case, it may
w ander from w hat appeared to be its trend
without any force to push it back again. If
velocity w e re a time series that did not exhibit
some stable long-term behavior, the Quantity
Th eory fram ew ork w ould not provide accurate
predictions o f long-run inflation on the basis o f
m oney stock movements.
Velocity, o f course, is calculated as V = (PT)/M.
The velocity o f circulation was calculated fo r
Divisia M I A and MMI, and, fo r purposes o f
comparison, fo r previous (M l) and current (M2)
m onetary targets as well. The Augmented Dickey-

27See Friedman and Kuttner (1990) for evidence on the ef­
fects of a changing sample period and Thornton and Bat­
ten (1985) for a discussion of the sensitivity of such results
to lag length selection.

http://fraser.stlouisfed.org/
FEDERAL RESERVE BANK OF ST. LOUIS
Federal Reserve Bank of St. Louis

Fuller (ADF) test fo r a unit root then was ap­
plied to each velocity series b y estimating a
regression o f the form :
(1) AX, = a + bAX,_, + cAX,_2 + dX,., + £„
w h ere X, is the natural logarithm o f the vari­
able, and A is the difference operator.28 The
ADF test examines w hether the coefficient, d,
on the lagged level o f X is significantly different
from zero. If it is not, then w e cannot reject the
null hypothesis that a unit root is present in the
data series. Although the estimated t-statistic
from this regression can be used in the ADF
test, the critical values must be m odified from
those o f the standard t-distribution. Because o f
the sample sizes in this study and because the
null hypothesis is that a unit root is present in
the data, the unit root hypothesis w ill be re­
jected fo r calculated t-values that are negative
and exceed 2.90 in absolute value.
If the unit root hypothesis is not rejected,
equation 1 is re-estimated after including a
trend term; in this case the critical value fo r the
coefficient d becomes -3 .5 and the null hypoth­
esis o f a unit root w ill be rejected by a test
statistic that is a larger negative number. This
test is conducted because a time series may be
stationary around a deterministic trend. Finally,
because it has been alleged that data from the
1960s dominate relationships betw een m oney
and other variables or that data from the 1980s
distort the measurement o f M l and M2, the
ADF tests w ere conducted over three sample
periods: I/1963-IV/1980; I/1970-IV/1989; and
I/1963-IV/1989.
Table 1A reports the ADF test statistics fo r
the logarithms o f each o f the fou r m onetary vel­
ocities during the three sample periods; the r u
and r T columns refer to the cases without and
w ith trend, respectively. As the results indicate,
no velocity measure is stationary in the log
levels even after accounting fo r trend. Thus, w e
difference the data again and estimate a regres­
sion o f the form:
(2) A2X, = a + bA2X,_i + cA2X,_2 + dAX,., + e„
w h ere the test again is w hether the coefficient,
d, is significantly different from zero. As before,

28For more detail on the test, see, for example, Dickey, et
al. (1986), especially pages 16-17. An FPE criterion was
used to determine the number of lagged values for AX,_,

31

Table 1A
Augmented Dickey-Fuller Tests on Monetary Velocities
Monetary
variable

1/1963 - IW1980
t u

tt

M1
M2
Divisia M1A
MMI

0.49
-1.25
1.44
1.53

-1.19
-1.99
-0.32
-0.23

1/1970 - IV/1989
T,

-1.92
-2.33
-0.64
-0.91

Tt
-1.26
-2.33
-1.66
-1.46

1/1963 - IV/1989

T,
-1.88
-2.72
-0.34
-0.73

-1.05
-2.69
-1.79
-1.72

Table 1B
Augmented Dickey-Fuller Tests on Growth Rates of
Monetary Velocities
Monetary
variable

1/1963 - IV/1980
t u

tt

1/1970 - IV/1989
T»

Tt

M1
-5.84* -5.87*
-3.10* -3.29*
M2
-4.50* -4.51*
-4.34* -4.31*
Divisia M1A
-5.59* -5.96*
-4.55* -4.52*
MMI
-3.54* -3.86*
-3.01* -3.00*
NOTE: An asterisk denotes statistical significance at the 5 percent level.
if the null hypothesis is not rejected, equation 2
is re-estimated w ith a trend term added.
The results o f this estimation are reported in
table IB. In this case, the results again find
nothing to distinguish any o f the m onetary in­
dexes as indicators o f long-run trends in infla­
tion as each measure apparently exhibits a
stable grow th rate o f velocity in each sample
period. I f one raises the significance level o f the
ADF test statistic to the 0.01 level, how ever,
critical values fo r the x u and x T tests o f about
-3 .6 and -4.15 w ould indicate that only the
velocities o f M2 and Divisia M IA show the ex­
istence o f a single unit root over time. In con­
trast, doubts w ould be raised about the stability

29As was noted in the paper’s introduction, a monetary ag­
gregate also must be controllable by the central bank to
be useful as an intermediate target variable. Applying a
test used by Belongia and Chalfant (1989) to each of the
four aggregates used in this article produces the following
results:
(1) Div M1A, = -2.25 + 1.05 * AMB,; R2 = .34; DW = 1.53
(2.16) (7.49)
(2) M1, = -3.21 + 1.37 *AMB,; R2 = .56; DW = 1.39
(3.69) (11.72)
(3) M2, = 2.98 + 0.78 * AMB,; R2 = 0.23; DW = 0.86
(2.98) (5.76)
(4) MMI, = -4.79 + 1.38 * AMB,; R2 = 0.18; DW = 1.19
(2.32) (4.96)



1/1963 - IV/1989
T„

-3.86*
-5.31*
-5.38*
-3.62*

Tt

-4.08*
-5.29*
-5.36*
-3.61*

o f the grow th rate o f MM I velocity in each o f
the three sample periods and about M l velocity
at least in the 1970-89 interval; the stability o f
M l velocity over the entire 1963-89 sample
w ould be a borderline call.
The upshot o f these cursory results is that,
fo r the specific final use criterion chosen (sta­
tionary velocity), a badly designed m onetary ag­
gregate (M2) may perform better than one that
is measured in a manner consistent w ith both
economic and statistical theory (MMI). Converse­
ly, some aggregates that have considered the
question o f asset collection and weighting (Divi­
sia M IA ) perform better than some o f the o f­
ficial aggregates (M l).29

The intuition of this test is to look for a close, contemporaneous
relationship between the instrument of monetary control (the
monetary base-AMB) and the target. By this standard, Divisia
M1A and simple-sum M1 both appear to be controllable.
Although the coefficient for AMB is statistically different from
zero for M2 and MMI, the fits (R2 = 0.23 and 0.18) are poor
and the low DW statistics suggest that other variables may
have important influences on this relationship. Overall, these
results do not make a strong case for the controllability of
either M2 or MMI.

NOVEMBER/DECEMBER 1990

32

SUMMARY
The breakdow n in the 1980s o f historical rela­
tionships betw een the grow th rates o f M l and
both the price level and nominal income has led
researchers to look fo r explanations o f these
changes. Some have argued that the historical
relationships w ere spurious and that the cur­
rent experience demonstrates the true relation­
ship. Others have argued that the traditional
relationships still hold after m inor revisions to
the estimating equations. Still others have used
indirect evidence to make adjustments to the o f­
ficial measures o f various m onetary aggregates.
None o f these efforts has been com pletely suc­
cessful in resolving what has been term ed “the
velocity problem ” .
This article has review ed still another ap­
proach to this problem: constructing m onetary
aggregates from first principles, using economic
and index num ber theory to determine which
assets should be included in an aggregate as
w ell as how they should be w eighted in ag­
gregation. The tw o measures discussed, Divisia
M IA and the M oney M etric Index, exhibited
quite d ifferen t velocity behaviors; Divisia M IA
velocity is consistent w ith a stable long-run rela­
tionship w ith the aggregate price level w hile
M M I velocity appears not to be. Similarly, M2
velocity appears to be stable over time whereas
M l velocity does not. If price stability is an im­
portant objective o f m onetary policy, monitoring
the grow th rates o f certain m onetary aggregates
can be a viable approach to policymaking. At
the same time, there is still no definitive answer
in terms o f all its final uses to the question:
W hat is money?

REFERENCES

Afriat, S. N. “The Construction of Utility Functions from Ex­
penditure Data,” International Economic Review (1967),
pp. 67-77.
Barnett, William A. “The User Cost of Money,” Economics
Letters, Vol. 1 (1978), pp. 145-49.
_______ . “ Economic Monetary Aggregates: An Application
of Index Number and Aggregation Theory,” Journal of
Econometrics (September 1980), pp. 11-48.
_______ . Consumer Demand and Labor Supply: Goods,
Monetary Assets and Time (Amsterdam: North Holland,
1981).
_______ . “ The Optimal Level of Monetary Aggregation,”
Journal of Money, Credit and Banking (November 1982,
part 2), pp. 687-710.

http://fraser.stlouisfed.org/
FEDERAL RESERVE BANK OF ST. LOUIS
Federal Reserve Bank of St. Louis

_______ . “ New Indices of Money Supply and the Flexible
Laurent Demand System,” Journal of Business and
Economic Statistics (January 1983), pp. 7-23.
_______ . “ Recent Monetary Policy and the Divisia Monetary
Aggregates,” The American Statistician (August 1984), pp.
165-72.
_______ . “The Microeconomic Theory of Monetary Ag­
gregation,” in William A. Barnett and Kenneth J. Singleton,
eds., New Approaches to Monetary Economics (London:
Cambridge University Press, 1987).
_______ . “ Developments in Monetary Aggregation Theory,”
Journal of Policy Modeling (Summer 1990), pp. 205-58.
Barnett, William A., and Seungmook Choi. “A Monte Carlo
Study of Tests of Blockwise Weak Separability,” Mimeo,
Department of Economics, University of Texas at Austin,
December 1986.
Barnett, William A., Melvin Hinich and Piyu Yue. “ Monetary
Policy with the Exact Rational Expectations Monetary Ag­
gregates,” in, Michael T. Belongia, ed. Monetary Policy on
the 75th Anniversary of the Federal Reserve System (Kluwer
Academic Publishers, forthcoming).
Barnett, William A., Edward Offenbacher and Paul Spindt.
“ New Concepts of Aggregated Money,” Journal of Finance
(May 1981), pp. 497-505.
_______ . “The New Divisia Monetary Aggregates,” Journal
of Political Economy (December 1984), pp. 1049-85.
Barnett, William A., and Paul A. Spindt. Divisia Monetary Ag­
gregates: Compilation, Data and Historical Behavior, Staff
Study #116, Board of Governors of the Federal Reserve
System, May 1982.
Batten, Dallas S., and Daniel L. Thornton. “Are Weighted
Monetary Aggregates Better than Simple-Sum M1?” this
Review (June/July 1985), pp. 29-40.
Belongia, Michael T., and James A. Chalfant. “ The Chang­
ing Empirical Definition of Money: Some Estimates from a
Model of the Demand for Money Substitutes,” Journal of
Political Economy (April 1989), pp. 387-97.
Clower, Robert W. “ Is There an Optimal Money Supply?”
Journal of Finance (May 1970), pp. 425-33.
Darby, Michael R., Angelo R. Mascara and Michael L.
Marlow. “ The Empirical Reliability of Monetary Aggregates
as Indicators: 1983-87,” Economic Inquiry (October 1989),
pp. 555-85.
Davis, Richard G. “ How Much Does Money Matter? A Look
at Some Recent Evidence,” Federal Reserve Bank of New
York Monthly Review (June 1969), pp. 119-31.
Deaton, Angus, and John Muellbauer. Economics and Con­
sumer Behavior (London: Cambridge University Press,
1980).
de Leeuw, Frank, and John Kalchbrenner. “ Monetary and
Fiscal Actions: A Test of Their Relative Importance in
Economic Stabilization—Comment,” this Review (April
1969), pp. 6-11.
Dickey, David A., William R. Bell and Robert B. Miller. “ Unit
Roots in Time Series Models: Tests and Implications,” The
American Statistician (February 1986), pp. 12-26.
Ewis, Nabil A., and Douglas Fisher. “ The Translog Utility
Function and the Demand for Money in the United States,”
Journal of Money, Credit, and Banking (February 1984),
pp. 34-52.

33

_______ . “Toward a Consistent Estimate of the Substitut­
ability Between Money and Near Monies,” Journal of
Macroeconomics (Spring 1985), pp. 151-74.
Fayyad, Salam K. “A Microeconomic System-Wide Approach
to the Estimation of the Demand for Money,” this Review,
(August/September 1986), pp. 22-33.
Feenstra, Robert C. “ Functional Equivalence Between Li­
quidity Costs and the Utility of Money,” Journal of Monetary
Economics (March 1986), pp. 271-91.
Fisher, Irving. The Purchasing Power of Money, Reprints of
Economic Classics (Augustus M. Kelley, 1963).
Friedman, Benjamin M. “ Even the St. Louis Model Now
Believes in Fiscal Policy,” Journal of Money, Credit, and
Banking (May 1977), pp. 365-67.
Friedman, Benjamin M., and Kenneth N. Kuttner. “ Money,
Income, Prices and Interest Rates After the 1980s,” Work­
ing Paper, WP-90-11 Federal Reserve Bank of Chicago
(July, 1990).
Friedman, Milton, and Anna J. Schwartz. “ Money and
Business Cycles,” The Review of Economics and Statistics
(February 1963, Part 2), pp. 32-64.
_______ . Monetary Statistics of the United States (Columbia
University Press, 1970).
Garfinkel, Michelle R. “The FOMC in 1989: Walking a
Tightrope,” this Review (March/April 1990), pp. 17-35.
Goldfeld, Stephen M., and Alan S. Blinder. “Some Implica­
tions of Endogenous Stabilization Policy,” Brookings Papers
on Economic Activity (3:1972), pp. 585-640.
Gordon, Robert J. “The Short-Run Demand for Money: A Re­
consideration,” Journal of Money, Credit and Banking (Part
1, November 1984), pp. 403-34.
Gramlich, Edward M. “The Usefulness of Monetary and
Fiscal Policy as Discretionary Stabilization Tools,” Journal
of Money, Credit, and Banking (May 1971), pp. 506-32.
Hafer, R. W. “The Money-GNP Link: Assessing Alternative
Transactions Measures,” this Review (March 1984),
pp. 19-27.
Hafer, R. W., and Joseph H. Haslag. “The FOMC in 1987:
The Effects of a Falling Dollar and the Stock Market Col­
lapse,” this Review (March/April 1988), pp. 3-16.
Hallman, Jeffrey J., Richard D. Porter and David H. Small.
“ M2 Per Unit of Potential GNP as an Anchor for the Price
Level,” Staff Studies, No. 157, Board of Governors of the
Federal Reserve System, April 1989.
Hancock, Diana. “Aggregation of Monetary Goods: A Pro­
duction Model,” in William A. Barnett and Kenneth J.
Singleton, eds., New Approaches to Monetary Economics
(London: Cambridge University Press, 1987).
Hetzel, Robert L. “ M2 and Monetary Policy,” Federal Reserve
Bank of Richmond Economic Review (September/October
1988), pp. 14-29.
Ishida, Kazuhiko. “ Divisia Monetary Aggregates and Demand
for Money: A Japanese Case," Bank of Japan Monetary
and Economic Studies (June 1984), pp. 49-80.
Klein, Benjamin. “ Competitive Interest Payments on Bank
Deposits and the Long-Run Demand for Money,” American
Economic Review (December 1974), pp. 931-49.
Laidler, David. “The Legacy of the Monetarist Controversy,”
this Review (March/April 1990), pp. 49-64.




_______ . “ The Quantity Theory is Always and Everywhere
Controversial—Why?,” Seminar Paper 89-03, Centre for In­
ternational Economic Studies, University of Adelaide (July
1989).
Offenbacher, Edward K. The Substitutability of Monetary
Assets, unpublished Ph.D. thesis, University of Chicago,
1979.
_______ . “ Disaggregated Monetary Asset Demand
Systems: Estimation and an Application of the Preference
Independence Transformation,” Special Studies Paper No.
144, Board of Governors of the Federal Reserve System,
July 1980.
Serletis, Apostolos. “ Monetary Asset Separability Tests,” in
William A. Barnett and Kenneth J. Singleton, eds., New Ap­
proaches to Monetary Economics, (London: Cambridge
University Press, 1987).
Serletis, Apostolos, and A. Leslie Robb. “ Divisia Aggregation
and Substitutability Among Monetary Assets,” Journal of
Money, Credit, and Banking (November 1986), pp. 430-46.
Spindt, Paul A. “ Money is What Money Does: Monetary Ag­
gregation and The Equation of Exchange,” Journal of
Political Economy (February 1985), pp. 175-204.
Stone, Courtenay C., and Daniel L. Thornton. “ Solving the
1980s’ Velocity Puzzle: A Progress Report,” this Review
(August/September 1987), pp. 5-23.
Swofford, James L., and Gerald A. Whitney. “ Flexible Func­
tional Forms and the Utility Approach to the Demand for
Money: A Nonparametric Analysis,” Journal of Money,
Credit and Banking (August 1986), pp. 383-89.
_______ . “ Nonparametric Tests of Utility Maximization and
Weak Separability for Consumption, Leisure, and Money,”
Review of Economics and Statistics, (August 1987),
pp. 458-64.
_______ . “A Comparison of Nonparametric Tests of Weak
Separability for Annual and Quarterly Data on Consump­
tion, Leisure and Money,” Journal of Business and
Economic Statistics (April 1988), pp. 241-46.
_______. “ Bounding an Economic Monetary Aggregate
Under Nonhomothetic Preferences,” Journal of Business
and Economic Statistics (January 1990a), pp. 137-42.
_______ . “A Money Metric Index of an Economic Monetary
Aggregate,” Economic Inquiry (1990b), forthcoming.
Tatom, John A. “Are the Macroeconomic Effects of Oil-Price
Changes Symmetric?, “ in Karl Brunner and Allan H.
Meltzer, eds., Stabilization Policies and Labor Markets,
Carnegie-Rochester Conference Series on Public Policy,
(Spring 1988), pp. 325-68.
Thornton, Daniel L., and Dallas S. Batten. “ Lag-Length
Selection and Tests of Granger Causality Between Money
and Income,” Journal of Money, Credit, and Banking (May
1985), pp. 164-78.
Varian, Hal R. “The Nonparametric Approach to Demand
Analysis,” Econometrica (July 1982), pp. 945-74.
_______. “ Non-parametric Tests of Consumer Behaviour,”
Review of Economic Studies (January 1983), pp. 99-110.
Wenninger, John. “The M1-GNP Relationship: A Component
Approach,” Federal Reserve Bank of New York Quarterly
Review (Autumn 1984), pp. 6-15.

NOVEMBER/DECEMBER 1990

34

Alison Butler

Alison Butler is an economist at the Federal Reserve Bank of
St. Louis. Lora Holman provided research assistance.

The Trade-Related Aspects of
Intellectual Property
Rights: What Is At Stake?
The pirating o f U.S.-financed research and developm ent discourages
innovation, denies markets to Am erican exports, and threatens technological
progress. Protection o f intellectual property rights preserves Am erica’s
technological edge, which is a key to our continued international
competitiveness.”
—Clayton Yuetter,
U.S. trade representative, 1986

J L HE SENTIMENTS expressed by the form er
U.S. trade representative above could have easily
been made by trade representatives from any
industrialized country about their ow n country.
In the last fe w years, many industrial countries
have becom e increasingly concerned over the
lack o f international protection o f intellectual
property rights (IPR).
In a 1988 study, the U.S. International Trade
Commission (USITC) attempted to estimate the
econom ic effects o f intellectual property rights
infringem ent by foreign countries on U.S. firm s.1
In its survey, firm s estimated their losses at
$6.2 billion in exports sales in 1986. These losses
accounted fo r approximately 1.4 percent o f the
total export sales o f products which are covered
by some form o f intellectual property rights that
year. Firms estimated additional losses o f $1.8
billion in sales in 1986 due to imports that in­

1USITC (1988). For limitations of the survey and data, see
the study.
2See also Feinberg (1988).

FEDERAL RESERVE BANK OF ST. LOUIS


fringed on their domestic IPR. In a related study,
Feinberg and Rousslang (1990) estimate that the
ratio o f lost profits to w orld w id e sales o f firms
selected from the USITC study ranged from
0.05 percent fo r the extractive sector to 3.6 per­
cent fo r the entertainment sector.2 Since firms
incur research and developm ent (R&D) expenses
to develop intellectual property, these lost sales
and profits defer firm s from undertaking the
risky process o f developing intellectual property
in the future.
Frustrations over the lack o f effective interna­
tional agreements regarding IPR have led the
United States to unilaterally create measures to
deal w ith w hat it perceives as unfair trading
practices. In particular, the so-called “ Special
301” clause o f the Omnibus Trade and Com­
petitiveness Act o f 1988 requires the U.S. Trade
Representative to identify countries that do not

35

adequately protect or enforce IPR. The trade
representative then has the authority to bring
an unfair trade practice case against that country.
Although several countries have been put on a
“priority watch list,” no cases have yet been in­
itiated by the U.S. Trade Representative under
the “ Special 301” designation.
Countries throughout the w orld, particularly
the m ore industrialized ones, are interested in
increasing the amount o f protection and en force­
ment o f IPR internationally. Th eir concerns are
seen in the current negotiations o f the General
Agreem ent on Tariffs and Trade (GATT). G A TT
is the principal rule-making body fo r interna­
tional trade, whose goal is to eliminate trade
barriers that reduce the fre e flo w o f goods. In
the current round o f negotiations, the so-called
Uruguay Round scheduled to end in Decem ber
1990, one o f the 15 negotiating groups is devoted
to developing an agreem ent regarding "traderelated aspects o f intellectual property rights,
including trade in counterfeit goods.”3 Current­
ly, intellectual property rights are explicitly ex­
cluded from G A T T ’s auspices.
There are many economic and legal issues
related to the protection o f intellectual prop er­
ty.4 The analysis in this paper focuses only on
the trade-related aspects o f IPR, that is, the e f­
fect o f differential (and uncertain) IPR across
countries on trade and the benefits and costs o f
creating international standards fo r protecting
intellectual property.

INTELLECTUAL PROPERTY
Intellectual property is an invention, idea, p ro­
duct or process that has been registered with

3GATT (1990), p. 11.
“For a discussion of some of the legal issues regarding IPR
in international trade, see Meessen (1987). To narrow the
focus of the paper, two issues have been ignored: which
items should be protected by intellectual property regula­
tions and what type of protection—patents or copyrights—
is appropriate for certain types of goods, such as software.
HJSITC (1988).
6Other types of intellectual property, such as trade secrets
and mask works, constitute a small percentage of all in­
tellectual property and are not discussed here. In addition,
because there are no internationally agreed-upon defini­
tions for intellectual or industrial property, these definitions
should be considered chiefly as guidelines.
intellectual property protection is important insofar as in­
novation is desirable. Some analysts argue (see, for exam­
ple, Nordhaus, 1969; Grossman and Helpman, 1989) that
innovation increases growth worldwide, improving the



the governm ent and that awards the inventor
(or applicant) exclusive rights to use the inven­
tion fo r a given period o f time. It confers
“ ...the right to exclude others from making, us­
ing, or selling the invention w ithin the national
territory."5 (For a discussion o f w hy intellectual
property is awarded protection, see shaded in­
sert on pages 39 and 40.) Industrial property
protection is generally awarded only to new
and useful products and production processes.
Intellectual property is protected by the govern ­
ment in a variety o f ways. Copyrights are
awarded to protect original works o f author­
ship, such as literary, artistic and musical
works; trademarks allow a manufacturer ex­
clusive rights to a distinguished name, symbol
or mark.6 As noted above, these rights are only
valid in the countries in which a patent applica­
tion has been awarded. This lack o f an interna­
tional IPR system could have a significant im­
pact on the amount o f innovative activity.7

CURRENT PROTECTION OF IPR
The protection awarded different types o f in­
tellectual property varies substantially across
countries. As shown in table 1, most countries
have patent lengths o f 15 years or m ore.8 His­
torically, an innovation was awarded patent pro­
tection fo r the num ber o f years approximately
equal to the amount o f time it took to train tw o
apprentices (14 years).9 Although the apprentice
system has been obsolete fo r many years, patent
lengths have rem ained essentially unchanged.
Often, how ever, the actual period in which a
firm can sell its product under patent protection
is shortened considerably; fo r example, tests fo r

quality of life through such things as better medicines, im­
proved living conditions and safer production processes.
Of course all innovations may not be beneficial, nor do all
people necessarily benefit from all innovations (see, for ex­
ample, Kamien and Schwartz, 1982 and Maskus, August
1990). In this paper, it is assumed that innovation has a
net positive effect on a country’s growth and on growth
worldwide.
8AII countries referred to are members of the World Intellec­
tual Property Organization (WIPO). All statistics in this sec­
tion, unless otherwise cited, are from WIPO (1988).
9Benko (1987).

NOVEMBER/DECEMBER 1990

36

product safety (which, fo r pharmaceuticals, can
take up to 10 years) are included in the life o f a
patent. Recently, both the United States and the
European Community proposed legislation that
w ould increase the effective length o f patent
protection—that is, the amount o f time a firm
can actually market or use a product before its
protection expires.10

Table 1
Domestic Protection
Number of
countries

General length

Patents1

18
28
57

5-10 years
10-15 years
15-20 years

Copyrights2

8
75
9

25 years
50 years
50+ years

Intellectual property

Several additional factors affect the amount o f
protection intellectual property is aw arded in
any given country. As tables 2 and 3 show, many
countries, including some industrial ones, ex­
clude certain industries from patent protection.
Those excluded are prim arily high-technology
industries w ith very high R&D intensities that
tend to produce expensive products. Although
currently a substantial amount o f innovation oc-

'From date filed.
2From author’s death.
SOURCE: World Intellectual Property Organization (1988)

10See Congress and the Nation (1985) and “ Patents for
Pharmaceuticals” (1990).

Table 2
Patent Exclusions_______________
Product

Pharmaceutical products (PHARM)
Animal varieties (ANIM)
Methods for treatment of humans
or animals (TREAT)
Plant varieties (PLANT)
Biological processes producing
plant or animal varieties (BIO)
Food products (FOOD)
Computer programs (COMP)
Chemical products (CHEM)
Nuclear inventions (NUC)
Pharmaceutical processes
Food processes
Microorganisms
Substances from microbiological
processes
Cosmetics
Fertilizers
Mixture of metals and alloys
Agricultural machines
Anticontaminants
Methods of agriculture or
horticulture

Number
of countries
that exclude

Industrial
countries1

47
59

8
18

17.0%
30.5

59
57

18
18

30.5
31.6

56
34
48
21
13

18
8
20
2
2

32.1
23,5
41.7
9.5
15.4

10
9
8

2
3
1

20.0
33.3
12.5

6
2
2
2
1
1

1
0
0
0
0
0

16.7
0.0
0.0
0.0
0.0
0.0

1

0

0.0

Percent
industrialized

1These are the Western European countries, Japan, Canada, Australia, United States, New Zealand
and Iceland.
SOURCE: World Intellectual Property Organization (1988)


http://fraser.stlouisfed.org/
FEDERAL RESERVE BANK OF ST. LOUIS
Federal Reserve Bank of St. Louis

37

Table 3
Selected Patent Exclusions by Country
Industrialized
Countries

Australia
Austria-Belgium
Canada
Denmark
Finland-Norway
France
Germany
Greece
Iceland
Ireland
Italy
Japan
Luxembourg
Netherlands
New Zealand
Portugal
Spain
Sweden
Switzerland2
United Kingdom
United States
Total

PHARM

ANIM

TREAT

PLANT

X
X
X
X
X
X
X1

X
X
X
X
X
X
X1

X
X
X
X
X
X
X’

BIO

X

X
X
X

FOOD

X
X
X
X
X
X
X
X’

X
X
X

COMP

CHEM

NUC

X
X
X
X
X
X
X
X1

X
X

X
X
X’
X

X'

X

X
X

X
X

X
X
X
X
X

X
X
X
X

X
X
X
X
X

X
X
X
X
X

8

18

18

18

18

8

20

2

2

13

6

8

6

5

7

2

6

3

X
X
X

X

X

X

X
X

X

X

X

8

20

18

20

20

5

18

1

0

Western Hemisphere
Brazil
Mexico
Venezuela

10

8
X
X

7

X
X

8

8
X
X

X
X
X

8

2
X
X

6

X
X
X

3

European
Eastern

8
7

7
4

8
5

5
4

5
2

6
6

6
3

6
6

5
5

Total

39

41

41

39

38

26

28

19

11

TOTAL

47

59

59

57

56

34

48

21

13

X
X
X
X
X

X
X
X'
X1
X
X

X
X
X
X3
X

X

X
X

X

Developing
Countries

Asian
India
Indonesia
Pakistan
Rep. of Korea
Thailand
African

X

X
X
X

X

X

X
X

X

X

X

X
X

NOTE: See table 2 for definitions.
1lf filed through the European Patent Convention (EPC).
includes Liechtenstein
Liechtenstein excludes if filed through the EPC.
SOURCE: World Intellectual Property Organization (1988)




NOVEMBER/DECEMBER 1990

38

curs in some excluded industries—fo r example,
genetic engineering—it generally takes place in
countries that do not exclude them from protec­
tion. For example, although European firm s ac­
count fo r 82 percent o f w orld investment in in­
dustrial plant and other biotech assets, only
2 percent o f it was spent in Europe. Similarly,
77 percent o f all patents in biotechnology w ere
issued in the United States and Japan, which o f­
fe r the most extensive patent protection.11
The length o f copyright protection, on the
other hand, is a m ore standardized measure
across countries, w ith the m ajority o f countries
issuing copyright protection fo r the rem ainder
o f the author’s life plus 50 years.
Other factors that influence the extent o f pro­
tection fo r intellectual property are the en force­
ment o f existing laws and the amount o f copy­
right and patent protection in other countries.
These factors are particularly important w hen
examining the trade-related aspects o f IPR.

Company $5.7 billion; this amount was reduced
to $909.5 million in 1990 on appeal.
Th e second type o f patent infringem ent w ould
occur if a company produced a camera similar
to Polaroid's and labeled and sold it as a Polaroid.
This is an example o f a "cou nterfeit” product. In
this instance, Polaroid’s tradem ark as w ell as its
patents w ould be infringed upon.
International trade can affect even non-traded
products that are protected by IPR. For example,
suppose a pharmaceutical firm , call it SAW, re ­
cently developed a new product called NOCOLD,
that cured the common cold and had no side e f­
fects. If the firm intended only to sell NOCOLD
in its ow n country (call it the North), it w ould
file fo r a patent only in the North.

The problem w ith the lack o f an international
system o f rules regarding IPR occurs w hen the
cost o f copying an innovation (including the cost
o f penalties if caught) is less than that o f either
purchasing or leasing the technology itself.

Even if SAW does not export NOCOLD, h ow ­
ever, the rate o f return earned could be affected
b y international trade. For example, a firm in
another country could create either a counter­
feit version or a cheaper imitation o f the p ro­
duct and export it to the North.12 Similarly, a
country that is producing NOCOLD legally could
export it back to the North.13 I f sales o f the
foreign-produced product in the North reduce
SAW ’s sales o f NOCOLD and its profits on the
product, the firm w ould expect a low er return
on NOCOLD as w ell as all prospective products.
Because the expected rate o f return affects the
decision to develop new products, the firm
undertakes fe w e r projects in the future. Thus,
protecting intellectual property is not just a
domestic issue.

Th ere are prim arily tw o ways that a product
patent can be infringed upon: First, a copy o f
the product can be produced, w ith essentially
the same characteristics (although not necessari­
ly the same quality) as the original, but w ith no
pretense o f being the original product. Typical­
ly, such copies are sold at a low er price. An ex­
ample o f this occurred in 1976 w hen Kodak in­
troduced a line o f instant cameras that w ere
similar to those already patented by Polaroid. In
1986, U.S. courts ruled that Kodak had infringed
on Polaroid’s patents and awarded the Polaroid

I f the company that infringed upon a patent
in either o f the tw o ways was not a U.S. com ­
pany, the patent laws in both the United States
and the country in w hich it was located w ould
have to be considered. Unless an agreement
states otherwise, a patent or copyright is valid
only in the country in which it was issued.14
Thus, if Kodak had produced and sold its cam­
eras in countries that did not honor Polaroid’s
patent, Polaroid w ould have been unable to
sue Kodak. Because many patent violations oc­
cur across national borders, differences in pa-

TRADE-RELATED ASPECTS OF
INTELLECTUAL PROPERTY:
SOME THEORETICAL ISSUES
Why IP R Is an International Issue

11“ Bugs that Divide” (1990).
>2Not all countries have made it illegal to import products
that infringe on domestic patents.
13Parallel trade, which occurs when a product produced
under a patent in one country is exported to another coun­
try and competes with the patented product in that coun­
try, is legal in many countries and for intra-EC trade.

http://fraser.stlouisfed.org/
FEDERAL RESERVE BANK OF ST. LOUIS
Federal Reserve Bank of St. Louis

14There are surprisingly few agreements among countries to
honor each other's patents. This is discussed in greater
detail later.

39

The Economics of Innovation
Th ere are essentially tw o types o f techno­
logical innovations: process innovations, which
are new production processes or im prove­
ments on existing technology, and product in­
novations, which are the creation o f new p ro­
ducts or improvements on existing products.1
Both types o f innovation are patentable. Be­
cause the economics o f these tw o are essen­
tially the same, the discussion focuses on p ro­
duction innovation fo r simplicity.
Intellectual property has the unusual (al­
though not unique) property that the know ­
ledge it contains is not depleted w ith use. For
example, no matter how many times the fo r ­
mula fo r aspirin is used, the formula itself
(that is, the knowledge contained in the pa­
tent) remains unchanged. As a result, the
marginal cost o f using this knowledge (e.g.,
the formula fo r aspirin) is zero. For economic
efficiency, this knowledge should be made
available to anyone interested, because doing
so does not diminish the stock o f knowledge
(or reduce the number o f times aspirin can
be made). O ver time, how ever, such a policy
would have some unfortunate consequences.
Generally, innovation is the result o f invest­
ment expenditures on research and develop­
ment (R&D). Because expenditures on R&D
occur before a new product is created, the
firm's decision to incur these costs involves
considerable uncertainty. The expected rate
o f return on R&D, which is the present dis­
counted value o f the stream o f net operating
profits divided by the present value o f the
R&D costs, has to be at least as great as the
opportunity cost o f resources devoted to
R&D—that is, the expected rate o f return
that would have been earned if the same
resources allocated to R&D w e re invested
elsewhere.

W hile the opportunity cost o f capital is easy
to determine (it is simply the interest rate),
the rate o f return on R&D is m ore difficult
to ascertain. It depends on h ow much R&D
must be spent before a new product is dis­
covered and developed, how much demand
there w ill be fo r the new product, and how
much production costs w ill be. The return on
R&D also depends on the time the firm can
produce the product exclusively and there­
fo re earn econom ic profits. W hile the first
three factors are outside the scope o f this
paper, the last factor demonstrates how IPR
can effect the incentive to innovate.
In the absence o f governm ent intervention,
maintaining exclusive rights to an innovation
fo r any period o f time is often difficult. Given
that the marginal cost o f using the knowledge
created b y the innovation is zero, one could
conclude that governm ents have no reason to
award these rights. W ithout assigning ex­
clusive rights to produce the innovation, h ow ­
ever, the amount o f time the innovating firm
can produce the product is both less certain
and likely shorter; any other firm that can
figure out how to make the product could
also produce it without changing the know ­
ledge associated with the innovation. For ex­
ample, a firm that did not discover the fo r­
mula fo r aspirin but, instead, was able to
produce it w ould reduce the return earned
by the innovating firm. This is true even
though entry by the noninnovating firm in
this market does not diminish the innovating
firm ’s ability to produce aspirin. This reduced
return on the investment in R&D appears to
increase efficiency by prom oting competition;
how ever, it also reduces the number o f R&D
projects that w ill be undertaken. If, however,
the governm ent assigns property rights to

1The economic reasons for copyright protection, essen­
tially the same as those for patent protection, are not
discussed separately. The reasons for protecting
trademarks, however, differ from those for patent and
copyrights, as trademarks are thought to provide infor­
mation for consumers about the quality of a product.
Protection of trademarks is intended to ensure that
they have some informative value to consumers, rather
than to protect an idea itself.




NOVEMBER/DECEMBER 1990

40

innovations (and enforces them), then the
amount o f time the product can be produced
exclusively w ill increase, raising the rate o f
return on R&D, which in turn has a positive
effect on the amount o f innovation.2
O f course, the w orld is not certain. Th ere
is no w ay o f knowing in advance w hether
the R&D expenditures w ill produce an eco­
nomically viable product. Intellectual prop er­

2There is some disagreement as to how much IPR in­
creases innovation (see, for example, Maskus, 1990).
The question is a difficult one to measure empirically,
because it requires determining what a firm would do
in a situation that has not occurred. One attempt to do
so is Mansfield (1986), who finds some role for patents
in encouraging innovation.

tent laws across countries and the lack o f an in­
ternational enforcem ent system affects the in­
centives associated w ith innovation.

IP R and International Trade: An
Example
Suppose that SAW wanted to sell NOCOLD in­
ternationally as well. A firm can do this one o f
three ways, all o f which are affected by the state
o f international property rights. For simplicity,
the example below compares the case in which
only one o f the tw o countries protects IPR.
D ir e c t E x p o r t s —First, SAW can export
NOCOLD directly to another country, which w e
w ill call the South. If the product is not p ro­
tected by a patent in the South, either because
SAW has not filed fo r protection there or
because the South does not protect IPR, cheaper
copies that use the same formula as NOCOLD
could be legally sold. Similarly, if the trademark
is not protected, counterfeit products that are
indistinguishable from NOCOLD can also be
legally sold. These copies decrease SAW's total
sales and profits associated w ith producing and
selling NOCOLD. In addition, if the product has
a tradem ark that is strongly associated with
SAW, the counterfeit product, if o f significantly
poorer quality, could adversely affect the
reputation o f the firm and further harm the
sales o f both NOCOLD and future products o f
SAW.

15For a discussion of the effects of international piracy in
motion pictures in the United States, see USITC (1988)
and Plock (1989).

FEDERAL RESERVE BANK OF ST. LOUIS


ty rights are a w ay o f rew arding firms fo r in­
curring the risk associated w ith R&D by in­
creasing the expected rate o f return on R&D,
thereby making m ore projects possible.
As long as innovation is considered
desirable, assigning property rights to in­
tellectual property is one w ay to encourage
firms to innovate.3

3For a general discussion of the role of R&D on innova­
tion, see Kamien and Schwartz (1982). In addition, the
issue of determining the optimal length and breadth of
a patent (within a country), has recently been dis­
cussed in the RAND Journal of Economics (Spring
1990) and Economic Inquiry (October 1984).

L ic e n s in g —SAW can also license the technolo­
gy to produce NOCOLD to a firm in the South.
Th e Southern firm, in turn, pays SAW a royalty
fee. W ithout protection o f IPR, how ever, a firm
has no incentive to pay licensing fees if it is
cheaper to copy the product than pay those
fees. For industries w h ere licensing represents a
significant proportion o f its revenues, the lack
o f IPR can cause a substantial loss o f revenue
fo r a firm . For example, motion picture firms
receive considerable revenues from licensing
fees paid by foreign distributors. Every
“pirated” copy o f a m ovie that is sold or shown
represents a loss o f revenue fo r the m ovie com­
pany.15
F o r e ig n D ir e c t In v e s tm e n t—The third possi­
bility, called foreign direct investment, occurs if
SAW locates production facilities in the South to
produce NOCOLD. Again, how ever, the lack o f
IPR in the South has a similar effect on innova­
tion as discussed previously. Because SAW has
no guarantee that it can control the production
o f NOCOLD in that country, the expected rate
o f return on the innovation is smaller, reducing
the profitability o f foreign investment in the
South; as a result, fe w e r firm s w ill engage in
such foreign direct investment. A country’s lack
o f adequate protection o f IPR could be particu­
larly costly in this case because the country is
foregoing some o f the benefits o f foreign direct
investment, such as n ew resources, training and

41

employment. Obviously, counterfeiting firms
might provide some o f these same benefits; since
they develop no new products themselves,
how ever, they are dependent on others fo r in­
novations to copy.16
As long as the direct cost o f counterfeiting (or
copying), including the likelihood and penalties
associated w ith being caught, is less than the
profits earned by the firm s doing the copying,
firm s w ill continue to pirate technology. Copy­
ing, how ever, low ers the rate o f return earned
by innovators and th erefore the overall amount
o f innovation. In the long run, the resulting
decline in the amount o f innovation reduces the
counterfeiting that can be done by firms in the
South. On an aggregate level, the citizens o f
both countries end up w orse off. Th ere is less
innovation, fe w e r products overall, and w o rld ­
w ide grow th is th erefore lower.
W hile the vast majority o f countries have laws
protecting intellectual property, the arguments
presented above describe what occurs in coun­
tries that either have w eak IPR or simply fail to
enforce w hatever IPR they do have. Given the
apparent advantages o f protecting intellectual
property, w h y are there so many problems with
trade in goods affected by intellectual property
rights? Part o f the answer to this question can
be found by examining the current international
system o f regulating intellectual property rights.

CURRENT SYSTEM OF INTERNA­
TIONAL IPR
The main organization responsible fo r interna­
tional agreements on IPR is the W orld Intellec­
tual Property Organization (WIPO), which is ad­
ministered under the auspices o f the United Na­
tions. W IPO's objectives include administering
international treaties and agreements on intellec­
tual property rights and encouraging the protec­
tion o f IPR w orld w id e.17 Currently, 125 coun­
tries are members o f W IPO .18 Table 4 provides
a partial list o f these countries and o f signatories

16Whether or not foreign direct investment is desirable for
developing countries is a separate issue and is discussed
elsewhere. (See, for example, Hood and Young, 1979.) For
the purposes of this paper, we assume that the lack of
IPR does not mean a country doesn't want foreign direct
investment—there are more straight-forward ways to
reduce or eliminate foreign investment.
17For a list of organizations that are solely administered by

o f several treaties administered by W IPO. The
tw o major international agreements on IPR are
the Paris and Berne Unions, w hich deal w ith in­
dustrial property (including patents) and copy­
rights, respectively. (For a description o f the
tw o agreements, see shaded insert on page 43.)
In addition to these unions, other international
property right agreements administered by W IPO
cover such things as industrial design, satellite
transmissions and registration o f trade and ser­
vice marks.19 Not all members o f W IPO, h o w ­
ever, are signatories to these unions and treaties,
nor are the number o f signatories fo r each
agreement identical (see table 4).20

Differing Patent Regulations
One important issue regarding the trade-related
aspects o f intellectual property is the problem
o f differential patent regulations across coun­
tries. For example, w hile most countries deter­
mine patent eligibility based on a first-to-file
basis, the United States employs a first-to-invent
rule. As a result, patent protection fo r the same
invention could be awarded patent protection to
different applicants depending on w hether the
actual inventor was the first to file.
W IPO has had some success in trying to stan­
dardize the process o f obtaining patents interna­
tionally through the Patent Cooperation Treaty
(PCT). This treaty allows applicants to file fo r a
patent in a central office and specify in which
o f the signatory countries (shown in table 4) it
wishes the application to have effect. This p ro ­
cess reduces the costs o f filing fo r patents by
centralizing the search and examination w ork
associated w ith determining patent eligibility.
The PCT is also designed to increase the time
an applicant has to decide w hether to w ithdraw
the application fo r foreign patents. A firm might
choose to w ithdraw its patent application and
avoid the expense associated w ith translating
the patent application into the local language
and finding a local patent agent if the demand
fo r the product to be patented is likely to be
too small.

18These statistics exclude the German Democratic Republic
because of the unification with the Federal Democratic
Republic that occurred on October 3, 1990.
19See WIPO (1990a).
20The United States, for example, did not sign the Berne
Union until 1988.

I, see W lr U (1988), p.




NOVEMBER/DECEMBER 1990

42

Table 4
Signatures of Selected International Intellectual Property
Agreements
Industrialized
Countries

Australia
Austria-Belgium
Canada
Denmark
Finland-Norway
France
Germany
Greece
Iceland
Ireland
Italy
Japan
Luxembourg
Netherlands
New Zealand
Portugal
Spain
Sweden
Switzerland2
United Kingdom
United States
Total

WIPO

X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X

Members of Treaties
Berne
Paris1

X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X

X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X

PCT

X
X
X
X
X
X
X

X
X
X
X
X
X
X
X
X

24

24

24

19

Asian3
India
Indonesia
Pakistan
Rep. of Korea
Thailand

27

9

17

4

African

40

25

35

14

Western Hemisphere
Brazil
Mexico
Venezuela

23

14

11

2

European
Eastern

11
7

11
7

12
7

4
3

Total

101

59

75

21

TOTAL

125

83

99

43

Developing
Countries

X
X
X
X
X

X
X
X

X
X
X

X
X
X

X
X

X
X

'Dominican Republic, Iran, Nigeria, San Marino and Syria are not members of WIPO.
includes Liechtenstein.
3Hong Kong, Singapore and Taiwan are not members of WIPO.
SOURCE: World Intellectual Property Organization (1990a)


http://fraser.stlouisfed.org/
FEDERAL RESERVE
Federal Reserve Bank of St. Louis

BANK OF ST. LOUIS

X

43

The Paris and Berne Unions: Some
Background Inform ation
“Th e Paris Union, the official name o f w hich
is the International Union f o r the Protection o f
Industrial Property, was founded b y a Conven­
tion signed in Paris in 1883 and last revised
in 1967.
Under the Convention, each m em ber State
must accord the same protection to the in­
ventions, trademarks and other subject mat­
ter o f industrial property o f the nationals o f
the other m em ber States as it accords to
those o f its ow n nationals. The Convention
also provides certain facilities to foreigners;
fo r example, it allows them, without losing
their claim to novelty, to file their applica­
tions fo r patents up to a year after first filing
in the country o f origin. It contains provi­
sions concerning the conditions under which
a State may license the use o f a patented in­
vention in its ow n territory if, fo r example,

Ninety percent o f all patent applications are
filed in the 43 countries that make up the m em ­
bership o f the PCT.21 The usefulness o f the PCT
is demonstrated by the surge in the number o f
applications received by the PCT, from 2,625 in
1979 to 14,874 in 1989.22 In addition, the
number o f countries designated fo r patent p ro­
tection by applicants rose from 6.6 percent to
15.8 percent o f m em ber countries during the
same time period.

Enforcem ent o f IP R
As mentioned previously, the enforcem ent o f
existing IPR is another big problem. Although
W IPO administers several international accords
regarding IPR, fe w laws currently en force these
treaties. The Paris Union only requires signato­
ries to give foreigners national treatment, that

21The statistics in this section are from WIPO (1990a,
1990b).
22The potential of this agreement has just begun to be
realized. For example, the United States, which is a
signatory of the PCT, had 66,850 applications for patents
by foreigners. (U.S. Department of Commerce, Patent and
Trademark Office, 1989).



the ow ner o f the patented invention does not
exploit it in such territory.
Th e Berne Union, the official name o f which
is the International Union f o r the Protection o f
Literary and A rtistic Works, was founded by a
Convention signed in Berne in 1886 and last
revised in 1971.
Under the Convention, each m em ber State
must accord the same protection to the copy­
right o f the nationals o f the other m em ber
States as it accords to that o f its ow n nation­
als. Th e Convention also prescribes some
minimum standards o f protection; fo r exam­
ple, that copyright protection generally con­
tinues throughout the author's life and fo r 50
years thereafter. It includes special provisions
fo r the benefit o f developing countries.”
W IPO, General Information, 1990

is, award them the same rights as they give their
ow n citizens.
The Berne Convention includes tw o en force­
ment provisions. First, it established the p re­
sumption o f authorship so that the author does
not have to prove it; rather, challengers o f the
copyright w ould have to provide evidence to
the contrary. Second, copies that infringe on a
copyright are subject to seizure in any country
in which the w ork has a copyright. This also
applies to imports o f reproduced materials from
countries w h ere the w ork is not protected.
W IPO does not have an international dispute
settlement mechanism w hereby an applicant (or
country) can file a complaint against another
country’s implementation o f the treaties. The
only recourse under the agreem ent is to bring a
case before the International Court o f Justice.23
That court, how ever, can only arbitrate cases

23The court’s ruling is not binding for all members, however.
See WIPO (1988).

NOVEMBER/DECEMBER

1990

44

that relate to the interpretation or application o f
the Paris and Berne Unions. Th ere appears to be
no penalty fo r ignoring the ruling o f the Court.
W IPO does require countries to give applicants
access to the same legal remedies fo r patent in­
fringem ent as they do their ow n nationals. This
requirement, how ever, is subject to existing
laws fo r patent violations. In fact, many coun­
tries do not have explicit penalties associated
w ith violations o f IPR, and fe w impose civil
penalties. According to W IPO, as o f 1988, only
14 countries had laws requiring the seizure o f
infringing patent articles, 20 countries granted
compensation o f damages and 11 destroyed in­
fringing goods. Th ere is somewhat m ore protec­
tion fo r products w ith trademarks. W IPO is
now preparing a study that w ill examine the
possibility o f establishing a new treaty which
creates a dispute mechanism to arbitrate possi­
ble violations o f international IPR agreements.

Other Issues
The problems described above explain w h y
there is considerable concern about the current
international system o f IPR. Having an interna­
tional set o f rules regarding the protection o f
intellectual property reduces the uncertainty
associated w ith innovation and increases the ex­
pected return earned on those innovations. This
does not mean that all countries must have the
same degree o f protection—GATT, fo r example,
has different rules regarding the acceptable
amount o f tariffs fo r industrial and developing
countries. Rather, an explicit set o f agreements,
along with an effective mechanism to mediate
disputes, could significantly decrease the loss o f
earnings associated w ith copying and counter­
feiting innovated products. In addition, such
agreements could reduce the information costs
o f determining the amount o f protection o f in­
tellectual property. These costs can be substan­
tial fo r firm s that intend to sell their products
internationally, and, as described above, must
file fo r a patent in each country w h ere it wants
to sell the product. Similarly, given the current
system in which countries can choose to ex­

24This terminology comes from Maskus (1990). This analysis
can also be used within a country, where some regions
have industries with a lot of innovation (such as the Silicon
Valley in California), and others have very few innovating
industries. For an example of a model that looks at innova­
tion and technology transfer both within and across coun­
tries, see Butler (1990).
25Chin and Grossman (1988), for example, find that the

http://fraser.stlouisfed.org/
FEDERAL RESERVE BANK OF ST. LOUIS
Federal Reserve Bank of St. Louis

clude specific products from patent protection,
firm s are faced w ith the choice o f selling their
products in some markets w ith no protection or
avoiding certain markets altogether. Developing
countries that protect IPR w ill increase its access
to new technology because innovating firms w ill
have stronger incentives to produce and sell
their products in countries that protect intellec­
tual property.

PROBLEMS IN REACHING INTER­
NATIONAL AGREEMENTS ON IPR
W h y are there so fe w (and such weak) interna­
tional agreements on IPR? One prim ary difficul­
ty in protecting and enforcing IPR is that the in­
centives to do so d iffer across countries, par­
ticularly betw een countries that are technology
exporters and those that are technology im­
porters.24 Generally, less innovation occurs in
developing countries; instead, in the absence o f
licensing or foreign direct investment, firm s in
these countries tend to produce goods whose
production technology has becom e standardized
or whose patent protection has expired.
Firms that successfully pirate technology in
many developing countries are often able to
produce essentially the same product at sub­
stantially low er production costs. Because there
is less innovation in developing countries, the
cost o f not protecting IPR (reduced future inno­
vation) is often less, at least in a static sense,
than the gain associated w ith selling these
products.
An argument often made by developing coun­
tries is that there is "excessive” protection o f
IPR in industrial countries. The optimal amount
o f protection o f intellectual property is difficult
to determine w ithin a given country. This issue
becomes m ore complicated in an international
context, because w hat is optimal from a domes­
tic perspective may not be optimal from an in­
ternational standpoint. For example, even from
a long-run perspective, the optimal amount o f
IPR protection can d iffer across innovating and
non-innovating countries.25

desired amount of protection between innovating and non­
innovating countries depends on the specification of social
welfare used and that strong IPR may not improve global
efficiency. Diwan and Rodrik (1989), in a different
theoretical framework, also find that the optimal amount of
protection between the innovating and non-innovating
countries coincide only if welfare in the two regions is
weighted equally.

45

Many poorer developing countries find it dif­
ficult to pay the higher price innovated goods
w ould cost if they are given patent protection
in their country. Developing countries th erefore
have little incentive, at least in the short run, to
protect commodities they need (or want) but
could not afford to buy if protection is aw ard­
ed. An example o f this is seen in the pharma­
ceutical sector, which is awarded patent protec­
tion in essentially all industrial countries, but
not in many poorer developing countries, which
might otherwise have difficulty purchasing
medical supplies.
Another argument fo r not protecting IPR in
technology-importing countries is that reverse
engineering, the process by which firms take
products apart to learn how to produce them,
enables firms to learn how to develop new pro­
ducts themselves and th erefore aids in a coun­
try ’s development. As the technological kn ow ­
how im proves in a country, these firm s begin to
innovate themselves. W hen this stage is reach­
ed, protection o f IPR generally begins to in­
crease.26
Given the current level o f protection o f IPR,
creating increased standards o f protection, at
least in the short run and fo r the least-developed
countries probably in the long run, w ill
redistribute income from technology-importing,
developing countries (whose residents w ill now
have to pay m ore fo r these types o f products)
to countries w hich innovate or already protect
IPR. As a result, success in negotiating increased
international protection fo r IPR w ill likely re ­
quire some concessions to the developing coun­
tries in other areas o f trade.27 These issues are
currently being discussed by W IPO and GATT.

CONCLUSION
The incentive to protect intellectual property
rights, as w ell as the actual amount o f protec­
tion awarded, differs w idely across industrial
and developing countries. Nevertheless, recent

26This has recently occurred in Korea, which, according to a
1987 study by the International Trade Commission, was
the third-largest source of U.S. losses resulting from viola­
tions of IPR. Recently, however, Korea has taken steps to
improve its enforcement of IPR, and the number of patents
issued in the United States to Korean nationals has risen
substantially, although it is still small in absolute value
(U.S. Department of Commerce, Patent and Trademark Of­
fice, 1989). Of course, U.S. pressure has provided addi­
tional impetus for Korea to increase IPR protection.



negotiations under W IPO and G A TT (which, as
o f this writing, is yet to be concluded) regarding
the trade-related aspects o f intellectual property
rights indicate increased support fo r interna­
tional agreements on intellectual property rights
and a realization that such agreements benefit
both industrial and developing countries. In ad­
dition, as the amount o f inventive activity and
the number o f countries engaged in innovation
increases, the trend tow ard m ore cooperation
and protection o f intellectual property is likely
to continue.

REFERENCES

Benko, Robert P. “ Intellectual Property Rights and the
Uruguay Round,” World Economy (June 1988), pp. 217-31.
_______ Protecting Intellectual Property Rights: Issues and
Controversies (American Enterprise Institute for Public
Policy Research, 1987).
“ Bugs that Divide.” Economist, July 28, 1990, pp. 57-58.
Butler, Alison. “ Endogenous Innovation in a North-North
Model of the Product Cycle,” Federal Reserve Bank of St.
Louis Working Paper 90-007 (1990).
Chin, Judith C., and Gene Grossman. “ Intellectual Property
Rights and North-South Trade,” Princeton University Discus­
sion Papers in Economics, No. 143 (November 1988).
Congress and the Nation. Vol. 6 (Congressional Quarterly
Inc., 1985), p. 548.
Diwan, Ishac, and Dani Rodrik. “ Patents, Appropriate
Technology, and North-South Trade,” National Bureau of
Economic Research Working Paper No. 2974 (May 1989).
Feinberg, Robert M. “ Intellectual Property, Injury, and Inter­
national Trade,” Journal of World Trade (April 1988), pp.
45-56.
Feinberg, Robert M., and Donald J. Rousslang. “ The
Economic Effects of Intellectual Property Right Infringe­
ments,” Journal of Business (January 1990), pp. 79-90.
Finger, Michael J., and Andrzej Olechowski, eds. The
Uruguay Round: A Handbook on the Multilateral Trade
Negotiations (The World Bank, 1987).
General Agreement on Tariffs and Trade. GATT: What It Is,
What It Does (1990).
Grossman, Gene M., and Elhanan Helpman. “ Product
Development and International Trade,” Journal of Political
Economy (December 1989), pp. 1261-83.
Hood, Neil, and Stephen Young. The Economics of Multina­
tional Enterprise (London: Longman Group United, 1979).
27For a review of some of the problems in negotiating IPR in
the Uruguay Round, see Maskus (1990), Benko (1988) and
Finger and Olechowski (1987).

NOVEMBER/DECEMBER 1990

46

Kamien, Morton I., and Nancy L. Schwartz. Market Structure
and Innovation (London: Cambridge University Press,
1982).
Mansfield, Edwin. “ Patents and Innovation: An Empirical
Study,” Management Science (February 1986), pp. 173-81.
Maskus, Keith. “ Normative Concerns in the International Pro­
tection of Intellectual Property Rights,” paper prepared for
meetings of the International Seminar in International Trade
at the National Bureau of Economic Research (Auqust 2-3,
1990).
_______. “ Intellectual Property,” in Jeffrey J. Schott, ed.,
Completing the Uruguay Round: A Results-Oriented Ap­
proach to the GATT Trade Negotiations (Institute for Interna­
tional Economics, 1990), pp. 164-79.
Meessen, Karl M. “ Intellectual Property Rights in Interna­
tional Trade,” Journal of World Trade Law (February 1987),
pp. 67-74.
Nordhaus, William D. Invention, Growth, and Welfare: A
Theoretical Treatment of Technological Change (MIT Press,
1969).
“ Patents for Pharmaceuticals.” Financial Times, May 15, 1990


FEDERAL RESERVE BANK OF ST. LOUIS


Plock, Ernest. “ International Piracy of Motion Pictures,” in
Ann Main, “ Pursuing U.S. Goals Bilaterally: Intellectual
Property and ‘Special 301’,” Business America (September
25, 1989), p. 7.
U.S. Department of Commerce, Patent & Trademark Office.
Commissioner of Patents and Trademarks Annual Report
(January 1989).
United States International Trade Commission. “ Foreign Pro­
tection of Intellectual Property Rights and the Effect on
U.S. Industry and Trade,” Report to United States Trade
Representative, Investigation No. 332-245, Publication 2065
(February 1988).
World Intellectual Property Organization. General Information
pamphlet (Geneva, January 1990a).
_______ . WIPO Newsletter (February 1990b).
_______. Existence, Scope and Form of Generally Interna­
tionally Accepted and Applied Standards/Norms for the Pro­
tection of Intellectual Property (Geneva, September 1988).
Yuetter, Clayton. News Release. “ Statement by Ambassador
Clayton Yuetter Concerning Intellectual Property Rights
Protection,” Office of United States Trade Representative
(April 7, 1986).

47

Michelle R. Garfinhel

Michelle R. Garfinkel is a senior economist at the Federal
Reserve Bank of St. Louis. Scott Leitz provided research
assistance.

The Economic Consequences
of Reducing Military Spending

B

EFORE THE CRISIS in the Middle East this
summer, the easing o f international tensions
had reduced, fo r many, the urgency fo r the
United States to continue building or even main­
tain its military strength. As support fo r allo­
cating the nation’s resources to defense w eak­
ened, people began to argue about the potential
fo r a significant "peace dividend” available to
the U.S. econom y.1
H ow should the savings from reduced defense
spending be put to use to enhance our nation’s
w elfare? Many analysts, concerned primarily
about the effects o f large public deficits, have
argued that the savings should be applied to
reduce the governm ent’s need to b o rrow .2
Others have voiced concern that a reduction in
defense expenditures w ill generate unem ploy­
ment, at least tem porarily, w hile resources are
reallocated to productive activities in the civilian
sector.3 Consequently, they have argued the in­
itial savings should be used to ease this adjust­
ment—perhaps by increasing expenditures on

'For example, see Pennar and Mandel (1989).
2See, for example, Schultze (1990).
3For example, see “ Peace Dividend or Recession?” (1990).
Also see Pennar and Mandel (1989), who report the results
of a study of the short- and long-term effects of reducing
the defense budget by 5 percent a year in real terms from
1991 to 1994 and keeping it constant thereafter. Although
this study predicts enhanced economic growth in the long
run, it also predicts some short-term losses. Also see Ellis



training programs.4 Many other policy recom ­
mendations have been made.
Although the current situation in the Middle
East raises doubt that there w ill be any signifi­
cant dividend in the near term, it does not de­
tract from the relevance o f such recom menda­
tions. Instead, it provides us with m ore time to
evaluate the various options associated with
future defense cuts.
In review in g the economic implications o f
reduced military spending, this article examines
some issues that have been overlooked b y those
in search o f ways to use the “peace dividend.”
The article begins w ith a b rief analysis o f re­
cent trends and the prospects fo r future cuts in
military spending to see how large a dividend
might be. Some simple economic principles are
then em ployed to assess how the peace dividend,
regardless o f its actual magnitude, might be used
to achieve diverse economic goals.

and Schine (1990), who report the results of a study in­
dicating that as many as 1 million defense-related jobs (or
20 percent of all jobs in defense-related activities) could be
lost by 1995. Although other analysts have argued that the
employment losses could be insignificant (see, for exam­
ple, Uchitelle (1990)), some defense contractors have
already begun to cut production and employment.
4See, for example, the bill proposed recently by Senator
Pell (S.2097).

NOVEMBER/DECEMBER 1990

48

decades. Its share in the 1980s was only onethird o f that in the 1940s.

Table 1
Trends in U.S. Defense Spending in the
Past 50 Years
Total defense spending1
Decade

1940s
1950s
1960s
1970s
1980s

Nominal

Real

$ 35.8
40.2
59.2
90.3
238.3

$285.5
187.0
216.4
200.0
274.4

Share of GNP

18.3%
10.1
8.4
5.8
6.1

'Billions of dollars averaged over the decade. The real figures
are in 1989 dollars.

H O W BIG MIGHT THE DIVIDEND
BE?
T o place the discussion o f the economic im ­
plications o f reduced defense needs into perspec­
tive, it is helpful to examine recent patterns o f
m ilitary spending. Table 1 shows three v e ry dif­
feren t perspectives o f U.S. m ilitary spending
over the past 50 years. W e can see that nominal
m ilitary spending in the United States has grow n
considerably since the 1940s, rising from an
average o f $35.8 billion per year during the
decade that included W o rld W a r II to an
average o f nearly $240 billion per year during
the 1980s. A fte r adjusting fo r inflation,
how ever, w e see a somewhat different picture.
Real military spending declined sharply im­
m ediately after the W W II decade; and, its pat­
tern since the 1950s has been m ore erratic,
w ith its net rise by the 1980s being considerably
less dramatic than suggested by the nominal
figures. Finally, as the table shows, military
spending as a fraction o f gross national product
(GNP) has fallen m arkedly over the past five

5During the Reagan Administration’s military build up,
defense spending grew at a 5.0 percent annual rate after
adjusting for inflation; in comparison, the annual growth
rate in real GNP over this same period was only 3 per­
cent. Earlier this century, before WWI, military spending’s
share of GNP was less than 1.5 percent. (Because data on
military spending before the 1940s are not entirely consis­
tent with the data presented in table 1, they are not shown
here.)
6NATO’s proposal then called for a reduction of its own and
the Warsaw Pact’s ground capabilities in Europe and for a
reduction in their tactical aircraft capabilities that would
leave an advantage for NATO. (Ground capability is mea­
sured by army units intended to fight on ground. Tactical
aircraft capabilities are measured by fighters and bombers

http://fraser.stlouisfed.org/
FEDERAL RESERVE BANK OF ST. LOUIS
Federal Reserve Bank of St. Louis

W hile m ilitary spending might now represent
a fairly small proportion o f GNP, it is still an im ­
portant component o f economic activity. Defense
spending was $301.1 billion in 1989—m ore than
$1200 per person in the United States.5 This
number suggests that there could be a consider­
able dividend from a large-scale disarmament.
Even without the Middle East crisis, how ever,
the military spending cuts likely to have occurred
in the near future w ould have been quite small
relative to the w hole economy. In a recent study,
fo r example, the U.S. Congressional Budget O f­
fice (1990) estimated the effects o f a proposal
by the North Atlantic Treaty Organization
(NATO) fo r limiting conventional forces in
Europe.6 It found that the treaty w ould
generate an annual savings o f about $3 billion in
1990 prices. Although $3 billion seems large, it
constitutes less than 1 percent o f the total
Defense Departm ent’s budget authority fo r
fiscal year 1990 and less than .06 percent o f
1989 GNP. This amounts to less than $13 per
year per U.S. citizen.
This estimated savings from the prospective
reduction in m ilitary spending pales in compari­
son to earlier U.S. disarmament efforts fo llow ­
ing w artim e periods.7 A fter W orld W a r II, fo r
example, defense spending fell by about $57.3
billion from 1945 to 1946, almost 27 percent of
GNP in 1945. From 1953 to 1954, after the K ore­
an War, defense spending fell b y $7.4 billion,
almost 2 percent o f GNP in 1953.
Forecasts o f the actual size o f future defense
cuts, o f course, are subject to much uncertain­
ty. The budget proposed b y President Bush in
January 1990 fo r fiscal year 1991 called fo r
reducing the defense budget by 2 percent after

intended to fight in the air and on the ground using con­
ventional weapons.) It called for a 27 percent reduction in
a selection of NATO weapons. Assuming an equally pro­
portionate reduction by all NATO members, this require­
ment implies that the U.S. would have had to remove and
destroy 600 tanks, 122 armored personnel carriers, 112
pieces of artillery, 189 helicopters and 105 aircrafts from
Europe. Furthermore, about 30,000 U.S. troops would
have had to be withdrawn from Europe and demobilized.
7As noted below, however, the economic implications of a
given reduction in military spending depend quantitatively
on whether the reduction follows a war or occurs during a
relatively peaceful period.

49

adjusting fo r inflation over the fiscal years 1991
to 1995.8 Others argued fo r defense spending
cuts as much as 5 percent in real terms per
year from fiscal years 1992 to 1994, achieving
an annual savings o f about $60 billion (in 1989
prices) starting in 1994.9 If cuts o f this magni­
tude w ere implemented, the implied savings
w ould constitute about 1.2 percent o f GNP fo r
the year 1989, nearly $250 per year on a per
capita basis.

Figure 1
PPC: Defense and Nondefense Goods

And the savings could be even larger. Indeed,
on the day o f the Iraqi invasion o f Kuwait,
President Bush announced that, although the in­
vasion indicates a need to maintain a strong
military force, U.S. armed forces could be
reduced b y 25 percent over five years given the
recent changes in Soviet-U.S. relations.10

EFFECTS OF THE DIVIDEND:
SOME BASIC ECONOMIC
PRINCIPLES
Generally speaking, the trade-off betw een
competing uses o f resources implies that a re ­
duction in real m ilitary spending w ould provide
a "dividend” in the fo rm o f increased real private
and public consumption and investment oppor­
tunities—that is, increased resources available
fo r the production o f consumption and invest­
ment goods. T o quantify these increased oppor­
tunities over time, the annual peace dividend is
defined here simply as the annual reduction in
real military spending.11
Figure 1 illustrates this trade-off given the na­
tion's resource and technology constraints by
means o f a hypothetical production possibility
curve (PPC) fo r defense goods (national security)
and nondefense goods (public and private in­
vestment and consumption). This curve depicts
the maximum quantities o f defense goods (M)
and nondefense goods (N) that can be produced
simultaneously fo r given amounts o f capital and
labor inputs.

s“ Peace Dividend or Peace Recession?” (1990).
9“ The Peace Economy” (1989).
10Dowd (1990). Specifically, the Pentagon plan then called
for cutting the armed forces by 500,000 troops from the
current level of 2.1 million. But, without a clear resolution
of the ongoing conflict in the Middle East, any reduction in
military spending might seem optimistic.



Assuming that resources are fully utilized, the
econom y is always operating on the fron tier
regardless o f the level o f military spending. If
no resources w ere allocated to the production
o f defense goods, fo r example, the total output
o f nondefense goods w ould be shown as N0.
Producing defense goods thus requires sacrific­
ing the production o f some goods fo r invest­
ment and consumption. The opportunity cost o f
providing a specific level o f national security,
fo r example, MA, is the value o f the lost produc­
tion o f nondefense goods, NA- N o. Conversely, if
the econom y w e re originally operating at point
A, cutting military spending out entirely would
imply increased production o f nondefense goods,
N a - N c. Thus, w ith a reduction in military spend­
ing (say M B- M A), the annual dividend w ould im­
ply increased opportunities fo r the production
o f nondefense goods (NB- N A). The shaded insert
on pages 50 and 51 contains a discussion o f the
w elfa re implications o f increased opportunities
fo r investment and consumption afforded by a
reduction in military spending.

11This definition envisions the dividend as a flow—i.e., as
the term is normally understood. Thus, a permanent cut of
X dollars in real terms per year implies an X dollar divi­
dend each year indefinitely into the future. Using a present
discounted value concept, these flows over time can be
translated into a stock concept: X/r, where r is the cons­
tant real interest rate.

NOVEMBER/DECEMBER 1990

50

W ill a Dividend Necessarily Enhance “Social
Welfare?”
Some interesting social w elfare implications
o f low er military spending can be illustrated
within the simple PPC fram ew ork. Figure 2
depicts the same PPC shown in figure 1; add­
ed to the figure are tw o indifference curves
that show specific combinations o f defense
and nondefense goods w hich yield constant
levels o f national “w elfa re” or "utility." The
curve labeled u indicates a higher level o f
utility than the one labeled u'. T h e shape o f
the indifference curves reflects the notion o f
diminishing social marginal utility. For exam­
ple, the nation is “less w illing” (must be given
considerably m ore nondefense goods) to
decrease consumption o f defense goods and
consume m ore nondefense good at point Y
compared w ith point Z.
At the point o f tangency (labeled X) be­
tw een the indifference curve associated with
utility u and the PPC, the nation’s utility is
maximized, given the available resource and
technology constraints. At this optimum level
o f consumption, the marginal utility trade-off

betw een defense and nondefense goods (the
slope o f the indifference curve, u) equals the
marginal rate o f transform ation betw een
them (the slope o f the PPC). Th e optimal
quantities o f defense and nondefense goods
are M x and Nx, respectively. If preferences
remain unchanged, a m ove to any other p ro­
duction combination on the PPC (or inside)
w ould result in a low er level o f national
w elfare. Thus, although a m ove from point X
to point Y, fo r example, produces a positive
dividend in the form o f increased resources
available to produce nondefense goods, it
w ould actually reduce the nation’s w elfare.
Th ere are tw o scenarios, how ever, under
w hich a decline in military spending could
enhance w elfa re by creating increased con­
sumption and investment opportunities. First,
the original quantities o f M and N could have
been suboptimal. If, fo r example, the econ­
om y w ere operating at a point to the left o f
X on the PPC (for example, point Z), military
spending w ould be considered too high from

Figure 2

Figure 3

Optimal Levels of Defense and
Nondefense Goods

A Preference-Induced Change in the
Optimal Levels of Defense and
Nondefense Goods


FEDERAL RESERVE


BANK OF ST. LOUIS

51

a social w elfa re point o f view. Alternatively,
the original point could have been inside the
PPC above point X, such as Z'. In this case,
military spending is also too high and, m ore
important, the econom y is operating below its
potential output levels.
Second, a cut in military spending could
generate w elfa re gains if it w e re associated
w ith an unanticipated decline in the severity
o f external threats. This seems to be the ex­
ample most relevant to the current situation
involving U.S.-Soviet relations. The reduced
value o f maintaining existing U.S. military
strength in the face o f detente can be cap­
tured best by presuming there has been a

The concave shape o f the PPC reflects dimin­
ishing marginal returns in productive transfor­
mation. That is, the amount o f nondefense
goods that must be sacrificed to produce one
m ore defense good increases as M increases.
For example, in the figure, the m ove from C to
A and the m ove from A to B involve identical
reductions in military spending. Starting at the
point w ith a higher level o f m ilitary spending
(Mc), how ever, that reduction in military spend­
ing implies greater additional production o f non­
defense goods than w hen starting at point B.
(That is, N a - N c > N b - N a.) Hence, assuming
resources are fully employed, a given reduction
in the production o f defense goods w hen the
current level is low —fo r example, during
peacetime—w ould imply smaller additions to
consumption and investment opportunities than
when the current level is high—fo r example,
during wartime.

THE CROWDING-OUT EFFECT OF
MILITARY SPENDING AND THE
LONGER-RUN EFFECT OF THE
PEACE DIVIDEND
These allocative effects o f the peace dividend
can also have implications fo r the amount o f
resources available fo r production in general.

,2Of course, one might argue that the level of military spen­
ding could positively influence the position of the PPC.
Because military spending enhances a nation’s ability to
protect its resources, it might serve to increase the na­
tion’s future resource base by encouraging more
investment.



general clockwise m ovem ent in the indif­
ference curves. This m ovem ent is illustrated
in figure 3, w h ere the new indifference
curves are labeled u. I f national security can
be maintained w ith less m ilitary spending, a
smaller quantity o f M with the same quantity
o f N (Nx) produces the same utility level
uo= uo. Further, the marginal value o f increas­
ing M relative to the cost o f foregone con­
sumption o f N falls. The new w elfa re max­
imizing combination o f M and N along the
PPC is denoted by Y, w h ere N increases to NY
and M decreases to M Y. The net "w elfare e f­
fect” o f the dividend, M Y- M X, w ould be
reflected in the increase from u to u;.

Specifically, any additional investment adds to
the future resource base, thereby enhancing
future output grow th and investment and con­
sumption opportunities. Hence, even if the de­
cline in military spending w ere tem porary, fo r
example lasting only one year, its benefits in
terms o f increased productive capacity could be
realized over many years. This longer-term e f­
fect can be m odeled in the fram ew ork presented
above by an outward shift o f the PPC curve
over time.12
Some analysts believe that, in recent decades,
military spending has been excessive, reducing
the residual supply o f productive resources (that
is, capital and labor) available fo r private and
nondefense public investment and thereby w eak­
ening the econom y.13 According to this "deple­
tion” theory, the effects o f higher levels o f
military spending are reflected in low er rates o f
investment and, consequently, low er rates o f
economic growth. Thus, the principal result o f
reduced military spending w ould be greater in­
vestment and econom ic growth.
Evidence on the "crow ding-out” effect of
military spending is based prim arily on empirical
analyses relating changes in military spending’s
share o f GNP to the GNP shares o f other broad
categories o f expenditures; typically, the studies
focus on the effect on private investment’s

13See, for example, Dumas (1987), Melman (1988) and Du
Boff (1989). In contrast, Weidenbaum (1990) argues that,
as defense spending has fallen relative to GNP, the effects
of such spending on the U.S. economy have become less
significant.

NOVEMBER/DECEMBER 1990

52

Table 2
Broad Categories of Real Expenditures
as a Share of Real GNP1
Decade

1940s
1950s
1960s
1970s
1980s

C

54.7%
58.4
59.5
62.2
64.5

1

12.1%
16.5
16.7
17.2
16.9

X

Gnm

Gm

10.1%
11.8
14.3
14.9
13.6

23.0%
13.2
10.3
6.2
6.2

0.5%
0.1
-0.4
-0.7
-1 .2

’ Each category is converted into real terms using its im­
plicit price deflator. Separate price deflators were used for
federal government spending and state and local government
spending, but both defense and non-defense federal expen­
ditures were deflated by the same number.

share o f GNP.14 Table 2 shows the trends in
personal consumption expenditures (C), private
domestic investment (I), nondefense public ex­
penditures (Gnm) on goods and services, defense
expenditures (G J and exports net o f imports (X)
as shares o f GNP over the past 50 years. To
focus on the real effects o f military spending,
each broad category o f spending is converted
into real terms b y dividing it b y its ow n price
deflator and by real GNP.15
As the table suggests, real military spending
has crow ded out all categories o f real expendi­

14See Gold (1990) for an extensive survey of this evidence.
Also see Adams and Gold (1987) and U.S. Congressional
Budget Office (1983). It should be noted that identifying
the degree to which military spending has resulted in
lower expenditures on other goods and services and lower
economic growth is difficult. The problem lies in determin­
ing how investment and other expenditures would have
behaved if military spending had been different—in an ex­
treme case, if it had been zero. Because reduced-form
parameters relating defense expenditures to other expen­
ditures would not be independent of the policy regime,
estimates of these parameters might provide little informa­
tion on how a permanent change in military spending (i.e.,
a policy regime change) would influence other expen­
ditures. In addition, it is important to note that, if higher
military expenditures result in higher levels of GNP, lower
shares of investment, for example, need not imply a
crowding-out effect of military spending.
15A comparison of tables 1 and 2 reveals that real military
spending as a fraction of real GNP was higher than
nominal military spending as a fraction of nominal GNP
from the 1940s to the 1970s, but slightly lower during the
1980s. This divergence reflects the difference between the
general price level of defense goods and that of all goods
and services. The price level for defense goods, on
average, was lower than the general price level between
the 1940s and the 1970s, but higher during the 1980s.
The focus on real military spending here is intended to

FEDERAL RESERVE BANK OF ST. LOUIS


tures, not just real private domestic investment.
As real military spending’s share fell from the
1940s to the 1980s by 16.8 percentage points,
real consumption’s share rose b y nearly 10 per­
centage points. Real private investment’s share
and real nondefense public spending’s share
also rose over this period, though less dramati­
cally.16 O f course, in a broad sense, the substitu­
tion observed betw een nondefense and defense
public expenditures is consistent w ith the
crowding-out notion; in this case, public invest­
ment on the nation’s infrastructure—that is,
highways, airports, mass transit, w ater
systems—was crow ded out.17
T o be sure, the size o f the peace dividend, as
defined here, is independent o f its allocation
among the production o f private and public
consumption and investment goods. But its
longer-term implications depend on that alloca­
tion. The best or "socially optimal” reallocation
o f resources among investment and current con­
sumption depends on the nation’s willingness to
forego current consumption in order to invest
and thereby enhance future consumption possi­
bilities. The greater this willingness, the m ore
likely the resources from a reduction in military
spending w ill be devoted to additional invest­
ment rather than additional current consump­
tion. The low er the nation’s willingness to
forego current consumption to enhance future
consumption, the low er w ill be the proportion

emphasize the importance of its real allocative effects.
Failure to account for relative price movements masks
these effects.
16Smith (1977) finds that the crowding-out effect of military’s
share of income on investment's share of income is nearly
one to one for 14 OECD nations during the 1960s. How­
ever, Boulding (1973), Edelstein (1989) and Aschauer
(1989a) argue that this effect is not empirically relevant for
the United States. Also see Browne (1989), who questions
the validity of the argument that military research and de­
velopment has crowded out civilian spending on research
and development by “ depleting” our nation’s scientists
and engineers. As is well-known, military R&D has pro­
duced important innovations that have been applied suc­
cessfully to production activities in the civilian sector; one
commonly cited example is the computer. In addition,
Browne argues that this crowding-out effect on R&D pre­
sumes that the supply of scientists and engineers is fixed.
A greater demand for highly skilled labor, however, has in­
fluenced its supply, although with the usual lag.
17As an annual average of GNP, net public infrastructure in­
vestment fell from approximately 2.3 percent in 1960-65 to
about 0.8 percent in 1980-85 (Du Boff (1989), p. 7, table
2). Aschauer (1989b) argues that the recent reduction in
public capital, including infrastructure, might be responsi­
ble for the recent decline in productivity.

53
o f the savings that is allocated to investment.
Thus, in contrast to the suggestion o f the deple­
tion theory briefly described above, the divi­
dend from reduced military spending need not
result in significantly greater rates o f invest­
ment and economic growth. The extent to
w hich the dividend w ill affect economic grow th
depends on h ow it is allocated among the p ro­
duction o f investment and consumption goods.

ALTERNATIVE USES OF THE
DIVIDEND
In recent decades, decreases in nominal
military spending typically have been associated
with increases in other public expenditures in
nominal terms. Indeed, after falling to 19.5 per­
cent in the 1950s from 25.1 percent during the
1940s, total public spending (federal plus state
and local) on goods and services has remained
roughly constant as a fraction o f nominal GNP,
around 20 percent; only the composition o f
those expenditures changed. Although there
could be reasons w h y this pattern might persist
in the upcoming decades, many analysts have
questioned w hether a continuation o f this pat­
tern is either likely or even desirable. N ever­
theless, the basic question to be addressed
should be couched in real terms; W hat should
be done with the peace dividend?

Increase Nondefense Public
Expenditures
Some analysts have argued that leaving
nondefense public expenditures alone and using
the reduction in military spending to either
decrease the deficit or low er taxes is not the
best use o f the peace dividend. Instead, many o f
them believe that at least part o f the savings
from arms reduction might best be used to in­
crease nonmilitary governm ent spending—
specifically, to rebuild the nation’s infrastruc­
ture.18 In terms o f the economic fram ew ork
above, this policy w ould shift out the PPC, in­

18For example, see Du Boff (1989) and Melman (1988).
19For example, see Bolton (1966), especially pp. 37-41.
20See Barro (1981). In studying the output effects of govern­
ment expenditures in the United States, he distinguishes
permanent from temporary components of military spen­
ding. He finds that the effect of increases in temporary
military spending (essentially wartime expenditures) on
output was nearly one-for-one; increases in permanent
military spending also increased output, but by less than
the change in military spending.
21See Barro (1981) for a theoretic discussion of the effects
of government expenditures on output. In support of this



creasing future production and consumption
opportunities.
Others have argued that, unless the fall in
military spending is somehow offset, resource
utilization and, hence, economic activity w ill fall
as w ell.19 In this view , w hich lies w ithin the
standard "Keynesian" paradigm, the governm ent
should use part (or all) o f the savings to finance
additional public expenditures, including noninfrastructure expenditures, such as w elfa re p ro­
grams, to offset the negative effect o f reduced
military spending on aggregate demand.
This argument assumes that military spending
in particular or public expenditures in general
enhance social w elfa re not only by providing
additional public goods, but by increasing
em ployment and thereby stimulating the econo­
m y—that is, by inducing the use o f idle resources.
It implies that, w ithout the increases in military
spending or, m ore generally, public expenditures
during the post-W W II period, the econom y
would have operated below its potential output
capability (that is, inside its PPC).
Although there is evidence that a permanent
decrease in military spending can produce a
permanent decline in aggregate output,20 the
decline in output could be generated, in part,
by a voluntary reduction in the supply o f labor.
In other words, this evidence does not necessari­
ly imply that a permanent decline in military
demand, without an increase in other public
spending, w ould cause these productive re ­
sources to becom e involuntarily idle on a per­
manent basis.

Reduce Taxes
Some analysts w ould like us to consider an
alternative policy that leaves other governm ent
expenditures unchanged and uses the dividend
to reduce taxes. By increasing individuals’ after­
tax income, this policy might induce individuals
to decrease their supply o f labor, which w ould
in turn decrease output without leaving labor
resources involuntarily idle.21

line of reasoning, Dunne and Smith (1990) find that for the
United States, military spending does not “ cause”
unemployment. Riddell (1988) argues, in a more Marxian
spirit, that the government's apparent bias for military over
non-military expenditures is driven by its objective to main­
tain international order so as to maximize the profitability
of U.S. capital. This possible endogeneity of military spen­
ding calls many empirical analyses that treat military spen­
ding as exogenous into question. Also see Garfinkel
(1990a), who uses a game-theoretic model to show how
military spending can be driven by aggregate economic
activity through the government’s motive to prevent other
nations from extracting its citizens’ resources.
NOVEMBER/DECEMBER 1990

54

It should be noted that a permanent decline
in measured output, triggered by the impact o f
reduced military spending (and reduced taxes)
on leisure, does not necessarily reflect a deterio­
ration in social w elfare. Because leisure has
value, w elfa re could increase even if consump­
tion did not. Further, the theory described
above suggests that, although individuals would
w ork less, they might actually consume m ore
nondefense private goods because their
disposable income has increased (their tax
liabilities have declined). On net, their w elfare
w ould have increased as long as this new out­
come w ere chosen voluntarily.22

Reduce the Public Deficit
Some analysts, w ho view large public deficits
as harmful to the economy, have argued that
the governm ent should use the peace savings to
reduce the public deficit.23 In particular, the
large deficits (public dissavings) o f the past
decade are thought to have caused a decline in
total national savings—that is, the sum o f
private and public savings. Since a decline in
total savings decreases the residual supply o f
credit available to private borrow ers, large
public deficits are considered by many to have
pushed up expected real interest rates (interest
rates adjusted fo r expected inflation).24 Thus,
using the dividend to reduce the public deficit
w ould decrease expected real interest rates and
thereby stimulate both investment activity and
the production o f goods, such as exports and
n ew homes, w hose sales are sensitive to
movements in interest rates.
Although the U.S. savings rate appears to
have declined in recent years, how much o f this
decline can be blam ed on large public deficits is

22As discussed below, however, labor and capital resources
could be left involuntarily idle temporarily as the economy
adjusts to the reduced military demand.
23See, for example, Schultze (1990). Also, see Chrystal and
Thornton (1988) for a related discussion of the effects of
deficit spending.
24lndeed, this effect on interest rates is thought to be the
mechanism through which military spending has crowdedout investment. By enhancing the productivity of private
capital, however, military spending could have had a
“ crowding-in” effect that would have offset its crowdingout effect. But Aschauer (1989a,b) presents evidence that
does not support the notion that additions to the stock of
military capital add to the productivity of private capital.
Moreover, the references cited in footnotes 14 and 16 pro­
vide evidence that military spending does not crowd out
private investment.
25Schultze (1990) estimates that national savings as a
percentage of national income (i.e., net national product)
has fallen from an average of 8 percent during the three

FEDERAL RESERVE BANK OF ST. LOUIS


unclear.25 The argument that public deficits in­
fluence the national savings rate is based on a
number o f potentially questionable assumptions.
One is that individuals do not view tax cuts that
increase public borrow in g (holding the level o f
governm ent spending constant) as increasing
their future tax liabilities. Instead, individuals
feel w ealthier and increase their consumption in
response to such tax cuts. Although they also
might respond by increasing their savings, the
increase in private savings is assumed to be in­
sufficient to keep total savings from falling. In
this view , fo r a given level o f governm ent ex­
penditures, public deficits decrease total savings
and increase aggregate demand.26

D oes the Timing o f Taxes Matter?
Other analysts argue that individuals believe
reductions in current taxes associated w ith addi­
tions to public debt must be financed eventually
with additional future taxes. In light o f the in­
crease in their future tax liabilities, individuals
increase their savings. Conversely, they respond
to a decrease in the public deficit, fo r a given
level o f governm ent expenditures, by decreasing
their savings. In either case, consumption is
unaffected.
In this view , often referred to as the “Ricar­
dian" view , individuals behave as if a decrease
in the public deficit results in an equal decrease
in the present discounted value o f their future
tax liabilities. This argument builds on the
assumption that public debt w ill be retired
eventually out o f future taxes. I f this view is
valid, private and public savings fo r a given
level o f governm ent expenditures should be
perfectly negatively correlated, w hile total sav­
ings should be unrelated to public savings.27

decades before 1980 to 3.3 percent during the first three
quarters of 1989. Some analysts, however, question the
notion that savings is too low in the United States; their
skepticism is based on problems with the conventional
measurements of savings. Cullison (1990) provides a
useful survey of this literature.
26See Thornton (1990) for a theoretical discussion of the link
between total national savings and public deficits.
27Against this Ricardian view, one might argue that deficits
could be financed either through increased seigniorage or
income taxes in future generations. In either case, the
“ burden” of the current deficit could be shifted and
budget deficits, holding government expenditures fixed,
could affect economic activity. See Barro (1989) for a brief
discussion of the empirical evidence on the effects of
budget deficits. While recognizing the problems associated
with testing the Ricardian proposition, Barro argues that
the existing evidence lends more support to the Ricardian
view than to the alternative view (p. 52).

55
To be sure, using part o f the annual dividend
to reduce the public deficit could increase total
investment and total consumption. According to
the Ricardian view , how ever, the amounts o f
these increases do not depend on w hether taxes
are cut or the deficit is reduced. A cut in the
deficit reduces future tax liabilities, but the tim ­
ing o f the tax cuts does not matter.
Because o f the distorting nature o f the income
tax system, how ever, the equivalence betw een
taxes and debt creation implied by the Ricardian
view w ould be, at best, a rough approximation.
Economic theory predicts that proportional in­
come taxation distorts individuals' decisions
about consumption and labor supply. These
distortions are costly and, other things being
equal, the severity o f the distortion increases as
the tax rate increases. Consequently, if the
federal governm ent wants to minimize the costs
associated with these distortions, given the path
o f future governm ent expenditures, it should
smooth income taxes over time.28 This m odified
version o f the Ricardian view suggests that tax
reductions, rather than deficit reductions,
w ould be a preferable use fo r the peace divi­
dend.29

THE TRANSITIONAL COSTS OF
ECONOMIC ADJUSTMENT
In thinking about how the savings from re­
duced military spending could be used, it is im­
portant to consider how the econom y adjusts to
unanticipated changes in resource uses. Reduced
military spending w ill produce a negative shortrun effect on production, as labor and capital
resources are shifted from military to civilian
uses. During the transition period, some re ­
sources w ill be unem ployed or underemployed.

A Historical Perspective
Previous disarmaments have been associated
w ith sizable reductions in economic activity.30

28See Barro (1979). Also see Garfinkel (1990b) for an exten­
sion of this theory to include conscription as an additional
tool for financing public expenditures to avoid the distor­
tions of income taxes, particularly during periods of severe
military needs.
29ln this view, deficits are necessary to smooth out the
distortionary effects associated with taxes. Hence, a tem­
porary increase in government spending should be financed
with debt and a temporary decrease in government spend­
ing should result in a budget surplus; this tax-smoothing
view of debt creation predicts that deficits are temporary
phenomena. Although historical evidence supports this



From the first quarter o f 1945 to the first quar­
ter o f 1946 (peak to trough), fo r example,
nominal GNP fell at a seasonally adjusted an­
nualized rate o f $22.8 billion or 10.3 percent.
These numbers understate the magnitude o f the
decline in output as, during this period, there
was a considerable acceleration in inflation.31 In
real terms, GNP fell 18 percent over this period,
and it was not until the third quarter o f 1952
that the level o f real GNP had fully recovered.
Although this decline in real GNP is large, it is
an overstatem ent o f the drop in national w el­
fare. W hile the level o f employment fell sub­
stantially, the unemployment rate rose very little.
Instead, a substantial num ber o f workers, partic­
ularly wom en, voluntarily w ith d rew from the
labor force.
The transition to peacetime after W W II was
facilitated, in part, by governm ent policies. Tax
reductions and transfer payments (unemploy­
ment and veteran benefits) left disposable (net
o f taxes) income nearly unaffected by the mas­
sive reduction in m ilitary spending. Thus, de­
mand fo r consumption goods rose to offset par­
tially the decline in military demand.
The sharp decline in military spending that
follow ed the end o f the Korean W a r was asso­
ciated w ith a mild recession. From the second
quarter o f 1953 to the second quarter o f 1954
(peak to trough), real GNP fell 3.2 percent. De­
clines in defense spending and other federal
governm ent expenditures, combined with inven­
tory decumulations, w e re the driving forces
here. By the first quarter o f 1955, how ever, real
GNP had climbed w ell above its previous peak.

Factors Influencing The Adjust­
ment Costs
The aggregate adjustment cost fo r any reduc­
tion in military spending can be measured by
the real value o f resources (labor and capital)
left idle involuntarily during the transition. For
a given cut in defense spending, the magnitude

positive theory of debt creation—see Barro (1979), for
example—it is unclear whether the current deficit is only a
temporary phenomenon. Indeed, the magnitude and per­
sistence of the peacetime deficits during the 1980s are un­
precedented in U.S. history.
30See Bolton (1966) and references cited therein for a more
detailed examination of these periods of disarmament.
Much of the discussion here draws from this work. Data
are taken from Balke and Gordon (1986).
31This acceleration was driven, in part, by the removal of
price controls in 1946.
NOVEMBER/DECEMBER 1990

56

o f these costs depends on the speed w ith which
labor and capital resources can be transform ed
to meet new demands.
The speed o f resource transformation, in turn,
depends on the degree o f specialization o f
resources used in the military sector. This spe­
cialization has tw o dimensions. First, certain in­
dustries, occupations and firms are highly de­
pendent on military demand. Second, military
production is highly concentrated in several
regions o f the United States. Such specialization
w ill slow the adjustment process.
A given reduction in military demand now
might generate relatively greater adjustment
costs than those associated w ith large-scale dis­
armaments follow ing wartim e periods. During
w artim e periods, resources norm ally used to
produce nondefense goods are m obilized quick­
ly and, presumably, on a tem porary basis; after
the war, resources are rechanneled easily into
their original civilian productive activities. In
contrast, during peacetime defense firm s and
their em ployees expect that demand fo r their
product is essentially permanent. T o the extent
that these firm s and employees have a compara­
tive advantage in the production o f defense
goods, they are less likely either to diversify
their operations into civilian markets or be able
to do so in the event o f an unanticipated perma­
nent reduction in military spending.32

EVALUATING ALTERNATIVE
PROGRAMS
T o evaluate society’s options fo r using the sav­
ings from reduced military spending, w e must
address tw o related economic issues. The first
issue, already discussed, concerns w hat are the
best uses o f the dividend or the new “ highestvalued” uses o f the resources previously used in
the military (presumably their previous highestvalued uses). The second issue concerns how to
rechannel resources efficiently from their mili­
tary uses to their new highest-valued uses.

32ln contrast, one might believe that defense contractors,
having learned from past experience with sharp declines
in military demand, would have diversified their operations
to exploit commercial opportunities. While such diversifica­
tion would provide insurance against large losses to these
firms and their employees in the event of an unexpected
decline in military demand, past efforts in this direction
have not been particularly successful. See Weidenbaum
(1973) and Ellis and Schine (1990).
“ Representative Boxer’s proposal (HR5327) is similar;
however, it would penalize defense contractors who close

FEDERAL RESERVE BANK OF ST. LOUIS


Some people have advocated establishing public
programs—fo r example, training programs—to
lessen the costs o f adjustment borne solely by
those closely linked to the m ilitary sector. A bill
introduced recently b y Senator Pell (S.2097), fo r
example, seeks to establish a program through
which grants w ould be made to assist state and
local governm ents in developing economic ad­
justment plans—fo r example, job retraining and
finding alternative uses fo r defense facilities.
These grants w ould be funded, in part, from the
savings from reduced military spending.33 Anoth­
er bill introduced by Senator Coats (S.2682) is
intended to aid defense contractors in diversify­
ing their operations into nondefense markets.
Through tax incentives, this bill would encourage
defense contractors and their employees to adopt
em ployee stock ownership plans (ESOPs) to fi­
nance the corporate restructuring necessary to
adjust to the reduction in military demand.
These programs are aimed at distributing the
adjustment costs and the benefits o f reduced
military spending equitably; how ever, they are
unlikely to effect an efficient reallocation o f
resources.34 Consider, fo r example, a program
that increases public nondefense expenditures
on goods that are most easily produced b y the
capital and labor resources originally em ployed
fo r the production o f defense goods. W hile this
program might w ell limit the adverse impact o f
reduced military spending otherwise borne by
those firm s and individuals highly dependent on
military demand, it is clearly inefficient from
the nation's point o f view. Unless increased
spending on these other goods w ere deemed
desirable on a permanent basis, this policy
w ould not provide firm s and individuals with
the incentives to channel their resources to new
higher-valued uses. Instead, it w ould merely
prolong the process o f adjustment and delay the
realization o f the full benefits from a perm a­
nent reduction in military spending.
Nevertheless, the redistributive effects o f
reduced military spending should not be

their firms unless a contract has been canceled. The
states of Washington and California already have initiated
adjustment plans. See Ellis and Schine (1990).
“ Although the evidence on the effectiveness of manpower
programs (for example, the Manpower Development and
Training Act) is mixed, some studies find that manpower
policies have been successful in raising the earnings of
training program participants. See, for example,
Ashenfelter (1978).

57

dismissed as unimportant or irrelevant in choos
ing how the peace dividend ultimately w ill be
used. The question o f w ho reaps the gains and
w ho bears the costs o f an unanticipated reduc­
tion in military spending is an important aspect
o f the problem and w ill play an important role
in the solution.

CONCLUSIONS
This article has examined some possible e f­
fects o f a permanent reduction in military spen­
ding. In principle, the present discounted value
o f the implied dividends from such a reduction,
in terms o f increased consumption and invest­
ment opportunities, could be substantial.
Through increased private and public invest­
ment, reduced military spending implies greater
economic grow th and, hence, greater consump­
tion and investment opportunities in the future.
The important economic questions are, How can
these resources be reallocated efficiently to non­
military uses? and, H ow can w e identify what
these uses should be? This article has introduc­
ed and discussed the economic issues that must
be addressed in answering these questions;
much further analysis and discussion w ill clear­
ly be needed before these questions can be
answered adequately.
O f course, given the recent course o f events
in the Middle East, the reduction in military
spending in the near future might not be large
enough to generate any sizable dividend. Thus,
debate over w hether the savings from reduced
m ilitary spending should be used to reduce the
public deficit, redistributed to taxpayers
through tax cuts or be used to rebuild the in­
frastructure might seem premature. Because
many communities, firms and individuals are af­
fected by even small reductions in military
spending, how ever, the "m icro” costs o f these
adjustments w ill not be ignored in the political
decision-making process.
But tem porary transitional costs do not justify
abandoning the effo rt to reduce the amount o f
resources allocated to military spending. M ore­
over, the Middle East situation is not a perma­
nent obstacle to realize a large dividend in the
future. W hile cuts in defense spending are not
expected to be particularly large now, a resolu­
tion o f the Middle East conflict w ill perm it a
much larger dividend in the future.



REFERENCES

Adams, Gordon, and David Gold. Defense Spending and the
Economy: Does the Defense Dollar Make a Difference?
(Washington, D.C.: Defense Budget Project at the Center
on Budget and Policy Priorities, July 1987).
Aschauer, David. "Does Public Capital Crowd Out Private
Capital?” Journal of Monetary Economics (September
1989a), pp. 171-88.
_______ "Is Public Expenditure Productive?” Journal of
Monetary Economics (March 1989b), pp. 177-200.
Ashenfelter, Orley. “ Estimating the Effect of Training Pro­
grams on Earnings,'"Review of Economics and Statistics
(February 1978), pp. 47-57.
Balke, Nathan S. and Robert J. Gordon. "Historical Data,” in
Robert J. Gordon, ed., The American Business Cycle: Con­
tinuity and Change (Chicago: The University of Chicago
Press, 1986), pp. 781-850.
Barro, Robert J. “The Ricardian Approach to Budget
Deficits,” Journal of Economic Perspectives (Spring 1989),
pp. 37-54.
_______ “Output Effects of Government Purchases,” Jour­
nal of Political Economy (December 1981), pp. 1086-121.
_______ “On the Determination of the Public Debt,” Journal
of Political Economy (October 1979), pp. 940-71.
Bolton, Roger E. “ Defense Spending: Burden or Prop?” in
R.E. Bolton, ed., Defense and Disarmament (Prentice-Hall,
Inc., 1966), pp. 1-53.
Boulding, Kenneth E. “The Impact of the Defense Industry
on the Structure of the American Economy,” in Bernard
Udis, ed., The Economic Consequences of Reduced Military
Spending (Toronto: D.C. Heath and Company, 1973), pp.
225-52.
Browne, Lynn E. “ Defense Spending and High Technology
Development: National and State Issues,” New England
Economic Review, Federal Reserve Bank of Boston
(September/October 1988), pp. 3-22.
Chrystal, K. Alec, and Daniel L. Thornton. “The
Macroeconomic Effects of Deficit Spending: A Review,” this
Review (November/December 1988), pp. 48-60.
Cullison, William E. “ Is Saving Too Low in the United
States?” Federal Reserve Bank of Richmond Economic
Review (May/June 1990), pp. 20-35.
Dowd, Maureen. “ Bush Says Military Can Be Cut 25% in 5
Years,” New York Times, August 3, 1990.
Du Boff, Richard B. “What Military Spending Really Costs,”
Challenge (September/October 1989), pp. 4-10.
Dumas, Lloyd J. “ National Security and Economic Delu­
sion,” Challenge (March/April 1987), pp. 28-33.
Dunne, Paul, and Ron Smith. “ Military Expenditure and
Unemployment in the OECD,” Defence Economics: The
Political Economy of Defence, Disarmament and Peace, Vol.
1, No. 1 (1990), pp. 57-74.
Edelstein, Michael. “What Price Cold War: Military Expen­
ditures and Private Investment in the U.S., 1890-1980,”
mimeo (January 1989).
Ellis, James E., and Eric Schine. “Who Pays For Peace?”
Business Week (July 2, 1990), pp. 64-70.
Garfinkel, Michelle R. “Arming as a Strategic Investment in a
Cooperative Equilibrium," American Economic Review
(March 1990a), pp. 50-68.
_______ “The Role of the Military Draft in Optimal Fiscal
Policy,” Southern Economic Journal (January 1990b),
pp. 718-31.

NOVEMBER/DECEMBER 1990

58

Gold, David. The Impact of Defense Spending on Investment,
Productivity and Economic Growth (Washington, D.C.:
Defense Budget Project at the Center on Budget and
Policy Priorities, February 1990).
Melman, Seymour. “ Economic Consequences of the Arms
Race: The Second-Rate Economy,” American Economic
Review (May 1988), pp. 55-59.
“ Peace Dividend or Peace Recession?” The Economist (April
7, 1990), pp. 31-32.
Pennar, Karen, and Michael J. Mandel. “ The Peace
Economy: How Defense Cuts Will Fuel America’s LongTerm Prosperity,” Business Week (December 11, 1989),
pp. 50-55.
Riddell, Tom. “ U.S. Military Power, the Terms of Trade, and
the Profit Rate,” American Economic Review (May 1988),
pp. 60-65.
Schultze, Charles L. “ Using the Peace Dividend to Increase
Saving,” Challenge (March/April 1990), pp. 11-17.
Smith, Ron P. “ Military Expenditure and Capitalism,” Cam­
bridge Journal of Economics (March 1977), pp. 61-76.


http://fraser.stlouisfed.org/
FEDERAL RESERVE BANK OF ST. LOUIS
Federal Reserve Bank of St. Louis

_______ “ Military Expenditure and Capitalism: A Reply,”
Cambridge Journal of Economics (September 1978),
pp. 299-304.
Stevenson, Richard W. “ Suppliers Brace for Arms Cuts,”
New York Times, June 18, 1990.
Thornton, Daniel L. “ Do Government Deficits Matter?” this
Review (September/October 1990), pp. 25-39.
Uchitelle, Louis. “ Economy Expected to Absorb Effects of
Military Cuts,” New York Times (April 15, 1990).
U.S. Congressional Budget Office. Defense Spending and the
Economy (Government Printing Office, February 1983).
_______ . Budgetary and Military Effects of a Treaty Limiting
Conventional Forces in Europe, Special Study (January
1990).
Weidenbaum, Murray. “ Defense Outlays Have Declining Im­
pact on Economy,” Christian Science Monitor, February 7,
1990.
_______ “ Industrial Adjustments to Military Expenditure
Shifts and Cutbacks,” in Bernard Uris, ed., The Economic
Consequences of Reduced Military Spending (Toronto: D.C.
Heath and Company, 1973), pp. 253-87.

Federal Reserve Bank of St. Louis
Post Office Box 442
St. Louis, Missouri 63166

The Review is published six
times per year by the Research
and Public Information
Department o f the Federal
Reserve Rank o f St. Louis.
Single-copy subscriptions are
available to the public fr e e o f
charge. Mail requests f o r
subscriptions, back issues, or
address changes to: Research
and Public Information
Department, Federal Reserve
Rank o f St. Louis, P.O. Roy 442,
St. Louis, Missouri 63166.
The views expressed are those
o f the individual authors and do
not necessarily reflect official
positions o f the Federal Reserve
Rank o f St. Louis or the Federal
Reserve System. Articles herein
may be reprinted provided the
source is credited. Please provide
the Rank’s Research and Public
Information Department with a
copy o f reprinted material.