Full text of Review (Federal Reserve Bank of Dallas) : First Quarter 2000

View original document

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

••••••••••••••••

• ••••

lepben P. A. Brou" and Daniel bIk

EMU at 1
Mark A w:'ynne

Measuring the Benefits of
Unilateral Trade Liberalization
Part 2: Dynamic Models
Carlos E.]. M Zarazaga

This publication was digitized and made available by the Federal Reserve Bank of Dallas' Historical Library (FedHistory@dal.frb.org)

[eonomie ~nd
fin~nei~1 Review
Federal Reserve Bank of Dallas

Robert D. McTeer, Jr.
President and Chief Executive Officer

Helen E. Holcomb
First Vice President and
ChiefOpemting Office,'

Robert D. Hankins
Senior Vice President, Banking Supelvision

Harvey Rosenblum
Senior Vice President and Director ofResearch

W. Michael Cox
Senior Vice President and Chief Economist

Editors
Stephen P. A. Brown
Senior Economist a nd Assistant Vice President

Jeffery W. Gunther
Research Officer

Mark A. Wynne
Research Officer

Director of Publications
Kay Champagne
Associate Editors
Jennifer Afflerbach
Monica Reeves
Design
Gene
Laura
Ellah

and Production
Autry
J. Bell
PUla

Economic and Financial Review (ISSN 1526-3940),
published quarterly by the Federal Reserve Bank of
Dallas, presents in-depth information and analysis
on monetary, financial, banking, and other economic
policy topics. Articles are developed by economists in
the Bank's Economic Research and Financial Industry
Studies departments. The views expressed are those of
the authors and do not necessarily reflect the positions
of the Federal Reserve Bank of Dallas or the Federal
Reserve System
Articles may be reprinted on the condition that the
source is credited and the Public Affairs Department is
provided with a copy of the publication containing the
reprinted material
Subscriptions are available free of charge Please
direct requests for subscriptions, back issues, and
address changes to the Public Affairs Department,
Federal Reserve Bank of Dallas, P.O Box 655906,
Dallas, TX 75265-5906; call 214-922-5254; or subscribe
via the Internet at www.dallasfed.org, Economic and
Financial Review and other Bank publications are
available on the Bank's web site, www dallasfed.org

Contents
ndturdl Rewurce ~cdrcity dnd
Tec~nolo~icdl (~dn~e
Stephen P. A. Brown and Daniel Walk
Page 2

[mu dt 1
Mark A. Wynne
Page 14

medwrin~ t~e Benefit~ of

Unildterdl Trdde Liberdlizdtion
Pdrt 2: Dyndmic model~
Carlos E. J. M. Zarazaga
Page 29

Nonrenewable natural resources, such as aluminum and
crude oil, exist only in fixed amounts on Earth. Consequently, some
observers are concerned that natural resource scarcity will eventually limit future economic growth and human well-being. Others
remain optimistic that technological change will overcome geophysical scarcity. Brown and Wolk examine the evidence for natural resource scarcity and find that over the past century reliance
on free markets has promoted sufficient technological change to
overcome geophysical scarcity for most nonrenewable natural
resources. Rather than rising-as would result from increased
scarcity-the relevant real prices of most nonrenewable natural
resources have fallen. Although declines in real prices have moderated since World War II, tl1e aumors find little evidence of increased scarcity in the postwar era. Increased reliance on markets
during the closing decades of the twentieth century is cause for
optimism that these trends will continue in the twenty-first.

Economic and monetary union (EMU) among eleven of the
fifteen members of the European Union began on January 1, 1999.
The national currencies of the eleven were abolished and replaced
with a new single currency, the euro. Responsibility for monetalY
policy shifted from the national central banks to me European
Central Bank. Many commentators in the United States thought
EMU would never come about or, if it did, that it would not last
long.
In this article Mark Wynne reviews EMU's first year. He looks
at how me economy of the euro area has fared under the single
monetalY policy, examines how successful me ECB has been in fulfilling its mandate for price stability, and considers the prospects for
the future. Despite the dramatic decline in the euro against me dollar over the course of 1999, the first year of EMU must be judged a
success. While it is still too early to say whether in the long run the
euro will be a strong currency like the Deutsche mark, the institutional design of EMU and the performance of those institutions over
the first year are promising.

This is the second of two articles examining the potential welfare gains or losses from a unilateral move toward free trade. Part
1 concluded that applied static models of international trade fail to
produce eye-popping positive welfare effects. In Part 2, Carlos
Zarazaga reviews available applied dynamic general equilibrium
models. He finds that the promises of larger welfare gains from unilateral trade liberalization do materialize in some dynamic models.
However, other models cannot completely dismiss some common
objections to the adoption of unilateral free trade policies.
Zarazaga also identifies the controversial theoretical and
empirical issues behind those objections that will have to be
resolved before unilateral trade liberalization is accepted as the
definitive, welfare-improving alternative to costly and prolonged
multilateral trade agreements.

In 1972, an interdisciplinary research group
called the Club of Rome predicted worldwide
catastrophe by 2050 (Meadows et al. 1972).
They based their prediction on three trends they
thought they observed: increasing scarcity of
nonrenewable natural resources, increasing
environmental degradation, and continuing
population growth. They saw the combination
of these trends as unsustainable and economic
misery as inevitable.
The Club of Rome was not original in its
pessimism about the future. English economist
Thomas R. Malthus raised similar concerns in
1798. His analysis led him to conclude that misery was the inevitable state of humans (Malthus
1798). According to Malthus, if per capita
income were above subsistence, population
would expand until per capita income was
reduced to subsistence level. (See the box entitled “An Overview of Malthus’ Principle of
Population.”) At the time Malthus was writing —
the early stages of the Industrial Revolution —
poverty was widespread in English cities, so
perhaps his pessimism was understandable.
Fortunately for us, Malthus was wrong.
Since at least the late 1800s, per capita income
in Western society has generally increased.
Technological change occurred at a rapid pace,
causing per capita income to rise even as the
population grew. In fact, per capita income rose
so much, the Club of Rome’s pessimism seems
hard to understand, except that Malthus’ original analysis did not take into account natural
resource scarcity or environmental degradation.
This essay examines whether the potential
scarcity of nonrenewable natural resources is a
reason for concern. Previous research (Barnett
and Morse 1963, Jorgenson and Griliches 1967,
Nordhaus 1973, Brown and Field 1978, Fisher
1979, Hartwick and Olewiler 1986, and Schmidt
1988) is mixed, but it generally has found that
the economic evidence is inconsistent with the
increasing scarcity of nonrenewable natural
resources. In fact, technological change driven
by free market forces has increased natural resource availability. Given the time elapsed since
the previous research was conducted, however,
it is appropriate to reexamine the evidence.

Natural Resource Scarcity and
Technological Change
Stephen P. A. Brown and Daniel Wolk

n this article, we examine

trends in the real prices of
nonrenewable natural
resources to determine whether
technological change is
outpacing geophysical scarcity
of these natural resources.

WHAT IS NATURAL RESOURCE SCARCITY?
Nonrenewable natural resources, such as
aluminum and crude oil, exist in fixed amounts
on Earth. When we use up all the crude oil on
the planet, we will have no more of this
resource. In addition, we tend to use the most
easily obtainable natural resources first. Over

Stephen P. A. Brown is director of energy economics
and microeconomic policy analysis in the Research
Department at the Federal Reserve Bank of Dallas.
Daniel Wolk is a research analyst in the Research
Department at the Federal Reserve Bank of Dallas.

FEDERAL RESERVE BANK OF DALLAS

An Overview of Malthus’ Principle of Population
Malthus thought an increase in population would reduce per capita income. His
conclusion followed from the law of diminishing marginal productivity: as population
increases, each worker has less land with which to work. Curve I in the figure represents this proposition for a given amount of land and level of technology. Curve II
represents this proposition for a higher level of technology and/or greater acreage.
The subsistence level of income is also represented in the figure.
For a given amount of land and level of technology, Malthus argued that a population would tend toward a subsistence level of income. If per capita income were
below the subsistence level (as illustrated by point A on curve I), starvation would
reduce the population. If per capita income were above the subsistence level (as
illustrated by point B on curve I), people would have more children and population
would grow. In either case, population would adjust until income just reached the
subsistence level (at point C on curve I). Therefore, he concluded that misery was
the inevitable state of humankind. This conclusion is often referred to as the “dismal
theorem” and may be the historical basis for calling economics “the dismal science.”
Malthus’ analysis is similar to that now made by ecologists studying animal populations and ecosystems. For example, if the deer population is smaller than a given
ecosystem can support, the deer will reproduce and multiply in number. If the population is greater than the ecosystem can support, the weak will die off and the population will be reduced. The deer population tends toward a subsistence level of nutrition.
Malthus further argued that — without moral restraint in human reproduction —
improved technology or increased resources would only increase human misery in
the long run. An increase in technology or land temporarily increases well-being (as
shown by a shift from point C on curve I to point D on curve II). Eventually, however,
the increased capacity of the economy will lead to population growth, which will only
be checked when per capita income reaches subsistence (point E on curve II).
Hence, Malthus concluded that increased technology or land availability would result
in more people living at subsistence, not an improvement in living conditions. This
conclusion is often referred to as the “utterly dismal theorem.”

time, natural resources become more difficult to
extract. For example, at the beginning of the
California gold rush, people were picking up
gold off the ground. Toward the end of the gold
rush, they were blasting the mountains with
water, using much more capital and labor.
Geophysical scarcity may be irrelevant,
however, if technological change increases
resource availability. Consequently, economists
prefer to measure scarcity in economic terms —
that is, through market prices. Economists
are interested in whether the prices of nonrenewable natural resources reflect increasing
scarcity. In other words, are the real prices of
natural resources rising to reflect increasing
scarcity?
The economics perspective can be illustrated by examining a production function for
the overall economy:
(1)

Q = Q (K, L, NR ),

where Q is output, K is capital, L is labor, and
NR is natural resource use.1 We expect normal
economic conditions for production, which mean
a positive marginal product for each input:
(2)

∂Q
> 0,
∂K

∂Q
> 0,
∂L

∂Q
> 0.
∂NR

For each input, output increases with its use, as
is shown by the positive first derivative.
Normal economic conditions for production also mean a diminishing marginal product
for each input:
∂Q
< 0,
∂K 2
2

(3 )

∂Q
< 0,
∂L2
2

∂2Q
> 0,
∂K ∂NR

∂Q
< 0.
∂NR 2

∂2Q
> 0,
∂NR ∂K

A
Population

In words, the marginal product of natural
resources is greater when either more capital or
more labor is used.
If we take increasing natural resource
scarcity to mean natural resource availability decreases over time, then as capital and labor grow
the production conditions described above can
explain the economic manifestation of natural
resource scarcity and why it might be expected
to limit economic growth. The conditions
expressed in inequalities 4 and 5 show that if
natural resource use declines while capital and
labor grow, the marginal productivity of natural
resources will rise and the marginal productivity
of capital and labor will fall. Hence, increasing
natural resource scarcity would imply that nat-

∂2Q
> 0.
∂L ∂NR

In words, the marginal product of capital and
the marginal product of labor increase when
more of the natural resource is used.
Similarly, the productivity of the natural
resources increases if either capital or labor
increases:
(5)

Per capita
income

For each input, output increases at a decreasing
rate with increased use of the input, as is shown
by the negative second derivative.
Economic theory also suggests how the
increased provision of capital, labor, and natural
resources affects the productivity of each other
input. For instance, the productivity of capital
and labor is expected to increase as natural
resource use increases:
(4 )

∂2Q
> 0.
∂NR ∂L

ECONOMIC AND FINANCIAL REVIEW FIRST QUARTER 2000

Subsistence income

ural resource prices rise relative to wages and
the return to capital.
The economic conditions described above
also suggest that in a world without technological change, output cannot keep pace with population growth unless natural resource use and
capital grow at the same rate. In fact, if natural
resource use grows more slowly than capital
and labor —as greater natural resource scarcity
would imply—output must grow more slowly
than capital and labor unless there is technological change.

source depletion (Solow 1974):
(6)

PNR,t = CNR,t + λe rt,

where PNR,t and CNR,t are the price and marginal
cost of producing the natural resource at time t,
respectively, r is the market interest rate, and
λe rt is the value of holding an additional unit of
the resource off the market until a future period
(a practice economists call “user cost”). The
relationship described by Equation 6 is commonly called the “Hotelling rule.”
With CX,t representing the effects of cumulative production on the cost of producing the
natural resource at time t, Peterson and Fisher
(1977) show

ANOTHER PERSPECTIVE ON
NATURAL RESOURCE SCARCITY

(7)

Hotelling (1931) develops a model to
explain how the prices of nonrenewable natural
resources —such as oil, natural gas, coal, copper, nickel, bauxite, zinc, and iron —would
evolve over time in the absence of technological change. Hotelling’s analysis exploits the
proposition that the quantity of nonrenewable
resources is fixed. The consumption of the
resource today reduces the amount available for
future consumption, and the owner of such a
resource must decide how to distribute its use
over time.
In an economy in which other investments
earn a market rate of interest, individuals saving
nonrenewable natural resources for future periods also must expect to earn the market interest
rate (including the appropriate risk premium). If
the expected return to saving a nonrenewable
natural resource for future periods is less than
the market interest rate, managers of that
resource will save less of it for the future. This
will make the resource more plentiful today and
less plentiful in the future, which will lower
today’s price, raise future prices, and increase
the expected return to saving the resource for
future periods.
On the flip side, if the expected return
is greater than the market interest rate, managers will save more of the resource for future
periods, making it less plentiful today and
more plentiful in the future. This will raise
today’s price, lower future prices, and decrease
the expected return to saving the resource for
the future. Only when the expected return is
equal to the market interest rate will managers
of the resource consider their production
plans finalized. Under these conditions, the
difference between the price and marginal
cost of producing a nonrenewable natural
resource will rise at the market interest rate
unless production costs are affected by re-

•

λ = –e –rtCX,t ,

which means λ is constant over time and the
user cost grows at the interest rate unless production costs change with cumulative extraction
(CX,t ≠ 0). If production costs rise with cumulative extraction (CX,t > 0), the user cost rises more
slowly than the interest rate.2 The price of the
natural resource is expected to rise over time,
however, whether or not production costs rise
with cumulative extraction (CX,t ≥ 0).3
Financial markets and forecasts of future
prices are generally consistent with theory
reflecting expectations that prices for nonrenewable natural resources will rise over long
periods of time.4 In fact, the Hotelling rule is
best interpreted as a market efficiency condition
describing how current and expected future
prices for these resources are simultaneously
determined by current market conditions and
expectations about future market conditions.
For nonrenewable natural resources, current
prices and expectations about future prices
depend on the information and technology
available at the time.
MARKET-INDUCED TECHNOLOGICAL CHANGE
As demonstrated above, if a nonrenewable
natural resource is expected to become more
scarce in an economic sense, its price will be
expected to rise. In a market system, expectations of higher prices increase the incentive to
find new technology that will offset geophysical
scarcity. When they expect higher prices, consumers have an incentive to look for new technology that lets them use less of a natural
resource. When they anticipate higher production costs, producers have an incentive to develop new technology to lower costs. In short,
the very mechanism that signals increasing economic scarcity of a nonrenewable resource helps

FEDERAL RESERVE BANK OF DALLAS

Table 1

Natural Resource Prices Deflated by the Consumer Price Index
Commodity
Aluminum
Anthracite coal
Bituminous coal
Copper
Iron
Lead
Natural gas
Nickel
Oil
Silver
Steel
Tin
Zinc

1870

1880

1890

1900

1910

1920

1930

1940

1950

1960

1970

1980

1990

1998

*
100.00
100.00
100.00
100.00
100.00
*
100.00
100.00
100.00
*
*
100.00

*
87.95
81.90
132.27
112.49
105.67
*
97.25
31.91
113.30
*
100.00
102.96

*
91.90
69.67
103.56
79.15
102.15
*
71.47
28.08
111.11
*
110.51
110.58

55.71
103.42
79.04
116.08
91.99
107.87
*
59.38
46.86
70.86
162.63
166.75
95.54

33.92
117.74
76.00
82.45
71.52
96.31
*
42.41
21.45
55.10
128.97
169.80
104.69

23.21
140.26
118.75
52.28
93.41
81.72
97.78
20.78
50.37
48.57
125.84
112.24
70.57

20.27
177.33
64.60
47.43
47.98
67.42
94.86
20.78
23.43
21.71
87.37
88.39
49.94

18.96
164.84
86.40
49.21
71.10
75.88
66.87
24.74
23.91
23.81
129.27
165.32
81.43

10.46
214.19
128.07
53.65
86.47
113.06
56.26
18.53
34.27
29.32
121.53
184.67
104.66

12.48
152.98
100.89
65.35
98.13
82.23
98.50
24.77
31.96
29.31
161.71
159.39
79.56

10.51
161.58
102.92
89.50
83.42
82.74
91.76
32.93
26.92
43.48
151.56
208.72
71.42

13.12
298.63
224.38
73.52
113.36
105.46
401.75
35.58
86.05
238.64
165.62
477.54
82.20

8.20
177.85
135.58
57.43
NA
73.34
277.63
31.05
51.30
35.83
134.21
140.11
105.35

5.87
122.07
96.32
29.64
NA
58.38
257.20
13.11
22.52
30.64
99.56
109.43
58.67

* All commodities indexed to 1870 = 100 except aluminum (1895 = 100), natural gas (1919 = 100), steel (1897 = 100), and tin (1880 = 100).
SOURCE: Authors’ calculations using data from Bureau of Labor Statistics, Department of the Interior, Department of Energy, and Manthy (1978).

stimulate the technological change that will offset that scarcity.5 Whether technology advances
rapidly enough to prevent a rise in the prices of
the resources, however, is a question best left to
the evidence.

prices and provides a conservative estimate of
the extent to which technological progress has
reduced the scarcity of nonrenewable natural
resources.
The second method, suggested by the production function and used by Nordhaus (1973),
deflates the prices of individual natural resources with the average manufacturing wage.
This approach shows how much human effort is
required to produce a given commodity and
provides an aggressive estimate of the extent
to which technological progress has offset resource scarcity.

WHAT IS THE EVIDENCE?
The conditions described above form a
basis to test whether nonrenewable natural
resources are becoming more scarce in an economic sense or whether technological advance
is making them more plentiful. Rising real prices
for nonrenewable natural resources would provide evidence that technological advance has
not offset increased geophysical scarcity; constant real prices would indicate that technological advance has just offset increased scarcity;
and falling real prices would signify that technological advance has more than offset increased geophysical scarcity.
In this article, we examine trends in the
real prices of twelve nonrenewable natural resources—aluminum, anthracite coal, bituminous
coal, copper, iron, lead, natural gas, nickel,
crude oil, silver, tin and zinc —and one basic
manufactured product, steel, to determine
whether technological change is outpacing geophysical scarcity for nonrenewable natural
resources. To obtain real prices from the nominal ones, we deflate the time series in two ways.
The first method, suggested by the Hotelling
rule and used by Fisher (1979) and Hartwick
and Olewiler (1986), uses an overall price
index, such as the U.S. Consumer Price Index
(CPI), to deflate the prices of individual natural
resources. This approach is the standard
method for converting nominal prices to real

ECONOMIC AND FINANCIAL REVIEW FIRST QUARTER 2000

An Overview of the Price Data
Under the conservative approach of deflating natural resource commodity prices by the
CPI, most series generally decline, as shown in
Table 1.6 All but three of the commodities —
anthracite coal, natural gas, and tin—had lower
real prices in 1998 than they did in the first year
for which data are available. In 1998, the prices
of anthracite coal and tin were 22.07 percent
and 9.43 percent above their respective initial
values. The price of natural gas was 157.2 percent above its 1919 value. The prices of steel
and bituminous coal were 0.44 percent and 3.68
percent below their initial values, respectively.
The prices for the remaining eight commodities
declined by more than 40 percent from the first
year for which we have data to 1998. Most
notable are nickel and aluminum prices, which
in 1998 were 13.11 percent and 5.87 percent of
their initial real prices, respectively.
Under the more aggressive approach of
deflating natural resource commodity prices by
manufacturing wages, we see stronger evidence
of downward trends, as shown in Table 2. By

Table 2

Natural Resource Prices Deflated by Manufacturing Wages
Commodity
Aluminum
Anthracite coal
Bituminous coal
Copper
Iron
Lead
Natural gas
Nickel
Oil
Silver
Steel
Tin
Zinc

1870

1880

1890

1900

1910

1920

1930

1940

1950

1960

1970

1980

1990

1998

*
100.00
100.00
100.00
100.00
100.00
*
100.00
100.00
100.00
*
*
100.00

*
67.12
62.50
100.94
85.85
80.65
*
74.22
24.35
86.47
*
100.00
78.57

*
65.30
49.50
73.58
56.24
72.58
*
50.78
19.95
78.95
*
102.88
78.57

55.71
68.04
52.00
76.37
60.52
70.97
*
39.06
30.83
46.62
162.63
143.75
62.86

29.99
68.49
44.21
47.96
41.61
56.03
*
24.67
12.48
32.05
114.04
129.43
60.90

15.19
60.40
51.14
22.51
40.22
35.19
96.78
8.95
21.69
20.92
82.37
63.33
30.39

11.06
63.64
23.18
17.02
17.22
24.19
78.25
7.46
8.41
7.79
47.66
41.56
17.92

7.24
41.41
21.70
12.36
17.86
19.06
38.61
6.21
6.01
5.98
49.36
54.41
20.45

3.14
42.33
25.31
10.60
17.09
22.35
25.56
3.66
6.77
5.80
36.51
47.83
20.68

2.94
23.70
15.63
10.12
15.20
12.74
35.08
3.84
4.95
4.54
38.08
32.36
12.33

2.19
22.14
14.10
12.26
11.43
11.34
28.90
4.51
3.69
5.96
31.57
37.48
9.79

2.67
40.05
30.09
9.86
15.20
14.14
123.85
4.77
11.54
32.00
33.76
83.92
11.02

1.75
24.92
19.00
8.05
NA
10.28
89.41
4.35
7.19
5.02
28.58
25.72
14.76

1.24
16.98
13.39
4.12
NA
8.12
82.21
1.82
3.13
4.26
21.04
19.94
8.16

αi and βi are parameters to be estimated, and et,i
is a normally distributed error term. As before,
we measure real prices for each of the thirteen
commodities by two methods —deflating with
the CPI and deflating by average U.S. manufacturing wages.
Estimating Equation 8 for the more conservative, CPI-adjusted data yields mixed results,
as shown in Table 3. Prices for five of the commodities —anthracite coal, bituminous coal, natural gas, steel, and tin —show significant positive annual trend rates of growth, varying from
a low of 0.2 percent for steel to a high of 2 percent for natural gas. Prices for iron and crude oil
show no significant trends. Prices for the other
six commodities — aluminum, copper, lead,
nickel, silver, and zinc —show significant negative annual trend rates of growth, varying from
–0.3 percent for lead and zinc to –2.2 percent
for aluminum.
Estimating Equation 8 for the more aggressive, wage-adjusted data yields stronger declines in commodity prices, as shown in Table
4. With the exception of natural gas, all the
commodity price indexes show significant negative trends. Annual rates range from –1.2 percent for anthracite and bituminous coal to –4.1
percent for aluminum. Natural gas has no significant trend.
To control for potential variation of commodity prices over the business cycle, we also
estimate Equation 8 by including measures of
world and U.S. GDP. Although business cycles
are shown to be significant in a few of the real
commodity prices, the signs and significance of
the trend coefficients are substantially similar to
those in Tables 3 and 4.

1998, all the commodities had lower real prices
than they did in the first year for which data are
available, and over half the commodities had
prices that were less than one-tenth of their
initial values. The 1998 prices of anthracite coal,
natural gas, and tin, which show gains in the
CPI-adjusted series, were 16.98 percent, 82.21
percent, and 19.94 percent of their initial values,
respectively. The real 1998 prices of steel and
bituminous coal stood at 21.04 percent and 13.39
percent of their initial values, respectively. The
prices of nickel and aluminum were 1.82 percent
and 1.24 percent of their first reported prices.
Because commodity prices vary over
the business cycle, we also analyze data that
coincided with peaks of both U.S. and world
business cycles. We find substantially similar
price trends to those reported in Tables 1 and 2.
Econometric Tests of Resource Scarcity:
1870–1998
Although prices for most nonrenewable
natural resources generally fell from the first
year for which data are available, they also
exhibited considerable volatility. Over short
periods, price data may reflect a number of market conditions other than resource scarcity and
technological advance, such as monopolization,
cartelization, taxation, and regulation. To
abstract from possible short-term fluctuations,
we test for time trends in the prices of
resources, using annual data from 1870 through
1998 as follows:7
(8)

lnPi = αi + βi t + et,i

for each nonrenewable natural resource i,
where Pi is the real price of resource i, t is time,

FEDERAL RESERVE BANK OF DALLAS

Table 3

Estimated Trends in Natural Resource Prices
Deflated by the CPI, 1870 –1998

Econometric Tests of Resource Scarcity:
Subperiods
When working with such a long time
series, breaks in the trends are possible. Casual
observation suggests the possibility of such
breaks for most price series around the end of
World War II. To test formally for breaks in the
individual series, we conduct Chow tests using
data from 1870 through 1945 in the first period
and 1946 through 1998 in the second period.8
The results show that at the 95 percent confidence level every price series, except lead and
tin deflated by manufacturing wages only, has a
significant break between 1945 and 1946.
Armed with this information, we repeat
the econometric exercises described in Equation
8 for two periods —from 1870 through 1945 and
from 1946 through 1998. For most of the commodities, strong downward trends in prices are
found from 1870 through the end of World War
II, but price declines moderate or reverse in the
postwar era.
With the CPI-deflated commodity prices,
ten of the thirteen pre-1946 series trend downward (Table 5 ). Anthracite coal and tin trend
upward, and bituminous coal shows no price
trend. After 1945, however, price declines moderate. Five of the commodity price series show
significant positive trends, four show no significant trend, and four show significant negative
trends.
With the wage-deflated commodity prices,
all eleven of the pre-1946 series trend downward (Table 6 ). As with the CPI-adjusted data,
price declines moderate after 1945. Four of the
commodity price series show no significant
trend, and six show significant negative trends.
Only natural gas shows a significant positive
trend after 1945.
As we did for the entire sample period, we
control for potential variation of commodity
prices over the business cycle in the subperiods
using measures of both world and U.S. GDP.
Although business cycles are significant in a few
commodity prices, the signs and significance of
the trend coefficients are substantially similar to
those in Tables 5 and 6.

Commodity
Aluminum
Anthracite coal
Bituminous coal
Copper
Iron
Lead
Natural gas
Nickel
Oil
Silver
Steel
Tin
Zinc

Trend growth rate

.73**
1.75**
1.26**
–.54**
4.65**
– 1.75**
– 3.34**
.98**
1.16**
1.05**
1.50**
–.26**
– 1.67**

–.022**
.007**
.006**
–.007**
.001
–.003**
.020**
–.012**
.001
–.007**
.002**
.006**
–.003**

** Denotes significance at the 95 percent confidence level.
SOURCE: Authors’ estimates using data from the Bureau of Labor Statistics,
Department of the Interior, Department of Energy, and Manthy (1978).

Table 4

Estimated Trends in Natural Resource Prices
Deflated by Manufacturing Wages, 1870 –1998
Commodity
Aluminum
Anthracite coal
Bituminous coal
Copper
Iron
Lead
Natural gas
Nickel
Oil
Silver
Steel
Tin
Zinc

Constant

Trend growth rate

1.71**
2.66**
2.18**
.38**
5.65**
–.83**
– 2.64**
1.89**
2.08**
1.97**
2.48**
.67**
–.75**

–.041**
–.012**
–.012**
–.025**
–.019**
–.022**
.004
–.031**
–.017**
–.025**
–.017**
–.013**
–.021**

the more aggressive wage-adjusted data, we
find no significant upward trends in commodity
prices. Breaking the series into two periods,
however, we find evidence that price declines
for nonrenewable natural resources may have
moderated (or reversed for some CPI-adjusted
price series) since World War II.9 Predicting
future price increases from this moderation is
unwarranted, however.10
At issue is whether the more conservative
or the more aggressive approach to analyzing
the price data is more appropriate for assessing
resource scarcity. The CPI-deflated price data
measure the scarcity of the nonrenewable natural

Econometric Tests of Resource
Scarcity Reconsidered
Econometric tests conducted for the entire
period or subperiods generally suggest similar
results for samples that include the post –World
War II data. Using the more conservative CPIadjusted data, we find that real prices for some
nonrenewable natural resources have positive
trends while others have negative trends. Using

ECONOMIC AND FINANCIAL REVIEW FIRST QUARTER 2000

Constant

Table 5

Estimated Trends in Natural Resource Prices Deflated by the CPI, 1870 – 1945 and 1946 – 1998
1870 – 1945
Commodity
Aluminum
Anthracite coal
Bituminous coal
Copper
Iron
Lead
Natural gas
Nickel
Oil
Silver
Steel
Tin
Zinc

Constant

Trend growth rate

1.32**
1.64**
1.43**
–.26**
4.92**
–1.71**
–.22
1.54**
1.52**
1.60**
1.84**
–.25**
–1.54**

–.033**
.010**
0
–.016**
–.008**
–.006**
–.028**
–.030**
–.010**
–.024**
–.006**
.005**
–.007**

1946 – 1998
Commodity

Constant

Trend growth rate

Aluminum
Anthracite coal
Bituminous coal
Copper
Iron
Lead
Natural gas
Nickel
Oil
Silver
Steel
Tin
Zinc

–.48**
2.39**
1.29**
–.41*
3.82**
–.37*
– 5.55**
–.51*
.49
–.50
1.80**
.50
–1.53**

–.010**
0
.006**
–.007**
.012**
–.016**
.041**
.003
.009**
.009**
–.001
–.001
–.004**

** Denotes significance at the 95 percent confidence level.
* Denotes significance at the 90 percent confidence level.
SOURCE: Authors’ estimates using data from Bureau of Labor Statistics, Department of the Interior, Department of Energy, and Manthy (1978).

Table 6

Estimated Trends in Natural Resource Prices Deflated by Manufacturing Wages, 1870 – 1945 and 1946 – 1998
1870 – 1945
Commodity
Aluminum
Anthracite coal
Bituminous coal
Copper
Iron
Lead †
Natural gas
Nickel
Oil
Silver
Steel
Tin †
Zinc

Constant

Trend growth rate

2.59**
2.58**
2.36**
.68**
5.85**

–.057**
–.008**
–.017**
–.034**
–.026**

1.37**
2.48**
2.45**
2.53**
3.14**

–.057**
–.048**
–.028**
–.042**
–.030**

–.61**

–.025**

1946 – 1998
Commodity

Constant

Trend growth rate

–.71**
2.16**
1.07**
–.64**
4.32**

–.017**
–.007**
–.001
–.014**
–.003

– 5.78**
–.74**
.26
–.73*
1.57**

.033**
–.004*
.001
.002
–.008**

–1.75**

–.011**

Aluminum
Anthracite coal
Bituminous coal
Copper
Iron
Lead †
Natural gas
Nickel
Oil
Silver
Steel
Tin †
Zinc

** Denotes significance at the 95 percent confidence level.
* Denotes significance at the 90 percent confidence level.
† Authors chose not to estimate this series in two periods because there was no break in trend.
SOURCE: Authors’ estimates using data from Bureau of Labor Statistics, Department of the Interior, Department of Energy, and Manthy (1978).

resources relative to a given basket of goods.
Because improved technology increases the
availability of all goods, the CPI-deflated measures of prices tend to underestimate the effect
of technological change in increasing the availability of the resources.11
Deflating the price data with manufacturing wages captures technological change that
increases the availability of all goods, but it also
reflects the rising educational attainment of
manufacturing workers from 1870 to 1998. As
such, the wage-deflated price measures tend to
overestimate the effect of technological change
in increasing the availability of nonrenewable

natural resources. The relevant real price—and
the correct assessment —lies somewhere between those found with the two measures.
Table 7 presents a summary of what we can
conclude from the relevant measures of the real
prices of the nonrenewable natural resources in
question. (Also, see the appendix.)
SUMMARY AND CONCLUSIONS
Some observers remain concerned that
increasing natural resource scarcity will limit
future economic growth and human well-being,
while others remain optimistic that technologi-

FEDERAL RESERVE BANK OF DALLAS

Table 7

Summary of Trends in the Real Prices of
Nonrenewable Natural Resources

cal change will overcome geophysical scarcity.
Reliance on free markets can promote the
requisite technological change. The increasing
scarcity of a natural resource increases its price.
When they expect higher prices, consumers
look for technology that lets them use less of a
natural resource. Producers turn to technology
that lowers production costs in expectation of
higher profits.
The question is whether technological
change can outpace geophysical scarcity, and
economic theory suggests a test. Rising real
prices for nonrenewable natural resources
would provide evidence that technological advance has not offset increased geophysical
scarcity; constant real prices would indicate
that technological advance has just offset increased geophysical scarcity; and falling real
prices would signify that technological advance
has more than offset increased geophysical
scarcity.
Using econometric tests to examine the
trends in the real prices of thirteen commodities, we find little evidence of increased natural
resource scarcity from 1870 through 1998. For
none of these commodities do we find conclusive evidence that the relevant real price has
risen. Our results indicate that the relevant real
prices could have risen or remained unchanged
for natural gas; could have risen or fallen for
anthracite coal, bituminous coal, steel, and tin;
could have remained unchanged or fallen for
iron and crude oil; and have fallen for aluminum, copper, lead, nickel, silver, and zinc.
Although we find evidence that price
declines for nonrenewable natural resources
may have moderated (or reversed for some CPIdeflated price series) since World War II, we
find little evidence of increased scarcity. For
only one of the thirteen commodities —natural
gas—do we find conclusive evidence that the
relevant real price has risen. The real price of tin
could have risen or fallen. The real prices could
have risen or remained unchanged for bituminous coal, iron, crude oil, and silver; could have
remained unchanged or fallen for anthracite
coal, nickel, and steel; and have fallen for
aluminum, copper, lead, and zinc.
In short, the evidence suggests that over
the past century, new technology driven by free
market forces has overcome the geophysical
scarcity of nonrenewable natural resources.
Increased reliance on markets during the closing decades of the twentieth century is cause for
optimism that these trends will continue in the
twenty-first.

ECONOMIC AND FINANCIAL REVIEW FIRST QUARTER 2000

Commodity
Aluminum
Anthracite coal
Bituminous coal
Copper
Iron
Lead
Natural gas
Nickel
Oil
Silver
Steel
Tin
Zinc

Whole period (1870 –1998)

Post – World War II

Falling
Rising to falling
Rising to falling
Falling
Unchanged to falling
Falling
Rising to unchanged
Falling
Unchanged to falling
Falling
Rising to falling
Rising to falling
Falling

Falling
Unchanged to falling
Rising to unchanged
Falling
Rising to unchanged
Falling
Rising
Unchanged to falling
Rising to unchanged
Rising to unchanged
Unchanged to falling
Rising to falling
Falling

SOURCE: Authors’ estimates using data from the Bureau of Labor Statistics,
Department of the Interior, Department of Energy, and Manthy (1978).

NOTES

The authors would like to thank W. Michael Cox for
providing manufacturing wage data.
For illustrative purposes, we assume constant returns
to scale for the world economy — that is, a doubling of
all inputs doubles output.
If CX,t is negative, the user cost rises more rapidly than
the interest rate.
For extremely high values of CX,t, the user cost and
price of the natural resource would fall over time.
These conditions do not generally exist. See Dasgupta
and Heal (1979).
Futures markets for nonrenewable natural resources
occasionally go into backwardation, reflecting shortterm supply constraints and the cost to users of stocking out.
Of course, technological advance may occur without
such stimulation, but a historical comparison of the
rates of technological growth in free market economies
with those occurring in the Communist-bloc countries
demonstrates the importance of incentives to technological change.
The 1980 prices show evidence of the commodity
price explosion in the 1970s, as prices for most commodities rise dramatically, then begin to fall.
Price data for aluminum, iron, natural gas, steel, and
tin cover the periods 1895 –1998, 1870 –1981,
1919 – 98, 1897–1998, and 1880 –1998, respectively.
The data may show additional or more-optimal breaks
than between 1945 and 1946, but exhaustive testing of
breaks is of relatively low power econometrically.
The commodity price explosion in the 1970s may
have contributed to the break in trend. Residuals for
trends estimated over the entire period and the
1870 –1945 subperiod are white noise, but residuals

for most trends estimated over the 1946 –98 period are
not.
Using CPI-adjusted data, Slade (1982) uses a quadratic time-trend to predict that prices for nearly all
nonrenewable natural resources would eventually
begin rising. Berck and Roberts (1996) show that other
specifications are preferred and that Slade’s conclusions are unwarranted.
Consider the case in which technology changes in
such a way that all goods and services, including nonrenewable natural resources, could be produced with
half as much effort. The CPI-deflated measure of
prices for nonrenewable natural resources would suggest no change in availability.

Hotelling, H. (1931), “The Economics of Exhaustible
Resources,” Journal of Political Economy 39 (April):
137–75.
Jorgenson, D., and Z. Griliches (1967), “The Explanation
of Productivity Change,” Review of Economics and
Statistics 34: 250 – 82.
Malthus, Thomas R. (1798), An Essay on the Principle of
Population (London: J. Johnson).
Manthy, Robert S. (1978), Natural Resource Commodities: A Century of Statistics (Baltimore: Johns Hopkins
University Press for Resources for the Future).
Meadows, Donnella H., Dennis L. Meadows, Jorgen
Randers, and William W. Behrens (1972), The Limits to
Growth: A Report for the Club of Rome’s Project on the
Predicament of Mankind (New York: Universe).

REFERENCES
Barnett, H. J., and C. Morse (1963), Scarcity and Growth:
The Economics of Natural Resource Scarcity (Baltimore:
Johns Hopkins University Press for Resources for the
Future).

Nordhaus, W. D. (1973), “World Dynamics: Measurement
Without Data,” Economic Journal 83 (December):
1156 – 83.

Berck, Peter, and Michael Roberts (1996), “Natural Resource Prices: Will They Ever Turn Up?” Journal of Environmental Economics and Management 31 (July): 65 –78.

Peterson, F. M., and A. C. Fisher (1977), “The Exploitation of Extractive Resources,” Economic Journal 87
(December): 681–721.

Brown, G. M., and B. C. Field (1978), “Implications of
Alternative Measures of Natural Resource Scarcity,”
Journal of Political Economy 86 (April): 229 – 44.

Schmidt, R. H. (1988), “Hotelling’s Rule Repealed? An
Examination of Exhaustible Resource Pricing,” Federal
Reserve Bank of San Francisco Economic Review, Fall,
41– 54.

Dasgupta, P. S., and G. M. Heal (1979), Economic
Theory and Exhaustible Resources (Cambridge:
Cambridge University Press).

Slade, M. E. (1982), “Trends in Natural Resource Commodity Prices: An Analysis of the Time Domain,” Journal
of Environmental Economics and Management 9 (June):
122 – 37.

Fisher, Anthony C. (1979), “Measurements in Natural
Resource Scarcity,” in Scarcity and Growth Reconsidered,
ed. V. Kerry Smith (Baltimore: Johns Hopkins University
Press).

Solow, Robert W. (1974), “The Economics of Resources
or the Resources of Economics,” American Economic
Review 64 (May): 1–14.

Hartwick, John M., and Nancy D. Olewiler (1986), The
Economics of Natural Resource Use (New York: Harper
and Row).

FEDERAL RESERVE BANK OF DALLAS

Appendix

Trends in Natural Resource Prices
Deflated by the CPI

Deflated by manufacturing wages
Aluminum
Logged values, $/lb.

Logged values, $/lb.

1
.5
0
–.5
–1
–1.5
–2
–2.5
1870 1880 1890 1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000

1
0
–1
–2
–3
–4
1870 1880 1890 1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000

By all measures, relevant real price fell during entire period, as well as after World War II.
Anthracite coal
Logged values, $/ton

Logged values, $/ton

3.5
3
2.5
2
1.5
1
.5
0
1870 1880 1890 1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000

2.5
2
1.5
1
.5
0
1870 1880 1890 1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000

Relevant real price could have risen or fallen during entire period but has remained unchanged or fallen since World War II.
Bituminous coal
Logged values, $/ton

Logged values, $/ton

2.5

1.5

0
1870 1880 1890 1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000

Relevant real price could have risen or fallen during entire period but has risen or remained unchanged since World War II.
Copper
Logged values, $/lb.

Logged values, $/lb.

1.5

–.5

–1

–.5

–1.5

–2

–2.5

–2.5
1870 1880 1890 1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000

–3.5
1870 1880 1890 1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000

By all measures, relevant real price fell during entire period, as well as after World War II.
Iron
Logged values, Index (1967 = 100)

Logged values, Index (1967 = 100)

6
5
4
3
2
1
0
1870 1880 1890 1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000

7
6
5
4
3
2
1
0
1870 1880 1890 1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000

Relevant real price could have remained unchanged or fallen during entire period but has risen or remained unchanged
since World War II.
Actual price

Trend whole period

Trend pre-1946 and post–World War II

(continued on next page)

ECONOMIC AND FINANCIAL REVIEW FIRST QUARTER 2000

Appendix (continued)

Trends in Natural Resource Prices
Deflated by the CPI

Deflated by manufacturing wages
Lead

Logged values, $/lb.

0
–.5
–1
–1.5
–2
–2.5
–3
–3.5
–4
–4.5
1870 1880 1890 1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000

–.5
–1
–1.5
–2
–2.5
–3
1870 1880 1890 1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000

By all measures, relevant real price fell during entire period, as well as after World War II. Chow test indicates price data
deflated by manufacturing wages should not be split into the two periods 1870 –1945 and 1946 –1998.
Natural gas
Logged values, $/1,000 cubic ft.

Logged values, $/1,000 cubic ft.

0
–.5
–1
–1.5
–2
–2.5
–3
–3.5
1870 1880 1890 1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000

–.5
–1
–1.5
–2
–2.5
–3
1870 1880 1890 1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000

Relevant real price could have risen or remained unchanged during entire period but, by all measures, has risen since World
War II.
Nickel
Logged values, $/lb.

Logged values, $/lb.

3
2.5
2
1.5
1
.5
0
–.5
–1
1870 1880 1890 1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000

4
3
2
1
0
–1
–2
–3
1870 1880 1890 1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000

Relevant real price fell during entire period but has remained unchanged or fallen since World War II.
Oil
Logged values, $/bbl.

Logged values, $/bbl.

4
3.5
3
2.5
2
1.5
1
.5
0
–.5
1870 1880 1890 1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000

2.5
2
1.5
1
.5
0
1870 1880 1890 1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000

Relevant real price could have remained unchanged or fallen during entire period but has risen or remained unchanged
since World War II.
Silver
Logged values, $/oz.

Logged values, $/oz.
2.5
2
1.5
1
.5
0
–.5
–1
1870 1880 1890 1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000

3
2.5
2
1.5
1
.5
0
–.5
–1
–1.5
1870 1880 1890 1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000

Relevant real price fell during entire period but has risen or remained unchanged since World War II.
Actual price

Trend whole period

Trend pre-1946 and post–World War II

(continued on next page)

FEDERAL RESERVE BANK OF DALLAS

Appendix (continued)

Trends in Natural Resource Prices
Deflated by the CPI

Deflated by manufacturing wages
Steel
Logged values, $/oz.

Logged values, $/oz.
2.5

2.5
2

1.5

1.5
1

0
1870 1880 1890 1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000

Relevant real price could have risen or fallen during entire period but has remained unchanged or fallen since World War II.
Tin
Logged values, $/lb.

Logged values, $/lb.

1.5

–.5

–1

–1
1870 1880 1890 1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000

–1.5
1870 1880 1890 1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000

Relevant real price could have risen or fallen during entire period, as well as since World War II. Chow test indicates price
data deflated by manufacturing wages should not be split into the two periods 1870 –1945 and 1946 –1998.
Zinc
Logged values, $/lb.

Logged values, $/lb.

0
–.5
–1
–1.5
–2
–2.5
–3
1870 1880 1890 1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000

0
–.5
–1
–1.5
–2
–2.5
–3
–3.5
–4
1870 1880 1890 1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000

By all measures, relevant real price fell during entire period, as well as after World War II.
Actual price

Trend whole period

Trend pre-1946 and post–World War II

SOURCE: Authors’ calculations and estimates using data from the Bureau of Labor Statistics, Department of the Interior, Department of
Energy, and Manthy (1978).

ECONOMIC AND FINANCIAL REVIEW FIRST QUARTER 2000

On January 1, 1999, the European Union
(EU) launched what will surely be one of the
most ambitious political and economic undertakings of the twenty-first century: economic
and monetary union (EMU), incorporating eleven
of the fifteen current members of the EU. A new
currency, the euro, replaced the national currencies of the eleven countries,1 and a new institution, the European Central Bank (ECB), took
over responsibility for monetary policy for the
euro area. Many commentators in the United
States thought EMU would never take place or,
if it did, that it would not last very long. The
successful launch of EMU was thus a surprise in
some quarters, and some of the skeptics have
been forced to reevaluate their positions. EMU
is now one year old, and it seems appropriate
to review what has happened during the first
year and assess the prospects for the future.
Over the course of 1999, the euro depreciated steadily against the dollar. The ECB made
its first rate moves, lowering interest rates in April
in response to deflation risk in the euro area
and raising them in November as the recovery
took hold and the inflation outlook deteriorated. The ECB successfully defended its independence against challenges from the finance
minister of one of the larger member states and
has worked to establish credibility for its commitment to price stability. The TARGET payments system, key to the integration of euro
area money markets, came online and has
operated without any major problems. The euro
has emerged as an important international currency, second only to the dollar. The volume of
international bonds denominated in euros
exceeded dollar-denominated issuance during
1999. The four EU countries that currently do
not participate in EMU all moved closer to eventual membership. However, there were few
moves toward the fiscal, labor, and product
market reforms that may ultimately determine
the fate of EMU.

EMU at 1
Mark A. Wynne

he economic and monetary
union is now one year old,
and it seems appropriate to
review what has happened

during the first year and assess
the prospects for the future.

MAIN DEVELOPMENTS DURING 1999
The euro officially became the currency of
the eleven participating nations on January 1,
1999. The rates to be used for converting
national currency units into euros were
announced on December 31, 1998. During the
changeover weekend, January 1 through
January 3, the financial community had to
reconfigure computer and accounting systems
to handle the new currency. Furthermore, all
government debt of the euro-area countries was
redenominated in euros, as were the share prices

Mark Wynne is a research officer and
senior economist in the Research Department
at the Federal Reserve Bank of Dallas.

FEDERAL RESERVE BANK OF DALLAS

Figure 1

ment decided which countries would participate
in EMU—and the actual launch of EMU.
Perhaps more important for the evolution
of the dollar –euro exchange rate was the fact
that over the course of 1999 the U.S. economy
continued to grow at a robust pace, while the
euro area experienced a growth recession.
Through the third quarter, GDP increased only
2.3 percent in the euro area, and in autumn
1999 the European Commission forecast an
increase of only 2.1 percent for the year as a
whole. Unemployment in the euro area remained stubbornly high, declining from 10.6
percent of the labor force in December 1998 to
9.6 percent at the end of 1999. Evidence
strengthened that trend productivity growth was
accelerating in the United States, but there were
few signs that much-needed structural reforms
were being undertaken in Europe.
It is too early to take the decline as symptomatic of fundamental problems with the new
currency. Over the long run, the nominal
exchange rate of the euro against the dollar will
reflect the relative success of the ECB in maintaining the euro’s purchasing power, but over
the short run, cyclical and other factors will be
more important.
The ECB made its first rate moves in 1999,
lowering its repo rate from 3 percent to 2.5 percent in April and then raising it back to 3 percent in November.4 It is significant that in neither case was there much political opposition
from the countries most likely to have opposed
these moves. The rate cut in April was probably

Dollar–Euro Exchange Rate
Dollars per euro
1.25
1.2
1.15
1.1
1.05
1
.95
.9
1/4/99

3/1/99

4/26/99

6/21/99

8/16/99

10/12/99

12/6/99

SOURCE: Policy Analysis Computing & Information Facility in
Commerce (PACIFIC) Exchange Rate Service
<http://pacific.commerce.ubc.ca/xr/>. Copyright 1998 by
Prof. Werner Antweiler, University of British Columbia,
Vancouver, Canada. Reprinted by permission.

of all companies listed in the euro area, along
with millions of bank accounts.
The most striking and oft analyzed development during 1999 was the steady depreciation of the euro against the dollar. The euro also
declined against the yen and the pound sterling.
When the euro made its debut on world financial markets on January 4, 1999, it was trading
at $1.18. It immediately began to depreciate
against the dollar, coming close to parity (and
briefly below in intraday trading) by December
1999 (Figure 1).2 The depreciation took many
commentators by surprise and was contrary to
the confident predictions of many that the euro
would rapidly appreciate against the dollar,
given the relative current account positions of
the United States and the euro area.
However, if we take a longer-term perspective, the decline of the euro against the dollar over the past year is less remarkable. Figure
2 shows the exchange rate of the euro’s predecessor, the European Currency Unit (ECU),
against the dollar from 1996 through 1998,
along with the exchange rate of the euro against
the dollar during 1999.3 Under the terms of the
transition to EMU, one ECU was required to
equal one euro at midnight December 31, 1998.
As Figure 2 shows, in late 1998, the ECU, or
rather the legacy currencies of the euro, experienced a strong appreciation against the dollar in
the wake of Russia’s default and the failure of
the hedge fund Long Term Capital Management
in the United States. Some of this appreciation
may also have been driven by the “europhoria”
in the period between the Brussels summit in
May 1998 — at which the EU heads of govern-

ECONOMIC AND FINANCIAL REVIEW FIRST QUARTER 2000

Figure 2

Dollar– ECU, Dollar– Euro
Exchange Rates, 1996– 99
Dollars per ECU, dollars per euro
1.4
Launch of EMU
1.3

1.2

1.1

.8
1/2/96

7/2/96

1/2/97

7/2/97

1/2/98

7/2/98

1/2/99

7/2/99

the U.S. CPI, for example, beginning with the
pricing concept. While the U.S. Bureau of Labor
Statistics uses the theory of the cost of living
index as the framework for constructing the U.S.
CPI (U.S. Bureau of Labor Statistics 1997), the
HICP uses “household final monetary consumption,” which means that only the prices paid in
monetary transactions are included. The HICP
does not, therefore, include the imputed costs
of agricultural products grown for personal consumption or the services of owner-occupied
dwellings. The latter is included in the U.S. CPI
and accounts for approximately one-fifth of the
basket.7
A more serious problem from the ECB’s
perspective is that the HICP program only
began in 1997. Aggregate HICP data are available for a slightly longer period, but the fact
remains that the ECB must work with price statistics for which there are a limited number
of observations. Even if a long time series on
prices were available, it is not clear how useful
it would be to the ECB. Since Lucas (1976),
economists have been sensitive to the instability
of estimated empirical relationships in the face of
policy regime changes. While there is some
debate in macroeconomics as to what exactly
constitutes a regime change, few would deny
that EMU is a major change in the monetary policy regime for all the participating countries.

the last thing the rapidly growing economies on
the fringe of the euro area (Ireland, Finland,
Spain and Portugal) needed. Indeed, Ireland,
which has come to be known as the “Celtic
Tiger,” seems to be exhibiting the symptoms of
a classic asset price bubble, with house prices
rising by as much as 20 percent to 30 percent a
year. Likewise, when it came time to raise rates
in November, the sluggish German economy
probably could have benefited from a longer
period of lower interest rates. However, the
ECB’s mandate is to maintain price stability in
the euro area as a whole. Thus, it has explained
its decisions to raise or lower interest rates on
the basis of developments at the euro-area level
rather than in terms of what has happened in
individual member states.5
THE CHALLENGE OF CONDUCTING MONETARY
POLICY FOR THE EURO AREA
One of the most important tasks prior to
EMU was to ensure that the ECB would have at
its disposal adequate statistical information to
make monetary policy decisions for the euro
area. This required some degree of harmonization of statistical practices across the EU, in particular for inflation and monetary statistics.
Primary responsibility for the production of official statistics in the EU rests with Eurostat, which
is one of the Directorates General of the European Commission. Eurostat produces statistics
for the euro area and the member states in conjunction with national statistical institutes and
plays a key role in ensuring that statistics are
harmonized. GDP estimates for the euro area
are constructed on a consistent basis using the
ESA95 version of the European System of
Accounts (ESA). Unemployment rates for the
euro area are calculated using a definition put
forward by the International Labour Office in
1982.6
The ECB defined price stability in terms of
the rate of increase in the Harmonised Index of
Consumer Prices (HICP) for the euro area. The
HICP program originated in the need for a common measure of inflation to assess EMU membership candidates’ compliance with the convergence criteria stipulated in the treaty. The
various national consumer price indexes (CPIs)
differ significantly in their concept and coverage. According to the European Commission
(1998), as much as 13 percent of expenditures
covered by the HICP are excluded from some
national CPIs, while as much as 17 percent of
expenditures covered by some national CPIs are
excluded from the HICP. The HICP differs from

PRICE STABILITY
Article 105 (1) of the Maastricht Treaty
states that the primary objective of the ECB shall
be to maintain price stability but leaves it to the
ECB to define what exactly, in terms of measured inflation, constitutes price stability. Prior
to EMU, the ECB announced that it would
define price stability as a “year-on-year increase
in the Harmonised Index of Consumer Prices
(HICP) for the euro area of below 2%.”
Furthermore, price stability is to be maintained
“over the medium term.” 8 At the launch of EMU,
HICP inflation in the euro area was running at
an annual rate of about 1 percent, having
slowed from rates in excess of 2 percent in early
1996. An energy price deceleration in 1997 and
decline in 1998 contributed significantly to the
favorable inflation situation at the launch of
EMU. However, as Figure 3 shows, during 1999
the inflation rate accelerated as energy prices
started to increase and the euro declined against
the dollar and other major currencies.
Furthermore, there has been some divergence of inflation rates across the euro area
over the past year. Figure 4 shows highest and

FEDERAL RESERVE BANK OF DALLAS

Figure 3

Figure 4

Euro-Area HICP Inflation

Inflation in the Euro Area

Percent

Percent
6

2.5
“Price stability”

Convergence
period

EMU launched

Launch of
EMU

4
1.5
3
1
2
.5

0
1/98

0
4/98

7/98

10/98

1/99

4/99

7/99

10/99

1/96

1/00

7/96

1/97

7/97

Highest

SOURCE: Eurostat.

1/98

7/98

1/99

Convergence value

7/99

Lowest

SOURCE: Eurostat.

lowest inflation rates across the eleven euro
area countries, along with the limit set down in
the Maastricht Treaty.9 Since mid-1998, inflation
in Portugal, Spain, and Ireland has exceeded the
limit set down in the treaty, although as of
December 1999 only Ireland’s inflation rate was
more than 1.5 percentage points above the
average of the three lowest. The ECB does not
yet include a measure of core inflation for the
euro area in the statistical appendix to its
Monthly Bulletin, although Eurostat, the EU’s
statistical agency, does include a core measure
(“All items excluding energy, food, alcohol, and
tobacco”) on its web site.10

These three assumptions, together with a
standard quantity theory view of the determination of the price level, led the Governing
Council to choose a reference value of 4.5 percent for M3 growth during 1999.12 The monthly
statistics on M3 growth are assessed in relation
to this reference value using a centered threemonth moving average of monthly growth rates.
It should be noted that the ECB’s derivation of
the reference value for the euro area’s M3
aggregate is similar to the Bundesbank’s procedure to derive its annual M3 target (see
Deutsche Bundesbank 1995).
As Figure 5 shows, M3 growth drifted
steadily away from its reference over the course
of the year. As of December 1999, M3 growth
was almost 2 percentage points above the reference value. The ECB discounted some of the
deviation as due to temporary factors associated
with the euro’s introduction. The ECB’s failure

THE REFERENCE VALUE FOR M3
The twin pillars of the ECB’s monetary
policy strategy are a reference value for the
growth rate of the broad money aggregate M3
and a broadly based assessment of the outlook
for future price developments and the risks to
price stability in the euro area. The choice of M3
rather than a narrower aggregate was based on
research indicating the M3 aggregate has desirable characteristics in terms of stability and
information about future inflation.11
The reference value for M3 is derived from
three assumptions:

Figure 5

Euro-Area M3 Growth
Percent
7
3-month moving average

Actual

5
4
Reference value

1. Price stability is defined as a rate of
increase in the HICP of 2 percent or less.
2. The trend rate of growth of real GDP
in the euro area is 2 percent to 2.5
percent.
3. The trend rate of decline in M3 velocity
is about 0.5 percent to 1 percent a
year.

ECONOMIC AND FINANCIAL REVIEW FIRST QUARTER 2000

3
EMU launched

2
1
0
1/98

4/98

7/98

10/98

1/99

SOURCE: European Central Bank.

4/99

7/99

10/99

1/00

Monetary Aggregates for the Euro Area
Before EMU, each of the EU member states constructed monetary aggregates
using national definitions that differed across countries. It was not possible to arrive
at a consistent aggregate for the euro area by simply adding together these differing
national aggregates. Thus, a key challenge prior to EMU’s launch was to harmonize
definitions to allow consistent measures to be constructed for the single currency
area. As part of this harmonization process, the European Monetary Institute and the
national central banks developed the concept of a Monetary Financial Institution
(MFI), consisting of three types. The first is central banks. The second is resident
credit institutions as defined by EU law, and the third is “all other resident financial
institutions whose business is to receive deposits and/or close substitutes for
deposits from entities other than MFIs and, for their own account…to grant credits
and/or to make investments in securities.” This third category consists primarily of
money market funds.
The main broad monetary aggregates for the euro area are defined below. The
M1 aggregate consists of currency in circulation and overnight deposits and differs
little from the old national definitions of M1. The category overnight deposits includes
balances on prepaid cards in those countries where prepaid card schemes exist. M2
adds to M1 deposits with agreed maturity up to two years and deposits redeemable
at notice up to three months. The M3 aggregate adds to M2 repurchase agreements,
liabilities of money market funds and debt securities up to two years. Note that prior
to EMU, repurchase agreements were excluded from the national definitions of monetary aggregates in France and Italy, while money market fund shares/units were
included only in the national monetary aggregates of France. For further information
on the new euro-area aggregates and how they relate to old national definitions, see
European Central Bank (1999b).

votes, arguing that making such information
public would increase pressure on council
members to vote along national lines rather than
in the interests of the euro area as a whole (see
Issing 1999).
Transparency in monetary policymaking
has many dimensions, and much of the criticism
of the ECB seems unwarranted. Table 1 compares practices of the ECB, the Federal Reserve
System, and the Bank of England as they relate
to transparency and accountability. The policymaking committee of the ECB — the Governing
Council—meets much more frequently than the
Federal Reserve System’s Federal Open Market
Committee (FOMC) or the Bank of England’s
Monetary Policy Committee (MPC). Through
1999 the ECB’s Governing Council met every
two weeks (except during August) at the ECB’s
headquarters in Frankfurt, although the
Maastricht Treaty requires only that it meet at
least ten times a year (Protocol No. 3 on the
Statute of the European System of Central Banks
and the European Central Bank, Article 10.5). A
press conference was held after the first of the
two meetings in each month, and the tradition
seems to be evolving that rate moves are only
made at the meetings that are followed
by a press conference. At the press conference
the president of the ECB summarizes recent
economic developments, then he and the vice
president hold a question-and-answer session
with journalists. The opening statement and the
Q&A are posted on the ECB’s web site
(http://www.ecb.int) within hours. The ECB
views the press conference, along with the editorial that appears in each issue of its Monthly
Bulletin, as a substitute for the publication of
minutes. (Neither the FOMC nor the MPC holds
a press conference after its meetings.) Transparency is a slippery concept, and there is no
meaningful way to evaluate whether a press
conference following a policy decision constitutes more or less transparency than the publication of votes and minutes.13
The second issue concerns the publication
of forecasts. The Bank of England has been an
innovator in this regard, publishing on a regular
basis its inflation forecast and not just a point
forecast. The FOMC does not publish forecasts
(although the chairman does report the range of
forecasts of committee members in his twiceyearly Humphrey–Hawkins testimony).
Article 109b.3 of the Maastricht Treaty
requires that

Definitions of Euro-Area Monetary Aggregates
Currency in circulation
Overnight deposits
Deposits with agreed maturity up to two years
Deposits redeemable at notice up to three months
Repurchase agreements
Money market fund shares/units and money market paper
Debt securities up to two years

M1
✔
✔

M2
✔
✔
✔
✔

M3
✔
✔
✔
✔
✔
✔
✔

SOURCE: European Central Bank.

to raise interest rates aggressively in response to
the deviation suggests that it may take a pragmatic view of the reference value for M3, much
as the Bundesbank did of its M3 target. From
the time the Bundesbank set its first monetary
target (in 1974) until the start of EMU, it succeeded in hitting its target only about half the
time.
COMMUNICATION: TRANSPARENCY
AND ACCOUNTABILITY
One criticism levied against the ECB during its first year is that it is not sufficiently transparent in making monetary policy decisions
and is not held adequately accountable for
those decisions (see, for example, Buiter 1999
and Begg et al. 1998). The critics argue that
the ECB should publish the minutes of
Governing Council meetings, the votes of individual council members, and the reasoning and
forecasts that underlie council decisions. The
ECB has resisted publication of minutes and

The ECB shall address an annual report on
the activities of the ESCB [European System

FEDERAL RESERVE BANK OF DALLAS

Table 1

Transparency in Monetary Policymaking at the
Federal Reserve, the ECB, and the Bank of England

of Central Banks] and on the monetary policy of both the previous and current year
to the European Parliament, the Council
and the Commission, and also to the
European Council. The President of the
ECB shall present this report to the Council
and to the European Parliament, which
may hold a general debate on that basis.

ECB

Federal Reserve
System

Bank of
England

Governing
Council

Federal Open
Market Committee

Monetary Policy
Committee

Every two weeks

Every six or
seven weeks

Every month

Announced strategy

Yes

Quantitative definition of
price stability

Yes

Publication of forecasts

Not yet

Yes

Publication of minutes

Yes

Publication of votes

Yes

Press conference

Yes

Accountable to elected body

Yes

Policymaking committee
Frequency of meeting

The ECB submitted its first annual report
in April 1999, and the European Parliament’s
Committee on Economic and Monetary Affairs
reviewed it. In its response, the committee
called for greater transparency from the ECB
(see European Parliament 1999). Specifically,
the committee noted that it

SOURCES: European Central Bank, Federal Reserve System, Bank of England.

7. Regrets that the ECB has fallen short of
the transparency practiced by other leading central banks; notes that the U.S.
Federal Reserve Board [sic], Bank of Japan,
Bank of England and Swedish Riksbank
now report both sides of arguments about
monetary actions; and calls for summary
minutes taken at meetings of the ECB
Governing Council to be published shortly
after the following meeting reporting
explicitly the arguments for and against
the decisions taken, as well as the reasoning used in reaching these decisions;
8. Calls on the ECB to publish macroeconomic forecasts on a six-monthly basis
which set out the prospects and the risks
attached to those prospects for: domestic
demand and its principal components, net
exports, nominal and real gross domestic
product, consumer price inflation, unemployment and the current account balance,
together with such relevant data and research on which such forecasts are based,
in order to permit a reliable assessment of
monetary decisions, avoid market misinformation, ensure market transparency and
hence counter speculation;
9. Calls on the ECB to publish a regular
overall report of economic developments
in each of the participating euro-area
countries together with a summary of the
national data which will facilitate comparisons of best practice; enable early warnings of potential problems within the euroarea which might require policy action
by respective governments; and inform
national wage bargainers of sustainable
earnings developments given their own
productivity, price and competitiveness
trends.…

ECONOMIC AND FINANCIAL REVIEW FIRST QUARTER 2000

At the subsequent hearings the ECB president acceded to the request to publish forecasts
and promised they would be published during
2000, along with the economic models used to
produce these forecasts. However, he rejected
the request that the ECB publish summary minutes, arguing as before that the information the
ECB provided at its press conferences and in its
Monthly Bulletin came “very close in substance
to the publication of summary minutes.” He also
rejected calls for reports on each euro-area
country, arguing that the production of such
reports would impede the development of a
euro-area perspective. The Committee on Economic and Monetary Affairs called for publication of votes on monetary policy actions after a
two-year delay, but this proposal was rejected
when put to a vote of the full European Parliament.
Concerns about the ECB’s accountability
to the European electorate have two dimensions. The first is whether the provisions of the
Maastricht Treaty that require the ECB to report
to the European Parliament satisfy the need of
accountability in a democratic society. The second is whether the European Parliament has the
stature to represent the European electorate’s
concerns. Regarding the latter, two significant
developments took place during 1999. In
March, the Parliament for the first time forced
the resignation of the European Commission
over allegations of financial misconduct,
thereby enhancing the Parliament’s standing
among EU institutions and its authority as the
representative body of the EU electorate. And
on May 1, the Amsterdam Treaty entered into

Table 2

Functions of International Currencies
Private use

their currencies to the euro through currency
board arrangements. A larger group of countries
(Cyprus, Macedonia, Cape Verde, Comoros, and
the fourteen countries of the West African
Colonies Françaises d’Afrique [CFA] zone) had
more traditional fixed exchange rate pegs to the
euro. Denmark and Greece are also pegged to
the euro, albeit under a cooperative arrangement under the terms of ERM II, the successor
to the Exchange Rate Mechanism (ERM) of the
European Monetary System.15 A third group
(Croatia, the Czech Republic, the Slovak
Republic, and Slovenia) has managed floats visà-vis the euro. A fourth group (Hungary,
Iceland, Malta, Poland, Turkey, Bangladesh,
Botswana, Burundi, Chile, Israel, and the
Seychelles) has either fixed or crawling pegs to
baskets of currencies that include the euro.
Finally, a fifth group of countries pegs to the
Special Drawing Right (SDR) issued by the
International Monetary Fund in which the euro
has a weight of about one quarter. (The other
currencies in the SDR basket are the U.S. dollar,
the Japanese yen, and the pound sterling).
Perhaps the most significant benefit to the
EU from internationalization of the euro would
be the seigniorage revenue it would earn from
foreign demand for euros. Although euro notes
and coins will not be introduced until 2002, it is
worth considering the revenue this may generate. At the end of 1999, approximately $600 billion of U.S. currency was in circulation.
According to Porter and Judson (1996), more
than half the stock of U.S. currency —and possibly as much as 70 percent—was held outside
the United States at the end of 1995. If we
choose a conservative estimate of 50 percent
and assume that absent these foreign holdings
the federal government would have to issue an
equivalent amount of short-term debt at the
then-prevailing interest rate of 5.3 percent, the
flow of seigniorage to the U.S. Treasury from
the foreign holdings was about $15.6 billion (=
$600 billion × 50 percent × 5.3 percent). As of
November 1999, there was approximately €330
billion of currency outstanding in the euro area.
Since euro notes and coins have not yet been
introduced, this total consists of the notes and
coins of the ten legacy currencies (Luxembourg
was in a monetary union with Belgium prior to
EMU). It is unlikely that many of the legacy currencies circulated to a significant extent beyond
their national borders, with the exception of the
Deutsche mark. Seitz (1995) estimates that
approximately 40 percent of the stock of
Deutsche marks circulates outside Germany. In
November 1999, Deutsche mark notes and coins

Official use

Unit of account

Pricing/quotation currency

Pegging currency

Medium of exchange

Payment/vehicle currency
In exchanges of goods and services
In currency exchange

Intervention currency

Store of value

Investment/financing currency

Reserve currency

SOURCES: Cohen (1971), Hartmann (1998).

force, substantially extending the right of codecision of the European Parliament, making it
the council’s legislative equal in many areas.14
EMERGENCE OF THE EURO
AS AN INTERNATIONAL CURRENCY
Prior to the euro’s launch, there was much
discussion about the extent to which it would
compete with or even displace the dollar as
the world’s most important international currency. Some argued it would take a long time
for the euro to replace the dollar in international
transactions because of network effects. (I find
it more useful to conduct transactions
in dollars when more of my trading and investment partners also conduct transactions in
dollars). Others argued that EMU itself was a
shock of sufficient magnitude to trigger rapid
adoption of the euro (see, in particular, Portes
and Rey 1998).
The ECB has stated repeatedly that “internationalisation of the euro…is not a policy
objective…[and] will be neither fostered nor
hindered by the Eurosystem.” Table 2 lists the
main functions of international currencies, using
the traditional classification of the functions of
money (see Cohen 1971 and Hartmann 1998).
The U.S. dollar is used to quote prices for industrial commodities, and many countries maintain
some type of currency peg to the dollar. There
are significant holdings of U.S. dollars in countries that have experienced high inflation, while
foreign central banks typically use dollars to
intervene in foreign exchange markets to support their local currency. Until last year the dollar was the currency of choice for international
bond issuance, and most central banks continue
to hold the bulk of their foreign exchange
reserves in dollar-denominated assets.
Since the introduction of the euro, most
commodity prices continue to be quoted in dollars, but large European firms now use the euro
for quotation purposes. For instance, Airbus no
longer uses the dollar to quote aircraft prices. As
of the end of 1999, three countries (Estonia,
Bulgaria, and Bosnia–Herzegovina) were pegging

FEDERAL RESERVE BANK OF DALLAS

in circulation amounted to €126 billion, or about
38 percent of the euro-area total. Thus, the estimated seigniorage revenue currently accruing to
the euro area (specifically, to Germany) from
non-euro-area holdings of Deutsche marks
amounts to about €2 billion a year (= €126 billion × 40 percent × 4 percent, using the interest
rate on two-year euro-area government bonds
as of November 1999 as an estimate of what the
government would have to pay to raise the
funds by borrowing).16 This probably constitutes
a lower bound on the amount of seigniorage
the EU will earn from non-EU holdings of the
euro once the notes and coins are introduced.
The euro’s domestic habitat is significantly
larger in economic terms than that of the
Deutsche mark, making the euro more attractive
to non-EU residents than the Deutsche mark
was. The estimated foreign seigniorage revenue
currently earned by the United States is probably an upper bound on what the EU can expect
to earn.
Euro notes will include €100, €200, and
€500 denominations.17 Currently, the highest
denomination note issued by the Federal
Reserve is the $100 bill. Higher denomination
notes may make the euro an attractive alternative to the dollar as a store of value in countries
undergoing high inflation. It may also make
the euro more attractive for transactions in
the underground economy. The existence of
high-denomination euro notes in and of itself
will not cause individuals who currently hold
dollars as a secure store of value in high-inflation countries or for illicit purposes to immediately switch to euros. These individuals will
also have to be convinced that the euro will
retain its value as well as, or better than, the dollar. This, in turn, will depend on the ECB’s track
record in maintaining price stability in the euro
area.

ically, although most banks seem to be maintaining one or two correspondent accounts for
each euro-area country until the euro notes and
coins are introduced in 2002.
The TARGET system was created, first, to
provide a pan-European payments system that
would integrate national money markets and
support the monetary policy of the ECB, and
second, to safeguard financial markets and institutions from systemic events. The former was
accomplished by linking the existing national
payments systems. The latter was accomplished
by moving to a real-time gross-settlement standard for national payments systems prior to
EMU and away from end-of-day settlement, or
netting systems, in which participants accumulate large open positions against their counterparties.
On January 4, 1999, its first day of operation, the TARGET system processed about
156,000 payments, with a total value of about
€1.18 trillion. Of these, about 5,000 were crossborder payments, totaling about €245 billion.
The volume of cross-border payments rapidly
increased to 20,000 to 30,000 a day, with a total
value between €300 billion and €400 billion,
after only a week of operation. The successful
launch of TARGET —and the consolidation of
national money markets—was reflected in the
rapid reduction in interest rate spreads in
overnight money markets in January 1999.
Of the other systems available for processing payments in euros, the three largest are
Euro 1, Euro Access Frankfurt (EAF), and the
Système Net Protégé (SNP) (known since April
1999 as Paris Net Settlement, PNS). There are
also two smaller local systems: Servicio Español
de Pagos Interbancarios (SEPI) in Spain and
Pankkien väliset On-line Pikasiirot ja Sekit
(POPS) in Finland. Together these systems settle
a daily average volume of €400 billion, and the
Euro 1 system (a cooperative undertaking
between EU-based commercial banks and the
EU branches of foreign banks) is by far the most
extensively used alternative to TARGET. The
existence of competitively priced alternative
payments systems caused some concern (see,
for example, Prati and Schinasi 1999) that
TARGET might not attract the volume of highvalue payments needed to significantly contribute to a lowering of payments-system systemic risk. That concern appears to have been
unfounded: through September 1999, the average value of TARGET payments was €5.8 million. The average value of cross-border payments was €12.9 million, while the average
value of domestic payments was €4.4 million.

TARGET
The architects of EMU faced a key challenge in the creation of a payments system that
integrated money markets in all EU countries.
The TARGET system (TARGET stands for Transeuropean Automated Real-time Gross settlement
Express Transfer) consists of fifteen national
real-time gross settlement systems and the ECB
payment mechanism. It provides a uniform platform for processing cross-border payments.
Prior to EMU, payments between EU countries
relied almost exclusively on correspondent
banking arrangements. Since the beginning of
1999, these relationships have declined dramat-

ECONOMIC AND FINANCIAL REVIEW FIRST QUARTER 2000

The average values of the payments settled by
the three biggest other systems (Euro 1, EAF,
and PNS) were €2.8 million, €3.3 million, and
€4.5 million, respectively.

take the UK into EMU when the time is right. In
late 1997 the UK Treasury announced five economic tests that would be used to determine
when the UK should join (see HM Treasury
1997):

WHAT ABOUT THE OUTS?

1. Are business cycles and economic structures compatible so that the UK and
other members of EMU could live comfortably with a common interest rate on
a permanent basis?
2. If problems emerge, is there sufficient
flexibility to deal with them?
3. Would EMU membership enhance the
attractiveness of the UK to overseas
investors?
4. How would EMU membership affect
the competitive position of the UK’s
financial services industry?
5. Will EMU membership promote higher
growth, stability, and a lasting increase
in jobs?

Not all fifteen members of the EU chose to
participate in EMU from the outset. Greece
failed to meet the convergence criteria laid
down in the Maastricht Treaty, while the UK,
Sweden, and Denmark chose to stay out for
domestic political reasons. Greece formally
applied for membership in March and hopes to
become a member at the beginning of next
year. As part of the convergence process, the
Greek drachma was revalued on January 17,
2000. The situation in the UK, Sweden, and
Denmark as to eventual membership in EMU is
less clear.
When the Maastricht Treaty was first put to
a referendum in Denmark, it was decisively rejected by the electorate. The treaty was ratified
in a subsequent referendum, but only after it
had been amended to provide an opt-out from
the single currency for Denmark (Protocol No.
12 of the Maastricht Treaty). However, since the
start of EMU the Danish krone has been pegged
to the euro with a ±2.25 percent fluctuation band
under the terms of ERM II, meaning that, in
effect, Danish monetary policy is dictated by the
ECB. The Danish prime minister has already
launched a political campaign to bring Denmark
into EMU, and in September the ruling Social
Democrats will hold a referendum on Denmark’s
entry into EMU.
Although Sweden satisfied all the convergence criteria for participation in EMU, it did not
join at the outset because of domestic
“Euroscepticism.” Some of this skepticism
waned in the closing months of 1998, when
Denmark and Sweden were more adversely
impacted by fallout from the Russian default
than was Finland, which had elected to join
EMU. Over the past year, attitudes in Sweden
have wavered between joining and not joining.
However, in January the ruling Social
Democratic Party announced for the first time
that it formally supports Swedish membership in
EMU.
Which leaves only the UK. The government secured an opt-out from EMU when the
Maastricht Treaty was negotiated (Protocol No.
11 of the Maastricht Treaty). With the change of
government in the UK in 1997, official attitudes
toward the EU changed significantly, and the
new Labor government declared its intention to

These tests are sufficiently vague that the
government could easily announce that the tests
are satisfied at any time. The UK took a further
step forward in February 1999 with the publication of a National Changeover Plan (HM
Treasury 1999) that details how UK membership
in EMU might come about and presents a
timetable for replacing sterling with the euro.
A more binding constraint on UK membership is the Labor government’s commitment
to put the issue to a referendum. As Figure 6
shows, the UK public remains skeptical about
the single currency, and in the June 1999 elections to the European Parliament, the anti-euro
Conservative Party won 36 seats, compared with
the Labor Party’s 29 seats. However, while public opinion in the UK remains decidedly against
membership in EMU, a significant segment of
British industry believes it is in the UK’s interest
to join. A June 1999 survey of members of the
Institute of Directors revealed that 67 percent
were in favor of the UK joining the single currency (in principle). In July the Confederation of
British Industry (CBI) announced that it was in
favor of the UK joining EMU. The CBI adopted
a pro-EMU stance after a poll of its members
showed that some 52 percent backed eventual
membership.18 However, the CBI has subsequently announced that it will no longer
actively campaign for UK membership until the
government takes a more active role in promoting the issue.
Opponents of UK membership in EMU
often argue that the UK business cycle is more
closely aligned with the U.S. business cycle than

FEDERAL RESERVE BANK OF DALLAS

Figure 6

trade links between countries, the more highly
correlated their business cycles are.

UK Attitudes to the Single Currency
Percent
70
4/98

60
50

OUTLOOK

6/99

2/95

I noted at the beginning of this article
that many commentators in the United States
doubted EMU would ever happen or thought
that, if it did, it would be a source of conflict
within the EU and between the EU and the
United States (see Feldstein 1997a,b). The common thread in the skeptics’ arguments was that
the EU does not constitute an optimum currency area in the sense of Mundell (1961).19
While there were some differences in economic
performance across the euro area over the past
year, we did not see the kind of dramatic asymmetries the skeptics believe will cause EMU to
collapse. Despite sluggish growth in two of the
larger economies (Germany and Italy), unemployment continued to decline across the
euro area, although it does remain at unacceptably high levels. Germany, which accounts for
about one-third of euro-area economic activity,
only experienced one quarter of negative growth
(at the end of 1998) rather than a full-blown
recession. How well the institutions of EMU will
deal with more severely asymmetric cycles if
and when they occur is an open question.20
In the near term it is also essential that the
EU address the issue of lender of last resort for
the euro area. The ECB has a very limited role
in bank supervision and regulation, and the
Maastricht Treaty does not spell out what
exactly the responsibilities of the ECB are in the
event of a major financial crisis. Article 105 of
the Maastricht Treaty mandates that the
European System of Central Banks (ESCB) shall
“promote the smooth operation of the payments
system.” The same article also states that “the
ESCB shall contribute to the smooth conduct of
policies pursued by the competent authorities
relating to the prudential supervision of credit
institutions and the stability of the financial system” and that the European Council may confer
upon the ECB specific tasks related to supervision. Begg et al. (1998) argue that the current
arrangements are unsafe and that there is no
secure mechanism for creating liquidity in the
event of a crisis. Banking supervision remains a
national responsibility, and there are questions
about whether the ECB would have access to
the relevant information to allow it to make
quick decisions if a crisis occurs.21 The
European Parliament’s Committee on Economic
and Monetary Affairs (EPCEMA) recently noted
that “…the ESCB’s arrangements for the emer-

5/99
5/98

5/99

5/98

2/95

4/98

6/99
5/99

5/98

0
2/95

6/99

4/98

8/95

2/96

8/96

2/97

Vote not to join

SOURCE:

8/97

2/98

8/98

Vote to join

2/99

8/99

1/00

Don’t know

ICM Research.

with the cycle in continental European countries
and that the criterion of cyclical convergence
will never be satisfied. This fact is documented
by Wynne and Koo (forthcoming), among many
others. They show that the correlation between
the cyclical component of output in the UK and
the United States is 0.67, which exceeds the correlation of UK output with that in France (0.58)
or Germany (0.45). The relative magnitudes are
similar if we look at employment instead of output. However, the relevance of this fact to the
debate about UK membership in EMU is not
obvious. To begin with, we do not fully understand why the UK business cycle is more closely
correlated with the U.S. cycle than with the
cycle in the rest of Europe. The correlation may
reflect the significant volume of trade and
investment flows between the UK and the
United States (most U.S. foreign direct investment in Europe goes to the UK), or it may be
due to other factors.
These flows, in turn, may be influenced
over time by the UK’s attitude toward EMU. If
the UK were to remain outside EMU permanently, some of these investment flows might
shift to the euro area. Already a number of
Asian investors in the UK have indicated they
will rethink their location choices should the UK
delay for long its decision on EMU membership.
Rose (1999) presents evidence suggesting the
real effects of a monetary union may be substantial. Specifically, he shows that two countries that share a common currency tend to
trade three times as much as they would if they
had different currencies. Furthermore, Frankel
and Rose (1998) demonstrate that the closer the

ECONOMIC AND FINANCIAL REVIEW FIRST QUARTER 2000

gency provision of liquidity to financial institutions in distress have been called into question
by the International Monetary Fund and by private sector observers, and EPCEMA urges the
ESCB to make clear that the necessary procedures for approval and disbursement of such
‘lender of last resort’ facilities are in place and
have been rehearsed.”
In its convergence report prepared as part
of the transition to EMU, the European
Monetary Institute (the forerunner of the ECB)
drew attention to the long-term problems posed
by pay-as-you-go pension systems in the EU.22
The ECB reiterated this point in its January 2000
Monthly Bulletin, noting that “the ageing of
populations represents a serious challenge to
the sustainability of the pay-as-you-go financed
public pension schemes” in the euro area. To
give some sense of the scale of the problem
faced by the euro-area economy, Figure 7 presents projections of the number of potential
workers per retired person over the next fifty
years for the United States and the EU.23
The decline in the ratio in the United States
reflects the aging of the baby-boom generation
and is the primary demographic factor fueling
the debate over the long-term sustainability of
the Social Security program here. However, as
Figure 7 shows, the aging problem is more
severe in the EU than in the United States. The
figure presents four variants for the EU. The first
two are for the euro area (EU11) and the current
fifteen members of the EU (EU15). Variant 3
(EU21) shows the projections if the EU expands
to include the six current applicants considered
the most likely candidates for early membership
(Estonia, Poland, the Czech Republic, Hungary,
Slovenia, and Cyprus). The final variant (EU28)

shows what happens if the EU expands to
include all thirteen of the current applicants (in
addition to the six just mentioned, Latvia,
Lithuania, the Slovak Republic, Bulgaria,
Romania, Malta, and Turkey).
The rapid rise in the dependency ratio
(decline in the number of workers per retiree) in
the EU reflects declining birth rates and
increased longevity. The decline in the birth rate
in three of the largest euro-area economies
(Germany, Italy, and Spain) has been so dramatic in recent years that, were it not for immigration, the populations of these countries
would have fallen.24 The aging of the population
might not be so problematic were it not for the
extensive reliance on publicly funded pensions
in these countries and the relatively generous
nature of these pensions. In Germany, for example, workers are entitled to a public pension
equal to 72 percent of their average net lifetime
earnings. Additionally, public expenditure on
health care for the elderly is high and has risen
with recent costly advances in medical technology. In short, demographic developments over
the next decades could prove a serious threat to
the fiscal positions of many of the euro-area governments that will necessitate painful reforms at
some point. Some changes have recently been
made (France now indexes pensions to prices
rather than wages; Germany switched from
indexing to gross wages to indexing to net
wages), but more remains to be done.
Obviously the aging of the EU population
is independent of whether the countries share a
common currency. Rather, its significance stems
from the institutional framework of EMU and, in
particular, the restrictions on national fiscal policies as set out in the Maastricht Treaty and elaborated upon in the Growth and Stability Pact.
Article 104 of the Maastricht Treaty states
that

Figure 7

Aging Populations in the EU and U.S.
1. Member States shall avoid excessive government deficits.
2. The Commission shall monitor the development of the budgetary situation and
of the stock of government debt in the
Member States with a view to identifying
gross errors. In particular it shall examine
compliance with budgetary discipline on
the basis of the following two criteria:
(a) whether the ratio of the planned or
actual government deficit to gross domestic product exceeds a reference value….
(b) whether the ratio of government debt
to gross domestic product exceeds a reference value….

Workers per retiree
4.5
4
3.5
3
2.5
2
1.5
1
U.S.

.5
0

1990

SOURCE:

2000

EU11

2010

EU15

2020

EU21

2030

EU28

2040

2050

United Nations.

FEDERAL RESERVE BANK OF DALLAS

5. If the Commission considers that an
excessive deficit in a Member State exists
or may occur, the Commission shall
address an opinion to the Council.
6. The Council shall, acting by a qualified
majority on a recommendation from the
Commission…decide after an overall assessment whether an excessive deficit exists.
7. Where the existence of an excessive
deficit is decided…the Council shall make
a recommendation to the Member State
concerned with a view to bringing that situation to an end within a given period….
9. If a Member State persists in failing to
put into practice the recommendations of
the Council, the Council may decide to
give notice to the Member State to take,
within a specified time-limit, measures for
the deficit reduction which is judged necessary by the Council in order to remedy
the situation….
11. As long as a Member State fails to comply with a decision taken in accordance
with paragraph 9, the Council may decide
to apply or, as the case may be, intensify
one or more of the following measures:
– to require the Member State concerned
to publish additional information, to be
specified by the Council, before issuing
bonds and securities;
– to invite the European Investment bank
to reconsider its lending policy towards
the Member State concerned;
– to require the Member State concerned
to make a non-interest-bearing deposit of
an appropriate size with the Community
until the excessive deficit has, in the view
of the Council, been corrected;
– to impose fines of an appropriate size.

successfully taken over monetary policy for the
euro area. The ECB faced the first serious challenge to its independence and effectively defended its status. It also conducted its first policy moves, easing monetary policy in April in
the face of a growing threat of deflation and
weak real activity in the euro area. In November
it reversed course, tightening policy as the balance of risks shifted to higher inflation, and the
euro-area recovery took hold.
The success of the first year does not
mean that it will be all plain sailing from here
on. Many challenges remain, and how the EU
and the ECB tackle these will determine the ultimate fate of EMU. One issue highlighted in this
article is the rapidly aging population of the EU.
The aging of the population over the coming
decades in conjunction with generous pension
provisions will put a severe strain on the public
finances of the euro-area economies. One solution might be to admit large numbers of immigrants, but Europe does not have a tradition of
encouraging large-scale immigration. The only
alternative is drastic reform of the public pension programs in all the countries, something no
government has yet been willing to tackle. More
generally, structural reforms of labor and product markets are crucial if the EU is to address
the high unemployment rates and sluggish
growth that have plagued it for the past decade.
Small moves have been made in this direction,
but a lot more needs to be done.
NOTES

The Growth and Stability Pact adopted at
the Dublin Summit in December 1996 is intended to clarify and strengthen the provisions
of the treaty in regard to excessive deficits by
strengthening fiscal discipline under EMU.25 The
existence of large, unfunded public pension liabilities will certainly complicate EMU participants’ ability to abide by the terms of the treaty
and the Growth and Stability Pact.26

CONCLUSIONS
By any reasonable standards, the first year
of EMU must be judged a success. The
changeover weekend went by without incident,
the TARGET payments system was launched
without any major problems, and the ECB has

ECONOMIC AND FINANCIAL REVIEW FIRST QUARTER 2000

I thank Bill Gruben, Evan Koenig and Carlos Zarazaga
for comments on an earlier draft and Eric Millis for
research assistance. Martin Boon at ICM Research in
London kindly supplied the results of the ICM
Research/Guardian polls of UK attitudes to the single
currency. Responsibility for remaining errors rests with
the author.
The eleven countries participating in EMU are Austria,
Belgium, Finland, France, Germany, Ireland, Italy,
Luxembourg, the Netherlands, Portugal, and Spain.
On January 27 the euro closed at below parity for the
first time ($0.9883 in New York).
The European Currency Unit (ECU) was a synthetic currency defined on the basis of a basket of the currencies
of the EU member states. Specifically, on December 31,
1998, one ECU consisted of 3.301 Belgian francs,
0.6242 German marks, 0.1976 Danish krones, 6.885
Spanish pesetas, 1.332 French francs, 0.08784 British
pounds, 1.44 Greek drachmas, 0.008552 Irish punts,
151.8 Italian lira, 0.13 Luxembourg francs, 0.2198
Dutch guilders and 1.393 Portuguese escudos (see
European Central Bank 1999a, 72).

10
11

mists in early 1999 and found that about 65 percent
favored UK membership in EMU.

Arguably the first policy action of the ECB was taken
in December 1998, when the eleven euro-area central
banks (the so-called Eurosystem) coordinated a
reduction in their short-term interest rates to a
common 3 percent level before the formal launch of
EMU.
To this end, in its Monthly Bulletin the ECB publishes statistics only for the euro area as a whole and not for individual member states. Statistical information is provided
on developments in the four EU countries that do not participate in EMU (Denmark, Greece, Sweden, and the UK)
and also on developments in the United States and
Japan.
Formally, people are counted as unemployed if they are
without work, are available to start work in the next two
weeks, and have actively sought employment at some
point during the previous four weeks.
The relative importance of owner’s equivalent rent in the
U.S. CPI as of December 1997 was just over 20 percent.
Interestingly, the ECB does not define how long the
“medium term” is.
The Maastricht Treaty stipulates that, as one of the
convergence criteria for assessing suitability for EMU
membership, a country’s inflation rate should not
exceed the average rate of the three best performers
by more than 1.5 percentage points.
See http://europa.eu.int/comm/eurostat/.
See, for example, the recent working paper by Coenen
and Vega (1999), which builds on other research conducted by the ECB’s predecessor, the European
Monetary Institute.
In December 1999, the Governing Council announced
that this value will also be used for 2000.
Note also that the president of the ECB has indicated
that none of the decisions to change interest rates
were made by a formal vote.
One of the objectives of the Intergovernment
Conference that drew up the Amsterdam Treaty, which
was signed in October 1997, was to enhance the
democratic accountability of EU institutions.

21
22

The main components of the European Monetary
System, which existed prior to EMU, were the Exchange
Rate Mechanism, which was essentially a system of
fixed exchange rates between the currencies of the participating countries, and the European Currency Unit,
which has now been replaced by the euro.
Note that the seigniorage revenue will be distributed
among participating countries using a formula prescribed in the Maastricht Treaty Protocol No. 3 on the
Statute of the European System of Central Banks and
the European Central Bank, Articles 29 and 31.
The denominational structure of the euro will consist of
coins at the 1, 2, 5, 10, and 20 euro cent denominations, coins at the €1 and €2 denominations, and
notes at the €5, €10, €20, €50, €100, €200, and €500
denominations.
The Economist newspaper surveyed British econo-

Ironically, the critics seem to overlook the later papers
by Mundell (1973a,b) in which he proposes additional
criteria for evaluating the suitability of a single currency for a group of countries. As a result of these
works, he has been referred to in some circles as the
father of the euro. See also Mundell (1998a,b).
The studies of Frankel and Rose (1998) and Rose
(1999) just cited are also relevant to this question.
Insofar as sharing a common currency enhances trade
flows within the euro area and these trade flows lead
to more synchronous business cycles, the concern
about asymmetric shocks may prove unfounded.
However, within a monetary union as long-standing
and fully credible as the United States, asymmetric
cycles may occasionally emerge. Through the 1980s
and 1990s different regions of the United States experienced shocks that caused localized recessions of
varying degrees of severity; the term “rolling recessions” entered policy debates to describe this phenomenon.
Prati and Schinasi (1999) articulate similar concerns.
See also the recent report by the G-10 (Group of Ten
1998).
Specifically, the figure shows the ratio of the population aged 25 to 64 to the population aged 65 and
older and is taken from the “medium variant” projections in United Nations (1998).
In its most recent forecasts the United Nations (1998)
projects that the population of Italy will fall from 57.3
million in 2000 to 41.2 million in 2050, that of Germany
from 82.2 million to 73.3 million, and that of Spain from
39.6 million to 30.2 million.
For further details see the May 1999 issue of the ECB’s
Monthly Bulletin.
The need for fiscal rules under a monetary union is a
contentious issue. Artis and Winkler (1997) argue that
the excessive deficit provisions of the treaty can be
justified on the grounds that under monetary union the
costs of an overly expansionary fiscal policy will be
borne by all members of the monetary union and not
just by the country pursuing the policy, creating an
incentive for countries to be more lax with their fiscal
policy. Bergin (2000), arguing from the perspective of
the fiscal theory of the price level, makes a similar
point.

REFERENCES
Artis, Michael J., and Bernhard Winkler (1997), “The Stability
Pact: Safeguarding the Credibility of the European Central
Bank,” European University Institute Working Paper RSC no.
97/54 (Florence, Italy: European University Institute).
Begg, David, Paul De Grauwe, Francesco Giavazzi, Harald
Uhlig, and Charles Wyplosz (1998), “The ECB: Safe at Any

FEDERAL RESERVE BANK OF DALLAS

Speed?” Monitoring the European Central Bank, no. 1
(London: Center for Economic Policy Research).

Hartmann, Philipp (1998), Currency Competition and
Foreign Exchange Markets: The Dollar, the Yen and the
Euro (Cambridge: Cambridge University Press).

Bergin, Paul R. (2000), “Fiscal Solvency and Price Level
Determination in a Monetary Union,” Journal of Monetary
Economics 45 (February): 37– 53.

HM Treasury (1997), UK Membership of the Single
Currency: An Assessment of the Five Economic Tests
(London: HM Treasury).

Buiter, Willem H. (1999), “Alice in Euroland,” Center for
Economic Policy Research Policy Paper no. 1.

——— (1999), Outline National Changeover Plan
(London: HM Treasury).

Buiter, Willem H., Giancarlo Corsetti, and Nouriel
Roubini (1993), “Excessive Deficits: Sense and Nonsense
in the Treaty of Maastricht,” Economic Policy 16 (April):
57–100.

Issing, Otmar (1999), “The Eurosystem: Transparent and
Accountable or ‘Willem in Euroland,’ ” Center for
Economic Policy Research Policy Paper no. 2.

Coenen, Günter, and Juan-Luis Vega (1999), “The
Demand for M3 in the Euro Area,” European Central
Bank Working Paper no. 6.

Lucas, Robert E., Jr. (1976), “Econometric Policy
Evaluation: A Critique,” Carnegie – Rochester Conference
Series on Public Policy 1: 19 – 46.

Cohen, Benjamin J. (1971), The Future of Sterling as an
International Currency (London: Macmillan).

Mundell, Robert A. (1961), “A Theory of Optimum
Currency Areas,” American Economic Review 51
(September): 657– 65.

Deutsche Bundesbank (1995), The Monetary Policy
of the Bundesbank (Frankfurt am Main: Deutsche
Bundesbank).

——— (1973a), “A Plan for a European Currency,” in The
Economics of Common Currencies, ed. Harry G. Johnson
and Alexander K. Swoboda (London: George Allen and
Unwin), 143 –72.

European Central Bank (1999a), Annual Report 1998
(Frankfurt am Main: European Central Bank).

——— (1973b), “Uncommon Arguments for Common
Currencies,” in The Economics of Common Currencies,
ed. Harry G. Johnson and Alexander K. Swoboda
(London: George Allen and Unwin), 114 – 32.

——— (1999b), Euro Area Monetary Aggregates:
Conceptual Reconciliation Exercise (Frankfurt am Main:
European Central Bank).
European Commission (1998), Report from the
Commission to the Council: On the Harmonisation of
Consumer Price Indices in the European Union
(Brussels: Commission of the European Communities).

——— (1998a), “The Case for the Euro – I,” Wall Street
Journal, March 24, A22.
——— (1998b), “The Case for the Euro – II,” Wall Street
Journal, March 25, A22.

European Parliament (1999), Report on the Annual
Report for 1998 of the European Central Bank, European
Parliament Session Document, October 15, 1999.

Porter, Richard D., and Ruth A. Judson (1996), “The
Location of U.S. Currency: How Much Is Abroad?”
Federal Reserve Bulletin 82 (October): 883 – 903.

Feldstein, Martin (1997a), “EMU and International Conflict,”
Foreign Affairs 76 (November/December): 60–73.

Portes, Richard, and Hélène Rey (1998), “The
Emergence of the Euro as an International Currency,”
Economic Policy 26 (April): 307– 43.

——— (1997b), “The Political Economy of the European
Economic and Monetary Union: Political Sources of an
Economic Liability,” Journal of Economic Perspectives 11
(Fall): 23 – 42.

Prati, Alessandro, and Garry J. Schinasi (1999),
“Financial Stability in European Economic and Monetary
Union,” Princeton Studies in International Finance no. 86
(Princeton, N.J.: Princeton University Printing Services).

Frankel, Jeffrey A., and Andrew K. Rose (1998), “The
Endogeneity of the Optimum Currency Area Criteria,”
Economic Journal 108 (July): 1009 – 25.

Rose, Andrew K. (1999), “One Money, One Market:
Estimating the Effect of Common Currencies on Trade,”
NBER Working Paper Series, no. 7432 (Cambridge,
Mass.: National Bureau of Economic Research,
December).

Group of Ten (1998), The Macroeconomic and Financial
Implications of Ageing Populations (Basle: Bank for
International Settlements).

ECONOMIC AND FINANCIAL REVIEW FIRST QUARTER 2000

Seitz, Franz (1995), “The Circulation of Deutsche Mark
Abroad,” Discussion Papers of the Economic Research
Group of the Deutsche Bundesbank, no. 1/95.

U.S. Bureau of Labor Statistics (1997), BLS Handbook of
Methods (Washington, D.C.: U.S. Government Printing
Office).

United Nations (1998), World Population Prospects: The
1998 Revision, vol. I, Comprehensive Tables (New York:
United Nations).

Wynne, Mark A., and Jahyeong Koo (forthcoming),
“Business Cycles Under Monetary Union: A Comparison
of the EU and U.S.,” Economica.

FEDERAL RESERVE BANK OF DALLAS

The drive for a multilateral trade agreement encompassing the Americas gained
momentum about two years ago, with the U.S.
Congress poised to grant the president fast-track
authority to negotiate Chile’s inclusion in the
North American Free Trade Agreement
(NAFTA). But the series of severe financial crises
that rattled the world almost immediately upon
NAFTA’s inception frustrated the fast-lane
approach and slowed progress toward the
agreement. Perhaps this delay reflected that policymakers, businesspeople, and even the general public were reconsidering the benefits of
trade agreements with crisis-prone partners.
With the prospect of agreement postponed indefinitely, would countries in the area
benefit from lowering, even if unilaterally, their
trade barriers? This is the issue addressed in this
series of two articles begun in the third quarter
1999 Economic and Financial Review.
Part 1 concluded that static applied general
equilibrium models could make a mild case for
unilateral trade liberalization. However, the article raised the possibility that dynamic models,
which incorporate the dimension of time, might
do substantially better. That conjecture was partially inspired by numerical experiments with
models in which the level of capital after the tariff
reduction is changed exogenously (from outside
the model).
For example, a static applied general equilibrium model by KPMG Peat Marwick (1991)
delivers larger welfare gains when the level of
capital in Mexico is increased exogenously to
make the rate of return to capital the same both
before and after NAFTA. The study starts by
assuming that the level of capital is the same
before and after the inception of NAFTA.
Mexico’s gains from NAFTA are negligible in
this exercise, the equivalent variation of 0.32
percent of GNP.1 But the assumption of a constant level of capital implies a higher real rate of
return to capital after NAFTA. Because this is an
unrealistic outcome under free capital mobility,
the study lets capital rise to the level needed to
ensure that the rate of return is the same before
and after NAFTA. Under this assumption, the
welfare gain is equal to 4.6 percent in GNP.2
Two qualifying comments are in order.
First, the capital increase necessary to make the
rate of return the same before and after NAFTA
is about 8 percent, which is substantial and, for
all we know, has not yet materialized, even six
years after NAFTA’s inception. Second, this
expansion in capital is introduced from outside
the model. It is impossible to determine, therefore, whether this new capital level is consistent

Measuring the Benefits of
Unilateral Trade Liberalization
Part 2: Dynamic Models
Carlos E. J. M. Zarazaga

his series of two articles has

examined the potential gains or
losses from unilateral trade
liberalization predicted by
available general equilibrium
models of international trade.

Carlos Zarazaga is a senior economist and executive
director of the Center for Latin American Economics
at the Federal Reserve Bank of Dallas.

ECONOMIC AND FINANCIAL REVIEW FIRST QUARTER 2000

with the optimal consumption (and, thus, saving) decisions of the agents populating the artificial economy. To answer this question, it is
necessary to formulate dynamic models that not
only lay out the microeconomic foundations of
consumers’ and firms’ behavior by specifying
preferences, endowments, and technology but
also incorporate the dimension of time in their
decisions. The solution to consumers’ and firms’
maximization problems will dictate the society’s
desired level of savings and, therefore, of capital, after trade reform. In other words, in
dynamic models the level of capital following a
trade reform is determined endogenously—that
is, within the model—rather than in some ad
hoc fashion from outside of it.
Part 2 of this series investigates the extent
to which applied dynamic general equilibrium
models deliver on their promise of large welfare
gains from unilateral trade liberalization.

The intertemporal dimension of the problem also appears in consumers’ budget constraint, which typically takes the form

∑i pitcit
N

for each period t, where yt is the household’s
endowment in terms of, say, good 1; bt is the
household’s net holding of assets measured also
in terms of good 1 (positive if the household is
a net creditor, negative if it is a net debtor) at
the beginning of period t; and rt –1 is the real
interest rate consumers receive (if net lenders)
or pay (if net borrowers) on their previousperiod asset holdings.
Rearranging the equation as

bt ≤ y t + (1 + rt − 1 )bt − 1 −∑i pitcit
N

makes it apparent that bt is the current period
savings: the excess of revenues from all sources,
the endowment yt, and interest payments from
assets (1 + rt –1)bt –1, minus current consumption
of goods and services, ∑ i pit cit .
The presence of savings in the budget
constraint makes clear that, in contrast to the
static case, a consumer can now borrow or lend
(depending on whether a negative or a positive
bt is chosen) to increase or decrease consumption from the level that current income would
otherwise permit.
In a dynamic setting, the consumer’s problem is no longer to choose the single consumption bundle ci, but rather, the whole sequence
∞
of consumption bundles {ci,t }t =0 that maximizes
utility. Correspondingly, the consumer will
∞
choose the sequence of asset holdings {bi,t }t =0
consistent with the ability to finance that optimal consumption stream.

APPLIED DYNAMIC GENERAL EQUILIBRIUM
MODELS OF INTERNATIONAL TRADE
Applied dynamic general equilibrium
models, unlike static ones, can address investment issues because they introduce the dimension of time in consumers’ and firms’ decisions.
As a result, consumers can postpone consumption today and save to be able to consume more
tomorrow. Recall that in dynamic general equilibrium models, in contrast to static ones, capital accumulation is determined endogenously
rather than exogenously.
Operationally, this difference between the
models is most apparent in the utility function
and budget constraint used to represent consumers’ behavior. Intuition suggests that a simple dynamic version of the static utility function
presented in Part 1 could be
(1)

Welfare =

∑t β ∑i
∞

≤ y t + (1 + rt − 1 )bt − 1 − bt

Technical Challenges of International Trade
Models with Endogenous Capital Accumulation
The addition of the dynamic dimension
could potentially increase welfare because the
removal of tariffs can prompt a decline in the
cost of the imported goods used in the production of domestic capital goods. The corresponding declines in the unit cost of production of capital goods and, therefore, in their prices provide
the necessary incentives for a higher rate of
investment. The resulting expansion of the capital stock increases labor productivity and, hence,
the output of consumer goods and services. But
the introduction of the temporal dimension also
raises technical complications worth exploring to
understand the limitations of the measures of the
benefits from trade liberalization reported below.
One of the challenges facing international
economists is the problem posed by a constant

αi log cit ,

where ci denotes real consumption of good i, αi
is a parameter that measures the relative importance the representative consumer attaches to
each good, t indexes the time of consumption,
and β is the factor by which consumers discount
the future, with 0 < β < 1.
This formulation of the utility function
conveys the idea that consuming a unit of a
good in the future is less attractive than consuming this same unit today. Postponing consumption is costly in terms of utility, and that is
why a bundle of goods consumed today yields
utility ∑ αi log ci, while that same bundle consumed tomorrow yields the utility β ∑ αi log ci.
(Recall that β < 1 by assumption.)

FEDERAL RESERVE BANK OF DALLAS

discount factor, β. The assumption of a constant
discount factor is standard in many intertemporal
models but is problematic in international applications, particularly for small open economies.
This is because models with a constant β usually
generate an explosive (implosive) consumption
path, in the sense that consumption as a fraction
of income constantly increases (declines) over
time. Such paths are highly counterfactual, as
consumption–income ratios tend to be stable in
actual economies.
The reasons for the odd outcome are outlined in the box entitled “Undesirable Implications
of the Constant Discount Factor Assumption.” Here,
it suffices to say the source of the mischief is the
combination of a constant discount factor and the
small open economy assumption. Under this latter
assumption, a small economy is capable of borrowing and lending unlimited amounts at a constant world interest rate. Of course, this assumption is a good approximation to reality only within
certain limits. Eventually, if the economy keeps
borrowing without bounds, it will absorb all
worldwide available savings, at which point the
economy will cease to be small and either the
interest rate will rise or the country will be unable
to continue borrowing.
International economists dealing with
dynamic models—that is, models of endogenous capital accumulation—have tried to solve
the problems created by the small open economy assumption in several ways. One popular
route has been to abandon the assumption, in
Equation 1, that the discount factor, β, is constant
over time and assume instead that it is a function of consumption.3 Mathematically, such an
assumption is represented as

welfare function. In particular, they propose
evaluating the welfare gains from free trade
according to the formula
(2)

∑t

∞

βt

[

σ
csαbs1− α
σ −1

]

σ −1
σ

where α and σ are standard parameters in the literature, assumed to have values that ensure the
concavity of the utility function, and bs is a composite of foreign and domestic assets in real
terms.5 This solves the problem of the lack of stationarity when β is constant because assets are
treated as just another good and subject to decreasing marginal utility. This will generally ensure a
stationary wealth–income ratio and generate a
stationary consumption–income ratio as well.6
The idea of including financial assets in
the welfare function is not new. In fact, many
models studying monetary policy issues assume
that money, an asset, is a determinant of the
utility function. This practice has met with
objections because what utility functions such
as that in Equation 2 say, if interpreted literally,
is that consumers derive pleasure from the mere
fact of holding money or bond issues. This is a
highly unattractive proposition, as consumers
clearly do not derive utility from the pieces of
paper but from what they can buy.7
Thus, international economists face the
difficulty that the assumption of a constant discount factor standard in closed macroeconomic
models is unappealing when applied to small
open economies, because it tends to produce
the counterfactual outcome that consumption as
a fraction of income permanently declines or
increases. This prediction has typically been
eliminated at the cost of counterintuitive preferences, a factor that must be taken into account
in evaluating the quantitative results of the
applied dynamic general equilibrium models
reported in the next section.

β = β(ct ),
which says the discount factor, β, at any point
in time is a function of consumption.
As the box explains, this alternative
assumption may give rise to stationary outcomes—that is, to equilibria with constant consumption–output ratios. Applied dynamic general equilibrium models typically assume the
function β(ct ) decreases in ct . In other words, as
consumption increases, β decreases. This is not
an entirely satisfactory specification because
there is little evidence that people discount the
future more as they become richer.4
Goulder and Eichengreen (1992) offer a
different solution to the problem posed by the
combination of the small open economy
assumption and a constant β. Instead of postulating a variable β, they introduce financial
assets (wealth) as a determinant of the utility or

ECONOMIC AND FINANCIAL REVIEW FIRST QUARTER 2000

Welfare =

Welfare Gains from Trade Liberalization
in Applied Dynamic General Equilibrium Models
Progress in quantifying the benefits of free
trade with dynamic models has been slow
because of the theoretical difficulties discussed
above and other computational issues. The few
such models available have a mixed record.
Goulder and Eichengreen (1992) find that
a U.S. move to unilateral free trade by removing
tariffs from an average rate of 4 percent would
cut consumption 0.32 percent, which in turn
implies a welfare loss equal to 0.44 percent of
GDP.8 The larger welfare gains dynamic models
anticipate do not materialize, therefore, in
Goulder and Eichengreen’s study.

Undesirable Implications of the Constant Discount Factor Assumption
or, equivalently, that β (1 + r ) = 1. This condition
will typically be satisfied only by chance. More
generally, either β (1 + r ) < 1 or β (1 + r ) > 1. In
the first case, Equation 1 implies that consumption
in each period will be lower than in the previous
one by the factor β (1 + r ). In other words, the
consumption–income ratio will decrease monotonically to 0. In the second case, the opposite is true:
the consumption–income ratio increases monotonically over time. The problem is that the implication
of the dynamic model in either of these cases is
grossly counterfactual, because observed consumption–income ratios are very stable over time.
As the text mentions, one possible way out of
this problem is to abandon the assumption of a
constant discount factor and assume, instead, that
it is a decreasing function of consumption. This can
be seen intuitively by replacing the function β with
β(ct ) in Equation B.1:

A constant discount factor in combination with
time-separable additive utility functions like the one
represented in Equation 1 of the text can lead to
counterfactual implications for the consumption
path of the model economy.
As explained, in a dynamic setting a consumer will typically maximize the utility function
∑∞t =0 βt log ct
subject to the intertemporal budget constraint
ct = y + (1 + rt –1) bt –1 – bt ,
where all symbols are as in the text, the endowment y is a constant, the real interest sequence
{rt –1}∞t =0 is exogenous, and there is only one good (i
= 1), with its units redefined so its price is 1. The
solution to this problem will be the selection of a
consumption sequence {ci,t }∞t =0 and an asset-holding sequence {bt }∞t =0 consistent with the ability to
finance that optimal consumption stream.
Substituting into the utility function the expression for ct given by the budget constraint yields the
maximization problem

ct +1
c
= β(ct ) ∗ (1+ r ) ∗ t .
y
y
Suppose that the function β(ct ) is decreasing
in ct and that for certain value c of ct , β(ct ) (1 + r ) >
1. This implies ct +1 > ct = c, which in turn implies
that β(ct +1) < β(ct ) and, therefore, that β(ct +1) (1 + r )
< β(ct ) (1 + r ). In other words, as consumption
increases, β(ct +1) decreases, and so does
β(ct +1)(1 + r ) until eventually β(ct +n )(1 + r ) = 1
for n large enough. At that point consumption
becomes stationary (in the sense that it repeats
itself over time) because

Max {βt ln[y + (1 + rt –1) bt –1 – bt ]
+ βt +1 ln[y + (1 + rt ) bt – bt +1]
+ ∑∞j =0 βt +j + 2 ln[y + (1 + rt +j + 1) bt +j + 1 – bt +j + 2]},
where the choice variable is bt in period t, bt +1 in
period t +1, bt +2 in period t +2, and so on, and j and
t are time indexes.
The first-order necessary condition with
respect to bt corresponding to this maximization
problem is
−

ct +n +1 ct +n
=
.
y
y

βt
y + (1+ rt −1)bt −1 − bt

However, this way of solving the lack of stationary equilibria in dynamic models with constant
discount factor β is somewhat of a mechanical
quick fix. Typically, any function β(ct ) will be continuous and, therefore, decreasing for some values of
ct (or eventually all of them, as in the example
above). At the same time, in any reasonable economic model, consumers want to consume more
the wealthier they are. This implies that when β(ct )
is decreasing in consumption, it is also decreasing
in wealth or, equivalently, that households become
more impatient to consume as they get wealthier.
Unfortunately, there is no empirical evidence to
support this rather ad hoc assumption. The opposite and equally arbitrary assumption that β(ct ) is
increasing in ct —that is, that a household’s desire
to accumulate wealth rises as it becomes richer—
cannot be empirically validated either (and introduces the additional technical difficulties mentioned
in footnote 4 of the text).

βt +1
+
∗ (1+ rt ) = 0.
y + (1+ rt )bt − bt +1
Dividing both sides by βt and using the budget
constraint again, the following equivalent expression results:
−

1
β
+
∗ (1+ rt ) = 0,
ct ct +1

which, assuming that rt = r, a constant, takes the
form
ct +1 = β * (1 + r ) * ct .
Assuming, for convenience only, that income is
constant over time, the above condition can be represented in terms of consumption–income ratios as
(B.1)

ct +1
c
= β ∗ (1 + r ) ∗ t .
y
y

In a stationary equilibrium, prices and real
consumption–income ratios are constant over time.
This implies that the above condition in any stationary equilibrium will take the form
c
c
= β ∗ (1 + r ) ∗
y
y

FEDERAL RESERVE BANK OF DALLAS

tion is introduced mainly to account for the puzzling “cross-hauling” in which many countries
appear to export the same goods they import.11
National product differentiation circumvents this
problem by assuming each country is the only
producer of the good it exports. However, this
also means tariffs could help a country exploit
its market power. Tariff elimination might be
damaging in this case because the optimal tariff
typically is not zero under this assumption. This
fundamental bias against free trade is absent
from Ahearne’s model but seems to prevail in
Goulder and Eichengreen’s.12
There are reasons to doubt the welfare
losses Goulder and Eichengreen’s model delivers because their preferences include assets as a
determinant of the welfare function. The resulting welfare measure may reflect consumption
changes as much as changes in asset holdings.
This is certainly an unappealing way to measure
welfare, in light of the general equilibrium theory standard that consumers do not derive utility directly from merely holding assets but from
the stream of goods and services those assets
can purchase.
Ahearne’s study may exaggerate the GDP
growth from unilateral trade liberalization
because he assumes perfect capital mobility.
This may not be the case in practice, as evidence suggests that households tend to invest
their savings in their home country rather than
in foreign ones. Goulder and Eichengreen capture more aptly this reality by assuming that
consumers have a bias for domestic assets, and
this implicitly limits the capital mobility responsible for the relatively large GDP and consumption gains in Ahearne’s model.13

Those gains do seem to materialize—at
least for developing countries—in a recent
model by Ahearne (1999). It is one of the few
dynamic models to quantitatively analyze unilateral trade liberalization’s effects for developing countries. Unfortunately, Ahearne focuses
on the performance of macroeconomic variables such as aggregate output, consumption,
and investment and does not report a measure
of welfare, such as that Goulder and
Eichengreen report. This omission makes welfare comparison of the two studies difficult. In
any case, to the extent the direction of change
of consumption and welfare are the same (as
they are in Goulder and Eichengreen), Ahearne’s
outcomes are more favorable to trade reform.
He finds that lowering tariffs from an average of
25 percent to 10 percent would result in an
increase in consumption of about 3 percent. A
reduction to an average rate of 5 percent would
raise consumption growth to about 4.5 percent,
while the complete removal of tariffs would
result in a 6 percent consumption increase.
Six percent consumption growth is by no
means negligible and could be seen as an indication that dynamic models can, after all, deliver
larger welfare gains from unilateral trade liberalization than their static counterparts. However,
it is important to emphasize that the relatively
large consumption growth of 6 percent is obtained from removing tariffs originally assumed
to be 25 percent. Ahearne’s study suggests that
consumption growth will be a more moderate
1 percent to 2 percent if the average initial
tariff is 4 percent, as in Goulder-Eichengreen.9
Still, this increase in consumption after trade
liberalization seems to reverse Goulder and
Eichengreen’s negative finding.
Unfortunately, it is difficult to pinpoint the
source of the opposite results of these two
dynamic models because their features are quite
different, from the specification of the utility
function to the underlying assumptions about
capital mobility.10 For example, Ahearne’s
assumption is that the discount factor depends
on the level of wealth (or equivalently, consumption), while Goulder and Eichengreen
assume that assets enter into the utility function.
Another important difference is that
Ahearne assumes the terms of trade are exogenous. Thus, changes in tariffs alter the relative
domestic prices but do not change the international terms of trade against the country that
liberalizes. This is not the case in Goulder and
Eichengreen because they adopt the so-called
Armington, or national product differentiation,
assumption. As Part 1 explained, this assump-

ECONOMIC AND FINANCIAL REVIEW FIRST QUARTER 2000

PRODUCT VARIETY AND GAINS
FROM TRADE LIBERALIZATION
It was argued that dynamic models of
international trade have the potential to deliver
the large welfare gains from trade liberalization
that their static counterparts have failed to produce. The preceding section suggests that
dynamic models cannot fulfill those expectations either, except under nonconventional representation of consumers’ preferences.
However, one often-heard criticism of all
the models discussed so far is that they fail to
incorporate the idea that free trade makes possible access to new technologies that enhance
the economy’s overall productivity. Perhaps this
is why dynamic general equilibrium models
produce only less-than-striking welfare gains
from unilateral trade liberalization.

Consider, for example, the constantreturns-to-scale production function presented
in Part 1 of this series:
(3 )

restrictions limiting quantities make trade
between two economies disappear. This implies
that consumers must make do with domestic
goods. Although a consumer would like that 27inch-screen TV and can afford it, he will have to
settle for the smaller domestic model without
remote control. Likewise, local producers will
have to adjust their technologies to the intermediate inputs domestic firms make available. A
construction company may prefer a special kind
of foreign-made insulation for a building that
will have to withstand extreme temperatures,
but the firm will have to use a more expensive
and less functional building design to achieve
the same results with the less suitable insulators
produced domestically.
Next, suppose all barriers to international
trade are lifted. Firms in this economy will be
able to use both domestic and foreign inputs.
The examples above suggest that a larger variety of goods, especially of intermediate inputs,
may be associated with aggregate productivity
gains not appropriately captured by conventional production functions.
To confront this limitation, economists
have started to play with less conventional production functions that incorporate the idea of
productivity gains from variety. Such production
functions can be constructed by a clever reinterpretation of the conventional constantreturns-to-scale production functions.
For simplicity, assume only one final consumption good is produced with the technology

Y = A L 1− αK α ,

where Y is aggregate output, L the amount of
labor input, K the amount of capital input, and
0 < α < 1.
In this specification, the total factor productivity, represented by A, is treated as a given
parameter, invariant to the trade regime. Therefore, this equation does not capture the idea
that trade liberalization will increase an economy’s overall productivity. Trade liberalization
can raise production only if it leads to the use
of more labor or capital inputs.14 Otherwise, the
same amount of labor and capital will produce
exactly the same amount of output.
A similar situation arises with the second
kind of technology, the increasing-returnsto-scale technology—or, equivalently, decreasingaverage-cost technology—considered in Part 1:
(4)

Total cost = F + bQ.

Again, notice that tariff policy changes can
reduce average costs only if they induce an
increase in the quantity of the good. But the
basic cost structure, defined by fixed cost F or
marginal cost b, is the same regardless of the
tariff regime under which countries operate.
This invariance of the overall productivity
to the trade regime implicit in conventional production functions has been challenged on several grounds. For instance, an important benefit
of international trade is that it gives consumers
more choices and offers producers more
options in terms of inputs. The advantage of
variety is particularly important for economies
that can produce only a limited range of goods
on their own. This is the case with economies
characterized by cost functions such as the one
in Equation 4—that is, economies with increasing returns to scale.
Recall from the discussion in Part 1 that in
such economies each good is produced by one
and only one firm. The number of varieties is
determined by the number of firms, which is limited when there are increasing returns to scale.
To see this, suppose all firms must pay the fixed
cost F in terms of a primary input z (for example,
land) and that each economy is endowed with Z
units of that good. Each economy on its own will
be able to produce, at most, Z/F varieties of
goods (for simplicity, we assume Z/F is an integer). The number of goods produced domestically will be limited by that upper bound.15
Thus, suppose prohibitively high tariffs or

(

)

Y = L 1− α x 1α + x 2α + … + x Mα ,
where L represents the amount of labor input, xi
represents the quantity of an intermediate input
i, i = 1, 2, …, M, and 0 < α < 1. Assuming each
intermediate input is used in the same quantity,16 the technology can be rewritten as
(5)

Y = M L 1− αx α.

This appears to be the same old constantreturns-to-scale technology of Equation 3, with
capital, K, relabeled x and the total factor productivity, A, relabeled M. Indeed, increasing
both the amount of labor input L and the typical
intermediate input x by h percent would raise
production of the consumption good by h percent, which is exactly what is supposed to occur
with a constant-returns-to-scale technology.
The trick is that relabeling A as M is not as
innocuous as it might appear because now A is
not necessarily fixed. In fact, A (or M )—the
number of varieties—can be regarded as an
input, just as L or each xi is. In other words,
according to this production function, aggregate

FEDERAL RESERVE BANK OF DALLAS

production of final goods Y requires combining
three inputs: the number of varieties of intermediate inputs (M ), the quantity used of each
of them (x ), and labor (L ).
The reinterpretation of A as the number of
intermediate-input varieties represents mathematically the old idea that one-size-fits-all
economies will be less productive than highly
specialized ones. The intuition is that access to
a larger variety of goods will make it more likely
that producers will find inputs that better fit the
characteristics of their production lines and that
consumers will find the products that best fit
their tastes and needs.17
The gains-from-variety effect can be better
understood by comparing the nonconventional
production function in Equation 5 with

With a production function like that in
Equation 5, the gains from freer trade will come
from two sources: the traditional one that tariff
reductions will make imported intermediate
goods cheaper and thus induce higher output
levels of the existing varieties of final goods and
services, and the nonconventional one of gains
from variety from ∆M. This second effect is a
good candidate for boosting the welfare gains
from unilateral trade liberalization beyond the
negligible to moderate results found by models
using more conventional production functions.
The remainder of this article reports the results
of recent work that has exploited this gainsfrom-variety approach to build a better case for
unilateral free trade.

Y = L 1− α ∗ M ∗ x .

Measuring the Welfare Gains from Variety
In analyzing these studies on welfare gains
from product variety, it is important to understand how tariffs reduce the product varieties
available to firms and consumers. All the studies discussed below assume that firms face cost
functions of the form in Equation 4. Equivalently, they assume all goods are produced with
increasing-returns-to-scale technologies.19 As
Part 1 explains, this is the only technology consistent with the national product differentiation
assumption, typically introduced to account for
the cross-hauling puzzle in trade statistics.
The introduction of increasing-returnsto-scale technologies is not inconsequential for
the potential gains from variety with a tariff
reduction. The attrition effect tariffs can have on
variety starts at much lower tariffs with increasing-returns-to-scale technologies than with constant-returns-to-scale technologies.
Consider a typical final-goods producer’s
demand for an imported intermediate input produced with a constant-returns-to-scale technology, as represented in Figure 1. Recall that the
cost function will look like Equation 4, with F
equal to 0, implying a constant marginal and
unit average cost of b. Suppose that initially
there is no tariff, and the equilibrium demand of
the input is at point E0, with price b equal to the
unit and marginal cost and quantity Q x 1. Next,
suppose an ad valorem tariff of τ percent is
imposed on this intermediate input. For the
sake of argument, assume the buyers absorb all
the burden of the tariff —that is, the foreign producers of those inputs still receive a price b
(equal to their unit cost of production) for each
unit of the intermediate good they sell to
domestic buyers. These buyers will have to pay
a price b (1 + τ/100) for the imported intermediate input. Suppose that at the new equilibrium

According to this production function, doubling
the varieties of intermediate-input goods will
have the same effect on output as doubling the
amount of each of those inputs, as can be seen
from the equalities
L 1− α ∗ (2M ) ∗ x = 2L 1− α ∗ M ∗ x = 2Y
= L 1− α ∗ M ∗ (2x ).
This is not the case with the proposed production function of Equation 5, in which doubling the number of varieties M doubles output,
but doubling the quantity of each intermediate
input x increases output only by a factor of 2α,
which is lower than 2 (recall that 0 < α < 1).18
In other words, in Equation 5 any increase
in variety has a larger impact on aggregate production than an identical percentage increase in
the quantities of the existing intermediate-input
varieties. Loosely speaking, this production function captures the idea that a society cannot easily compensate for the loss of variety with more
of the same old stuff. This issue is relevant to
measuring the gains from unilateral tariff removal
because freer trade policies (even if implemented
unilaterally) may give a country access to a larger
variety of goods. The welfare gains from such
policies may be important if a larger variety of
intermediate inputs, as the production function
suggests, increases the economy’s productivity in
the manufacturing of domestic goods.
In the reinterpretation (Equation 5 ) of the
conventional constant-returns-to-scale technology (Equation 3 ), the total factor productivity
parameter A in the latter would be equal to M
before trade liberalization but eventually equal
to (M + ∆M ) after trade liberalization, where ∆M
represents the additional varieties of intermediate inputs resulting from freer trade.

ECONOMIC AND FINANCIAL REVIEW FIRST QUARTER 2000

(1 + τ/100)(b + F/Q x 3 ) and the quantity demanded drops to Q x 4. It appears, then, that the
pair (b + F/Q x 3, Q x 4), represented at point E 3
of the demand curve in Figure 1, is a good candidate for the new equilibrium. But the appearance of two different quantities, Q x 3 and Q x 4, in
this pair suggests something is wrong with that
conjecture. Indeed, unlike in the constant-returnsto-scale case, producers of the intermediate good
experience an increase in the average unit
production cost by cutting production of the
intermediate input from Q x 3 to Q x 4. In fact, the
unit costs will be b + F/Q x 4, higher than the
b + F/Q x 3 per unit they will receive from the
price inclusive of tariff (1+ τ/100)(b + F/Q x 3 ).
In other words, at the price (1+ τ/100)(b +
F/Q x 3 ) producers of the imported intermediate
input will suffer a loss. Therefore, they will have
to increase the price (before tariff), say, to b +
F/Q x 4 . But this higher producer price will result
in a higher user price of (1+ τ/100)(b + F/Q x 4 ),
which in turn will reduce demand for the intermediate input even further. This will result in
a higher unit cost to produce the imported intermediate input and lead to another round
of increases in the domestic price of those
imports. Eventually, unit costs will keep rising at
a higher rate than the price. This shows up in
Figure 1 in the fact that for prices above b +
F/Q x 3 , the demand curve stays always to the left
of and below the unit average cost curve.21 This
implies producers will always suffer a loss if
they cut production below the pretariff level
Q x 3 . Since the tariff reduces demand below that
level of production, the producer of the intermediate input will be forced out of business
and that input variety will disappear from the
market.
Thus, in contrast to the constant-returnsto-scale case, the imposition of even a moderate
tariff in the presence of increasing returns to
scale may reduce M, the number of intermediate-input varieties available to final-goods
producers. This is because the tariff has a “market-size” effect on the intermediate input’s unit
production cost that was absent in the constantreturns-to-scale technology case.
The next section discusses how these general ideas have been implemented in recent
studies that attempt to take into account productivity gains from variety eventually introduced by
policies of unilateral trade liberalization.22

Figure 1

The Effect of Tariffs on Product Variety
Price of
imported
intermediate
input
Average cost function
under increasing returns
to scale

b+F
QX4
b + F (1 + τ )
QX3
100 E3

Demand

Average cost
function under
constant returns
to scale

b(1 + τ )
100
E2

b+F
QX3
b

E1
QX4 QX2

E0
QX3

QX1

Quantity of
imported
intermediate input

price the quantity demanded, Q x2, is one-fourth
Q x1, the equilibrium quantity before the tariff was
introduced. Foreign producers will have to cut
the quantity produced by three-fourths. The question is whether they can stay in business doing
that. The answer under constant returns to scale
is an unambiguous yes because producers will
always receive the price b for each unit, regardless of sales level. Since b is also the unit cost of
production, they will cover their costs and be able
to stay in business whether they sell Q x1 units (as
in point E 0 ) or Q x2 (as in point E 1). Thus, under
constant returns to scale, the available varieties of
intermediate inputs will be the same before and
after the introduction of the tariff. The only effect
of a tariff is that each intermediate input will now
be produced in a smaller amount, to match the
decline in the quantity demanded as a result of
the tariff-induced price increase. Thus, moderate
tariffs will tend not to have any visible consequences for product variety under constantreturns-to-scale technologies.20
Consider the alternative case in which intermediate inputs are produced with an increasingreturns technology, whose cost function will look
like Equation 4. The unit cost of producing Q x1 is
greater than b by F/Q x1. Therefore, b cannot be an
equilibrium price because foreign producers of
the imported inputs would suffer a loss. Assume,
then, that the equilibrium price for the imported
intermediate inputs under increasing-returns-toscale technology is b + F/Q x 3. The higher price
implies, of course, that the equilibrium quantity,
Q x3, will be to the left of the equilibrium quantity
under constant returns to scale, Q x1.
Now suppose the same tariff of τ percent
is levied on all imported intermediate inputs
and the tariff is borne entirely by the domestic
buyers of imported intermediate goods. As a
result, the price increases from b + F/Q x 3 to

Welfare Gains from Variety
in Static Models of International Trade
I report first a recent study by Klenow and
Rodriguez-Clare (1997) because, strictly speak-

FEDERAL RESERVE BANK OF DALLAS

ing, their model is static and thus belongs to the
class discussed in Part 1 of this series. However,
I deferred discussion of this work until now
because it is one of the few available studies
that explicitly considers the potential gains from
variety when measuring the benefits of unilateral trade liberalization.23
The Klenow and Rodriguez-Clare model
incorporates the product variety effect by assuming that importing firms must pay a fixed cost F
to operate and a constant price b for each unit
of imported good. For all practical purposes, it is
as if the imported goods are produced according
to a cost function like that in Equation 4.
Tariffs will make imported goods more
expensive and, hence, reduce the demand for
them. By the mechanism explained earlier, this
smaller market size will eventually leave some
importing firms unable to cover their fixed
costs, forcing them to shut down. Consumers
then suffer because they can no longer find the
varieties of goods they had been purchasing.
Likewise, local producers will become less productive, as Klenow and Rodriguez-Clare assume
a production function of the type in Equation 5
(although a much more complicated one), by
which a lower M (number of intermediate-input
varieties) results in productivity losses and, therefore, in lower output despite the same capital
and labor inputs.
The authors quantify the model using data
for Costa Rica and find that removal of a 10 percent tariff can quadruple the gains from unilateral trade liberalization compared with a model
in which product variety effects are absent. In
particular, they find that imposition of a 10 percent tariff on intermediate goods leads to welfare losses equal to about 2 percent of GDP, as
opposed to only 0.5 percent in lost GDP when
the variety effect is not taken into account.
Thus, incorporation of gains from variety
works in the expected direction of increasing
the welfare gains from trade liberalization but
keeps them within the moderate ranges of the
static models without the gains-from-variety
effect, as reported in Part 1.
One possible reason for the moderate
gains in the Klenow and Rodriguez-Clare model
is that the national product differentiation
assumption works against unilateral trade liberalization, as explained above. The model’s static nature also could be a factor. Thus, the next
step is to see if these limitations are overcome
by dynamic models—those that incorporate the
dimension of time and, hence, saving and
investment decisions—in the context of gainsfrom-variety effects.

ECONOMIC AND FINANCIAL REVIEW FIRST QUARTER 2000

Welfare Gains from Variety
in Dynamic Models of International Trade
Quantitative dynamic models measuring
the effects of unilateral trade liberalization do
not abound. Even fewer of them have tackled
the gains-from-variety effect. One such model is
a study on Austria by Keuschnigg and Kohler
(1996).
As explained earlier, all models of international trade that consider tariffs’ effects on product variety must introduce, in one way or
another, fixed costs of production. In the case of
Keuschnigg and Kohler, it is the local producers
of final domestic goods (and not importing
firms, as in Klenow and Rodriguez-Clare) that
face a fixed production cost. This assumption is
the same as in the Klenow and Rodriguez-Clare
model, except that tariffs will not change the
number of foreign varieties but of domestic
intermediate-input varieties supplied to local
producers.
Because of this fixed cost, and for the
same reasons as in the static models, tariff
reductions in Keuschnigg and Kohler increase
the market size for every good, eventually making it profitable to import or produce varieties
unavailable before. In addition to this static
effect, Keuschnigg and Kohler introduce a
dynamic one by linking the stock of capital with
the number of product varieties.
The intuition behind this additional effect
is similar to the one given above when describing why, in the presence of a fixed cost, a fixed
factor like land may limit the number of product
varieties an economy can produce. The same
logic works here, replacing land with capital.
Suppose each firm in the economy must pay the
fixed cost Keuschnigg and Kohler assume in the
form of k units of capital. A given capital stock
K could support at most K/k product lines or
varieties. Since K is implicitly assumed fixed in
static models, any increase in varieties must
come through reduction in production costs
rather than expansions in the capital stock.
But since Keuschnigg and Kohler allow
for investment, the capital stock is not fixed. In
fact, reductions of tariffs on intermediate and
capital goods can induce a process of capital
accumulation for the reasons discussed above.24
If the capital stock increases by ∆K as a consequence of a unilateral move to trade liberalization, the economy can eventually support the
higher number of product varieties (K + ∆K )/k .
This capital accumulation effect induces gains
from variety in Keuschnigg and Kohler beyond
those induced by the market-size effect present
in static models described earlier.

Dynamic models that incorporate gainsfrom-variety effects seem to have more potential
for delivering nonnegligible welfare gains. At
the same time, these models include significant
increasing-returns-to-scale technologies, a somewhat problematic feature because it opens the
door to government intervention and may
undermine the case for free market policies that
the gains-from-variety effect is meant to boost.
A clear message from the quantitative
experiments these two articles report is that neither the introduction of time nor of product
variety effects can completely overpower the
strong force against unilateral removal of tariffs
introduced in almost all models by the national
product differentiation assumption.
The strength of such a force is suspect,
especially in models that assume monopolistic
competition. As explained in Part 1, that
assumption puts the market power at the firm—
rather than at the country—level, which, in
principle, should weaken the case for trade barriers introduced by the national product differentiation assumption. Perhaps more weight
should be given to the models discussed in this
article that mitigate the country market power
effect of national product differentiation. These
models deliver moderate to sizable welfare
gains from unilateral trade liberalization.
Thus, the advantages of unilateral trade
liberalization are cause for optimism. But to the
extent that country market power is perceived
as important in evaluating alternative trade policies, countries may balk at the prospect of a
unilateral free trade policy. Hence, the suspension or slow progress of negotiations for multilateral trade agreements is cause for concern.
The measures of welfare gains (or costs)
from unilateral trade liberalization policies presented in this article should be considered with
caution for several reasons. First, dynamic models generally pose formidable technical challenges that—in the particular applications discussed in this article—have been circumvented
with not completely appealing shortcuts.
Second, the theoretical foundations of the
national product differentiation assumption and
the bias it introduces against unilateral trade liberalization remain controversial, especially in
models that assume a monopolistically competitive market structure. Furthermore, the empirical measures of the market power implicit in
such an assumption are imprecise. Certainly,
international trade researchers still have a lot of
work to do before the benefits of unilateral
trade liberalization policies can be confidently
assessed.

Keuschnigg and Kohler calibrate their
model to Austrian data and find that removing a
10 percent average tariff would result in welfare
gains equal to about a 4 percent increase
in GNP if the scale economies are fairly large (or
if fixed costs are fairly big).25 This is more than
two times the gains in the Klenow and
Rodriguez-Clare static model.
The larger welfare gains in Keuschnigg
and Kohler again demonstrate that omitting the
time dimension and capital accumulation may
lead to a fairly sizable underestimation of the
benefits of unilateral trade liberalization.
However, like many models reported
above, Keuschnigg and Kohler’s does not produce striking welfare gains. The potential for
positive terms-of-trade effects from the national
product differentiation assumption may be
responsible for this. In fact, Keuschnigg and
Kohler report that with a milder terms-of-trade
effect, the gains from removing a 10 percent
average tariff could be as large as 7 percent of
GNP.
It is important to remember that the gainsfrom-variety effect may be a dangerous concept
to play with. However beneficial to the case for
unilateral trade liberalization, it may paradoxically undermine the very case for free market
policies that it is meant to buttress. This literature typically appeals to increasing returns to
scale, and, in the presence of such technology,
markets cannot achieve the social optimum
without government intervention.26
CONCLUSION
Free trade advocates consider the denial
of fast-track authority to the U.S. president a
worrisome development. The concern is that
lack of interest in multilateral trade agreements
will create a backlash against the free trade
policies Latin American countries adopted in
the 1990s. The fear is warranted if each country
in the region perceives it will experience
welfare losses from adopting free trade
policies when some if its major trading partners
do not.
This series of two articles has examined
the potential gains or losses from unilateral
trade liberalization predicted by general equilibrium models of international trade. Negligible
to moderate gains are found in static as well
as dynamic models that do not incorporate
gains from product variety. The results confirm
that the omission of the dimension of time
and, hence, of capital accumulation can undermine the case for unilateral trade liberalization.

FEDERAL RESERVE BANK OF DALLAS

NOTES

The author wishes to thank David Gould, Gregory
Huffman, Evan Koenig, and, especially, Steve Brown
for comments on earlier drafts that substantially

improved the contents and organization of the ideas
in both articles of this series. Any remaining errors are,
1

of course, mine.
As Part 1 explains, the equivalent variation in income
is the change in income that consumers should experience without a trade liberalization to replicate the level
of utility they would attain with it. A negative equivalent
variation in income implies that consumers are worse
off after trade liberalization.
Although this model is frequently put in the applied
general equilibrium category, it does not strictly belong
there because it implicitly assumes an excess supply
of labor (that is, the labor market is in disequilibrium)
in Mexico, at least before NAFTA.
One of the first authors to implement this approach in the
context of a small open economy was Mendoza (1991).
The alternative assumption that β is an increasing
function of ct is not less problematic, as it creates the
possibility of multiple equilibria, an issue beyond the
scope of this article.
As explained in Part 1, concavity is a mathematical
property that captures the idea that consumers have a
taste for variety.
The counterpart for the lack of stationarity of the consumption–income ratio with a constant β is the lack of
stationarity of the wealth–income ratio.
The introduction of money in the utility function has
been justified as a shortcut to capture the notion that
money facilitates trade. In fact, Feenstra (1986) shows
that under certain conditions, transaction costs in
trade will operate as if money were an argument of the
utility function. However, the same reasoning applies if
transaction costs are associated with buying and selling bonds or equities. It is on these grounds that
Poterba and Rotemberg (1987), for example, include
short-term government debt (and not just fiat money)
in the utility function. This shortcut to modeling transaction costs explicitly may be useful for addressing
certain monetary policy questions, but its application
to the issue of the welfare consequences of alternative
trade policies is more controversial.
For the reasons given in Part 1, consumption growth
may not be a good measure of well-being, particularly
in models in which labor supply is endogenous. For
example, households may consume more after trade
liberalization but also work harder, so the welfare gains
may not be nearly as large as the increase in consumption would otherwise suggest. That’s why most
applied studies of international trade, like Goulder and
Eichengreen’s, report the equivalent variation in
income rather than the actual variation in income (or
GDP).

ECONOMIC AND FINANCIAL REVIEW FIRST QUARTER 2000

I arrived at this figure by assuming that the change in
consumption from a removal of tariffs will be a linear
function of the original average tariff rate for tariffs in
the 0 percent to 10 percent range.
Both models assume several sectors, but the details of
the disaggregation and technologies in each of them
differ. Both models introduce frictions in the investment
process but differ in the details. Goulder and
Eichengreen assume that changing the capital stock
from its current level is costly in terms of resources,
while Ahearne assumes it is costly in terms of time—
that is, that it takes several periods to bring the capital
stock to the desired level.
As Part 1 discusses, Japan may appear to import and
export cars simply because of the way trade
statistics are reported. For example, Japan could be
importing convertibles and exporting vans. Although
these are different products, they might appear simply
as “cars” in the broad categories used in trade statistics, giving rise to an apparent cross-hauling puzzle.
Unfortunately, none of these authors report an optimal
tariff for their models. One conjecture worth exploring
is that the 4 percent initial tariff Goulder and
Eichengreen assume in their benchmark case is much
closer to the optimal tariff than the 25 percent rate
Ahearne assumes. Obviously, removing an optimal tariff will cause welfare losses while removing a nonoptimal one might enhance welfare.
In fairness, Ahearne himself reports that GDP gains in
his model are more moderate under limited international capital mobility.
As explained in the previous section, trade liberalization can result in a higher capital stock, which implies
a higher use of capital input in the production process.
It is important to note that models of monopolistic
competition (as opposed to perfect competition) have
established rigorously how the number of goods and
the amount produced of each will be determined in a
decentralized economy, using two conditions: that
each firm will maximize profits by producing the output
level at which the marginal revenue equals the marginal
cost, and that free entry ensures that in equilibrium no
firm will capture monopolistic rents. For a didactic presentation of this material, see Krugman and Obstfeld
(1991), chapter 6.
This assumption is only for expositional convenience.
Strictly speaking, it is a result, not an assumption, that
can be obtained as the equilibrium outcome of a
monopolistic competition model in which final goods
are produced from inputs that enter symmetrically (that
is, have the same elasticity of substitution) in a constant elasticity substitution production function. For a
more formal discussion, see Romer (1987).
Ethier (1982) was among the first to propose production
functions of this type. For a nontechnical but persuasive presentation of the benefits of variety in production and consumption, see Cox and Alm (1999).

Thus, if α = 1/2, output will increase by 21/2 = √ 2 ≅
1.41.
Recall that under increasing-returns-to-scale technologies, in contrast to constant-returns-to-scale technologies, the average unit cost declines with the quantity produced.
The tariff will affect M only if it gets so high that the
price b (1+ τ/100) is at or above the demand curve’s
intersection with the vertical axes, where the equilibrium quantity demanded will be zero. As anticipated,
the tariff need not be that high to affect product variety
in the presence of increasing-returns-to-scale technologies.
For a situation like this to emerge, the demand curve
must intersect the vertical axis. Not all utility functions
will induce that property. For example, the demand
functions induced by the logarithmic utility functions in
Part 1 never intersect the vertical axis. Of course, the
studies in Part 2 specify utility functions that do induce
that property on the demand for the relevant goods.

Cox, W. Michael, and Richard Alm (1999), “The Right
Stuff: America’s Move to Mass Customization,” Federal
Reserve Bank of Dallas 1998 Annual Report, 3 – 26.
Dixit, A., and J. Stiglitz (1977), “Monopolistic Competition
and Optimum Product Diversity,” American Economic
Review 67 (June): 297– 308.
Ethier, Wilfred J. (1982), “National and International
Returns to Scale in the Modern Theory of International
Trade,” American Economic Review 72 (June): 389 – 405.
Feenstra, Robert C. (1986), “Functional Equivalence
Between Liquidity Costs and the Utility of Money,”
Journal of Monetary Economics 17 (March): 271– 91.
Goulder, Lawrence H., and Barry Eichengreen (1992),
“Trade Liberalization in General Equilibrium: Intertemporal
and Interindustry Effects,” Canadian Journal of
Economics 25 (May): 253 – 80.

Recall that the parameter M, the number of varieties in
the nonconventional production function (Equation 5),
can be interpreted as a measure of the overall efficiency of technology because it plays the same role
as A, the total factor productivity parameter, in the
more standard production function (Equation 3).
The Klenow and Rodriguez-Clare model contains
many interesting details that cannot be discussed here
without sacrificing the focus of the article. Therefore, I
sketch only those features of the model whose understanding is essential to trace the fundamental forces
behind its welfare results.
Because Keuschnigg and Kohler use an overlapping
generation model, they do not have to confront
Ahearne’s difficulty of how to introduce the time preference parameter β in agents that never die.
The welfare gains were computed taking into account
that the capital stock will gradually adjust to its new
long-run equilibrium level after the trade reform is
implemented.
In more technical terms, the Second Welfare theorem
does not hold under increasing returns to scale; therefore, a Pareto optimum cannot typically be implemented by a free market economy. Dixit and Stiglitz
(1977) show, for example, that corrective measures
could involve taxes on some goods and subsidies on
others. By analogy, it is not difficult to envision environments in which the remedies would involve tariffs on
some imports and subsidies on some exports.

Keuschnigg, Christian, and Wilhelm Kohler (1996),
“Commercial Policy and Dynamic Adjustment Under
Monopolistic Competition,” Journal of International
Economics 40 (May): 373 – 409.
Klenow, Peter J., and Andres Rodriguez-Clare (1997),
“Quantifying Variety Gains from Trade Liberalization,”
manuscript, Graduate School of Business, University of
Chicago.
KPMG Peat Marwick/Policy Economics Group (1991),
“The Effects of a Free-Trade Agreement Between the U.S.
and Mexico” (Washington, D.C.: U.S. Council of the
Mexico – U.S. Business Committee, May).
Krugman, Paul R., and Maurice Obstfeld (1991), International Economics: Theory and Policy, 2nd ed. (New
York: Harper Collins).
Mendoza, Enrique (1991), “Real Business Cycles in a
Small Open Economy,” American Economic Review 81
(September): 797– 817.
Poterba, James M., and J. J. Rotemberg (1987), “Money
in the Utility Function: An Empirical Implementation,” in
New Approaches to Monetary Economics, ed. William A.
Barnett and Kenneth J. Singleton (New York: Cambridge
University Press), 219 – 40.
Romer, Paul (1987), “New Theories of Economic Growth,”
American Economic Review 77 (May, Papers and
Proceedings, 1987): 56 – 62.

REFERENCES
Ahearne, Alan (1999), “Trade Liberalization and Capital
Accumulation in Developing Economies: A Quantitative
Analysis,” manuscript, International Finance Division,
Board of Governors of the Federal Reserve System.

FEDERAL RESERVE BANK OF DALLAS

Full text of Review (Federal Reserve Bank of Dallas) : First Quarter 2000

FRASER