CAVE.6607.cp25.

p525-542 6/6/06 1:25 PM Page 525

CHAPTER 25

Interdependence and
Policy Coordination

C
hapter 18 showed how floating exchange rates could help insulate countries
from each other’s policy changes or from other disturbances. This complete insu-
lation or independence held only under certain very special conditions, how-
ever—most importantly the absence of capital flows. This chapter will show that even
with freely floating exchange rates, countries are in fact interdependent. What happens
in the United States has important effects in Europe or Japan and vice versa.
The fact that the world is so interdependent leads to a new topic: international
macroeconomic policy coordination. National policy makers may be able to do better
by setting their policies cooperatively than they can when each acts independently.

25.1 International Transmission of Disturbances

Under Floating Exchange Rates
Recall that under floating exchange rates, the overall balance of payments must sum to
zero. We last considered the effect of one country’s expansion on another country’s
economy in Chapter 18. At that point we were assuming that the net capital flow was
zero. It followed that the exchange rate always adjusted automatically so as to ensure
that the trade balance was zero. In a model in which the trade balance was the only
channel through which one country’s disturbances affected another’s, floating exchange
rates provided complete insulation. It was almost as if each country were a closed econ-
omy with no trade. In practice, substantial international synchronization of business
cycles has continued since 1973: worldwide recession in 1974–1975, 1980 to 1982, 1991,
and 2001, and worldwide recovery after each. The extent of synchronization seems, if
anything, to have exceeded that during the fixed exchange rate era. This may be related
in part to supply shocks or other commonly shared disturbances.
This section examines the two major routes via which disturbances can penetrate
through the insulation provided by floating exchange rates. The first is the presence of
capital flows, introduced in Chapters 22 and 23, which allow international transmission
via the trade balance. The second consists of various effects that exchange rate changes
can have on national economies other than the effect through the trade balance.

525
CAVE.6607.cp25.p525-542 6/6/06 1:25 PM Page 526

526 Chapter 25 ■ Interdependence and Policy Coordination

Transmission via Capital Flows

Even when the exchange rate adjusts so that the overall sum of the trade balance plus
the capital account is zero, the existence of any kind of net capital flow implies that the
trade balance is not zero. If one country goes into trade deficit, the change will be
transmitted to the rest of the world as a trade surplus. The trade imbalance is financed
by a flow of capital from the surplus country to the deficit country.
We now examine the two-country version of the Mundell-Fleming model of float-
ing exchange rates introduced in Chapter 23 to show how monetary or fiscal expansion
in one country is transmitted to the other. Any degree of capital mobility would be suf-
ficient to establish transmission. Perfect capital mobility will be assumed here, in part
because it gives a simpler model than partial capital mobility and in part because this
assumption has accurately described the major industrialized countries in recent years.
There will be no more modeling of the capital account as a finite flow responding to a
given interest rate differential.
The foreign country is modeled analogously to the home country. Call the home
country the United States and the foreign country, Europe. Figure 25.1 shows the two
side by side. Recall that under perfect capital mobility, arbitrage equates the U.S. and
European interest rates (omitting for now any expectation of future changes in the
exchange rate): i 5 i*. This means that the two points representing the two countries’
equilibrium positions must lie on the same horizontal line. Otherwise, if one country
had a higher interest rate, there would be a potentially infinite demand for its assets,
with potentially infinite borrowing in the low-interest country. E and E* represent the
initial equilibrium points.

Fiscal Expansion in a Large Country

A U.S. fiscal expansion would shift the IS curve out to IS9 if there were no other change
in the exchange rate. However, as in Chapter 23, the large capital inflow that would be
attracted by the higher interest rate at point A causes the dollar to appreciate, worsen-
ing the U.S. trade balance and shifting the IS curve back to the left. Under perfect cap-
ital mobility, the appreciation of the dollar and the backward shift of the IS curve
continue until the parity condition, i 5 i*, is restored. Previous chapters have all taken
i* as exogenously fixed, with the implication that the equilibrium is back at the starting
point, E. But i* need not be exogenous, as we will now see.
Saying that the dollar appreciates is the same as saying that the European currency,
the euro, depreciates. Saying that the U.S. trade balance worsens is the same as saying
that the European trade balance improves. Therefore, as the U.S. IS curve shifts left, the
European IS* curve shifts right. Intuitively, expenditure has been switched from U.S.
goods to European goods. The two curves will shift until their intersections with their
respective LM curves occur at the same horizontal level—at points B and B*, respec-
tively. Only then are interest rates equalized. Both intersections lie on the country’s
BP curve, as floating rates imply they must; but the curve has shifted to a higher level.
Figure 25.1 shows how two of the (overly strong) results derived earlier must now
be modified. First, the insulation property of floating rates is undone by international
CAVE.6607.cp25.p525-542 6/6/06 1:25 PM Page 527

25.1 ■ International Transmission of Disturbances Under Floating Exchange Rates 527

FIGURE 25.1
Fiscal Expansion in a Two-Country, Mundell-Fleming Model
Despite perfect capital mobility, the United States can drive up the interest rate because of its
large presence in world capital markets. After a U.S. fiscal expansion (to A) causes the dollar
to appreciate, the European IS* curve shifts right as the IS curve in the United States shifts left.
Equilibrium entails expansion both for the United States at B and Europe at B*.

(a) United States (b) Europe

i i*
LM LM*

A
B B*
BP ′′
BP
E IS ′ E*

IS ′′ IS* ′′
IS IS*
0 Y 0 Y*

capital mobility. Through the channel of the trade balance, the fiscal expansion is trans-
mitted positively to Europe as an increase in the demand for European output.
Furthermore, a property of fiscal expansion established in Chapter 23, that it is
ineffective at raising domestic output under perfect capital mobility, is now also
undone. Previously it was assumed the home country was sufficiently small in world
capital markets that it could take the foreign interest rate, i*, as given. If the country is
as large as the United States, however, then it is large enough to drive up interest rates
everywhere in the world simultaneously. The fiscal expansion succeeds in raising i to
the extent that it also raises i*. Thus it succeeds in raising Y to the extent that it also
raises Y*, without violating equilibrium in the financial markets. The large-country
assumption restores effectiveness to fiscal policy despite perfect capital mobility.1
The U.S. fiscal expansion that generated U.S. recovery in 1983–1984 is a perfect
illustration of international transmission. As already noted, U.S. interest rates rose,
attracting a capital inflow from abroad, and the dollar continued to appreciate against
the European currencies and the yen. The U.S. trade deficit rose sharply, resulting in a

1
It is the assumption of capital mobility, however, that restores transmission between countries, regardless of
their size. Even if the home country is small, its fiscal expansion will have a positive dollar effect on the rest of
the world that, although small as a proportion of foreign GDP, is significant relative to domestic GDP.
Similarly, a fiscal expansion abroad will have a positive effect on the home country that is significant relative
to home GDP. The original reference is Robert Mundell, “A Reply, Capital Mobility and Size,” Canadian
Journal of Economics and Political Science, 30 (1964): 421–431.
CAVE.6607.cp25.p525-542 6/6/06 1:25 PM Page 528

528 Chapter 25 ■ Interdependence and Policy Coordination

corresponding improvement in trade balances in Europe, Japan, and almost every-

where else. The U.S. expansion thus did much to pull the rest of the world out of reces-
sion at the same time as it did so domestically.
Another example is Germany’s 1990–1992 increase in spending in association with
its absorption of the former East Germany. The higher spending drove up German
interest rates, attracted a capital inflow, appreciated the mark, and in a short time
changed a large German current-account surplus into a deficit. Germany’s trade bal-
ance loss was its trading partners’ gain.
The U.S. fiscal expansion that contributed to U.S. recovery after the recession of
2001 did not fit the pattern of higher interest rates and a stronger dollar. Why not? It
was accompanied by a strong monetary expansion.2

Monetary Expansion in a Large Country

We next consider, in Figure 25.2, a U.S. monetary expansion. It would shift the LM
curve out to LM9 if there were no change in the exchange rate. However, as was
explained previously, the large capital outflow induced by the lower interest rate at
point A causes the dollar to depreciate. This improves the U.S. trade balance (relative
to what it would be at point A) and shifts the IS curve out to the right. Under perfect
capital mobility, the dollar depreciation and the outward shift of the IS curve must con-
tinue until i 5 i* is restored. Previously i* was taken as exogenously fixed, with the
implication that the equilibrium was all the way out to point B in Figure 23.4(d). There
was very strong stimulus to output, all coming from net foreign demand. This need no
longer be the case, however. Saying that the dollar depreciates is the same as saying
that the euro appreciates. Saying that the U.S. trade balance improves is the same as
saying the European trade balance worsens. Therefore, as the U.S. IS curve shifts right,
the European IS* curve shifts left. The two curves will shift until their intersections
with their respective LM curves occur at the same horizontal level: at points B and B*,
respectively, in Figure 25.2.
As with the fiscal expansion, two of the strong results derived earlier are over-
turned. First, the monetary expansion is transmitted abroad. However, the transmission
is now negative, or inverse: European income falls because of the lost net exports. As a
result of lower transactions demand for money in Europe, i* falls as well.
Because the U.S. monetary expansion succeeds in lowering i*, it lowers i as well:
The United States is large in world financial markets, so it can drive down interest rates
everywhere simultaneously. This allows the second new finding. The monetary expan-
sion does not cause as big a depreciation as in the small-country case of Chapter 23, so
there is not as large a stimulus to net foreign demand. The lower interest rate means
that some of the expansion will come from domestic demand. This is a more realistic
result than when it appeared that all of the expansion had to come from a large
increase in net foreign demand. The British monetary contraction that began in 1979,

2
The monetized fiscal expansion of 2001 thus had more in common with the Vietnam-related expansion of the
late 1960s than with Reaganomics.
CAVE.6607.cp25.p525-542 6/6/06 1:25 PM Page 529

25.1 ■ International Transmission of Disturbances Under Floating Exchange Rates 529

FIGURE 25.2
Monetary Expansion in a Two-Country, Mundell-Fleming Model
After a U.S. monetary expansion (to A) causes the dollar to depreciate, the European IS* curve
shifts left as the IS curve in the United States shifts right. Equilibrium entails expansion for the
United States at B but contraction for Europe at B*.

(a) United States (b) Europe

i i*

LM LM*

LM ′

BP
E E*
BP ′′
B B*
IS ′ IS*
A IS IS* ′

0 Y 0 Y*

for example, resulted in a subsequent loss of net exports as a result of the higher value
of the pound, and it also resulted in a loss of domestic demand in construction and
other sectors, as a result of the higher interest rates. The same was true of the U.S. mon-
etary contraction of 1980 to 1982.

Transmission via (Nontrade) Exchange Rate Effects

Whether or not there are international capital flows, and therefore nonzero trade bal-
ances, developments in one country can be transmitted to the other country’s economy
if the exchange rate has effects in addition to its effect on the trade balance. To begin
with, an increase in the euro/dollar exchange rate will certainly be felt in Europe as an
increase in the euro prices that Europeans have to pay for imports. For there also to be
an impact on European output requires some additional effect. Four possibilities are
effects on saving, money demand, prices of imported inputs, and wages. All four can
result from higher import prices in Europe.
The effect on European saving is the Laursen-Metzler-Harberger effect, devel-
oped in the appendix to Chapter 18. A rise in the euro/dollar rate is an adverse shift in
the terms of trade for European households: It is a fall in the purchasing power of a unit
of European output over a basket of consumer goods that includes imports. Europeans
react as they would to any loss in real income, by reducing saving (for any given level
of real income measured in domestic units) so as to smooth consumption over time.
CAVE.6607.cp25.p525-542 6/6/06 1:25 PM Page 530

530 Chapter 25 ■ Interdependence and Policy Coordination

The higher level of expenditure raises real output and employment in Europe. If the
increase in the euro/dollar rate originated in the U.S. fiscal expansion, then the rise in
European output means that transmission is positive. If it originated in a U.S. monetary
contraction, then the rise in European output means that transmission is inverse. In
both cases, the Laursen-Metzler-Harberger effect reinforces the same pattern of trans-
mission that we have just seen brought about by high capital mobility.
The next three possible exchange rate effects, however, go the other way. They are
each reasons why an increase in the euro/dollar rate might lower output in Europe.
First is a possible effect via the demand for money. Previous chapters have viewed
P, the price level for domestically produced goods, as the appropriate variable for
determining money demand. In the Mundell-Fleming model, P is fixed in the short run.
Thus the exchange rate does not enter into the money-demand equation. However,
the CPI could be considered the appropriate variable for determining money demand
as easily as P. If imports have a weight of a in the European CPI, then a 1 percent
increase in the euro/dollar rate that raises European import prices by 1 percent will
raise the CPI and reduce the real money supply by a percent. Thus it will shift the LM
curve to the left and have a contractionary effect on European output. If the dollar
appreciation originated in a U.S. fiscal expansion, then the potential decrease in Euro-
pean output represents inverse transmission. If it originated in a U.S. monetary con-
traction, then it represents positive transmission. In either case, the effect via money
demand is the opposite of the effect via the trade balance that appears in the standard
Mundell-Fleming model. The contractionary effects in Europe was one of the argu-
ments open to those Europeans who claimed that the U.S. policy mix of the early
1980s—tight money and a loose budget, resulting in a strong dollar—had adverse
effects on European growth.
The effects just mentioned come via aggregate demand. There are two remaining
effects, both of which come via aggregate supply rather than aggregate demand. If the
price of oil or other imported inputs is set in dollars, then the increase in the euro/dol-
lar rate will raise the price of the input for European firms. Finally, if European wages
are tied or indexed to the European CPI, then the increase in the euro/dollar rate will
raise European wages relative to the price of goods produced in Europe. Either way,
European firms find that their input costs have gone up relative to the price of the
goods they produce, which will cause them to cut back on output. The contractionary
supply effects, like the contractionary demand effects, can reverse the transmission
results of the Mundell-Fleming model. To understand aggregate supply effects fully
requires a more detailed examination of the supply relationship than we have previ-
ously carried out. The next chapter is a convenient place to do it.

25.2 Econometric Models of the Interdependent World Economy

Having looked at a bewildering variety of possible routes for transmission of monetary
and fiscal policy, some affecting other countries’ levels of economic activity positively,
some affecting them negatively, we naturally might wonder which effects are likely to
dominate in practice.
CAVE.6607.cp25.p525-542 6/6/06 1:25 PM Page 531

25.2 ■ Econometric Models of the Interdependent World Economy 531

Economists have built a number of econometric models of the world macroecon-

omy, each including the major countries or blocs of countries. Often these models are
quite large in terms of the number of equations or the amount of work they require.
Some of the models are built and maintained in private consulting firms that make
economic forecasts for corporate clients. Some are at agencies of national governments
or at multinational public institutions. Some are at universities.
The models also differ in their economic philosophies. Some are extremely Keynes-
ian, showing little or no effect of a monetary expansion on prices. Others represent
the New Classical school of thought, featuring rational expectations and frictionless
determination of wages and prices. Most adopt the intermediate synthesis view taken
in this text. This still allows for tremendous divergence among the models, however.
Even within the same overall model specification, different estimates of parameter
values can have very different implications concerning issues such as whether interna-
tional transmission is positive or negative.
The models are often used in simulations to predict the effect of a given policy
change, relative to some baseline predicted path for the world economy. It can be diffi-
cult to compare the results of any two models because the policy experiment being
conducted may differ between the two. The simulation results from one model say that
a Japanese fiscal expansion would appreciate the yen, and those from another say it
would depreciate the yen. One possibility is that the models truly differ; the first, for
example, incorporating a high degree of capital mobility for Japan and the second a
low degree. However, another possibility is that the first simulation is considering the
experiment with the M1 money supply held constant, so that the fiscal expansion
pushes the interest rate far up, whereas the second is holding something else constant
(the monetary base, or even the interest rate itself), with the result that a fiscal expan-
sion automatically leads to an accompanying increase in M1.
A project undertaken under the auspices of the Brookings Institution asked twelve
leading international econometric models to perform simulations for some carefully
specified macroeconomic policy experiments.3 Tables 25.1 and 25.2 show the results in
the second year after a fiscal expansion and a monetary expansion, respectively. The
twelve models with their abbreviations are as follow: MCM—the Federal Reserve
Board’s Multi-Country Model; COMPACT—the European Community Commission’s
model; EPA—the Japanese Economic Planning Agency’s model; LINK—Project Link,
which puts together the various models of national economies that had already been
built in the respective countries; LIV—the Liverpool model of Patrick Minford, a new
classical British economist who advised Prime Minister Margaret Thatcher; MSG—the
McKibbin-Sachs Global model, which assumes rational expectation, but is otherwise
somewhat Keynesian, built by Jeffrey Sachs of Columbia University and Warwick
McKibbin of Australia National University; MINIMOD—a smaller approximation of
the MCM, built by staff of the International Monetary Fund; VAR—estimates by

3
The model simulations were presented and evaluated in Ralph Bryant, Dale Henderson, Gerald Holtham,
Peter Hooper, and Steven Symansky, eds., Empirical Macroeconomics for Interdependent Economies (Wash-
ington, DC: Brookings Institution, 1988).
CAVE.6607.cp25.p525-542 6/6/06 1:25 PM Page 532

532 Chapter 25 ■ Interdependence and Policy Coordination

TABLE 25.1
Fiscal Policy: Simulation Effect in Second Year
of Increase in Government Expenditure (1 Percent of Output)
Currency
Y CPI i Value CA CA* i* CPI* Y*
Effect in United States Effect in Rest of OECD
Fiscal Expansion
in United States (in percentage) (Pts.) (in percentage) ($B) ($B) (Pts.) (in percentage)
MCM 11.8 10.4 11.7 12.8 216.5 18.9 10.4 10.4 10.7
COMPACTa 11.2 10.6 11.5 10.6 211.6 16.6 10.3 10.2 10.3
EPAb 11.7 10.9 12.2 11.9 220.5 19.3 10.5 10.3 10.9
LINK 11.2 10.5 10.2 20.1 26.4 11.9 NA 20.0 10.1
LIV 10.6 10.2 10.4 11.0 27.0 13.4 10.1 10.6 20.0
MSG 10.9 20.1 10.9 13.2 221.6 122.7 11.0 10.5 10.3
MINIMOD 11.0 10.3 11.1 11.0 28.5 15.5 10.2 10.1 10.3
OECD 11.1 10.6 11.7 10.4 214.2 111.4 10.7 10.3 10.4
TAYLORc 10.6 10.5 10.3 14.0 NA NA 10.2 10.4 10.4
WHARTON 11.4 10.3 11.1 22.1 215.4 15.3 10.6 20.1 10.2
DRI 12.1 10.4 11.6 13.2 222.0 10.8 10.4 10.3 10.7
Currency
Y CPI i Value CA CA* i* CPI* Y*
Effect in Rest of OECD Effect in United States
Fiscal Expansion
in Rest of OECD (in percentage) (Pts.) (in percentage) ($B) ($B) (Pts.) (in percentage)
MCM 11.4 10.3 10.6 10.3 27.2 17.9 10.5 10.2 10.5
COMPACTa 11.3 10.8 10.4 20.6 29.3 13.0 10.0 10.1 10.2
EPAb 12.3 10.7 10.3 20.7 213.1 14.7 10.6 10.3 10.3
LINK 11.2 10.1 NA 20.1 26.1 16.3 10.0 10.0 10.2
LIV 10.3 10.8 10.0 13.3 217.2 111.9 10.8 13.1 20.5
MSG 11.1 10.1 11.4 12.9 25.3 110.5 11.3 10.6 10.4
MINIMOD 11.6 10.2 10.9 10.6 22.2 13.2 10.3 10.2 10.1
OECD 11.5 10.7 11.9 10.9 26.9 13.3 10.3 10.2 10.1
TAYLORc 11.6 11.2 10.6 12.7 NA NA 10.4 10.9 10.6
WHARTON 13.2 20.8 10.8 22.4 25.5 14.7 10.1 20.0 10.0
a
Non-U.S. short-term interest rate NA; long-term rate reported instead.
b
Non-U.S. current account refers to Japan, Germany, United Kingdom, and Canada.
c
CPI NA; GNP deflator reported instead.
Source: Frankel and Rockett (1988).

Christopher Sims and Robert Litterman obtained by Vector AutoRegression (a tech-

nique that uses no economic theory, but merely looks for regular patterns in the data);
OECD—the Interlink model built by staff members at the Organization of Economic
Cooperation and Development (an agency with a membership of thirty industrialized
countries and a Secretariat in Paris); TAYLOR—a rational expectations model by John
Taylor of Stanford University, an official in both the first and second Bush administra-
tions; WHARTON—a generally Keynesian model, originally built by Nobel laureate
Lawrence Klein of the University of Pennsylvania; and DRI—the model of Data
Resources, Inc., Lexington, Massachusetts, a firm that sells economic forecasts to cor-
porations and government agencies.
CAVE.6607.cp25.p525-542 6/6/06 1:25 PM Page 533

25.2 ■ Econometric Models of the Interdependent World Economy 533

The Results for Fiscal Policy

Table 25.1 summarizes the effects of a fiscal expansion, an increase in government
spending equal to 1 percent of income, according to eleven models in the simulations.
The variables shown are output, the consumer price index, the short-term interest rate,
the exchange rate, and the current account. The first five columns show the variables in
the region originating the fiscal expansion, the last four columns the foreign region.
As expected, the models all show a positive effect on output. The numbers in the
first column can be read as fiscal multipliers.4 They are mostly in the range of 1 to 2.
Almost all the models show increases in the price level and the interest rate, from
which follows some crowding-out of interest-sensitive sectors such as construction.
The main ambiguity in theory, as we saw in Section 23.1, is whether the fiscal expan-
sion causes the currency to appreciate: whether capital mobility is sufficiently high that
the capital inflow attracted by higher interest rates is more than enough to finance
the increased imports resulting from higher income. However, the eleven models in
Table 25.1 show relatively little disagreement in practice. All but two show an apprecia-
tion of the dollar when the United States is the country initiating the fiscal expansion.
This would not have been the case in the 1960s; it reflects the high degree of capital
mobility that has evolved.5
In almost all the models, the simulations show that fiscal expansion is transmitted
positively to the foreign region.This is not surprising because the current account worsens
in the originating region and thus improves in the foreign region. The positive transmis-
sion does indicate, however, that the three possible contractionary effects of a currency
depreciation (the ones studied in Section 25.1, via money demand, wages, or imported-
input prices) either are not operating, or at least are not operating strongly enough to
outweigh the increase in net export demand falling on the goods of the foreign region.

The Results of Monetary Policy

Table 25.2 summarizes the effects of a monetary expansion equal to 4 percent of the
money supply (phased in over the first year). The simulations show more conflict
among the models than do the results of a fiscal expansion. They all agree that the
monetary expansion drives down the interest rate and thereby stimulates domestic
income, and they generally agree that it depreciates the currency. Yet they divide
almost evenly on the question of whether the domestic trade balance improves, causing
the foreign trade balance to worsen and foreign income to decrease. That is, they dis-
agree on whether international transmission is inverse.6

4
Because (DY / Y) / (DG / Y) is the same as DY / DG.
5
When the fiscal expansion originates in other countries, the appreciation of their currency is not as great as
when the fiscal expansion originates in the United States. Indeed, there are four models that indicate a depre-
ciation of the foreign currencies against the dollar. This largely reflects a belief that Japan and Europe are not
as open financially as the United States (and perhaps also that the LM curve is steeper in the United States, so
that interest rates tend to rise more easily than in the rest of the world).
6
The Mundell-Fleming model says that this inverse transmission should occur. As we saw in the preceding
chapter, the lower interest rate that results from a monetary expansion leads to a net capital outflow, which
corresponds to a current-account deficit abroad. However, the introduction of expectations into investors’
asset preferences, as in Chapter 27, can reverse this effect.
CAVE.6607.cp25.p525-542 6/6/06 1:25 PM Page 534

534 Chapter 25 ■ Interdependence and Policy Coordination

TABLE 25.2
Monetary Policy: Simulation Effect in Second Year
of Increase in Money Suppy (4 percent)a
Currency
Y CPI i Value CA CA* i* CPI* Y*
Effect in United States Effect in Rest of OECD
Monetary Expansion
in United States (in percentage) (Pts.) (in percentage) ($B) ($B) (Pts.) (in percentage)
MCM 11.5 10.4 22.2 26.0 23.1 23.5 20.5 20.6 20.7
COMPACTb 11.0 10.8 22.4 24.0 22.8 11.2 20.5 20.4 10.2
EPAc 11.2 11.0 22.2 26.4 21.6 210.1 20.6 20.5 20.4
LINK 11.0 20.4 21.4 22.3 25.9 11.5 NA 20.1 10.1
LIV 10.1 13.7 20.3 23.9 213.0 10.1 20.1 20.0 20.0
MSG 10.3 11.5 20.8 22.0 12.6 24.4 21.2 20.7 10.4
MINIMOD 11.0 10.8 21.8 25.7 12.8 24.7 20.1 20.2 20.2
VARd 13.0 10.4 21.9 222.9 14.9 15.1 10.3 10.1 10.4
OECD 11.6 10.7 20.8 22.6 28.4 13.1 20.1 20.1 10.3
TAYLORd 10.6 11.2 20.4 24.9 NA NA 20.1 20.2 20.2
WHARTON 10.7 10.0 22.1 21.0 25.1 15.3 21.3 20.1 10.4
DRI 11.8 10.4 22.3 214.6 21.4 114.5 21.1 21.3 20.6
Currency
Y CPI i Value CA CA* i* CPI* Y*
Effect in Rest of OECD Effect in United States
Monetary Expansion
in Rest of OECD (in percentage) (Pts.) (in percentage) ($B) ($B) (Pts.) (in percentage)
MCM 11.5 10.6 22.1 25.4 13.5 10.1 20.2 20.2 20.0
COMPACTb 10.8 11.0 21.0 22.3 25.2 11.9 10.0 10.1 10.1
EPAc 10.0 10.0 20.1 20.1 20.1 10.1 20.0 20.0 10.0
LINKe 10.8 20.6 NA 22.3 21.4 13.5 10.0 20.0 10.1
LIV 10.4 12.8 20.9 28.4 17.1 28.2 21.1 23.4 11.6
MSG 10.2 11.5 20.7 21.4 215.9 112.0 21.2 20.6 10.3
MINIMOD 10.8 10.2 21.8 24.8 13.6 21.4 20.6 20.5 20.3
VARb 10.7 20.5 23.0 25.5 15.2 210.0 10.6 20.7 11.2
OECD 10.8 10.3 21.3 22.1 21.6 12.3 20.2 10.1 10.1
TAYLORb 10.8 10.7 20.3 23.5 NA NA 20.2 20.5 20.1
WHARTON 10.2 20.1 20.8 10.2 12.6 10.5 10.0 10.0 10.0
a
The increase in the money supply is phased in over four quarters.
b
Non-U.S. short-term interest rate NA; long-term rate reported instead.
c
Non-U.S. current account is Japan, Germany, United Kingdom, and Canada.
d
CPI NA; GNP deflator reported instead.
e
Appreciation of non-U.S. currency NA; depreciation of dollar reported instead.
Source: Frankel and Rockett (1988).

Many of the models say that the higher imports drawn in by higher income are
more than enough to offset the effect of the exchange rate on the trade balance, with
the result that the expansion is transmitted positively, rather than negatively, to the for-
eign region. In large part this comes from observing the effect in the second year after
the change in policy. The full effect of the exchange rate on the trade balance is not
CAVE.6607.cp25.p525-542 6/6/06 1:25 PM Page 535

25.3 ■ International Macroeconomic Policy Coordination 535

felt until the third year or later. However, it is possible to sum up the results of all the
models by saying that under floating exchange rates, one country’s monetary expan-
sion appears to have only small effects on other countries’ incomes because the income
and exchange rate effects on the trade balance roughly cancel each other out.

25.3 International Macroeconomic Policy Coordination

We have examined a wide variety of channels whereby policy changes in one country
have effects in other countries. Policy makers have become increasingly aware of this
interdependence of their national economies.

The Institutions of International Cooperation

A number of institutions have been established to facilitate discussion of economic
issues that concern all countries and to facilitate coordination of their policies. The
IMF conducts “surveillance” of the policies of the major industrialized countries,
although its influence on them is inevitably far less than on the poorer, indebted coun-
tries who have little choice but to listen to the fund’s advice. Each year the OECD
sponsors meetings of cabinet ministers from its member countries, supported by regu-
lar meetings of the Economic Policy Committee and Working Party 3, in addition to a
plethora of other meetings of specialists from the member countries dealing with par-
ticular economic sectors. Central bankers from the Group of Ten industrialized coun-
tries meet regularly, often in association with the Bank for International Settlements.7
In 1975, at the suggestion of French president Valéry Giscard d’Estaing, the heads
of states of large industrialized economies met at Rambouillet, France. The purpose on
that occasion was to ratify politically the movement from fixed exchange rates to float-
ing exchange rates, which market forces had imposed on the world monetary system a
few years earlier. The summit meetings have continued each year since then, bringing
together leaders from the group of seven largest industrialized countries, known as the
G-7: the United States, Japan, Germany, France, the United Kingdom, Italy, and
Canada. The most substantive G-7 summit meeting took place in Bonn, Germany, in
1978. There Japan and Germany agreed to the U.S. plan for joint expansion, according
to which the three countries would be the “locomotives” pulling the world economy
out of the stagnation that had followed the 1974 oil shock. The U.S. motive behind the
locomotive theory was the fear that if the United States continued to expand on its
own, it would suffer an enlarged trade deficit.
In recent years, the earlier spirit of informal discussion has been lost and the sum-
mit meetings have become mammoth media events. (The group was expanded to
include Russia in 1997, so it is now the G-8.) Beginning in September 1985, the focus
shifted—for the purpose of serious economic policy-making—to the regular meetings

7
The BIS was originally set up after World War I to facilitate the reparations payments that appeared in the
discussion of the transfer problem in Chapter 17. From its headquarters in Basel, Switzerland, the BIS contin-
ues to function as the central bankers’ exclusive club.
CAVE.6607.cp25.p525-542 6/6/06 1:25 PM Page 536

536 Chapter 25 ■ Interdependence and Policy Coordination

of the finance ministers. The occasion then was a meeting that took place at the Plaza
Hotel in New York and produced the Plaza Accord, under which the United States
agreed to cooperate in bringing down the value of the dollar. The finance ministers
meet regularly to discuss the macroeconomic and financial interactions of their
economies.

The Theory of Gains from International Policy Coordination

How should all these meetings and institutions be viewed? Are the meetings just
media events, opportunities for the heads of state to escape domestic political difficul-
ties and be seen on television looking statesmanlike? Are the institutions simply over-
paid bureaucracies whose principal mission is the sampling of continental cuisine?
Although sometimes it might seem that way, there are some good arguments in favor
of international cooperation.
There is an elegant theory of the economic gains from international macroeco-
nomic policy coordination: Two or more countries will in general be better able to
attain their economic objectives if they set their policies jointly than if they set them
independently. The alternative, in which each country independently sets its own poli-
cies, taking the policies of the others as given, is called the noncooperative equilibrium,
also termed the Nash equilibrium.
There are a number of ways in which spillover effects among countries can render
the noncooperative equilibrium unsatisfactory. Each defines a “game” between national
policy makers.
The game that comes up most often, particularly when the world is in recession
because of inadequate demand, could be called “exporting unemployment.” Consider
the United States and Europe. Each must decide whether or not to follow expansion-
ary demand policies. Table 25.3 shows the four possible outcomes. If Europe has a
trade balance objective, it will be reluctant to expand, for fear that the United States
will be less expansionary and leave Europe with a trade deficit. Similarly, the United
States will be reluctant to expand, for fear that Europe will be less expansionary and
leave the United States with a trade deficit. The result is that each country will hold
back its level of demand in an effort to improve its trade balance at its neighbor’s
expense. This policy is self-defeating when the countries try it simultaneously, plunging
the world into a recession where everyone loses. This noncooperative equilibrium
occurs in the first cell of the table.
The solution to the exporting unemployment problem is the “locomotive strat-
egy”: Both countries should agree to expand simultaneously, whether by means of fis-

TABLE 25.3
The Game of “Exporting Unemployment”
United States Contracts United States Expands
Europe Contracts Recession in both countries; TB 5 0 TB favors Europe
Europe Expands TB favors United States Boom in both countries; TB 5 0
CAVE.6607.cp25.p525-542 6/6/06 1:25 PM Page 537

25.3 ■ International Macroeconomic Policy Coordination 537

cal policy or by some combination of fiscal and monetary policy, so that output is
higher everywhere with no change in the trade balance.8 This is the logic behind the
policy package adopted at the Bonn summit of 1978.
A particular variety of the exporting unemployment game, called “competitive
depreciation,” arises when fiscal policy is the tool used and exchange rates are floating.
Then each country has an especially strong temptation to contract because a fiscal con-
traction will lower interest rates, cause its currency to depreciate, and provide further
improvement in its trade balance at its neighbor’s expense. In the 1930s each country
devalued in a (vain) attempt to gain a trade advantage against the other. (Recall
“beggar-thy-neighbor” policies.)
Other games are possible as well. Under a system of floating exchange rates, one
possibility is the game of “competitive appreciation” (the opposite of the competitive
depreciation game). It is illustrated in Table 25.4. This game depends on the assump-
tion that each country has as its ultimate objective, in addition to high output, low
inflation as measured by the CPI. It can, of course, be difficult to attain both of these
objectives simultaneously. There is a trick, however, whereby a country can attain both
objectives. It can keep the overall CPI stable, even if output is growing rapidly and
thereby putting upward pressure on the prices of domestically produced goods. The
trick is to appreciate the currency—for example, through a combination of tight mon-
etary policy and loose fiscal policy that drives up interest rates and makes the coun-
try’s assets attractive to international investors. The point is that the strong currency
will reduce the price of imports, when expressed in domestic currency. To the extent
that imports have a share in the CPI, the overall inflation rate can be kept down, even
if the prices of domestic goods are rising. Some economists have attributed such a
motive to the U.S. government’s adoption of its 1980s policy mix of tight money and
loose fiscal policy.9
Notice, however, that this trick can only be brought about at the expense of the
country’s neighbors—by exporting inflation. If the first country experiences an appreci-
ation and downward pressure on its CPI, then its neighbors are experiencing deprecia-
tion and upward pressure on their CPIs. The noncooperative equilibrium appears again
in the first cell of Table 25.4. Both countries are keeping interest rates high in unsuc-
cessful attempts to appreciate their currency. The result is worldwide recession. The
cooperative solution is that both agree simultaneously to lower interest rates. Then they
can attain stronger economies with no adverse effect on their exchange rates or CPIs.
A more permanent solution to problems of competitive appreciation or deprecia-
tion would be for the countries to agree to a system of fixed exchange rates. Then the
leaders do not have to get together to negotiate over specific macroeconomic policies.
Perceptions that competitive devaluation had helped prolong the Great Depression of
the 1930s were a major reason why the delegates to the Bretton Woods conference of
1944 chose a system of fixed exchange rates for the postwar international monetary

8
The supplement to this chapter presents the more complete analysis of the exporting unemployment game
that is relevant when each country has a continuous range of macroeconomic expansion or contraction
options from which to choose, as opposed to the simple choice presented in Table 25.3 (expand vs. contract).
9
Jeffrey Sachs, “The Dollar and Policy Mix: 1985,” Brookings Papers on Economic Activity, 1 (1985): 117–186.
CAVE.6607.cp25.p525-542 6/6/06 1:25 PM Page 538

538 Chapter 25 ■ Interdependence and Policy Coordination

TABLE 25.4
The Game of “Competitive Appreciation”
United States Raises United States Lowers
Interest Rate, i Interest Rate, i
Europe Raises i* Recession in both: no change Dollar depreciates: U.S. import prices and
in exchange rate or CPI CPI go up; European CPI goes down
Europe Lowers i* Dollar appreciates: U.S. CPI Boom in both countries: no change in
goes down; European CPI exchange rate or CPI
goes up

system. In the language of the Articles of Agreement of the International Monetary

Fund, the members agreed to refrain from manipulating their exchange rates to seek
“unfair advantage.”

Obstacles to Successful Coordination

If international policy coordination was really as easy as Tables 25.3 and 25.4 make it
appear, it might seem odd that agreements do not take place more often than they do.
A number of obstacles make coordination difficult in practice. Even if the setting is as
simple as we have laid out, there is first of all the problem of dividing the gains from
cooperation between the two countries. In any game there is the possibility that both
parties will bargain “tough,” with the result that the potential gains are lost to both.
Then there is the issue of enforcement of the agreement. The United States, knowing
that Europe has set its money supply at the level agreed on, may be tempted to reduce
its own money supply because that will move it to higher levels of economic welfare. Of
course, if the bargain were explicit, this deviation from the agreement would constitute
cheating. The gains would be at most short run; when Europe realizes that America has
broken the agreement, it too will change its policy settings, causing a return to the non-
cooperative state. Even if no automatic penalty is built in for cheating, America will be
discouraged from breaking the agreement if it is concerned that it would acquire an
undesirable reputation as an untrustworthy party in potential future agreements.
A different difficulty arises from the fact that in the games described so far, policy
makers are maximizing their objectives only period by period. If coordination consti-
tutes joint expansion, as in the locomotive game, then this will raise inflation in the cur-
rent period. If the current period is the only one that matters, then the policy makers
will already have factored in the inflation correctly when mapping out their policy
preferences. However, the expansion will also raise expected inflation in the next
period, so workers will demand higher wages and there will be a higher level of actual
inflation in the future for any given level of output, as we see in Chapter 26. In such cir-
cumstances, coordinating period by period may actually be harmful in the long run.10

10
The damage to inflation-fighting credibility is offered as an argument why countries might be better off
renouncing coordination altogether, in Kenneth Rogoff, “International Macroeconomic Policy Coordination
May Be Counterproductive,” Journal of International Economics, 18 (May 1985): 199–217.
CAVE.6607.cp25.p525-542 6/6/06 1:25 PM Page 539

25.3 ■ International Macroeconomic Policy Coordination 539

A final obstacle to successful macroeconomic policy coordination arises from

uncertainty. So far we have assumed that policy makers know precisely (1) what their
proper objectives are (for example, what weight should be placed on full employment
versus inflation); (2) where their economies are relative to the target optimums (the
baseline forecast); and (3) what effect given changes in the policy instruments will have
on the economy (the size of the multipliers in the correct model of the world macro-
economy). In reality, however, policy makers are uncertain about each of the three. The
third kind of uncertainty is illustrated in Table 25.2 by the disagreement among the
major econometric models as to the effects of monetary policy. All three kinds of
uncertainty make it difficult for each country in the bargaining process to know even
what policy changes it should want its partners to make. A number of pessimistic con-
clusions emerge. Given differing perceptions, the policy makers may not be able to
agree on a coordination package. Even if they do agree, the effects may be different
from those anticipated.11
The standard German view of the joint expansion agreed on at the 1978 Bonn
summit is that it turned out to have been undesirable because by the end of the decade
the priority had shifted back to fighting inflation. One possible way to understand this
view is to see it as an example of uncertainty about the baseline position of the econ-
omy relative to the optimum: The 1979 oil price increase associated with the fall of the
shah of Iran moved the world economy to a more inflationary position than had been
anticipated at the time of the summit meeting. Another way to understand it is to see it
as an example of disagreement over the correct model. In the model that the United
States has in mind, a monetary expansion can raise output and employment, whereas
in the Germans’ model, monetary expansion simply goes into prices. Conflicting per-
ceptions as to how the economy works make international coordination difficult, as
much today as in 1978.
Thus the gains from international coordination are not as automatic as is sug-
gested by the simple model illustrated here. The potential gains are still there, how-
ever. There are also other, more broadly defined arguments in favor of cooperation
that include, for example, the exchange of information among countries. Sometimes
the international meetings help give the individual countries the clarity of vision, sense
of purpose, and political momentum needed to accomplish tasks (like cutting budget
deficits) that some leaders in the individual governments considered to be in their indi-
vidual national interests all along but were unable to accomplish in isolation.
It is inevitable that national leaders will, to an increasing extent, have to work
together, particularly in time of crisis. It is good that policy makers maintain steady
contact and do not wait for a crisis to become acquainted.

11
Jeffrey Frankel and Katharine Rockett, “International Macroeconomic Policy Coordination. When Policy-
Makers Do Not Agree on the True Model,” American Economic Review (June 1988): 318–340. Furthermore,
even if the effects of coordination are as anticipated, the gains are generally estimated to be small, as was first
shown in Gilles Oudiz and Jeffrey Sachs, “Macroeconomic Policy Coordination Among the Industrial Eco-
nomies,” Brookings Papers on Economic Activity, 1 (1984): 1–64. The reason is that the magnitude of interna-
tional transmission effects is estimated to be relatively small.
CAVE.6607.cp25.p525-542 6/6/06 1:25 PM Page 540

540 Chapter 25 ■ Interdependence and Policy Coordination

25.4 Summary
This chapter showed that even with freely floating exchange rates, countries are inter-
dependent. Only a small country can afford to ignore the effects its policy changes
have on its trading partner because for a large country some of the effects bounce
back. This chapter extended the Mundell-Fleming model of Chapter 23 to large coun-
tries and considered other extensions as well. The most important channel of transmis-
sion between countries is the trade balance. Wherever there are net capital flows, there
are nonzero trade balances. There are also other possible channels of transmission.
Econometric models suggest that the overall effect of a fiscal expansion in one country
is the obvious one: There is an increase in demand for the net exports of the other
country. Thus the expansion is transmitted positively. The overall transmission effect of
a monetary expansion appears to be small, however, because the income and exchange
rate effects on the trade balance tend to cancel each other out.
The fact that the world is interdependent leads to the topic of international macro-
economic policy coordination. National policy makers may be able to do better by
setting their policies cooperatively than they can in the (Nash) noncooperative equilib-
rium, where each acts independently. For example, in a worldwide recession, with each
country afraid to expand its economy on its own for fear of a deterioration of its trade
balance, there can be gains from a general agreement to expand cooperatively.

CHAPTER PROBLEMS

1. The country of Bretagne holds its real money supply, M / P, constant, but the rest of the
world undertakes a monetary expansion that drives down interest rates in Bretagne as
well. Which is greater: the stimulus to its economy from lower interest rates or the loss
of demand (net exports) when its currency appreciates against the rest of the world?
I.e., does Y rise or fall? (Hint: Consider the money market equilibrium condition.)
2. When the country of Euphoria adopts a combination of easy fiscal and tight monetary
policy, and exchange rates are flexible, is a foreign country suffering from unemploy-
ment likely to be pleased with the consequences? A foreign country suffering from
inflation? A foreign country with a large external debt denominated in the currency of
Euphoria?

Extra Credit
3. This problem concerns interdependence and the coordination of fiscal policy between
two countries: Melanzane and Rigatoni. The country of Melanzane has two target vari-
ables: the domestic price level, p, and the exchange rate, s, valued in Melanzane-per-
Rigatoni currency units (because the country wants to stabilize the two components of
the CPI: domestic prices and import prices), both in log form. These target variables
are affected both by the government spending of Melanzane, gM as a percentage of
CAVE.6607.cp25.p525-542 6/6/06 1:25 PM Page 541

Suggestions for Further Reading 541

GDP, and the government spending of its trading partner, Rigatoni, gR as a percentage
of GDP.
p 5 A 1 CgM 1 FgR
s 5 B 1 D(gR 2 gM)
a. In a standard Mundell-Fleming model, on what would the sign of D depend?
What do most multicountry econometric models say about the signs of C, D,
and F?
b. Assume that Melanzane wishes to reduce both s and p to zero. This could be the
aftermath of an increase in oil prices that has raised A and B above zero. Solve for
the optimal combination of gM and gR that would be preferred by Melanzane if it
had its first choice. Assume that the signs of C, F, and D are as in (a). What is the
sign of (gM 2 gR), that is, would Melanzane prefer that it cut spending more or that
Rigatoni cut spending more? Why? Show the optimal point for Melanzane on a
graph analogous to that in the chapter supplement.
c. If Melanzane seeks to minimize a quadratic loss function
L 5 p2 1 vs2
where v is the weight placed on the exchange rate objective, derive its reaction
function, giving gM as a function of gR. How will Melanzane react to a fiscal con-
traction by Rigatoni, if exchange rate effects are not very important (i.e., if D and
v are low), and why? If they (D and v) are high?
d. Assume that Rigatoni has a similar objective function, with its prices determined
analogously. Indicate on a graph what optimal combination of gM and gR would be
preferred by Rigatoni, and its reaction function. Describe the Nash noncoopera-
tive equilibrium. What sort of cooperative bargain would raise economic welfare
and why?

SUGGESTIONS FOR FURTHER READING

Fischer, Stanley. “International Macroeconomic Policy Coordination.” In Martin Feld-

stein, ed., International Policy Coordination (Chicago: University of Chicago Press,
1988). A relatively comprehensive yet concise survey of the literature, including
some of the skeptics.
International Monetary Fund, World Economic Outlook (Washington, DC: IMF, April
2006). Twice a year the IMF Research Department presents forecasts of the major
economies and reviews current policy issues such as current account imbalances.
Oudiz, Gilles, and Jeffrey Sachs. “Macroeconomic Policy Coordination among the
Industrial Economies,” Brookings Papers on Economic Activity, 1 (1984): 1–76.
A readable introduction to international policy coordination and the first serious
attempt to quantify the gains.
CAVE.6607.cp25.p525-542 6/6/06 1:25 PM Page 542

542 Chapter 25 ■ Interdependence and Policy Coordination

Swoboda, Alexander, and Rudiger Dornbusch. “Adjustment Policy and Monetary

Equilibrium in a Two-Country Model.” In M. Connolly and A. Swoboda, eds.,
International Trade and Money (London: George Allen and Unwin, 1973). A clear
exposition of the two-country Mundell-Fleming model, using a graph with Y and
Y* on the axes, as in our Figure 17.6.
Willett, Thomas. “Developments in the Political Economy of Policy Coordination,”
Open Economies Review, 86, no. 2 (1999): 221–253. A review of twenty-nine con-
clusions from the literature.