The “Exorbitant Privilege”: A Theoretical Exposition

Wenli Cheng
Department of Economics, Monash University
Dingsheng Zhang
China Economics and Management Academy
Central University of Finance and Economics

Abstract: This paper develops a general equilibrium model to study how the
“exorbitant advantage” works, whether it is sustainable, and what may be the
consequences if it is removed. Its main findings are: (1) the center country that issues
the reserve currency enjoys the “exorbitant advantage” in the sense that her current
account deficit can be financed by the periphery country’s reserve holdings. The
“exorbitant privilege” is predicated on the overvaluation of the reserve currency
caused by a higher rate of money growth in the center country; (2) the “exorbitant
advantage” is not likely to be sustainable in the long run; (3) if the “exorbitant
advantage” is removed, the value of the reserve currency will depreciate, the terms of
trade will change against the periphery country and sector composition will change in
favour of the tradable sector in the center country and in favour of the non-tradable
sector in periphery country. These changes will be more pronounced if the center
country repays her debt by printing money instead of raising taxes.

JEL classification: F11, F31, E34, E42

Keywords: “Exorbitant Privilege”, international currency, external imbalances,

foreign debt, internal adjustment

1
1. Introduction

The term “exorbitant privilege” was coined in 1965 by the then French Minister

Valery Giscard d’Estaing to describe the US’s unique ability to finance large external

deficits (Gourinchas and Rey, 2005). The French economist Jacques Rueff (1971)

explains the working of the exorbitant privilege as follows:

“…when a country with a key currency runs a balance-of-payment deficit - that

is to say, the United States, for example - it pays the creditor country dollars,

which end up with the latter's central bank. But the dollars are of no use in Bonn

or in Tokyo or in Paris. The very same day, they are reloaned to the New York

money market, so that they return to the place or origin. Thus the debtor country

does not lose what the creditor country has gained. So the key-currency country

never feels the effect of a deficit in its balance of payments. And the main

consequence is that there is no reason whatever for the deficit to disappear,

because it does not appear.” (p.78)

Rueff (1971) argues that the American balance-of-payment deficits (over the 1950s)

he analyses was not brought about by the United States; rather it was “the outcome of

an unbelievable collective mistake” (p.24), namely, the adoption of the Bretton

Woods System of gold-exchange standard which accepts the US dollar as a substitute

for gold as reserves. In Rueff’s words,

“By accepting this substitution, the banks of issue put themselves in a position

of being able to lend immediately to the United States, wholly or in part, the

dollars they received in settlement of American balance-of-payments

deficits. …

2
 To the extent that the new system was applied, it in fact released the United

States from the obligation of settling its foreign debts. Thus the United States

was in a position to lend, to give, and even to buy outside its own frontiers

without having to worry about its own capacity.” (p.65)

More generally, Rueff (1971) contends that the gold-exchange standard “allowed the

countries in possession of a currency benefiting from international prestige to give

without taking, to lend without borrowing and to acquire without paying.” (p.23)

Fifty years on, the Bretton Woods System that Rueff criticizes has long broken

down, but the US appears to continue to enjoy the “exorbitant privilege” as evidenced

by its apparent ease in acquiring foreign debts to finance its persistent current account

and fiscal deficits. This “exorbitant privilege”, as Rueff might argue, is attributable to

the “international prestige” that the US dollar possesses. Besides being the main

vehicle currency in foreign exchange markets, the main invoice currency for

international trade, the US dollar is the most important reserve currency and the US

Treasury securities are widely held by central banks around the world as official

exchange reserves. Given the US dollar’s dominant position and the fact that the US

dollar is no longer tied to gold, the current international monetary order is referred to

as the “dollar standard” (McKinnon, 2001). A notable feature of the dollar standard in

recent decades is that some fast growing Asian countries (including those that have

officially adopted floating exchange rate systems) choose to fix the value of their

local currencies against the US dollar (Calvo and Reinhart, 2002). The outcome of

these unilateral choices of exchange rate policies resembles the arrangement under the

Bretton Woods system of fixed exchange rates, thus Dooley, Folkerts-Landau, &

3
Garber (2004) characterise it as the “Revived Bretton Woods system”, with the US

being the “center country” and Asia in the “periphery”.

How does the “exorbitant privilege” work under the “Revived Bretton Woods

system”? Is the system sustainable? What might happen to the real economies of the

world should the centre country lose its “exorbitant privilege”? These are the

questions this paper attempts to tackle. To answer these questions, we develop a

simple two-country, three-good trade model with money. The setup of our model is

similar to that in Lucas (1982) and Stockman & Dellas (1989), but our model makes

four specific assumptions to characterise some stylised features of the “Revived

Bretton Woods system”: (1) the currency of the centre country (the “dollar”) is the

medium of exchange used in international trade and is the reserve currency of the

periphery country; (2) the periphery country’s currency is pegged to the dollar, with

the peg maintained by the periphery country’s intervention in the foreign exchange

market; (3) the periphery country’s foreign exchange reserves are invested in country

2’s government bonds with no well-defined date of net repayment; and (4) the funds

raised from the government bond issue are given to the citizens of the center country

as transfer payments.

Based on our model, we come to three main conclusions. First, the centre

country that issues the reserve currency has the “exorbitant privilege” in the following

sense: once the center country’s current account deficits translates into increases in

reserves of the periphery country, the increased reserve is reinvested in the center

country, thus the center country’s deficits are more or less automatically funded.

However, having access to automatic funding does not mean that the center country

will necessarily run a current account deficit. The presence of a current account deficit

4
is an indication that the currency of the centre country is overvalued. The

overvaluation may be caused by a higher rate of money growth in the centre country

or by a desire on the part of the periphery country to devalue its currency for

mercantilist purposes or to increase its reserves for precautionary considerations. This

finding supports Aizenman’s (2010) contention that it is a fallacy to regard global

imbalance as a “necessary requisite and consequence” (p.23) of the dollar standard. It

is also broadly consistent with the casual observation that the increase in global

imbalances during the years prior to the global financial crisis of 2008 coincided with

both an easy monetary policy environment in the US and a fast rate of official reserve

accumulation in Asia.

Second, if international trade is balanced, reserves in the periphery country do

not change, the center country does not utilise its “exorbitant privilege” although it

continues to benefit from the existing reserves invested by the periphery country. If

the policies of the center country or the periphery country or both make the dollar

overvalued, trade imbalance will occur. The imbalance cannot be sustained in the long

run because a sustained imbalance requires chronic overvaluation of the reserve

currency which undermines the standing of the dollar as a reserve currency.

Moreover, a lasting trade imbalance leads to an accumulation of debt in the center

country and an accumulation of reserves in the periphery country to increasingly

undesirable levels. It also leads to a change in sector composition in the center

country in favour of the non-tradable sector.

Third, if the “exorbitant privilege” were to be removed in the sense that the

demand for the dollar as a reserve currency falls significantly and the periphery

country withdraws (a part of) its reserves invested in the center country, the dollar

5
would fall in value, the terms of trade would change against the periphery country and

the sector composition would change in favour of the tradable sector in country 2 and

the non-tradable sector country 1. The dollar depreciation and country 1’ terms of

trade deterioration would be more drastic, but the changes in sectoral composition

would be less pronounced if the centre country chose to repay its debt by “printing

money” instead of raising taxes.

Our paper connects with three related strands of literature. The first literature

concerns the “exorbitant privilege”. Gourinchas and Rey (2005) find that the US

enjoys an “exorbitant privilege” in the sense that US’s return on its external assets

was notably higher than foreigners’ return on US liabilities for over half a century

since the early 1950s. Gourinchas, Rey and Govillot (2010) further argue that the

“exorbitant privilege” comes with an “exorbitant duty” as indicated by a substantial

worsening of US foreign asset position during the recent global financial crisis when

there was a net selling of US external assets. Gourinchas et al.’s definition of

exorbitant privilege is somewhat different from ours since they emphasise the excess

return of US external assets whereas we focus on the center country’s ability to obtain

automatic funding for its external deficits. However our analysis points to an

explanation for the excess return of US external assets, namely that, government

policies under the dollar standard can create a powerful force that draws large

amounts of official reserves from periphery countries to the US. Since safety is the

main concern for the owners of official reserves, the reserves are typically invested in

low-risk low-yield government securities. In contrast, US investments overseas

involve higher risks and enjoy higher returns.

6
 Secondly, our paper relates to the literature concerning the direction of

international capital flows. The flow of official reserves from the poor periphery to

the rich center country in our model constitutes a significant part of the “global saving

glut” which Bernanke (2005, 2007) suggests may be a contributing factor to the

global financial crisis. The direction of capital flow in our paper is also consistent

with the empirical findings documented by Lucas (1990) and Alfaro et al. (2008).

Thirdly, our paper contributes to the ongoing debate on the causes and

implications of the global imbalances. On the optimistic side of the debate, it is

argued that the imbalances may be an equilibrium outcome resulting from the fact that

different regions of the world differ in their capacity to generate financial assets from

real investments (Caballero, Farhi, and Gourinchas, 2008). It is further suggested that

the pattern of global capital flow associated with the imbalances may be “optimal”

from the emerging economy’s perspective of developing a competitive tradable goods

sector (Dooley and Garber, 2007). On the pessimistic side, there are concerns that the

global imbalances cannot be sustained and that its eventual correction may trigger a

dollar crisis and impose substantial adjustment costs on the world economy. For

instance, Blanchard, Giavazzi, & Sa (2005) argue that the increases in US demand for

foreign goods and in foreign demand for US assets are behind the US current account

deficits. They predict that the current account deficit must be reversed in the future

accompanied by a substantial depreciation of the US dollar. Similarly, Obstfeld &

Rogoff (2005) present a model of an adjustment process through which global balance

is returned by a change of demand for tradable versus non-tradable and domestic

versus foreign goods. They also predict a large depreciation of the US dollar. The

conclusions of our paper are closer to the pessimistic view of the debate. However,

7
our focus is on the role of the US dollar as the reserve currency and the ramifications

of government policies in both the center country and the periphery for the global

imbalances.

2. A Model of the “Exorbitant Privilege”

Consider a world with two countries, country 1 (the periphery country) with a

population of N1 and country 2 (the centre country) with a population of N2.

Individuals in both countries derive utility from consuming three goods X, Y and Z

over a time horizon T. Good X is produced in country 1, good Y is produced in

country 2 and good Z is a non-tradable good produced in both countries. Domestic

trade is mediated by each country’s domestic currency, the “yuan” in country 1 and

the “dollar” in country 2; and international trade is meditated by the dollar. As a result,

consumers in country 1 have to purchase “dollars” to pay for imported good Y, so

their budget constraint has a domestic currency component and a foreign exchange

component. In contrast, consumers in country 2 only use dollars. The yuan-dollar

exchange rate is fixed at e and country 1’s central bank maintains this rate by buying

or selling dollars in the foreign exchange market if necessary.

We assume that all of country 1’s foreign exchange reserves are in dollars and

are invested in one-period bonds issued by the government of country 2 (at an

exogenously determined interest rate i). Bond repayments are financed by new bond

issues. The funds raised by issuing bonds are distributed to country 2’s consumers in

transfer payments.

The decision problem of a representative consumer in country 1 is:

max ∑ β (α1 ln x1t + α 2 ln y1t + α 3 ln z1t ) ( α1 + α 2 + α 3 =

t −1
1 ), (1)
{ x1t , y1t , z1t , FX It }Tt =1 t =1

8
subject to the budget constraint:

domestic currency: pxt x1t + pz1t z1t =

w1t − eFX It ,

foreign exchange: p yt y1t = FX It .

The decision problem for a representative consumer in country 2 is:

max ∑ β (α1 ln x2t + α 2 ln y2t + α 3 ln z2t ) ( α1 + α 2 + α 3 =

t −1
1 ), (2)
{ x2 t , y2 t , z2 t }Tt =1 t =1

subject to the budget constraint:

pxt FX Gt
x2t + p yt y2t + pz2 t z2t =w2t + .
e N2

In the above decision problems, subscript t denotes time t and subscript i (i=1, 2)

denotes country i; xit , yit , zit are quantities of goods X, Y and Z consumed;

pxt , p yt , pzi t are money prices of goods X, Y and Z denominated in the currency of the

country of production; wit is the wage rate; FX It is the foreign exchange (dollars)

purchased in the foreign exchange market to pay for imports (good Y); e is the fixed

exchange rate; FX Gt is the amount of dollars purchased by country 1’s central bank in

order to maintain the exchange rate at e . Since FX Gt represents an increase in

reserves and the reserves are invested in country 2’s government bond, FX Gt / N 2 is

the per person transfer payment from the country 2’s government funded by its bond

issue.

On the supply side, producers in country 1 produce goods X and Z using the

following technologies:

X 1t = a1x L1xt , Z1t = a1z L1zt , (3)

where L1xt and L1zt are labor devoted to producing goods X and Z in country 1.

9
 Under the assumption of perfect competition, the following conditions hold:

π z t = pz t Z1t − w1t L1zt = 0 ,

1 1
(4)

pxt X 1t + pz1t Z1t − w1t ( L1xt + L1zt ) =

FX Gt e (5)

Equation (4) is a standard zero-profit condition. Equation (5) stipulates that the

difference between country 1’s total value of production and its total value of

consumption is the value of trade surplus, which is equal to the amount of the foreign

exchange purchase by the central bank. If trade is balanced, FX Gt is zero. If country 1

has a trade surplus, country 1’s central bank buys FX Gt dollars (with FX Gt e yuan),

invests the same amount in country 2’s government bonds, enabling country 2 to

finance its trade deficit.

In country 2, producers make goods Y and Z using the following technologies:

Y2t = a2 y L2 yt , Z 2t = a2 z L2 zt . (6)

Perfect competition implies:

π yt = p yt Z 2t − w2t L2 yt =0 , π z t = pz t Z 2t − w2t L2 zt = 0 .
2 2
(7)

The market clearing conditions in the economy are summarised below.

(1) The markets for good Z:

N1 z1t = Z1t , N 2 z2 t = Z 2 t . (8)

(2) The market for good X:

N1 x1t + N 2 x2t =X 1t =a1x L1xt . (9)

(3) The market for good Y:

N1 y1t + N 2 y2t =
Y2t =
a2 y L2 yt . (10)

(4) Labor markets:

L1xt + L1zt =
N1 , L2 yt + L2 zt =
N2 . (11)
10
(5) The foreign exchange market:

pxt
N1 FX It + FX Gt = N 2 x2t . (12)
e

The left-hand side of equation (12) is the sum of private demand for dollars (to pay

for imports) and central bank’s purchase of dollars; the right hand side is country 1’s

dollar earnings from exports.

(6) Money markets:

Strictly speaking, there are no separate markets for money – money, as the medium of

exchange, features in all markets. In country 1, money supply is affected by

government intervention in the foreign exchange market, but the government can still

choose a rate of money growth π 1 (although this will affect the amount of foreign

exchange intervention required to maintain a given exchange rate). The time path of

money supply for country 1 is described by:

(1 + π 1 )m1(t −1) + FX Gt e ,
m1t = (13)

where π 1t is a measure of money growth rate (on the basis of last period’s money

supply, excluding growth due to foreign exchange market intervention); FX Gt e is the

increase in the quantity of money due to foreign exchange market intervention,

assuming no sterilisation.

After a foreign exchange market intervention, country 1’s reserve holdings are:

R=
t FX Gt + (1 + i ) Rt −1 , (14)

where Rt −1 is the reserve holdings in the previous period, and i is the (exogenous)

interest rate paid on country 2’s government bonds.

The total money supply is equal to the total wage income in each country:

m1t = N1w1t , (1 + π 2 )m2(t −1) =

m2t = N 2 w2t , (15)
11
where π 2t is the rate of money growth in country 2.

Solving the consumers’ decision problems (1) and (2), and applying zero

profit conditions and market clearing conditions, we obtain the solutions for all

endogenous variables in the model, which are presented in Table 1. 1 Based on the

model outlined above and the solutions presented in Table 1, we now attempt to

answer the questions set out in the introduction section, namely, how the “exorbitant

privilege” works, whether it is sustainable, and what would likely happen to the

economies of both the center and the periphery if it were removed.

3. Analysis

3.1. How does the “exorbitant privilege” work?

In our model, the “exorbitant privilege” is characterised as the central country’s

ability to obtain automatic funding for its trade deficit because the dollar is the

periphery country’s reserve currency and the reserves are invested in the center

country’s government bonds. By this characterisation, the “exorbitant privilege” is

only utilised when country 2 runs a trade deficit which is funded by country 1’s

increase in reserves, that is, when FX Gt > 0 . From Table 1, we have:

α1 (1 + π 2 )m2(t −1) e − α 2 (1 + π 1 )m1(t −1)

FX Gt = , (16)
(1 − α1 + α 2 )e

which implies,

α 2 (1 + π 1 )m1(t −1)
FX Gt ≥ 0 ⇔ e ≥ . (17)
α1 (1 + π 2 )m2(t −1)

1
These solutions describe a temporary “state of rest” in the sense that consumers behave optimally and
markets clear; they are not real equilibrium solutions as external trade is not balanced. There is a
market tendency to regain balance as explained in section 3.

12
Since the fixed exchange rate ( e ) coincides with the free market exchange rate when

FX Gt = 0 , equation (17) suggests that the “exorbitant privilege” is utilised when e is

higher than the free market rate, or in other words, when the dollar is overvalued.

Suppose e is equal to the free market rate at time t-1, an overvaluation of the dollar

may occur at time t if country 1 devalues its currency ( e becomes larger), and/or if

the rate of money growth in country 2 is higher than that in country 1 ( π 2 > π 1 ).

Once the exorbitant privilege is utilised at time t (i.e., FX Gt > 0 ), the money

supply in country 1 increases by FX Gt e (assuming no sterilisation). Other things

equal, the increase in country 1’s money supply lowers the market value of country

1’s currency relative to the dollar, thereby driving the fixed exchange rate towards the

free market rate in the next period. Therefore, in the absence of continued devaluation

by country 1 or excess money growth in country 2, there is a tendency for external

imbalances to correct themselves. This suggests that if external imbalances persist, it

is likely that the policies in the periphery country or the centre country or both have

kept the dollar over-valued.

3.2. Is the “exorbitant privilege” sustainable?

While government policies in the center country and/or in the periphery can

lead to persistent external imbalances, it does not mean that “exorbitant privilege” is

sustainable. A necessary condition for the “exorbitant privilege” is that the periphery

country is willing to use the dollar as its reserve currency. An important reason why

the dollar is chosen as the reserve currency is that it is perceived to be a stable

currency based on past experiences. If, as shown above, the utilisation of the

“exorbitant privilege” requires a continued over-valuation of the dollar, then the

13
process of utilising the “exorbitant privilege” weakens the very foundation which

supports the dollar as a reserve currency, therefore the system cannot be sustainable in

the long run.

The use of the “exorbitant privilege” also has other consequences that lower

the sustainability of the system. From the periphery country’s perspective, the

“exorbitant advantage” leads to a continued increase in its foreign exchange reserves

which over time tend to go above its desired level. As the increase in reserves is lent

back to the central country on an indefinite basis, it leads to the peculiar direction of

long-term capital flow from the relatively poor periphery country to the relatively rich

centre country, and since the capital flow is the outcome of a net flow of goods in the

opposite direction, it has the effect of transferring real resources from the periphery to

the centre.

While the centre country benefits from the real resource transfer in the short

run, there are long run costs. First of all, the flip side of the increase in reserves by the

periphery country is the centre country’s accumulation of debt. Although the debt is

denominated in dollars and has no well-defined time frame for net repayment,

sovereign debt problems can escalate and result in financial crises as the history of the

last eight centuries amply testifies (Reinhart and Rogoff, 2009). Secondly, the use of

the “exorbitant privilege” also leads to a change in sectoral competition in country 2.

From Table 1, we have:

L1xt α1 + α 2
= (18)
L1zt α3

L2 yt (2 − α 3 )α 2 e (1 + π 2 )m2(t −1) + α 2α 3 (1 + π 1 )m1(t −1)

= (19)
L2 zt α 3 (1 + α 2 )e (1 + π 2 )m2(t −1) − α 2α 3 (1 + π 1 )m1(t −1)

It is easy to derive from equations (18) and (19) that

14
L2 yt L1xt α (1 + π 1 )m1t
< ⇔e> 2 , (20)
L2 zt L1zt α1 (1 + π 2 )m2t

which implies that if the “exorbitant privilege” is at work, then country 2 has a lower

proportion of resources devoted to the tradable sector relative to country 1. In other

words, the use of the “exorbitant privilege” results in a change in the sectoral

composition in the centre country in favour of the non-tradable sector at the expense

of the tradeable sector. This is politically undesirable as it raises concerns about the

“hollowing out” of the manufacturing sector in the centre country. The sectoral

composition bias is also economically inefficient as it is driven by debts other than

comparative advantage, and entails future adjustment costs to the extent that the bias

has to be corrected when debts are eventually paid.

3.3. What would happen if the “exorbitant privilege” were removed?

If, as argued above, the “exorbitant privilege” cannot be sustained in the long run, the

logical next question is what would happen if it were removed. By the removal of the

“exorbitant privilege”, we mean that the standing of the dollar as a reserve currency is

in question and the periphery country not only stops accumulating dollar reserves, but

also gradually reduces its dollar reserves over time. As the dollar reserves are held in

the form of the centre country’s government bonds, a reduction of reserves amounts to

demanding the centre country’s government to repay its debt. The centre country’s

government may raise taxes to repay the debt (case 1), or it may simply print money

to meet its debt obligations since its debt is denominated in dollars (case 2). In the

following we present two simple models to illustrate what would likely happen to the

economies of the centre and the periphery countries in these two cases.

15
 For simplicity of illustration, our models assume: (1) the government of the

periphery country an equal amount of principal in each period; and (2) the debt

repayments are distributed equally to citizens of the periphery country in transfer

payments.

Consider case 1 where debt repayments by country 2 are financed by taxes.

The model for case 1 is similar to that presented in section 2 except for 3 differences.

First, an individual consumer in country 1 now receives a transfer payment

which changes the foreign exchange component of the budget constraint. The decision

problem of the representative consumer is:

max ∑ β (α1 ln x1t + α 2 ln y1t + α 3 ln z1t ) , ( α1 + α 2 + α 3 =

t −1
1 ), (21)
{ x1t , y1t , z1t , FX It }Tt =1 t =1

subject to the budget constraint:

domestic currency: pxt x1t + pz1t z1t =

w1t − eFX It ,

∆Rt
foreign exchange: p yt=
y1t FX It + ,
N1

where ∆Rt / N1 is the per capita transfer payment.

Second, an individual consumer in country 2 now has to pay a tax to fund the

debt repayment. The decision problem of the representative consumer is:

max ∑ β (α1 ln x2t + α 2 ln y2t + α 3 ln z2t ) , ( α1 + α 2 + α 3 =

t −1
1 ), (22)
{ x2 t , y2 t , z2 t }Tt =1 t =1

subject to the budget constraint:

pxt ∆R
x2t + p yt y2t + pz2 t z2t =w2t − t ,
e N2

where ∆Rt / N 2 is the per capita tax imposed for the purpose of debt repayment.

16
 Third, the periphery country no longer intervenes in the foreign exchange

market. As a result, the clearance condition for the foreign exchange market becomes:

pxt
N1 FX It = N 2 x2t , (23)
et

which determines the free market exchange rate et .

Correspondingly, country 1’s money supply equation becomes:

m1=
t (1 + π 1 )m1(t −1) (24)

Now consider case 2 where country 2’s government simply discharges its debt

obligations with “printed” dollars. The model for case 2 differs from that for case 1 in

two aspects. First, consumers in country 2 are no longer required to pay tax to fund

the debt repayments, thus the decision problem for the representative consumer in

country 2 is:

max ∑ β (α1 ln x2t + α 2 ln y2t + α 3 ln z2t ) ( α1 + α 2 + α 3 =

t −1
1 ), (25)
{ x2 t , y2 t , z2 t }Tt =1 t =1

subject to budget constraint:

pxt
x2t + p yt y2t + pz2 t z2t =
w2t
e

Secondly, the dollar “printing” changes country 2’s money supply, which becomes:

m2t = (1 + π 2 )m2(t −1) + ∆Rt . (26)

We solve for the equilibrium for each of the models outlined above and

present the solutions in Table 2. Based on the results in Table 2, we investigate how

the economies would respond to the removal of the “exorbitant privilege”, focusing

on the exchange rate, the terms of trade, and sectoral composition in each country.

Intuitively, the removal of the “exorbitant privilege” means that the demand

for the dollar would fall, therefore the value of the dollar would fall as well. The value
17
of the dollar would fall by more if country 2 also increases its supply of dollar to

discharge its debts. As shown in Table 2, the exchange rates under the two cases are:

α 2 (1 + π 1 )m1(t −1)
Case 1: (et )1 = , (27)
α1[(1 + π 2 )m2(t −1) ] + α 3∆Rt

α 2 (1 + π 1 )m1(t −1)
Case 2: (et ) 2 = . (28)
α1[(1 + π 2 )m2(t −1) + ∆Rt ] + α 3∆Rt

From equation (17), we know that for the “exorbitant privilege” to work, the dollar

needs to be overvalued, that is,

α 2 (1 + π 1 )m1(t −1)
e> . (29)
α1 (1 + π 2 )m2(t −1)

It is easy to show that

e > (et )1 > (et ) 2 . (30)

Inequality (30) implies that if the “exorbitant privilege” were removed, the dollar

would depreciate in value; and that the dollar would depreciate more in the case 2

(where country 2 resorts to money printing to meet its debt obligations) than in case 1

(where debt repayment is financed by tax revenue).

Now consider the terms of trade. From Table 1, we have the prices for tradable

goods in the model of the “exorbitant privilege”:

α1 (1 − α1 − α 2 )(1 + π 1 )m1(t −1) + α1 (1 + α1 + α 2 )(1 + π 2 )m2(t −1) e

( pxt )ep = , (31)
(1 − α1 + α 2 )(α1 + α 2 )a1x N1

(1 + π 2 )m2(t −1)
( p yt )ep = , (32)
a2 y N 2

where subscript ep indicates the “exorbitant privilege”.

From Table 2, we have the prices for tradable goods in the two models with the

removal of the “exorbitant privilege”:

18
 (1 + π 1 )m1(t −1) (1 + π 2 )m2(t −1)
Case 1: ( pxt )1 = , ( p yt )1 = , (33)
a1x N1 a2 y N 2

(1 + π 1 )m1(t −1) (1 + π 2 )m2(t −1) ∆Rt

Case 2: ( pxt ) 2 = ,=( p yt ) 2 + , (34)
a1x N1 a2 y N 2 a2 y (1 − α 3 ) N 2

where subscripts 1 and 2 indicate case 1 and case 2, respectively.

From equations (31)-(34), it is easy to show that

pxt p α 2 (1 + π 1 )m1(t −1)

( )ep > ( xt )1 if e > , (35)
p yt p yt α1 (1 + π 2 )m2(t −1)

pxt p
( )1 > ( xt ) 2 . (36)
p yt p yt

Equations (35) and (36) suggest that the removal of the “exorbitant privilege” would

leads to a fall in country 1’s terms of trade; and the fall would be more pronounced in

case 2 than in case 1. The fall in country 1’ terms of trade is driven by a change in

relative demand for goods X and Y due to income changes in country 1 and country 2.

The relative price change is larger in case 2 because the increase in money supply in

country 2 further drives up the price of good Y.

Finally we consider how each country’s sector composition (as described by

the ratio of labor employed in the tradable sector relative to that in the non-tradable

sector) would respond to a removal of the “exorbitant advantage”. From Table 1, we

have each country’s sector composition in the model of the “exorbitant privilege”:

L1xt α + α2
Country 1: ( )ep = 1 , (37)
L1zt α3

L2 yt (2 − α 3 )α 2 e (1 + π 2 )m2(t −1) + α 2α 3 (1 + π 1 )m1(t −1)

Country 2: ( )ep = . (38)
L2 zt α 3 (1 + α 2 )e (1 + π 2 )m2(t −1) − α 2α 3 (1 + π 1 )m1(t −1)

19
From Table 2, we have each country’s sector composition with the removal of the

“exorbitant privilege” in case 1 and case 2:

α 2 m2t
(α1 + α 2 )
L1xt α 2 m2t + α 3 ∆Rt
Case 1: ( )1 = (39)
L1zt α 2 m2t
(1 − α1 − α 2 )
α 2 m2t + α 3∆Rt

∆Rt
(α1 + α 2 + α 3
)
L2 yt m2t
( )1 = (40)
L2 zt ∆R
(1 − α1 − α 2 − α 3 t )
m2t

α 2 ∆Rt
(1 − α 3 − α 3 )
L1xt α1m2t + α 3 ∆Rt
Case 2: ( )2 = (41)
L1zt α 2 ∆Rt
(α 3 + α 3 )
α1m2t + α 3∆Rt

L2 yt 1 − α3
( )2 = (42)
L2 zt α3

From equations (37)-(42), we derive

L1xt L L
( )ep > ( 1xt ) 2 > ( 1xt )1 (43)
L1zt L1zt L1zt

L2 yt L2 yt L2 yt α 2 (1 + π 1 )m1(t −1)
( )ep < ( )2 < ( )1 if e > (44)
L2 zt L2 zt L2 zt α1 (1 + π 2 )m2(t −1)

Inequalities (43) and (44) suggest that with the removal of the “exorbitant privilege”,

labour resources would move away from the tradable sector (producing good X) in

country 1 and move towards the tradable sector (producing good Y) in country 2. The

changes in sectoral composition are driven by changes in relative demand for good X

and good Y. The extent of change in sector composition would be smaller in case 2

than in case 1 because in case 2, debt repayment financed by dollar “printing” drives

20
up of the price of good Y, therefore the relative demand increase for good Y is

smaller.

4. Concluding remarks

In this paper, we have presented a set of three simple models to study how the

“exorbitant advantage” works, whether it is sustainable, and what would be the likely

consequences if it were removed. We have shown that the presence of the “exorbitant

privilege” is an indication of the over-valuation of the dollar which may be caused by

policies in the periphery and/or the center country. We have argued that the

“exorbitant privilege” is unsustainable in the long run, and that its removal would

likely lead to a dollar depreciation, a change in the terms of trade against the

periphery, and a change in sector composition in favour of the tradable sector in the

center country and in favour of the non-tradable sector in the periphery.

Our model has some simplifying assumptions which we hope to relax in future

research. Although relaxing these assumptions does not qualitatively change the main

conclusions of the paper, it will give us a better representation of the economic

phenomena we want to study so that we can gain a deeper understanding of them. For

example, in this paper we assume no private lending or borrowing. This greatly

simplifies the model as the consumers’ decisions become time independent. With the

inclusion of private lending and borrowing in the model, the extent of the “exorbitant

privilege” may be greater because the capital flowing into the center country can be

used to finance both the center country’s current account deficit and its external asset

acquisition, enabling the center country to become the world’s “venture capitalist”

(Gourinchas and Rey, 2005).

21
 Another simplifying assumption in this paper is that there is no private

demand for money therefore money does not enter the consumers’ utility function.

Under the dollar standard, there is a demand for dollars in the periphery but virtually

no demand for the periphery countries’ currencies in the center country. If this

asymmetry in money demand is included in the model, a higher monetary growth rate

in the center country (which results in the utilisation of the “exorbitant privilege”) can

lead to a greater resource transfer from the periphery to the center country. This is

because higher money growth leads to higher prices which increase the transaction

demand for dollars in the periphery, and this increase in demand for dollar holdings

needs to be met by additional exports earnings (Cheng and Zhang, forthcoming).

Finally, we assume no sterilisation following the periphery country’s purchase

of dollars in the foreign exchange market. With the introduction of a partial (but not

complete) sterilisation, it will slow down but not stop the process through which the

fixed exchange rate approaches the free market level, consequently the exorbitant

privilege can persist for longer but not forever in the absence of continued policy

interventions.

References

Aizenman, J. (2010). On the causes of global imbalances and their persistence: myths,
facts and conjectures. In S. J. E. a. B. H. Stijn Claessens (Ed.), Rebalancing
the global economy: a primer for policymaking: CEPR London.
Alfaro, L., Kalemli-Ozcan, S., and Volosovych, V. (2008). Why doesn't capital flow
from rich to poor countries? An empirical investigation. The Review of
Economics and Statistics 90(2).
Bernanke, B. S. (2005). The Global Saving Glut and the U.S. Current Account Deficit
speech delivered for the Sandridge Lecture at the Virginia Association of
Economists, Richmond, March 10,
www.federalreserve.gov/boarddocs/speeches/2005/200503102/default.htm.

22
Bernanke, B. S. (2007). Global Imbalances: Recent Developments and Prospects.
Speech delivered at the Bundesbank Lecture, Berlin, Germany, September 11.
http://www.federalreserve.gov/newsevents/speech/bernanke20070911a.htm.
Blanchard, O., Giavazzi, F., and Sa, F. (2005). International Investors, the U.S.
Current Account, and the Dollar. Brookings Papers on Economic Activity(1),
1-49.
Caballero, R. J., Farhi, E., and Gourinchas, P.-O. (2008). An Equilibrium Model of
'Global Imbalances' and Low Interest Rates. American Economic Review,
98(1), 358-393.
Calvo, G. A., and Reinhart, C. M. (2002). Fear of Floating. Quarterly Journal of
Economics, 117(2), 379-408.
Cheng, W. and Zhang, D. (forthcoming). International Transmission of Monetary
Shocks and the Non-neutrality of International Money, Review of
International Economics.
Dooley, M., Folkerts-Landau, and Garber, P. (2004). The Revived Bretton Woods
System. International Journal of Finance and Economics, 9(4), 307-313.
Dooley, M., and Garber, P. (2007). Is It 1958 or 1968? Three Notes on the Longevity
of the Revived Bretton Woods System. In R. H. Clarida (Ed.), G-7 Current
Account Imbalances: Sustainability and Adjustment (pp. 103-132). Chicago:
University of Chicago Press.
Gourinchas, P.-O., and Rey, H. (2005). From World Banker to World Venture
Capitalist: US External Adjustment and the Exobitant Privilege. NBER
working paper 11563.
Gourinchas, P.-O., Rey, H., and Govillot, N. (2010). Exorbitant privilege and
exorbitant duty. IMES Discussion Paper Series 10-E-20, Bank of Japan.
Huang, X.D., and Suchada, T. (2003). Sources of Exchange Rate Fluctuations: The
Cases of Mexico and Thailand in the Aftermaths of their Recent Currency
Crises. Annals of Economics and Finance 4, 375-400.
Lucas, R. E. (1982). Interest rates and currency prices in a two-country world. Journal
of Monetary Economics, 10(3), 335-359.
Lucas, R. E. (1990). Why doesn't capital flow from rich to poor countries? American
Economic Review 80, 92-106.
McKinnon, R. I. (2001). The International Dollar Standard and the Sustainability of
the U.S. Current Account Deficit. Brookings Papers on Economic Activity,
2001(1), 227-239.
Obstfeld, M., and Rogoff, K. S. (2005). Global Current Account Imbalances and
Exchange Rate Adjustments. Brookings Papers on Economic Activity(1), 67-
123.
Reinhart, C. M., and Rogoff, K. S. (2009). This time is different : eight centuries of
financial folly. Princeton: Princeton University Press.
Rueff, J. (1971). The Monetary Sin of the West. New York: MacMillan.
Stockman, A., and Dellas, H. (1989). International Portfolio Nondiversification and
Exchange Rate Variability. Journal of International Economics, 26, 271-289.

23
 Table 1. Solutions for the model of “Exorbitant Privilege”

Equilibrium solutions
Purchase of “dollars” α1 (1 + π 2 )m2(t −1) e − α 2 (1 + π 1 )m1(t −1)
by country 1’s central FX Gt =
bank (1 − α1 + α 2 )e
(1 − α1 )(1 + π 1 )m1(t −1) + α1 (1 + π 2 )m2(t −1) e
w1t =
(1 − α1 + α 2 ) N1
Wages
(1 + π 2 )m2(t −1)
w2t =
N2
α1 (1 − α1 − α 2 )(1 + π 1 )m1(t −1) + α1 (1 + α1 + α 2 )(1 + π 2 )m2(t −1) e
pxt = ,
(1 − α1 + α 2 )(α1 + α 2 )a1x N1

w1t (1 − α1 )(1 + π 1 )m1(t −1) + α1 (1 + π 2 )m2(t −1) e

p= =
(1 − α1 + α 2 )a1z N1
z1t
a1z
Prices
w2t (1 + π 2 )m2(t −1)
=
p yt =
a2 y a2 y N 2

w2t (1 + π 2 )m2(t −1)

p=
z2 t =
a2 z a2 z N 2

(1 − α1 )(1 + π 1 )m1(t −1) + α1 (1 + π 2 )m2(t −1) e

x1t a1x (α1 + α 2 )
=
(1 − α1 − α 2 )(1 + π 1 )m1(t −1) + (1 + α1 + α 2 )(1 + π 2 )m2(t −1) e
N 2 (1 − α1 )(1 + π 1 )m1(t −1) + α1 (1 + π 2 )m2(t −1) e
y1t = α 2 a2 y
N1 (1 − α1 + α 2 )(1 + π 2 )m2(t −1) e
z1t = α 3 a1z
Consumption N (1 + α 2 )(1 + π 2 )m2(t −1) e − α 2 (1 + π 1 )m1(t −1)
x= (α1 + α 2 )a1x 1
N 2 (1 − α1 − α 2 )(1 + π 1 )m1(t −1) + (1 + α1 + α 2 )(1 + π 2 )m2(t −1) e
2t

(1 + α 2 )(1 + π 2 )m2(t −1) e − α 2 (1 + π 1 )m1(t −1)

y2t = α 2 a2 y ,
(1 − α1 + α 2 )(1 + π 2 )m2(t −1) e
(1 + α 2 )(1 + π 2 )m2(t −1) e − α 2 (1 + π 1 )m1(t −1)
z2t = α 3 a2 z
(1 − α1 + α 2 )(1 + π 2 )m2(t −1) e
L1xt α1 + α 2
=
Labor ratio: tradable vs. L1zt α3
non-tradable
L2 yt (2 − α 3 )α 2 e (1 + π 2 )m2(t −1) + α 2α 3 (1 + π 1 )m1(t −1)
=
L2 zt α 3 (1 + α 2 )e (1 + π 2 )m2(t −1) − α 2α 3 (1 + π 1 )m1(t −1)

24
Table 2. Equilibrium Solutions for the models of debt repayment

Case 1 Case 2
α 2 (1 + π 1 )m1(t −1) α 2 (1 + π 1 )m1(t −1)
Exchange rate et = et =
α1[(1 + π 2 )m2(t −1) ] + α 3∆Rt α1[(1 + π 2 )m2(t −1) + ∆Rt ] + α 3∆Rt
(1 + π 1 )m1(t −1) (1 + π 1 )m1(t −1)
w1t = w1t =
N1 N1
Wages
(1 + π 2 )m2(t −1) (1 + π 2 )m2(t −1)
w2t = w2t =
N2 N2
(1 + π 1 )m1(t −1) (1 + π 1 )m1(t −1)
pxt = pxt =
a1x N1 a1x N1
(1 + π 1 )m1(t −1) (1 + π 1 )m1(t −1)
pz1t = pz1t =
a1z N1 a1z N1
Prices
(1 + π 2 )m2(t −1) (1 + π 2 )m2(t −1) ∆Rt
p yt = =p yt +
a2 y N 2 a2 y N 2 a2 y (1 − α 3 ) N 2
(1 + π 2 )m2(t −1) (1 + π 2 )m2(t −1)
p z2 t = p z2 t =
a2 z N 2 a2 z N 2
α 2 ∆Rt
=x1t a1xα1 (1 + ) α 2 ∆Rt
α1[(1 + π 2 )m2(t −1) ] + α 3∆Rt =x1t a1xα1{1 + }
α1[(1 + π 2 )m2(t −1) + ∆Rt ] + α 3∆Rt
N 2 α1 (1 + π 2 )m2(t −1) + (α 2 + α 3 )∆Rt
y1t = a2 y
( ) N α1 (1 + π 2 )m2(t −1) + ∆Rt
N1 (1 + π 2 )m2(t −1) y1t= (1 − α 3 )a2 y 2 [ ]
N1 (1 − α 3 )(1 + π 2 )m2(t −1) + ∆Rt
α 2 ∆Rt
=z1t α 3a1z {1 + } α 2 ∆Rt
α1[(1 + π 2 )m2(t −1) ] + α 3∆Rt = z1t α 3a1z {1 + }
Consumption α1[(1 + π 2 )m2(t −1) + ∆Rt ] + α 3∆Rt
N [(1 + π 2 )m2(t −1) ] − ∆Rt
x2t = α1α 2 a1x 1 N [(1 + π 2 )m2(t −1) + ∆Rt ] − ∆Rt
N 2 α1[(1 + π 2 )m2(t −1) ] + α 3∆Rt x2t = α1α 2 a1x 1
N 2 α1[(1 + π 2 )m2(t −1) + ∆Rt ] + α 3∆Rt
∆Rt
= y2t α 2 a2 y {1 − } (1 + π 2 )m2(t −1)
[(1 + π 2 )m2(t −1) ] y2t α 2 (1 − α 3 )a2 y [
= ]
(1 − α 3 )(1 + π 2 )m2(t −1) + ∆Rt
∆Rt
= z2t α 3a2 z {1 − } z2t = α 3 a2 z
[(1 + π 2 )m2(t −1) ]
α 2 [(1 + π 2 )m2(t −1) ]
α1 + α 2 α 2 ∆Rt
L1xt α 2 [(1 + π 2 )m2(t −1) ] + α 3∆Rt 1 − α3 − α3
= L1xt α1[(1 + π 2 )m2(t −1) + ∆Rt ] + α 3∆Rt
L1zt α 2 [(1 + π 2 )m2(t −1) ] =
Labor ratio:
1 − α1 − α 2 L1zt α 2 ∆Rt
tradable α 2 [(1 + π 2 )m2(t −1) ] + α 3∆Rt α3 + α3
vs. non- α1[(1 + π 2 )m2(t −1) + ∆Rt ] + α 3∆Rt
tradable ∆Rt
α1 + α 2 + α 3
L2 yt [(1 + π 2 )m2(t −1) ]
= L2 yt 1 − α3
L2 zt ∆Rt =
1 − α1 − α 2 − α 3 L2 zt α3
[(1 + π 2 )m2(t −1) ]

25