Swoboda 1973

Download as pdf or txt
Download as pdf or txt
You are on page 1of 20

London School of Economics

The Suntory and Toyota International Centres for Economics and Related Disciplines

Monetary Policy under Fixed Exchange Rates: Effectiveness, the Speed of Adjustment and
Proper Use
Author(s): Alexander K. Swoboda
Source: Economica, New Series, Vol. 40, No. 158 (May, 1973), pp. 136-154
Published by: Wiley on behalf of The London School of Economics and Political Science and
The Suntory and Toyota International Centres for Economics and Related Disciplines
Stable URL: http://www.jstor.org/stable/2551776
Accessed: 27-06-2016 07:08 UTC

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
http://about.jstor.org/terms

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted
digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about
JSTOR, please contact support@jstor.org.

London School of Economics, The Suntory and Toyota International Centres for
Economics and Related Disciplines, The London School of Economics and Political
Science, Wiley are collaborating with JSTOR to digitize, preserve and extend access to Economica

This content downloaded from 128.42.202.150 on Mon, 27 Jun 2016 07:08:44 UTC
All use subject to http://about.jstor.org/terms
[MAY

Monetary Policy under Fixed Exchange Rates:


Effectiveness, the Speed of Adjustment
and Proper Use'

By ALEXANDER K. SWOBODA

Some, though far from total, agreement has begun to emerge as to


the role and effects of monetary policy in a closed economy. At least
major issues have been delineated and the battle joined in terms of fairly
well-defined analytical frameworks. The impact of changes in the stock
of money (or its rate of change) on prices, output and interest rates has
been discussed at the theoretical level and investigated empirically.
Much dispute remains as to the lag-structure of response to monetary
disturbances, as to the division into output and price effects, and as to
proper monetary targets and policy indicators. Nevertheless, most
economists would agree that monetary policy can be used as a counter-
cyclical device, and that the stock of money (or its rate of growth) can,
in some average sense over the medium run, be controlled, however
difficult it may be to exercise such control in the very short run and
however poorly monetary authorities have actually performed in this
respect.
Discussion of monetary policy in the open economy, on the other
hand, has proceeded at a higher level of abstraction (or over-simplifica-
tion) and empirical work has remained scarce. The reason is close at
hand: with some notable exceptions, recent developments in monetary
theory and policy analysis have been, largely, the work of economists
based or trained in the United States; and, from an American vantage
point (especially a Middle-Western one and before the so-called "dollar"
crises), what more natural simplifying assumption than that of the
closed economy? Yet, in recent years, under the pressure of events and
following the rediscovery of Hume and Ricardo and the work of,
among others, Meade, Alexander, Polak, Prais, Tsiang, Johnson and
Mundell, the analysis of monetary policy in the open economy has made
much progress, at least on a theoretical plane. The focus and conclu-
sions of that work, especially that of Mundell, have been rather differ-
ent from those of analyses dealing with the closed economy: the
monetary balance-of-payments adjustment mechanism and the role of
capital mobility and of the size of countries are emphasized; severe
' Preliminary versions of this paper were presented at the Second Konstanz
Seminar on Monetary Theory and Policy (June, 1971) and at the February 1972
Money Study Group Conference held at Bournemouth. I am indebted to Leonall
Anderson and Karl Brunner for their incisive discussion at Konstanz and to
Harry G. Johnson for helpful comments.
136

This content downloaded from 128.42.202.150 on Mon, 27 Jun 2016 07:08:44 UTC
All use subject to http://about.jstor.org/terms
1973] MONETARY POLICY UNDER FIXED EXCHANGE RATES 137

limits to the use (and controllability) of the money stock as a counter-


cyclical device are found and one asks not only whether monetary policy
can, but also whether it should, be used for anti-cyclical purposes.
This paper attempts a brief summary of the analytical conclusions
reached as to the effectiveness and proper use of monetary policy in an
open economy under fixed exchange rates. Its usefulness, if any,
should lie in clarification of some implicit assumptions and conclusions
that have perhaps received insufficient attention in the literature. Much
confusion and controversy can be avoided by precise specification of
definitions and assumptions on the one hand, and by explicit delineation
of the exact aims and limits of a particular piece of analysis or conclusion,
on the other hand.
The first section below discusses the proposition that the money
supply is an endogenous variable in an open economy under fixed
exchange rates in terms of comparative-statics analysis. The second
section focuses on the determinants of the length of time required to
regain equilibrium after a monetary disturbance. The third deals with
the proper use of monetary policy under fixed exchange rates, while
some concluding remarks are offered in the last Section.

I. THE EFFECTIVENESS OF MONETARY POLICY

In a fundamental sense, monetary policy can have no lasting impact


on the income level of a small open economy under fixed exchange
rates. It is important to understand the meaning and limitations of this
proposition as well as the assumptions required to make it hold.
First, monetary policy must be defined as an exogenous once-and-for-
all change in the domestic assets of the consolidated banking system (or
of the domestic source component of the base) and not as an exogenous
change in the money supply or the rate of interest (as the point of the
analysis is to show that these two variables are endogenous and not
exogenous). Second, an economy will be said to be small if it cannot
influence foreign interest rates, income levels, and so forth. Third, by
lasting influence is meant a permanent change in income after the
economy has adjusted fully to the change in the domestic assets of the
banking system (i.e. the proposition is stated in terms of full-equilibrium
comparative statics). Fourth, by fixed exchange rates is meant that the
spot rate is rigidly fixed by the exchange-stabilization operations of the
government. For simplicity, assume that spot rates, and all other
variables, are expected to remain at current levels so that the forward
rate coincides with the spot exchange rate.1
Given these definitions, three assumptions only are needed to prove
the proposition stated at the beginning of this section, namely: (1) that

'This assumption is not strictly necessary to the proof of the proposition stated
at the beginning of this section. If we assume that the system is stable and that
people learn, the full equilibrium to which the system eventually converges will
not be affected. However, some assumptions result in an unstable model.

This content downloaded from 128.42.202.150 on Mon, 27 Jun 2016 07:08:44 UTC
All use subject to http://about.jstor.org/terms
138 ECONOMICA [MAY

the economic system is stable; (2) that an increase in the money supply,
from equilibrium, tends to create a balance-of-payments deficit; and (3)
that the associated reserve loss tends to reduce the money supply.' For,
suppose that, starting from a position of full equilibrium in the economy,
the monetary authorities increase the domestic assets of the consoli-
dated banking system; for a given stock of foreign-exchange reserves
the money supply increases and a balance-of-payments deficit emerges
(by assumption 2); reserves fall and the money supply contracts (by
assumption 3); as long as the money supply has not returned to its
initial value there must be a balance-of-payments deficit and a further
contraction of the money stock; final equilibrium will occur when the
system has returned to its initial equilibrium position, as it will under
assumption 1. In the final equilibrium, the money supply is returned
to its initial level, as are all real variables and prices. Only the composi-
tion of the consolidated banking system's assets has changed, the
increase in domestic assets being matched by an equal decrease in
foreign reserves.2
Note that this conclusion has been reached independently of any
specific assumnptions as to the existence of capital movements, the
responsiveness of the latter to interest-rate changes, the ratio of traded
to non-traded goods, or the extent to which transitional changes take
the form of real output or price variations. These factors will affect the
path of adjustment (and their role in this context will be discussed in
the next section) but not the final equilibrium. That this should be so is
readily explained. Our assumptions imply that the money supply in an
open economy and under fixed exchange rates is an endogenous
variable in the "long run", that is, that, other things equal, there is
only one money stock compatible with payments equilibrium, and that
the monetary mechanism of adjustment works properly, i.e. that it will
ensure that the money stock converges to its equilibrium value. In an
important sense, our conclusion is nothing but the small-country
counterpart of the Ricardian "natural distribution of specie".
It may be useful to illustrate the above general proposition in the
specific context in which it was originally put by Mundell and amplified

1 Under most "reasonable" dynamic postulates, if the third assumption is


satisfied, then satisfaction of the second becomes a necessary condition for
stability of the system, i.e. for satisfaction of the first. This interdependence is
merely noted here but plays a fundamental role in the analysis of Section III,
below.
2 This assumes, of course, that the composition of the banking system's assets
does not per se affect the public's behaviour. In addition, it is assumed implicitly
that open-market operations do not engender wealth effects in full equilibrium
(XX in Figure 1 does not shift) as they would not if changes in future tax liabilities
are entirely discounted by the public. Furthermore, for a small open economy, the
assumption that the relationship between real variables and the payments balance
is unaffected by changes in the level or composition of wealth is sufficient to ensure
fixity of FF in Figure 1 and, in the case of capital immobility, invariance of full-
equilibrium income to monetary policy, whatever happens to XX. In a world
model this need not be the case as wealth effects would affect aggregate world
expenditure.

This content downloaded from 128.42.202.150 on Mon, 27 Jun 2016 07:08:44 UTC
All use subject to http://about.jstor.org/terms
1973] MONETARY POLICY UNDER FIXED EXCHANGE RATES 139

by the present writer.' Figure 1 represents the combinations of interest


rate and income level that equate the supply of domestic output with
the demand (the sum of domestic spending and the balance of trade)
along XX, the demand for money with the supply along LL, and for
which the trade balance is equal but opposite in sign to the capital
account, leaving the balance of payments in equilibrium, along the
curves FF. These last three curves correspond to various degrees of
capital mobility (defined as the interest sensitivity of capital inflows),
ranging from capital immobility along F1F1 to perfect capital mobility
along F3F3. Three interpretations can be given to variations in "income":
changes in real output with prices constant, changes in domestic prices
(foreign prices assumed constant) with real output constant,2 or a
mixture- of both. An increase in the domestic assets of the banking
system temporarily pushes LL out and to the right to, say, L'L'. A

x F

F3 A3___
B7L

/F2/ , X
do/' X
F1

0Y
FIGURE 1

payments deficit emerges and the money supply contracts until LL is


re-established. That is, an integral component of this type of model, in
addition to the excess demand functions for money, domestic output
and foreign exchange is the equation that ensures endogeneity of the
money supply, an equation of the form:

dM/dt = h(F), h'> O,


where F is the excess supply of foreign exchange and M the domestic
money supply.
Before turning to the influence of the size of countries on this con-
clusion, the meaning of the expression "endogeneity of the money stock"
should be clarified and contrasted with that of "controllability of the

I See Mundell [8], as reprinted in [10]; and Swoboda [13], from which Figure 1
is reproduced with minor changes. (References in square brackets are listed on
pp. 153-4.)
2 For the variable price case see, in particular, Mundell [9], also reprinted in [10].

This content downloaded from 128.42.202.150 on Mon, 27 Jun 2016 07:08:44 UTC
All use subject to http://about.jstor.org/terms
140 ECONOMICA [MAY

money stock". Endogeneity of an economic variable is always defined


in relation to a particular model and to a relevant length of run. Thus,
in terms of the models discussed so far, the domestic assets of the bank-
ing system are considered to be policy-determined and exogenous over
any relevant time period and endogeneity of the money stock in full
equilibrium (the long run after all adjustments have taken place) is
implied by the way in which the model is specified, especially with res-
pect to the monetary mechanism of payments adjustment. Alternatively,
one could treat changes in the stock of money as exogenous in the very
short run (that is, one could treat the impact effect-before any other
adjustment has taken place-of a change in domestic assets of the
banking system on the stock of money as exogenous in the very short
run), the rate of change of the money stock as endogenous in the inter-
mediate run (defined by reference to an ongoing balance-of-payments
adjustment process), and the equilibrium money stock as endogenous
to the system in the long run (defined by reference to complete adjust-
ment in all markets).' Endogeneity, however, should not be confused
with "uncontrollability". In buffer-stock analysis, the quantity and
price of coffee bought and sold are endogenous variables, yet they may
or may not be "controllable" depending on the buffer stock's inventories
of coffee and money relative to flow private excess demands or supplies.
In the present context, controllability of the money stock will depend
(a) on the limits set to the decumulation or accumulation of foreign-
exchange reserves by the availability of foreign-exchange reserves on the
one hand and that of domestic assets on the other hand, and (b) on the
speed of adjustment of the system to a discrepancy between the actual
and equilibrium money stock. This topic will be pursued further in
Section II below.
Finally, note once again that the conclusion- that monetary policy
affects only the composition of the assets of the banking system is of
course valid only for the definition of monetary policy adopted here,
namely, a once-and-for-all change in the stock of domestic assets of the
consolidated banking system. This definition rules out systematic
neutralization operations by the monetary authorities, that is, the
creation of aflow of domestic assets equal in size but opposite in sign to
the flow of foreign-exchange reserves.
The above conclusion as to the ineffectiveness of monetary policy in
terms of the full-equilibrium values of real variables depends crucially
also on the "small country" assumption. Defining the effectiveness of
monetary policy as the change in domestic income or interest rate that
results from a one "dollar" change in the domestic assets of the banking
system, it is possible to show that this effectiveness is directly propor-
tional to the relative size of the country undertaking the open-market
operation. This proposition has been advanced by Mundell for the case

1 This three-fold classification follows on a suggestion made by Karl Brunner


at the Konstanz meeting.

This content downloaded from 128.42.202.150 on Mon, 27 Jun 2016 07:08:44 UTC
All use subject to http://about.jstor.org/terms
1973] MONETARY POLICY UNDER FIXED EXCHANGE RATES 141

of perfect capital mobility within a "Keynesian" framework.' We will


show that it is in fact quite general and that it follows from the proposi-
tion that, under fixed exchange rates, the final impact on all variables,
except the distribution of the world's reserves, of an increase in the
domestic assets of country A is the same as that of an equivalent in-
crease in domestic assets in any other country. That is, the final outcome
is entirely determined, with the exception noted above, by the impact of
the increase in domestic assets on the "world money supply"-the sum
of the money stocks in the hands of the publics of various countries.
In terms of a two-country model, we can write:

(1) Mw=M+M2=D,+D2+Rl+R2
D1+D2+R1+(Rw-R1)j D1+ D2+Rw,
where M refers to money supply, D to the domestic assets of the
consolidated banking system, R to foreign exchange reserves-, the sub-
subscripts 1, 2 and w refer to country 1, country 2 and the world,

PI ~~T
2 K'

K B

A c K'

T ~K

0 1
FIGURE 2

respectively, and where world reserves RW are assumed to be given. The


proof of our statement is particularly simple if we can assume that each
country is specialized in the production of one good, that full employ-
ment prevails in both countries, and that the income elasticity of the
demand for money is unity in both countries. This case is illustrated in
Figure 2. TT shows those combinations of the money prices of the two
goods which keep the balance of trade in equilibrium (this line goes
through the origin and stays fixed as long as "real" forces do not change
the equilibrium terms of trade). KK shows those combinations of the
two money price levels that would keep the world money supply equal
to the sum of the demands for money in the two countries. A movement
along KK corresponds to a redistribution of the world money supply
between the two countries. For a given world money supply, the distribu-
tion of specie is unique; suppose we take money away from the residents
l See [10], appendix to ch. 18.

This content downloaded from 128.42.202.150 on Mon, 27 Jun 2016 07:08:44 UTC
All use subject to http://about.jstor.org/terms
142 ECONOMICA [MAY

of country 2 to give it to those of country 1 and let prices vary so that


we are at C. At that point there would be an excess demand for the
goods of country 2 (a trade surplus for country 2) and money would
flow from 1 to 2 until equilibrium is re-established at A. Suppose now
that the central bank in country 1. increases the domestic assets of its
banking system. By Eq. (1) this increases the world money supply and
KK shifts up to K'K'. Country 1 experiences a temporary balance-of-
payments deficit until the world money supply is redistributed and prices
adjusted to make B the new equilibrium point. Note that the final
equilibrium would also be at B had country. 2 initiated the money
supply increase.
It is now a simple matter to prove that the impact of a given change in
the money supply of country 1 is proportional to the size of that country.
The quantity equation for the two countries is given by (2) and (3) below,
and trade equilibrium requires (4):

(2) Pi Y1 = M1 V1;

(3) P2 Y2 =M2V2;

(4) P11P2= Q,
where V1 and V2 are income velocities of circulation, Y1 and Y2 are
real income levels, and Ql is a constant. Substituting from (2) and (3)
into (4) and carrying out log differentiation yields

(5) dlogM1+dlog V1+dlog Y2-dlog M2-dlog V2-dlog Y1


=0.

Noting that Y1, Y2, V1 and V2 are assumed constant (we assumed unit
income elasticity of the demand for money), (5) yields: dlogM1=
d log M2; the increase in the world money supply is distributed propor-
tionately to existing money stocks, the latter being obviously related to
the size of countries (and exactly related if V1 = V2). For instance
suppose that, initially, the money stock of country 1 is $10 and that of
country 2 is $90. Now let country 1 increase the money supply by $1; in
the final equilibrium the money supply of country 1 will have increased
by $0'10, the other $0 90 spilling out to country 2 to increase the latter's
money stock. Money prices will have increased by only 1 per cent.,
even though the stock of money was initially increased by 10 per cent.
in country 1.
Much the same type of result can be obtained in a "Keynesian world"
where each country's economy is represented by the type of model
depicted in Figure 1 above. The bare bones of such a model are given in
Figure 3.1 Income levels replace prices on the abcissa and ordinate,
and the slope of the balance-of-payments equilibrium schedule TT is
equal to m1/m2, the ratio of marginal propensities to import. Assuming
1 A complete analysis of this model can be found in Swoboda and Dornbusch
[15].

This content downloaded from 128.42.202.150 on Mon, 27 Jun 2016 07:08:44 UTC
All use subject to http://about.jstor.org/terms
1973] MONETARY POLICY UNDER FIXED EXCHANGE RATES 143

that the income elasticity of the demand for imports is unity in both
countries and that the capital account is initially balanced, an increase
in the world money supply (whatever its origin) again shifts the KK
curve to K'K' and changes income levels in the two countries in the
same proportion. If we now assume that the income and the interest
elasticities of the demand for money are the same in the two countries,
the increase in the world money supply is again distributed among the
two countries in proportion to their income levels (and initial money
stocks) as the system moves from A to B. This result is independent of
the degree of capital mobility since, as Dornbusch and the present
writer have shown [15], the shift in KK is independent of the degree of
capital mobility and the equilibrium interest-rate differential is invariant
with respect to changes in the world money supply.'

Y2

K?, f < K' T


K~~

K'

0 Y
FIGURE 3

To summarize, the effectiveness of monetary policy, defined as the


full-equilibrium impact on domestic money income of a one dollar
open-market operation is directly proportional to the size of the home
country relative to that of the rest of the world, this effectiveness
tending to zero as the country becomes very small in relative terms.
So far, the analysis has been carried out under the assumption of
stationary expectations and once-and-for-all changes in the money
supply. As has been shown by Mundell, similar results hold in terms of
the comparative dynamics of equilibrium growth and inflationary
paths.2 Assuming for simplicity that all goods are perfect substitutes
in the world economy and that the money-income elasticity of the
demand for money is unity, the equilibrium percentage rate of growth
of the money supply is equal to the sum of the percentage rates of
growth of output and prices (in the short run, stock adjustments
make for a more complicated result). Any attempt at increasing
'See [15], Section IV.
2 The most important contribution is Mundell [11], ch. 15. See also Mundell
[10], ch. 9; Komiya [4]; Johnson [2]; and a forthcoming paper by A. Laffer.

This content downloaded from 128.42.202.150 on Mon, 27 Jun 2016 07:08:44 UTC
All use subject to http://about.jstor.org/terms
144 ECONOMICA [MAY

(decreasing) the rate of monetary expansion above (below) its equili-


brium level through changes in the rate of domestic credit expansion
results only in a balance-of-payments deficit (surplus) in the small open
economy. Equilibrium in the balance of payments (no change in the
IldD
stock of reserves) requires a rate of domestic credit expansion, - -d-,

equal to D (P dt + Tt ) with symbols defined as before.' Note that


in this type of model, inflation in the small country is always of the
imported type, except in the very short run. Prices of goods and services
are determined in the international market and domestic monetary
policy has but a negligible influence on international prices. Inflation
(or deflation) is imported through goods arbitrage; capital movements
are neither necessary nor sufficient for the process to take place. They
will, however, in practice influence the speed of adjustment to inflation-
ary or non-inflationary equilibrium, the subject to which we now turn.

I1. THE SPEED OF ADJUSTMENT


The discussion so far has been concerned with full-equilibrium
comparative statics and once-and-for-all changes in the domestic
source component of the monetary base. For practical purposes, it is
important to ask how long it will take for full equilibrium to be re-
established after an initial monetary disturbance and whether and to
what extent disequilibrium policies can be effected. In terms of Figure 1,
the questions we want to ask are: (a) how long will it take to restore the
initial equilibrium after a shift of LL to, say, L'L'; (b) how feasible is it,
and what is required, to keep the money supply at a level consistent with
maintenance of the system at a point such as B? To answer these
questions satisfactorily would require, at the analytical level, building a
complete dynamic model of the adjustment process and, empirically,
determining the value of the parameters that enter the analytical model.
The approach below is a much more modest one; we discuss, intuitively
and separately, the probable influence of two factors-the degree of
capital mobility and the proportion of traded to non-traded goods-on
the speed of adjustment of the system. We also consider briefly the role
of exchange-rate margins. The analysis is non-rigorous and subject to all
the usual caveats about implicit dynamics.
One possible procedure to obtain partial answers to the questions
raised above is as follows: suppose that somehow the system temporarily
settles at point B in Figure 1; then ask how the magnitude of the

1 The expression in the text is obtained simply by differentiating M= D + R =


L ( Y, P) proportionally with respect to time, setting dR/dt equal to zero, and noting
that the elasticity of the demand for money with respect to each of its arguments is
assumed equal to 1. With given growth rate of output and (world) rate of inflation,
the balance of payments as a proportion of the stock of reserves, (1/R)(dR/dt), is
entirely determined by the percentage rate of domestic credit expansion, (I/D)
(dD/dt).

This content downloaded from 128.42.202.150 on Mon, 27 Jun 2016 07:08:44 UTC
All use subject to http://about.jstor.org/terms
1973] MONETARY POLICY UNDER FIXED EXCHANGE RATES 145

disequilibrium at B (in a flow per unit of time sense) is affected by the


two factors mentioned above. Presumably, the larger the disequilibrium
at B, the more speedily, other things equal, would the (stable) system
tend to return to equilibrium and the harder would policies have to be
applied to maintain it at B.' One way of analysing this issue is to ask
what policies are required to turn the temporary disequilibrium point B
into a "quasi-equilibrium" point.
Consider point B: the demand for domestic output is equal to the
supply, and the demand for domestic money by residents is equal to
the supply, but there is a balance-of-payments deficit. Other things
equal, there would be a tendency for the money supply to decline as the
monetary authorities intervene in the foreign-exchange market to
prevent a depreciation of the home currency. However, the authorities
can keep the money supply at the level implied by L'L' by neutralizing
the monetary effects of reserve losses.2 Denoting the base by B, the
money multiplier (assumed constant at the given interest rate) by m,
foreign-exchange reserves by R and the domestic securities held by the
central bank by D*, we have:

(6) M=mB=m(R+D*);

1 dM 1 (dR dD*)
(7) Mdt= 7dt+
Thus, keeping the money supply constant requires

(8) dD*Idt = - dR/idt.

In words, the monetary authorities must increase (decrease) the domestic


source component of the base by the same amount per unit of time as
the foreign source component decreases (increases), namely, by an
amount equal to the balance of payments deficit (surplus), dR/dt. Thus,
the extent of the balance-of-payments disequilibrium at B determines
the rate of neutralization operations required to keep the system in a
state of quasi-equilibrium, and, in the absence of neutralization, will
determine in part the time interval needed to restore full equilibrium.
In this context, consider, first, the role of capital mobility, defined for
simplicity as the interest-rate responsiveness of international flows of

'This statement is approximative rather than exact since the length of time
required to reach equilibrium after a temporary disturbance depends not only on
the speeds of adjustment in various markets but also on whether the approach to
equilibrium is direct or cyclical.
2 It may seem, at first, that treating points like B as quasi-equilibria violates
Walras' law. For how can there be a disequilibrium in the foreign-exchange
market when both the money and goods markets are in equilibrium? The answer
is that a flow excess supply of securities by the public matches their flow excess
demand for foreign exchange. The authorities prevent these disequilibria from
affecting the goods and money markets by absorbing the flow excess supply of
securities at the existing rate of interest through their neutralization operations
and by satisfying the excess private demand for foreign exchange at the existing
exchange rate through their exchange-stabilization operations.

This content downloaded from 128.42.202.150 on Mon, 27 Jun 2016 07:08:44 UTC
All use subject to http://about.jstor.org/terms
146 ECONOMICA [MAY

capital.' In the simple, variable real income "Keynesian", framework of


Figure 1, the balance-of-payments disequilibrium at a quasi-equilibrium
point like B is given by expression (9):

dR mE - (s +m)K1
(9) dt-sm) -) E (M-M*)<O,
where s and m are the marginal propensities to save and import,
respectively, E is domestic expenditure, L the demand for money, K net
capital imports, Y money income, i the rate of interest, M and M* are,
respectively, the actual and full-equilibrium stocks of money (M> M*),
and subscripted variables denote partial derivatives with respect to the
subscript.2 Obviously, the higher the interest responsiveness of capital
flows, Ki, the larger the payments disequilibrium created by a dis-
crepancy between the actual and full-equilibrium money stock. In the
limit, the payments disequilibrium tends to infinity as capital mobility
becomes perfect (Ki-*oo). This is the case where neutralization becomes
impossible and self-contradictory as it would require open-market
operations to be undertaken at an infinite rate; this is inconsistent with
positive and finite reserve stocks-a deficit country would rapidly lose
all its reserves, a surplus country accumulate the world's entire stock of
reserves. To summarize, the higher the degree of capital mobility the
less scope for the disequilibrium effects of monetary policy and the
larger the rate of neutralization operations required to maintain a given
quasi-equilibrium stock of money.3
Consider, next, the role of non-traded goods. To focus on the point at
issue, imagine that there are only three goods in the system; traded or
international goods, non-traded or domestic goods, and money.
Assume that the three goods are substitutes. Assume further that the
foreign-currency price-and hence the domestic-currency price at a
given exchange rate-of international goods is fixed or exogenously
determined. From equilibrium, let the monetary authorities increase
the money supply; the impact effect is to create an excess supply of
money and an excess demand for both domestic and international goods.
The excess demand for foreign goods is reflected in an excess demand for
foreign exchange, and the money supply will tend to decrease as the
authorities sell foreign exchange to stabilize the exchange rate. In the
end, full equilibrium will be re-established when the money supply has
returned to its original level, with international prices and domestic
prices unchanged. However, there will have been a transitory increase
1 Conceptually, it would be preferable to treat capital movements as resulting
from a stock-adjustment process. The flow approach could be considered as a
special short-run version of the stock-adjustment one, the interest sensitivity of
flows depending, in part, on the speed at which portfolios are adjusted.
2 For a derivation and further explanation see Swoboda [13].
3 If the stock-adjustment view of capital movements is adopted, the required
extent of neutralization policies becomes independent of the degree of capital
mobility in the long run; after the stock adjustment has been completed, the
neutralization rate depends only in the trade disequilibrium associated with the
quasi-equilibrium point B.

This content downloaded from 128.42.202.150 on Mon, 27 Jun 2016 07:08:44 UTC
All use subject to http://about.jstor.org/terms
1973] MONETARY POLICY UNDER FIXED EXCHANGE RATES 147

in domestic goods prices. The length of time it will take for the system to
return to equilibrium will depend partly on the size of the impact effect
on the excess demand for foreign exchange and hence on the balance of
payments and the rate of change of the money supply. To the extent
that part of the excess supply of money is absorbed by a rise in the
prices of domestic goods, the excess demand (per unit of time) for
international goods will be smaller than it otherwise would be (by
Walras' law). Therefore, other things equal, we would expect the
rapidity with which the system adjusts to a monetary disturbance to be
directly related to the ratio of traded to non-traded goods, even though
the final equilibrium is not.
The same type of reasoning can be applied to the analysis of the role of
non-traded goods in the transmission of "imported inflation". We will
show that, contrary to a belief sometimes expressed, the presence of

PD

0 P*~/

0 po1 p*
FIGURE 4

non-traded goods does not affect the full transmission of inflation in the
long run and in the absence of neutralization policies; it does, however,
affect the "length of the short run" and the rate at which neutralization
operations need to be carried out to maintain the domestic below the
international rate of inflation.
These points are illustrated in Figure 4, which is based on a diagram
used by Mundell to analyse the effects of devaluation (an issue with
which we are not concerned here).' The curve DD shows those combina-
tions of domestic-goods prices, PD, and international-goods prices, PI,
that equate the demand and supply of non-traded goods, II those
combinations of PI and PD that leave the excess demand for international
goods equal to zero, and MM those combinations of the two prices that
equate the demand for money with the existing stock. The three curves
are drawn on the assumption that the three goods are substitutes: in
' See Mundell [11], ch. 9.

This content downloaded from 128.42.202.150 on Mon, 27 Jun 2016 07:08:44 UTC
All use subject to http://about.jstor.org/terms
148 ECONOMICA [MAY

addition, assume that the real excess demand functions are homogen-
ous of degree zero in the two money prices and the nominal quantity of
money. Initial equilibrium is at Q. Suppose now that the price of
international goods rises in the rest of the world; the domestic-currency
price of international goods must rise in the same proportion, say from
P?' to P* if the rate of exchange is fixed and if goods arbitrage takes
place. Before the price of domestic goods changes, the impact effect of
the rise in foreign prices is to move the commodity prices to point R;
there is now an excess demand for money and domestic goods and an
excess supply of foreign goods. The price of domestic goods begins to
rise, and the country experiences a balance-of-payments surplus that
shifts the MM curve up and to the right pulling the DD and IH curves
in its wake. Final equilibrium is established at T; by the homogeneity
postulate the money prices and the money supply will all have increased
by PoP*/OPo.'
The movement to T, however, will tend to be slower with non-traded
goods present in the system. For, as domestic-goods prices rise, part of
the initial excess supply of money will be eliminated and the deficit per
unit of time be lower. Similarly, keeping the money supply at its initial
level involves a lower rate of sterilization operations when non-traded
goods are present. As foreign prices rise and the money supply is kept
constant, domestic goods prices rise until the market for domestic goods
is cleared, i.e. the system moves to point S. At S, the flow excess supply
of international goods is equal to the payments surplus and to the rate
of sterilization operations required to keep MM fixed. Define an
aggregate price index by P= aPD + (1- a)PI, where the quantity weights
are, respectively, the shares of domestic and international goods in
total expenditure. Sterilization policies keep the percentage increase in
P below the percentage increase in PI by preventing PD from rising in
the same proportion. The higher the share of non-traded goods in total
expenditure, the less P will increase for given increases in PI and lesser
increases in PD (such as those involved in moving to point S). The greater
the elasticity of substitution between traded and non-traded goods,
the closer the percentage increase in PD will be to that in P, (the closer
point S will lie to point T) and the less successful will a sterilization
policy be in moderating the rise in the price index p.2
I It may be interesting to note that Mundell shows that a devaluation of
P,P*IOPI- leads to a final equilibrium at T. Thus we,have shown that an x per
cent. devaluation is equivalent in its effects to an x per cent. "exogenous" increase
in foreign prices, a point that is consistent with a proposition developed by
Kuska [5] in an analysis of devaluation based on a Patinkin-type disaggregated
model. Kuska shows that an x per cent. devaluation is equivalent to a (lOOx/100-x)
per cent. increase in foreign money stocks. With due attention paid to the defini-
tion of percentage changes, the two propositions are equivalent.
2 In the discussion in the text capital movements were neglected. If we allow for
the existence of a bond market, the excess demand for money created by the
imported inflation at point Q would tend to result in a rise in the interest rate. As
the rate of interest would be higher when the system is at S than when it is in full
equilibrium, capital inflows will tend to take place and make the task of keeping the
domestic below the world rate of inflation more difficult.

This content downloaded from 128.42.202.150 on Mon, 27 Jun 2016 07:08:44 UTC
All use subject to http://about.jstor.org/terms
19731 MONETARY POLICY UNDER PIXED EXCHANGE RATES 149

The analysis above was carried out in terms of a once-and-for-all


increase in international goods prices. It could easily be re-cast in
terms of rates of increase of international goods prices, domestic goods
prices and money supply, care being taken to allow for the effects on the
stock demand for real balances of changes in the expected rate of
inflation.
Finally, it has been assumed so far that the spot rate of exchange was
perfectly fixed and expected to remain so, and that there was no differ-
ence between the spot and forward rates. This is clearly an "unrealistic"
assumption and an inappropriate one in certain circumstances. In the
real world, the spot rate is allowed to fluctuate, albeit within narrow
intervention limits, and the forward rate is not systematically pegged.
Complete discussion of the complications introduced by this flexibility
is impossible here. Suffice it to say that exchange-rate margins make it
possible for the system to behave somewhat like a flexible rate system as
long as the margins are not reached. For instance, an increase in the
money supply would lead to a permanent increase in prices or income
(with fixed money wages and money illusion) as the home currency
depreciates, that is, as long as the price of foreign exchange does not
rise to the upper intervention point; moreover, the higher the degree of
capital mobility, the greater the effectiveness of monetary policy under
flexible exchange rates.' The flexibility of forward rates gives some
additional scope for the use of monetary policy even in the face of a
high degree of capital mobility and pegged spot rates. For, if the flow of
arbitrage capital is in perfectly elastic supply with respect to the
covered interest-rate differential, it is still possible to create a divergence
between domestic and foreign interest rates by affecting the premium or
discount on forward exchange. However, the scope for affecting
premia on forward exchange is severely limited as long as confidence in
the parity is maintained; for the speculative supply or demand for
forward exchange will tend to become perfectly elastic at the exchange-
rate margins.
The main strands of the argument up to this point can be summarized
by four propositions. (1) The full-equilibrium effect on incomes,
prices and interest rates of a once-and-for-all change in the domestic
assets of the banking system of an open economy is directly proportional
to the economic size of the country relative to the rest of the world. A
corollary of this proposition is that t-he effectiveness of monetary policy
thus defined tends to zero as the country becomes smaller and smaller.
(2) This conclusion is independent of the degree of capital mobility
and the proportion of traded to non-traded goods. (3) In the short run,
however, it is possible for a country to affect its price and income level
by maintaining the money stock at a given level through neutralization
policy, that is, by combining a money-stock policy with a flow-
neutralization policy-where the flow of open-market sales (purchases)

1 See Mundell [10], chs. 17 and 18; and Sohmen [12].

This content downloaded from 128.42.202.150 on Mon, 27 Jun 2016 07:08:44 UTC
All use subject to http://about.jstor.org/terms
150 ECONOMICA [MAY

equals the balance-of-payments surplus (deficit) implied by the main-


tenance of the money supply below (above) its full-equilibrium value.
This possibility is not available when capital is perfectly mobile or
when there are no non-traded goods as the required rate of neutraliza-
tion policy would rapidly become infinite. (4) In the long run, however,
the maintenance of quasi-equilibrium positions is incompatible with
fixed exchange rates; the limits are reached in theory when only domestic
assets back the money supply in case of a deficit, and when only foreign
assets back the money supply in case of a surplus. As the ratio of
domestic to foreign assets is usually larger than one, this means that
a quasi-equilibrium involving a surplus can, potentially, be maintained
for a longer lapse of time than one involving a deficit.'
The effects of monetary and neutralization policy under fixed ex-
change rates have now been outlined. It remains to try to define the
appropriate use of monetary policy in an open economy.

III. THE APPROPRIATE USE OF MONETARY POLICY2


There are, we have argued, severe limits to the effectiveness of monet-
ary policy as a counter-cyclical device under fixed exchange rates except
in the short run. However, within the limits set by the available stock
of reserves and the openness of the economy as measured by its relative
size, the degree of capital mobility and the ratio of traded to non-traded
goods, monetary policy coupled with neutralization of reserve flows
can still be used to stabilize income and prices in the short run. Under
what circumstances is it "appropriate" to use the monetary instrument
in such a fashion under fixed exchange rates?
The answer, I would argue, is "in those cases where the balance of
payments takes care of itself in the long run". There are two principal
cases where this statement would apply.
First, suppose that the initial position is one of internal and external
balance; then introduce cyclical variations in the balance of payments,
deficits alternating with surpluses but averaging out to zero over a
finite time period. In that case it may be appropriate to maintain the
money supply at its long-run "average" equilibrium level, neutralizing
temporary reserve changes through open-market operations. A pre-
requisite for the success of such a policy is that the reserve stock be
'This statement is subject to at least two qualifications. First, in practice, the
limits are reached faster than in theory as speculation as to exchange-rate changes
sets in. Second, as payments imbalances are usually defined, the foreign-exchange
assets that are accumulated (decumnulated) in the case of a surplus (deficit) are
those of the monetary authorities; the ratio of domestic to foreign-exchange assets
of the monetary authorities is quite low in several open economies; in this case the
scope for neutralization of a surplus through traditional open-market operations
will be limited. A substitute to open-market operations is to induce commercial
banks to absorb into their portfolios the foreign-exchange assets that would other-
wise flow to the central bank; this is one of the goals of the German Bundesbank's
policy of offering commercial banks special swap rates.
2 In what follows, I abstract from the problem of the proper use of monetary
policy by the centre or nth country of the system.

This content downloaded from 128.42.202.150 on Mon, 27 Jun 2016 07:08:44 UTC
All use subject to http://about.jstor.org/terms
1973] MONETARY POLICY UNDER FIXED EXCHANGE RATES 151

large enough relative to disturbances to finance temporary deficits


without causing anticipations of devaluation. Moreover, this policy will
succeed in stabilizing income and prices only if the payments disturb-
ances originate predominantly in the capital account, that is, if in
Figure 1 the curve FE shifts while the XX curve stays put. If this is not
the case, stabilizing the money supply will not suffice to stabilize prices
and income.' Second, consider disequilibrium situations where the
monetary policy requirements of internal and external balance coincide,
that is, situations where inflation is coupled with a deficit or where
deflation is accompanied by a payments surplus. In these cases, directing
monetary policy towards internal balance will also relieve the existing
payments disequilibrium.
On the other hand, the attempt to use monetary policy as a counter-
cyclical device in so-called dilemma cases, unless these reverse them-
selves fairly rapidly, is incompatible with a regime of fixed exchange
rates. For, suppose that deflation is accompanied by a deficit; using
monetary policy (coupled with neutralization operations) to restore full
employment will eventually lead to exhaustion of foreign exchange
reserves, and conversely for the case where a payments surplus is
coupled with inflation. These are, of course, the dilemma cases empha-
sized in Meade's classic work [6].
The preceding remarks should not be interpreted to mean that there
is no important role for monetary policy in the open economy under
fixed exchange rates. For, though monetary policy is rarely appropriate
as a counter-cyclical device and taken by itself, unless the level of
reserves is of no concern, it does represent a very powerful instrument
of balance-of-payments policy. In terms of full-equilibrium comparative
statics, a $1 open-market sale of securities increases the stock of foreign
exchange reserves by $1 (MW-M1)/Mw under the assumptions listed
in Section I above. As M1 becomes very small relative to Mw, the
reserve gain tends to equality with the open-market sale.
This suggests that some other instrument be used for internal balance,
leaving monetary policy to take care of residual payments imbalances.
This is precisely the purpose of Mundell's well-known monetary-fiscal
policy-mix analysis which consists of two separate (though related)
propositions :2 (1) that fiscal and monetary policies have different
impacts on internal and external balance and hence represent two
separate instruments with the help of which it is possible to achieve
simultaneously internal and external balance, a comparative-statics
proposition or possibility theorem; (2) that, in a decentralized system
of policy responses, assigning monetary policy to external balance and
fiscal policy to internal balance leads to convergence to the desired
position of simultaneous internal and external balance while the

1 A case can also be made for the use of monetary policy to stabilize income and
prices when disturbances originate in finite shifts in the public's portfolio prefer-
ences.
2 See Mundell [7], reprinted in [10].

This content downloaded from 128.42.202.150 on Mon, 27 Jun 2016 07:08:44 UTC
All use subject to http://about.jstor.org/terms
152 ECONOMICA [MAY

reverse assignment of instruments to targets does not, a proposition in


economic dynamics.
Whereas the first of these propositions suggests that to achieve
internal and external balance simultaneously it is sufficient to find those
values of the monetary and fiscal instruments that will do so and set
these instruments accordingly, the second proposition asserts that in a
system of decentralized response-or more fundamentally in a system
of limited information-the proper use of monetary policy under fixed
exchange rates is for external balance. The fundamental basis for this
assignment of the monetary instrument resides in the automatic
monetary mechanism of payments adjustment. As a matter of fact, it can
be shown that stability of the open economy requires as a necessary
condition that an increase in the money supply, from equilibrium, leads
to a deterioration of the balance of payments. It turns out that fulfil-
ment of that condition is necessary and sufficient to insure stability of
the assignment of monetary policy to the balance of payments and of
fiscal policy to internal balance.'
To conclude this section, a few remarks on the interpretation of the
expression "appropriate" or "proper" use of monetary policy are in
order. The normative content of these expressions is limited even
though, I would argue, important. The appropriateness of the particular
recommendations discussed above is conditional on certain maintained
hypotheses and restricted to a narrow concept of "appropriateness".
The principal maintained hypothesis is that a fixed exchange regime
prevails and that no alternative system is available. Appropriateness
basically means (a) "possible" in terms of a fixed-target framework
limited to fairly narrow definitions of internal and external balance, and
(b) convergent under a system of decentralized decision-making under
limited information. In addition, tying monetary policy to the balance
of payments is seen to be the only governing principle for monetary
policy consistent with maintenance of a fixed exchange-rate system in
the long run. These conclusions are quite consistent with (a) a stock-
adjustment view of capital movements-the latter implying simply that
reconciliation of internal and external balance through manipulation of
the capital account is possible only in the short run but would imply
ever-widening interest rate differentials in the long run, and (b) with the
argument first advanced by Johnson and pursiled by Williamson that
the composition of the balance of payments that results from application

1 This point is demonstrated and related to the problem of limited information


in Swoboda [14]. Note also that the expression "effectiveness" has sometimes
been used in the policy-mix literature in a different sense from that given to it in
Section I above. Thus, monetary policy is sometimes said to be more effective than
fiscal policy in dealing with external balance (and vice versa for internal balance)
under fixed exchange rates. This is a notion of relative effectiveness, or comparative
advantage, that is most useful in a discussion of convergence of decentralized
assignments. Formally, the assertion above about comparative effectiveness means
that the ratio of the effect of a given change in monetary policy on external
balance to its effect on internal balance is greater than the equivalent ratio for
fiscal policy under fixed exchange rates.

This content downloaded from 128.42.202.150 on Mon, 27 Jun 2016 07:08:44 UTC
All use subject to http://about.jstor.org/terms
1973] MONETARY POLICY UNDER FIXED EXCHANGE RATES 153

of the policy mix may, or even is likely to, run counter to the dictates
of the welfare considerations imnplied by maximization -of a social
welfare function over time.' In the context of the policy-mix model, this
last consideration implies that a third instrument be found to make the
composition of the balance of payments consistent with the steady-
state one that maximizes world welfare over time.

IV. CONCLUDING REMARKS


The main thrust of this analysis has been to argue that monetary
policy is an inappropriate instrument of anti-cyclical policy under
fixed exchange rates in a small open economy in anything but the short
run. That is, systematic use of monetary policy to stabilize incomes and
prices is, save under exceptional circumstances, incompatible with the
proper functioning of a system of fixed exchange rates. This argument
does not require that the stock of money be uncontrollable; except
under extreme circumstances (such as perfect mobility of all forms of
capital), some measure of control over the stock of money can be re-
tained in the short run; rather, the issue concerns the use to which the
remaining measure of control should be put. The simultaneous pursuit
of internal and external balance requires the use of at least two instru-
ments of economic policy, say,. fiscal and monetary policy, and the
assignment of instruments to targets is not a matter of indifference, the
monetary instrument having a clear comparative advantage in affecting
the balance of payments and the stock of reserves.
These are well-known conclusions; this paper has merely tried to make
their basis quite explicit and to deal in more detail with the role of
capital mobility and non-traded goods. Much further work needs to be
done on these issues, notably, at an empirical level, on the role of size,
capital mobility, forward markets and non-traded goods.2 Finally, new
issues for analysis and policy arise in an inflationary world where all
goods are close substitutes; for, in such a world, fiscal policy of the
counter-cyclical variety loses much of its power to affect income and
price levels in small countries and the issue of control of the overall
inflationary process, and the role of large and reserve-issuing countries
therein, becomes of paramount importance.

Graduate Institute of International Studies, Geneva.

REFERENCES
[1] Courchene, T. J., "The Price-Specie-Flow Mechanism and the Gold-
Exchange Standard: Some Exploratory Empiricism Relating to the
Endogeneity of Country Money Balances", in H. G. Johnson and A. K.
Swoboda (eds.), The Economics of Common Currencies, 1973.

1 See Johnson [3] and Williamson [16].


2 For some interesting empirical work on the issue of controllability and
endogeneity of the money stock, see Courchene [1] and Willms [17].

This content downloaded from 128.42.202.150 on Mon, 27 Jun 2016 07:08:44 UTC
All use subject to http://about.jstor.org/terms
154 ECONOMICA [MAY

[2] Johnson, H. G., "The Monetary Approach to Balance-of-Payments Theory",


in M. B. Connolly and A. K. Swoboda (eds.), International Trade and
Money, 1973.
[3] -, "Theoretical Problems of the International Monetary System", in
Further Essays in Monetary Economics, 1973.
[4] Komiya, R., "Economic Growth and the Balance of Payments: A Monetary
Approach", Journal of Political Economy, vol. 77 (1969), pp. 153-70.
[5] Kuska, E. A., "The Pure Theory of Devaluation", Economica, vol. XXXIX
(1972), pp. 309-15.
[6] Meade, J. E., The Balance of Payments, 1951.
[7] Mundell, R. A., "The Appropriate Use of Monetary and Fiscal Policy under
Fixed Exchange Rates", IMF Staff Papers, vol. 9 (1962), pp. 70-9.
[8] -, "The International Disequilibrium System", Kyklos, vol. 14 (1961),
pp. 153-70.
[9] -, "The Monetary Dynamics of International Adjustment under Fixed
and Flexible Exchange Rates", Quarterly Journal of Economics, vol. 74
(1960), pp. 227-57.
[10] -, International Economics, New York, 1968.
[11] , Monetary Theory: Inflation, Interest and Growth in the World Economy,
Pacific Palisades, 1971.
[12] Sohmen, E., Flexible Exchange Rates, rev. ed., Chicago, 1969.
[13] Swoboda, A. K., "Equilibrium, Quasi-Equilibrium, and Macro-Economic
Policy under Fixed Exchange Rates", Quarterly Journal of Economics,
vol. 86 (1972), pp. 162-71.
[14] , "On Limited Information and the Assignment Problem", in E.
Claassen and P. Salin (eds.), Stabilization Policies in Interdependent
Economies, Amsterdam 1972.
[15] - and R. Dornbusch, "International Adjustment, Policy, and Monetary
Equilibrium in a Two-Country Model", in Connolly and Swoboda (eds.),
op. cit.
[16] Williamson, J. H., "On the Normative Theory of Balance-of-Payments
Adjustment", in G. Clayton, J. C. Gilbert and R. Sedgwick (eds.), Monetary
Theory and Policy in the 1970s, Oxford, 1971.
[17] Willms, M., "Controlling Money in an Open Economy: The German Case",
Federal Reserve Bank of St Louis Review, vol. 53, (1971).

This content downloaded from 128.42.202.150 on Mon, 27 Jun 2016 07:08:44 UTC
All use subject to http://about.jstor.org/terms

You might also like