Krugman Taylor (1976) PDF

Download as pdf or txt
Download as pdf or txt
You are on page 1of 12

Journal of International Economics 8 (1978) 445-456, 0 North-Holland Publishing Company

CONTRACTIONARY EFFECTS OF DEVALUATIOl+l

Paul KRUGW~N and Lance TAYLOR*


Yale University, New Haven, CT 06520, U.S.A.
M.I.T., Cambridge, MA 021.19, U.S. A.

Received November 1976,, revised version received February 1977

A simple model is developed to illustrate a number of contractionary effects of currency


devaluation, some of which have been noted prs,viously. In a Keynesian model, it is shown thaa
depreciation can lead to a reduction in national output if(i) imports initially exceed exports;.
(ii) there are differences in consumption propensities f:oG:.profits and wages; (iii) government
revenues are increased by devaluation, e.g. when there are significant export taxes. Similar
efYects are also shown to exist in monetarist models, via reduc:tions in both real balances andi
the nominal nzcney supply. A numerical example *llustrate:s the results for an economy
typical of semi-industrialized countries.

1. Inihduction
Theoretical treatments of currency devaluation generally conclude that it
stimulates economic activity. The ilnitial increase in th,e price of foreign goods
relative to home good.; is presumed to produce an exce:;s demand for home goods.
Models differ on how the system reacts, but in general home goods output,
domestic prices, or both rise. The possibility that price movements caused by
devaluation will create enough losers in real terms to cause an initial excess
supply of home goods is almost always left out.
This oversight persis& even though there rs substantial empirical evidence
suggesting that devauation often reduces aggregate: demand (vide Cooper
( 1971a)). Even a few theorists like Hirschman ! I949), Diaz-Alejandro ( 1963),

Cooper (1971b) and others have suggested that fatling output and employment
after devaluation are qtiite frequently to be expec:ted. These analyses, however,

*The authors are grateful to the referees ant1 to Ruciiger Dornbusch, Jagdish Bhagwati,
Edmar Bacha and members of the h4.l.T. Trade: and IkveloprnenI Workshop for comments
on previous drafts.
The possibiijty that dev,:iluation will produc:: a fali in out ?ut has been known to inter-
nationa. economists for many years, but theoretic4 treatments are t-in-e.!n the model developed
by Meade (1951), wkh WEISthe basis of most trreoretical analysts of devaluation. until t.le
mid-60s, devaluation could produce contraction only if rhe Marshall-Lerncr condition WC,:
not satisfied, sormething which was ruled1 out by assumption. Hirschman (1949) had poin!ed
out earlier that the MarsMl-Lerner condition must be n;,odifie#Jwhen trade is not initia.lly in
balance, and he argued that th:.s made al contractidnary effect from devaluatioq more likely
starting Ifrom an initial deficit. This point was late:,(. _~nfirmed in 2~ge:%ecaleyuilibrum model by
have had little impact on thinking about exchange rat.es, The prevailing view
is ckarly presented in ,the recent s~rnr%,~ryby Johnson I( S Cl%) of approaches to
devaluation : &zvaluatim can be expected to raise output if there are ur.employed
resources; to raise domestic prices if ehere arent.
Qur aim in this paper is to establish the plausibility of a third outcome - a
fa31in output. We attempt to do this by combining and extending the results of
previous authors in Q~K:simple, pedagogically appealing model. The model is
designed to bring out. clearly the income effects suppressed in conventional
approaches - for neglec,hingthe contractionary impacts olFdevaluation amounts
to ignorin% i~-t~::f: tffe&, especially those transferring real purchasing power
toward ecoflotnic a,ctorsswith high marginal propensities to save. By redirecting
income to high savers, clevaluation can create an excess of saving over planned
investment ex ante, and reductions in real output and imports ex post. The three
most important circumstances are t& following:

(i) When devaluation takes place with an existing trade deficit, price increases
OFtraded goods :mmediately reduce real income at home ;Ind increase it abroad,
since foreign currency payments exceed receipts. Within the home country the
value of foreign sa.vingLgoes up ex ante, aggregate demand goes down ex post,
and imports fall along with it. The larger the initial defiicit, the greater the COP;,-
t ractionary 0u.tcom.e.

$1 Even if fcjre,igc trade is initially in balance, devahration raises prices of


traded goods rel;;tise to home goods, giving rise to windfzill profits in export and
import-competing indus,tries. If money wages lag the p-ilce increase pqd if the
marginal propensi!ty to save from profits is higher than from wage:!, ex. ante
national savings goes up. The magnitude of the resu.Ling cc ntraction :+nds on
the difference between sal.vingspropensities of the two classes.

(iii) Finally, if Ihere a.re ad valcrem taxes on exports ar imports, devaluation


.
f&& ;&t;; frr~~rnthe private sector to the govel*nment, which has a
ir__?.>:??e

saving prope~~siry OT pity in the shori ii;;;. -?nce again,


a_ the final outcome is
reduciion in aggregate demand.

Casual empiricism suggests th;:t ali three circumst.inces prevail in many


countGes, especially the less deveioped ones. In these countries a deflationary

Coopr (1971b!. Daz-Alejartdro (1963) advanced another argument for contraction follo~+ing
devaluation, arising from the: redistributitjn of income from wages tcl profits. We are not awaire
of dny papers Analyzing the contractiona: y role of what we call the f ,scal effect of devaluation.
A word should. also be said about the extensive recent work on the monetary approach to
devaluation, of -ah ch Dornbusch (1973, is a good example. This approach does argue for P,
of contraction from devaluation - reduced real expenditure vi:i the effect of devalual.iDn
ces. Since thf.:s< models assume full employment cf resources, however, no
re dr kwn about the consequences for output and erqlloyment.
impacr from devaluaflon is more than a remote possibility; it is close to a pre-
sumpiion.

2. A anacroecomanic model
In this section we develop a simple Keyne+Kalecki model of an open economy
with the following characteristics :
(i) There are two distinct sectors, an export sector producing for the world
market and a home goods sector proclu.cing for domestic demand.
(ii) Prices of exports and imports are fixed in foreign currency ; home goods
prices are cletermined by a markup on direct costs of labor and imported
inputs required to sustain production (think off petroleum in an oil-shot t
country).
(iii) The wage rate is fixed in domestic currency.
(iv) In the short run, substitution responses of both exports and imports. to
price changes are negligible. Export volume is determined by available
capacity, while imports enter with fixed coeffncients into dome:;tic pro-
duction.
(v) Interest rates are kept constant by action of the monetary authority, ~11that
we need only consider income-expenditure relationships.
Assumptions (i)-(v) c,re chosen for analytical convenience, but the:; appear to
correspond fairly well to the stylized characteristics of many partially industri-
alized. countries. :ln these countries most export earnings come ifom an agri-
cultural or mining sector producing for world rnark-ZtP;.Domestic: industry has
been built up bj, import substitution via protectii,n, so that the remaining
imports are noncl3mpetitive.. chiefly interrnt:diz?.e goods and raw materials, for
which little substitution is palssible in the short r~.
The assumption of an accommodating monetary policy is made temporarijy
in order to allow us to focus on the incone effects of kvaluation. We will return
lo monetary analysis in a later sectjon. We begin with :Ln equation for the price
of ho&megoods : .

:n;-**+ cr)efiicients
where:aLtl, a,MHart, ll+y.. __ of labor and imports respectively into home
goods, w is the wage rate, Piti the E;.:.;medomestic price of in-lports. and Z a
markup factor.
Prices of imports and exports are determined by world prices, t;t?tes. and the
excizaLngerate :
where e is the exchange rate of domestic currency for dollars, i, and zlwthe ra.te!s
of ad valorem tax on exports and imports, and Ipi, P,$ the dollar prices of exports
and imports on world markets. Notice that (l)-(3) implly tlhat a change in th.e
exchange rate changes traded go4s prices relative to the wage rate and the price
of home goods, but does not affect the terms of trade.
Recipients of income may be divided into two classes : those who receive wages
and those who receive profits or rents. The nominal income of each class is
determined by the equations

YR = =(u,~,,,N~+~~~P,.,)H+(P~-~Lx~)X. (5)
Here H and X are outputs of borne goods and exports, gLxis the input of labor
per unit of exports.
For simplicity of exposition, it will be assumed that aiII imports are inputs into
home goods production, i.e. that there is no direct fin;~l demand for imports.
This implies that P,, is the proper deflator if we want to measure real income of
ti,orkers or capitalists. We will assume separate consumption functions for the
two groups. so the demand side ofthe model may be written

Here, Af stands for real imports. I is the interest rate, which we assume to be
held fixed, / is real invrstment, and G is real government consumption. For
convenience, define C,c-,ic7(
Y~;P,i) = yw, X,J7( Y,/P,,) = yx.
When the exchange rate is held fixed, equ*;aLIons (4)--(7) make up a Stan&r3
Keynesian open-economy model. It is a simple matter to co:mpute muitipliers on
home gaods production and iFTorts. The multiplier elects of a change in
go\rernment expenditure, for exmiple., are

dH 1 dM
-=- - = :,$fH,-10,
CG D dG

So tr this is familiar ground; tie unfamiliar results appt:a,r when the exchange
rate changes, rhifting the values of multipliers along with itself. The model
excludes, by assumption, substiwtion and monetary elkts of devaluation,
beawg only mcome etrects. The next section will examine jiust ihese.

r-act from stock changes. Ir. practice, a good part of the rrespome tu devaluation
ut to sketc uouM take place via inventory adjCJstirlent. Again, the unfaihqgly
for 5implicits.
P. Krugrnan and L. Taylor, Dcwalucafion 449

3. Income effects of &valuation


Even when dievaluation does not change a countrys terms of trade, it has a
number of other income effects. Unless the trade account is initially balanced, a
devaluatica changes the real income of the country as a whole. Within the
country it produces redistribution from workers to capitalists, and from the
private sector as a whole to the government. These real income adjustments do
not take place independentiy - they interact, and there is no way to decompose
the impact of a devaluation into separable components. In order to study the
different income effects of devaluation individuallly, it is necessary to consider
special case:s in -which only one r,bfthem is operating. That will be tlhe procedure
followed in this :section.

3.1. Devaluatiorzfrom (lzninitial trade imbulance


Hirschman (1,349)and Cooper (1971b) have shown - though with little impact
on other theoretical work - that devaluation from an initial trade deficit reduces
real national Income and may lead to a fall in aggregate demand. The argument
is straightforward. De,valuation gives with one hand, by raising export prices,
while taking away with the other, by raising import prices. If trade is balanced,
and the terms of trade are not changed, these price changes offset each other.
But if imports exceed exports, the net result is a reduction in real income within
the country.
The income loss can be quantified using the model of this paper by considering
a special case. Let yR = yw =: y, eliminating within-country distribution effects,
and let tX =::t, = G = 0, eliminating fiscal effects. After a good deal of mani-
pulation WCcan derive the elasticity of home goods output with respect to the
exchange rc,te (the result is g%en as an elasticity because the economic meaning,
of the:expression ii; clearer in that form) :

dki P PJ-P*M
-*.- = K*
de H P,H 9

where K ==(r/D) (1-(&$4/P&!) (1 f-z)).


Since &M( 1i-z) < I$$/ from equations (1) and (7), K is positive. Therefore,
output of home goods - and hence total oul put, employment, and imports - will
rise or fall depending on whether tra:;le is initially in surplus or deficit. Since
countries which devalue are usually in deficit at the time, there is contraction.
Its magnitude for a given percentage devaluation is proportional to the ratio of
the initia 1 trade deficit to domestic production.

3.2. Distributional eflkcts


Devaluation redistributes income from wages tc; profits and rents, for two
reasons. First, if money wages are rigid in the short run, devaluation lowers the
real wage as home goods prices go up in response to increased intermediate
port costs. Secondly, domestrc currency export recsipis ;.ilso go up, producing
windfall profits from the second _erm of equation (5). If, as is widely believed,
the marginal propensit!y to save out of profits is larger than the marginal pro-
pensity to save out of wages, this change in income shares will reduce aggregate
demand and therefore imports, as pointed out by Diaz-Alejandro (EM). We
again demonstrate by cons.idering a spe.cial case of the mc:jded.Suppose trade is
initially balanced, P,*X = P$M, so there is no trade balanl::e effect, and also that
IX = I!4 = G = 0. We z8ssumeyw > yR. We can solve, once again, for the elasti-
city of home goods output with respect to the exchange rate, which after some
substitutions becomes

where Y = Ywi- Y,! is total private income. Thus the elasticity of output (and
importsj with respect to dev;aluation is proportional to the difference in marginal
propensities to consum :, to the share of wages in income, aqd to th,e share of
imports in income. It is ~lso an increasing function ofthe marku]p. If consumption
propensities are equai, de4uation has no short-run effect on output, employ-
ment or trade, but mer: ;i Gifts income from wages to profits. The tradition.al
reluctance 3,i leftist gov zrnmcnts to devalue may have something to do with this
fact.

Theoretical models of the balance of payrnents ordinarily ignore the govern-


ment budget, bIut the f&al implications of devaluation may be of great practical
importance. There are 2.number of possibilities: if the government budget is not
initialiy b;rlanced, thercbis an income effect comparable to the income effect of
devaluation via the tracle deficit: if the..eare progres!;ive income taxes, or higher
taxes on profits than cn wages, the government claims an. increalsed share of
income; finally, if there are ad valorem taxes on exports or imports, higher
traded gods prices w II redistribute income to the govlerment. (One way of
losking at this last point is to say that rhe private sector pa:ys more for impor*ts
than it earns from expor?s, even though trade is balanced for the country as a
whole, so Ihe get ancthle - version of the wade balance effect ;[%lready
discussed.
To illustrate how 6scal reactions can make at devaluation deflationary, con-
of an export tax and assume tM = 8, *JR:= J,,,==y. Further assume
trade account and the g~nvernment budget a.re initially baianced,
x = P*~.~~ar YW+- Y, = cP~ X + ?,&l- dl$U. TP$renwe can solve
P. Krugttrmt md L.. Tnyh, Devahratim 451

for the result

The devaluation elasticity is proportional to the tax rate on exports and the share
of imports in income. Although the rnl;3delassumes a proportional tax function,
what is relevant in general is the marg,iinal rate, which may be very high. In one
fairly common case, where agricultural exports must be sold to the state at fixed
prices, the marginail tax rate is one, SC the fiscal drag from devaluation is quite
strong.

4. A numericalexample
It seems worthwhile to provide a, numerical example at this point, for two
reasons. IFirst, an efftirt to treat the general case of the model analytically pro-
duces very complicated algebra. Second, an example may help persuade th.e
reader that the effects we have been considering are in fact important, not
merely curiosities.
To produce a computable model recluires some, though not much, specializa-
tion of the functional forms. Assume that workers and capita.lists have pro-
portional consumption functions with constant consumption shares yiVand yR
respectively. Then (4) may be writen in the special form

H = Y~(~u@,,)fy,! Y,lP,)+I+G (6)

The equations ( l)-( 7) then form a solvable system.


The assumed vaiues of parameters and ex.:jgenous variables are given in
table I. The numbers chosen are arbitrary, though they are meant to fall within
a reasonable range for semi-industrialized coun-tries. 3
Devaluation increases the value of e while 1esJing all of the other parameters
and exogenol s variables unchanged. Suppose the currency is devalued by 25
percent. We can assess the effects by computing the values of some important
quantities for both e = 1.Oand c = 1.25. The results are shown in table 2.
Before devalua,lt;on the economy has a trade deficit of 8.4 percent of GDP -
large: but by no means uncommon. It exhibits all of the features that tie h.ave
<seencan make a devaluation deflatLonary - initial deficit, differential savings
behavior, ad valorem taxes on tradeid goods .- but none to an unusua.1 degree.
When the currency is devalued, there is a substantial deflation. Real GLP and

3For apph ca ti 0 n of similar models in practice, see the papers by Abel et al. (1976) on
Portugal and Taylor (8974) on Chile. The e:<amp!e here abstracts from much dehail on lax
systems, differential tariffs and export subsidies and so on, but its results are broadly similar to
those from the country studies.
Table 1
Assumed values of paramete;s and exogenous
variables.
-_.- - --
Li&H IO.75 x* 1
a.k4H 0.25 1

QLX 0.25 px* I


7W 1.0 I :!(I
?a 0.5 G 1.0
lx 0.5 X il5
CM 0.2 e 1.0
2 2.4

Table 2
Effects of a devaluation.
-_ ---
e = i.0 e = 1.25 % change
---- -I_-
Nominal GDP al factor cost 127.7 124.5 -2.5
GDP at consjam p-ices 127.7 119.8 -6.2
Prise of home gc ~0s 1.47 1.575 +7.1
0utput 43fhome ,;oods 102.7 96.0 -6.5
Trade b:jiance in dollars - 10.7 -9.0 + 15.9
Trade b&we in domestic money -. 10.4 - 11.2 -4.7

the output of hd)rne goods fall, while the trade balance improves in dollar terms
because imports decrease along with output. The loss in real GDP of 7.91 might
in practice be offset bl some export responsiveness to devaluation. However,
even on our unrealistic assumption that thie import content ot exports is nil, the
relevant elasticity wouId have to be close to two in the short run to restore GDP
in initia: prices to its pre- devahJation level. l[n a semi-industrialized country,
such a responsive export industry is lanlikely.
Finally, note that a;;gregate 2neasures behave quite difTerentiy in real and
nominal terms. The fald in current price GDP is less than half tthe fall in constant
prices, while the trade balance actually worsens when measured in domestic
currency. Phe differenL:e Mween reai and nominal movlements [aiso pointed
out by Hirschman ( 1949) and Cooper (197 1 b)] has ob\Gous importance for
monetary analyss of dee/aIuation, to which se now turn.

5. Manetary effects of devaluation

The analysis. of previous sections, with its purely Keynesian approach, may
seem dated and largely irrelevant to leconomists accustomed to the monetary
to thie bala$ncctof payments. It might be argued that the inco:ile effects
:;i be u~~mportar~t if the monetary authority, instead of
P. Krugman and L. Taylor, Devalubtion 453

pegging the interest rate, were to keep some monetary aggregate constant. How-
ever, if the possibility of unemployment is permitted in lieu of the usual monetar-
ist assumptions of full employment and flexible price Eevel, our analysis agrees
qualitatively with the usual models. Devaluation, by raising prices, increases the
demand for nomirral money at any given level of outpslt and employment. The
impact effect is contractionary, either more or less so than when interest rates
; tre held constant.

To illustrate the contractionary effect on the monetary side, suppose we were


to adopt an extreme quantity theory position, under which there is a strictly pro-
portional relationship between some monetary aggregate and income :

A = kc&+ u,), (8)

where A is a monetary aggregate fixed in the short run. Using equations (l)-(5)
and (8), we can derive the result

dH e
-*-= -k(PxX-+zPMM)/(A-kP,y:Y),
de .v

which will always be negative. Thus devaluation is deflationary in the short run
in monetarist as well as in Keynesian models.
Another numerical example may be in order. Suppose that the initial state of
the economy is the same as in the last section, but that now, because the central
bank holds M2 (say) constant, nominal GDP does not change following de-
valu.ation. The results of a 25 percent devaluation are displayed in table 3.

Table 3
Effects of devaluation hoIding nominal income constant.
_-

e = 1.0 e = 1.25 % change


-. -
iPeal GDP 127.7 122.9 -3.8
Output of home goods 102.7 98.6 -4.0
Tzde balance in dollars - 10.7 -- 9.7 +9.3
Yrzde balance in domestic currer: :y - 10.7 - 12.1 - 13.1
__P -- -

In this case the contraction resulting from ldevaluation under monetarist


assumptions is less than under Keynesian assumptions, but is still substantial.
A.11(jf this has assumed that the monetary authority really can determine
moneta:;y aggregate:;, something which is not necessarily so. In many countries
open-market ~?perationf; are not available, and th;e government must rely on its
own deficit :s:td - in rare instances - a balance oi payments surplus to create
new monetary bask. The identity in the absence of open-market operations (and
substitutes such as rediscount znd overdraft) is base creation = government
deficit + balance of pal cqents. rUow we have already observed that devaluati.on
will often increase government revenues through its effect on indirect taxes. We
have also seen that devaluation can cause the trade deficit to worsen in dome:;ti;:
c:rrrency although it im.ploves in dollars [as noted also in Cooper (197 I b)]. So it
is possible, even likely, that devaluati.on will lead to a reduchvz in th,e rate of
growth of the monetary base - an additior al deflationary influence. This res.ult
is exactly the oppo4te of what comes out of orthodox monetary approach
models, like that of Johnson ( t972).

6. Implications for p&icy

The purpose of tl)is paper has been to argue that, in the short run at least,
devaluation may noit work the way we usually assume; that 1:aken ibyitself it is
quite likely to have the presumably undesirable effects of shifting; the income
distributron against jabor and reducing employment and output. W:hat does this
do to our recommendations to countries with balance of payments problems?
Should we abanCon devaluation as a prescription because of its undesirab:le sicle
effects ?
The tlleorists answer - and he has a point - would be that the effe:crs $
devaluation or aggregate demand are irrelevant. Governments have other toots
with which they can manage demand. If they dont like the demand effects of
devaluation, let them compensate with fiscal or monetary policy, lea-&fin;
de\ tluat ion to accomplish its primary purpose of inducing substitution.
FractiGza: men would answer that ma:ters are not tha!t simple. Governments,
especially in less-developed countries, arc not sufficiently flexible to fine-!;une
their economic;. Thus one cannot ta.ke ii! Ior granted that devaluations \viill be
xcolnpe.nied by appropriate stabilization measures, and one therefore canraot
dl Ass cl:valuations demand effects.
There is a reasonable argument whiich 5l:artsfrom ,this point and continues as
fo!lous:

(i) In t!se shoi-t run the balance of paymems deficit is structural - that is, both
imports ar:d, exports are not very sensitive to price changes for a given level
of domes! 16output.
iii) As a con5equence, any favorable short-run elects of devalu;ttion 07 tht:
trade bal;l-:ce come primarily throw& economic contraction rather i.hial:l
substitution.
t iii) Devaluation not only reduces output and employment, but redistri Jutes
income frlDr?tlabor to capital as well.
ev84uaGon is a costly curl:, awd a devaluation big enough to re&lce
ieyments deficit subs;antialiy in the short fun may bc un-
e government should beg or borrow to meet
P. Krugman and L. Taylor, Ikmlrration 4%;

the short-term deficit and work toward eliminating its structural dificulties
by expansion of traded goods production in the medium run.4
The question is how one goes about correcting structural problems. In
economics whicla are closely tied to the world market, direct government in-
vestment is not likely to be too helpful. Governments can build and manage
roads, darns, and even steel plants; but there are few countries where they can
effectively ,1roduce wigs, or false teeth, or cosmetics, or peasant agricultural
products; yet these may be precisely the goods that the country has much chance
of exportikq or substituting for imports. Co a policy designed to expand the
capacity of ;he traded goods sector will probably have to rely on encouragement
of private investment. This can be accomplished with a variety of tools: subsidies,
tariffs, preferential credit, multiple exchange rates. It can also be accomplished,
without the microeconomic distortion, that these measures create, by devalu-
ation, which increases pro&ability in traded goods production. Perhaps, then,
one should think of devaluati(Jn as a measure designed to rectify balance of pay-
ments difficulties in the medium rather than the short run.
In challenging the establish.ed vieilr I-jfthe effects of devaluation on aggregate
demand, then, this paper does not deny its usefulness as a policy tool. It is impor-
tant, however, that policymakers be au ;u-eof its contractionary effects. Normally,
devaluation is regarded as an expenditure-switching measure, which should be
combined with an offsetting expenditure-reducing policy. Wha,t we have seen
is that devaluation Aelf may have an expenditure reducing effect. A stabiliza-
tion plan which, say, combines devaluation with tax increases may thus be
piling deflation on deflation, and the governvent may find itself confro;ated with
a steeper decline in output than it wanted. Devaluation should in many cases be
accompanied by measures to increase demand.
In any ca,se, it is not the purpose of this pa,per to give policy advice valid for all
countries at alI times. The important point is tha.t bevaluat.ion may be deflationary
and one should be COt.heailert for that possibility.

If one grants the proposition that in the short run there is little that less-developed countries
can or should do to reduce the balance of payments deficit, one must also grant an inqortant
corollary about the appropriate fiscal polcy when the external deficit is large. With investment
limited by all the factors which develo;)merrt economists sum. up under the rubric absorptive
capacity constraint, at full employment the government I, c forced to run a deficit to satisfy the
identity: investment + balance of payments = private saving+ government current surplus,
because the foreign deficit is so large. Far from being Inflationary finance, a government
dekit in such circumstances supports demand :or home goods against unavoidable leakages
of purchasing power aI>road through the trade gap.

_4be!, Andrew, Miguel P. Beleza, Jeffrey Frankel, Raymc.)nd Hill and Paul Krugman, 1976.
Tire Portugu ese economy: Current situation and proy,pects, mimeo (Banco de Portugal,
Lisbon) August.
Alexander, Sidney S., 11952,The effects of devalustim on a trade balance, Intc~rnaticmai
Monetary Fulnd. Statf Papers 2, X3-2781.
Cooper, Richard N., 197la, Currency devaluatic)n in developing countries, Essays in Inter-
national Finance No. 86, International Finance Section, Princeton University.
Coaper, Rkhard N., 1971b, Devaiuation and aggregate demand in aid-receiving countries, in:
J.N. Bhagwati et al., Trade, balance of payments and growth (North-Holland, Amsterdam).
Diaz-Alejandro, (Zarlos F., 1963, A. note on the impact of devaluation and the redistributive
effect, JournalI of Political Economy 71, 577-%%_I.
Dornbusch, Rudiger, 1973, Deva!uation, mcsney, and non-traded goods, tqimerikan Economic
Review 63,87 l-880.
Hirschman, Albert O., 1949, Devaluation and the trade bpjance: A note, r-review of Econom.ics
and Statistics 31, 50-53.
Johnson, Harry C.i., 1972, The monetary approach to balance-of-payments theory, Journal IaIf
Financial and Quantitative Analysis 7, 1555-l 572.
Johnson, Harry Ci., 1976, Elasticity, absorption, Keynesian multiplier, Krynesian policy, and
monetary apyiroaches to devaluation theory: A simple geometric el,positiorri American
Economic ReGew 66, 448-452.
eade, James E ., #951., The theory of imernational economic policy, I : The balance of
payments (Oxford I_niversity Press, Oxford).
Taylor, Lance, iSi(74,Short-term policy in semi-industrialized countries: <he narrow limits of
the possible, Journail of Development Economies 1, 85-104.

You might also like