Devaluation-Bias and the Bretton Woods System

Fee Nee ante mene

eal

(f675 in KFD Series)

INTERNATIONAL FIANCE DiSCUSSTOR PAPERS

DEVALUATPLION-BIAS AND THE BEETLON WOCDS SYSTEM by

Samuel] i, Karz

Discussion Paper No, 2, August 31, 1¢74

Division of International Finance

Board of Governcrs of the Federal Reserve Systen

The analysis and conclusions of this paper represent the views of the author and should not be interpreted as veriacting the views cf the Board cf Governors of the Federz) Poserve system or its staff. Discussion papors jm many cases are circuiated in preliminary form to stinulete Gisezsston éud conment and are not to be cited or quoted wirhout the permin~ sion cf the author.

creamer ae oN OF SARA ALP on Ne IEEE TER RY IRN RY RETRO ORE MERRET GRIESE “itt reer en ee

anes

ore ae, eta wa erates eee, NORTE ERT OTE TRS SAY

Ae tn 8 ae, EE OP Oe TOR OS TON

a etna aon rete mah meet hae ene

ee ed

Devaluation-Bias and the Bretton Woods System

Devaluation-bias under adjustable-pez? « « « « « «© «© wo ©

Exchange rate policies among the industrial countrics

Credits now available to deficit countries

Revaluation as an alternative to domestic inflation

European concern about domestic costs of external

‘ payments surpluseS . « 6 © « « «swe Is there evidence of a devaluation-bias?

Devaluation-bias: the statistical evidence .

Devaluation-bias against the dollar? ..

eo

Is the Hirsch-Niggias test an appropriate measure

devaluation-bias? eo e« «© © © © © © @

Have exchange-rate changes offset.relative

domestic prices? . « 6 « © © ew eo Consumer price comparisons rr er er) Wholesale price compariscnS ..«..«-ee«e Export price comparisons re er ee

Concluding observations ...-e.s-e-e-eeeve

TR Ser an Se Ne A PRIS BELA Oe ARRAY SN RE cm Mem grate hte en cm i aA pA ewe eam com

e

of

e

Exchange rates alone as a measure of devaluation-bias changes

in e e « e

e °

a wd

10 11 13 14

17

20 29 22 24 27

26

29

OE TS TE

Devaluaticn-Bias and the bretton Woods System Samuel I, Katz

The exchange-rate system established at Bretton Woods -usually called the par-value or the adjustable-peg system -- was intended to combine a fixity of the par-value of each member's currency in the short-run with a flexibility of the parity in the longerrun. That is, the member was expected to maintain the spot exchange rate within 1% on either side of the parity and to alter the parvalue only where its officials could dcmonstrate to the Fund that a state of "fundamental disequilibrium" existed. Because of this

emphasis upon the fixity of spot rates and of the parity in the short-

run, the Fund's par-value system can be regarded as a variant of

fixed exchange rates.

. The presumption that there is a devaluation-bias in present international monetary arrangements rests primarily on the hypothesis, widely affirmed in the standard literature in international econcmics, that, under any variant of a fixed exchange rate, “the more urgent

need for action and the bulk of the adjustment burden [is imposed] on

1/ _ deficit rather than surplus countries."" That is, the deficit country

(which is the reserve loser) is more likely to be forced unwillingly to devalue than the surplus country (which is the reserve-gainer) is to be forced unwillingly to appreciate. Furthermore, there is so high een erence

l/ Leland B. Yeager, International Monetary Relations (New Yerk: Harper and Row, 1966), p. 104.

-2-

a degree of substitution among the products of the industrial countries in world markets that an individual country can delay a decision to devalue (when its domestic prices have become no longer competitive at the current par-value) only at the risk of a significant loss of place in these markets. By contrast, a delay in revaluing by a surplus country after its domestic price trends have lagged far enough behind those abroad to make the existing parity cut-of-date is actually beneficial to the export industries.

As a result of these considerations, it is suggested, an industrial country is more likely to be forced_unwillingly to devaiue than it is to appreciate, When the country does eventually decide to alter the par-value, it is added, the authorities are more likely to make relatively larger changes in the damvard, than they would tend to make in the upward, direction. Let us assume that officials decide, in a situation of "fundamental disequilibrium," to alter the par-valve in accordance with international economic theory which assigns to the exchange rate the function of adjusting for differential rates of domestic price increases ameng trading partners. The devaluation~bias hypothesis suggests that they are likely to act differently ina devaluation than in a revaluation situation. When a devaluation is needed, it is argued, there are incentives for them to select a new par-value which is Lower than would be needed merely to correct existing price disparities. They usually choose a lower. par value on four grounds:

to avoid a second devaluation; to anticipate the inflationary feedback

Saad

ee OTE

-3-

‘of the devaluation process on local prices/costs; to obtain the support of the business community which will welcome the improved competitive position of exports in foreign markets; and to induce short-term capital inflow (on the basis of expectations that the country will be able to maintain the reduced parity). “These considerations are much altered in the case of a revaluation, In the case of an appreciation, the authorities can: a. Be less fearful of a second revaluation than the authorities in a devaluing country need be of a second devaluation; b. Select a lower increase in parity-to the extent that they anticipate the deflationary feedback from the revaluation; ce. Anticipate that the busiacss community (fearing domestic deflation and loss of foreign competitiveness) vill be more critical, the larger the amount cf the revaluation; and d. Expect capital outflows -~ not inflcws -- after revaluation, which would be larger, the greater the appreciation. a It is proposed in this paper to consider the validity of the presemption that a devaluation-bias cogs exist under the Fund's system of the adjustable peg. We will review the actual experience among Fund members in making changes in par-value and then consider the

analytical and statistical evidence which might throw light on this

experience.

ws ems

-4-

Soe nner

The argument that there is a devaluation- -bias under the adjustable-peg system is usually put in terms of the number and direction of exchange-rate actions. Since 1946, the changes in par value by members of the Fund have been one-sided: “the fact that in the past 25 years the Fund has seen 60 devalvations, and only three revaluetions, suggests movements upward may be even more sticky than

1 movements downward. ''=

~

But this evidence must be qualified in at least two major respects. First, the devaluations of the ma jor European and numerous cther currencies in Séptember 1949 might be excluded on the grounds that they wore unavoidable transitional adjustrents to the near-term effects of World War II: they were, in essence, diffecential rates of revalvation of the dollar agzinst the principal European currencies. Cn the other hand, they would reflect a devaluation bias in those cases where the amount e£ devaluation could be regarded with hindsight to have been excessive.

Secondly, by count, most parity actions ameng Fund members have been by the less-developed countries (the LDC's) which are widely known to have been resisting the decision to devaluc. Woodley has

characterized the exchange-rate policies of these countries in these

ere ee cn A,

1/ Speech by U.K. Chancellor of the Exchequer Roy Jenkins in Sunmary Proceedings, Twenty-Fourth Ann Annual Meetins of Board_of Governors “(September 25 = Ccteber 3, 1569) , “Taternational Monetary Fund, Washington, 1969, p- 38.

w-

-5-

words: "the principal policy question ... is why less-developed : ) countries almost consistently err on the side of maintaining overobs . valued currencies. 7 In practice, the less-developed countries have often been reluctant te devalue their currencies because of the high domestic political costs of such a decision. In addition, officials

in the LDC's have often come to regard the gains accruing to their

particular country from the terms-of-trade benefits of an overvalued

parity to exceed any improved craderbalance gains which could be

expected frem a lower and more realistic exchange rate. If there is any evidence of a devaluation-bias in wor ld_payments arrangements , then, it will have to be sought not among the LDC's but only in the exchange~ rate decisions of the small group of highly-compet itive industrial countries.

“Exchange rate policies among, the industrial countries - Between 1960 and mid-1971, there were among the industrial countries four decisions to devalue (Canada in 1962; the United Kingdom and Denmark in 1967; and France in 1969) and five to revalue (Germany and the Netherlands in 1961; Germany in 196%; and Austria and Switzer land in 1971). (See Table 1.) In addition, the Canadian authorities in. 1970 and German and Dutch authorities in May 1971 permitted their currencies to float; each of them was quoted above the former parity

ee neem

lf W. John R. Woodley "Some Institutional Aspects of Exchange Markets in the Less-Developed Countries" in The International Market for Foreian Exchange, edited by Robert Z. Aliber (New York: Praeger, 1969), p» 177.

Table 1

The Balance Between Revaluations and Devaluations Among Industrial Countrics, 1960-71 (in per cent)

Date and Country

1961 Germany Nether lands 1962 Canada 1967 United Kingdom Denmark 1969 Trance Germany 1971 Austria Switzerland c/

Source: 1s 1979, Pp.

a/ In year of pa

1960-69, The Role of FE» International Payr ents, 39; 1971, Internationa

Change in par value

(1)

Country share in exports of industrial

countries a/

— -10,-4— 1.8

8.3 16.1

Weighted change in par value

(1) x Q) 100

(3)

to. or mw SD fy

4 ~ e

uw =

Net weighted change in par values of industrial countries, 1960-69

+ 5,05 + 7,07

1.37 2.46

- 0.89

+ 0.07 + 0.17

Net weighted change in par values of industrial countries,

based on 1970 trade values.

b/ Par value adopted in May 1962, exchange rate in January 1990.

“Tateria tiona

1960-71

+ 0,55

ee Ce ead

hhange Ra Rates_ int the Ae Adjustment of al Monetary Fund, Washington, 1 Financial Statistics.

a ancora nT TE OE ad

rity change, except for 1971 computetions which are

compared with level of floating

~ ¢/ Not a member country of the International Monetary .Fund.

an - .

7 7-

on July 1 (1971). In magnitude, the devaluations ranged between & and 14 percent and the revaluations between 5 ‘and 9 percent.

| One way to summarize this experience js to weight the amount of parity change by the share in exports of the industrial countries supplied by the country whose parity is altered. On this basis, largely because of the heavy weight given to the two German revaluations, there is a near-balauce between devaluations and vevaluations: the net weighted change is -0.89 percent for the pericd from 1950-69 and -9.65 percent if the Austrian and Swiss revaluations are included, (See Table 1.) This negative sum would have been_increased if the cbserva~ tion period had been extended to include. the French devaluation in 1957-58; on the other hand, it would be reduced if there were any allowance made for the effects of the rise in the Canadian dollar in 1970 or the temporary de facto rises in the market values of the Gerran and Dutch currencies at mid-1971.

This position of near-balance in exchange-rate decisions as between devaluation and appreciation would seem to be evidence contrary to the existence of a devaluation-bias in current monetary arraigements, This denial of devaluation-bias on the basis of. this exper fence is reinforced by arguments along two general lines. ” -

Credits now available to deficit countries = In the first place, it can be suggested, this hypothesis does not take adequate

account of the access to official credits now available to deficit

countries which have been developed, especially over the past decade.

eens cep et vee etme — ae i

-B8-:

This argutsent has been stressed by Russell:, ‘The common assumption

with respect to deficit countries see that they cannot permit their

reserves to decline infinitely oe- neglects the possibility of infinite

borrowing of reserves. "2! It cannot be doubted that the possibilities

of official borrowings which were deve Loped during the 1960'

s have

helped to allocate the adjustment burden somewhat more evenly between

deficit and surplus countries. It is no longer realistic to assert

country on a pr ori grounds.

unequivocally that the burden of adjustwent falls mainly on the deficit

But there are limitations to the access which deficit countries have to external credits and their continuing use of such

, financing cver @n extended period has proved to be a costly strategy.

Consider the case of the United Kingdom as an example. In connection

with credits from the Fund in 1964, for example, the U.K. Chance Lior

supplied a letter of intent which “detailed the policies that the

— United Kingdom was pursuing, and would pursue. "=! Even ‘more specific”

qe eS

Books, 1970), PP» 76-77.

Vol, 1, Chronicle, (Washington: Lu.F., 1969) pe 572.

Soe Ninna . een a

commitments were accepted in a May 1965 credit arrangement ond again in

November 1967 and there were close and cunt inuing consultations between So

1/ Robert W. Russell "Hultilateral Surveillance, Consultation, and * the Adjustment Process" in The Future of The International Monetary

System edited by Hans W, J. Bosman and Frans A. Me Alting von Ceusau, Publication of the John F, Kennedy Institute Center for International Studies, Tieburg, the Wether lands (Lexington, Masse: Yeath Lexington

2/ J. Keith Horsefield, The Internationa) Monetary Fund 1945-1965,

Bat.

be ® y

-~9-

the U.K. and the Fund over current U.K. economic developments while the credits were being used. In addition, the Prime Minister gave as a reason for the decision to revalue in his television address to the British nation on November 19, 1967, an unwillingness to try "to borrow this time in conditions in which our creditors abroad inight well insist on guarantees about this or that aspect of our national policies."

As it turned out, it was not until duly 21, 197) that the U.K. Chancellor of the Exchequer could inforn: Parliament that the Government ‘will act

now have to consult the staff of the Ii about the progress of the U.K.

“economy'' because “the obligation to consult the Fund has been removed

as a result of the accelerated debt repayment to the Fund announced

te last week,

In practice, the Fund has been able to establish effective cormunication with borrowing countries conceraing their current econecmic

Sag 2/ 1 : * .

policies. By contrast, the Fund's Managing Director has asserted: "st is clear that the influence which the Fund or any other organiza~ tion can exercise on a surplus country is Limited. "2! So far as surplus

countries are concerned, their compliance with the Fund's code of

‘ recent

i/ IM Morning Press, July 26, 1971 quoting from a report in the Daily Telesraph (London} of July 21, 1971, pe 17.

2/ See my review The Internaticenal Monetary Fund, 1945-1965 in The

— ——

Journal of Finance, December 1970, especially PP. 1218-19.

3/ ‘Pierre-Baul Schweitzer, "Stamp Memorial Lecture," London, December 2, 1969 (mimeo) pe 9

eon ergo merippercaigens ace 4 Teemenmamsee nm egg N pon ae en anna

‘s - 10

behavior has depended largely on moral suasion and upon the country "Ss

perception of its own longer-run self interest. For this reason, the

greater availability of international credits and since 1970 the

initiation of Special Drawing Right allocations through the IMF may have helped to bring more balance in allocating the burden of adjust-

ment between surplus and deficit countries, but they have not removed

entirely the unequal constraints upon unwilling deficit countries to

take corrective action which have always been charecteristic of a fixed

exchange-rate system.

Revaluation _as_an alternative fo domastic inflation

A second argumant against a devaluation-bias hypothesis rests

on the incentives which surplus countries have in a highly inflaticnasry

world economy to make use of revaluation on purely domestic considera-

tions: to shield the internal economy from the "imported inflation"

preduced by large balancc-of-payments surpluses. The record since

1959 found in Table 1 shows five instances of appreciation a among the

major industrial countries as compared with four instances of devalua-

tion. (See Table 1.) In historical terms, this frequency of revalua-

tion can be regarded as little short of remarkable. For "before the

advent of -the Bretton Woods system eee explicit revaluation of a

currency was extremely rare."

Rates in the he Adjustment _ of i international

———— - Lf ‘the Role of Exchange Washington, 3970), p. 38.

Payrents ts (internaticnal Monetary Fund:

i

"4 . ee

\

ee een em RRR RE SEE RN ER be a at RN ema Ee

‘ -ll-

The several instances of revalsation undoubtedly reflect an overriding concern in industrial countries about the domestic economic costs of “imported inflation" produced by large balance-of-payments surpluses. This concern contrasts with the period of the 1630's when deficit countrics experienced pressures not only {rem losses of external reserves but also from internal derands for currency depreciation as a support fer employment and domestic-inccme goals. In the prevailing environment in which countries have been facing inflationary pressures

and excess demand instead, the recent Fund report on exchange rates

“motes that;

"exchange adjustment will more often contribute to domestic stabilization in the countries whose external payments positien permit an appreciation of their currency, rather than in the countries whose external pcsitions require a depreciation." (p. 38)

European concern about domestic costs of external payments surpluses - It is striking that the major country which has perhaps the greatest sensitivity to domestic inflationary dangers -- Germany -~ has been the chief proponent of revaluation as a means of protecting the internal economy from external inflation. The change in attitude about revaluation expressed by the late President Blessing of the German central

bank in the late 1960's was prompted largely by a concern about mini-

a

mizing the domestic effects of "imported inflation." He stated:

"It has been asked whether it would not be more appropriate for the sick to devalue than for the healthy to revalue ... Until a few years ago, it

ere

et, Mla alts ae ltl

. ~12~-

had been my opinion that the sick ought to undergo an operation and not the healthy." i/

ra

It was only gradually, after he had faced the domestic inflationary effects of Germany's enormous export surpluses over a period of years, that Herr Blessing caine to change his mind. He was led to recommend the revaluation of the DM in 1968 because, in his words,

"T have since been forced to admit that we live

jn a world which is no longer «+e prepared to

accept really severe disin£laiionary measures, . and that the healthy can preicct himseil against

inflation only by means of a change in parity."

By contrast with this view, President Holtrop of the Dutch central bank has argued that there is an asyirmetry in the allocation of the economic costs of the adjustment-burden as between surplus and deficit countries in the post-war exporience. In his view, the “surplus” countries «.. generally lived up to the prescription of the Brookings

Institution report and allewed their economics to be inflated by their

2/

surpluses without putting up too much resistance." To the extent

that this asymmetry reflected an unwillingness of surplus countries to appreciate, even when experiencing balance-cf-payments surpluses in a

period of excess internal comand, it could be heid to support, rather ec nee

l/ Herr Karl Blessing, Fresident of tne Bundesbank, speech before the German Cooperatives at Kainz, Germany, on October 10, 196%.

2/ M. W. Holtrop, "Phe Balance of Payments Adjustmont Process, Its Asymmetry, and Possible Consequences for the Internaticnal Payments System" in Approaches to Greater Flexibility of Exchanse Rates: The Burpenstcck Pupers edited by George il, Halm (Princeton: University

ress, 1979), p» 158.

od

1 ne ge enero iene eee eevee eens i mt

‘ -13-

than to deny, the existence of a devaluation-bias in the international monetary system. It certainly cannot be used to reject that hypothesis. Is there evidence of a devaluation-bias? - There have been at least two major recent developments which have tempered the built-in incentives in any system of fixed exchange rates to place greater constraints to adjust on deficit than on surplus countries. But neither of them can be regarded as grounds for concluding that there is no devaluation-bias in internaticnal monetary arrangements. The greater possibilities that deficit countries have. to borrow do not enable them to avoid adjustment indefinitely since foreign credits are often available only on the basis of conditions laid down by the lender, whether it is one or more countries cr an international institution, Similarly, @ concern about inflation has encouraged surplus countries to reva ive on a scale which can only be regarded as unprecedented; but these decisions have often been delayed to an extent which can only be regarded, in retrospect, as excessive. If the traditional analytical basis for the hypothesis that there is a devaluation-bias in any fixed-rate system has been ereded, it has not been altogether

invalidated by the post-war experience. Accordingly, let us review

‘the statistical evidence to see whether it can throw light on whether

a devaluation-bias does, or does not, exist in current monetary arrange~

ments.

-4-

Devaluation-bias: the statistical evidence

‘ Perhaps the major attempt to approach the question of devalua-

. tion~-bias in current international monetary arrangements through an analysis of the statistical evidence has been the work of Hirsch and Higgins on “offective™ exchange rates.2/ In their article, the authors distinguished between actual changes in parity (or (nominal adjustnents) compared with the "effective adjustments which measured the effects on each country of “changes in the exchange rate, as customarily expressed, of other currencies, whether these changes are large or small, and whatever their timing." (p. 453.) As a technical matter, "effective" exchange ratcs were calculated for 14 industrial countries on the basis of the formal (or ucminal) parity changes for each of them minus the weighted indirect changes produced by parity adjustments made by tradin partners. These adjustments were intended to allow for economically significant movements in the parities of a country's trading partners as they might affect the international competitive position of each of 14 industrial countrics. Accordingly, the computed "affective" exchange rate was intended to measure the "iinpact of the concurrent nominal rate changes for the other currencies'’ on the internation2l pesiticn of

each country.

a RR

1/ Fred Hirsch and Tlse Higgins “An Indicator of Effective Exchange . Rates" IF Staff Papers, November 1970, pp. 453 - 487.

Pe ee a RY YE eet aah eee eer een

table 2. Industrial coumerics: Chances in eCfective

exchange rates, in dollar parities, and in crficial reservess 1959-1969

ee ene

Actual ‘ ; Cumulative changes Proposed Official reserves changes in in test of (in billions of dollars): _ effective dollar devaluation December December Change parity l/ parity bias 2/ 1959 — 1969 in period 2. Other Countries with devaluation-bias: 2/ Germany +17.3 +14.8 42.5 4.8 7-1 42.3 Japan +O.9 -. “+ +0.° 1.4 3.7 +2.3 Norway + 0.6 -- +0.6 0.3 0.7 +0.4 3. Countries with revalvation-bias: 2/ cn Italy - 1.0 -- -1.0 3.1 5,0 +1.°9 Belgium - 1.3 -- “3.3 1.3 2.4 +i.1 Switzer lard oe LG -- ~1.4 2.1 4.0 +1. 9 Denmark - G1 - 7.2 “1.9 0.3 04 40.1 ‘A. United Kingden -13.5 14.3 +0..8 2.8 2.5 -0,3

Source: (1) Changes in effective parity and in dollar parity: Fred Hirsch and llse Higgins "An Indicator of Effective Exchange Rates" International Monetary Fund Staff Papers, Hoverber 197, Table 3 on page 473. (2) Changes in official reserves: International Financial Statistics, International Monetary Fund. The eetimates include: Gold, SDRS reserve positions in the Fund and foreign exchange.

o |

?

o

1/ Wirsch-Wiggins define the change in effective parity as “the percentage “@irec change in its numeraire rate minus the weighted percentage ‘indirect! change in the numeraire rates of other currencies." Because the U.S. dollar serves as the aumera i. currency, its effective rate is affected only by the indirect effects of other parity chenges. The effects of parity changes are calculated for the 14 industrial countli. specified in International Financial Statistics. The authorge have calculated an ines. of effective exchange parities for cach of them; they applied to the direct changes in parities the indirect effects calculated on the basis of weights which "reflect

the share of cach ef these countries in the given country's exports of manufactures to and imports of manufactures fron the 13 other countries combined” (p. 459). The

- formula for the index is specified in footnote 3 on page 455 and in Appendix I on

- pages 47°-480.

2/ The authors have proposed that “a devaluation bias is present if there is a positive difference between the change in a country's exfective parity aud aay parity change of its sun, '' (Footnote 12, page 474.)

- 16 -

The body of the Hirsch-Higgins article was concerned with the calculation of the "effective" exchange rates for 14 industrial countries between 1959 and 1969 and an exploration of the significance of. this novel and useful concept. As a by-product of this work, houwever, they reported that "the movement in the effective exchange rate js also observed as an indicator of whether, from the standpoint of particular currencies, the changes in parities of other currencies have involved a devaluat joa bias cr a revalyation bias" and were led to assert: “contrary to some general impressions, no general devaluation bias is found in the system as a whole." (p. 454.) It is the statistical evidence on which their comments on devaluation-bias are’ based which concern us in this papere

The authors have proposed an explicit statistical test of devaluation-bias on the basis of a coxpariscn of the "effective" and the "nominal" changes in parity for each of the industrial]. countries found in Table 2. The steps taken by the authors in calculating the "effective" exchange rates are sunmar ized at the bottom of Table 2 and explained in detail in their article. In tzble 2, the "effective" exchange rates in the first column are compared with the changes, in nominal parity by each country in the second column: the difference between them, shown in the third column, constitute the test of

devaluation-bias proposed in the Hirsch-Higgins article. In their view, there is a devaluation-bias when the changes

in effective parity exceed the changes in nominal parity and a

-

Seem eno ng teem tN A pee ON nee ine arn RNR ARAN Sate ene RTI SCR SS meRN anes emg ne

: | -11- :

revaluation-bias when they are negative. That is, there is a devaluation-bias when the effective parity value of the currency is higher than the announced parity change and a revaluation-bias when the effective value is less than the parity change.

By this test, there is a devaluation-bias in substantial amounts against the United States, Germany and Canada and in amounts of less than one percent apainst the United Kingdom, Japan, Sweden and Woerway. Fer the cther seven countries in Table 2, there was a revaluation-bias by this test. The denial by the authors of evidence

of a devaluation-bias in current international financial arrangements is based on what appears by their test to be an even-balance as between

devaluation-bias and revaluation-bias among these countries during

the observation period.

a ee re ne er en NES Ane ARS

Higgins calculations also show that there was -- by their test and on

the basis of the concept of the effective exchange rate -- an effec-

tive appreciation of the dollar and a devaluation-bias in the system vis-a-vis the dollar of 4.7% between 1959 and 1959, (See Table 2.)

The United States (as the numeraire currency in the system) was passively affected by the indirect impact of both revaluations and devaluations

of the other 13 countries. As a result of all the parity changes made, the dollar was effectively revalucd by 4.7%. Furthermore, the deva)uation-bias in the system vis-a-vis the dollar was nearly double the 2.5%

bias against a strong surplus country like Germany.

-

This appreciation of the dollar should be recognized for what it is: a purely statistical statement of the effects on the United States of the various changes in parity by each of the other 13 industrial countries. These effects are measured by the difference in each case between actual (effective) and the intended (nominal) parity changes for each of them

At the same time, it must be noted, this upward thrust in the dollar's effective competitive valuation came during a period when the United States was losing, and the other industrial countries as a group (excluding Britain) were gaining, reserve-assets at an unsustainable rate. (See Table 2, Column 6.) Seven of the reserve-gainers experienced a revaluation-bias by this test (that is, the effective exchange rate changed less than the actual changes in their dollar parity) and five of them experienced a much smaller devaluation-vias than did the United States. The question must therefore be raised: were these results evidence of a general devaluation-bias in the

system as a whole or merely of a devaluation-bias against the dollar?

The authors regard their results as evidence that there is

“no general devaluation-bias in the system as a whole. If we consider

the changes in the effective parities of the seven countries in Table 2

which made no formal parity adjustments, we find that there were four

cei ili had

which made no formal parity adjustments, we find that there were four cases of revaluation-bias by this test and three cases of deva luat ionbias. Had there been a devaluation-bias in the system as a whole, the argument would run, the changes in effective exchange rates (for countries which made no change in their own parities) ought to have been primarily only in one direction and not been so evenly-balanced. However, there remains the fact that there was a devaluationbias against the dollar of 4.7% by the Hirsch-Higgins test and therefore the question: even if there were no general devaluation-bias in the system as a whole, was there evidence of a devaluation-bias against the dollar? On this point, the authors argue that the devaluation- ‘bias against the United States nasa not result from any general devaluation bias in the adjustment of par values." (p. 474) Instead, the effective appreciation for the United States "reflects rather the particular orientation of its trade, and particularly its large trade with Canada." In their denial of any general devaluation-bias, Hirsch- Higgins point out that there would have been a devaluation-bias against the United States of only 0.5% for this period had a different system of weights been used (pp. 474-5) and, in.addition, that "the effective appreciation of the U.S. dollar in this period, on either basis of weighting, falls well short of the apparent relative appreciation in

its internal value." (Footnote 14, p. 475.)

- 20+

Is the Hirsch-Hiegins test an appropriate measure of devaluation-bias?

The main thrust of the Hirsch-Higgins work is to measure the cross-impact of parity changes by 13 trading partners on the effective exchange rate of each of 14 industrial countries. By themselves, these calculations were not intended to throw Light on the question of devaluation-bias, and the conclusions in the article about the existence of devaluation-bias are only a by-product of their ccmputations. The question must therefore be raised: how appropriate is the statistical test of devaluation-bias which they have proposed? Furthermore, is it altogether consistent to attribute their findings of devaluation-bias for the United States only to the pattern of U.S. foreign trade (rather than to any degree of devaluation-bias in the system as a whole) when their findings for each of the other 13 countries are also determined by the particular pattern of foreign commerce of each of them? Exchange rates alone as a measure of devaluation-bias - The authors conclusions are based upon their proposed test of devaluationbias. Because their results show that seven countries experienced yevaluation-bias and seven devaluation-bias, they conclude that there can be no tendency toward devaluation-bias in the system as a whole. (See Table 2 and the Hirsch-Higgins article, footnote 12, >. 474.) It is striking that the Hirsch-iggins proposed test is made up of a comparison of the differences between two values of the same variable: ‘that is, the explicit ("nominal") and the actual

(effective) changes in the exchange rate for each country. But

oll ME ai iy nas,

»~ 2] =«

exchange rates changes among a group of countries ought to have a built-in tendency toward rough balance; after all, exchange rates are merely price ratios and any change in the value of A‘s currency necessarily alters the value of B's currency in the opposite direction. In terms of the Hirsch-Higgins computations, a revaluation of currency

A will produce a relative devaluation for cach of the other 13 currencies in proportion to the trade-value weight selected; conversely, a devalustion will produce relative revaluations of the other currencies on the

same basis. In the case of Austria, for example, there was a deva lua-~

tion of the "effective" exchange rate for the _schilling between 1959

and 1969 (even though there was no change in its parity) merely because

-the German mark had been revalued. A tendency for changes to be roughly

in balance as between revaluation and devaluation when effective

changes in exchange rates among a group of countries are compared should

, not be regarded as altogether unexpected; on the contrary, purely on

a_ priori ‘grounds, a tendency toward one-directional movements in these rates would be a surprising outcome.

The main objection to the Hirsch-Higgins test, from an analytical p: point of view, is the absence in it of any indicator of changes

in the internal value of each of the currencies. By concentrating

exclusively on two measures of external value, the test appears to

treat changes in relative exchange rates as a phenomenon entirely

“ {Independent of developments in the internal economies or even in the

balance of trade or payments of the group of countrics being studied.

~~ —_

apenas

7 no ease pennie eter empties psa

ae elena.

- 22 -

Have exchanee-rate changes offset relative changes in domestic prices? - Perhaps a more broadly-based test of devaluation-bias would compare the changes in the external and the internal values of the several currencies. This test would be closer to the traditional concern in international economic theory about the function of changes in exchange rates as the means of adjusting for differential rates of domestic price increases among trading partners. Under this approach, the question would be posed: to what extent have the changes in external values of these currencies between 1959 and 1969 served to offset relative changes in their internal values | dur ing this period?

Interestingly, the authors make such a comparison for the United States against the other 13 countries, but they did so only in passing in a footnote digression. They found that "the effective appreciation of the U.S. dollar in this period ..- falls well short of the apparent relative appreciation in its internal value, if the latter is measured by the comparative increase in the consumer price index in ‘the United States against the weighted increase in the other industrial countries." (Footnote 14, pp. 475-76.) Cn the basis of the consumer price index, prices outside the United States rose by 45% between 1959 and 1969 compared to a rise of only 34% in the United States. Thus,

there was "excess inflation outside the United States of 11%" on the

‘basis of this calculation. - re vee ate pe tees

This alternative test of devaluation-bias needs to be

explored more fully. For this purpose, price relatives for each of

~

- 23 ~

the 13 other industrial countries visea-vis the United States have

been computed on the basis of:

Py (69) Pug (59) ; ——— KO index

Pys (69) Pp (59)

where F is each of the 13 other countries in turn. By this measure,

the index will be ¢. 100 when U.S. prices have risen more rapidly (and > 109 when they have risen less rapidly) than those in country F during the observation period. Separate indexes have been computed for: |

- Consumer prices;

- Wholesale prices or home- and imported goods prices; and

- Export prices or average (unit) values. The second step is to measure, on the basis of 1959 = 100,

the change in the dollar value of each F-currency in terms of the

dollar on the basis of

where X is the dollar-value of the F currency. The exchange-rate calculation will yield a value > 100 only when the dollar-value of

the F-currency is higher in 196° than it was in 1959.

2 it cea, . wt ‘ « Oh

We can then compare the*changes in the external and internal values of each F currency against the dollar by combining these two

components into a form of competitiveness-index in which

Pp(6S) Pys (59) Xp (69) Pyg (69) Pp(59) p59)

t 6K 2 Oo

when the change in external value exactly offsets the change in internal value, An improvement in the relative position of the United States vis-a-vis country F would be demonstrated by an index-value > 100

since it would be mean that relative price movements abroad have been

greater than the changes in the dollar-value of the F-currency. A

--geterioration in the U.S. relative position (marked by an index-va Lue

£100) would signify. either a higher rate of inflation in the United States than in country F or a depreciation of the F-currency against

the dollar.

‘Consumer price comparisons - The computations in Table 3

.

demonstrate that the three price measures produce contradictory

findings. Accordingly, a judgment from this evidence that there is,

or. is” not, a devaluation-bias against the dollar depends upon the

particular price measure which is chosen, By our calculations, the

comparative changes in the externa l-versus- -the internal values of the

dollar vis-a-vis each of the F-countries: (a) improves substantially

on the basis of the consumer price indes 5 &) improves marginally on

the basis of the “wholesale price index; and (c) deteriorates substantially

ood \

“kayunos vord fq pa zcdoa sv s3z0dx0 jo onpeA (TUN) aBervae x0 odfad-azodxy /Z% *mopSut oytuq fgndyno TeTaysnput 40 sootad {Aucwia9g ©syonpoad [etaqysnput 10: saotad p3uty peatun Ter t 3 T BS t _ dT esdTOyn funtSytoq SAT esl ‘ucdep ‘epcurg 20j gdotad opTesatoum s:sMoT [OF se ydaoxo saotad yrodwr pue suo} /t

“OL61 oquoson ‘siddeg FyOIs AW T (OOF IAe suplSTy-yoszTH Wory $eaPs aBucyoKXa IATZIVzJO

ut pue Aqtacd ur sesury) “d*N I ‘SoTast yess Tetoueuld [euot eudazuy UT pozzoder se eyep VIFAg 299A1N0S

G*°S6 6 °EOT T°Ctt Z°SS G6 °ECT Z°CIT (sax0dxe 1961) perystem ‘seseraay °S C°€S Z°7CT °C 6°26 T “SCT S*utt | payystenun ‘soseray *7 ve . eee $°96 2°66 1°95 9°S5 €°85 8°S6 swopsuTy pa3tun °€ 5°OL C*€ecl $*0T c°18 g°g0T- €°Ort ° epaqsny €°Ss L°86 T*CGT | 6°15 2°20 L°ect aourld . . S*L8 9°COT 9°O2T 6°58 8 °ZOT zwezt . . a — yarewued 8°60T omer aan T°9ct 7°1Il 4° ECT 9°21 pueyTiszatAs £°S6 Z*ect S°IGT o°L6 9°4CT g°ccl- wntTSTog 7°C6 € LCT ¢*ztt €°1S 7°8CT 9°CIT =. . Ay eat . €°C6 C°L6T (omy A ana) - €°SGT g° ctl | Spuey 29439N , ESUTQ-UOTJENT LAST YITM Sot zqUnoD > An . £°66 6°SiT T°Stt €°56 H°STl - 9*HTt UdpPaMs : z°88 °601 Z°Ztt L£°L8 L°3ot S*lIt - Avion - ce 8°o8 9°66 6 *SET 0°98 £°85 L°vet 7 ueder 6 °€6 8°86 6°1S L°t 9°L5 8 °C6 epeueg SO £°“01 8 °6CT L°Stt ¢*zct g*LO1 9°CtT : _ fAuewre9 7 . 2 SBT Q-UOTI ENT eASD YITA sot xz,UNndD *T . i Saotad SdoT ad Sootad ‘Sootad Sootad sodotaid : Jiodeq Spesopou Wwnsu0d yiodNq +«spesoyoyy swnsuod |

rdgea oouvyoXa ne

DAT JOO; FO UT SosueYys uO posed

rae Uy Sesueyo [ewIoZ uo paseg

(CGI = 665T) 5951 03 6661 ‘Setz3UNOD TePrISNPUT ET LOJ S3REIS POITUH SyI SPA-e-STA gosurys oZer-oSueyoxe pue Vd} *€ BTqQeL

men enapatary enti en Mem

~ €Z -

cet dig ere ee ee

-~ 26 -

on the basis of the export-price (or unit-value) index. These results are consistent with the Hirsch~Higgins calculations in that the changes in the external value of the dollar were less than the changes in its internal value as measured by the consumer price index. The difference between the average excess inflation outside the United States of 10% in Table 3 and the 11% reported by them is probably to be explained by differences in weighting and/or in computation.

However , there are doubts about the validity of the CPI as a realistic measure of relative price trends for international comparisons. This index is usually rejected because tr has so large a services component and because it includes so many products which do not enter into international trade. Hence, changes in the CPI have been of only limited value as a measure of the comparative changes in the local currency's internal value for international purposes. For example, Junz-Rhomberg did not use it at all in their study of prices and export performance of industrial countries; instead, they considered ‘two alternative price variables (export unit values and wholesale prices) and one cost variable (unit labor costs).2/

NicKinnon has pointed out a second objection to the CPL for international comparison in a recent Essay in the Princeton Series: that the difference between consumer and either wholesale or export

er

1/ Helen B. Junz and Rudolf R. Rhemberg "Prices and Export Performance of Industrial Countries, 1959-63," IMF Staff Papers July 1965,

a are me Re ECR OE Re aR arm reine sme Me ae. °

en ante tte seen ner ected nnteratnemn tintin rennnnee

- 27 -

prices (in any country and, hence, as between countries) can be a function of the rate of growth of per capita output.~ In fast-growing countries, rapid increases in real wages lead to increases in the

costs of services to final consumers which mainly affect the CPI. He

cited Japan, Germany and Italy as examples of rapid-growth economies

as compared to Canada, the United Kingdom and the United States as slow-growth ones. Using Japan as the extreme example, he found that

consumer prices rose by 97.3% in Japan between 1953 and 1969 compared

to a rise of only 41.4% in the United States. (p. 22.) By contrast, however, export prices actually declined by~5.2% in Japan but rose by

29.6% in the United States in this period. Because rapid increases in

real wages in fast-growing countries lead to increases in the cost of

services to final consumers, which mainly affect the CPI and are

largely nontradable, the consumer price index cannot serve as a measure

of the change in a currency's interval value for international purposes.

“(p. 21.) It is also not a satisfactory indicator oft the change in

relative competitiveness of an industrial country -- the kind of

calculation which might be made in-an attempt to determine whether a

change in parity between two countries might be indicated on the basis

of price variations.

- When we turn to the calcula-

Wholesale price comparisons

tions based on wholesale prices in Table 3, the evidence continues

ne nanmenentl

1/ Ronald I. McKinnon HNonetary Theory and Controlled Flexibility in the Forcign Exchanges, Princeton Essay No. 84, April, 1971, pp. 21 ff.

\

\ | |

!

rh EE ob ee

~ 28 «

to be contrary to the hypothesis that there has been a devaluation-bias against the dollar in the post-1959 period. However, the competitive advantage in favor of the United States vis-a-vis the other 13 countries is reduced -- on a weighted average basis --~ from 10% to 4% (See

Table 3.)

There are doubts about the usefulness of the wholesale price index as a comparative international measure on at least two grounds. On one side, the wholesale price index includes numer ous commodities (both of domestic and imported origin) which do not enter into export costs or move in internaticnal commerce. It also includes widelytraded standarized international goods whose prices tend to move together. On the other hand, it gives little weight. to the wide range of finished manufactured goods which comprise the bulk of the export trade of the industrial countries. In these respects, the wholesale price index can be regarded as a faulty indicator of changes over time of a currency's internal value for purposes of international comparison.

Export price comparisons - When we turn to the computations based on export prices, however, we find that the competitive-index has a value of 95 on a weighted average basis, whether based on changes in formal parities or in effective exchange rates. (See Table 3.) By this criterion, accordingly, a devaluation-bias against the dollar vis-a-vis the other 13 currencies can be identified.

Again there are doubts about the relevance of these results

for our purposes. In particular, there is concern about the technical

“~

~ 29 «+

properties of export-price indexes (whether of the price or unit«value variety) because of weighting and valuation problems over time. Furthermore, they are heavily weighted by goods in which the country has a strong comparative advantage. On the other hand, however, Junz-Rhomber g found that "regression equations with wholesale price relatives indicate that in many cases changes in this measure of price competitivencss are less closely associated with changes in market shares than are changes in relative unit values" of exports (p. 245) and that ‘'on the whole, unit value indices are the most useful indicators currently available

for the measurement of price competitiveness—in_international trade."

(p. 259)

Concluding observations

We have attempted to measure the extent to which changes in exchange rates have, or have not, offset differential price movements as between the United States and each of the 13 other industrial countries. The evidence from these purely statistical exercises is mixed. A judgment about whether there has, or has not, been a devaluation-bias against the dollar in the observation period depends upon the arbitrary selection of one price measure over the other two.

Hirsch-Higgins found, as an incidental by-product of their

construction of "effective" exchange rates for 14 industrial countries,

“that there had been a devaluation-bias against the dollar between 1959

- 30 - . |

and 1969, By their proposed test of devaluation-bias, however, they denied that this bias could be attributed to a general bias in the way par-values were adjusted. Both analytical and statistical doubts can be raised about the use of their test as an appropriate measure of whether there is a general bias in the way par-values have been adjusted. On analytical grounds, the primary evidence offered in the Hirsch-Higgins test encompassed only two ways of measuring the same economic indicator =" the country's exchange rate. We have indicated a preference for a devaluation-bias test which follows the traditional concern in international economics for a measure of exchange rate changes related to differences in domestic price or cost fluctuations. Let us assume, as a limiting case, for example, that we have measures

for the 14 countries in which each country's exchange rate was changed

in the observation period to the exact extent needed to offset dif-

ferential internal price variations in each country. Under this assumption, a comparison between each country's formal parity changes and its computed "effective" exchange rate might reveal differences, but these differences would merely be a statement of statistical results. It would be difficult, it is suggested, to attribute analytical significance to such deviations between parity and “effective” exchange-rate values. |

| On statistical grounds, the evidence as to the existence of , devaluation-bias was mixed. That is, any conclusion about devaluation~

_bias, one way or the other, which is based only on price data, depends

value for international comparative purposcs.

-gituation, the statis

-~31on the arbitrary selection of one price measure over the other two.

There is no analytical concensus which would justify an exc Lusive

concentration on any one of them. However, the measure the authors

chose to explore in a passing footnote -- the CPI -- is probably the

least widely accepted measure of changes ina currency's internal

Furthermore, the measure

found by Junz-Rhomberg to be the best measurement of price competi-

tiveness -~ export prices or unit-values -- yield results which

contradict the findings based on international comparisons of CPL

trends. Loe

Apart from the disparate evidence of the several price

measures, there is some question about the relevance of any statistical

exercise to a meaningful conception of devaluation-bias. From a

theoretical point of view, a tendency toward devaluation-bias could

take the form of a delayed adjustment of exchange rates by surplus

countries to differential rates of change in internal prices. So long

as the delayed adjustments were made within the observation period,

the statistical evidence would show no confirmation of that hypothesis.

Furthermore, a protracted reluctance of surplus countries to revalue

promptly in accordance with a devaluation-bias hypothesis could lead

to accelerated price inflation within the surplus countries. In that

tical correction of a devaluation~-bias could take

the form of an adjustment of internal prices in them. In. both: these

_ eases, the statistical results could fail to record the effects of the

cee Ne RMR OE me ane ets fe cma

'

‘either. Accordingly, support fo

-

- 32 +

devaluation-bias merely because the adjustments did occur within the

observation period, even though the processes of adjustment were

admittedly delayed in ways consistent with a devalvation-bias hypothesis.

We come therefore to the conclusion that the statistical

evidence that is availabie cannot be interpreted as a categorial denial

of the existence of a devaluation-bias in current international monetary

arrangements, either as 2 general bias in the way par values have been

adjusted or as a particular bias against the United States. But it

also cannot be regarded as categorial support for such an hypothesis

r the devaluation-bias hypothesis

must continue to be looked for in the concepts of international economic

theory which postulate that the greater part of the adjustment burden

under a fixed-rate system is likely to be borne by the deficit country

and in the practical world of affairs where officials in surplus countries

widely regard it as appropriate that deficit countries ought to bear

the greater part of the burden of international payments adjustments.

te