Professional Documents
Culture Documents
Structural Determinants of Real Exchange Rates and National Price Levels Some Empirical Eveidence
Structural Determinants of Real Exchange Rates and National Price Levels Some Empirical Eveidence
net/publication/4724855
CITATIONS READS
269 588
1 author:
Jeffrey H. Bergstrand
University of Notre Dame
65 PUBLICATIONS 10,448 CITATIONS
SEE PROFILE
All content following this page was uploaded by Jeffrey H. Bergstrand on 02 November 2015.
Structural Determinants of Real Exchange Rates and National Price Levels: Some Empirical
Evidence
Author(s): Jeffrey H. Bergstrand
Reviewed work(s):
Source: The American Economic Review, Vol. 81, No. 1 (Mar., 1991), pp. 325-334
Published by: American Economic Association
Stable URL: http://www.jstor.org/stable/2006806 .
Accessed: 25/05/2012 16:43
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .
http://www.jstor.org/page/info/about/policies/terms.jsp
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.
American Economic Association is collaborating with JSTOR to digitize, preserve and extend access to The
American Economic Review.
http://www.jstor.org
StructuralDeterminantsof Real Exchange Rates and
NationalPrice Levels: Some EmpiricalEvidence
By JEFFREY H. BERGSTRAND*
325
326 THE AMERICAN ECONOMIC REVIEW MARCH 1991
(R2 = 0.85, RMSE = 0.17; t statistics in pa- At higher [real per capita] incomes,
rentheses). products of different kinds, although
The predominant alternative explanation filling the same basic needs, are likely
to replace less sophisticated types of
to productivity differentials involves relative products; furthermore, products filling
factor endowments. In the two-good, two- new needs are added.... But the more
we divide total production into sub-
3Bracketed terms added. In the Kravis et al. (1982) groups, the greater will be the varia-
data, the empirical distinction between nontradeables tions in income elasticity. [pp. 94-5]
and services and between tradeables and commodities
is fairly minor and rests entirely upon the treatment of I formalize the Linder claim that per
construction. Tradeables consist of all commodities capita income has a dominant influence on
except construction; nontradeables consist of all ser-
vices plus construction (see Kravis et al., 1982 p. 193).
the structure of demand by assuming the
Consequently, reference here will be made to non- following nonhomothetic, nested Cobb-
traded services and traded commodities. Douglas-Stone-Geary utility function for the
VOL. 81 NO. 1 BERGSTRAND: STRUCTURALDETERMINANTS 327
tive factor intensitiesin production.If non- ment for traded commoditiesexceeds that
tradedservicesare relativelylabor-intensive for nontradedservices,implyingan income
in production (i.e., |A|> 0), a 1-percent- elasticity of demand for nontradeables
highercapital:laborratio in B relativeto A (tradeables)greater(less) than one.
will cause B's supplyof nontradedservices Thus, all three hypothesespotentiallycan
relative to traded commoditiesto be lower explain structuralvariation in the real ex-
than A's. change rate. Since per capita income is
correlated positively with the capital:labor
C. Equilibrium ratio and the level of productivityin com-
modities relative to services across coun-
Demand function(9) and supplyfunction tries, earlier studies have not tried to dis-
(15) can be solved for the equilibriumrela- tinguish empirically among the relative
tive price level (or real exchange rate), p, importancesof these three channels. How-
and the equilibriumrelativeoutputlevel, X, ever, if capital:laborratios and relative-pro-
in the representativecountry: ductivity measures are available, a ready
method of distinguishingbetween the de-
( 16) ^ + ' S II 1 A mand and supplyroles of per capita income
is to examine empiricallythe cross-country
relationshipbetween per capita income and
the output of nontradedservicesrelativeto
+ [(1- 8) -T - 8PXNI(XT + PXN) A
traded commodities, as demonstrated by
( OrD+ crS)( XTPXN) equation(17). A 1-percent-higherproductiv-
ity in traded commoditiesrelative to non-
A =X D(1+ S) [ UD A
(17)k traded services in B comparedwith A will
O-D + kS ( D + CJS)lkAk cause B's output of nontraded services
relative to traded commodities to be
OrD(1 + o(S)/(oD + US) percent lower than
+4S[(1
+ S)xT 8PXN] (XT + PXN) .
( ?D + S ) ( XT PX N))
A's. A 1-percent-highercapital:laborratio
in B relative to A will cause B's output of
Equation (16) demonstrates how the pro- nontraded services relative to traded com-
ductivity-differentials, relative-factor-endow- modities to be oD /[(OD + rs)IAI]percent
ments, and Linder hypotheses are all poten- lower than A's. However,a 1-percent-higher
tially relevant for explaining variation across per capita income in B relative to A will
countries in the equilibrium relative price of cause B's relative output of nontradedser-
nontraded services to traded commodities vices to traded commodities to be higher
(and the general price level). A 1-percent- than A's if, to be consistent with equation
higher productivity in traded commodities (16), the income elasticity of demand for
relative to that in nontraded services in nontradeables(tradeables)is greater (less)
country B compared with country A will than one.
cause B's relative price level to be (1 + The theoreticalargumentsare illustrated
cS)/(AUD + O') percent higher than A's, sup- in Figure 1. CountryB might have a higher
porting the productivity-differentials model. relativeprice level than A because B has a
A 1-percent-higher capital:labor ratio in B higher productivityin traded commodities
relative to A will cause B's relative price relative to nontraded services or a higher
level to be 1/[(0-D + os)jjAI]-percenthigher capital:laborratio (assumingnontradedser-
than A's if nontraded services are relatively vices are labor intensive)or both, causinga
labor-intensive in production (tAI> 0), sup- lower supply of nontradedservices relative
porting the relative-factor-endowments to traded commodities. However, if non-
hypothesis. A 1-percent-higher per capita traded services (traded commodities) are
income in B relative to A will cause B's luxuries (necessities) in consumption,
relative price level to be higher if the higher-per-capita-incomecountry B might
weighted minimum-consumption require- have a higher relative price level than A
330 THE AMERICAN ECONOMIC REVIEW MARCH 1991
A. Reduced-FormEstimates
P PN /X
Given these 1975 estimates of capital:
PT \/
labor endowment ratios (k) and levels of
productivityin commoditiesrelative to ser-
vices (H), econometric analogues to re-
duced-formequations (16) and (17) could
be estimated using ordinaryleast squares
//\ (OLS). Estimationof the log-linearversion
of (16) yields:5
(18) lnp=-2.81 + 0.171nfl + 0.231nk
\/ D (12.68) (2.25) (2.84)
B
xD
XA +0.181ny
(1.67)
X..XN/XT
(R2 = 0.90, RMSE = 0.14; t statistics in
FIGURE 1. RELATIVE DEMAND AND RELATIVE
SUPPLY CURVES FOR COUNTRIEs A AND B
parentheses). Equation (18) suggests that
each supply-orientedhypothesisfor the re-
lationship between the real exchange rate
because of a higher relative demand for and per capita GDP has partialexplanatory
nontradedservices.Only by examiningem- power. The level of productivityin com-
piricallyboth reducedformsor their under- modities relative to services has the ex-
lyingstructuralequationscould one hope to pected positive effect on the price of ser-
disentanglethese three influences;these are vices relative to commodities,accordingto
examinednext. the productivity-differentialshypothesis;the
coefficientestimate is statisticallysignificant
II. EmpiricalResults at the 2.5-percentlevel (one-tailed t test).6
Recent estimates of capital and labor en-
dowments and of levels of productivity in ratio of national output in commodities industries to
commodities and in services across coun- the level of employment in commodities industries
tries now make it possible to distinguish divided by the ratio of national output in services
empiricallyamong the three competingex- industries to the level of employment in services indus-
planations for the robust positive correla- tries, that is, (XT/LT)/(XN/LN), the inverse of the
calculation in Kravis et al. (1983) for the level of
tion between countries' price levels and their productivity in services relative to commodities. Em-
per capita GDP's. Edward B. Leamer (1984) ployment data are from the International Labour Or-
provides measures of capital and labor en- ganization's (1979) Year Book of Labour Statistics, as in
dowments for numerous countries circa Kravis et al. (1983). The output of commodities relative
to services and all bther data are from table 6-12 in
1975,of which 23 countriesoverlapwith the Kravis et al. (1982).
34 in Kravis et al. (1982) and Kravis and sTreating the coefficients of II, k, and y in (16) and
Lipsey (1983). The technique of Kravis et al. (17) as constants, indefinite integration of those equa-
(1983) could be used to approximate the tions yields log-linear forms with constants appended.
level of productivity in commodities relative 60ne referee noted that differences across countries
to services for 21 of these countries.4 in (XT /LT)/(XN /LN), the proxy for I'T / IN, could
largely reflect differences across countries in their capi-
tal per unit of labor in (traded) commodities relative to
(nontraded) services- the latter denoted
4This explains the 21 countries for regressions (1) kT/kN -rather than differences in 11T/ rIN. As noted
and (2). Capital and labor (LABOR1) for 1975 are in Section I, each country's I3LT ( = LT /XT) and /3LN
from Leamer (1984 appendix table B.1). The level of (= LN /XN) are negative functions of the country's
productivity in (traded) commodities relative to (non- states of productivity (TLT and 7LN, respectively) and
traded) services, II = 11T / 11 N is approximated by the of the country's relative factor price (W/R). Assuming
VOL. 81 NO. 1 BERGSTRAND: STRUCTURALDETERMINANTS 331
structural equations. Since this particular Second, the estimate of the elasticity of
model is overidentified, the underlying substitution in demand (oD), 0.9, is statisti-
structuralparameterestimatesfor the rela- cally significantand has a plausible magni-
tive demandand supplyvariablescannotbe tude. As noted in Section I, for the special
determined immediatelyfromreduced-form case in which the minimum consumption
equations(18) and (19).8 requirementfor nontradedservices (XN) iS
However, two-stage least squares (2SLS) zero, SD should equal 1 or - oD = -1. In
convenientlyenables determinationof the fact, the linear restriction that the coeffi-
structuralparameterestimates of theoreti- cient estimate equals - 1 could not be re-
cal equations (9) and (15).9 Two-stage jected even at the 10-percent significance
least-squares estimation of log-linear ver- level (F[1, 18]= 0.056). This suggests that XN
sions of (9) and (15) yields the following is not significantlydifferentfrom zero, con-
demand (D) and supply (S) equations for sistent with earlier observations noted in
the relativeoutput of servicesto commodi- Section I.
ties: Third,the relativeprice level, the level of
productivityin servicesrelativeto commodi-
(20) ln XD = - 1.96 - 0.901np ties, and the capital:laborendowmentratio
(1.87) (2.10) all have the expected relationshipswith the
supply of services relative to commodities,
+0.401n y althoughtheir coefficientestimates are not
(1.79) significant. The linear restriction on this
supply equation's coefficient estimates im-
plied by theoretical supply function
(21) lnXS= 4.61 + 1.481np (15)-that the coefficientsof the price and
(1.15) (1.07) fl variablesshould sum to - 1-was tested
also. This linear restriction was rejected
-0.351nfl - 0.601nk at the 10-percent significance level, but
(1.00) (1.20) could not be rejected at the 5-percentlevel
F[1,17]= 3.838).
(RMSE = 0.11 and 0.33, respectively; t Fourth,as noted in footnote 2, the empir-
statisticsin parentheses). ical distinction between services and non-
Several points are worth noting. First, tradeables and between commodities and
empirical implementation of theoretical tradeablesis a fairly minor one, resting on
structuralequations (9) and (15) yields re- the treatment of construction activity.
sults consistent with the Linder-type de- Equations(18)-(21) were also estimatedus-
mand hypothesis.The demand for services ing nontradeables/tradeablesdata instead
relative to commodities is significantly(at of services/commodities data. The results
the 5-percent level, one-tailed t test) re- are similarbut are omitted here for brevity;
lated to per capitaincome, holdingconstant these are available from the author upon
the influencesof relativefactorendowments request.
and productivity differentials on relative In concludingthis section, I emphasizea
supply. In the context of the model, this key empiricalresult that has been omitted
suggests that the income elasticity of de- in this literature:the null hypothesis that
mand for nontraded services (traded com- per capita income has no effect on real
modities)is greater(less) than one. exchangerates and nationalprice levels via
demandis rejected.
8This system would be just identified if one of the III. Conclusions and Policy Implications
supply variables were deleted.
9The t statistics reported are asymptotically valid.
Given that there are only three exogenous variables, This study has provided empirical evi-
2SLS estimation can only be used appropriately here to dence that the systematiccross-countryre-
determine relative outputs and prices. lationshipbetween real per capita incomes
VOL. 81 NO. 1 BERGSTRAND: STRUCTURALDETERMINANTS 333
and national price levels (or real exchange exchangerate. The resultshere suggestthat
rates), commonlyattributedto the supply- relative productivitylevels, capital:laborra-
oriented productivity-differentials and rela- tios and tastes can explain as much as 90
tive-factor-endowments hypotheses,can also percent of the variationacross countriesin
be attributedpartly to a demand-oriented real exchangerates. Consequently,the vari-
"Linder-type"hypothesis.Assumingnonho- ation in such diverse real economic vari-
mothetic tastes, countries with higher real ables over time will tend to obscure the
per capita income will exhibit, in equilib- policy identificationof the equilibriumreal
rium, stronger demand for nontraded ser- exchangerate.
vices relativeto tradedcommodities,raising
their relative price. Empirical evidence
showed that, even when differences across REFERENCES
countriesin capital:laborendowmentratios
and levels of productivityin commodities Balassa, Bela, "The Purchasing-Power-Parity
relative to services are accounted for, real Doctrine: A Reappraisal," Journal of Po-
per capitaincome still has a significantposi- litical Economy, December 1964, 72,
tive correlationwith relativeprice levels and 584-96.
outputs. The results are consistentwith re- Bhagwati, Jagdish N., "Why are Services
cent studies emphasizingthe importanceof Cheaper in the Poor Countries?" Eco-
nonhomotheticityof tastes for explaining nomic Journal, June 1984, 94, 279-86.
consumption and trade patterns across Clague, Christopher, "Determinants of the
countries (cf. Jerry G. Thursbyand Marie National Price Level: Some Empirical
C. Thursby, 1987; Hunter and Markusen, Results," Review of Economics and Statis-
1988).Yet with only 21 usable observations tics, May 1986, 68, 320-3.
from the Kraviset al. (1982) data set, the Dornbusch,Rudiger,"Purchasing Power Par-
resultsare only encouraging,not conclusive, ity," in John Eatwell, Murray Milgate,
and more empirical research along these and Peter Newman, eds., The New Pal-
lines seems warranted. grave Dictionary of Economics, London:
Finally,two policy implicationsare raised Macmillan, 1988, 1075-85.
by the issues addressedin this study. First, Hunter, Linda C. and Markusen, James R.,
some of the U.S. dollar'sdramaticreal ap- "Per-Capita Income as a Determinant of
preciation vis-'a-visforeign currencies be- Trade," in Robert C. Feenstra, ed., Em-
tween 1980 and 1985 and the subsequent pirical Methods for International Trade,
large real depreciationfrom 1985 to 1988 Cambridge, MA: MIT Press, 1988, 89-
shouldnot be attributedto "disequilibrium," 109.
or exchange-ratemisalignment.Per capita Jones, Ronald W., "The Structure of Simple
GDP of most other major industrialized General Equilibrium Models," Journal of
countriesfell relative to that of the United Political Economy, December 1965, 73,
States in the firsthalf of the 1980'sand rose 557-72.
subsequently,although not nearly by the Kravis, Irving B., Heston, Alan W. and Sum-
magnitude of actual real exchange-rate mers, Robert, World Product and Income,
changes. Second, the Plaza Accord of Baltimore: Johns Hopkins University
September 1985, which indicated that the Press, 1982.
majorindustrializedcountries'governments 9 , and , "The Share of
coordinatedan effort to facilitate the U.S. Services in Economic Growth," in F.
dollar's real depreciation toward some Gerard Adams and Bert G. Hickman,
"zone," and massive 1988 coordinatedcen- eds., Global Econometrics, Cambridge,
tral bank interventions,which revealedthat MA: MIT Press, 1983, 188-218.
the same countriescoordinatedan effort to and Lipsey,RobertE., Towardan Ex-
preventthe dollar'sdepreciationbelow this planation of National Price Levels, Prince-
zone, both suggest that the governments ton Studies in International Finance No.
have some notion of the "equilibrium"real 52, Princeton,NJ: InternationalFinance
334 THE AMERICAN ECONOMIC REVIEW MARCH 1991