Jwrnal of Etono~nic Inrepnriun 8(1). Spring 1993. 1-32.
Factor Price Equalization: Theory and Evidence
Farhad Rassekh' U ~ ~ i w r ~ i y of Hutfwd
Henry Thompson* Auburn University
Abstract
Sr~rder~ts of intcrnariorro~ economics have absorbed rhe facrorprice equalization
(FPE) theorem over recent decades. I f may be time ro reconsider rhar facrvrprices
~vould likely converge ourside rhe contexf ofthe formal Heckrcher-Ohlin-Lcrner
Sarnuelsun rnodrl. This paper reviews the thcorerical evolution of FPE and the
rnlpiricul evidence regordins ;he influence vf trade on the in~rrnational pbtlern of . .. . facrvr.prices.
This paper examines the status of factor price equalization (FPE) as a scienrific hypothesis. Every college :etu&nt af inlemational economics is exposed to FPE in one fom or imorhet Intematiooal wde theorisu doing rekmh in competitive models typically accept FPE as a pw~digm and favor frer Ira&. Trade rheorists doing research in smtegic models of imperfect competition swm less inclined to .. do either. Ultimately, empirical tcsts of FPE should carry the ddy. . ..: , .: . .
The first section of this paper reviews the theoretical background of FPE and notes rhe tendency for trade to cause factor price convergence (FPC) outside [he
*~;uhud Rassekh: Depanmcnr of Econarnics. Univrrsiw of Haford. We% HarTford. CT. 061 17. U. S , A,. Henry Thompson: Depmment ol~Econornics. Auburn Univeqicy. &hb.m, AL. 36819. U. S . A,. The au~hurs thank Roy Rutlin. Bruce Elmlie. William Milbcrg. aid an
onanyolous refent for orvaiuble mmrncnta. The usudl discluimer applies.
. .
Farhad Rusekh end Henry Thornpaor1 ! . . . . .:. . ,i
. . . ... .# -: ' - this point. Samuelson's R E ilnicIes reminded Robins of a seyinarlpnper that Lemer, as n student, had ppresented in 1933 at the London School of-Eco.nomics. . .
Lemer [I9521 recreates that paper w d clearly lays outsuficirn~.c~nditions for FPE . -, -
7 - . . in the 2x2 model. , .
. . The hearr of FPE in competitive general equilibrium & e l s is an inyertible
. . ..,+
mapping from the vector of pricPs p to the vector of l^aclorprioes w. Under+me .. . . , . .
conditions, prices of goods uniquely determine prices of pmductivef~tg@. The
geneml equilibrium R E literature is rich, vast, and vaiid: ~amuei~o11i~ l9~;~1%7, :
19711, McKenzie [1955j, Johnson [1957], L.anc&er [19571;Umwa , [1959]; .. man ..,. [1966], Bhagwati and Srinivasm [1971], Kuga 119721, N i d o [I9721 Balassa
. . !::& 119741, Joshi [19871. McMhon 119881, d o t h e r s . Controversy . o v e i ~ ~ - ~ ~ d '
early, as exemplified by Peace and kames 1195 11 and P+me [19>2!.:.wnfcee , . . <...- trade equaliiits' the price of each good between t m d i nilions, equdi-nt-tional . . ' ,
prices of the same produc~ive factors are implied if the rnap~irig-fmm . . p to w is singular.
The formnl prouf of FPE is based on a neoclassical production strucQq w p t w r , . . . .4 . .> ,1_7 ".
final traded gwds and two primary . . domestic facton: of pN"ction i*%e HOLS.. model. ~ x a n s i o n to mdelswi th more goods, more factors; nontraded goods, . .
internationally supplied factors. and market imperfections are. built upon the HQLS .
model. With competitive pricing, the price of agwd (q) equals . . itsiYrra.dgg+scqf . ,. . -.
pmductiun (c,). In the background, not modelled explicitly, must bii &fi%f: '' ' i f ..,?,,
entry and exit which forces average cost to equal price.. Firms in eao~:i%&stry .. .
minimize cost by choosing inputs to produce a profit. ~ i m i z ~ , d t i l ~ & & Y , p . o . . -
exogenous (lo the f m ) iacbctor prices. Cost rninimizukion . is . a key l k i b - A t l i 6 f ~ ~ . . . . .-
qumen t . Competitive pricing is not necessaty for cost minirniziiion, ju@jting ,..
FPE could occur ~ ~ a i d e the context of competilive pricing:. . . . ,-- -
. , . ,
Rices of [he final trdcdgoods ate avsurned exog6nousat world leuels: Ihe 51ndl counhy wsumprion. The null hyprhrsis of the small'countty assumpfipn has nut
been freely rejrcted in the f=w enlpirical tests to h i e . ~couiikfloses market power . .
;~li the degme ut'aggregdtion increases l f a factor priceis &en at the world level. 9.:: -.,-:
the economy is a price taker in that international factor mai'ker. The ~mrnediare
effect of exogenous prices, either of g o d s ~r -~ roduc t i b r f i l c io~ , . .. is , to reduce the .:.... ,.
degrees of freedom in the general equilibrium mod&.. ' :. .. . . .: . . . .. ... . . . . .. ., . ;..,
The critical cundition fur FPE is decep'tively . simple: . . . the number . . o f ~ & t e s . . must
nor be zreairr tban the number of inrernalionil marker ( e ~ o g ~ n a u s p n ~ ~ s ) as, , . . - . . . . . .
' ' .: 5:.
devel'oped by Samuelson [19531. Jones and Scheinkman [1977], and Ethier and Svenshn [1986]. If the nunlber of factors is greater than the number of exogenous
, ,
prices. FPE does no1 follow since differat vectors w of factor prices suppofl the
c o m ~ t i t i v r equilibrium. Trnde theorists lypically avoid the yiloalion of more factors than intemtional markets, although it has not been eliminated onemphicd pounds. According to tests of separability. the number of factors is evidently quite large. Clark, Hofler, and Thompson[1988] point out hat US manufacturing has: ill kast eigh~ types of labor which cannot be consistently aggregated in any way.
If the number of international make& is greater than the number of factors, the wmpetitive model is overdetermined and some induslries will shut down. This situation is described by Johnson 11967. 19701, B e d [1970], Racier [1979]. and Wu [1987]. The classic enample is the complete specialization which occurs in a Ricardian coostant cost economy in which two traded goods nre prduced with a single domestic factor.
Anothei implicit assunlption is that there is no joint production. Each of the finished goods is produced via m independent production function. Joint pmduction (gas atid heating ail, for enample, as outputs of oil refining) is similar to disaggm gating finished gwds in its effects on the mapping from prices to fa& prices. If dl goods are traded and the number of goods (nanjoint plus joint) exceeds the numbcr of factors..rhe model is ovrrdetenninwl. Samuelson [I9921 and Jones [I9921 address the issue of joint production in models of production and wade.
A. FPE in the 2x2 HOLS Model
A picture af cost minimk;ltion for a two sector economy is presented in h e ~erner -~e&ce d i a g m of Figure 1. Inpub of factors 1 and 2 are measwed along
eithp aak. UNI ~illuf: isoqults are labelled x, ( j = I. 2). Smooth convex unit
value isaquants represen1 amouo~s of the two facmrs of production required to produce one dollar's worth ( l l p , ) of either good. Rices of goods are fixed at exog- enous world levels. Suppose further that production functions are identical across
counuies, a suenucus &sumption considerins the literature on applied production
malyyis. E;lch unit vdue isoquant would then be identical for each of the n-ding
I partners.
Productive factors are fwely nlobile between sccrors. which implies each factor
price' (w,) will be h e r m xross sectors. Finns in each indusuy will produce
Farhad hslekh and H c q Thompwn
fucmr 2
Figure 1
where the unit value iwquant is supported by an isocost Line. With bob indushies employing both fdaors, a common isocost line must supportborh unit vdue
isoquants. The common unit i ~ c o s t line is written
Equilibrium is pictured with cost mi~mizing inputs which are functions of the vector of factor prices: a,, =a,(w). Factor prices w, and w2 ate uniquely &rer- mined in h e general equilibrium. There is no room for different f s l o r prices and no way for different factur supplies to influence factor prices given this structure of production. evidently even wilh redundant factor supply in complete specialization.
B. FPE andHig11 Dimcmuiuna[ Mode.% of Pmducriun and M u
Suppose the nun~ber r of productive factors is greater than 2 and all factor prices
are endogenous. If h e number n of goods equals rand all goods are traded at world prices, FPE follows. When r = n = 3. for instmce, three bowl shaped isoquanrs are
suppomd by a unique isocost plane. . .
If r = 3 and n = 2 as in Heckscher's original model, FPE loses its logical compul-
sion and technical conditions would determine whether free rrade tends to equalize
factor prices. Two bowl shaped isoquanh rest within the three factor axes. The
isocosrplaoe could pivot and remain tangent to the isoquants. F a c t ~ ~ p r i m s are nor uniquely delemnined. This theoretical possibility is called factor price polairation
6 Fanor Prim Equdization: Theory und Evidence
(FPP) by Thompson [1986]. A negative correkation can arise between the price of a good and h e price of the factor used most intensively in its production. If all three
pairs of factors are technical substitutes, free wide causes factor price convergence (FPC) except i n situations of pronounced differences in facmr intensity. Both factor substitution and factor intensity play a role in determining the etfecls of changing
prices of goods on factor prices, as described by Jones and Easton [I9831 and
'Ihompson [1985u]. Factors in the 3x2 model can be renumbered and unit factor inputs ranked
Factor 1 is the extreme factor in sector 1. Factor 3 is extreme in sector 2, and factor
2 is the middle fjclor. using the terminology of Ruffin [1981]. If dw, 1 dp, and dw,l
dp, rue both positive. FPC would occur and relatively abundant (cheap) facton would enjoy increasing prices with t d e . Suppose, on the other hand, conditions favor FPP. If factor 1 is relatively cheap, good 1 should be relatively cheap in autarchy. When va8e opens, the price of good 1 rises bur the price of e&me factor 1 may fall. Empirical work would either examine trchnical conditions (substitution elasticities and factor intensities) or look for direct evidence of FPP versus FPC.
If r = 2 and n = 3, an addir iod isoquant would be added to Figure 1. Cost minimization in every indusuy would be 'impossible unless h e Lhree isoquants
happened to align, and one indusuy would generally be forced to shut down. An implication is that as an economy opens to uade and goods switch from being nonvaded to traded, industries would shut down. While shutting down may seem
fwfelched, industtie5 (defined fmely enough) regularly do just that. Political oppo- nenb to free trade are in fact inclined to srress this pint. ~ k d [I9591 and Johnson [1967,1970] explicitly develop h e 2x3 model. Ifone of therhreegwds is nontraded,
its price would adjust so ha t cost minimhation could occur. Rodriguez [I9751 and Deatdorff and Courant [I9901 examine WE in the presence of a nonrraded good.
I C. Near FPE
...
Specifications of general equilibrium models wherc FYE does not hold suggest
that FPE will nearly hold between trading partners. Near FPE (NFPE), developed
by Thompson [1990'1, applies across a wide range of underlying production func-
dons. Elasticities of Fictor prices wih -spa to factor utouwments, if not zero, ate nearly zero. Different endowments actors competidve economies would account for only small quantitative differences in factorprices. In the computable general equilibrium (CGE) models sunreyed by Shoven and Whalley [1984], the effeccs of
changing endowments on factor prices are cansistently small. An implication of NWE is lhar much of the fuss Over conditiow leading to FPK
versu FIT or even PPP may be of little empirical weight. Free trade would at any
nte nearly equalize faclor prices across economics. The international equalizing OF prices of goods has a powerful quantitative influence, which should perhaps come
as lit* suprise. Imagine an economy with I0 pductive factom under conditions
which Ieidd to FPE. Disaggregating one facror would relan the suff~cient conditions
for FPE, and the Jw I & elasticities (where v represents rhe vector d facior en-
dowments) would no longer be zero. This disagbmgation, however, would have
only a small quantitative effect on the model's sratic solution of factor prices.
D. FPE and Exogenous Factor Prices
Suppose rhe price af one of the c d C t 0 ~ ~ in Figure I is exogenized, with input. 1
(capital) bought on an inremationd market at w;. The endpoint I / k; of the isocost
line is fixed. With each OF the prices of goods exogenously given, the model i s
overdeter~nined in that the arbikxy position of 1 / rv;may be inconsistent with the
two isoquants. If one of the goods wen nonuaded, its price would be endogenow
and the number of factors (two) would equal the number of i n t e~a t i ond markets (one for factor 1 plus one for good 2). , . .
In models with more factors than goods, exogenous factor prices would reduce the degrees of freedom .and increase the likelihood of FPE. Insrirutional fcrcrs nlay exogenize factor pnce6: l a k r contracts, subsidized capital returns, and so on. In a model where FPE holds, opening one factor market to an exogenous international
price would result in industrial shutdown.
. - .
E. FPE and the Specific Faclors Mudel
In the specific factors (SF) model, each sector shares a common factor (labor)
whiIe empIoying a scctor specific input (capital). Figure 2 piccures the SF model, - with two sets of fuctor prices for identical unit value isoquants. FPE cloes not hold
Factor Price Eqrwlizaioo: 'Theory and Evidcnce
LI
Figure 2
since factor prices depcnd on factor endowmenls. Samuelson [I9711 shows hat with identical homothetic demand a w s s coun-
tries. FPC o c c m wirh free trade. imagine two SF economies pmducing the same
two goods in autarchy. Assume the only difference between the two countries is Ihiit
the home country has mare sector 1 capital. The relative price of good in the home
country would be lower, and the home country would export good 1 when trade opens. With p, rising, the reNm to capitrd in sector 1 rises. FPC occurs since
Jr, I Jp, > Q where r;- is the return 10 capital in sector J.
FPC is closely linked wilh the Heckscher-Ohlin (HO) pattern of trade in rhe SF mcdel. Arising price for an exported good leads to a higher return ro Actor specific ,
capiull. A country would tend to export goods using specific factors which were I : I
relatively abundant (cheap) in autarchy. The SF madel occupies n nliddlo position between the 3x2 model and the 2x2
HOLS model. In tk 3x2 model. neither FPE nor FPC is necessary outcome. In the i !
2x2 model. FPE is necessary In the SF model, FPC occurs given regular assump-
rions about demand, bur WE is no[ a necessary outcome. These relationships generalize to higher dimecsiond models in a suaighriorward way.
W E would be implied in the SF model ifone faclor were mobile inkmationally. If capital i n one sector comes tiom an intrm[ionnl market. employmeni of inter- ! national cipital adju~rs to clear the model. The cost minimirution is then lied to a
unique set of factor prices. Tne SF modd in Figure 2 with intsmtiona: capild
leads to FPE as in Snnivasan [1983j andThompson [1985b].
F. FPE and Aggregahn
The FPE argumenl may hold in higher dimensions for various numbers of facton, goads, and internationally uaded factors, a3 described by Erhier 119741,
Chaog 119791, Takayama 119821, mdTho~npson [1983.1987]. Suppose there are r
factors of production, with m of the r factors employed at international prices. K less than s = r - m goads are traded internationalIy, FPE does not hold and FPCor
FPP could be tbe mle. If s goods are traded inttmationally, W E will hold. If more
than s goods are traded internationally. industrial shutdown will ocCui.Viewed
from the perspective of aggregation schemes, FPE can be called a razor's edge
proposition. .. .
The issue of the number of goods and factors is mote than idle academics. Industries and factors of production smggle to establish their economic identity.
Aggregation is critical to theFPE qurnent, asdiscussed by Hicks [1959]. Chipman
[1%6], and Ktueger [19681. Theoretical investigation comes to the impasse of having ro specify a model with a cenain number of facton and goods. Learner
11992a. p. 151 rugues that 'Doe rather silly assumption that cries outfor change is equal numbers of commodities and fdctors." Aggregnuon is h e critical pmcess leading to conditions which determine whether to exped FPE, but few researchers
are drawn to improve its difficult theory or application. A convenient aggregation
scheme is typically chosen lo suit the theoretical model with little else for a guide. A wlevant quesrion concerns how much FPE would be relned if a~gregation is
uonble to meet strict requirements for FPE to hold Put another way, how much of
h e dispariry in inremationd wages can be altribut~id to conditions regilrding the
numbers of F~ctors and international markets? Tests of a theoretical propsition are impIicitly tests of the assumprions implying the proposi~ion. If a t~eory is rejected.
some of the assumption leading ro il must be altered and the theory reformuhtd
. .
G. FPE and Impcrfgcf Conwetition . . . . -. ... .,:
.. ;+." The building blocks of FPE me cost minirniza~ion and interjector f&or rnobil- ., , . .
=:. ity. Conipetition in the factor markets, more p i n a i v e than cornpetition in oulput
markets, is fundamental. Models with monopolies. international duoplies, and product differentiation in the output markers have no necessary impact on the W E mgumenr. In any industrial oqanization, firms will minimize cost.
Imagine he economy is a single monopolist hiring two facrors of production and producing a good which is borh cowumed a! home and exponed. The monapolist
equates rnllrginal revenue (from foreign and home demand) with marginal cost to
maximize profir. Profit is maximized profit by hiring inputs where the marginal rate
of technical substitution along the w e r e d isquant equala the ntio of factor prices. In suaregic models, Nnsh or Berhiand equilibria d e t d n e oulpur, price and
perhaps quality of competing f m 5 . If factor markets ace comprtitive and f m minimize their cost of production with all inpuls vacinble, srrdtegic behdvior does not necessatily affwt the FPE argument. Strategic m&ls are rypically panid A
equilibrium while FPE is a general equilibrium result. Games theoreric models would ideally be imbedded in u genenl squilibrium model which could elucidate
inmme redistribution.
H. N e c e s s q Conditions for FPE
' If some goods ace uaded f-eely between two counrries and FPE holds, some
necessary condidous follow. Free made implies p, = for each Laded good, where n represents a foreign variable. FPE implies wi = w; for each factor. Marginal product (MP) for each factor in producing each good must then be equal across
countries. Iffactors are paid their marginal value or revenue produce, M e j = Mt;',
regardless of the indusmal structure. One test of FPE would k to esrimare and compare MPs of the same factor across '
counmes. Over time. MPs should converge. The issue is not whether production
functions are identical, bur whether IMPS at existing input levels are eqLILI1 (FI'E) or converge (FPC). Differen1 production functions for the some good could result in identical MPs. The obsen,ation or different wages across countries has no neces-
sary bearing on labor's MPs in the vilriot~s industries. . .
I . A Gerreral Argurnent for I;PC - - . . .
FF'C is ven imporcant as a generul working h$pothesi~.kln Oh!$'i [I9331 irrhrmal manner. a compelling argulnenr proce&ls~fro~n relative failor obundice
FarhuJ R~rrekh and Henry Thomp&ol~ 1 I
to relative factor cheapness and then to relatively cheap goods which use those
factor inlensively. Hicks [1959, p. 2671 argues that FPE could be viewed as a long
N n tendency. When a counuy begins to expon;-demand rises for its relatively cheap g o d s which ace likely cheap because theu.production'relies on relatively cheap factors uf pmduction. With the opening of trade, t lwe is an incmase in the
demand for these abundant and cheap factors used intensively in export production. Few economists express ovcnvbelming exception to this argument, alhough
t h ~ r e lire numerous plwes where it might go astrdy and it might nut explain every
instance of uncle. This general q u r n e n t for FPC lacks logical compulsion but
carries weight as a practical guide to understanding the effece of lrade. FPC has broad general appeal.
Economists who believe h a t markets work we11 (or better ihao m y alternative)
believe the general argument that free uade will tend to equalize the return lo
similar faclorv interndonally. Economists who are suspicious of markets will not
bust ha hide to create a more equitable international disu-ibution of income. For each group, empirical testing muy be more crilical than theoretical refiement.
Ill. hpfrlcal Studles of FPE
R E would ideally be tested across different countries under various conditions.
ScienMic opinion would gradually form on one side of the proposition or the other. The FPE theorem would then be generally wcepted or rejected. Hit were rejected,
alternative theories would be formulated and tested in turn, R E has no1 received
such scientific scrutiny, and has been dismissed offhand by some of the profession. It is wonh noling thit of the four propositions of the HOLS model. only the HO
vade pattern has been subjected to syskmatic empirical scrutiny. Bhngwati [l964]. Michaely [1964], Stern [1975], and Dedor f l . 119841 survey the field. The Stolper- Samuelson rheorem has been directly rested only by Magee [L989] and Gaston ru~d
Trefler [1991]. Papers ssrimnting h e enects of protection (or b ide liberalization)
on wages and income distribution include Evrns [1971], Bwgess [1976], Hrvrigan and ~ o w e r [1982l, and Thompson [1990]. The computable genenl equilibrium
(CGE) livrilture prcduces empirical estimates of ~ t o l ~ e r - ~ m u ~ l ~ o n compardtive
static clasdcities. The Kybczynski theorern has never been directly tested, but
Learner [lY9?b] presents some preliminwy uysulrs.
Systemdtic empirical work on FPE began during the 1980s. It is curious that FPE
12 Famr Price Equuliration: Theory and Evidence
had not previously received enlpiricd attention, especially since the HO and FPE theorems are regarded as two sides of the same win. Heckscher and Ohlin never treated the two propositions as separate. Much of the subsrqueat theoretical work considers the twe propcsitions inseparable, as in Minhas 119631 and Jones LIY881. The HO theorem has been tested numerous times since Jxontief [1953] and continues to a m c t empirical alknrion. For a recent m j o r contribution, see Learner
[1984]. . ..
A. E d y Pessimism abour llre Relevance of FPE
A reason for the lack of empirical work on FPE lies in rhe peusimistn hat leading trade fiwrists expressed about its empiricd validity. According to Flam and FiandL!rs [1991. p. 91. Heckscher "did not accept factor-price equalization as an empirical
fact." Ohlin 11933, p. 261 went huther and declncd h a t "Complete equality of factor prices is ... almost unthinkable nnd cellainly highly impmbable."In Samuelson (19491, where arigorous proof of FPE is provided, there is at !east mild doubt in its
empirical validity. After laying down the assumptions of FPE. Samuelson writes
[p. 81.01 "...our problem is f n m now on a purely logical one. Is 'If H, hen inevita-
bly C' a correct statement? The issue is not whether C (FPE) will a c u d y hold." Trade lire&- in the 1960s i s replele with negetive remarks about the ernpiricnl
relevance of WE. Samuelson [lM. p. 1521 calls FPE an "unrealistic model," and .,
notes that it requires n production function that "would nor seem realistic enough for empirical calculations.'' Caves L1960, p. 921 offers the following view: "One may well wonder why the mid factor price equalization theorem has artracted so much anention. The whole discussion is, for better or worse, a supreme example of
non-opmtional theorizing." Bhagwari [19M, p. 321 notes. while reviewing W E : "Although the subject is . . . of historic interest and stiii contin.ue6 to aluact fresh minds, one cannot help feeling that perhaps 100 great o propodon of in~cllectual
energy hm bwn directed r o w d s a quesrion of limited utility." Travis (1964, p. 2161 comrncnls that "lhe abilily of trade to equalize fiictor returns i d thu's -to dlucate world production optimally is limited, even if transpon cosir werezero."
Kemp 11964, p. 451 expresses the opinion that FPE "is impormt if only because it
focuses attention on the obstncles LO equalization."
An empirical study by Minhas [1Y63] questions the relevance of FPE b e c ~ u s e o i factor intensify xversnls (FIRS). Minhas noles thnt when the elzitici~y of subsrim-
Fwh(ld Rmekh md Heury Thompson 13
tion between two facton diffem across counnies, FIRS may occur. After rstima~inp damticities for several indusbies, Minhas concludes that FIRS cannot be ruled out. Minhas points out that h e lame set of commodity prices can be consistent wirh
ditTetent factor price rauos in the presence of F I k . "Hence the equaliry of com-
modity price obtained through trade will not. in geocral, g u m t e e an equalization of F x b r prices in each country" [p. 451. Minhas did not, however, directly test FPE in his work.
B. Tests of FPE in the IY7Vs
The negative attitude of trade rheorisrs about the empirical relevance of W E began to moderate in the 1970s. This moderation stemmed, in pan at leas;; from
e u m e n t s raised in an empirical study by Krueger (19681 and a theoretical analysis
by Samuelson [IYIl].
Krueger [I9681 anenlpts to determine the sources of per capira income . . differ-
ences among counuies and draws conclusions about the valid@ of WE. She
assumes rhdt all countries share the same aggregate pmduction function with underlying idenucal pmduction functions for each good as in the HOLS mod-I.
Using 1959 data, Krueger suggests that more than half of the difference in per capita income between the US and many less developed countries (LDCs) can be
anribured to dift'erences in enduwnlents of human capital.
Although Krueger does not directly test FFE, her results began to tnr~rfo'orrn the
negiuive ouitude towards FPE. Srem [1973, p. 351, for example, writes that FPE "has never been subjscted to direct empirical inves~igation, presumably because i t
appears so obviously Violated by the sizable differences in factor prices that exist among couohies." Howaver, he interprets Krucgrr's results a s suggesting "hat perhaps we were not as fax from factor-price equaliwdon as might have been
thought. This is an inlrigui~~g suggestion that would ,be inkresting to investigate fmher." This invrpretation represents a shift to the point of encouraging empirical
. . . ..
"-*c.-* . - ,,- work on FPE. ..
S m u ~ l s o n [1Y71] directed inention from FPE' to FPC, .- & noted by-!ones and .,. . .. Neary 11984, p. 241. Samuelson utilizes the specific ?&tor rnodel$nrraducing identical homothetic preferences across countries. Samuelson's insight. inspired by
Ohlin 119331, has implicntions for empirical testing. p&rbaps the empirical rel-
evance of FPE cnnnot bejudged by comparing static factor returns across countries. - , - . .
14 Fuclur Price Equalizariun: Theoty and Evidcncc
The critical ks l i s whether factor prices converge (diverge) as uddt expands
(contrxcts). If such b n d s ate found, forces of FPE musl be at work. Officer [I9711 combines FPE witb purchasing power parity (YPP) to inttoduce
micro FPE (MFPE), built on the aggregaV: production function , . . which Hicks 119591 q u e s would provide validity for FPE as a long run tendency. OFficer w u e s rhal "many kinds and quulilies of ficiors of production are aggregated into two broad fwtors, labor and capiral" [p. 8701. This departure fmm conventional FPE stems from the contention that the assumption of identical production functions across countries is tw unrealistic. He argues that MFPE is based on reasonable assumptions and suggests hat esting MFPE is a proper way to tesl FPE. He
presents a model that boils down to h e satisfaction of
where 1 denotes aforeign variable, a,reptesen!s labor input per unit of output and p is the price level. A test across 10 induskid countries over theperiod 1952-70 yields results predicld by MFPE. Officer concludes MFPE "comes close to fulfill- ment in the real world, which implies h t Ule equilibrium exchange rates &fined
by the unit-factor-cost and purchasing-power-parity theories are close to identical? [p. 8771. Balasa [I9711 criticizes Officer's approach on the grounds that a conclu- sion about the obsolute version oCPPP has to lx drdwn.
floystad [I9741 tests two related assumptions of FPE: theequality of tv and r among different industries within a county. and the technical efficiency of resource ! allocation. He focuses on 17 nranufacturing and construction industries in Nowiy
for 1955, 1961, and 1965. The dam display significanrdifferences / n u: and r acroG
indusuies. Floystad estimates the marginal products oCL and K under rechnical ef- i
ficiency and compares them against observed marginal products (MPs ) . Estimates reveal that for 15 of 17 industries, the difference between the actual and:optimal w
! is less than 10%. The difference between ths actual and optimal r is less than 10%
' - . for only 9 of rhr 17 indusrries. Roystad then rejects rhe assurnpiWaf intersector Factor mobility. A tesr for the equalicy of I,, and of r in different indus$tjc5 ac~oss the
countries in the European Free Trade Area ( E R A ) a n d k ~ u r o ~ t a n E c o l t o n ~ i c Community (EEC) does not reject the hypothesis "that the iuucturr of w3gts in'rhe
manufacturing industries tend to be h e s a n e in the trading cuuntries because of
equol prices on the products and that wages are morp or less equdto the marginal
product of inbor in a technically efficient situation" ip. 5771. Floysr~d's pioneering .
paper reptesents the first attempt to test the smpiricd validiry of FPE:
.,. . FPE generates testable implicatioos that can be divided h l ~ two broad catego-
ries: static implioations rhn; would hold at iny given p i n t in &e, and dynamic
implicaliom that should tend to hold over time. The implication of . FPE . . that factor
returns are independent of factor endowments is a static properry. On the ooteer hanQ convergence of factor productiviry or factor prices overtime is's dynamic
propem'. . .
KoUoE, Leolller. and Sachs [US. 19811 uahered in empirical work on FPC in the 1980s. showing that the process of FPC hns k e n occurring for some counbies. Across the US, West Germany. Japan, and South Korea, mufacruring wages (w) converge over [he period 1967-77 and capital-labor i t i o s (K/L I k) converge over the period 1958-75. KLS argue lhat the conveeence ink only panidly explains the
observed convergence of w. The KLS data show [p. 231
rhat [he ratio of the lopanew capiral-labor mtio to rhml ofthe ~S ' inc rAed by almust 160 percent brlwcen 1967 and IY75. If one assumed hat wagw were determiwd by a linear homugeneous Cobb-lhuglm production functio~i with a capital coefficint of .3, this p w r h in relative capital labor ratios would imply a 18 percent increase in relauve wages. However. from 1967 u, 1975 Jupanese relative wages hemselves incresrd by almost 16Upetcent. Asimi- Q set of numbm holds true for Germany.
The nuthors report the percentage changes in relative wages in 23 industries over the 1967-77: b. 171 "The process of international w g r equalization appears. for
[he most pan, to have oc~urred uniformly across indusuies within th= various countries. which suggests a freely mobile intcrnal domestic labor market." Despi~e
wage convergence. KLS argue that the HOLS model cannot account for (among
other things) the observed large disparity in w across counlries. ' . .
They develop a muditied model in whichit is costly~q~alter , . @e input. Mar-
ginal revenue products (MRPs) of K would br unequol in the short win. Under this
condition. w would be unequal across domestic industries despite identical ttch-
nologies and the equalization of MKPs uf labor across counlries. The KLS mcdel
generates FPE in its long mn sready slne. Simulation under myopic expectations
suggcsrsrapid international wage converEence would occur. Hdfof h e gap berween , . . .
16 F u u r Rice Equalization: Theor?. and Evidence
the initial domestic w and the iong mn iv closes within 6 yeus. The remaining gap takes much longer to close.
Tovias [I9821 uses data from the EEC over h e period 1950-75 to test FPE, wguing h a t the assumptions which led to WE hold reasonably well in rhe EEC- Tovias repons standard deviations and cwfficic~tts of variation of manufacturing wages for several yean before and after the formation of the EEC in 1937. Calculations n v e d relative convergence until 1968, but divergence afreiwards. - Tovias claims that resulb provide overall empirical support lor RE, and raises a question: "By how much did free movement of goods andor workers contribute to
the convergence of labor costs?" [p. 3881
Gremmen [1985] criticizes the method of rovias for resting FPE and pmvides a ... . . . -*
partial amwer to the above question. Gremmen constructs h e following qua&on:
where a denotes aforeign variuble, k = WL, and b is h e level of bilakrd trade. The ratio o l bilateral impora to GNP is nl. and ba rn+m'. Gremmen estimates this equation for the EEC over the period 1959-79 when there are high levels of trade,
md for 26 olher countries among which there is less trade. According ta FPE, a should be negative (if w > w') and zero. AIiuger b would mean a higher level of bilntrrdl lrade, which would cause 1v and w' to approach each olher. The world
equution, which includes the EEC and the other 26 cuu~~uies, is estimated for 197.6.
Estimates of parunzters suppon FPE. The ekasticity a is about the same{-0.07) for both the EEC and the world.
Muurik I19871 raises questions about Grernmen's work, challenging the basic nution that relative resource endownlens should not influence relative farior prices significantly even when trade i s iflrmsc. Mowik objects on theoretical and empiri-
cal erounds, noting that the validity of FPE hinges on the convergence of k in indi- vidual industries across counlrirs. I.te reeslimnte~ Gremnien's model for the EEC. showing that k is an ir~iponant detsrminan~ of factor pries differences and tinding
.high pmial comlutiun coeficients between relative real wweu and k. Muurik
maintains that the first step in truly tesrina FPE would be the r~rinrrrrion of secroral prc~d~icliu~t fiin ~.tiur~s.
Dollarand \Voltl'lDW. 19881 note that forccs behind FPE should cause indusrrial labor productivity to converge ucruss counhes. The); report coefficienrs of variation of value added per work hour (trverage pmductivitg) across 28 indushes in 13
Funhad Rmrekh and Henry Thompson 17
industrial counlries for 1963, 1979, and 1982. A convergence of average
productiviries (APs) occurs for all industties over rhe sample period. Moreover, DW show thal variation in the employment mix among countries does not play an imponant role in explaining croswountry differences in aggregare productivity in
r.. -.-
all rnanufafturing, nor have changes in employment mixes been an important source of convergence.' [p. 5jOl. DW spculate that the sources of convergence ~ A P S have
been convergence of both production techniques and of ~echnology, although no evidence of such convergencies is presented. They find that in 1982 different countries held produclivity leads in different industries, which is more compatible
with trade [heories based on dilfereilces in technology.
Makhlari and Rasekh [MR, 19891 select 16 OECD countries over the p e r i d
1961-83 to test the proposition that international twde influences factor prices. MR
show rhar during the sample period t r ~ d e significantly increased, while borh w and
k converged. The increase in uade and convergence of k both conbibute to ex-
plaining convergence in ~ v . MR argue that a measure of trade openness should k included among the variables explaining w, and find it significant in most cases. A
classification of high wage and low wugr countries is suggested. Specifically. re-
gression resulls indicate that Canada. the US, Denmark, Gemany, the Reherlands,
and Sweden are h i ~ h wage countries, while. Japan, New Zealan& Austria, Bel- gium, Finland, France, Ireland. Norway. Swilzerland, and the UK ate low wage countries.
Mokhtaii [1992'J, using the MR data, si~ows that the dispersion in manufacturing wages across the sample countries 1,esponds asymmeuically to the expansion of
impork/output (~rdy) and rxportsfoutpur (x/y) in the short run. An increase in ~IL+
1e;lds lo international diver~ence of w whereas an expansion in d y induces a con-
vergence. In the Long run, however, an increae in borh measures of wade openness leads to wage convergence.
Williamson [I9911 calculates real wages of unskilled urban workers for 15
countries since 1830. His sample includes Austrdlia. Argenrinn, Canada, and (he US (the New World) and I I counlries in Western Europe (the Old World). The data
reveal that wal wages diverge across Frmce, Ireland, the Netherlands, Sweden, the
UK and the US, the countries for which dara are available uver the period 1830-56. A process of convergence then ssrs in, beconling more dramatic wound 1870 and
lasting througlt 19118. The interwar period upsets the convergence process, bur it
resumes iri mid-IYbUs. Willimson suggests rhar rhr observed cunvzrgeltce could
18 Factor Ricc Equaluution: Thmry und Evidenw
be due to factor movements, made, and reuhnology ~mnsfer.
In a subsequent work. O'Rourke and Wilrinmson [OW, 19921 show that the integration of internationnl commodity markets over the period 1870-1913 conlrib- ukd to factor price convergence between the UK and the US. OW atnibuu about
half of the obsrrved convergenoe of the wage-renlal d o (w/r) to the convergence
of commodity prices. They test FPE directly by linking fdctor prices to commodity prices rather rhan to increased mde . as in several of the papers reviewed above. Moreover, OW substantiate Ohlin's contention that equalization of commodity
prices between the US and Europe had led to FPE during the last hdf of the nineteenth century.
Rassekh [I9931 finds thar FPE is capable of explaining cross-country variation in
indusuy-level wnges for a sample of 11 indusuies in 14 OECD countries over the period 1970-85. He suggests thar diffusion of rechnology. relative dispersion of
production techniques, and to a lesser extent international wade at h e industry level explain the dispersion of wages across countries. Moreover, examination of wages and pmduction !scbniques in 7 industries of the nontraded goods sector indicates that variation in r h e ~ variables across countties in much leas than the variation in
the traded gmds sector.
D. Tests of the Static ln~plications of FPE
Kotlikoff and Leiuner [KL, 19871 inuoduce and compue three models of intw- national wade. First, the HOLS model cream FPE and by implication equalization
of the growth rues of factor prices. 'lhe second model assumes complete special: izativn in an uneven model with more goods h n factors. The third model is a specific facwr (SF) mo&i witb capital immobile between sectors in the shon run.
In the HOLS model. the accumublion of capitid (K) would tesult in rhe production of tnorr K-intensive goods with no change in the cclpiml-labor ratio k in any indus-
tq. Thus, The marginal product of labor and rhe wage w would also ternain un- '
changed. Human capital, however, is capable of raising the averap w. lo the un-
even model, more K would increaw w and change IhepclNern of specialization.
The HOLS model implies a zero correl;lrion cocfiicient "between each country's
capital per worker and its industry-specific capiurl per m n , value added per man,
and emlings per man" [p. '481. KL report these coefficients for 28 industries in 28 couniries for 1978. Of the 83 correlation coefficients. 75 exceed 0.5 and 33 exceed
Fabad Rasrekh und Henry Thompron 19
0.8, casting some doubt on the HOLS model. Although KL find some support for
PFE in rpgtession analysis, the evidence against FF'E in taken ro be overwhelming.
They claim that b e ather two models receive stronger empirical suppon. Two
o b ~ e ~ a t i o n s are offered: "each of the l i m e model6 plays an important role in determining trade, gmwth, and factor returns" [P. 2691; ,and a model with mare goods than facrors and adjusment costs is superior to any of the h r e r proposed models [p. 2301.
Dollar, Wolff, and Baumol [DWB, 19881 present evidence against the static implications of WE. DWB choose 28 industries in 13 indusrrial countries in 1980, and find that labor productivity (value added per employ=) in each induslry varies
considerably across countries. DWB note sat 'Yor the average industry the crass country productivity differential is of the order of 1 0 0 % [p. 3 I]. Data reveal rhat "a country that ha6 relatively high pmductivity in one industry tends to have high pmductiviry in dl industries" [p. 331. FPE implies thal counrries employing a
highrr k in the aggregate must be producing relalively more cltpilal intensive goods,
rather lhan using more capital per worker in each industry. To test this implication,
DWB calculate the cmfficient ~Pcorrelation between k in all manufacturing and in each industry. Most of (he correlation caeff~ciens exceed 0.5 and ;Ire sigdicantly
greater than zero, rejectiog this implication of b FPE mdel . Estimalion of a vonslog prcduction function wirh variable r e m to scale leads
DWB w suggest that "the economies of scale hypothesis does help considerably 'in accounling for the observed deviations from the predictions of the FPE, but it still
leaves much ro be explained" [p. 371. Tests suggest that h e model based on constant retums to scale and productivity differences (xross cauntries at the
industry level) outperforms "the model with identical technology and economies of scale" Ip. 401. DWB find that larger economies tend to have higher pmductivity, which could be the "result of external economies of scale operating at the national
level" [p. 431. While almost all of rhe empirical work on FPE u&s mnnufacluring wages or
indusnii labor productivity, Alston and Iohnsori [&I. 19881 focus on farmland markers. Their sample countries includeArgentina, Australia, Canada. New Zcaland,
and he US over the period 1960-81. AJ argue that since land is u heterogeneous factor of production, a comparison of growth iates in land prices rather than price levels is apprwpriale. Empiricd resuln show that "among the five countries the
annual growth rates of land prices (poolcd across land types) wcre approximntely
. . ,. 20 Factor I'ticc Equ;llizntiun: Thwry urul Evidence
equd between 1961 and I980 and betwecn 1970 and 1980" [p. 1511. N also test
FPE for the Staresin the US. Among five combelt Stales. land prices urem not at all
equal but theu growth rateb-wen similw. Similar relqtionships are detected for land renu. Moreover, the annual nominal exponential growrh rates bfland prices for 314
ofthe 48 continentid States fall within a 95% confidence interval, providing strong
suppon for the implication that wade equalizes the =turn to similar immobile
factors across trading regions (Stntes in this rnvdd).
Peterson [I9891 estimates rmtginal mtes of return OF capital for high. middle,
and low income counrries, employing a Cobb-Douglas production knction. Rates of return are 1596,242, and 21% for the three groups ~spen ive ly . Peterson argues
that in the presence of full iind atcurore information and in the absence of distortions,
the rates of mum would be equal. No conclusion re~arding FPE is drawn, but Peterson's work could serve as part of h e basis of a more complett. test of WE.
Elmslie and Milberg [EM, 1992) use a method developed by Brecher and Choudhi [I9881 to test the assumption of identical production functions, one of rhe building blocks of FPE. EM also test a pmposition of Kemp and Shimomura 119881
[hat technological differences in autarchy will c~uickly dissiplite with (rade: Tests
arr conducted by calculating the coeficient of variation of a technology mamix
actors countries over time. Elements of h e technology manix measure "the total
use of factor i in the production of all commudities produd in the United States"
[p. 91. EM repon the coefficient of v i a t i o n for 14 sectors in Germany. Italy, Japan, Noway, and Portugal for lOj9, 1965. 1970, and 1975, and obswve that "as trdde
openness increased, technologies neither diverged nor dramatically converged"
[p. 111. Accordingly. EM conclude that the assumption of identical production functions is not warranted. EM formulate an altemetive model in which technologi-
cal differences drive uade flows, but do no1 iest theu model. Since EM test only 5 counrdes. it is difficult to accept their finding as n severe blow to the FPE nlvdel. Their results cannot, however, be dismissed because FPE itself . . has received mixed
.. - empirical reviews.
It should be mentioned that a convecence of technologies across countries is expaced as international trade expands. The rcason is that increases in rrdde will
enhance con~pctition, forcing firms to employ mote advanced rechnolory. Such a
process should lead to [he convergence of technologies across countries. The
literature on the converEence of per cr~pircl GDP rei:olves around thc diffusion of
technology fmm he inorc advancsd to the less udvanced economies. For seminal
Farhad R ~ s e k h and Henry Thompson 7.1
papers on lhe convergence of per capita GDP, see Bnurnol [1986]. Abramowin
[1986], and Dowrick and Nguyen [1989]. For the role of intemat iod hade in the convergence process, see Barm and Sda-i-Martin .. [1990]. Ben-David [1991]. and Rassekh 11 9921.
N. Pollcy lrnplicetlons of FPE
Trade theories generally suggesr that international trade is mutually beneficial.
although the gains from made may not be evenly dismbuted. FPE implies that workers (at least in some sM1 p u p s ) in high wage countries may lose from trading with low wage counuirs. Thiu is a crucial issue for potential free trade areas
with a wide m g e of wages across men~bers. -.. Learner [19Eb] documents the convergence of industrial wages among devel-
oped countries during the period 1960-89, showing her unril1978 wage convergence
was mainly due to "extraordinary real wage growth in Japan. West Germany, France and Korea but relatively less wagegrowth in the United Stares" [p. 71. From 1978 to 1989, wages in the US and Germany declined while wages in Japan
continued to rise, surpussing wages in the US. Moreover, Leaner presents evidence of wage convergence among countries in the EEC and the EFTA. Between 1978
and 1989, real w a g s of the high wngr countries (Denmark, Germany, Belgium, the Netherlands) fell while rral wages of the low wage countries (Ireland, Spzin, Greece, Pothigal) generally rose [p. 8J. Although Learner supports free trade, he warns hat ". . . in the abarncv of very substmud increases in trade barriers, red
wages in the U.S. are virtually certain to decline over the next decade because of the forces of wage equalization" [p. 91. However, this panern of wage movements is
not necessarily evidence suppcrring FPE because the figures an: average industrial
wages which do not reflect shifting skill levels. Marrover, no econometric effort is made to attribute wage movemenn to thc forces underlying FPE. Nrverchrless,~the evidence and arguments ate cornpelIing enough to cast doubt overthr political
popularily of free uride in the high wage countries. Can the HOLS paradigm provide policy guidelines'? An implication of {he
Stolper-Samuelson theorem is that the move toward free trade would makt a
counuy's relatively scarce Fdctors (unskilled and low-skilled workers i n the vS)
worse off and the relatively abundant fnctors (skilled workers and cdpiral in the US)
betrer off. Maintaining protection or abando~~ing free wade to protect low wase . .
27 Fmwr Price Equuliiilrion. Theor). unJ Evidence .;
workers is an inefficient economic alternative. A more sensiblebut politically
difficult policy would k to redislribute some of h e gains from free trade to ihe
losers. McCurdy and Mroz [I9911 presenl evidence that h e income of college
educated labor in the US increased by 8% while the income of workers with less
education declined between 1978 andl989.While other forces may be irwork, [he falling wages of unskilled and low skilkd workers in rhe US my be due to
increased free b d e . If the HOLS model is a guide for policy, rhe ab'unao. nt facton
which gain from free wade could compensate l h e s c m e f i c ~ 6 for lh$ibsses. FPE has deep implications for the less developed c o ~ t r i e ~ (LZ)Cs) Hirschman
[I9771 refers to the publication of Sarnueluon's FPE papers [1948; 1 9 4 % ~ ~ "one of
the important. though badly ever mentiond dates i n h e emergencegf . .., .?. . . &;elopmenr - economics" [p. 671. Hirschman argues h a t the significance 0 f f P E . t ~ econolnic
development goes fnr beyond the general pmposi~ion~hat i"temaijona1 . ,.. . .... -. . . trade-is
mumally beneficial, because "Samuelson's results.. . poinred 1'0 trade & apkntid force toward h e equalization of inconies around the world'. [p. 681.
Whercas FPE might sensibly make some workers in the developed countries
(DCs) uneasy about trading with the LDCs, it shouldentice LDCs to w e t free wade _-.*-..I
with the DCs. Development economists, however, have generally taken h e opposite
position in advising the LDCs. Streeten [I9791 notes lhat leading development
emnomiss have expressed disbelief in the ability of FPE to btinng about international
wage. equality: "Raul Prebisch, Nan6 Singer und ~ u n n & ~ ~ r ~ ~ ~ t ~ k i r s s . . e l e g a n t l y . ., but more realistically, challenged not only Snrnuel'son's findivgs blur h e more
general view tha equilibrating forces showed a &dency for h e ." fruits ... ~f econo'mic
progress to be widely and, after time-lag, evenly shared" [p 581. : . ' . .. There have been, of cuurse, orher attempts to cast doubk or! FPE is_'= b~neticial
pmdigm for LDCs. For example, Moore [I9901 argues ihat a 3%3'model . - is more appropriate ha11 the 2x2 model for the ;maiysis of Kid~.policj;'i~.~&i~l.Xmeric~
where economies can be described by n model with manufac&kngi.agriculrurnl .-
expo*, md Vnditionrd rubsistence@cultm. In Moore's modei, tnde liberalization
would reducc h e rerum to labor that constiturrsup 10 60% of rhe population in
some puts of Cenrral America. He suggesb.-th&..a land . ,~disuibut ion .. propam could make vide liberalizntion beneficial to the poor in rhk t eg j h~~ - -~
For the LDCs. 3. belief in FPE amounts to a belief in the benefits from free wade.
Economists who believe in the overall efficiency of'thc mi&et'.system ., , reyard FPE
a a beneficial force fo:rhe LDCs. Krucser [I9901 vurrhat LIX: which haw . .
. . .,.::. . . , - . . . ., .. . . . . . .
F& Rsrsckl~ and Henry Thomy-run 23
resorled to pprtectionist uade regimes in the postwar era have been disillusioned by
results and most have changed their policies. It is unrealistic, however, to expect LDCs to continue a move to free wade as long as their trade partners in the DCs prokct the relatively high wages of their unskilled workers.
V. Suggestion fur Fu~ihw Research
Where do we stand on the scientific scatus of R E ? This review of h e Iireraolre
suggests that systematic testing of FPE ha5 only just k ~ u n . If FPE is granted the
stntus of a null hypothesis, ir has not been conclu~ively rejected. Potential avenues for further r r s k h discussed below.
Empirical tests of FPE have been confined to DCs. Efforl should be mdde to include the newly industrialized counuies WICs) and the LDCs. Kotlikoff and
'Leamcr [1987. p. 2291 suggest that large differences in factor returns could be due
to sufficiently dissimilar endowments which phce counaies in different cones of
diversification. FPE would then have to be tested separateIy for DCb. LDCs. and NICs within particularly same endowment cone. Convergence of factor prices
within groups but not necessarily between groups would be expected. Cue should be taken to ensure that only instances of increasing mdde in simiIar products is
included.
Faccorj need to be disaggregated to the greatest extent pussible. A change in
relative mmufachuing wages would provide evidence regarding W E only if manufacturing in each country were made up of similar industries with similar production techniques. Trade theoris15 have long recognized the need to include
skilled labor as a seprlrate input. Labor skill p u p s are not readily separable, and shouId be tteated as separate inputs. Capital 5150 comes in varieties: machinery, equipment, and buildings, some of which are sector specitic. Land is hardly a homogeneous input. Learner [1984] sets a minimum srandard for the classification
of productive inputs. Dollar and Wolf [I9881 and Kotlikoti and Learner 119871
come close to catching the spirit of W E in the application of high dimensional
models. The role of different forces in the convergence of labor productiviry uncovered
by Dollar and Wolff 119881 desenres study. Both production techniques (capitnl-
labor ratios) and technologies (production functions) may converge across coun- rries in each indusny. Mokhtari and Rassekh llY891 find that cxpmsion in intema-
21 Factor Ricv Equaliratina: Theory und Evidence
tional uade and convergence of capital-labor ratios across counlries are associated wirh convergence of manufacturing wages. The evolution of capital-labor ratios in
individual industries and countries need?i to be carefully examined in relation to the evolution of factor prices. In this regard, Rassekh [I9931 provides some prelimi-
nary Icsults.
Using exchange rates to converr income or wages lo a common currency may
overstate h e difference in purchasing power across counbies. In lesting FPE, wages have 10 be adjuskd for purchasing power to yield a reliable comparison.
Tovias L1982, p. 3811 argues that "foreign exchange rates of conversion seem to be sufficient in particular because we me looking into long-term uends." This would be correct if inflation rates in ihe sample counhies converge in the long run and shon run or spot exchange rates were nor our of line. Prices of non~d&d goods may also be critical in calculating differences in real wages across countries.
Production functions in various industries in different countries need to be
reliably estimated. Althouou$h it is nor nece s sq tn eslirnate production functions to rest the implications of FPE, knowle'dge of underlying production functions would
flesh out skletal factor propofiions theory. The vast literature on applied produc-
tion analysis provides a fooriog on which interntionid economists can build more
complete resfs and applications of factor proponions theory. It can be anticipated b a t estimated production functions will be somewhat
different acmss countries. FPE could nevertheless be stated as an approxinlate rhevrem: If production functions are similar across oilions and other suficirnt conditions hold, factor prices will become approximately equal with mde. Ob- served inrernational differences in production functions would be insufficient reason lo completely abandon !actor proporrions theory, the spirit of the FPE theorem, or the generic belief in FPC.
There is an informal view of an economy which underlies the FPC argument. Algebmic general equilibrium models formalize this picture, but may cause t r d e chwrists to lose sigh1 of the bnsic issues. Models of imperfect competition provide
detail about rhe strategic workings of patticular industries, but fiil to pr6Gide broad
insight. FPE remains a focal poinl of research in international economics.
VI. Conclusion
This paper surveys the ~heorerical cvulution and empirical ~nvestigation of the
Farhdd Rasekh nnwrlHenry Thompsan 25
factor price equalization theorem since its inceprion early this century. Theoretical
advancements poinl lo the richness of FPE as a conceptual-framework. Empir id
investigations provide tangibiliry for the proposition by dispelling some of the abstraction that surrounds FPE.
Learner [1992a] presents a powerful argument against testing trade models.
Rather than attempting to accept or reject a model, Learner suggests thatthe focus of empirical work should be on identifying "circumstances in which h e model is
useful and other circumstances in which it is misleading" Ip. 21. The reviewed
empirical works on FPE can be seen as following iwamer's insigh~. While much remains to be done, the profession is realizing that empirical investigation is critical
for the science to progress.
Valuable ideas and worthy pmpositiom rend lo survive in scholarly circles. FPE has conrinued to attract fresh minds despite its dismistd on u priori grounds as an unrealistic proposition by a ponion d the profession. This survey is meanr ro
moderate the view of the skeptics and enmumge s c i enac progress on FPE.
. .. ..;- References - .
. . .
Abramovitz, Moses [1986]. 'Patching Up, Forging Ahead, and Falling Behind.'.
Jourml of Economic History 46; pp. 385-406. ... . .. .
Alston. Julian M. and Paul R. Johnson [1980'], "Factor Price Equalization among Intemtional Farmland Markers," Australian Journaf of Agricultuml Eco- nomics 32; pp . 142-152.
Balassa Bela [1974j, "Purchasing Power Paky and Factor Price Equalization:
Commem." K.vkios 27; pp. 879-883. B a r r ~ . Robed 1. and Xavier Sala-i-Martin [1990], "Economic Growrh and Conver-
gence Across the United States," NBER Wurking Paper, No. 3419 Baurnol, William J. [1986]. 'Productivity Growth. Convergence. and Weifarc:
Whit the Long-Run Data Show." Arnericun Ecunornic Review 76; pp. 1072-
1085. Ben-David, Dan [199l]. ''Equalizing Exchange: A Study of the Effects of Trade
Liberalization," NBER Working Paper, No. 3606
Benrand, Trent 1. [1970], "On Factor Price Equnliz-don When Comrnpdities Ournumber Fnctors: A Note," Econornica 37: pp. 89-90.
Bhagwari. Jagdish 119641. 'The Pure Theory of International Trade: A Survey,"
-mlUl J0IX-J PUT! aPPJ1 ~UO!lELUJlUI~, [2661] aaq~!~ UlE!l[!M prrr! a3Ng 'a!lSIUlg
'OEOT-0101 'dd :6L Ma!nay 3!ruouoq myarupr ,,'ax1a%a~uo3 pm dn-qaa3 :~c-0$61 'qfin13 ~youoq ah!1mduro3 (1330,. '[6961] uaXn%~ 041 sna pm anals 'q~u~o(l
'LV-EZ -dd :ssard 1~ :aBpuqma '8qsuaag '3 yaqoa
Xq paupa 'apvj poynwaiuj of spoy~am las.u.rdrr13 m ,,'sayuno3 sscuay 1-1 p3u!du13 uv :Xl!hpnpoq Joqq hsnpul pw lapox nogrnpnlq
aVd-JolXd aqJ,,, '[88611 lomneff 'f m!ll!M PUB BPM 'N PMPa 'P!hBCI 'rnIIo[T -gSS-6pS 'dd !OL "!I
-?rwjS pur, S~,TWOUD.~~~O dla!nay ,,'~861-~961 'sa"onoq pamhw Buov
h!hy~npold ~oqq JO asua3~anuo3, '[8861] BIOM .N pJemm pm p!ma 'qoa
'965-685 -dd : 1 E Mayay ~!wovoq 1wuo~~wwapq ,,'$poo~ paparluo~ q!& uoguz~pbz~
a~w JWX~JO poowaw TI uo,, 'Lo6611 1~~no3 'N lmd pm .A uu1v 'wopnaa 'PWnOH WON :~J~lSlUV '( '[OA 'UaUq '6 MIad
pue sauof 'M ppu0.d Lq palpa 'rnwouoq ~mo!~nwa~u~Jo yoyoou~~ aw
y .,'wow aprq Srya!paq pun sauoau 3uysa~, '[H6r] 1 mpj'popna~ 'IOZ-L6I 'dd !92 EJ~IlJ7SJ~~OUOJ~ ,,'8~;ulrmSP~nIEly Sn " 1oq~lJO &!I!w~~s,, '[88611 U~~WO~L ha^ Pm 'JWOH Pm4a!?I 'UW 'TW3
-SL-~I 'dd :p~ a~!~~awouo~z ,:hrasu arapo~ ay~
'E md :apL pogmo~u~ JO hoau aw 30 dams v>. '[gg61] 'S voy '~~d!q3
'9ZL-60L 'dd :LP DJ!JlaUIOUOJ~ ,:SJOIIO~JP.~ PW 3p009 ~UBK
gr!~ run~rqn!n@ lelaua9 P ~VLJO su=oaql mas,, '[6L611 uo~su!~ '2~3 'SSW dl!~Jah~
p~sm~ :aEipuqm3 'un1JnaS 3!zuouoq pprto apml '[0961] 'a pmqq .S~AB~
'StrL1 'dd !B ~WOUO.?~ ]~3!l!fodl0 PUlmunOf ,,'S-WS palrun am JOJ SaJUap pu!du13 ~UIOS :uo!lnqys!a ~WOJUI pue S~YL, '[g~611 'd p!ma 's$a%na
'LI-5 .dd :sssq LIW :a8puqum3 'snsuaad .3 uaqox
Xq paqpa 'ap~ f~~o!l"~alfq ~.(T~J~WID~!J!~W~ ,.'I~POI.V ym~-qqo -1ayaq~a~ aqi jo IsaL huno3-ow v :salns papn aq pm! spem3 y uopdrunruo3jo aualua3mlJq aq~, '[g861] gnom ww pw W3M 'Jaqm~fl
$c-61 'dd : 1 EJ?WOU033 lLJU0.t)
-wAa1ulJo [muno( .,'uogqpnbg a3g3 lo~md pm asucdoq uo!rsnpoU
:s[syuaJajj!a a%~ 30 hoam au. '[I ~6 11 msnh!uus 'N 'L pro rls!pafit 'pm%qa 'M-1 .dd !pL )Dunor 3?rummy
aluaph3 pue Lmau :uopzficnb aspd ror3ei 9.2,
Farhad Rassekh and Henry Thqrnpson 27
sity Uniformity: An Empirical Assessment." Welhvirtschajlichrs Archiv. Erhier. Wilfred 119741, "Some of the Theorems of Lotemational Trade with Many
Goods and Factors." Jourrra[ of Intemotiurrol Ecunomic*~ 13; pp. 201-224. Ehier, Wilfred and Lars E. 0. Sveosson [1986], 'The Theorems of International
Trade and Factor Mobility," Journal uf lntemarional economic^ 20; pp. 2 1- 42.
Evans, H. D. [ 1971 I. "Income Dishbution. Welfare, and the Ausrralian Tariffs," Australiun Ecorronric Papers 10: pp. 89-1 13.
Floystad, Gunnar [19741. "Factor-Rice Equalization in 'Ihwry and Practice,"
Welnvirschafiliches Archiv 9; pp. 554-578. Ham, Harry and M. June Flanders [1991], HeckscherOhlin Trade Theory, Cam-
bridge: The MIT P~ess.
Gaston, Noel and Daniel Trefler [1992], "Rutection and Real Wages: Some Evi- dence," manuscript, TUlane University.
Gremmen. Hans J . [1985], 'Testing the Factor Rice Equalh ion Theorem in the W3: An Alternative Approach." Journal of Common Market Srvdics 23: pp. 277-286.
Hartigaa James C. and Edward Tbwer 119821, 'TI-& Policy and the American Income Diskibution," Review of Economics u d Slahtics 64; pp. 261-270.
hckxher, Eli. F. I19191, 'The Effects of Foreign Tmde on the Dis~but ion of
Income,'' translated and included in Ram and Flanden [1991]. Hicks. John R. 119591, 'The Factor Price Equalization Theorem." in Essay in World
Ecorronrics, Oxford: Oxford University Press. Hirschman, Albm [1977], "A Generalized Linkage Approach ro Development,
With Special Refwence to Staples," Economic Developnrerrr and Cultural Chmge 25, Supplement; pp. 67-98.
Johnson. Harry G. [1957], "Factor Eudowments, international Trade and Factor Ptices," The Mancl~ester Sclrwl ~ ~ E c o M I ~ ~ ~ c and Sociui Studies 25; pp. 270-
283. Johnson, Hatry G. [1Y67]. "The Possibility of Factor-Price Equalization when
Commudities Ournumber Fnctors," Econorni~u 34; pp, 282-288. Johnson. Harry G. [1970]. "On Factor Price Equalization when Commodities
Outnumber Factors: A Commenl," Ecor~omicn 37; 89-90.
Jones, Ronald W. and Stephen Eaton [1983], "Factor Intensities and Factor Sub-
stitution i n General Equilibrium," Jurrr.r~ul of tnrrrnatiur~al Econonrics 15;
28 Fucuor Price Equalir;liion. Theory and Evidencc
pp. 65-99. Jones, Ronald W. and Jose A. Scheinkmm [1977], 'The Relevance of the Two-
Sector Production Model in Trade Theory." Jourrral of Political Economy 85; pp. 909-935.
Jones, Ronald W. and J. Peter Neary [19841, 'The Positive Theory of Inlemarional Trde," in Harrdbook of lnren~arionul Economicj. Edited by Ronald W. Jones and Peter B. Kenen, Amsterdam: North Holland; pp. 1-61.
Jones. Ronald W. [1988]. "Heckscher-OhlinTra&demry.I_ in The New PalgmveA Dicrionury of Econofnics, edited by John Eatwell. Murray Milgate. and Petw Newman, Vol. 2, London: Macmillan Press Limited; pp. 620-627.
Jones, Ronald W. 119921, "Joinredness in Pmduction and Factor-Price Equaliza- ! tion," Review ofinrernational Economics 1; pp. 10-18.
Joshi, Ptavin R. [1987], "Heckscher-Ohlin Model and Factor Price Equalization Theorem Under Alternative Conditions: A Geometrical Exposition," The I Indian Ecor~ornic J o u r ~ L 34: pp. 92-108.
Kemp, Murray C. [1Y641, The Purr Theory of In remal io~l Trade, Englewood Cliffs: Prentic+Hall.
Kemp, Murray C. and Kogi Shimomura [198R], 'The Impossibility of Global Absolute Advanrage in h e Hechcher-Ohlin Model of Trade," O.&ni Eco- nomic Papers 40: pp. 575-576.
Kotlikoff, Lawrence J., Edward E. Learner and Jeffrey Sachs [1981]. 'The Intema- 1 tionlil Economics of Transitional Growth: The Case of the United States," ! NBER Working Paper, No. 773. 1
Kotlikoff, Lawrence J. and Edward E. Learner [1987], "Empirical Tests dAltemn- tive Models of International Growlh," in Trade and Srrucrural Changr in Pacific Asia, edited by Colin I. Bradford and Willinm H. Branson, Chicago: The University of Chicago Press: pp. 227-269.
Krueger,Anne 0. [1968], "Factor Endowments and Per Capita h o m e Differences Among Countries," Econornic Jourtrul78; pp. 641-659. I
Krueger, Anne 0. [1990], "Theory and Practice of Commercial Policy: 1945-
1990,'' NBEK Wurking Paprr, No. 3569. Kuga, Kiyoshi [1972J, 'The Factor-Price Equalization Theorem.'' Ecunomrrrica
40: pp. 723-736.
Lanctter, Kelvin [1957], "The Heckscher-OhIin Trade Model," Ecotlarr~errico 24; pp. 19-39,
Farhnd Ri166ekh u d He~iry Thompson 29
Lcmd, A. H. [1959], "Factor Endowment and Faclor Prices." Ecartomicu, pp. 137- 144.
Learner. Edward E. [1984], Souires of Internarioml Cumpururive Advnrrrage,
Cambridge: The MIT Press. Learner Edward E. [199Za], 'Testing Trade Theory," NBEA Working Puprr, No. 3957.
Learner Edward E. [1992b]. "Wage Effeclu of a U.S.-Mexicon Free Trade A g w
ment". NBER Working Paper, No. 3991. Lemer, Abba P. [ 19521. "Factor Prices and International Tnde," Ecunomica 19; pp.
1-18.
Leontief, Vassily [1953], "Domestic Production and Foreign Trade: The American
Position Re-examined," Pmceedings of the American Ph~losophic~al Sociery
97: pp. 332-349.
Magee, Srephen P. [1989]. 'Three Simple Tkts of the Stolper-Samuelson Theo- rem." in Black Hole Tarias and Endogenous Policy Theory, edikd by Stephen
P. Magee, William A. Brock, and Leslie Young, Cambridge: Cambridge
University Press; pp. 101-110. McKenzie, Lionel W. [19551, "Equality of Factor Prices in World Trade,"
Ecor~onrerrica 23; pp. 239-257.
McCurdy, Thomas, and Thomas Mroz [1991], "Measuring &croeconomic Trends in Wages from Cohort Specifications," Manuscripl Stanford Univmity.
McMahon, Gary [1988], "Dud Factor F'rice Equalizarion: Elements of a Theory,"
World Development 16: pp. 903-912.
Michtrely, Michael [1964], "Factor F'mpordons in International Trade: Current
State of the Theory," Kyklos 17; pp. 529-550. Mi&. Bagicha A. [1963], AII ln~remutiu~l Conrparison uf Facror Cosrs und
Factor (/ye, Amsterdam: Nonh-Holland.
Mokhrari. Manoucher and Fahad Rassekh [1989J, '?he Tendency Towards Factor Price Equalization Amung OECD Countries," Review of Econonrics arid Sra-
risrics 7 1 : pp. 636642.
Mokhtnri, Manoucher L19921, "Do Wages and lnrernational Trade Fullow An
Equilibrium Path?" Manusc~ipr. University of Maryland. Moore, Robert E. [1990], "Can Trade Liberalization Lead lo rm Increase in Poverty
in C e n t ~ I herica?'Jurmicrl of'Drrmelopnimt Ecor~on~ics; pp. 83-92.
hlourik. Add V. [1987]. "Testing h e Fictor Price Equalization Theorem in the EC: An Alternative Approuch: A Comriirnt." Juurnul of' Common Murker Sfudies
30 Fac~or Price E q d z a ~ i u n : Tllwly and Evidence
25: pp. 79-86. N i d o , Hukukane [1972], "Relative Shms and Factor Rice Equalization." Journal
uf ln~rrnational Economics 2 ; pp. 257-164. OWm, Lawrcnce H. 119741, "Purchasing Power Pwity and Factor Rice Equalizb
tion," Kyklos 17; pp. 868-883.
Ohlin, Betti1 119333, Itrterregional and International Trade. Cambridge: tiarvard
University Press. O'Rouke, Kevin and Jeffrey G. Williamson [1992], "Were Heckscher and Ohlin
Right? Puning the Factor Price Equnlhtion Thwrem Back inm History", NBER Working Paper, No. 37.
Pearre, LF. (19521, "A No& on Mr. Lemer's Paper," Econornica 19: pp. 16-1 8.
Pearoe, LF. and S.F. James [1951-521, 'The Factor Price Equalization My!li." The
Review of Economic Srudies 19; pp. I I I - 120.
Pelenon, Willis 119891. "Rates of R e m on Capital: An International Compari-
son." Kyklos 42; pp. 203-3 17. Rader, Trout 119791, "Factor Price Equalization with More Industries than Fac-
tors," Grncrul Equilibrim. Growth, and Trade. edited by Jerry Green and
Jose Scheinkman, Amsterdam: N o h Holland; pp. 347-354. Rassekh, Fahad [1992], "The Role of infernational Trade in the Convergence of
Per Capita GDP in the OECD. 1950-1985," Inrenlatiunul Economic Journal
6 ; pp. 1-15. Rase!&, Farhad [1993a]. "International T d e md ' Ihe RelaCive Dispersion of
Industrial Wages and Production Techniques in 14 OGCD Counlries. 1970-
1985," forthcoming in Open Economies Review. July.
Rodriquez, Carlos Alfredo [ 1975'1, "International Factor Mobility. Nontraded Goods,
and the International Equnlization of Prices of Goods and Facrors,"
Economrrrica 43; pp. 115-1 24. Ruffin. Roy [1981], "Factor Movements and Prices wirh Three Factors and Two
Goods," Economics Lerrers 7: pp. 177-1112,
Samuelson. Paul A. LIY481, "internationd Trade and h e Euqdization of Factor Prices," Econo~nic Journal 58: pp. 163-1 84.
Somuelson, Paul A. 119491. "Inlemational Factor-Price Equnlization Once Again,"
Economic Jour~~o/ 59; 18 1- 197. Sdmuelson, Paul A. 119531, "Prices of Factors and Goods in Grneral Equilibrium."
The Review of Ecorrurnic~ Srrrdiex Z 1; pp. 1-20.
Furhad Rusckh wJ Henry Thompson 31
Samuelson. Paul A. [1951]. "A Comment on Factor Price Equalization." The Re-
view uf Economic Studies 19; pp. 12 1 - 122. Samuelson. Paul A. 119641, 'Theorelical Notes on Tmde Problem," Review of
economic^ and Siarirrics 46; pp. 145-154. Samuelson. Paul A. [IY671, "Summary on Factor-Rice Equalization," InrrrnationuI
Economic Revierv 8; pp. 286-295. Samuelwn, Paul A. 11 971 I, "Ohlin Wfi Right" Swedish Jourrral of Economics 73:
pp. 365-384. Samuelson, Paul A. [1992), "Factor-Price Equaliaation by Trade in Joinr and Non-
Joint Production," Review of Inrernurional Economics 1 ; pp. 1-9.
Shoven. John, B. and John Whalley [1984], "Applied General Equilibrium Mcdels of Taxation and International Trade," Journal of Economic Literdure 22: pp. 1007-1051
Srinivaran, T. N. [1983], "Xutemational Factor Movements, Commodity Trnde and Commcscial Policy in a Specific Factor Model," J o u m l of Internarional
Ec~unornics 14; pp. 289-3 12. Stern, Roben M. [1964], 'Theoretical Notes on Trade Roblems." Review of Ecu-
. . nomics und Sratirrics 46; pp. 145- 154.
Swm, Roben M. [1975], 'Yesring Trade Theories." in lnrer~r iunal Trade and Fi- nance, edited by Peter B. Kenen Cambridge: Cambridge University b s ~ ;
pp. 3-49. Streeten, Pdul'[l979], "Development Ideas in Historical Perspective: The New
Interest in Development," in Economic Growth undResouizes, Vol. 4, edited by h a Adelman, New York: St. Martin's Press; pp. 56-68.
TaliJyarna. Akira [1982]. "On Theorems of G m w d Comgetitive. Equilibrium of Production and Tide : A Survey of Some Rrcmt Developments in the Theory of Internacion;ll Trade," Keio Economic Sludies 19; pp. 1-37.
Thompson, Henry 119831, "Factor Migration and income Redishbution in Interna- tional Trade," Kcio Econon~ic Siudies 20: pp. 63170.
Thompson, Henry [1985n], "Cornplcmcnfarity in a Simple General Equilibrium Production Model," Canadian lournnl oj'Eco~~omics 18: pp. 616-62 1.
Thompson, Henry [1985b]. "lntemational Capital ~ c b i l i l y in i speckc Factor Model," Atlur~tic Econornic Jur~rnril 13; pp. 76-79.
Thompson, Henry 11 9 861, ''Free Trdde and Fac~or Price Polarization," Eltmpruri
Econumic Rwiew 30: pp. 419-425.
32 Fucmr Rice Equalization: Theory and Evidence
Thompson, Henry 119871, "AReview of Advancements in the General Equilibrium
Theory of Production and Trade," Keio Econumic Studies 24: pp. 43-62.
Thompson. Henry [1990], "Simulating a Mullifactor General Equilibrium Model of Production and Trade," Inrcrnatiunal Econornic J o u m l 4 ; pp. 21-34.
Tovias, Alfred [1982], "Rsting Factor Price Equ&zation in rhe EEC," Journal of Cornmon Market Sludies 20; pp. 375-388.
Trnvis, William P. 119641, The Theory of Trade and Pm~cc~ion, Cambridge: Harvard University Press.
Uzawa, Himfumi [1959]. "Prices of the Facton of Produnion in International Tmde," Ec~onomicu 27; pp. 448-468.
Williamson, Jeffrey G. [1992], 'The Evolution of Global Labor Marken in the F i t
and Second World Since 1830: Background Evidence and Hypotheses," forthcoming, Journal of Economic Hi~rory.
Wu, Ho-mou [1987], "The EquPlization of Factor Prices in General Equilibrium
When Commodities Outnumber Factors," Journal of Infernational Econom- ics 23; pp. 343-356.