Exchange Rate Shock on Malaysian Prices of Imports and Exports: An Empirical Analysis

Transcription

1 Journal of Economic Cooperaion and Developmen, 30, 3 (2009), Exchange Rae Shock on Malaysian Prices of Impors and Expors: Jaria Duasa 1 This sudy examines he significan impac of exchange rae shock on prices of Malaysian impors and expors. In mehodology, he sudy adops vecor error correcion (VECM) model using monhly daa on nominal exchange raes, money supply, prices of impors and prices of expors covering he period of M1:1999 o M12:2006. For furher analysis, we adop an innovaion accouning by simulaing variance decomposiions (VDC) and impulse response funcions (IRF). VDC and IRF serve as ools for evaluaing he dynamic ineracions and srengh of causal relaions among variables in he sysem. In fac, IRF is used o calculae he exchange rae pass-hrough on impor prices and expor prices. The findings indicae ha, while he exchange rae shock is significanly affec he flucuaion of impor prices, he degree of passhrough is incomplee. 1. Inroducion In recen years, i is found ha inflaion in a number of indusrial and developing counries has remained surprisingly sable in he face of wide swings in exchange raes. Malaysia is no an excepional. Prior o he mid 1990s, he moneary sraegy in Malaysia was based on argeing money aggregaes. Up unil 1987, M1 was he main policy arge and laer, greaer imporance placed on M3 during financial liberalizaion. Due o fundamenal changes o he financial sysem in he early 1990s, he Cenral Bank shifed from moneary argeing o ineres rae argeing. The shif was precipiaed by he financial deregulaion and 1 Deparmen of Economics, Faculy of Economics and Managemen Sciences, Inernaional Islamic Universiy Malaysia, P. O. Box 10, Kuala Lumpur, Malaysia.

2 100 Exchange Rae Rhock on Malaysian Prices of Impors and Expors: liberalizaion as well as srucural changes in he economy o a more ineres-sensiive marke. Moneary policy in he counry remains supporive of economic aciviy wih inflaion moderaing o 3 per cen in Even hough he counry faced he episode of financial crisis in , i did no push up inflaion in he counry as well as in oher counries in Asia. Since hen, inflaion has moved a similar pace in four ASEAN counries despie differing exchange rae movemens. In mos counries, impor prices have moved a a considerably slow pace. This developmen has drawn aenion o he issue of exchange rae passhrough o domesic prices or he impac of exchange rae movemen on impor/expor prices of he counry. Is here any evidence of he impac of exchange rae on domesic prices? Is his impac or exchange rae pass-hrough is declining? These are he focal quesions in he curren sudy. A decline in exchange rae pass-hrough can have imporan macroeconomic implicaions. Exchange rae pass-hrough on impor prices, for example, affecs expendiure swiching in he domesic marke by changing he relaive prices of impored and domesically produced goods. This raises he quesion of wheher a decline in he exchange rae pass-hrough has weakened a channel hrough which curren accoun imbalances can be adjused. Exchange rae pass-hrough (ERPT) refers o he exen o which exchange rae changes aler relaive prices. I is used o refer o he effec of exchange rae changes on one of he following: (1) impor and expor prices, (2) consumer prices, and (3) rade volumes. In his sudy, we focus on he effecs of exchange raes changes on impor and expor prices. The exbook definiion of exchange rae pass-hrough is he percenage change in local currency impor/expor prices resuling from a one percen change in he exchange rae. Simply, ERPT is given by: P P ( EP ) EP and is inerpreed as he degree o which changes in he nominal exchange ranslae ino domesic prices; where α refers o he esimaed coefficien of pass-hrough for counry i, P is he domesic price level (expor prices or impor prices or consumer prices) for counry i a ime, denoes changes, E denoes nominal exchange rae

3 Journal of Economic Cooperaion and Developmen 101 of counry i in ringgi per US dollar and P * denoes foreign prices (Anaya, 2000). Theoreically, complee pass-hrough occurs when α = 1 and incomplee pass-hrough occurs when 0 < α < 1. One sandard way o esimae ERPT is by obaining he coefficien from regressing changes in prices indices on movemens in nominal exchange raes. The ineres in pass-hrough began in he 1960s and early 1970s when open economy moneary models assumed absolue (or relaive) Purchasing Power Pariy (PPP) o explain he behaviour of exchange raes. Since hen, heoreical sudies and empirical work in his field have gone hrough few sages (See Yang (2003), Dornbusch (1988), Krugman (1987), Kneer (1993), Yang (1997)). Recen heoreical developmens in open-economy macroeconomics enable a deeper analysis of pass-hrough issue. Obsfeld and Rogoff (1995), for example, provide a deailed micro-founded framework ha enables an assessmen of welfare crierion in he form of uiliy of he represenaive household. Their sudy has been exended in subsequen empirical works, paricularly by Tille (2000) and Bes and Devereux (2000). Several sudies have invesigaed he degree of pass-hrough and wheher exchange rae pass-hrough is sable across. This includes sudies by Taylor (2000), McCarhy (2000), Campa and Goldberg (2002) and ohers. Unforunaely, hese sudies are largely focused on he OECD counries. Ineress on his issue on emerging economies only sared afer he 1997 Asian financial crisis. The crisis necessiaes a re-look ino he exen of exchange rae pass-hrough o domesic prices in he Eas Asia region because i is a key facor in he ransmission of shocks (Parsons and Sao, 2005). Thus, his sudy aemps o examine he significan impac of exchange rae shock on prices of Malaysian impors and expors or simply ERPT. In mehodology, wihin VAR framework, he sudy adops vecor error correcion (VECM) model o analyze he impac of nominal exchange rae shock on impor/expor prices in shor run and long run. For furher analysis, we adop an innovaion accouning by simulaing variance decomposiions (VDC) and impulse response funcions (IRF). The degree of ERPT is hen calculaed using impulse response funcions. The res of he paper is organized as follows: he nex secion presens he empirical mehods and preliminary analysis of he daa. Secion 3 highlighs he empirical findings including he daa preliminaries and he

4 102 Exchange Rae Rhock on Malaysian Prices of Impors and Expors: resuls based on he unresriced VAR and VECM ess. Furher inferences are hen made based on he VDC and IRF analysis. Finally, secion 4 concludes and draws several policy recommendaions from he major findings of he paper. 2. Daa and Mehodology Daa and Variables Daa on nominal exchange raes, money supply M3, impor prices and expor prices are monhly, ranging from M1:1999 o M12:2006 and sourced from Bank Negara Malaysia s Quarerly Bullein and Inernaional Moneary Fund s IMF Financial Saisics of various issues. The raw daa obained for mos variables are in indices (impor prices, expor prices and nominal exchange raes) excep money supply M3, which is in RM million. The base period for impor and expor price indices is 1999 and he base year of nominal exchange raes is All variables are ransformed ino logarihm, denoed wih ln and Δ denoes he firs difference operaor. To evaluae he inegraion properies of he variables, we employ sandard augmened Dickey-Fuller (ADF) and Phillips-Perron (PP) ess (Dickey and Fuller, 1981; Phillips and Perron, 1988). A variable is said o be inegraed of order d, wrien I(d) if i requires differencing d imes o achieve saionariy. For coinegraion, we employ he VAR based ess of Johansen (1988) and Johansen and Juselius (1990). The Model To es he impac of changes in exchange rae on prices of expors and impors, he vecor auoregressive (VAR) model is adoped. In his analysis, here is a se of p=4 endogenous variables, z = [lnneer, lnm3, lnpx, lnpm] where lnneer,lnm3, lnpx and lnpm refer o log nominal exchange raes, money supply, expor prices and impor prices, respecively. Following Johansen(1988,1991) and Johansen and Juselius(1990,1992), we consider a p-dimensional vecor ime series z and model i as an Unresriced Vecor Auoregression (VAR) involving up o k-lags of z.

5 Journal of Economic Cooperaion and Developmen 103 z A1 z 1... Ak z k, ~ niid (0, ) (1) where z is a (px1) marix and each of he Ai is a (pxp) marix of parameers. The Johansen approach is used wih he consideraion ha i enables hypoheses ess concerning he marix and he number of equilibrium relaionships o be carried ou. Prior o coinegraion es, he maximum lag lengh, k, is chosen for he Unresriced Vecor Auoregression Model (VAR). Choosing he appropriae lag lengh is imporan since a k oo small will invalidae he ess, whereas a k oo large may resul in a loss of power (Kanioura, 2001). The appropriae lag is chosen by checking he residuals of VAR model wih one lag afer anoher and he selecion of lag is based on he one ha has he absence of serial correlaion in he residuals. Being aware of he lag order, hen we consruc he long-run equaions (Unresriced VAR model) for he series. The analysis is carried ou furher by doing he Johansen coinegraion es wih k-1 lag. The deerminaion of he number of coinegraing vecors is based on he maximal eigenvalue and he race ess. The vecor error correcion model (VECM) resrics he long-run behaviour of he endogenous variables o converge o heir coinegraing relaionships while allowing for shor-run adjusmen dynamics. In his case, he coinegraion erms are he correcion erms since a series of parial shor-run adjusmens correc gradually he deviaion from longrun equilibrium. The VECM corresponds o a resriced VAR of order k- 1 for he firs differenced series, wih he inclusion of error-correcion erms for he coinegraing vecors. We wrie a p-dimensional vecor error correcion model (VECM) as follows: y k 1 i y i y 1 i, = 1,...T (2) where y is he se of I(1) variables discuss above; ~niid(0, ); is a drif parameer, and is a (p x p) marix of he form where and are boh (p x r) marices of full rank, wih conaining he r coinegraing vecors and carrying he corresponding loadings in each

6 104 Exchange Rae Rhock on Malaysian Prices of Impors and Expors: of he r vecors. The adjusmen coefficiens in marix α refer o he coefficiens of he Error Correcion (ECM) erms. Addiionally, we adop an innovaion accouning by simulaing variance decomposiions (VDC) and impulse response funcions (IRF) for furher inferences. VDC and IRF serve as ools for evaluaing he dynamic ineracions and srengh of causal relaions among variables in he sysem. The VDC indicae he percenages of a variable s forecas error variance aribuable o is own innovaions and innovaions in oher variables. Thus, from he VDC, we can measure he relaive imporance of flucuaion of impor or expor prices in accouning for flucuaion in nominal exchange raes. Moreover, he IRF race he direcional responses of a variable o a one sandard deviaion shock of anoher variable. This means ha we can observe he direcion, magniude and persisence of impor or expor prices o variaion in nominal exchange raes. In fac, IRF could be used o calculae he exchange rae passhrough on impor prices and expor prices. Wihin a vecor auoregression (VAR) framework, ERPT is defined as he response of impor/expor prices o an impulse in nominal exchange rae. The degree of pass-hrough is calculaed eiher by he immediae impac of changes in he exchange rae on impor/expor prices (shorrun pass-hrough) or by he cumulaive effec of such changes (long-run pass-hrough, calculaed as he esimaed shor-run coefficien divided by one minus he esimaed inflaion lags). To be more specific, passhrough coefficien is calculaed using as follows: ERPT j 1 j 1 imp neer 3. Empirical Findings As a preliminary sep, we firs subjec each variable o Augmened Dickey Fuller (ADF) and Phillip-Perron (P-P) uni roo ess. The resuls of he ess are displayed on Table 1. The resuls generally sugges ha mos variables, in paricular, nominal exchange raes, money supply and expor prices, are inegraed of order one as he null hypohesis ha he

7 Journal of Economic Cooperaion and Developmen 105 series are no saionary is acceped a level bu rejeced a firs difference. In oher words, he variables are saionary a firs difference or I(1). Only series of impor prices is saionary a level or I(0). Since lnneer, lnm3r and lnpx are non-saionary or random walk sochasic processes and hey are inegraed of he same order, he linear combinaion of hese variables migh be saionary or we could say ha hey are coinegraed. Since hey are coinegraed, here is a long-erm equilibrium relaionship beween he variables. Variable ADF es saisic Table 1: Uni Roo Tess (wih rend and inercep) Level Firs Difference P-P es saisic (wih rend and inercep) Level Firs Difference lnpm -5.61*** *** -5.77*** *** lnpx *** *** lnneer *** *** lnm *** *** Noe: ***, ** and * denoe significance a 1%, 5% and 10% level, respecively. In order o es for coinegraion, he VAR model is developed which consiss of 4 endogenous variables : z = [lnneer, lnm3, lnpx, lnpm]. The maximum lag lengh, k, of 2 is chosen. Based on Maximum Eigenvalue and Trace ess of coinegraion, here is one coinegraing vecor exised among he variables. Table 2 provides deail resuls of hese coinegraion ess. Normalising lnpx, following is he suggesed vecor or error correcion erm (ec) or long-run relaionship among he variables (he number in parenheses are -raios): ec = lnpx lnm lnpm lnneer (-2.94) (3.91) (1.95)

8 106 Exchange Rae Rhock on Malaysian Prices of Impors and Expors: Table 2: Johansen Coinegraion Tess Resuls Null Hypohesis Max-Eigen 5% Criical Trace 5% Criical abou Rank (r) Saisic Value Value Saisic r= r r We hen proceed wih an esimaed error correcion model o illusrae how he coinegraion resuls migh be uilised. The vecor error correcion model (VECM) resrics he long-run behaviour of he endogenous variables o converge o heir coinegraing relaionships while allowing for shor-run adjusmen dynamics. Table 3 displays shor-run equaions. All coefficiens of shor-run equaion are coefficiens relaing o he shor run dynamics of he model s convergence o equilibrium and coefficiens of lag ec represen he speed of adjusmen. The negaive sign of he ECM erms is he correc sign of he error correcion and is significance shows he evidence of causaliy in a leas one direcion. However, from boh equaions, here is no evidence of significan role of exchange raes in changes of impor and expor prices in shor-run. To es he robusness of he error correcion model, we apply a number of diagnosic ess. We find no evidence of normaliy failure, serial correlaion, heeroskedasiciy and ARCH (Auoregressive Condiional Heeroskedasiciy) effec in he disurbances for equaion 1 and equaion 2 only suffers problem of normaliy failure in is residuals. From an esimaed VAR, we compue variance decomposiions and impulse-response funcions, which serve as ools for evaluaing he dynamic ineracions and srengh of causal relaions among variables in he sysem. The resuls of variance decomposiion and impulse response funcions are displayed in Table 4 and Figure 1, respecively.

9 Journal of Economic Cooperaion and Developmen 107 Table 3: The Vecor Error Correcion Model Equaion Ind. Variable Dependen Variable (1) lnpx (2) lnpm consan lnpm *** lnpm lnpx *** lnpx lnm lnm *** lnneer lnneer ec ** *** Included observaion Adjused R Diagnosic es: (F-saisics) JBnormal Far Farch Fhe ** * Noes: 1. Far is he F-saisic of Breusch-Godfrey Serial Correlaion LM Tes Farch is he F-saisic of ARCH Tes JBnormal is he Jarque-Bera Saisic of Normaliy Tes Fhe is he F-saisic of Whie Heeroskedasiciy Tes 2. ***, ** and * denoe significance a 1%, 5% and 10% level, respecively. From Figure 1, he IRF can produce he ime pah of dependen variables in he VAR, o shocks from all he explanaory variables. I could be seen ha impor prices reac significanly o nominal exchange rae innovaions as i respond negaively for he firs 10 monhs before subsides o zero. Shock in nominal exchange raes, however, does no give significan impac on boh expor prices and money supply. These resuls imply ha impor prices are more sensiive han expor prices o shock in nominal exchange raes. However, he significan response could only be raced in shor-run bu no in long-run.

10 108 Exchange Rae Rhock on Malaysian Prices of Impors and Expors: Figure 1: Impulse Response Funcions Response o Cholesky One S.D. Innovaions ± 2 S.E..03 Response of LNPM o LNPX.03 Response of LNPM o LNM3.03 Response of LNPM o LNNEER Response of LNPX o LNM3 Response of LNPX o LNPM Response of LNPX o LNNEER As discussed earlier, he variance decomposiion is an alernaive mehod o IRF for examining he effecs of shocks o he dependen variables. I deermines how much of he forecas error variance for any variable in a sysem is explained by innovaions o each explanaory variable, over a series of ime horizons. Usually own series shocks explain mos of he error variance, alhough he shock will also affec oher variables in he sysem. From Table 4, looking along he main diagonal, he resuls reveal ha he own shock is relaively high for money supply, nominal exchange raes and expor prices. This implies he exogeneiy of hese variables in variance decomposiions as afer he shock, he variance appears o be less explained by innovaions in oher explanaory variables. On he oher hand, he resuls shows ha he percenage of variance explained by own shock for impor prices are relaively smaller especially. In fac, he own shock s conribuion is declining in long run from 92% in second monhs o 60% in 36 monhs period which indicaes ha impor prices are highly endogenous as compared o oher variables.

11 Journal of Economic Cooperaion and Developmen 109 Table 4: Variance Decomposiions Variance Decomposiion of LNM3: Period (mhly) S.E. LNPX LNM3 LNPM LNNEER Variance Decomposiion of LNNEER: Period (mhly) S.E. LNPX LNM3 LNPM LNNEER Variance Decomposiion of LNPM: Period (mhly) S.E. LNPX LNM3 LNPM LNNEER Variance Decomposiion of LNPX: Period (mhly) S.E. LNPX LNM3 LNPM LNNEER Cholesky Ordering: LNPX LNM3 LNPM LNNEER

12 110 Exchange Rae Rhock on Malaysian Prices of Impors and Expors: The VDC subsaniae he significan role played by nominal exchange raes in accouning for flucuaions in impor prices. A 2 monh horizon, he fracion of exchange raes forecas error variance aribuable o variaions in impor prices is only 2%. I hen increases a longer horizon and a 36-monh horizon he conribuions are almos 19%. Obviously, he conribuion of nominal exchange raes o flucuaions in expor prices is increasing bu less han 10%. The resuls hus srenghen he findings earlier ha nominal exchange raes significanly affec impor prices as compared o expor prices in a case of Malaysia. The ERPT on impor/expor prices are calculaed for 12-monh (1 year) and 36-monh (3 years) period using accumulaed response. ERPT on impor and expor prices are found o be less han 1 or incomplee. Specifically, ERPT on impor prices are 41% and 43% in 12 monhs and 36 monhs, respecively. ERPT on expor prices are 8% and 30%, respecively. The overall findings, hus, indicae ha, while he exchange rae shock is significanly affec he flucuaion of impor prices, he degree of passhrough is incomplee. This is due o he fac ha Malaysia is having low inflaion environmen paricularly in mid 1990s and early This low inflaion environmen iself is induced by a credible moneary policy of he counry. The average inflaion rae was 3.3 per cen during he early 1980s and declined o 2.01 per cen during This is consisen wih Ca Zorzi e.al (2007) findings ha for emerging economies wih single digi inflaion, ERPT is low and similar o ha of he developed economies. There are also possibiliies ha low ERPT is subjec o significan non-linear behaviour of he exchange rae which could probably be esed in fuure sudies. Since i is found ha ERPT is low and incomplee, depreciaion of he currency is no associaed wih an increase in impor prices or decrease in expor prices for he period under review in Malaysia. Therefore, domesic policies sill have a significan role in conrolling domesic prices. Thus, if here is an increase in inflaion environmen in he counry, i is recommended ha policy-makers o pay aenion o exchange raes, as hey have an effec on he inflaion developmen. The degree of pass-hrough is also crucial in analyzing inernaional ransmission of shocks. Wih complee pass-hrough, moneary policy shocks produce a posiive cross-counry correlaion of consumpion, bu

13 Journal of Economic Cooperaion and Developmen 111 he sign of his correlaion is reversed when pass-hrough becomes sufficienly low (See Devereux and Engel(2002)). This suggess, o he exen ha moneary policy shocks are imporan in explaining business cycles, he declining pass-hrough implies ha hose business cycles are becoming more synchronized. The moneary policy rade-off beween inflaion and oupu variabiliy is eased as he degree of ERPT decreases, since he exchange rae channel hen ransmis moneary policy, and foreign disurbances, o a smaller exen. Hence, policymakers in Malaysia may no need o be overly sensiive o exchange rae flucuaions resuling from regional or global financial crisis. The low degree of pass-hrough also implies ha he use of exchange rae adjusmen o improve rade balance may be less effecive. The exchange rae policy may be a blun insrumen when used o resore exernal balance since relaive price adjusmens is limied. Besides, he lower ERPT, he less is he subsiuion beween foreign goods and domesic consumpion and his leads o a dampening of expendiureswiching effecs. 3. Conclusion By employing he VECM model he sudy finds evidence supporing he imporance of nominal exchange rae shock o flucuaion of impor prices in a case of Malaysia afer he crisis period. This implies ha a small open economy like Malaysia is highly suscepible o exernal shock such as a shock in exchange raes. If he degree of volailiy in impor or expor prices is high due o he shock, i migh give greaer effec on real aciviies of he economy ha are exogenous o moneary policy. The exen o which exchange rae changes aler impor/expor prices is known as exchange rae pass-hrough. The sudy finds ha he degree of ERPT on boh impor and expor prices are lesser han 1 or incomplee. The findings reflec ha ERPT is endogeneous o Malaysian inflaionary environmen and moneary policy credibiliy. Thus, he policy makers are suggesed no o be very sensiive o exchange rae flucuaions. Imporanly, he use of exchange rae adjusmen o improve rade balance is found o be less effecive wih low ERPT.

14 112 Exchange Rae Rhock on Malaysian Prices of Impors and Expors: References Anaya, Jose Anonio Gonzalez (2000). Exchange Rae Pass-Through and Parial Dollarizaion: Is There a Link? CREDPR Working Paper, 81. Bank Negara Malaysia. Monhly Saisical Bullein. Various issues. Bes, C and Devereux, M.B. (2000) Exchange Rae Dynamics In A Model of Pricing-o-Marke, Journal of Inernaional Economics 50(1), Ca Zorzi, M., Elke Hahn and Marcelo Sanchez (2007). Exchange Rae Pass-Through in Emerging Markes, Working Paper Series No. 739, European Cenral Bank Campa, J.M. and L.S. Goldberg (2002). Exchange Rae Pass-Through ino Impor Prices: A Macro or Micro Phenomenon, NBER Working Paper No Devereux, M.B. and Engel, Charles (2002). Exchange Rae Pass-Through, Exchnage Rae Volailiy and Exchange Rae Disconnec, Carnegie Rocheser Conference, Carnegie Mellon Universiy. Dickey, D.A. & Fuller, W.A. (1981) The Likelihood Raio Saisics for Auoregressive Time series Wih a Uni Roo. Economerica, 49(4): Dornbusch, Rudiger (1988). Exchange Raes and Inflaion. MIT Press, USA. Inernaional Moneary Fund, Inernaional Financial Saisics, Various issues. Johansen, Søren. and Juselius, Kaarina. (1990). Maximum likelihood esimaion and inference on coinegraion: wih applicaions o he demand for money. Oxford Bullein of Economics and Saisics, 52, Johansen, S. and K.,Juselius (1992) Tesing Srucural Hypoheses in a Mulivariae Coinegraion Analysis of he PPP and he UIP for UK. Journal of Economerics, 53, Johansen, S. (1988) Saisical analysis of coinegraion vecor. Journal of Economics Dynamics and Conrol. 12:

15 Journal of Economic Cooperaion and Developmen 113 Johansen, S. (1991) Esimaion and Hypohesis Tesing of Coinegraing Vecors in Gaussian Vecor Auoregressive Models. Economerica, 59, no. 006, Kanioura, A. A Coinegraion Analysis of US Ineres Raes. Working Paper no. 3. Deparmen of Economics, Universiy of Sheffield, 2001 Kneer, Michael M. (1993). Inernaional Comparisons of Pricing-o- Marke Behavior, American Economic Review, vol. 83, Krugman, Paul K. (1987). Pricing o Marke When he Exchange Rae Changes in Real Financial Linkages Among Open Economies, ed. By Sven W. Arnd and J. David Richarson. MIT Press, Cambridge, Massachuses. Also available as NBER WP. McCarhy, Jonahan (2000). Pass-Through of Exchange Raes and Impor Prices o Domesic Inflaion in Some Indusrialized Economies, Saff Repors of Federal Reserve Bank of New York, No Obsfeld, M. and Rogoff, K. (1995) Exchange Rae Dynamics Redux, The Journal of Poliical Economy 103(3), Parsons and Sao (2005), Exchange Rae Pass-Through in Eas Asia, Cener for Inernaional Trade Sudies (CITS) Working Paper, Faculy of Economics, Yokohama Inernaional Universiy. Phillips, P.C.B. & Perron, P. (1988) Tesing for a Uni Roo in Time Series Regression. Biomerika, 75(2): Taylor, J.B. (2000) Low Inflaion, Pass Through and The Pricing Power of Firms, European Economic Review 44(7), Tille, Cedric (2000) Beggar-Thy-Neighbor or Beggar-Thyself? The Income Effec of Exchnage Rae Flucuaions, Federal Reserve Bank of New York. Yang, Jiawei (1997). Exchange Rae Pass-Through in U.S. Manufacuring Indusries, Review of Economics and Saisics, Yang, Jiawei (2003). Is Exchange Rae Pass-Through Symmeric? Evidence From US Impors, School of Business and Public Managemen, George Washingon Universiy, Washingon DC.