Exchange Rate Pass-Through to Domestic Prices in Iran ( )

Transcription

1 Iranian Economic Review, Vol.13, No.20, Fall & Winer 2007 Exchange Rae Pass-Through o Domesic Prices in Iran ( ) Ali Taiebnia Armaghan Rahimi Absrac This aricle invesigaes he pass-hrough of exchange rae shocks ino impor, wholesale and consumer price indexes in Iran by using a monhly daa se for he period The baseline analysis is carried ou wih idenified an unresriced vecor auoregressive model. Impulse response funcions show ha pass-hrough is incomplee. Moreover, he price effec of an exchange rae shock is more pronounced in he case of impor price relaive o wholesale and consumer prices. Variance decomposiion mehod indicaes exchange rae shocks are imporan in explaining he variance of he prices. Also he variance of exchange rae has he bigges share in explaining he CPI inflaion. These resuls are remarkably robus o a number of alernaive specificaions of he model. Key words: Exchange Rae, Pass-hrough, Price indexes, VAR - Corresponding Auhor, Assisan Professor of economics, Faculy of Economics, Universiy of Tehran. - MA in Economics, Faculy of Economics, Universiy of Tehran

2 86 /Exchange Rae Pass-Through o Domesic Prices in Iran ( ) 1- Inroducion Exchange rae pass-hrough ino prices plays a key role in he moneary policy ransmission. In an open economy like Iran s economy, he domesic price level canno remain immune o exernal price shocks such as exchange rae movemens and changes in impor prices. Any depreciaion or appreciaion of he exchange rae no only resuls in significan changes in he prices of impored final goods, bu also affecs impored inpus and herefore, he cos of he final goods and services. Specifically, exchange rae movemens can influence domesic prices hrough direc and indirec channels. In case of direc channel, exchange rae movemens can affec domesic prices hrough changes in he price of impored final goods and impored inpus. In general, when a currency depreciaes, i will increase impor prices while appreciaion in a price aker counry like Iran will cause impor prices o lower. The poenially higher coss of impored raw maerial and capial goods associaed wih an exchange rae depreciaion push up marginal coss and lead o higher prices of domesic produced goods. In case of indirec effec, he exchange rae depreciaion affecs he ne expors which in urn influence he domesic prices hrough he change in aggregae demand which pus upward pressure on domesic prices. In addiion, impor-compeing firms migh increase prices in response o foreign compeior price increases in order o mainain profi margins. However, The exen and he speed of exchange rae pass-hrough depends on several facors such as marke srucure, pricing policies, general inflaionary environmen, involvemen of non-radable goods in he disribuion of radable goods, relaive share of impors in WPI and CPI baske, ec. 1 The exbook definiion of exchange rae pass-hrough is he percenage change in local currency impor prices resuling from a one percen change in he exchange rae beween he exporing and imporing counries. (Goldberg and Kneer, 1997, p. 1248) 1- See Zulfiqar Hyder, 2004

3 Taiebnia, Ali & Armaghan Rahimi. / 87 Since, Changing in impor prices passed on o wholesale and consumer prices, in he presen research we use a broader definiion of exchange rae pass-hrough, which is seen as he change in impor and hen domesic prices. Thi research examines how he exchange rae changes affec price indexes in Iran. Sspecifically, we have focused on he changes in impor, wholesale and consumer price indexes by analyzing monhly daa for he period 1990:3 o 2006:6 and he baseline empirical model is a Vecor Auo Regressive (VAR) model. In order o quanify he effec of an exchange rae shock on prices, Impulse-Response Funcions are used. Then, we apply he Variance Decomposiion mehods o rank he relaive share of exchange rae movemens for he explanaion of changes in price indexes. This sudy conains 7 secions. In he nex secion, we have a review of he relevan lieraure. Adoped empirical approach is inroduced in secion 3. The daa se and preliminarily saisical and he economeric model are also presened in his secion. Secion 4 coninues by he analysis and empirical resuls of he sudy, including he resuls of esimaion, applying he impulse-response funcions and Variance Decomposiion mehods. In secion 5, we explain he sensiiviy analysis. Summery and conclusion are presened on secion 6. Finally, references are menioned in secions Review of Lieraure There are various heoreical researches abou pass-hrough exchange rae. Some of hem ry o explain exchange rae pass-hrough by microeconomic models and he ohers discuss i in he conex of macroeconomics. For insance, Goldberg and Kneer (1996) provides a microeconomic heoreical explanaion for he incompleeness of passhrough from an indusrial-organizaion perspecive. This research indicaes ha common currency relaive prices for similar goods expored o differen markes are highly correlaed wih exchange raes beween hose markes. I concludes ha incomplee pass-hrough is a consequence of imperfec compeiion i.e. hird-degree price discriminaion. Krugman (1987) inroduces he concep of pricing-o-marke, which sands for exchange rae induced price discriminaion across counries. In oher words, i describes a siuaion where exporing firms adjus heir

4 88 /Exchange Rae Pass-Through o Domesic Prices in Iran ( ) (desinaion-specific) markups o compensae for exchange rae changes. Pricing-o-marke may hus be considered as a microeconomic explanaion for incomplee exchange rae pass-hrough ino impor prices. By conras, radiional open macroeconomic models posulae perfec compeiion, fully flexible prices and purchasing power pariy. In hese models, he pass-hrough is necessarily complee. For example, Obsfeld and Rogoff (1995) inroduces nominal rigidiies and marke imperfecions ino a microfounded dynamic general equilibrium model. However, purchasing power pariy is sill mainained a all imes, and he pass-hrough is complee. In Obsfeld and Rogoff s model, nominal prices are se in producers currencies (producer currency pricing) and consequenly, nominal exchange rae flucuaions cause one-o-one reacions in prices of impored goods ha means in he shor-run exchange rae pass-hrough is complee. Bu sudies on exchange rae pass-hrough have almos unanimously rejeced he assumpion of purchasing power pariy and, hus, a one-o-one relaionship beween exchange rae changes and changes in he prices, paricularly in he shor run. If significan lags exis in he ransmission of exchange rae changes o domesic prices, exchange-rae depreciaion would only have limied impac on he rae of domesic inflaion. Menon (1995a) which is one of he mos comprehensive surveys of he lieraure on exchange rae pass-hrough presens an overview of 43 empirical sudies on indusrialized economies, of which he mos ofen sudied is he Unied Saes. The majoriy of sudies conclude ha exchange rae pass-hrough is incomplee, indeed alhough he degree of pass-hrough does vary significanly across differen counries. The size and he openness of he individual economies are he main facors ha influence he degree of pass-hrough across counries. I furhermore, repors ha pass-hrough relaionships have remained almos sable over he ime.1 Differen resuls for a counry sem primarily from he use of differen model specificaions and variable selecions and differen mehodologies for esimaing he models, raher han from differen ime periods sudied. Some sudies have 1- See also Parsley (1993). Some sudies have, neverheless, challenged his resul. See, for example, Taylor (2000), and Gagnon and Ihrig (2001)

5 Taiebnia, Ali & Armaghan Rahimi. / 89 also found pass-hrough o be asymmeric, which implies ha he rae of pass-hrough is differen during exchange rae depreciaions and appreciaions. 1 All excep one of he sudies reviewed by Menon (1995a) use he OLS esimaion echnique and hey did no consider ime-series properies of he daa, paricularly he non-saionariy. By ha ime, Kim (1991) had been he only sudy on exchange rae pass-hrough which applied a vecor auoregressive (VAR) framework. I addresses he problem by employing mehod developed by Johansen (1988). In a sysem of five-equaion vecor error correcion model, his paper finds ha he US inflaion, exchange rae, money supply, income, and ineres rae are coinegraed. The coinegraion analysis of he daa shows ha he dollar exchange rae has a significan negaive impac on he inflaion measured by he producer price index. I is furher esablished ha he exchange rae Granger causes he inflaion. Afer Kim (1991), many sudies use VAR frameworks o invesigae exchange rae pass-hrough such as Kenny and McGeigan (1998), Rowland (2003), Hyder (2004) and Choudhri, E. U., Faruqee, H. and D. S. Hakura. (2005). A more recen sream of lieraure examines he pass-hrough exchange rae in conex of cross-counry analysis. These researches like McCarhy (2000) and Mihailov (2005) sudy he pass-hrough ino differen prices for differen counries. In general, hey find ha he pass-hrough ino consumer prices, which is he major concern for moneary policy, is small and ofen even insignifican. In summaries, McCarhy (2000) examines he impac of exchange raes and impor prices on he domesic PPI and CPI in seleced indusrialized economies. The empirical model is a VAR incorporaing a disribuion chain of pricing. Esimaing he model over he pos-breon Woods era, impulse responses indicae ha exchange raes have a modes effec on domesic price while impor prices have a sronger effec. I finds ha pass- hrough is 1- Some of hese sudies are Mann (1986), Kreinin, Marin and Sheehey (1987), and Marson (1990). However, some ohers have found no evidence for such an asymmery, like Ahukorala (1991)

6 90 /Exchange Rae Pass-Through o Domesic Prices in Iran ( ) larger in counries wih a larger impor share. I is also larger in he counries wih more persisen exchange raes and impor prices. Mihailov (2005) compares exchange rae pass-hough on aggregae prices in he US, Germany and Japan across a number of dimensions. By using he recen empirical approach, he conribuion of his sudy is o perform a pass-hough sensiiviy analysis of alernaive pass-hrough esimaes. Mihailov (2005) finds ha he economeric mehods, daa and frequency and variable proxy employed maer for precision of deails, ye hey ofen agree on some general rends. Thus his research indicaes ha pass-hrough on impor prices has declined in he 1990s relaive o 1980s, pass-hough on expor prices remains counry-specific and pass-hrough on consumer prices is nowadays negligible in all hree considered counries. Feinberg (2000) and Hüfner and Schröder (2002) are some oher researches which apply he cross-counry analysis. In order o deermine he linkage beween exchange rae movemens and domesic producer prices, Feinberg (2000) examines Colombia, Korea, and Morocco as hree developing counries and Hüfner and Schröder (2002) sudies he passhrough of exchange rae changes o consumer prices for he euro area by esimaing vecor error correcion models for Germany, France, Ialy, he Neherlands and Spain. In hese researches, generally, incomplee passhrough ino domesic prices is found and Feinberg (2000) finds his impac is greaer for developed economies. In addiion, Campa and S. Goldberg (2005) provides cross-counry and ime series evidence on he exen of exchange rae pass-hrough ino he impor prices of 23 OECD counries. I finds compelling evidence of parial pass-hrough in he shor run, especially wihin manufacuring indusries. I also shows ha over he long run, producer-currency pricing is more prevalen for many ypes of impored goods. This research concludes ha counries wih higher raes of exchange rae volailiy have higher passhrough elasiciies and macroeconomic variables have played a minor role in he evoluion of pass-hrough elasiciies over ime.

7 Taiebnia, Ali & Armaghan Rahimi. / Empirical Approach In order o examine he exchange rae pass-hrough ino he prices in each sage of disribuion, his research uilizes an unresriced Vecor Auoregressive (VAR) model which accoun for endogeneiy of he variables. The major benefi from using unresriced VAR model is ha i remains usable when heoreical prescripions are insufficien. Tha is why we use his model and preven o do Johansen procedure or even srucural VAR models. Afer esimaing he VAR model, impulse-response funcions will be used. A he end, in order o quanify he effec of an exchange rae shock on prices we will apply Variance Decomposiion mehods o rank he relaive share of exchange rae movemens for explanaion of changes in he price indexes. Chosen Variables are moivaed by many recen researches especially by McCarhy (2000) and Faruqee, e al (2005) Daa and Preliminarily Saisical Properies In his sudy, daa is colleced from saisical resources published by cenral bank of Iran (CBI). Whole sample includes 196 monhly observaions for he period 1990:3 o 2006:6 Our model is based on he following series: Nominal Exchange Rae (EXR), Impor Price Index (MPI), Wholesale Price Index (WPI), Consumer Price Index (CPI), Oupu Gap (GAP) and broad Money (M2). The firs four variables are he cener variables in our analysis. Oupu Gap which is defined as he difference beween poenial and acual oupu, added o he model for considering real flucuaions in he economy. In order o consider he effec of moneary policy, we also add Money variable o our model. All series, excep he oupu gap are used in logarihm form 2. A deailed descripion of he daa is provided in Appendix A. 1- Faruqee, Hakura & Choudhri (2005) examines he performance of differen new open economy macroeconomic models in explaining he exchange rae pass-hrough in a wide range of prices. In his research prediced responses are compared wih he evidence based on VAR models o see how well differen models fi he daa. 2- Since oupu gap is negaive for some years, i can no be used in logarihm form.

8 92 /Exchange Rae Pass-Through o Domesic Prices in Iran ( ) Before we proceed o esimae he model, i is imporan o find he order of inegraion of each variable and hen selec he opimal lag lengh for he VAR model. The resuls of Augmened Dickey Fuller (ADF) and Philips Peron (PP) uni roo ess 1, used o deermine he saionariy of he variables, are presened in ables (B-1) and (B-2). They show ha all variables are I (1) i.e. inegraed of degree one excep GAP which is saionary. The opimal lags for esimaing he VAR model are chosen based on wha indicaed by he majoriy of available crieria. (Opimal lags are 8) 3-2- Economeric Model Our VAR model is esimaed wih six endogenous variables. The reduced form represenaion of he model can be wrien as Z( L) y = c + ε (1) Where E( ε ε ) = Σ Σ = σ ij, i, j = 1, 2,..., T Where y = [ DLEXR, DLMPI, DLWPI, GAP, DLM 2, DCPI ] is he vecor of m=6 endogenous variables, c is a vecor of consans and ε is a vecor of residuals. DL in he beginning of he name of each variable denoes difference of ha variable in naural logarihm form. p Z(L ) = I + Z L + Z L Z L 1 2 p is a marix polynomial in lag operaor and in our model p is equal 2. The corresponding Vecor Moving Average of his model is: 2 y = B( L) E( u u ) I u = 1- See Dickey and Fuller (1979) and Phillips P. C. B (1987). You can also see Hamilon (1994), pp See for example Lukepohl (2005, chaper 2) and Hamilon (1994, chaper 11 and 12)

9 Taiebnia, Ali & Armaghan Rahimi. / 93 ( = k k = 0 k Where B L) B L = B u + B u + B u +... To idenify he srucural shock, we need some resricions. This idenificaion can be achieved by seing B 0 equals o he Cholesky decomposiion ha means Σ = B 0B 0. In his sudy, he following ordering of he variables has been chosen for he Cholesky decomposiion: DLEXR DLMPI DLWPI GAP DLM2 DLCPI Choosing of his ordering is based on boh heoreical and empirical reasons. The heoreical reasons were explained expansively in secion 1. In addiion, since changing in wholesale prices affec GDP hrough he cos of producion, i is considered before GAP. DLM2 is pu afer GAP and almos a he end of raking because we assume a reacive funcion for he cenral bank, as i is noed by McCarhy (2000). In he empirical researches, he decision on he ordering of he variables is ofen based on he pairwise granger causaliy ess. The resuls of doing hese ess for our variables are repored in Table (B-3) and (B-4). They confirm chosen causaliy beween all variables in he above ordering. 4- Resuls In he curren secion we explain Impulse-Response Funcions (IRF) and Variance Decomposiion resuls Impulse-Response Funcions Afer esimaing he explained VAR model, he Impulse-Response Funcions (IRF) have been used o quanify he degree of exchange rae pass-hrough. By his mehod, we can esimae he effec of an imposed innovaion o he equaion of exchange rae ino he oher variables of he sysem like as impor, wholesale and consumer prices. Since he imposed shocks (impulses) in he log-differenced exchange rae variable are normalized o one, we can inerpre he esimaed responses as he passhrough elasiciy.

10 94 /Exchange Rae Pass-Through o Domesic Prices in Iran ( ) IRF resuls for he period of sudy are repored in figures (A-2-1), (A-2-2), (A-2-3),(A-2-4) and ables (B-5), (B-6). Also, he exchange rae passhrough ino differen prices is paerned a appendix B.1. These resuls indicae ha all prices response quickly o an exchange rae shock. The accumulaive exchange rae pass-hrough amouns o 27 percen o impor prices and 17 percen for boh wholesale and consumer prices afer 7 monhs. I increases such ha afer 12 monhs, i comes o 40.6, 27.5 and 21 percen for MPI, WPI and CPI respecively and almos afer 24 monhs (wo years) i is going o be consan a 61.8, 43 and 33.5 percen for hese indexes. These amouns are significanly differen from zero bu less han one. Therefore, he exchange rae pass-hrough is incomplee. Figure (B-6) compares he exchange rae pass-hrough for he differen prices. I exhibis ha he exchange rae shock is more pronounced in he case of MPI relaive o WPI and CPI. Consumer Price Index (CPI) is affeced less han oher price indexes. 1 This evidence is obvious because he share of radable commodiies in MPI is bigger ha in WPI and CPI. Moreover, CPI includes services ha are generally no raded and so less affeced by exchange rae changes direcly Variance Decomposiion Alhough he impulse responses indicae he exen of pass-hrough o he prices, hey do no indicae how imporan hese shocks have been in price flucuaions. If he exchange rae shocks in a counry are small, hen pass-hrough could be large bu exchange rae would have lile influence on domesic inflaion. In order o invesigae he imporance of exchange rae volailiy on inflaion, we examine he variance decomposiions of he price variables. Resuls of Variance decomposiion, which show he conribuion of innovaion in he exchange rae o he variabiliy of MPI, WPI and CPI, are presened in ables (B-7), (B-8) and (B-9). As we can see, he exchange rae shocks are imporan in explaining he variance of he prices. For insance, a 1- This resul is in line wih many oher exchange rae pass- pass hrough researches. See for example McCarhy (2000), Choudhri, Faruquee and Hakura (2005) and Mihailov (2005)

11 Taiebnia, Ali & Armaghan Rahimi. / 95 a six monh forecas horizon, heir share is abou 20.5,19 and 25 percen for MPI, WPI and CPI respecively. Therefore, alhough he pass-hrough exchange rae ino CPI is less han oher price indexes, he variance of exchange rae has he bigges share in explaining he CPI inflaion. The percenage of variance explained by he exchange rae comes ino he highes amoun afer four monhs for CPI (26.29 percen), eigh monhs for WPI (19 percen) and nine monhs for MPI (22 percen). Then i declines as he forecas horizon increases. 5- Sensiiviy Analysis One way for esing he robusness of he baseline resuls is examining he sensiiviy of he esimaed model wih respec o he differen idenificaions. In our baseline model all of variables are nominal excep GAP. The alernaive model is he same as he explained model in secion 3.2 bu he difference is in endogenous vecor y. In he alernaive model, we consider y as follows: y = [ DLEXR, DLMPI DLWPI, DLM 2, DCPI ]. Tha is we esimae he model wihou real variable and omi he GAP variable. Tables (B-10) and (B-11) and figure (A-2-5) compare he resuls of alernaive esimaed model wih he resuls of he baseline model. To prevening he confusion, variables of he model wihou GAP are noed wih*. As i is shown, excep for CPI, whose resuls changed slighly, here is no significan difference beween he resuls of wo models. We also esimaed he baseline model wih differen Cholesky ordering of he variables proposed by he Granger Causaliy es. In all of hese cases he resuls are no noiceable differen. For abbreviaion, we do no repor hem. Since he baseline VAR model is no sensiive o he differen specificaions, he resuls of variance decomposiion, Impulse Response Funcions (IRF) and he esimaed pass- hrough elasiciies are reliable. 6- Summary and Conclusion This sudy invesigaed he pass-hrough of exchange rae shocks ino impor, wholesale and consumer price indexes in Iran by using a monhly

12 96 /Exchange Rae Pass-Through o Domesic Prices in Iran ( ) daa se for he period Firs, we reviewed he relevan lieraure. Then, empirical approach was explained. The baseline analysis was carried ou wih idenified a Vecor Auoregressive (VAR) model and pass-hrough effecs were quanified by means of impulse response funcions. Evidence showed ha pass-hrough elasiciies are significanly differen from zero bu less han one. Therefore, hey are incomplee. Moreover, exchange rae shock is more pronounced in he case of impor price relaive o wholesale and consumer prices such ha afer wo years, he accumulaive pass- hrough comes o 61.8,43 and 33.5 percen for MPI, WPI and CPI respecively. To invesigae he imporance of exchange rae volailiy on inflaion, we examined he variance decomposiions of he price variables. This mehod indicaed ha exchange rae shocks are imporan in explaining he variance of he prices and alhough he pass-hrough exchange rae ino CPI is less han oher price indexes, he variance of exchange rae shocks has he bigges share in explaining he CPI inflaion. These resuls were remarkably robus o a number of alernaive specificaions of he model. For insance, we esimaed he baseline model wihou GAP variable as he represenaive of real flucuaions and he resuls were no significanly differen. References 1- Ahukorala, Premachandra (1991), Exchange Rae Pass-Through: The Case of Korean Expors of Manufacures, Economic Leers, Vol. 35, No. 1, January, pp Campa, J. M. and L. S. Goldberg. (2005), Exchange rae pass-hrough ino impor prices, The Review of Economics and Saisics, Vol. 87, No. 4, pp Cenral Bank of Iran (CBI), Economic Trends, No Choudhri, E. U., Faruqee, H. and D. S. Hakura. (2005), Explaining he exchange rae pass-hrough in differen prices, Journal of Inernaional Economics 65(2):

13 Taiebnia, Ali & Armaghan Rahimi. / Dickey, David N., and Wayne A. Fuller. (1979), Disribuion of he Esimaors for he Auoregressive Time Series wih a Uni Roo, Journal of American Saisical Associaion, Vol. 74, pp Feinberg, Rober M. (2000), The Role of Inernaional Discipline in Three Developing Economies: Exchange Rae Effecs on Domesic Prices in Colombia, Korea and Morocco, Review of Inernaional Economics, Vol. 8, No. 1, pp Gagnon, Joseph E. and Jane Ihrig (2001), Moneary Policy and Exchange Rae Pass-Through, Inernaional Finance Discussion Paper No. 704, Board of Governors of he Federal Reserve Sysem, Washingon D.C. 8- Goldberg, Pinelopi K. and Michael M. Kneer (1997), Goods Prices and Exchange Raes: Wha Have We Learned?, Journal of Economic Lieraure, Vol. 35, No. 3, Sepember, pp Hamilon, James (1994), Time Series Analysis, Princeon, New Jersey: Princeon Universiy Press. 10- Hüfner, Felix P., and Michael Schröder (2002), Exchange Rae Pass- Through o Consumer Prices: A European Perspecive, Discussion Paper No , Cenre for European Economic Research. 11- Hyder, Zulfiqar, and Shah, Sardar, (2004), Exchange Rae Pass- Through o Domesic Prices in Pakisan, Working Paper No. 5, Research Deparmen, Sae Bank of Pakisan. 12- Kenny, Geoff, and Donal McGeigan (1998), Exchange Raes and Impor Prices for a Small Open Economy: The Case of Ireland, Applied Economics, Vol. 30, No. 9, Sepember, pp Kim, Ki-Ho (1998), US Inflaion and he Dollar Exchange Rae: A Vecor Error Correcion Model, Applied Economics, Vol. 30, No. 5, May, pp Mann, Caherine L. (1986), Prices, Profi Margins and Exchange Raes, Federal Reserve Bullein, Vol. 72, No. 6, June, pp Menon, Jayan (1995a), Exchange Rae Pass-Trough, Journal of Economeric Survey, Vol. 9, No. 2, June, pp Menon, Jayan (1995b), Exchange Rae Pass-Trough Prices for a Small Open Economy, Applied Economerics, Vol. 27, No. 3, March, pp

14 98 /Exchange Rae Pass-Through o Domesic Prices in Iran ( ) 17- Mihailov, Alexander (2005), Exchange Rae Pass-Through on Prices in Macro daa: A Comparaive Sensiiviy Analysis, Discussion Paper No. 568, Universiy of Essex, Deparmen of Economics. 18- McCarhy, Jonahan (2000), Pass-Through of Exchange Raes and Impor Prices o Domesic Inflaion in Some Indusrialized Economies, Working Paper No. 79, Bank for Inernaional Selemens, Basel. 19- Rowland, Peer (2003), Exchange Rae Pass-Through o Domesic Prices: The Case of Colombia, Banco de la Republica. 20- Taylor, John B. (2000), Low Inflaion, Pass-Through, and he Pricing Power of Firms, European Economic Review, Vol. 44, No. 7, June, pp

15 Taiebnia, Ali & Armaghan Rahimi. / 99 Appendix A- Daa Appendix A.1- Definiions EXR: Nominal (Marke) Exchange Rae in erms of Dollar/Rial, Monhly, Economic Trends, Published by Cenral Bank of Iran (CBI) MPI: Impor Price Index, Quarerly, (1997=100), Published a Economic Trends, Cenral Bank of Iran (CBI) WPI: Impor Price Index, Quarerly, (1997=100), Published a Economic Trends, Cenral Bank of Iran (CBI) CPI: Consumer Price Index, Quarerly, (1997=100), Published a Economic Trends, Cenral Bank of Iran (CBI) GAP: Defined as he difference beween poenial and acual Gross Domesic Producion Acual GDP: Gross Domesic Producion, a consan prices 1997, afer seasonal adjusmen in erms of billion Rials, Quarerly, Published a Economic Trends, Cenral Bank of Iran (CBI) Poenial GDP: Calculaed by using Hodrick-Presco Filer (lambda=1600) for acual GDP M2: Liquidiy in erms of Billion Rials, Quarerly, Published a Economic Trends, Cenral Bank of Iran (CBI) Noe: Dae of he beginning of he year in Iran is 21 March.

16 100 /Exchange Rae Pass-Through o Domesic Prices in Iran ( ) A- Resuls Appendix A.2- Figures Exchange Rae Pass-Through ino Impor Price Index (MPI),Wholesale Price Index (WPI), Consumer Price Index (CPI) Figure(A-2-2):Accumulaed Response o one Uni DLEXR 1.6 Accumulaed Response of DLMPI o DLEXR 1.2 Accumulaed Response of DLWPI o DLEXR 1.0 Accumulaed Response of DLCPI o DLEXR DLEXR ± 2 S.E. DLEXR ± 2 S.E. DLEXR ± 2 S.E. Figure(A-2-1):Response o one Uni DLEXR Innovaions.4 Response of DLMPI o DLEXR Response of DLWPI o DLEXR.16 Response of DLCPI o DLEXR DLEXR ± 2 S.E. DLEXR ± 2 S.E. DLEXR ± 2 S.E.

17 Taiebnia, Ali & Armaghan Rahimi. / Percen DLMPI DLWPI DLCPI Monh Figure (A-2-3): Exchange Rae Pass-Through ino Price Indexes Percen DLMPI DLWPI DLCPI Monh Figure (A-2-4): Accumulaed Exchange Rae Pass-Through ino Price Indexes ] DLMPI DLMPI*

18 102 /Exchange Rae Pass-Through o Domesic Prices in Iran ( ) Figure (A-2-5): Accumulaed Exchange Rae Pass-Through ino Price Indexes in he Models wih and wihou GAP DLWPI DLWPI* DLCPI DLCPI* B- Tables Variable Table ( B-1) Augmened Dickey Fuller Uni Roo Tes (ADF) Wih Inercep Wih Inercep and Trend Level Firs Difference Level Firs Difference LEXR ADF(0)=-1.41 ADF(0)= ADF(1)=-0.92 ADF(0)= LMPI ADF(2)=-2.99 ADF(1)=-8.02 ADF(0)=-1.10 ADF(1)=-8.75 LWPI ADF(0)= ADF(0)=-0.25 ADF(0)= GAP ADF(2)= ADF(2)= LM2 ADF(12)=-0.02 ADF(11)=-2.67 ADF(12)=-2.41 ADF(11)=-2.66 LCPI ADF(1)=-2.16 ADF(0)=-8.62 ADF(1)=-0.57 ADF(0)=-8.94 Noe: The value in parenheses is he order of he lag used, which is decided by using he Schwarz informaion crieria. The null hypohesis in each case is ha he variable is inegraed of order one and herefore i is non-saionary. In he firs case (Wih Inercep) he 5% rejecion region for he Dickey-Fuller saisic is ADF < -2.89, and he 1% rejecion region is ADF < In he firs case (Wih Inercep and Trend) he 5% rejecion region for he Dickey-Fuller saisic is ADF < and he 1% rejecion region is ADF <

19 Taiebnia, Ali & Armaghan Rahimi. / 103 Table (B-2) Philips- Peron Uni Roo Tes (PP) Variable Wih Inercep Wih Inercep and Trend Level Firs Difference Level Firs Difference LEXR PP(2)=-1.60 PP(6)= PP(2)=-0.75 PP(7)= LMPI PP(6)= PP(6)=-1.15 PP(7)= LWPI PP(7)=-3.41 PP(8)= PP(7)=-0.49 PP(7)= GAP PP(4)= PP(4)= LM2 PP(11)=0.39 PP(10)= PP(3)=-2.41 PP(10)= LCPI PP(5)=-2.08 PP(1)=-8.59 PP(5)=-0.26 PP(3)=-8.91 Noe: he value in parenheses is he number of bandwidh (according o he Newey- Wes using Barle kernel). The null hypohesis in each case is ha he variable is inegraed of order one and herefore i is non-saionary. In he firs case (Wih Inercep) he 5% rejecion region for he Philips-Peron saisic is PP < -2.89, and he 1% rejecion region is PP < In he firs case (Wih Inercep and Trend) he 5% rejecion region for he Philips-Peron saisic is PP < and he 1% rejecion region is PP < Table (B-3): Resuls of Pairwise Granger Causaliy Tess Variable DLEXR DLMPI DLWPI GAP DLM2 DLCPI DLEXR DLMPI DLWPI GAP DLM DLCPI Noe: indicaes he direcion of Granger Causaliy and means variable in row does no cause variable in column.

20 104 /Exchange Rae Pass-Through o Domesic Prices in Iran ( ) Table (B-4): Pairwise Granger Causaliy Tess Null Hypohesis: Obs F-Saisic Probabiliy DLEXR does no Granger Cause DLCPI DLCPI does no Granger Cause DLEXR DLM2 does no Granger Cause DLCPI DLCPI does no Granger Cause DLM DLMPI does no Granger Cause DLCPI DLCPI does no Granger Cause DLMPI DLWPI does no Granger Cause DLCPI DLCPI does no Granger Cause DLWPI E-06 GAP does no Granger Cause DLCPI DLCPI does no Granger Cause GAP DLM2 does no Granger Cause DLEXR DLEXR does no Granger Cause DLM DLMPI does no Granger Cause DLEXR DLEXR does no Granger Cause DLMPI DLWPI does no Granger Cause DLEXR DLEXR does no Granger Cause DLWPI GAP does no Granger Cause DLEXR DLEXR does no Granger Cause GAP DLMPI does no Granger Cause DLM DLM2 does no Granger Cause DLMPI DLWPI does no Granger Cause DLM DLM2 does no Granger Cause DLWPI GAP does no Granger Cause DLM DLM2 does no Granger Cause GAP DLMPI does no Granger Cause DLWPI DLWPI does no Granger Cause DLMPI GAP does no Granger Cause DLMPI DLMPI does no Granger Cause GAP GAP does no Granger Cause DLWPI DLWPI does no Granger Cause GAP Exchange Rae Pass-Through o ino Price Indexes

21 Taiebnia, Ali & Armaghan Rahimi. / 105 Table (B-5): Response of DLMPI, DLWPI and DLCPI o One Unie Exchange Rae Innovaion Period DLMPI DLWPI DLCPI Table (B-6): Accumulaive Response of DLMPI, DLWPI and DLCPI o One Unie Exchange Rae Innovaion Period DLMPI DLWPI DLCPI Noe: In Varianec Decomposiion considered Cholesky Ordering is:dlexr DLMPI DLWPI GAP DLM2 DLCPI

22 106 /Exchange Rae Pass-Through o Domesic Prices in Iran ( ) Table(B-7): Variance Decomposiion of DLMPI Period S.E. DLEXR DLMPI DLWPI GAP DLM2 DLCPI Table(B-8): Variance Decomposiion of DLWPI Period S.E. DLEXR DLMPI DLWPI GAP DLM2 DLCPI

23 Taiebnia, Ali & Armaghan Rahimi. / 107 Table(B-9):Variance Decomposiion of DLCPI Period S.E. DLEXR DLMPI DLWPI GAP DLM2 DLCPI Exchange Rae Pass-Through o ino Price Indexes in he Model wihou GAP Table (B-10):Response of DLMPI*, DLWPI* and DLCPI* o one uni DLEXR Innovaion Period DLMPI* DLWPI* DLCPI* E

24 108 /Exchange Rae Pass-Through o Domesic Prices in Iran ( ) Table (B-11):Accumulaed Response of DLMPI*, DLWPI*, DLCPI* To on uni DLEXR Innovaion Period DLMPI* DLWPI* DLCPI*