Relationship between Trade Openness and Inflation: Empirical Evidences from Pakistan ( )

Transcription

1 The Pakisan Developmen Review 50:4 Par II (Winer 2011) pp Relaionship beween Trade Openness and Inflaion: Empirical Evidences from Pakisan ( ) SEHAR MUNIR and ADIQA KAUSAR KIANI * 1. INTRODUCTION Inflaion has always been an imporan issue for he policy-makers as i creaes uncerain siuaion in he economy ha may badly affec economic growh. Therefore, high and sable economic growh in addiion wih low inflaion is he main objecive of macroeconomic policies. Sric moneary policy wih fiscal consolidaion appears o have conribued o low price levels. The concern wih inflaion has no only o balance whole macroeconomic siuaion, bu also from he fac ha increase in inflaion raes hurs he poor severely as heir consumpion baske becomes significanly decreased. A general rise in prices in he economy is usually called inflaion. Inflaion was occurred due o some demand and supply side facors. Inflaion can be resuled due o supply shocks of differen food iems and world wide oil prices. Rising oil prices always increase prices of almos all oher commodiies for consumers. These supply shocks are volaile and can occur huge changes in food and oil prices. There are following demand side issues which increase price level in Pakisan. Firsly, increased local demand due o foreign remiances and demand managemen policies oupaced he local producion, esablishing posiive oupu gap, which in urn pu burden on prices o increase. Growh in privae consumpion remained above 10 percen during 2003 o 2006, showing sympoms of demand side burdens on prices. [Khan, Bukhari, and Ahmed (2007)] Secondly, he widening gap among local demand and producion was filled by growh in oal impors; i was increases above 40 percen in FY05 and by 24 percen in FY06 as compare o ha gap of impors, expors increased by only10 percen in FY05 and 13 percen in FY06. 1 Which resul ino increase in rade defici and high expeced inflaion in fuure? Thirdly, broad fiscal policy enhances local demand and add burden on curren accoun defici. This means, i increases gap among saving and invesmens, which has o be financed. Moreover, financing of fiscal defici hrough money creaion adds inflaionary burden. On he oher side, governmen borrowing from Sae Bank of Sehar Munir <ssahar_munir@yahoo.com> is MS Fellow a Inernaional Islamic Universiy (IIUI), Islamabad. Adiqa Kausar Kiani <a.kiani@fuuasisb.edu.pk> is Associae Professor, Head of Economics Deparmen a Federal Urdu Universiy of Ars, Science and Technology (FUUAST), Islamabad. 1 Economic Survey

2 854 Munir and Kiani Pakisan (SBP) also increased, which have serious effecs on price level. Fourhly, broad moneary policy wih high growh rae in money supply and loose credi policy was also conribuing o large prices. [Khan, Bukhari, and Ahmed (2007)] The exensive survey of Inernaional Moneary Fund, suggess ha excessive credi growh in developing counries can have bad impacs on real variables. Increasing impor prices is also a major reason in enhancing inflaion and in his scenario he depreciaing exchange rae can pu upward pressure on prices. 2 Similarly, Khan and Qasim (1996) and Hasan, e al. (1995) suggesed ha indirec axes are also he basic reason of inflaion in Pakisan. Trade Openness is defined as a phenomena of sharp economic inegraion beween counries capure hrough rade liberalisaion, invesmen and capial flows, as well as echnological changes 3. Trade Openness associaion wih falling prices is he mos popular proposiions found in inernaional rade and here has been unique urn in favor of higher economic inegraion of world. Openness suggess he economic benefi from inernaional rade, inernaional capial ransacions, and he inernaional exchange of knowledge and informaion. The lower he hurdles o inernaional rade ransacions he higher level of inegraion and benefis. The new growh heory suggess ha openness widens he marke, induc an increase in developmen, reallocaes employmen o new aciviies ha need more human capial and enhances knowledge flow beween counries. Oher han benefis, some expenses are also aached wih i. A main problem arises from decreasing rade hurdles is he loss in ariff revenue ha is percen of governmen revenue in developing economies. If ariffs are decreased or vanished, hese economies will have o implemen oher axes in order o keep heir budges a desire level. Objecives of he Sudy: The main objecive of his research is o deermine he naure of he relaion among inflaion and rade openness for Pakisan. The core focus of his sudy is o apply he coinegraion approach of Johansen (1998) and Johansen and Juselius (1990) in order o examine wheher he Romer s findings (1993), ha he negaive link among inflaion and rade openness, holds for Pakisan or no. Hypohesis: The null hypohesis (H0) of his sudy is o esimae he exisence of Romer s Hypohesis in Pakisan and alernaive hypohesis (H1) is oherwise. 2. REVIEW OF LITERATURE Romer (1993) esed he hypohesis ha here was negaive relaionship beween rade openness and inflaion. Romer s regressing inflaion on openness for cross secional daa of 114 economies over he Pos-Breon Woods period. 4 He assessed he srong relaionship beween inflaion and openness in poliically unsable counries wih independen cenral banks. Lane (1997) emphasised on differen channel hrough which openness and inflaion relaed, especially he degree of imperfec compeiion, degree of cenral bank independence, poliical insabiliy and price rigidiy in he non-raded secor.15-years average annual daa from 1973 o 1988 have underaken for cross secional analysis using 2 IMF (2004). 3 Torres (2001). 4 From 1973 o he early 1990s.

3 Relaionship beween Trade Openness and Inflaion 855 OLS and finding shows he saisically significan negaive link beween openness and inflaion. Terra (1998) challenged Romer s empirical findings using regression on 20 sample counries which were dividing ino 4 groups according o indebness level. The ime frames used in sudy were pre-deb crisis 5 and deb crisis period 6 for severely, moderaely and less debed counries. Negaive bu significan link beween inflaion and openness was found among severely indebed counries in Lain America bu ha was no exiss in moderaely and less debed counries. Bleaney (1999) esimaed relaionship beween inflaion and rade-openness for 100 counries hrough regression from and Resuls indicaed he negaive correlaion beween inflaion and openness for cross-secional daa of 1970s and 1980s ha has disappeared in 1990s. The same resuls were obained if per capia income levels, populaion, area and exchange rae regimes were conrol. 7 Cavallari (2001) insered he relaion of rade openness and inflaion in monopolisic producion model and unionised labour marke of domesic secor. The resul of heoreical model showed ha rade openness can affec inflaion in a posiive or negaive way and final resul depends on level of concenraion of wage bargaining in counry. Resuls indicaed ha in counries where wage bargaining concenraed here did no exiss any relaion among openness and inflaion. However, in counries where wage bargaining decenralised, here exiss negaive link beween openness and inflaion. Alfaro (2001) esimaed panel daa of 146 counries from by using fixed effec of counry and ime effec regression among openness and inflaion. Resuls indicaed ha in he shor run, here was no influence of openness on inflaion and fixed exchange rae was an imporan facor o reduce inflaion. In he long run, she concluded ha negaive and saisically significan relaionship exised among openness and inflaion. Temple (2002) ried o esablish relaion of rade openness and he Phillips curve for 44 counries from Regressions resuls indicaed ha Phillips curve will be more inclined in open economies. Ashra (2002) used muliple regressions by aking panel daa from 1980 and 1990 of 15 counries o discuss relaion beween inflaion and openness. He concluded ha inflaion was effeced by openness no maer eiher an economy possessing hyper-inflaion or i is big. Jin (2002) focused on he openness-growh and openness-inflaion relaions for Korea by applying variance decomposiions (VDC s) 9 and impulse response funcions (IRF s) 10 which were based on moving averages of quarerly daa from o As a resul of disinflaion in indusrial counries, he negaive correlaion beween per capia GDP and inflaion was srong in , whereas i was weak in Phillips curve slope aached wih openness is depend on small open economy sysem wih nominal rigidiy. 9 Shows he quaniy of informaion of each variable conribues o he oher variables in a vecor auoregression (VAR) models. I deermines how much error variance of each variable can be explained by exogenous shocks o oher variables. 10 Impulse response funcions show he effecs of shocks on he adjusmen pah of he variables. I shows how an unexpeced change in one variable a he beginning affecs anoher variable wih he passage of ime. In ime series analysis i is imporan in deermining he effecs of exernal shocks on he variables of he sysem.

4 856 Munir and Kiani Resuls of IRF s indicaed ha openness has inverse impacs on oupu growh bu no long run effecs, i furher showed ha financial marke and rade openness has inverse effecs on he oupu growh and prices. Resuls of VDC s showed ha effecs of openness were significan and increase in openness reduced ariffs and hence lower impor prices. Bowdler (2003) used cross secional daa of 20 counries o es he shor erm inclinaion of Phillips curve relaes posiively wih rade openness. He concluded ha if cambial regime aken ino consideraion hen degree of rade openness in a counry exered posiive effec on inclinaion of Phillips curve. Sachsida, Carneiro, and Loureiro (2003), used fixed and random effecs model in order o verify he Romer s findings by using he daa of 152 counries for he period of 1950 o They concluded ha negaive relaion among openness and inflaion was neiher specific o counries nor o cerain ime period. Agarwal and Narayanan (2003) used he daase of 53 developing counries locaed a five differen regions for he period of 1975 o GMM Findings showed ha openness had significan negaive effec on inflaion afer The analysis of pre 1989 daa showed ha only fixed exchange rae regime had significan negaive effec. Gruben and McLeod (2004) used panel regression for conrolling counry specific effecs and confirmed abou negaive relaion among inflaion and rade openness. The ime varying coefficiens suggesed ha counries wih more openness o rade enjoyed greaes deducion in heir inflaion during he 1990s. Empirical specificaion also provided coefficien of variaion for inflaion, ha afer 1985 he more open economies have less volaile inflaion. Kim and Beladi (2005) examined he relaion among inflaion and rade openness for 62 economies which consiss of 28 OECD and 34 developing economies and seleced on he basis of cenral bank dependency index form 1947 o Panel analysis indicaed posiive relaion among prices and openness for advanced economies such as U.S., Belgium, and Ireland and inverse relaion for developing counries as in line wih Romer s (1993). Nunziaa and Bowdler (2006) hypohesised negaive relaion among openness and probabiliy of huge increase in prices using daa from 19 OECD economies from A range of probi regressions shown empirical suppor for greaer openness reduces he probabiliy of an inflaion sar even afer conrolling variables. The openness impac on lagged GDP growh and inflaion in U.S. were posiive bu saisically insignifican. Bowdler and Malik (2006) suggesed ha openness may change srucure of consumpion and producion of goods whose prices were more sable inernaionally by using panel daa of 96 counries from Resuls of ordinary leas squares suggesed ha opening of economy more sharply han he average has experienced huge deducions in inflaion. Sachsida (2006) esimaed relaion among inflaion and rade openness o verify Romer hypohesis for 152 counries wih division in 7 differen groups from Fixed and random effec resuls given suppor o Romer s ha inverse relaion among inflaion and openness were resricing neiher o subse of economies nor o ime period. Chung-Shu Wu and Jin-Lung Lin (2006) invesigaed openness-inflaion relaionship using panel daa of 13 counries ha included Asian 4 Newly Indusrialised

5 Relaionship beween Trade Openness and Inflaion 857 Economies (NIE s) 11 and he G7 12 from 1973 o Panel regression resuls clear ha models wih or wihou consan consrain give differen relaionships beween openness and inflaion. Wih resriced consan erms, he resuls were similar o Romer s (1993) however, if relax ha resricions, empirical resuls does no show a cerain relaionship. They concluded ha openness has significan negaive relaionship wih inflaion for NIEs, bu has mixed resuls for G7. Aisen and Veiga (2006) analysed panel daa of more han 100 counries from 1975 o 1999 and found ha less economic openness along wih higher degrees of poliical insabiliy generaed more volaile inflaion raes. Resuls indicaed ha higher openness was relaed o lower inflaion bu his canno be found in all counries a all imes and hey also suppored he exisence of impor price effec. Hanif and Baool (2006) esed Romer s hypohesis for Pakisan using ime series daa from 1973 o They found ha real gross domesic produc, moneary growh, ineres rae, whea suppor price and openness (he raio of growh in rade o GDP) has inverse effec on inflaion in Pakisan. Resuls from Regression Analysis clear ha supply facors were imporan han moneary facors in he process of inflaion. Gopal (2007) discussed he effec of openness on ariff srucure, expor compeiiveness, prices and economic growh for 11 counries of Lain American region 13 during Ordinary leas square resuls indicaed he exisence of significan posiive relaion and higher openness beween Lain American counries would enhance o upgrade insiuions. The opening up of markes could play vial role in decreasing economic rens aeched wih economic and insiuional arrangemens. Evans (2007) focused on level of imperfec compeiion ha affecs he relaion among openness and inflaion boh wihin a counry and beween counries by using 2 counry overlapping generaions (OLG) 14 model from Resuls indicaed ha level of imperfec compeiion among he producers plays a subsiue for marke power enjoyed by counry s moneary auhoriy in obaining monopoly rens available in inernaional srucure. 15 Badinger (2007) assessed he relaion among inflaion and openness measured in erms of financial openness using cross-secional daa of 91 counries from SLS resuls indicaed ha larger rade and financial openness reduced cenral bank s independency which yield o less inflaion ha is aached wih larger oupu-inflaion radeoff. Daniels and Vanhoose (2007) considered open economy wih degree of incomeax progressiviy influenced on he ineracion beween openness, cenral bank independence and prices by using daa of 17 counries from 1979 o Regression analysis of cross-counry inflaion provided favor inverse relaionship beween inflaion 11 Hong Kong, Korea, Mexico, Philippines, Singapore, and Taiwan. 12 Canada, France, Germany, Ialy, Japan, U.K. and he Unied Saes. 13 Consis of various sub regional groups: Mexico, Cenral America (Cosa Rica, El Salvador, Guaemala, Honduras, Nicaragua and Panama), and he Caribbean 13 counries; Souh America conains he Andean Communiy (Colombia, Ecuador, Bolivia, Peru) and Mercosur (Argenina, Brazil, Uruguay, Paraguay and Venezuela) and Chile. 14 In which agens live counable ime span long enough o live one period a leas wih he nex generaions of agens. 15 Tha is, greaer level of imperfec compeiion among producers decreases he benefis from inflaion generaed by counry s moneary auhoriy.

6 858 Munir and Kiani and income ax progressive sysem. OLS Resuls indicaed ha higher openness and cenral bank independency reduced he income-ax progressiviy effecs on price levels. Berumen, Dogan, and Tansel (2008) assessed he role of openness on inflaion for 4 MENA counries 16 hrough EGARCH model 17 from 1952 o 2006 by using expor and impor openness separaely. Resuls suggesed ha increase in expor openness 18 reduces inflaion volailiy for all MENA counries. However, incremen in impor openness 19 reduces price level for Jordan and Morocco bu increases for Algeria and Turkey. The effec of inflaion on openness was posiive for Jordan, Morocco and Turkey and saisically significan jus for Morocco. Menghan (2008) esimaed shor and long run effec of openness on inflaion hrough changes in produciviy and ineres rae by using indusrial panel daa of 20 indusries in each of 6 OECD counries 20 from 1980 o Resuls indicaed ha openness reduced inflaion rae, produciviy and mark up in shor run while; long run resuls were ambiguous. Furuoka and Mun Ho (2009) examined relaion beween openness, unemploymen and inflaion by choosing 3 Asian economies 21 wih differen degrees of openness from 1980 o OLS resuls indicaed ha as counry opened up o world by rising he quaniy of impors hen coefficien of Phillips curve slope become smaller. They concluded ha more open counries end o have flaer Phillips curve wih higher sacrifice rae. Lin (2010) invesigaed relaion among rade openness and inflaion of 106 counries using quanile regression from Resuls refleced inverse impac of openness on inflaion when price level was larger bu no effec when i was less. He concluded ha relaion among openness and inflaion appeared o be srenghening in larger prices period and was exremely robus o consider 1980s deb crisis and conrol he exchange-rae regime. Mukhar (2010) applied mulivariae coinegraion approach and vecor error correcion model o examine he Romer s hypohesis for Pakisan. He esimaed ime series daa from 1960 o 2007 on budge defici (BD), GDP, rade openness (TO), exchange rae (ER) and inflaion (CPI). The empirical findings show ha here was significan inverse long run relaion among prices and openness which confirmed he exisence of Romer s hypohesis in Pakisan. Zakaria (2010) empirically examined relaion among rade openness and prices in Pakisan using annual ime series daa from 1947 o Generalised Mehod of Momens (GMM) resuls shown ha posiive relaion holds among openness and inflaion in Pakisan and he conrol variables i.e. money supply, fiscal defici, exchange rae depreciaions, foreign inflaion, erms of rade, foreign deb and democracy significanly affec inflaion. 16 Middle Eas and Norh African (Algeria, Jordan, Morocco and Turkey). 17 GARCH models assumed ha posiive and inverse error erms effec on volailiy. From empirical poin exponenial GARCH (EGARCH) volailiy performs asymmerically o he sign of shocks. 18 Expor-GDP raio. 19 Impor-GDP raio. 20 USA, Japan, Canada, Porugal, Finland and Ausralia. 21 Japan (9.8 percen), Souh Korea (32.9 percen) and Malaysia (77.2 percen).

7 Relaionship beween Trade Openness and Inflaion 859 Evans (2011) proposed ha rade openness enhanced counry s incenive o creae inflaion by esimaing daa hrough regression from 1973 o 1987 and 1988 o He concluded ha openness was inflaionary beween developed counries in which moneary policy can roughly approximaed by conrolling for imperfec compeiion and inelasiciy of labor supply wihin counry. 3. THEORETICAL FRAMEWORK AND METHODOLOGY Now, we designed he suiable model and explain how he variables are consruced and described he sources from where he daa has been aken. Afer ha explain he economeric mehodology for esimaion and inerpreaion of resuls Mehodology Inflaion is a complex phenomena and i is no easy o esablish an empirical model for a counry. However, i is possible o find he key variables effecing he inflaion in Pakisan. The mos common empirical mehodology for examining he rade openness and inflaion relaion had been o apply single equaion model for inflaion, reaing rade openness as an independen variable wih ohers. Solomon and dewe (2004) use four variable single equaion model where budge defici (BD), gross domesic produc (GDP) and exchange rae (ER) were reaed as independen variables and inflaion (CPI) as an dependen variable. Solomon and de We (2004) model is also used by Mukhar (2010) in his sudy. To his, we add real agriculure value added (Agr), financial marke openness (FMO), money and quasi money (M2), rade openness (TO) impor openness (IO) and expor openness (EO) as an independen variable wih Gross Domesic produc (GDP) and Exchange Rae (ER) are used in Real Terms. We also include Two Dummy Variables of 1982 and 1990 in Solomon and de We (2004) model for changes in Exchange Rae Regimes and Financial and Srucural Reforms respecively. In order o obain he objecives of a sudy, model is expressed as follows; Where, CPI 0 1RealAgr 2RealER 3LnRealGDP 4TO 5 LnM2 FMO 6 (a) CPI shows Inflaion rae Real Agr shows Real Agriculure Value added Real ER shows Real Exchange Rae Ln RealGDP shows Naural logarihm of Real Gross Domesic Produc TO shows Trade Openness FMO shows Financial Marke Openness LnM2 shows Money and Quasi money TO shows Trade Openness CPI 0 1RealAgr 2RealER 3LnRealGDP 4IO 5 LnM2 FMO 6 (b)

8 860 Munir and Kiani Where, IO shows Impor Openness. CPI 0 1RealAgr 2RealER 3LnRealGDP 4EO 5FMO 6 LnM2 (c) Where, EO shows Expor Openness Daa Sources In his sudy we have aken annual ime series daa ha covers he period of 1976 o 2010 from various sources including Inernaional Financial Saisics of Inernaional Moneary Fund (IMF s). World Developmen Indicaors (WDI). Saisical Appendix 2010 of Sae Bank of Pakisan (SBP). In independen variables, naural logarihms of real GDP and Money and Quasi Money are aken because he daa is in Rs millions while, all ohers variables are aken as % of GDP excep exchange rae and inflaion rae which are index numbers wih base year Selecion and Consrucion of Variables Following are he variables used in his sudy Table of Variables Descripions Code Variables Definiions Formula Unis Agr Real Agriculure Value added ER FMO Real Exchange Rae Financial Marke Openness Includes foresry, huning, fishing, culivaion of crops and livesock producion. Value added is whole secor oupu afer adding all oupus and subracing inpus. I is esimaed wihou making reducions for depreciaion or depleion of fabricaed asses and degradaion of naural resources. The rae a which one currency will be exchanged for anoher. I is also considered as he value of one counry s currency in erms of anoher currency Scenario where exising adminisraive and marke resricions on capial movemen across borders have been vanished. When capial accoun liberalizaion implemens, i should creae Openness, hen financial inegraion will gradually be obained.[rober sehrer] All oupus -Inermediae inpus (No deducing depreciaion of fabricaed asses and degradaion of naural resources) (Marke rae)*foreign Inflaion Domesic Inflaion Source of Daa and Definiions % of GDP WDI, World Bank naional accouns daa, and OECD Naional Accouns daa files. Index Number wih base Year 2005 Inernaional Moneary Fund, Inernaional Financial Saisics. FDI (Ne Inflows) % of GDP Saisics & DWH Deparmen, SBP. Coninued

9 Table of Variable (Coninued) GDP Real Gross Domesic Produc Relaionship beween Trade Openness and Inflaion 861 The marke amoun of goods and services produced by a counry in a given year. TO Trade Openness Value o which counries allow rade wih oher counries. Broad economies generally have higher opporuniies, a he same ime hey also face compeiion from ohers economies Trade Openness is he sum of expors and impors of goods and services measured as a share of gross domesic produc. CPI M2 IO EO Consumer Price Index Money and Quasi Money The annual percenage change in he value of fixed baske of goods and services ha may be fixed or changed afer specified periods. Includes currency ouside banks, demand deposis oher han hose of cenral governmen, he ime, savings, and foreign currency deposis of residen secors oher han cenral governmen. Impors Openness The value of all goods and services received from he res of he world. Expors Openness The value of all goods and services provided o he res of he world. Nominal GDP Domesic Inflaion (Expors + Impors) GDP LnCPI LnCPI 1 (Impors of goods & services GDP) *100 (Expors of goods & services GDP) *100 Rs Million Inernaional Moneary Fund, Inernaional Financial Saisics and daa files. % of GDP WDI, World Bank naional accouns daa, and OECD Naional Accouns daa files. Index Number wih base Year Rs Million Inernaional Moneary Fund, Inernaional Financial Saisics Inernaional Moneary Fund, Inernaional Financial Saisics. % of GDP WDI, World Bank naional accouns daa, and OECD Naional Accouns daa files. % of GDP WDI, World Bank naional accouns daa, and OECD Naional Accouns daa files Esimaion Techniques Usually many macroeconomic variables are non-saionary for his purpose we can apply uni roo esing echnique in order o see ha wheher he variables are saionary or no. Then, he variables which are saionary a I (1) we have used Johansen (1998) and Johansen and Juselius (1990) Maximum Likelihood Coinegraion Technique and Vecor Error Correcion Model in our sudy o check he long run relaionships in beween hem Univariae Analysis (a) Uni Roo Tes Many variables are non saionary for his we can use Uni Roo Tes in order o verify is order of inegraion. Then, only hose variables are incorporaed in he sudy which is saionary a 1 s difference I (1).

10 862 Munir and Kiani (b) Augmened Dickey- Fuller Tes (ADF) The Augmened version of Dickey Fuller Tes is used for larger and complicaed models which adjus he DF es from serial correlaion in he error erm μ by puing lagged values of dependen variable Y Mulivariae Analysis In order o find he exisence and number of long-run relaionship(s) he economeric framework we used in he sudy for analysis is he Johansen (1998) and Johansen and Juselius (1990) Maximum Likelihood Coinegraion Approach. Two or more series are coinegraed if hey observe same kind of sochasic behavior. I is saisical propery of ime series variables and uses when all he variables are saionary a I (1). The coinegraion approach in a mulivariae sysem is similar o he ADF es, bu requires he use of vecor auoregressive (VAR). A vecor auoregressive (VAR) model wih a lag lengh of 1 was used o es for he number of coinegraing relaionships beween he variables. When wo series are coinegraed i suggess ha even boh processes are non saionary, here is some long run relaionship linking boh series so ha i is saionary. The AIC or SBC is used o deermine he number of lags in he coinegraion es (order of VAR). 22 There are wo likelihood raio es saisics in he Johansen (1998) and Johansen and Juselius (1990) Maximum likelihood Coinegraion Approach; he race and he Maximum Eigenvalue boh can be used o deermine he exisence of number of coinegraing vecors and hey don always indicaed he same number of coinegraing vecors. The disribuion of boh es saisics is non-sandard. The Trace es is a join es wih null hypohesis of number of coinegraing vecors is less han or equal o r, agains alernaive hypohesis ha here are more hen r coinegraing vecors. The Maximum Eigenvalue es conduced separae ess on each eigenvalue wih null hypohesis ha here are r coinegraing vecors exis agains he alernaive hypohesis ha here exiss (r + 1). The Johansen s maximum eigenvalue and race ess indicae he coinegraing vecor (eq s) in model and rejec he null hypohesis of no coinegraion a 5 percen significance level. Then consider he 1s coinegraing equaion having normalised coefficiens of all variables wih sandard error (S.E) in parenheses and calculae T value by dividing coefficien wih S.E. T value greaer hen 2 indicae he significance of hose variables a 5 percen confidence level Vecor Error Correcion Model A main qualiy of coinegraed variables is ha heir ime pahs are effeced by he exen of any deviaion from he long-run equilibrium [Anders (2004)]. The error correcion mechanism (ECM) erm presens he percenage of correcion o any deviaion in he long-run equilibrium price in a single period and also represens how fas he deviaions in he long-run equilibrium are correced. Depending on he presence of how many coinegraing vecors, we can hen es for he shor run dynamics using a vecor 22 Gujarai, N. Damodar, Basic Economerics (Fourh Ediion).

11 Relaionship beween Trade Openness and Inflaion 863 error correcion model. A vecor error correcion model (VECM) is a process wih he qualiy of deviaion from presen sae means is long-run link will pu ino is shor-run dynamics i.e., how changes in rade openness in shor run conribued o is long run relaion wih inflaion. 4. ESTIMATION RESULTS The firs sep in coinegraion analysis is o es he saionariy of variables. Table 2 in Appendix presens he Resuls of Augmened Dickey Fuller Tes. I shows ha all he variables incorporaed in his sudy are found o be saionary a firs difference I(1). To obain opimal lag lengh for coinegraion analysis, basically wo crieria are used namely he AIC and he SBC. The SBC has suggesed lag lengh of 1 as opimal, while he AIC indicaes 3 as an opimal lag lengh. However, we have seleced opimal lag lengh 1 as suggesed by he SBC because when we use he lag lengh 3 for coinegraion analysis we find no coinegraing vecors under boh Trace and Max-Eigen saisics. While wih lag lengh 1, we may obain same and differen numbers of coinegraing vecors under boh hese saisics. Firs, we explain he resuls of inflaion rae wih openness by using he proxy of Trade raio (Expors + Impors) from equaion (a). The coinegraion relaionships beween inflaion rae, Real Agr, Real ER, Real GDP, FMO, M2 and TO has been invesigaed assuming linear rend in daa wih an inercep in coinegraing equaion using he esimaion echnique. Table 3 in Appendix repors Johansen (1998) and Johansen and Juselius (1990) Maximum Likelihood Coinegraion Resuls. The Trace saisics (λ race) and Maximum-Eigenvalue (λ max) saisics indicae ha here is Four coinegraing vecors in seven ime series under boh saisics. We can rejec he null hypohesis of no coinegraing vecor in favour of four coinegraing vecors under Trace and Maximum-Eigenvalue saisics a 5 percen level of significance. Under he assumpion of no deerminisic rend in daa and inercep and no rend in coinegraion equaion, we can obain he equaion which is normalised for inflaion o obain meanings from he coefficiens are given below; CPI RealAgr RealER LnRealGDP T Val (0.2212) ( ) ( ) ( ) TO FMO LnM 2 ( ) ( ) ( ) Normalised coefficiens wih T value shows ha excep wo variables all he independen variables reflec significan and sandarised relaionships a 5 percen level of significance. The coefficien of Trade Openness carries a posiive sign and saisically significan a 5 percen level of significance, which shows ha a 1 percen increase in rade openness brings abou percen increase in inflaion rae. This finding is suppored by he empirical resuls of Kim and Beladi (2005), Pehnel (2007), Gopal (2007), Evans (2007), Razin and Loungani (2007), Berumen, Dogan, and Tansel (2008) and Zakaria (2010). There is significan posiive long run relaionship among inflaion and rade openness in Pakisan and coefficien cleared ha 1 percen incremen in rade openness increases he inflaion by percen. Which confirms he rejecion of our null hypohesis.

12 864 Munir and Kiani The coefficien of real GDP carries a negaive sign bu saisically insignifican a 5 percen level of significance and shows ha a 1 percen increase in real GDP brings abou percen decrease in inflaion rae. This finding is in line wih Agarwal and Narayanan (2003) which shows ha GDP has a significan negaive effec wihou dummies for counry, ime and exchange rae regimes. Mukhar (2010) also suppored he significan negaive relaionship beween inflaion rae and GDP such ha a 0.42 percen decrease in he inflaion is associaed wih a 1 percen increase in GDP. While, Menghan (2008) found a posiive long run relaionship beween GDP and prices. The coefficien of real ER carries a posiive sign and saisically significan a 5 percen level of significance, which shows ha a 1 percen increase in real ER brings abou percen increase in inflaion rae. This finding is no suppored by he resuls of Agarwal and Narayanan (2003) ha he fixed exchange rae regime has significan negaive effec on inflaion if he daase is analysed in wo differen ime spans indicaing ha i is a shor-run phenomenon. Bu, Mukhar (2010) found a significan posiive relaionship beween inflaion rae and ER such ha a percen increase in he inflaion is associaed wih a 1 percen increase in ER. Rogoff (1985) proposed ha increased inflaion has an exra cos and he opimal rae chosen by moneary auhoriies was lesser as he deerioraing effec on exchange rae increases. The coefficien of real Agr carries a significan posiive sign a 5 percen level of significance and shows ha a 1 percen increase in real Agr brings abou percen increase in inflaion rae. This finding is in line wih Hanif and Baool (2006) ha growh in suppor prices of whea is found o be posiive and significan. And, Ashra (2002) also suppored ha rae of growh of agriculural oupu have saisically significan impac on he local inflaionary process. The coefficien of money and quasi money carries a negaive sign bu saisically insignifican a 5 percen level of significance, which shows ha a 1 percen increase in money and quasi money brings abou percen decrease in inflaion rae. Bu, Agarwal and Narayanan (2003) and Ashra (2002) found a significan posiive robus effec of he money growh on inflaion and suppors he heoreical argumens of he moneariss. Broad moneary policy increases GDP and depreciaes he exchange rae, and he laer adjusmen pus up impor prices and inflaion in proporion o he openness of he economy [Romer (1993)]. The coefficien of FMO carries a significan posiive sign a 5 percen level of significance and shows ha a 1 percen increase in FMO brings abou percen increase in inflaion rae. Our resuls are no suppored by Jin (2002) which shows significan negaive shor-run effecs of financial marke openness on he growh raes of he price level. And, Badinger (2007) also found ha increase in financial openness by one percenage poin leads o a decrease in inflaion by 0.36 percen. Vecor error correcion mechanism (VECM) erm represens he speed of adjusmen back o he long run relaionship among he variables. Table 4 in Appendix presens he resuls of he error correcion model for Pakisan under sudy for Inflaion wih Trade Openness. The esimaed coefficiens show he immediae impac of differen independen variables i.e., (real agriculure value added, real exchange rae, financial marke openness, real GDP, rade openness, money and quasi money) on Inflaion Rae. The ECM erm for Pakisan is which is negaive bu insignifican in he

13 Relaionship beween Trade Openness and Inflaion 865 analysis a 5 percen level of significance and suggess ha inflaion is correced by per annum. In he shor run, i can be observed ha flucuaion exiss in general. While, all adjusmens ake place wih in he same or following ime periods, implying ha he sysem seles down quickly. The coefficien of he ECT of inflaion variable carries he negaive sign and saisically insignifican a 5 percen level wih he speed of convergence o equilibrium of percen. This means ha, whenever here is any disurbance in he sysem in he long run, in every shor-run period, a percen correcion o disequilibrium will ake place. More specifically, ECT coefficien shows ha a deviaion from he long run equilibrium value in one period is correced in he nex period by he size of he coefficien. This indicaes he sabiliy of he model. While, FMO and M2 are saisically insignifican and TO is saisically significan bu hey carry a posiive sign. This means ha, in case of any disurbance, divergence from he equilibrium pah will ake place and he whole sysem canno be brough o equilibrium posiion in each case. Then, we explain he resuls of inflaion rae wih openness by using he proxy of Impor raio from equaion (b). The coinegraion relaionships beween inflaion rae, Real Agr, Real ER, Real GDP, FMO, M2 and IO has been invesigaed assuming linear rend in daa wih an inercep in coinegraing equaion using he esimaion echnique. Table 5 in Appendix repors Johansen (1998) and Johansen and Juselius (1990) Maximum Likelihood Coinegraion Resuls. The Trace saisics (λ race) and Maximum-Eigenvalue (λ max) saisics indicae ha here is Five and Three coinegraing vecors respecively in seven ime series. We can rejec he null hypohesis of no coinegraing vecor in favour of five and hree coinegraing vecors under Trace and Maximum-Eigenvalue saisics a 5 percen level of significance. Under he assumpion of no deerminisic rend in daa and inercep and no rend in coinegraion equaion, we can obain he equaion which is normalised for inflaion o obain meanings from he coefficiens are given below; CPI RealAgr RealER LnRealGDP T Val ( ) ( ) ( ) ( ) IO FMO LnM 2 ( ) ( ) ( ) Normalised coefficiens wih T value shows ha excep M2 all he independen variables reflec significan and sandardised relaionships a 5 percen level of significance. The coefficien of Impor Openness carries a posiive sign and saisically significan a 5 percen level of significance, which shows ha a 1 percen increase in impor openness brings abou percen increase in inflaion rae and confirmed ha if impors share rises in oal rade hen i posiively effec inflaion. These resuls are no in line wih he empirical resuls of Berumen, Dogan, and Tansel (2008) as coefficiens of Impor openness is negaive which suggess ha higher impor openness decreases inflaion volailiy for Jordan and Morocco and his effec is saisically significan jus for Jordan. However, i is posiive for he oher wo counries bu saisically significan jus for Turkey. While, Wu and Lin (2006) suppors posiive relaionships beween impor openness and inflaion wihou consan consrain.

14 866 Munir and Kiani Bu, Agarwal and Narayanan (2003) shows he mixed resuls ha before 1989 only fixed exchange rae regime had significan negaive effec on inflaion and afer 1989 openness had significan negaive effec on inflaion. There is posiive long run relaionship among inflaion and impor openness in Pakisan and coefficien cleared ha 1 percen incremen in impor openness increases he inflaion by percen. Which reflecs ha impored inflaion increases in Pakisan because of increase in demands of impors and confirmed he rejecion of our null hypohesis. The coefficien of real GDP carries a negaive sign and saisically significan a 5 percen level of significance, which shows ha a 1 percen increase in real GDP brings abou percen decrease in inflaion rae. This finding is suppored by Agarwal and Narayanan (2003) which shows ha GDP has a significan negaive effec wihou dummies for counry, ime and exchange rae regimes. And, Mukhar (2010) also suppor a significan negaive relaionship beween inflaion rae and GDP such ha a 0.42 percen decrease in he inflaion is associaed wih a 1 percen increase in GDP. While, Menghan (2008) found posiive long run relaionship beween GDP and prices. The coefficien of real ER carries a posiive sign and saisically significan a 5 percen level of significance, which shows ha a 1 percen increase in real ER brings abou percen increase in inflaion rae. This is no suppored by Agarwal and Narayanan (2003) ha he fixed exchange rae regime has significan negaive effec on inflaion if he daase is analysed in wo differen ime spans indicaing ha i is a shorrun phenomenon. Bu, Mukhar (2010) suppors our resuls ha here exiss significan posiive relaionship beween inflaion rae and ER such ha a percen increase in he inflaion is associaed wih a 1 percen increase in ER. Rogoff (1985) proposed ha increased inflaion has an exra cos and he opimal rae chosen by moneary auhoriies was lesser as he deerioraing effec on exchange rae increases. The coefficien of real Agr carries a significan posiive sign a 5 percen level of significance and shows ha a 1 percen increase in real Agr brings abou percen increase in inflaion rae. This finding is in line wih Hanif and Baool (2006) ha growh in suppor prices of whea is found o be posiive and significan. And, Ashra (2002) also suppored ha rae of growh of agriculural oupu have saisically significan impac on he local inflaionary process. The coefficien of money and quasi money carries a posiive sign bu saisically insignifican a 5 percen level of significance, which shows ha a 1 percen increase in money and quasi money brings abou percen increase in inflaion rae. Our resuls are suppored by Agarwal and Narayanan (2003) and Ashra (2002) ha a significan posiive robus effec of he money growh on inflaion which also suppors he heoreical argumens of he moneariss. Broad moneary policy increases GDP and depreciaes he exchange rae, and he laer adjusmen pus up impor prices and inflaion in proporion o he openness of he economy [Romer (1993)]. This shows ha money remains an imporan facor of he inflaionary process in Pakisan. The coefficien of FMO carries a significan posiive sign a 5 percen level of significance and shows ha a 1 percen increase in FMO brings abou percen increase in inflaion rae. Our resuls are no suppored by Jin (2002) which shows significan negaive shor-run effecs of financial marke openness on he growh raes of he price level. And, Badinger (2007) also found ha increase in financial openness by one percenage poin leads o a decrease in inflaion by 0.36 percen.

15 Relaionship beween Trade Openness and Inflaion 867 Vecor error correcion mechanism (VECM) erm represens he speed of adjusmen back o he long run relaionship among he variables. Table 6 in Appendix presens he resuls of he error correcion model for Pakisan under sudy for Inflaion wih Impor Openness. The esimaed coefficiens show he immediae impac of differen independen variables i.e.; (real agriculure value added, real exchange rae, financial marke openness, real GDP, impor openness, money and quasi money) on Inflaion Rae. The coefficien of he ECT of inflaion variable carries he posiive sign and saisically insignifican a 5 percen level and suggess ha long-run equilibrium condiions of inflaion does no influence he shor-run dynamics in Pakisan wih impor openness which indicaes he insabiliy of he model. While, he coefficiens of he ECTs of impor openness, FMO, Agr and M2 carries a posiive sign bu excep impor openness all ohers are saisically insignifican a 5 percen level of significance. This means ha, in case of any disurbance, divergence from he equilibrium pah will ake place and he whole sysem canno be brough o equilibrium posiion in each case. Lasly, we explain he resuls of inflaion rae wih openness by using he proxy of Expor raio from equaion (c). The coinegraion relaionships beween inflaion rae, Real Agr, Real ER, Real GDP, FMO, M2 and EO has been invesigaed assuming linear rend in daa wih an inercep in coinegraing equaion using he esimaion echnique. Table 7 in Appendix repors Johansen (1998) and Johansen and Juselius (1990) Maximum Likelihood Coinegraion Resuls. The Trace saisics (λ race) and Maximum-Eigenvalue (λ max) saisics indicae ha here is same Five coinegraing vecors in seven ime series. We can rejec he null hypohesis of no coinegraing vecor in favour of five coinegraing vecors under boh Trace and Maximum-Eigenvalue saisics a 5 percen level of significance. Under he assumpion of no deerminisic rend in daa and inercep and no rend in coinegraion equaion, we can obain he equaion which is normalised for inflaion o obain meanings from he coefficiens are given below; CPI RealAgr RealER LnRealGDP T Val ( ) ( ) ( ) ( ) EO FMO LnM 2 ( ) ( ) ( ) Normalised coefficiens wih T value shows ha only Real ER and EO reflecs insignifican relaionships a 5 percen level of significance. The coefficien of Expor Openness carries a posiive sign and saisically significan a 5 percen level of significance, which shows ha a 1 percen increase in expor openness brings abou percen increase in inflaion rae. This finding is no suppored by empirical resuls of Berumen, Dogan, and Tansel (2008) as expor openness reduces inflaion for all Middle Eas and Norh African (MENA) counries. While, Agarwal and Narayanan (2003) shows he mixed resuls ha before 1989 only fixed exchange rae regime had significan negaive effec on inflaion and afer 1989 openness had significan negaive effec on inflaion. Bu, Ashra (2002) shows ha openness has significan posiive effecs on inflaion no maer eiher an economy is experiencing hyper-inflaion or i is large. There is

16 868 Munir and Kiani posiive long run relaionship among inflaion and expor openness in Pakisan and coefficien cleared ha a 1 percen incremen in expor openness increases he inflaion by percen. The coefficien of real GDP carries a posiive sign and saisically insignifican a 5 percen level of significance, which shows ha a 1 percen increase in real GDP brings abou percen increase in inflaion rae. These resuls are suppored by Menghan (2008) which shows posiive long run relaionship beween GDP and prices. While, Agarwal and Narayanan (2003) shows ha GDP has a significan negaive effec wihou dummies for counry, ime and exchange rae regimes. Mukhar (2010) also found a significan negaive relaionship beween inflaion rae and GDP such ha a 0.42 percen decrease in he inflaion is associaed wih a 1 percen increase in GDP. The coefficien of real ER carries a negaive sign and saisically significan a 5 percen level of significance, which shows ha a 1 percen increase in real ER brings abou percen decrease in inflaion rae. This finding is suppored by Agarwal and Narayanan (2003) ha he fixed exchange rae regime has significan negaive effec on inflaion if he daase is analysed in wo differen ime spans indicaing ha i is a shor-run phenomena. Bu, Mukhar (2010) found a significan posiive relaionship beween inflaion rae and ER such ha a percen increase in he inflaion is associaed wih a 1 percen increase in ER. Rogoff (1985) proposed ha increased inflaion has an exra cos and he opimal rae chosen by moneary auhoriies was lesser as he deerioraing effec on exchange rae increases. The coefficien of real Agr carries a posiive sign bu saisically insignifican a 5 percen level of significance and shows ha a 1 percen increase in real Agr brings abou percen increase in inflaion rae. This finding is in line wih Hanif and Baool (2006) ha growh in suppor prices of whea is found o be posiive and significan. And, Ashra (2002) also suppored ha rae of growh of agriculural oupu have saisically significan impac on he local inflaionary process. The coefficien of money and quasi money carries a negaive sign bu saisically insignifican a 5 percen level of significance, which shows ha a 1 percen increase in money and quasi money brings abou percen increase in inflaion rae. While, boh Agarwal and Narayanan (2003) and Ashra (2002) found a significan posiive robus effec of he money growh on inflaion and suppors he heoreical argumens of he moneariss. Broad moneary policy increases GDP and depreciaes he exchange rae, and he laer adjusmen pus up impor prices and inflaion in proporion o he openness of he economy [Romer (1993)]. The coefficien of FMO carries a posiive sign bu saisically insignifican a 5 percen level of significance and shows ha a 1 percen increase in FMO brings abou percen increase in inflaion rae. Our resuls are no suppored by Jin (2002) which shows significan negaive shor-run effecs of financial marke openness on he growh raes of he price level. And, Badinger (2007) also found ha increase in financial openness by one percenage poin leads o a decrease in inflaion by 0.36 percen. Vecor error correcion mechanism (VECM) erm represens he speed of adjusmen back o he long run relaionship among he variables. Table 8 in Appendix presens he resuls of he error correcion model for Pakisan under sudy for Inflaion wih Expor Openness. The esimaed coefficiens show he immediae impac of differen

17 Relaionship beween Trade Openness and Inflaion 869 independen variables i.e., (real agriculure value added, real exchange rae, financial marke openness, real GDP, expor openness, money and quasi money) on Inflaion Rae. The ECM erm for Pakisan is which is negaive and insignifican a 5 percen level of significance in he analysis and suggess ha inflaion is correced by per annum. In he shor run, i can be observed ha flucuaion exiss in general. While, all adjusmens ake place wih in he same or following ime periods, implying ha he sysem seles down quickly. The coefficien of he ECT of inflaion variable carries he negaive sign and saisically insignifican a 5 percen level wih he speed of convergence o equilibrium of percen. This means ha, whenever here is any disurbance in he sysem in he long run, in every shor-run period, a percen correcion o disequilibrium will ake place. More specifically, ECT coefficien shows ha a deviaion from he long run equilibrium value in one period is correced in he nex period by he size of he coefficien. This indicaes he sabiliy of he model. While, he coefficiens of he ECTs of expor openness carries a posiive sign and real ER carries a negaive sign bu hey boh are saisically significan a 5 percen level of significance. While, all ohers variables carries a negaive sign and saisically insignifican. This means ha, in case of any disurbance, divergence from he equilibrium pah will ake place and he whole sysem canno be brough o equilibrium posiion in each case. 5. CONCLUSION The paper empirically explores he relaionship beween rade openness and inflaion in Pakisan using annual ime series daa for he period of 1976 o Since Pakisan s economy has a considerable degree of rade openness, he local price level canno remain immune from abroad shocks. The expeced empirical findings shows ha here is a significan posiive long-run relaionship beween inflaion and rade openness, impor openness and expor openness which rejecs he exisence of Romer s hypohesis in Pakisan. The posiive insignifican effec of money and quasi money on inflaion wih impor openness proxy is somehow follows he moneariss who argue money o be he mos imporan variable influencing he inflaionary process. An increase in he developmen level of he counry and a shif from fixed o flexible exchange rae regime are also found o pu up he counry s inflaion rae. The sudy also shows he significan posiive effec of financial marke openness (FMO) on inflaion wih rade and impor openness proxy as capial accoun liberalisaion implemens which should creae openness, hen financial inegraion will gradually be obained. As, Pakisan has rich agriculure base wih large share of agri-producs in expors and real agriculure value added also shows he significan posiive effec on inflaion wih rade and impor openness proxy. The sudy shows he significan posiive effec of Real ER on inflaion wih rade and impor openness proxy. This implies ha i is no advisable for policymakers o implemen a flexible exchange rae sysem because ha could lead o a major depreciaion ha would creae inflaionary problems. The challenges for he fuure is o find ways of combine flexible exchange rae wih low inflaion in Pakisan.