ANALYZING THE SUSTAINABILITY OF INDIA S CURRENT ACCOUNT POSITION FOLLOWING THE REFORMS OF THE EARLY 1990S

Transcription

1 ANALYZING THE SUSTAINABILITY OF INDIA S CURRENT ACCOUNT POSITION FOLLOWING THE REFORMS OF THE EARLY 1990S Niloufer Sohrabji * Absrac Following he 1991 crisis, India underook reforms ha liberalized rade and invesmen. India faced curren accoun deficis for mos of he period following hese reforms. This paper analyzes susainabiliy of India s curren accoun posiion over he las decade using he ineremporal solvency model of Hakkio and Rush (1991) and Hused (1992). In his heoreical framework, he ineremporal solvency consrain is saisfied if here is coinegraion beween inflows and ouflows of he curren accoun. This paper finds coinegraion beween he series when allowing for a srucural break using he Gregory and Hansen (1996) procedure. Dynamic GLS esimaion shows a srong relaion beween India s curren accoun inflows and ouflows. Based on he empirical resuls, his paper concludes ha here has been an improvemen in rade paerns and despie experiencing deficis, India s curren accoun posiion is susainable. Keywords: Coinegraion, curren accoun susainabiliy, dynamic GLS, India, ineremporal solvency JEL Classificaion: C32, F32 and F41 * Correspondence o: Simmons College, Fenway, Boson 02115, sohrabji@simmons.edu, phone: , fax: I am indebed o wo anonymous referees for heir insighful feedback. All errors remain mine. 1

2 1. INTRODUCTION This paper examines he solvency of India s curren accoun posiion over he las decade. Large and deerioraing curren accoun deficis have played a role in he Mexican peso crisis in 1994, he Eas Asian crisis in and Turkey s crisis in India experienced large curren accoun deficis in he lae 1980s and suffered a balance of paymens crisis in Following he crisis, India underook reforms ha liberalized rade and invesmen. For mos of he period since hese reforms, India has experienced rade and curren accoun deficis. This raises concerns abou he liberalizaion reforms, including he possibiliy of anoher crisis, and makes i imporan o sudy India s curren accoun posiion in his period. Trade and curren accoun deficis need no be problemaic. If rade deficis are linked o poor erms of rade or because of a weak producion base, hen high impor bills are a drain on he economy. On he oher hand if a counry is increasing impors of capial goods i indicaes improvemen in producive capaciy which helps fuure rade balances. This is he basis of he ineremporal solvency model proposed by Hakkio and Rush (1991) and Hused (1992) which analyzes coinegraion beween expors and impors and oher ouflows of he curren accoun. Using he above heoreical framework his paper analyzes susainabiliy of he Indian curren accoun since he reforms of he early 1990s. While coinegraion beween inflows and ouflows is rejeced for he period as a whole, here is evidence of coinegraion when allowing for a srucural break. Taken ogeher hese resuls indicae ha here is an improvemen in he curren accoun posiion in his period. Based on coinegraion resuls, he long run relaionship beween expors and impors and oher debis is esimaed. This paper finds evidence of a srong relaion beween curren accoun inflows and ouflows and concludes ha despie he deficis, India s curren accoun posiion is susainable. 2

3 The paper is organized as follows: he nex secion provides background on India s curren accoun posiion which is followed by a review of he lieraure. Secion 4 presens he heoreical framework and empirical mehodology for examining curren accoun susainabiliy. Secion 5 discusses he empirical resuls and he las secion concludes. 2. BACKGROUND To deermine he curren sae of India s exernal posiion i is useful o analyze he major changes in he curren accoun over he las few decades. Figure 1 plos he rend in he overall curren accoun posiion for India from 1980 o 2006 using wo measures, levels of curren accoun balances and he raio of curren accoun balances o GDP. Figure 2 maps ou he differen componens of he curren accoun including merchandise rade balance, services rade balance, ne unilaeral ransfer accoun and ne invesmen accoun. Togeher hese graphs provide an ineresing porrai of he Indian curren accoun posiion over he las hree decades. India had a deerioraing curren accoun posiion in he lae 1980s (figure 1). The curren accoun defici was approximaely $ 2 billion in he early 1980s ( ) which jumped o over $ 4 billion by This rend coninued for he remainder of he decade wih he defici exceeding $ 7 billion by This same rend is observed in he curren accoun defici o GDP raios. This raio was approximaely 1% in he early 1980s which jumped o 2% by he mid- 1980s and ending up a 2.24% in Three of he four componens of he curren accoun recorded deficis in his period (figure 2). These include merchandise rade balance, services rade balance and ne invesmen income. While he ne unilaeral ransfer accoun was in surplus for he enire decade, i was no enough o reverse he deficis in he oher hree accouns. Expecedly, merchandise rade accoun had he larges defici and was he main driver of he curren accoun defici. 3

4 Figure 1: India s curren accoun posiion [ ] Billions of U.S. dollars Percenage Curren accoun balances Curren accoun o GDP raio Noes: The GDP series was expressed in Indian rupees and curren accoun series in U.S. dollars. GDP was convered o dollars using he marke exchange rae. Source: Inernaional Financial Saisics daabase. Figure 2: Componens of India s curren accoun balance [ ] Billions of U.S. dollars Merchandise rade balance Services rade balance Ne unilaeral ransfers Ne invesmen income Noes: The GDP series was expressed in Indian rupees and he componens of he curren accoun series in U.S. dollars. GDP was convered o dollars using he marke exchange rae. Source: Inernaional Financial Saisics daabase. Given he deerioraing rend in he curren accoun and is componens during he 1980s i is no surprising ha India experienced a balance of paymens crisis in Following he crisis, India embarked on a liberalizaion program which resuled in increased foreign rade and 4

5 invesmen. While boh expors and impors increased, he laer oupaced he former and hus India coninued o experience rade deficis. From 1995 onwards here was once again a much worsened curren accoun posiion (figure 1). By 1998, he curren accoun defici had reached he levels observed prior o he 1991 crisis (approximaely $ 7 billion). A 1.67% however, he curren accoun defici o GDP raio in 1998 was lower han hose observed in he lae 1980s (figure 1). The improved curren accoun defici o GDP raios despie worsened curren accoun deficis were due o he high growh levels observed in India in he 1990s. The rend in he componens of he curren accoun observed in he previous decade acceleraed in he 1990s. Ne invesmen income which was declining in he 1980s worsened furher in he following decade. This showed ha in he 1990s, foreign invesmen in India yielded higher reurns han Indian invesmen abroad. This is no surprising given he high levels of growh in India during his period and is rise as a significan emerging economy. Increased impors due o dismanling of rade barriers led o higher merchandise and services rade deficis in he 1990s compared o he 1980s. As in he 1980s, ne unilaeral ransfer surpluses grew in he following decade. However, hese surpluses were once again insufficien o overurn he deficis in he oher hree accouns. Thus, India experienced high curren accoun deficis. The 2000s sared off wih a significanly improved curren accoun posiion for India. This was parly relaed o he dramaic increase in ne unilaeral ransfer surpluses which reached $ 22 billion by 2003 (figure 2). Also, he growh of expors in boh goods and services was significan leading o a decline in merchandise and services rade deficis. In fac, by 2003 rade of services recorded a surplus. Improvemen in hese hree accouns was more significan han he deerioraion of he ne invesmen income accoun and hus India experienced curren 5

6 accoun surpluses in he early 2000s. By 2003, curren accoun surplus peaked a $ 8.77 billion and he curren accoun surplus o GDP raio reached 1.45% (figure 1). The posiive curren accoun posiion urned negaive in he mid-2000s wih curren accoun defici o GDP raios of approximaely 1% in boh 2005 and 2006 (figure 1). The levels of curren accoun balances also saw a sharp reversal. From a surplus of $ 8.77 billion in 2003 he curren accoun balance changed o a defici of $ 7.84 billion in 2005 and worsened o $ 9.42 billion in 2006 (figure 1). In his period, ne expor of services and ne unilaeral ransfers improved, reaching surpluses of $ 11 billion and $ 26 billion in 2006 respecively (figure 2). However, merchandise defici grew more significanly and overurned he posiive impac of hese improvemens. While expor of goods was increasing rapidly, impors far oupaced expors and exacerbaed he merchandise rade defici which exceeded $ 43 billion in 2006 (figure 2). This deerioraing rend has coninued wih merchandise rade defici reaching $ 59.4 billion in 2007 (Indian Minisry of Finance, Economic Survey ). This in urn led o an increased curren accoun defici which reached $ 17 billion in 2007 (Reserve Bank of India, Handbook of Saisics on Indian Economy Ocober 2008). In recen years, India is one of he few major Asian economies suffering from a curren accoun defici. Undersanding he cause of hese deficis is criical in analyzing he susainabiliy of India s exernal posiion. One major reason for he increase in merchandise rade defici was he deerioraion of erms of rade (Indian Minisry of Finance, Economic Survey ). The value of expors could no keep pace wih he value of impors resuling in rade deficis. Moreover, he rise in inernaional crude oil prices which caused he value of peroleum, oil and lubricans impors o rise by 55% have conribued o his defici (Reserve Bank of India, Monhly Bullein

7 2009). While prices have declined, his remains a vulnerable componen of India s curren accoun posiion. On he posiive side, i is imporan o noe ha in 2004, here was an increase in he expor of high valued engineering goods (Indian Minisry of Finance, Economic Survey ). Furher increases in hese, and oher high priced expors, could lead o improvemens in India s rade posiion. The same repor noes ha he highes growh in impored goods was observed in capial goods in he early 2000s. This is imporan because capial goods add o producive capaciy. Thus, increases in impors were no financing consumpion, bu raher should be viewed as invesmen. In urn, his invesmen helps growh in expors and GDP which can conribue o fuure curren accoun susainabiliy. Over he las decade he curren accoun has worsened (lae 1990s), improved (early 2000s) and is once again deerioraing (mid-2000s). The recenly rising curren accoun deficis may sugges ha he curren accoun posiion is once again becoming weaker. However, an examinaion of he differen componens of he curren accoun reveals a more complicaed picure. While larger merchandise impor growh has conribued o deerioraing rade and curren accoun deficis, he ypes of impors deermine if his implies an unsusainable curren accoun posiion. To shed ligh on his issue, his paper employs an ineremporal solvency approach using quarerly daa from The nex wo secions discuss he lieraure and framework for sudying susainabiliy of he curren accoun posiion. 3. LITERATURE The ineremporal approach for exploring curren accoun susainabiliy is based on he heoreical model of Hakkio and Rush (1991) and Hused (1992). Hakkio and Rush (1991) discuss susainabiliy in he conex of budge deficis which has been used by Marin (2000) and 7

8 Baharamshah and Lau (2007) o examine fiscal susainabiliy for he U.S. and Eas Asian counries respecively. In he conex of curren accoun susainabiliy, he framework examines he long-run relaion beween inflows and ouflows in he curren accoun. This requires ha expors of goods and services (inflows) be coinegraed wih impors of goods and services, ne unilaeral ransfer paymens and ne income paymens (ouflows). The hree componens ha make up he ouflows of he curren accoun are referred o as impors+ for he remainder of he paper. Coinegraion beween hese wo series implies ineremporal solvency of he curren accoun. This model has been used by Hused (1992) and Founas and Wu (1999) for he U.S., Leachman and Thorpe (1998) for Ausralia, Apergis, Karakilidis and Tabakis (2000) for Greece, Arize (2002) for 50 developed and developing counries, Baharumshah, Lau and Founas (2003) for Indonesia, Malaysia, Philippines and Thailand, Irandous and Ericsson (2004) for 6 developed counries, Narayan and Narayan (2005) for 22 leas developed counries, Kalyoncu (2005) and Oğuş Binalı and Sohrabji (2009) for Turkey, Kim, Min, Hwang, and McDonald (2009) for Asian counries, Kónya (2009) for Czech Republic, Hungary and Slovenia and Lau, Oh and Wong (2009) for bilaeral rade susainabiliy beween Malaysia and wo of is rading parners, China and India. Curren accoun susainabiliy for India has been sudied by a few auhors including Upender (2007), Kónya and Singh (2008) and Holmes, Panagioidis and Sharma (2008). All papers es for coinegraion beween expors and impors+ and focus on he period beween 1950 and 2005 approximaely. Upender (2007) finds ha expors and impors+ are coinegraed. Conradicory o his resul, Kónya and Singh (2008) find no coinegraion beween expors and impors+ even afer allowing for a break period in due o a shif in he exchange rae regime. 8

9 Holmes, e al. (2008) find wo regimes, one ha lass unil lae 1990s when India s expors and impors were no coinegraed and he oher beginning in he lae 1990s and ending in 2003 where here is coinegraion. This paper adds o hese resuls and he larger lieraure by focusing on India s curren accoun susainabiliy in he 1990s-2000s following he liberalizaion of he economy. As noed earlier, major changes were observed in he rade and curren accoun posiion following he reforms in he early 1990s. Thus, his paper builds and exends he analysis of he second coinegraed period sudied by Holmes e al. (2008) and ries o reconcile he differing resuls by Upender (2007) and Kónya and Singh (2008). There are also some differences in he economeric mehodology. Wih he sandard Johansen es, his paper also employs he Gregory and Hansen (1996) procedure. The laer allows for he possibiliy of a break beween expors and impors+ and endogenously deermines his break. Resuls show ha here is a srucural break in he early 2000s which has changed he relaionship beween inflows and ouflows of he curren accoun in he las few years. Given his resul, he long run equilibrium relaionship beween expors and impors+ is esimaed using dynamic GLS. This paper finds evidence of a susainable curren accoun posiion over he las decade. The model and economeric mehodology for sudying curren accoun susainabiliy is discussed below. 4. INTERTEMPORAL APPROACH FOR CURRENT ACCOUNT SUSTAINABILITY This paper uses he ineremporal solvency heoreical framework of Hakkio and Rush (1991) and Hused (1992) o deermine curren accoun susainabiliy. The model assumes ha he amoun an open economy borrows or lends in he inernaional marke mus equal he presen value of surpluses or deficis (for a more complee discussion of he model see Oğuş Binalı and 9

10 Sohrabji, 2009). A susainable curren accoun posiion indicaes ha presen deficis can be repaid by fuure surpluses. This implies examining he following economeric relaion, X = α + β + ε (1) MM where X equals expors of goods and services and MM equals impors of goods and services, ne ineres paymens and ne unilaeral ransfer paymens (denoed as impors+). Coinegraion beween expors and impors+ sheds ligh on curren accoun susainabiliy. Tess for coinegraion require ha he series be nonsaionary. Several saionariy ess are used including he Augmened Dickey-Fuller (ADF) es, he Phillips Perron (PP) es, he KPSS es 1 and he Zivo-Andrews es 2. If he series are inegraed of order one, coinegraion ess can be conduced. Coinegraion is esed using he Johansen procedure (boh race and eigenvalue saisics). To allow for he possibiliy of a break in he coinegraing relaion his paper follows Baharumshah e al. (2003), Cook (2004) and Oğuş Binalı and Sohrabji (2009) in using he Gregory and Hansen (1996) coinegraion procedure (denoed as he Gregory-Hansen procedure for he res of his paper). This procedure endogenously deermines he break for hree differen models denoed as level shif (C), level shif wih rend (C/T), and regime shif (C/S) and are specified as: Model wih level shif (C) Model wih level shif and rend (C/T) Model wih regime shif (C/S) X X X = μ 1 + μ 2 D + β MM + ε (2) = μ 1 + μ 2 D + γ + β MM + ε (3) = μ 1 + μ 2 D + β1 MM + β 2 MM D + ε (4) where 0 if τ D = and τ is he srucural break poin. 1 if > τ 10

11 The Gregory-Hansen procedure uses a grid search procedure for breakpoins in he 0.15N and 0.85N region (where N is he sample size). For each model he null hypohesis of no srucural change is esed using he ADF saisic. If here is coinegraion wih a srucural break, he long-run relaionship beween expors and impors+ can be esimaed. Following he lieraure, he long-run equilibrium relaionship beween expors and impors+ can be esimaed using dynamic ordinary/generalized leas squares (DOLS/DGLS) proposed by Sock and Wason (1993) 3. Dynamic esimaion includes he srucural break informaion deermined by he Gregory-Hansen procedure as well as lags and leads of he firs difference of he regressors and is given as X ( MM MM ) D + γ d ( L) Δ MM ε = α + β MM + δ + (5) τ where MM τ is impors+ a he srucural break poin, Δ MM is he firs difference of impors+, d ( L) is a polynomial in L which represens lags and leads and all oher variables are as previously defined. The coefficien β capures he relaion beween expors and impors+ and hus sheds ligh on he susainabiliy of he curren accoun posiion. If he relaion beween expors and impors+ is no saisically significan ( β 0) hen he curren accoun posiion is unsusainable since impors+ do no add o expors. A saisically significan and srong relaionship beween expors and impors+ ( β 1) indicaes ha he counry has a susainable curren accoun posiion. In his case, he counry is saisfying he ineremporal budge consrain and can generae sufficien surpluses in he fuure o repay earlier curren accoun deficis. Finally, a saisically significan bu weak relaion beween expors and impors+ ( 0 < β <1) indicaes ha while impors+ conribue o he expor base, he counry s curren accoun posiion remains vulnerable. The relaionship is no srong enough o generae sufficien 11

12 surpluses in he fuure o repay presen deficis. The daa and resuls are repored and discussed in he following secion. 5. EMPIRICAL RESULTS Following he lieraure, expors and impors+ are measured in real erms as a percenage of GDP denoed as RXY and RMMY. This paper focuses on he period afer he 1991 crisis and subsequen reforms of he early 1990s. Lack of complee daa for earlier years in he 1990s and 2007 onwards resriced he saring and ending poin of he daa se from 1996:Q4 o 2006:Q4. The series include expors of goods and services, impors of goods and services, ne ineres paymens, ne unilaeral ransfer paymens and GDP. GDP is expressed in Indian rupees and all oher series are expressed in U.S. dollars. Real GDP is compued using he GDP deflaor (base year = 2000). The real value of expors and impors+ are calculaed by muliplying he series wih he real exchange rae. The laer is compued from he marke exchange rae (Indian rupee o U.S. dollar) and price levels (CPI) for India and for he U.S (base year = 2000). All daa is available from he Inernaional Financial Saisics daabase. The daa is seasonally adjused using he X11 addiive mehod. The series are ploed in figure 3. Figure 3: Real expors and impors+ as percenage of GDP [1996:Q4-2006:Q4] Percenage Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q4 RXY RMMY Source: Inernaional Financial Saisics daabase. 12

13 Figure 3 shows he changing rade paerns beween he 1990s and 2000s discussed earlier. There were deficis during he 1990s which urned ino surpluses a he beginning of he 2000s and revered o deficis in he mid-2000s. More imporanly, he graph also suggess a link beween inflows and ouflows which is esed laer in his secion. Table 1 repors he uni roo ess (assuming a consan only) for he series in levels and firs differences. The es resuls suppor expecaions ha he series are I(1) in levels and I(0) in firs differences. Resuls for ess assuming no deerminisic componen or wih a consan and rend (no repored here) show similar resuls. Given hese conclusions, he series are esed for coinegraion using he Johansen approach and he Gregory-Hansen procedure. Table 1: Uni Roo Tess Variables ADF PP KPSS ZA ZA break RXY [0] Δ RXY * [0] RMMY [0] Δ RMMY * [0] (10) * (4) (2) * (2) * (5) (5) * (5) (2) [0] * [3] [6] * [0] 2004:4 2004:2 2004:3 2003:4 Noes: * indicaes rejecion of he null hypohesis a 5% level of signifcance. The null hypohesis for all ess excep KPSS is ha he series is nonsaionary. All ess are conduced assuming a consan only. Saionariy ess assuming no deerminisic componen or a consan and rend which are no repored here o conserve space show similar resuls. Numbers in brackes for ADF and ZA ess denoe lags. Maximum lags were se a 6 and lag lengh is deermined using AIC. Numbers in brackes for PP and KPSS correspond o lag runcaion parameer. The lag lengh for he VAR used for he Johansen es was deermined o be 2 lags using AIC. Based on he daa he Johansen es was conduced assuming an inercep and rend in he daa and an inercep bu no rend in he coinegraing equaion. Following Baharumshah e al. (2003) his paper uses he Reinsel and Ahn (1988) degree of freedom correcion given as 13

14 ( T p k) T where T is he sample size, p is he number of variables and k is he number of lags. This correcion facor mus be muliplied wih he race and eigenvalue saisics before comparing hem o criical values. The Johansen coinegraion es resuls for no coinegraing equaions as well as a-mos-one coinegraion equaion are repored in able 2. They indicae ha here is no coinegraion beween he wo series. This implies ha he curren accoun is unsusainable for he sample period as a whole. Table 2: Johansen coinegraion es resuls for RXY and RMMY No. of CE(s) Adj. race value 5% c.v. Adj. eigenvalue 5% c.v. None A mos Noe: adjused values imply ha he saisics are muliplied by he small sample correcion facor. Given he changes in he rade and curren accoun balances noed earlier, his paper explores if here is a srucural difference beween he 1990s and 2000s. In erms of he ineremporal solvency model, his implies esing for a break in he coinegraing relaionship. This paper uses he Gregory-Hansen procedure o allow for he possibiliy of a srucural break in he coinegraing relaion. The resuls for he hree models hey discuss are presened in able 3. Table 3: Gregory-Hansen coinegraion es resuls for RXY and RMMY Model 1: C Model 2: C/T Model 3: C/S -saisic * * ** Break period 2001:1 1999:1 2001:1 Lag lengh [0] [3] [0] Noe: * and ** denoes rejecion of he null a 5% and 10% respecively. The Gregory-Hansen procedure shows evidence of coinegraion beween expors and impors+ wih a srucural break for India. This implies ha while expors are no relaed o impors+ for he period as whole, here is a change in he relaionship over he sample period. 14

15 The breaks are he firs quarer of 2001 for he firs and hird model and he firs quarer of 1999 for he second model (able 3). The break poin of he firs quarer of 2001 is seleced for he esimaion of equaion (5) because i coincides wih a few imporan changes in India s rade posiion. The break period marks a urning poin in he rade posiion. I is he firs ime here is a surplus recorded in curren accoun balances (figure 1) alhough hey do no las. This is relaed o high unilaeral ransfers (including remiances) and an improvemen in boh services and merchandise rade (figure 2). Alhough service expors remained less han impors for a few more quarers, he growh in expors of services began exceeding growh in impors of services in he early 2000s. This period also marks a significan growh of rade wih counries in Lain America. While he regional parnership has been imporan since he mid-1990s, expors from India o counries in Lain America jumped from $ 0.9 billion in 2000 o $ 1.5 billion in 2001 (Indian Minisry of Commerce repor, 2005). This was he larges annual jump in he enire period, alhough expors were growing hroughou. Impors also rose in ha period ($ 0.71 billion in 2000 o $ 1 billion in 2001) bu hey were less han he increase in expors. Finally, his period also coincides wih a changing impor posiion. As noed earlier, he highes growh raes in impors in he early 2000s are seen in capial goods which add o he producive capaciy of a counry. Thus, he evidence suggess ha India was no imporing primarily for consumpion bu raher invesing in is fuure. While coinegraion resuls indicae ha here are improvemens in India s rading paerns i does no indicae ha he curren accoun posiion is susainable. These resuls are used o esimae he long-run relaion beween expors and impors+ which can shed ligh on curren accoun susainabiliy. As noed earlier, he break poin from he firs and hird model (firs 15

16 quarer of 2001) is used in esimaing equaion (5). Iniially, dynamic OLS was used in he esimaion. However, since he LM es showed evidence of serial correlaion, equaion (5) was re-esimaed using dynamic GLS. The esimaion was robus for deparures from oher sandard assumpions such as heeroskedasiciy, non-sabiliy and non-normaliy. They are repored in able 4 wih he resuls of dynamic GLS esimaion. Table 4: Dynamic GLS esimaion Diagnosic es resuls Whie es (p-value) RESET es (p-value) Jarque-Bera es (p-value) Adjused Dynamic GLS esimaion resuls 2 R βˆ (SE) H A : β <1 -saisic (0.813) (0.267) (0.493) * (0.327) Noes: The break period from he G-H procedure was used in he esimaion. The lag and lead lengh for differenced RMMY erms were deermined o be one according o AIC and SC. A preliminary es (LM es) showed evidence of serial correlaion, so dynamic GLS was used. The able repors diagnosic ess of he null of homoskedasiciy, sabiliy and normaliy using he Whie es, RESET es and Jarque-Bera es respecively and resuls of he DGLS esimaion. * denoes rejecion of he null a 5% level of significance. The esimaed relaion beween expors and impors+ capured by βˆ is wih a sandard error of This is a high and saisically significan coefficien which suggess a srong relaion beween expors and impors. Moreover, he null of β 1canno be rejeced in favor of he alernaive ha β < 1 and hus he resuls show ha India s curren accoun posiion is srongly susainable. For mos of he sample period, India has recorded curren accoun deficis. These deficis were paricularly high a he beginning of he sample period. The empirical resuls indicae ha hese deficis can be viewed as an invesmen which has resuled in a healhy overall curren accoun posiion. Despie he posiive resuls here could be reasons o be concerned abou India s fuure curren accoun posiion. Firsly, he improvemen in rade and curren accoun deficis (o surpluses) in he early 2000s has been reversed and deficis coninue o grow (beyond he end of 16

17 he sample period). Moreover, while India s rade of services shows healhy surpluses and was helpful in improving rade balances, i remains a small componen of overall rade. The same is rue of he unilaeral ransfer accoun which is posiive and growing bu ulimaely a modes elemen of he overall curren accoun. Merchandise rade which coninues o dominae he overall rade and hus curren accoun posiion remains a vulnerable componen. Merchandise expors have grown bu have no kep pace wih he explosion of impors. This could be furher exacerbaed by he global financial crisis and recession in he U.S. The U.S. coninues o be a significan expor parner for India wih a share of approximaely 15% in he las few years. A decline in demand from he U.S. due o he recession and possible proecionis policy could hur he fuure susainabiliy of India s curren accoun posiion. Coninued curren accoun susainabiliy requires greaer diversificaion of he expor secor. Sofware services and engineering goods have been paricularly successful for expors. Branching ou ino oher dynamic secors such as financial services and pharmaceuicals could reap greaer benefis. In addiion, diversificaion of India s rade parners and a reduced dependence on he U.S. will also help fuure curren accoun susainabiliy. In recen years, here has been an increase in rade wih China, Japan and EU and Lain American counries. Deepening and exending hese rade relaions will reap benefis for India. Also, increasing India s share in Asia as well as Ausralia and New Zealand will be beneficial. 6. CONCLUSION This paper analyzes he susainabiliy of India s curren accoun posiion from using he ineremporal solvency model by Hakkio and Rush (1991) and Hused (1992). This framework examines he relaion beween curren accoun inflows (expors) and ouflows (impors+). Assuming boh are inegraed of order one, coinegraion beween expors and 17

18 impors+ implies ha he ineremporal budge consrain of he curren accoun has been saisfied and he curren accoun is on a susainable pah. The Johansen and Gregory-Hansen procedures are used o es for coinegraion. The Johansen es shows no coinegraion beween expors and impors+ and hus ineremporal solvency is rejeced for he period as a whole. The Gregory-Hansen es shows evidence of coinegraion beween he series wih a srucural break in he firs quarer of This indicaes a changed relaionship beween expors and impors+ in he 2000s. This relaionship beween expors and impors+ assuming he break (from he Gregory-Hansen resuls) is esimaed using DGLS. The resul reveals a srong and saisically significan relaion beween expors and impors+. This indicaes ha India s curren accoun posiion was susainable in he period following he reforms of he early 1990s. These resuls capure imporan aspecs of India s rading posiion. Earlier, he increased impor bill was relaed o rising oil impors. However, since he early 2000s, here has been a significan growh in impors of capial goods. This improved composiion indicaes ha impors are adding o producive capaciy and can hus conribue o increased expors. Moreover, his period marks an improvemen in some of he componens of he curren accoun noably in he rade of services. These improvemens have conribued o India s curren accoun susainabiliy in his period. To mainain fuure curren accoun susainabiliy, India mus diversify is merchandise expor base (for example, owards he pharmaceuical secor) and increase and build on is rade parnerships (for example in Asia and Lain America). NOTES 1. The KPSS (Kwiakowski, Phillips, Schmid and Shin) es is a es for saionariy around a level or a rend. As opposed o all oher ess used, he null hypohesis for KPSS is ha he series is saionary. 18

19 2. The Zivo-Andrews uni roo es allows for srucural breaks. 3. Sock and Wason (1993) noed ha Mone Carlo sudies show ha dynamic OLS/GLS has lowes roo mean square error of all asympoic esimaors. REFERENCES Apergis, Nicholas, Konsaninos P. Karakilidis, and Nicholas, M.Tabakis (2000), Curren Accoun Defici Susainabiliy: The Case of Greece, Applied Economics Leers, 7, Arize, Augusine C. (2002), Impors and Expors in 50 Counries: Tess of Coinegraion and Srucural Breaks, Inernaional Review of Economics and Finance, 11, Baharumshah, Ahmad Zubaidi and Evan Lau (2007) Regime Changes and The Susainabiliy of Fiscal Imbalance In Eas Asian Counries, Economic Modeling, 24, Baharumshah, Ahmad Zubaidi, Evan Lau, and Silianos Founas (2003) On he Susainabiliy of Curren Accoun Deficis: Evidence from Four ASEAN Counries, Journal of Asian Economics, 14, Cook, Seven (2004), Spurious Rejecion by Coinegraion Tess Incorporaing Srucural Change in he Coinegraing Relaionship, Applied Economics Leers, 11, Founas, Silianos and Jyh-Lin Wu (1999), Are he U.S. Curren Accoun Deficis Really Susainable?, Inernaional Economic Journal, 13, Gregory, Allan W. and Bruce E. Hansen (1996) Residual Based Tess for Coinegraion in Models wih Regime Shifs, Journal of Economerics, 7, Hakkio, Craig S. and Mark Rush (1991) Is he Budge Defici oo Large? Economic Inquiry 29, Holmes, Mark J., Theodore Panagioidis and Abhiji Sharma (2008) The Susainabiliy of India s Curren Accoun Universiy of Macedonia, Deparmen of Economics, Discussion Paper Series WP Hused, Seven (1992) The Emerging U.S. Curren Accoun Defici in he 1980s: A Coinegraion Analysis, Review of Economics and Saisics, 74, Irandous, Manuchehr and Johan Ericson (2004) Are Impors and Expors Coinegraed? An Inernaional Comparison, Meroeconomica, 55, Kalyoncu, Huseyin (2005) Susainabiliy of Curren Accoun for Turkey: Ineremporal Solvency Approach, MPRA Paper

20 Kim, Bong-Han, Hong-Ghi Min, Young-Sun Hwang and Judih A. McDonald (2009) Are Asian Counries Curren Accouns Susainable? Deficis, Even when Associaed wih High Invesmen are no Cosless, Journal of Policy Modeling, 31, Kónya, László (2009) The Susainabiliy of he Curren Accoun in he Czech Republic, Hungary and Slovenia, Empirical Economics, 36(2), Kónya, László and Jai Pal Singh (2008) Are Indian Expors and Impors Coinegraed? Applied Economerics and Inernaional Developmen, 8(2), Koureas, Georgios P. and Leonidas P. Zarangas (1998) A Coinegraion Analysis of he Official and Parallel Foreign Exchange Markes for Dollars in Greece, Inernaional Journal of Finance and Economics, 3, Lau, Evan Swee-Ling Oh and Ling-Na Wong (2009) Economeric Analysis on Malaysia- China and Malaysia-India Bilaeral Trade Susainabiliy, Inernaional Journal of Mahemaics and Compuaion, 2, Leachman, Lori and Michael Thorpe (1998) Ineremporal Solvency in he Small Open Economy of Ausralia, The Economic Record, 74(226), Marin, G.M. (2000) US defici susainabiliy: a new approach on muliple endogenous breaks, Journal of Applied Economerics, 15, Narayan, Paresh Kumar and Seema Narayan (2005) Are Expors and Impors Coinegraed? Evidence from 22 Leas Developed Counries, Applied Economics Leers, 12, Oğuş Binalı Ayla and Niloufer Sohrabji (2009) Ineremporal Solvency of Turkey s Curren Accoun, Izmir Universiy of Economics Working Papers Reinsel, Greg C. and Sung K. Ahn (1988) Asympoic Disribuion of he Likelihood Raio Tes for Coinegraion in he Nonsaionary Vecor Auoregressive Model. Working Paper, Universiy of Wisconsin Madison. Sock, James H. and Mark W. Wason (1993) A Simple Esimaor of Coinegraing Vecors in Higher Order Inegraed Sysems, Economerica, 61, Upender, Mulakala (2007) Long Run Equilibrium beween India s Expors and Impors during , Applied Economerics and Inernaional Developmen, 7(1),