Sustainability of Government Debt in Sri Lanka

Transcription

1 Susainabiliy of Governmen Deb in Sri Lanka OG Dayarana-Banda Ph.D, A. A. S. Priyadarshanee 2 ARTICLE INFO Available Online July 24 Key words: Governmen Deb; Deb Raios; Augmened Dicky-Fuller and Phillips-Perron Tess; Chow Tes. ABSTRACT This paperexamines he impac of governmen budge defici on deb susainabiliy in Sri Lanka by using a novel mehodological approach. The sudy used annual ime series daa from 96 o 22 in Sri Lanka for empirical esing. Susainabiliy of governmen debis esed by using face value, marke value and discouned marke value of governmen deb as a proporion of GDP. Discouned marke value of deb o GDP raio was calculaed using weighed average ineres rae. Resuls of Augmened Dickey Fuller and Phillips-Perron ess indicae ha deb raios are nonsaionary implying he exisence of an unsusainable deb oulook. Resuls of he Chow es, employed o es if a srucural break can be observed in 978 as a resul of moving away from he command economy o a markeoriened economy, indicae ha he policy change has no led o have any fixed change in he mean of deb serials. The resuls compel us o conclude ha public deb in Sri Lankan is no susainable, so ha a swich is required from foreign deb o oher sources of financing of fiscal defici or defici reducion. Inroducion Deficis in governmen budge and sock of deb are imporan facors of economic performance of a counry. Mos aggregae economic variables end o flucuae a differen levels wih effecs on various inerrelaed facors. A presen, a long erm budge defici is observed in Sri Lanka. For example, average budge defici as a percenage of Gross Domesic Produc (GDP) was -6. during 96-69, -7.3 in 97-79, -.28 in 98-89, and -7. during 2-22 (Cenral Bank of Sri Lanka22). Furher, he sock of deb in Sri Lanka has also been gradually increasing. Though from hese rends, one migh conjecure ha he public deb in Sri Lanka is no susainable, a formal analysis by applying appropriae mehods are required o undersand wheher public deb in Sri Lanka is unsusainable. As unsusainable public deb ends o negaively cause fuure economic developmen, he resuls of hese empirical ess would generae imporan insighs for effecive implemenaion of fiscal policy. This sudy aims o analyze he effecs caused byhe Sri Lankan budge defici on oal deb susainabiliy and is relaionship. There are many sourcesforfinancing governmen budge defici such as prining money, borrowing from local insiues, adjusing foreign reserves,and obaining foreign loans.each financing mehod would enail differen macroeconomic repercussions; money prining would be linked o inflaion, use of reserves migh propagae exchange rae movemens and possible balance of paymens crises, borrowing from foreign sources migh provoke exernal debcrises, and inernal borrowing wih higher ineres burden and poenially explosive deb dynamics would cause inflaion (Akcaye al 25). Unsusainable exernal deb ends o creae macroeconomic insabiliy, loss of inernaional compeiiveness leading o long erm economic downurn. Every counry is legally bound o sele heirforeign deb.for deb selemen o be susainable, he presen value of he resources ransferred o foreign counries should be equal o hesaring deb sock of a counry. A counry which is borrowing should creae a budge surplus for he possible repaymen of loans in he fuure. If he discouned value of he fuure budge surplus is less han he presen deb sock o be seled, he deb is no susainable a ha ime. As a resul, he deb services (ineres and he principle repaymen) end o be unseled. When he deb services are on operaion, welfare of he counry is Senior Lecurer. Deparmen of Economics & Saisics, Universiy of Peradeniya, Peradeniya, 24, Sri Lanka, ogdayab@gmail.com 2 Economics and Saisics, Universiy of Peradeniya, Peradeniya, 24, Sri Lanka, imaliwimalasiri@yahoo.com P age

2 Susainabiliy of Governmen Deb in Sri Lanka OG Dayarana-Banda/A. A. S. Priyadarshanee decreased. Some counries aemp o increase welfare by no paying he deb services. As a resul, he willingness o sele deb will be decreased. When here is coninues primary surplus, likelihood of he susainabiliy of deb end o increase due o a few reasons. Firs likelihood of he deb susainabiliy will increase by decreasing he crowding ou effecs hrough decreasing of he accrued ineres. I would also increase by increasing he efficiency of resource disribuion and increasing income hrough reducion of accumulaed of ineres paymens. Finally deb susainabiliy would be enhanced by increasing he demand of money base as a resul of decreasing inflaionary expecaion. Review of Lieraure The susainabiliy of deb is normally evaluaed by calculaing he governmen gross or ne deb as a percenage of Gross Domesic Produc (GDP). If he deb of he governmen is o be susainable, he deb o GDP raio should be sable or decreasing over ime. There are wo concepual approaches of calculaing he susainabiliy of deb. These are accouning&presen Value Consrain (PVC) approaches. The raio of deb o GDP is used mainly in accouning. Pasinei(998) and Goldsein (23) have expressed ha if a sable deb raio can be managed over ime hen fiscal policy would be susainable. For deb o be susainable according o Cuingnon(997), a sable raio of primary defici o GDP (or surplus has seleced he growh arges of gross domesic produc and when he real ineres rae is sable) and deb o GDP should be generaed.according o he PVC approach,he no Ponzi Game (NPG) condiionhe fac ha here is no way of new borrowings o sele deb is used o analyze deb susainable. The deb o GDP raio calculaed a a paricular ime o be susainable, he discouned presen value of expeced fuure surplus should be equal o unseled deb sock. This approach expecs o generae a budge surplus in fuure if here is no enough ne income o sele cumulaive deb and ineres a curren. As long as governmen pays ineres by obaining moredeb, he fiscal policy is no susainable. This approach is poining ou ha he presen defici should be covered by fuure surplus. Hamilon and Flavin(986) were pioneers of inroducing iner-emporal budge consrain (IBC). Kremers(998) furher developed his heory by poining ou ha if he IBC o be sauraed, he deb should grow a a lower rae han he ineres rae, and he budge defici and sock of deb should be saionary. Wilcox (989)furher showed ha i is beer o es he sabiliy of discouned deb. Based on hese approaches, he main deerminan facor of susainabiliy of he debsock of he governmen is he budge defici of a counry. Growh rae of GDP is also an imporan facor of measuring susainabiliy. The lieraure has exensively used saionary approaches and predicion mehods o empirically examine he deb susainabiliy. Several sudies relaed o he case of Sri Lanka have focused on susainabiliy of fiscal policy (Gupa 992, Siriwardana997, Jayawickrama24).Susainabiliy of foreign deb in relaion o fiscal policy has no been formally sudies wih reference o Sri Lanka. This paper focuses on filling his void. Maerials and Mehods The objecive of his research is o sudy he susainabiliy of Sri Lankan deb.we employ a novel mehod o empirically esimae he susainabiliy of public deb in Sri Lanka. The approach developed byjayawickrama(24), Jayawickrama&Abeysinghe(26), and Akcay e al (25) were used o build he analyical framework for he empirical sudy. I is assumed by his sudy ha he primary defici is financed by prining money and by financinghrough domesic and foreign bonds. The nominal value of budge defici for one year can be wrien as follows. G T + i B = M + B () Here, G, T, B, i, i B, M are oal expendiure of hegovernmen wihou ineres paying, ax Income of he governmen, sock of domesic and foreign deb of he governmen a he end of year, nominal P age

3 Inernaional Journal of Business and Social Research (IJBSR), Volume -4, No.- 7, July, 24 ineres rae on governmen deb, ineress which is o be paid for he year of -, and base money, respecively. ( + i ) B ( M M ) B = G T + (2) The following proporions were obained dividing his equaion by nominal gross domesic produc (Y ). G T B M M = + Y Y B Y ( + i ) Y Y Y (3) By subsiuing he economic growh rae o he soluion obained hrough above equaion Y Y λ =, he following soluion was obained: Y i ) b ( ) m m + λ λ) b = g + From above equaion, (4) d = g m m λ) (5) The value which is obained by he seignorage (he income generaed by governmen hrough prining of money) is subraced from he primary budge defici has been equaed o λ ) = ) i ρ is he discoun facor. d of equaion (5). b = ρ b ρ d (6) b = lim N E ( ρb) + N ( ρd ) j= + j (7) From here, i is considered ha no new deb are wihdrawn when seling he exising deb or No Ponzi Game.This requires he fuure limis of deb sock o be no posiive (or negaive). lim N ( b) E ρ (8) + N The idea of his condiion is he presen value of governmen sock of deb should converge o zero in he infinie fuure. According o ha, he presen value of expeced fuure budge surplus should be equal o he sock of deb of he governmen in any year where here exiss he above condiion. 2 P age

4 Susainabiliy of Governmen Deb in Sri Lanka OG Dayarana-Banda/A. A. S. Priyadarshanee b = j= ( ρd ) + j (9) According o his approach, he discouned raio of deb as a share of gross domesic produc should be sable for debo be susainable. The expeced raio of deb should be equal o zero a limi analysis. ( b) lim E ρ = N + N () I proves ha he debis no longer susainable if i is being a posiive facor wihou being equal o zero. The susainabiliy of governmen deb is esedby using face value deb as a share of GDP, marke valueof deb as a share of GDP and discouned marke valueof deb as a share of GDP. The nominal value of oal deb sock was divided by nominal gross domesic produc relaed o ha year o calculae he face value of governmen deb o GDP raio ( ) f b. The marke value of governmen deb o GDP raio ( b ) m was obained by dividing marke value of oal deb sock of he governmen by nominal GDP. To ge he marke value of governmen deb, he face value of governmen deb was discouned wih one plus he yield on governmen deb. I was a challenge o ge he yield on governmen deb because he ime aken for he deb o be maured is differen. Accordingly, he yield on governmen deb was derived by dividing he governmen oal ineress paymens of he considered year by oal ousanding sock of deb of he previous year. The b was derived based on he above discouned marke value of governmendeb o GDP raio ( ) dm calculaed face and marke values using he following equaion (equaion ). ( ) bdm = b m + ρ () ρ k + λ + r = k= k is discoun facor Here, λ and r are real GDP growh rae and ineres rae respecively. Weighed averageineres rae (WAIR) for 52 years from 96 o 22 was calculaed by using equaion (2) and he ineres rae ( r ) was calculaed by dividing he sum of WAIR by 52. The WAIR was calculaed as follows. WAIR + d f = ai bi (2) f Where, a, b, i d, i is he domesic deb sock as a proporionae of oal sock of deb, he foreign sock of deb as a proporionae of oal sock of deb, ineres rae for domesicdeb and ineres rae for foreign deb,respecively. Resuls and Discussion Augmened Dicky-Fuller (ADF) and Phillips-Perron (PP) ess were used o es he saionary and nonsaionary of he face value of deb o GDP raio, marke value of deb o GDP raio and discouned marke 3 P age

5 Inernaional Journal of Business and Social Research (IJBSR), Volume -4, No.- 7, July, 24 value of deb o GDP raio. The calculaed riple deb serials were esed for uni roo only wih consan erm and wih consan erm and deerminisic rend.resuls are summarized in able. According o he resuls, he null hypohesis ofuni roo of face value of deb o GDP raio including consan erm and no rend canno be rejeced. Tha is, when i is esed by reducing riple lags o one lag he ADF es saisics was smaller han 5% Mackinnon Criical Value a all he hree imes. Similarly, he consan erm was posiive a all he imes.accordingly, he deb of Sri Lankan governmen is no seemingly susainable. Graph : Governmen Deb o GDP Raio Percenage Face Value Marke Value Discouned Marke Value Source: Auhors calculaion based on CBSL s daa However, he resuls obained abou governmen deb susainabiliy by furher sudies by conducing above ess relaed ohe marke value of governmen deb o GDP raio (bm) and he discouned marke value of governmen deb o GDP raio (bdm)in addiion o he face value of governmen deb o GDP raio are presened in able for furher verifying he empirical resuls. According o hose resuls, he marke value of governmen deb o GDP raio wih he consan erm and wih he consan erm and rend were no saionary a boh imes. Even hough he uni roo es was done by reducing he number of lags from hree o one, he ADF es saisics failed o rejec he null hypohesis of which uni roos were presen. The mos appropriae deb serial was o measure he susainabiliy of Sri Lankan governmen s deb was he marke value of discouned governmen deb o GDP raio. I was concluded ha, he serial was non saionary due o he es saisics being failed of rejecing he null hypohesis. Tha is he presen value of governmen sock of deb was no convergen o zero in infinie fuure. Accordingly, he sock of deb of he governmen relaed o any year is no equal o he presen value of expeced fuure budge surplus. Lasly, according o ha, boh he marke value and discouned marke value of he governmen deb o GDP raio were found o be non-saionary. Also, he consan erm of all hose imes has shown a posiive value. According o hose resuls, he governmen debin Sri Lanka was proven o be unsusainable. Table : Augmened Dicky Fuller Tes Consan wih no Trend Consan wih Trend Variable Firs Lag Second Lag Third Lag Firs Lag Second Lag Third Lag bf (-2.99) (-2.922) (-2.925) -o ( ) (-3.55) (-3.525) bm (-2.99) (-2.922) (-2.925) ( ) (-3.55) (-3.525) bdm (-2.99) (-2.922) -.88 (-2.925) ( ) (-3.55) (-3.525) Source: Auhors calculaion based on CBSL s daa. 4 P age

6 Susainabiliy of Governmen Deb in Sri Lanka OG Dayarana-Banda/A. A. S. Priyadarshanee To furher verify his, we employed Phillips-Perron ess. The riple daa serials were re-esed by aking maximum lags as hree and reducing o one for esing he saionariy. Those resuls are presened in able 2. Any deb serials failed o ge enough es saisics o rejec he null hypohesis ha he deb serial consis of uni roo a any ime. Accordingly, hose resuls furher prove ha he governmen deb in Sri Lankan is no susainable. Table 2: Phillips-PerronTes Consan wih no Trend Consan wi Trend Variable Firs Lag Second Lag Third Lag Firs Lag Second Lag Third Lag bf bm bdm Source: Auhors calculaion based on CBSL s daa. When here is a srucural change in he daa serials, he resuls are being challenged. Tha is according o (Perron 989), when here is a srucural change in daa serials, he null hypohesis of he presence of uni roos explained by ADF es can be bias. Therefore,possible srucural breaks were esed by using Chow Tes o see if here is a srucural change in he riple variables of he face value, marke value and discouned marke value of governmen deb o GDP raio. I was esed ha if here is any break in governmen GDP in 978 due o he open economic policy inroduced in he laer par of 97s. However, no significan break in any of he variables - face value of deb of governmen deb o GDP raio, marke value of governmen deb o GDP raio, and discouned marke value of governmen deb o GDP raio were observed.the resuls indicae ha he F saisics relaed o hose variables and log likelihood raios were no significan. Even hough all of hose above ess were done relaed o he year 979 any of F saisics or log likelihood raios were no numerically significan. The gap beween income and expendiure of he governmen or he budge defici is deermined by he fiscal policy. From 96 o 22in Sri Lanka, many poliical paries wih differen economic policies have been ruling he counry. The fiscal policy implemened based on various economic policies which has no remained sound according o above resuls. However, i is needed o broadly sudy he reasons for he unsusainabiliy of governmen deb. Graph 2: Toal Borrowing in Sri Lanka Rs. Million Domesic Borrowing Foreign Borrowing Toal Borrowing Source: Auhors calculaion based on CBSL s daa. Firs, i is imporan o discusshow borrowings, ineres paymens, and srucure of he oal sock of deb are srucured and how i has affeced he susainabiliy of deb.according o he 2 nd graph, Sri Lanka s borrowings gradually grew afer he year 978. Bu i shows a growh of borrowing from local sources from he beginning of 99 decade. The unsusainable deb leads he ax and non-ax income of he economy o decline and increase he curren expendiure. Then, o generae he money needed o cover expense and pay 5 P age

7 Inernaional Journal of Business and Social Research (IJBSR), Volume -4, No.- 7, July, 24 deb services, governmen has o ge new loans. The local wihdrawal of loans could have lead for he oal debo be unsusainable. There is also a big risk of local ineres rae o increase, when he local wihdrawal of loans is higher. This would resul in crowding privae invesmen ou. Graph 3: Ineres Paymen on Governmen Deb in Sri Lanka Rs. Million Domesic foreign Toal Source: Auhors calculaion based on CBSL s daa. Graph 3 explains how he governmen deb ineres paymens srucured as locally and inernaionally.a larger porion of oal ineres paymens is domesic ineres paymens. Even hough he domesic wihdrawal of loans is comparaively higher, he domesic ineres paymens have conribued o oal ineres paymens a a higher rae han ha. The foreign ineres paymen is comparaively lower, and a considerable decrease can be seen in 25. This has adversely affeced he oal sock of deb of Sri Lanka o increase, ha i has obained more deb from domesic sources a higher ineres raes. Wih he higher ineres paymens and he higher he deb service paymens, i furher leads o increase he budge defici. This has significanly conribued for governmen deb in Sri Lankan o be unsusainable. According o graph 4, he domesic deb sock from he oal deb sock of he Sri Lankan governmen has increased afer 995. Even hough he wihdrawing of loans from domesic sources is comparaively higher han ha of he foreign sources, here is no big gap beween domesic and foreignsources on he cumulaive deb.however, he oal deb sock is shown a large increase afer 978. Meanwhile, when he oal sock of deb goes up, he governmen has o allocae more for expendiure o coverdeb, which appears o have conribued for governmen deb o be susainable. Graph 4: Toal Deb Composiion in Sri Lanka Rs. Million Domesic Foreign Toal Source: Auhors calculaion based on CBSL s daa. 6 P age

8 Graph 5: Governmen Toal Income o Deb Service Raio Susainabiliy of Governmen Deb in Sri Lanka OG Dayarana-Banda/A. A. S. Priyadarshanee Percenage Graph 5 shows ha he oal income of governmen o deb service raiohas slowly grown unil 978 and gradually grown a a higher rae afer 978. The governmen obained new loans o large infrasrucure projecs including educaion, and healh projecs.this resuled in increasing he deb services. This conribues o accumulae deb resuling in unsusainable sock of deb. The curren expendiure has grown faser han he growh of governmen gross domesic produc in many years according o he graph 6. The growh of curren expendiure in 978 was 69% and economic growh rae was 8.2%. When he gap beween economic growh and growh of curren expendiure increases, deb also ends o go up. This leads o furher borrowing making he public deb o be unsusainable. Graph 6: Curren Expendiure o GDP Raio 8 Rs. Million GDP Growh Rae Curren Expendiure Growh Rae Source: Auhors calculaion based on CBSL s daa The revenue o he gross naional produc raioof Sri Lanka is lower han ha of he recen indusrialized counries in he Norh Asia and developed counries. The ax income was 4.2% as a percenage of gross naional produc in 27 in Sri Lanka, which has decreased o 3.3% in 28. However, he ax income was approximaely 2% from 96 o he firs half of 99. Inefficiency in ax adminisraion and differences in wage srucure have led o decrease he aiudes of people abou ax auhoriy. Though ax income was expeced o increaserecenly, i has acually decreased. 7 P age

9 Inernaional Journal of Business and Social Research (IJBSR), Volume -4, No.- 7, July, 24 Conclusion The resuls of he empirical ess clearly indicae ha he governmen deb in Sri Lanka is no susainable. These resuls may be fuure verified if one uses he mehods such as Vecor Auoregressive mehod and impulse response funcions, which are no he subjec of focus in his paper. A number of facors were idenified as responsible for he governmen deb o be unsusainable over he years in Sri Lanka. Overreliance on various sources of deb o finance he budge defici has resuled in his predicamen. A large number of public works programmes were implemened during he pas few decades mainly hrough foreign commercial borrowing. The oupu generaed by hese invesmens was basically non-radable goods and services making i exremely difficul for he governmen o find means for financing deb services. Expor growh was no faciliaed mainly by hese invesmens. References Akcay, O. C., C. E. Alper, and S. Ozmucur.(25).Budge Defici, Inflaion and Deb Susainabiliy: Evidence from Turkey (97-2). Bogazici Universiy Research Papers,No. 22 Cenral Bank of Sri Lanka.(96-28). Annual Repors. Colombo: Cenral Bank of Sri Lanka. Cuingon, John T.(997). Analyzing he Susainabiliy of Fiscal Deficis in Developing Counries.Deparmen of Economics, Georgeown Universiy, Washingon D.C. Goldsein, M.(23).Deb Susainabiliy, Brazil and he IMF.Working Paper, Insiue of economics. Inernaional Gupa, K.L. (992). Budge Deficis and Economic Aciviy in Asia. London: Ruledge. Hamilon, J. D. and M. Flavin.(986). On he Limiaion of Governmen Borrowing: A Framework for Empirical Tesing.AmericanEconomic Review 76: Jayawickama, Ananda. (24). Fiscal Policy Susainabiliy: Evidence from Sri Lanka.Sri Lanka Economic Journal 5(): Jayawickrama, Ananda and TilakAdeysinghe.(26). Susainabiliy of Fiscal Deficis: The U. S. Experience SCAPE Working Paper Series, Paper No. 27/, Deparmen of Economics, Naional Universiy of Singapore: -28 Kremers, J. M.(988). Long Run Limis on he US Federal Deb.Economics Leers 28: Pasinei, L.L.(998). The Myh (or folly) of he 3% Defici/GDP Maasrich Parameer.Cambridge Journal of Economics 22:7-36. Perron, P. (989). The Grea Crash, he Oil Price Shock and he Uni Roo Hypohesis.Economerica 57: Siriwardana, K.K.M.(997). An Analysis of Fiscal Susainabiliy in Sri Lanka.Saff Sudies. Cenral Bank of Sri Lanka: 8-8. Wilcox, David W.(989). The Susainabiliy of Governmen Deficis: Implicaions of he Presen Value Borrowing Consrain. Journalof Money, Credi and Banking 2: P age