India: Why Fiscal Adjustment Now

Transcription

1 Public Disclosure Auhorized Public Disclosure Auhorized Public Disclosure Auhorized Public Disclosure Auhorized India: Why Fiscal Adjusmen Now Brian Pino and Farah Zahir 1 Absrac India s growh performance has been impressive over he las wo decades. Bu is susainabiliy has been in quesion, firs wih he 1991 fiscal-balance of paymens crisis (BoP), and hen again afer 1997/98, when fiscal deficis reurned o he 10 percen of GDP range and governmen deb grew. This paper analyzes he deerioraion in India s public finances and presens evidence suggesing ha, in he absence of a fiscal adjusmen, low inflaion and high reserves may have been pursued a he expense of long-run growh and povery reducion. Resolving his inflaionexernal vulnerabiliy-growh policy rilemma requires fiscal adjusmen. In making is case, he paper shows, firs, ha fiscal fundamenals have weakened afer 1997/98 even when compared wih he pre-1991 crisis period. This has coninued in spie of he recen record lows in ineres raes. Second, he fiscal sance is no conducive o long-run growh and povery reducion because capial spending has been cu o accommodae higher ineres paymens and oher curren spending, wih expendiures on he social secors sagnaing. Third, wihou a fiscal adjusmen, he deb burden is likely o reach unmanageable levels by he end of he Tenh Plan period. In conras, a phased adjusmen beginning now and focusing on a relaively small se of reforms is likely o improve deb dynamics subsanially over he same horizon, while also promoing faser growh and povery reducion. World Bank Policy Research Working Paper 3230, March 2004 The Policy Research Working Paper Series disseminaes he findings of work in progress o encourage he exchange of ideas abou developmen issues. An objecive of he series is o ge he findings ou quickly, even if he presenaions are less han fully polished. The papers carry he names of he auhors and should be cied accordingly. The findings, inerpreaions, and conclusions expressed in his paper are enirely hose of he auhors. They do no necessarily represen he view of he World Bank, is Execuive Direcors, or he counries hey represen. Policy Research Working Papers are available online a hp://econ.worldbank.org. 1 Boh auhors are a he World Bank. We are graeful o Michael Carer and Sadiq Ahmed for encouraging us o wrie his paper. I liberally draws on India Susaining Reform, Reducing Povery, a World Bank Developmen Policy Review released in July In revision, he paper has benefied subsanially from commens from he peer reviewers Shanayanan Devarajan, Kalpana Kochhar, Mihir Rakshi and Suresh Tendulkar. Helpful commens were also received from Ijaz Nabi and Sephen Howes.

2 1. Inroducion India s economic growh acceleraed in he second half of he 1980s along wih rising fiscal imbalances. The laer spilled over ino curren accoun deficis, depleing reserves and leading o a balance-of-paymens (BOP) crisis in As par of he program of sabilizaion and srucural reform ha followed he crisis, cusoms and excise duies were cu and axes on financial inermediaion lowered wih he objec of creaing more compeiive and efficien real and financial secors. This marked a break wih decades of proecion and financial repression. The fiscal defici was also cu and he capial accoun parially liberalized o arac foreign porfolio and direc invesmen. As a resul of hese reforms, he economy grew rapidly and fiscal and governmen deb indicaors improved dramaically during he five years 1992/ /97, which coincided wih he Eighh Plan period. Bu during he Ninh Plan period 1997/ /02, growh slowed and fiscal deficis wen back o heir pre-1991 crisis levels, raising concern abou he slowing pace of fiscal and srucural reform. 2 Governmen deb grew rapidly as revenues fell and ineres paymens climbed, alhough inflaion and ineres raes sared falling in line wih global rends oward he end of his period. Annex 1 conains background maerial on he 1991 crisis and summarizes sudies on governmen deb susainabiliy hrough he 1990s. Looking back over he las wo decades, India s macroeconomic repor card has been mixed. Is growh performance during he 1980s was exceeded by only 9 ou of 123 counries, and ha during he 1990s by jus 19 ou of 137 counries. 3 Bu India has also ranked high on fiscal deficis. In 2000, ou of 74 counries wih a populaion exceeding 10 million, only 7 had general governmen fiscal deficis exceeding 7 percen of GDP. India was one of hem and only Turkey and Zimbabwe had higher deficis. 4 Ahluwalia (2002a) presens daa showing ha India s fiscal and deb indicaors are comparable o or worse han hose of Argenina, Brazil and Turkey, counries ha have acually experienced a serious recen macroeconomic crisis. He neverheless concludes ha India is no vulnerable o an exernal crisis in view of is build up of foreign exchange reserves, capial conrols, flexible exchange rae sysem and widespread public ownership of banks. This sands in conras o 1991, when India suffered a fiscal-bop crisis wih fiscal deficis of comparable size bu much lower deb. Wha has changed is he reserve build-up. Bu does his mean ha concern abou he fiscal sance is misplaced? No, for hree key reasons. The firs is ha even hough growh rose from a 3.5 percen rend rae over he previous hree decades o close o 6 percen over he 1980s and 1990s, his has no been enough. India sars he weny-firs cenury wih per capia income around half ha of China and Indonesia, counries ha in 1970 were a comparable sages of developmen. 5 No surprisingly, he Naional Developmen Council mandaed a arge of 8 percen per year in real GDP growh for he Tenh Plan (2002/ /07); bu he macroeconomic chaper of he Plan makes i clear ha achieving his goal calls for a subsanial cu in he fiscal defici o raise public savings and invesmen. The second poin is ha ineres paymens are claiming a growing share of revenues, limiing governmen capial expendiure and spending on prioriy social programs. During he 2003/04 Union Budge discussions, he Finance Miniser informed he Rajya Sabha ha "Of our revenue, 50 percen is swallowed by paymen of jus ineres on (governmen) deb. Anoher 20 percen goes on subsidies and 25 percen on defense. Wha am I lef wih? 6 In oher words, he qualiy of he fiscal sance has deerioraed and opions for adjusmen are going o be consrained by he high level and non-discreionary naure of ineres paymens. Third, even hough he high reserves offer insurance agains a 1991-ype crisis, his does no rule ou a fiscal 2 Ahluwalia (2002b), Williamson and Zagha (2002). 3 Srinivasan (2001). 4 Acharya (2001, 2002a, 2002b). 5 World Bank (2003), Figure 1. 6 Times News Nework, Times of India, New Delhi, March 14,

3 crisis iself. India is no immune o a Sargen-Wallace (1981) oucome, where rising ineres paymens compel he governmen o resor o inflaionary finance, damaging macroeconomic credibiliy and raising real ineres raes. If he high reserves are used o preserve exchange rae sabiliy, ha counry could have a siuaion of appreciaing real exchange raes and high real ineres raes, huring he real secor and lowering growh prospecs. Thus a a minimum, failure o address he fiscal imbalance will hur growh prospecs and se povery alleviaion back. An alernaive viewpoin argues ha wih excess capaciy in indusry and record levels of reserves and food socks, his is he righ ime for a fiscal expansion o simulae growh. 7 As is shown below, he wo main argumens agains his reasoning are firs, ha governmen deb and ineres paymens have already reached exremely high levels, consraining couner-cyclical fiscal policy. Second, he build up in reserves has no been a foruious acciden bu driven by a macroeconomic policy mix designed o keep inflaion low and minimize exernal vulnerabiliy in spie of he fiscal deerioraion. The need for fiscal adjusmen now arises no because a crisis is imminen bu because posponing reform is likely o resul in an unmanageable deb and ineres burden by he end of he Tenh Plan period nowihsanding he curren low ineres raes and burgeoning reserves. In conras, a phased adjusmen beginning now and focusing on a relaively small se of reforms is likely o yield subsanial, posiive benefis over he same horizon and be more conducive o longrun growh and povery reducion. By adjusmen is no mean an immediae, drasic cu in he fiscal defici. The only way o do so given he exraordinarily high level of ineres paymens would be o cu capial expendiure even more, which would be undesirable. Insead, he advocaed focus is on revenue mobilizaion and redirecing non-meri subsidies oward capial and developmen expendiure. The paper makes is case firs by showing ha a key facor in he fiscal deerioraion during he Ninh Plan period (1997/ /02) has been he significan reform-induced losses in revenue (boh indirec ax revenue and financial-repression revenue), and also he governmen's inabiliy o adjus public spending (as ineres paymens ake up an increasing share of he budge). Deb dynamics worsened over 2002/03 despie he record lows in ineres raes. The reason for his and he inadvisabiliy of gambling on he persisence of low ineres raes are spelled ou. Second, i demonsraes ha governmen deb dynamics are showing signs of being unsusainable. The primary defici (non-ineres fiscal defici) is large and real ineres raes have converged o growh raes. Third, he Reserve Bank of India s (RBI) policy of building up reserves and serilizing hem (o cushion agains exernal shocks and keep inflaion low, respecively), coupled wih consrains on governmen spending (because of fiscal deficis and public deb) has resuled in low public and privae invesmen, puing long-run growh in jeopardy relaive o he levels needed for rapid povery alleviaion. Moreover, while higher reserves have increased liquidiy, he join balance shee of he general governmen and he RBI has weakened. Fourh, projecions show ha posponing fiscal adjusmen will push he deb burden o unmanageable levels by he end of he Tenh Plan period; bu a phased adjusmen ha begins immediaely will lower deficis and grealy improve spending composiion. Before proceeding, wo poins are worh noing. Firs, srucural reforms o underpin faser growh are crucial o any fiscal reform scenario. These are no discussed here; World Bank (2003a) conains a deailed discussion. Second, he Governmen of India (GoI) and he RBI are cognizan of he need for a fiscal adjusmen, as evidenced by numerous iniiaives and policy pronouncemens. The key is o find he momenum for implemenaion in he presen environmen of high reserves and low ineres raes, a momenum his paper hopes o advance. 7 Chandrasekhar (2000). 3

4 2. Macroeconomic Fundamenals 2.1 Public Finances and Exernal Vulnerabiliy Figure 1 depics rends in he general governmen s (cener plus saes consolidaed) gross fiscal defici and deb as percenages of GDP in 1985/ /03. I shows ha he gross fiscal defici averaged 9 percen of GDP before he crisis. I hen fell sharply during he period of high growh and fiscal resrain ha marked he Eighh Plan period (1992/ /97), bu resumed growing equally sharply afer 1997/98, reurning o he 9-10 percen of GDP range during he Ninh Plan period (1997/ /02). Figure 1: General Governmen Defici and Deb Sock (% GDP) 1985/ / Fiscal defici /GDP Eighh Plan (92/93-96/97) Ninh Plan (97/98-01/02) Deb sock /GDP 85/86 87/88 89/90 91/92 93/94 95/96 97/98 99/00 01/ Fiscal de fici Source : Budge Documens, RBI, Minisry of Finance, IMF and Saff Esimaes Deb sock Figure 1 also shows ha he raio of general governmen deb rose from approximaely 58 percen of GDP in 1985/86 o 85 percen in 2002/03. 8 This deb was he equivalen of $440 billion wih a rupee-o-foreign currency mix of 86:14 compared o 77:23 on he eve of he 1991 crisis. Deb of cenral public enerprises was anoher 10 percen of GDP, and guaranees by he cener and saes, he laer largely in suppor of borrowing by loss-making public enerprises in power and irrigaion, amouned o 12 percen of GDP. Wih fiscal defici levels in he 10 percen of GDP range and governmen deb more han 25 percenage poins of GDP higher han he mid-1980s, public finance fundamenals could be considered worse han in he pre-1991 crisis period. Table 1 shows rends in public finance and exernal vulnerabiliy indicaors. Daa values over 1985/ /90, he five years preceding he 1991 crisis, serve as a benchmark. The 1990s are divided ino he period of rapid growh, 1992/93-96/97, which coincided wih he Eighh Plan (shown in he able as 8 h Plan avg.); and he slowdown, 97/98-01/02, which coincided wih he Ninh Plan period (9 h Plan avg.). 9 8 Throughou his paper, he componen of general governmen deb denominaed in foreign currency is valued a curren exchange raes. 9 This division of he ime periods is used hroughou he paper. 4

5 The fiscal indicaors shown in Table 1 depic a percepible deerioraion during he Ninh Plan period. Gross fiscal deficis reurned o he levels of he mid-1980s afer being cu sizably during he Eighh Plan period. Revenue deficis (curren spending minus revenues, noe ha privaizaion proceeds are reaed as financing, no as revenue) doubled from less han 3 percen of GDP in he second half of he 1980s o 6 percen during he Ninh Plan period. Primary deficis were reduced subsanially during he Eighh Plan period relaive o he second half of he 1980s; bu hen rose during he Ninh Plan period. Comparing he Ninh Plan period wih he mid-1980s, he fac ha he primary defici was much smaller while he revenue defici was much larger and he gross fiscal defici abou he same can only mean ha rising ineres paymens displaced capial expendiure. General governmen deb fell from 71 percen of GDP a he end of 1989/90 o 65 percen a he end of he Eighh Plan period bu hen rose by 15 percenage poins o he end of he Ninh Plan, and anoher five percenage poins he following year. Table 1: Key Macroeconomic Indicaors % of GDP (unless oherwise defined) Fiscal and Deb Pre-crisis period 85/86-89/90 Crisis 90/91 8h Plan avg. 9h Plan avg. Gross fiscal defici 1/ Revenue defici 1/ Primary defici 1/ Deb ousanding 2/ Exernal Vulnerabiliy Forex reserves $ bn 2/ M3/Forex Reserves % 2/ Shor-erm exernal deb 3/ /Forex reserves % 2/ Curren accoun balance Real GDP growh % Noes : 1/ For he general governmen. The figures for 2002/03 are revised esimaes for cenral gov. and saff esimaes for sae governmens. 2/ For end of las fiscal year in he period. 3/ Exernal deb wih original mauriy less han a year plus all long-erm deb falling due over he nex one year. Source: Governmen budge documens, Handbook of Saisics on Indian Economy and Annual Repor , Saff esimaes 02/03 RE In conras, exernal vulnerabiliy indicaors srenghened impressively. Foreign exchange reserves seadily buil up afer he 1991 crisis wih a jump of $21 billion in 2002/03, and now surpass $95 billion. The able also shows wo sandard measures of inernaional liquidiy: broad money, or M3, divided by foreign exchange reserves; and shor-erm exernal deb divided by reserves. 10,11 M3 was over 3,600 percen, and shor-erm exernal deb 365 percen of reserves he year before he BOP crisis. The firs liquidiy measure fell o less han 500 percen by March 31, 2003, and he second o a safe 18 percen. Boh measures peaked a or around he ime of he 1991 crisis and have fallen dramaically since. Their behavior explains why India is no on he verge of anoher BOP crisis in spie of he deerioraion in he fiscal fundamenals. 10 M3 is defined as currency wih he public, demand deposis wih banks, oher deposis wih Reserve Bank of India, pos office savings bank deposis and ime deposis wih banks. 11 IMF (2000) idenifies he raio of shor erm exernal deb o reserves as a key indicaor of vulnerabiliy o an exchange crisis. Chang and Velasco (1998) idenify inernaional liquidiy as a key facor in he Eas Asian crisis. 5

6 The shif oward long-erm domesic (rupee) deb in governmen financing afer he 1991 crisis has also boosed liquidiy. The weighed average mauriy of deb issued hrough daed securiies more han doubled from 6.6 years in 1997/98 o 14.3 in 2001/ Moreover, over 85 percen of cenral and sae governmen securiies are held by naionalized banks, The Sae Bank of India, he sae-owned Life Insurance Corporaion and RBI while he remainder is held by he muual fund, Uni Trus of India, he Naional Bank for Agriculure and Rural Developmen (NABARD), employees providen fund and privae banks. 13 In shor, much of he domesicallyissued deb is in he hands of governmen-owned or conrolled eniies. 2.2 Fiscal Trends 1985/ /02 Table 2 shows ha general governmen revenues fell considerably during he Ninh Plan period relaive o he second half of he 1980s. 14 Compared o he second half of he 1980s, capial expendiure fell by over hree percenage poins of GDP during he Ninh Plan period, while he sum of ineres, adminisraion and pensions rose by hree percenage poins of GDP and a 22 percenage poins of revenue. Developmenal spending on healh and educaion sagnaed, while ha on irrigaion, power and ranspor acually declined. Even he fiscal improvemen secured during he Eighh Plan period involved a large compression of capial spending, as can be seen from he able. There was a hope ha his compression would be offse by rising privae invesmen in infrasrucure; bu wih he excepion of elecom, his has no happened on he desired scale. As noed in RBI (2003a), page III-46, a larger privae role in infrasrucure would require insiuional reform and economically efficien user charges o ensure he reasonable reurn on invesmen. Table 2: General Governmen Fiscal Trends 1985/ /02 % of GDP 85/86-89/90 8 h Plan 9 h Plan 02/03 (unless oherwise defined) Avg. Avg. Avg. Es. 1/ Revenue Curren expendiure 2/ Capial expendiure 3/ Gross fiscal defici Memo Ineres Ineres/Revenue Irrigaion+ power+ ranspor Ineres+ admin.+ pensions (Ineres+ admin.+ pensions)/revenue Noes: 1/ Revised esimaes for he cener and budge esimaes for he saes 2/ Refers o Revenue expendiure in he budge 3/ Refers o Capial oulay and ne loans and advances from he cener o he saes Sources: GOI budge documens(2003/04 and various issues), RBI bulleins, CSO, Handbook of Saisics on Indian Economy 2001/02 (RBI), Saff esimaes 12 RBI (2001a), Chaper XI Public Deb Managemen. 13 RBI (2001b), Table While 2002/03 shows a large revenue increase, i is based on he budge esimae for saes, which ends o be opimisic. 6

7 Table 3 presens more deail on he fiscal adjusmen during he Eighh and Ninh Plan periods. 15 The second column shows ha, relaive o he mid-1980s, he fall in revenues of 1.5 percenage poins of GDP and he rise in ineres paymens of 1.3 percenage poins was neuralized by a cu in capial expendiures of offseing magniude. The fall in revenues a a ime of acceleraing growh, especially of indusry and manufacuring, he mos ax-buoyan componens of GDP, is aribuable o he cus in cusoms and excise duy raes underaken as par of he reforms following he 1991 crisis. The objec of hese cus was o increase efficiency in producion and rade, wih he fall in indirec axes o be recouped by improvemens in direc ax collecions, alhough he laer has no so far occurred. Table 3: Fiscal Adjusmen Dynamics, 1985/ /02 (based on period averages) % of GDP 8 h Plan (92/93-96/97) versus 85/86-89/90 9 h Plan (97/98-01/02) versus 8 h Plan Revenues Primary defici Deb Ineres Capial expendiure Gross fiscal defici The rise in ineres paymens was due a rise in average deb/gdp levels and a rise in average ineres raes. Conrolling for he former, he rise in ineres raes accouned for abou 1 of he 1.3 percenage poins increase in ineres paymens. We conservaively use 1 percenage poin of GDP as an esimae of he loss of financial repression revenues as a resul of inroducing aucions for governmen securiies in 1992/93, a general freeing up of deposi and lending raes and a large reducion in he cash reserve raio and sauory liquidiy raio for banks. Financial repression has hisorically played an imporan role in India s public finances. Giovannini and de Melo (1993) esimae ha revenues from financial repression averaged 2.86 percen of GDP per year for he years Using a similar mehodology, Klezer and Kohli (2001) esimae ha financial repression revenues fell from an average of 6 percen of GDP over o 2.9 percen over We do no inend o revisi hese compuaions or address relaed mehodological issues; based on he rise in ineres paymens during he Eighh Plan period, we sugges ha easing financial repression cos he budge 1 percenage poin of GDP. 17 Thus, he simple sory of fiscal adjusmen during he Eighh Plan period is ha revenues fell by abou 2.5 percenage poins of GDP as a resul of ax reform and financial liberalizaion, and was compensaed by cuing capial expendiure insead of finding alernaive revenue sources as originally envisaged. 15 We omi he crisis year, 1990/91, and he year following i. 16 Compued as he ex pos differenial beween foreign and rupee ineres raes imes he amoun of governmen deb held ouside he cenral bank. This calculaion ignored he effec of ax concessions on governmen deb. 17 The rise in ineres raes during he Eighh Plan period migh also reflec higher inflaionary expecaions following he devaluaion of 1991; bu only he liberalizaion of ineres raes allowed his o be refleced in realized ineres raes. Furher, i would ake a few years for he full effec o feed hrough, unil all he hisorically issued deb maured. 7

8 Table 3 also shows ha he revenue decline coninued and ineres paymens rose furher during he Ninh Plan period, compensaed by addiional cus in capial expendiure. Falling revenues accoun for 0.9 of he 1.4 percenage poins of GDP rise in primary defici relaive o he Eighh Plan period. As capial expendiure endured a furher cu of 0.4 percenage poins, i follows ha he rise in non-ineres curren spending in he form of higher salaries, pensions and adminisraive coss accouned for abou 1 percenage poin of GDP. This was probably driven by he Cenral Fifh Pay Commission award, he implemenaion of which began in 1997 a he cener, followed he nex year by he saes. Summarizing, we can say ha relaive o he pre-crisis mid 1980s, ax and financialrepression revenues fell by 2.5 percenage poins of GDP during he 1990s as a resul of needed reforms, offse by a compression in capial expendiure; 18 while he Fifh Pay Commission award probably had a negaive impac of abou 1 percenage poin of GDP. The impac of he laer is expeced o aenuae over ime, because of he discree naure of civil service wage increases. Moreover, he issue is no so much ha he civil service is over-saffed or he wage bill excessive when scaled by populaion and GDP respecively, bu more ha here is a skills imbalance, wih mos of he employmen concenraed in low-skill jobs, which pay considerably more han comparable jobs in he privae secor; wih fewer and probably under-paid saff a he higher skills level needed for efficien social service delivery. 19 Hence, reversing he fiscal slide needs o focus on mobilizing revenue and resoring governmen capial and developmen expendiure; cuing he wage bill is a secondary issue. 2.3 Governmen Deb Dynamics The change over ime in he general governmen deb-o-gdp raio can be decomposed ino wo pars: he firs aribuable o he primary defici/gdp raio (pd); and he second, a erm proporional o he difference beween he real ineres rae paid on governmen deb and he growh rae of real GDP, (r-g). During financial repression, we would expec he rise in deb o be less han he primary defici (as raios of GDP), because ineres raes are kep arificially low; bu wih financial liberalizaion, ineres raes would end o rise and migh evenually exceed growh raes, puing governmen deb on a much faser upward rajecory. We shall show ha his is wha happened during he Ninh Plan period. Figure 2 presens he primary defici and he erm proporional o (r-g) based on he sandard difference equaion for governmen deb shown in Annex 2. I can be seen ha afer 1994/95, he erm proporional o (r-g) becomes smaller in magniude and over he las four years, has been eiher posiive or close o zero, no longer offering he offse o he primary defici ha i did when financial repression was he norm. 18 While revenues fell and ineres paymens rose furher during he Ninh Plan period, i is impossible o say wha par of his is cyclical and wha par due o reforms, so he 2.5 percenage poins should be reaed as a lower bound. 19 World Bank (2003) conains an analysis of he wage bill issue and he need for civil service reform. 8

9 Figure 2 : Deb Dynamics Percen /87 88/89 90/91 92/93 94/95 96/97 98/99 00/01 02/03 Deb impac of (r-g) Primary defici r-g The implied value for (r-g) iself can be obained hrough appropriae soluion from he deb difference equaion shown in Annex 2, and his is also ploed in Figure Implied (r-g) has chared an upward rend afer 1994/95 and has been eiher close o zero or posiive since 1999/00, in spie of he record lows in ineres raes over he las 18 monhs. On average, he implied real ineres rae was a subsanial 6 percenage poins below he growh rae during he Eighh Plan period and explains why he raio of general governmen deb o GDP (GGD/GDP) fell in spie of primary deficis averaging over 2 percen of GDP. Bu during he Ninh Plan period (i) he implied ineres rae converged o or exceeded he growh rae, as anicipaed in Serven (1996) and Reynolds (2001), while (ii) he primary defici rose o an average of 3.5 percen of GDP. The combinaion of (i) and (ii) is by definiion unsusainable wihou a fiscal correcion, which, as argued in Box 1, is required in spie of he curren low ineres raes. 20 Lahiri and Kannan (2002) direcly calculae he ineres rae as oal ineres paymens divided by oal deb, alhough sricly his should be augmened by a erm capuring he exchange rae losses/gains on he exernal deb componen of public deb. To do he laer accuraely, one would need deailed informaion on he currency composiion of public deb. We have calculaed he effecive rae of ineres assuming ha all exernal deb is dollar-denominaed, and excep for he years immediaely before and afer 1991, when he rupee depreciaed significanly agains he dollar, he implied ineres rae racks he direcly calculaed ineres rae quie closely. 9

10 Box 1: Will Low Ineres Raes Solve he Problem? Do he record lows in ineres raes mean ha he emphasis on unsusainable deb dynamics is misplaced? No, for he following reasons: Low ineres raes sem no from beer macroeconomic fundamenals bu from weakness in he global economy and capial inflows. To he exen ha capial flows ino India have been driven by one-off evens since Sepember 11, 2001 including fears of increased scruiny of accouns held overseas as par of ani-money laundering drives or by he insabiliy in Iraq more recenly, i would be risky o slow fiscal reform on a gamble ha such flows will coninue indefiniely. If capial flows subside and he global economy picks up, ineres raes will once again be deermined by medium-run inflaionary expecaions as molded by fiscal and macro fundamenals, which have weakened. Low ineres raes for he governmen do no auomaically mean low ineres raes for he privae secor, excep for he bes credis, as argued laer in his paper. Low ineres raes do no auomaically improve deb dynamics. In fac, he increase in GGD/GDP has acceleraed from less han 2 percenage poins of GDP per year over he firs hree years of he Ninh Plan period (1997/98 o 1999/00) o over 4.5 percenage poins of GDP per year over he las hree years (2000/01 o 2002/03) nowihsanding he low ineres raes. Why? Firs, primary deficis have gone up during he Ninh Plan period relaive o he Eighh, increasing borrowing needs. Second, wha maers is he difference beween ineres raes and growh raes. Slowing growh could offse he beneficial impac of lower ineres raes. 21 Third, wha maers for deb dynamics is he average cos of GGD. Low ineres raes a he margin obviously help, bu will have o persis for sufficienly long (depending upon he inheried cos and mauriy srucure of he ousanding sock of GGD) o make a den on he average rae. On he oher hand, since he low ineres raes have resuled from low global inflaion, domesic inflaion has also come down wih India s increasing openness, which hurs o he exen ha high nominal coupon deb has been issued in he pas. 22 The ne effec is herefore ambiguous: low marginal ineres raes help, low inflaion hurs. One soluion ha has been proposed is o engage in deb swaps in order o capialize on he curren low ineres raes. However, hese have o be carefully approached. To he exen ha hese are beween he cener and saes, hey will no affec he posiion of he general governmen unless he cener can pre-pay he ulimae credior wihou penaly; and o he exen ha hey are wih banks, bu no marke-based (as for example in a scheme where he resuling capial gain is shared beween banks and he governmen or fully capured by he governmen), hey will weaken boh he balance shees of banks and erode heir income wih negaive consequences down he road. Bu aligning small savings raes and providen fund (PF) reurns wih marke indicaors will lower he cos of new borrowings. 23 Anoher opion is o prepay exernal deb. The firs wo possibiliies are o use divesmen receips or raise funds by borrowing in he domesic marke o pay off exernal deb. The wo are equivalen. If divesmen receips are used o prepay deb insead of financing he fiscal defici hen new borrowing would go up exacly by he amoun of he divesmen receips. Thus, prepaymen is equivalen o replacing an exernal liabiliy wih a domesic one; i does no change he balance shee of he governmen. Ineres paymens will go down if he exernal deb carries a much higher ineres rae (apar from posing exchange rae risk) han he new domesic borrowing. However, engaging in such prepaymen also has is limis. Domesic financial insiuions are already carrying some of he highes levels of governmen paper in he world relaive o deposis and addiional borrowing in he domesic marke o pay off exernal deb means less funding for he privae secor. The hird possibiliy is o borrow from he RBI which hen uses is reserves o pay off deb. This will improve he join income posiion of he governmen and RBI (bu no he join balance shee) if reurns on reserves are lower han he raes paid on exernal deb. 24 Thus, while deb swaps and prepaymen may lower ineres paymens a he margin, a more fundamenal fiscal adjusmen is needed. 21 Bu why doesn growh pick up wih lower ineres raes? An answer is aemped below. 22 Thus, RBI (2003a), page IV-20, noes ha he average ineres rae on new Cenral Governmen securiies issued during he year reached a peak of percen in 1995/96, declining gradually o percen in 1999/00 and 9.44 percen in 2001/02. Even hese laer raes are high relaive o oday s inflaion raes and growh. 23 However, pre-paying deb now held by he NSSF ha creaes a wedge beween is obligaions and wha i receives from he governmen will only creae a coningen liabiliy. 24 I will also replenish RBI s holdings of governmen securiies, which have been depleed by he serilizaion of capial inflows. Of course o he exen his happens, i worsens he income posiion of RBI as foreign exchange reserves ypically have lower reurns han domesic governmen paper. 10

11 2.4 Quasi-Fiscal Issues In addiion o he deerioraion in explici public finances, quasi-fiscal aciviies have become a growing hrea since he mid-1990s. These may be grouped ino hree caegories: losses associaed wih off-budge aciviies, noably, supply of irrigaion and power; coningen liabiliies, which may eiher be explici, e.g., guaranees of bonds issued by loss-making public enerprises; or implici, e.g., financial secor bailou coss resuling from implici deposi insurance; and unfunded pension liabiliies. These could be inerrelaed, e.g., he losses of Sae Elecriciy Boards (SEBs) may be financed by bonds issued wih governmen guaranees. SEB losses were esimaed a Rs.332 billion in 2001/02, riple he level in 1996/ To pu his in perspecive, he gross fiscal defici of he saes was 4.4 percen of GDP in 2001/02, or abou Rs.1183 billion, a number which included only a porion of he power secor loss, namely, Rs. 83 billion. If he full amoun of he loss were included, he fiscal deficis of he saes would rise by an average of 30 percen. Some saes have sared reporing heir deficis wih and wihou he loss of he respecive SEB in recogniion of he growing problem. If SEB losses are explicily included in he general governmen defici, hese would increase by a significan percenage poins of GDP, i.e., he gross fiscal defici would be 11.5 percen of GDP insead of 10 percen. Such inclusion appears o be only a maer of ime - he Ahluwalia Commiee 26 recommended replacing pas-due arrears of SEBs o cenral governmen agencies by sae governmen bonds, while he riparie agreemen signed in March 2003 by RBI, sae governmens and Minisry of Power allows RBI o use amouns due o he saes o sele any unpaid bills presened o he respecive SEB by cenral agencies. Guaranees have risen sharply afer 1996, mosly in connecion wih borrowings by SEBs and special purpose vehicles (SPVs) esablished for large irrigaion projecs. 27 CRISIL esimaes ha of he marke borrowings by sae-level eniies guaraneed beween , Rs.440 billion will be called over he nex five years. While his may no seem much, his esimae excludes guaraneed loans from banks/fis such as NABARD, HUDCO, LIC, NCDC, REC, public secor banks and regional rural banks. Moreover, he guaranors hemselves, i.e., he sae governmens, are already heavily indebed. Financial secor bail-ou coss have no so far been significan, unlike in Eas Asia and some of he ransiion counries of Cenral and Easern Europe. Since 1992/93, relaively small amouns of money have been spen o assis naionalized banks, regional rural banks, UTI, IDBI and IFCI, wih anoher 0.8 percen of 2002/03 GDP idenified o help UTI, IFCI and IDBI. The banks repored gross non-performing loans (NPLs) were 10.4 percen of loans as of March 2002; ne of provisions he figure was 5.5 percen; and ne non-performing loans as a percen of asses were 2.3 percen or abou 1.5 percen of GDP. While some financial analyss have suggesed ha NPLs are subsanially undersaed by various evergreening mehods, a doubling of he gross figure would pu NPLs a less han 7 percen of bank asses, or abou 4.5 percen of GDP. Anoher indicaor of risk, lending for socks, real esae, and commodiies, is less han 4 percen of lending. While off-balance shee operaions are equal o nearly 60 percen of he balance shee, hree-fourhs are forward exchange conracs (half by he foreign banks) mosly relaed o 25 Source: Planning Commission, Governmen of India: Annual Repor ( ) on he Working of Sae Elecriciy Boards & Elecriciy Deparmens, May, The acual financial loss is even greaer because of collecion problems. 26 Repor of he Exper Group, Selemen of SEB Dues, May CRISIL (2002). 11

12 expors and impors and ofen wih RBI as he couner-pary. 28 While hese numbers are low in comparison wih Eas Asia and China, his has o be looked a in he conex of public secor deb ha is over 90 percen of GDP in an environmen of worsening governmen deb dynamics. Pension liabiliies have been growing rapidly boh a he cener and saes. A he cener, pension spending for civil servans grew by more han 20 percen per year during he 1990s, and has now reached 1 percen of GDP. The siuaion is even more difficul a he sae level. For example, in Uar Pradesh 29, he raio of pension spending o sae GDP increased from 0.4 percen o 1.2 percen during he 1990s. The Whie Paper on Tamilnadu Governmen s Finances issued in 2001 (pages 19-20) noes ha pension commimens have been growing a an incredible rae of 30 percen per annum on an average during he las five years and consiue he fases growing componen of he oal revenue expendiure of he Sae Governmen. Pensions as a percenage of revenue expendiure grew from 5 percen in 1991/92 o 13 percen in 2000/01. Pension expendiures a he cenral and sae levels are likely o keep on growing a a fas pace, especially a he sae level where employmen more han doubled over he pas 30 years. Preliminary esimaes conduced in 2001 using he World Bank PROST acuarial model sugges ha he presen value of cenral and sae pension liabiliies could amoun o 25 percen of GDP Crisis Proofing In spie of he rise in he level of governmen deb and he deerioraion in is dynamics, India is no on he verge of a macroeconomic crisis such as ha endured recenly by Russia, Turkey and Argenina, which apparenly exhibied beer fiscal fundamenals on he eve of heir crises han India has over he pas few years. 31 There are several reasons for his. Firs, here is only limied capial accoun converibiliy. Second, cross-counry comparisons are complicaed by differences in he currency composiion of deb, is mauriy, wheher or no he real exchange rae is in equilibrium and he size of coningen liabiliies and poenial balance shee currency/mauriy mismaches in he banking and corporae secors. In India, here has been a shif owards long-erm rupee deb in financing he fiscal defici afer 1991, while sricures on exernal borrowing by corporaes and banks auomaically lower he poenial for currency mismaches. Third, in conras o he recen crisis counries, India has an impeccable credi hisory (no defauls, no deb resrucurings) and a rack record of low inflaion and high macroeconomic sabiliy, which permis i o issue long-erm deb denominaed in is own currency, an opion unavailable o counries such as Russia, Turkey and Argenina. 32 Fourh, counries wih unsusainable public deb dynamics which evenually experienced crisis, such as Russia and Argenina, simulaneously had low inernaional liquidiy. This combinaion proved inracable, even wih subsanial bailou packages. In conras, India s inernaional liquidiy has improved subsanially afer 1991 as noed above. Afer he crisis, here was a conscious decision o build up reserves o provide greaer insurance agains exernal secor sresses and uncerainies. 33 Acharya (2001, page 17) noes ha he capial accoun surplus reached is peak in 1993/94 (a 3.5 percen of GDP) and has been well below ha level in subsequen years. Neverheless, excep for 1995/96, he capial accoun has been large enough in relaion o he corresponding curren accoun defici in each of he las en years (i.e., 1991/ /01), o ensure accreion o foreign exchange reserves. 28 While no a serious issue a presen, his could poenially encumber RBI reserves and bears monioring. 29 See Zahir (2002) 30 A deailed analysis of he pension liabiliy is beyond he scope of his paper, alhough we flag i as an imporan unfunded liabiliy. See also World Bank (2003a). 31 For he comparison of fundamenals, see IMF (2002), chaper III, page 37, and Ahluwalia (2002a). 32 These facors, ogeher wih he degree of dollarizaion have been sressed as criical facors in deermining defaul risk in Rogoff, Reinhar and Savasano (2003). 33 Acharya (2001), page

13 Reddy (2002, page 1913) noes ha.wo facors are responsible for significan addiion o reserves in he recen pas viz, a far lower level of curren accoun defici (CAD) han he expeced susainable level of abou 2 percen of Gross Domesic produc (GDP) each year and a coninued inflow of capial. Table 4 confirms ha he reserve increase was driven by he fall in he curren accoun defici afer he 1991 crisis; he surplus on he capial accoun is more or less consan excep for a small, apparenly emporary increase during he Eighh Plan period. I divides he pos-crisis reserve accumulaion ino hree phases. The firs is he Eighh Plan period, during which fiscal and curren accoun deficis fell, while privae invesmen rose from a rae of 13 percen of GDP during he second half of he 1980s o over 15 percen. The second is he Ninh Plan period, when fiscal deficis jumped upward, bu he curren accoun defici fell furher and reserve accumulaion acceleraed. Privae invesmen rose slighly. The hird phase is pos-april 2002, when reserves have boomed and a small curren accoun surplus has appeared. Table 4: Sages in Reserve Accumulaion % of GDP 8 h Plan 9 h Plan (period avg.) 85/86-89/90 92/93-96/97 97/98-01/ /03 Fiscal defici Invesmen 1/ public privae Naional savings public privae Curren accoun balance 2/ Capial accoun balance deb-creaing direc and porfolio inv Increase in reserves ($ bn.) Noes: 1/ : Adjused for errors and omissions. Hence, he componens do no sum o oal invesmen. 2/ : There are differences in esimaion of CAB as per he Naional Accouns and he RBI. Thus he numbers do no yield he ideniy CAB = naional savings minus invesmen. Source : Naional Accouns 2002 and various issues, Reserve Bank of India Bulleins The fac ha privae invesmen grew by much less han privae saving during he Ninh Plan relaive o he Eighh has been used o argue ha he rising fiscal defici absorbed par of he increase in privae savings and o his exen crowded ou privae invesmen, hereby helping o keep he curren accoun defici in check. 34 We reurn o his poin laer in he paper. The acceleraion in reserve accumulaion during 2002/03 and afer raises wo quesions: (i) wha is driving i? and (ii) how o manage he macroeconomic consequences. Abou 40 percen of he reserve increase during 2002/03 happened afer November An unknown par of his migh have been driven by capial inflows relaed o fears of he Iraq war, and herefore, his pace of accreion migh no be mainained. 35 Anoher possibiliy is we are seeing capial 34 Ahluwalia (2002), Acharya (2001) and IMF (2002). 35 See RBI (2003b) and Kapur and Pael (2003). Reddy (2002) conains a comprehensive accoun of foreign exchange reserves policy hrough he 1990s. 13

14 reflows from overseas as a resul of he insabiliy in he Middle Eas and money araced by he rupee-dollar ineres differenial given expecaions abou he exchange rae. Whaever he reason, we show nex ha he reserve build-up has improved liquidiy, bu no solvency. The high reserves have no srenghened he join balance shee of he general governmen and RBI because he RBI has been serilizing he capial inflows o keep inflaion low. Le δ denoe he consolidaed liabiliies of he general governmen and RBI, which we call ne public deb, d general governmen deb, f RBI s ne foreign asses and c is ne domesic asses, all expressed as raios o GDP. By definiion, δ is given by: 36 (1) δ = d f c, where denoes ime. Now consider a siuaion where increases in reserves are serilized by RBI sales of governmen securiies, i.e., reducions in is ne domesic asses, c, so ha f = c. 37 I hen follows from (1) ha: (2) δ = d, i.e., he rajecory of ne public deb and general governmen deb should be expeced o be he same. In oher words, while growing foreign exchange reserves boos inernaional liquidiy, hey do no srenghen he combined balance shee of he governmen and RBI. Essenially, domesic deb is being issued o finance he defici and buy reserves. 38 Table 5 brings his ou clearly. We see ha ne public deb increased by 23 percenage poins of GDP beween 1996/97 and 2002/03; from Table 1, general governmen deb increased by 20 percenage poins of GDP over he same period. We also see ha beween 1997/98 and 2001/02, he sum of ne foreign asses and ne domesic asses ( base money ) flucuaed in a small range of abou percen of GDP, bu fell sharply in 2002/03. This is consisen wih a policy of serilizaion o keep inflaion low, wih he fall in base money in 2002/03 aribuable o he appreciaion of he rupee. 39 Table 5 : Ne Public Deb (End March) 90/91 96/97 97/98 98/99 99/00 00/01 01/02 02/03 Crisis General gov deb (Rs. billion) (-) Ne foreign exchange asses of RBI (-) Ne domesic asses of RBI (+) Ne non-moneary liabiliies of RBI Ne public deb Memo : % GDP Ne foreign exchange asses of RBI Ne domesic asses of RBI Ne public deb Source: Reserve Bank of India, Monhly Bulleins 36 The definiion of ne public deb also includes some oher erms on RBI s balance shee, as shown in Table 5, which we ignore here for simpliciy. 37 Sricly speaking, his holds approximaely because ne foreign asses and ne domesic asses are boh expressed as shares of GDP. 38 Anoher way of making his poin is o say ha genuine asses can be creaed only by saving, no borrowing. To quoe Flood and Jeanne (2000) reserves are expensive window dressing by accumulaing hem he governmen behaves as an insolven household ha would ry o impress crediors by financing large cash balances wih credi card deb. 39 Tha is, RBI was exchanging an appreciaing asse (rupee governmen securiies) for a depreciaing one (dollars) wih a lower nominal coupon. For an analysis of serilizaion and is effecs, see Lal, Bery and Pan (2003). 14

15 4. Macro Policy and Growh The policy of accumulaing reserves while keeping inflaion low in he face of rapidly deerioraing public finances suggess ha India has had is own macroeconomic policy rilemma afer 1997/98, seemingly oping o choose he firs wo ou of he following hree objecives: Keep inflaion a single-digi levels 2. Minimize vulnerabiliy o exernal crisis by accumulaing reserves 3. Maximize long-run growh and povery reducion. Objecives 1 and 2 seem o have been pursued a he expense of 3. Given he posiive link beween growh and income povery reducion in India (World Bank 2003), a reconfiguraion of fiscal policies ha promoes growh would speed povery reducion while also improving deb dynamics. We pursue he fiscal policy-growh link by asking wo quesions: will India grow ou of is deb problem, or does i need a fiscal adjusmen o achieve saisfacory growh in he firs place? And wha has been he impac of macro policy on privae invesmen? To anicipae a laer secion, we also noe ha unlike he rilemma in Obsfeld and Taylor (1998), he above is no a rue rilemma, in ha a fiscal adjusmen which reduced deficis and improved he composiion of governmen spending would permi all hree objecives o be aained. 4.1 A Growh Crisis? I was argued above ha nowihsanding he record lows in ineres raes, a fiscal correcion would be needed o place governmen deb on a susainable pah. The Tenh Plan documen sheds more ligh on his opic. Chaper 2 of he plan documen noes ha in order o mee he growh arge of 8 percen per year he invesmen rae will need o rise by 4 percenage poins o a lile over o 28 percen, wih he incremen in domesic savings conribuing 3.5 percenage poins and he rise in he curren accoun defici (CAD) he balance of 0.5 percenage poins. The ICOR is assumed o fall from 4.5 o 3.6, invesmens o grow a 14 percen per annum compared o he long-run growh rae of 6.5 percen, and consumpion by 6.9 percen per annum. The chaper noes ha privae household savings rose over he Eighh and Ninh plans in response o he cu in ax raes and consequen rise in disposable income. This could slow down in he Tenh plan period because of he need o raise he ax/gdp raio, and highlighs he need o increase public savings from -1.7 percen in he base year of he Tenh plan (2001/02) o +2.1 percen in is las year (2006/07). Unless his happens, he plan documen is quie specific ha he growh arge is unlikely o be reached. I also cauions ha he CAD should no be used o slacken he public savings arge, and ha a safe upper limi is 3 percen of GDP. The plan documen hen compares he fiscal deficis projeced under he above savingsinvesmen scenario (he desirable defici from he perspecive of aaining a compound average growh rae of 8 percen during he Tenh Plan period) wih he fiscal defici ha would achieve susainabiliy, defined as a sabilizaion of he governmen deb/gdp raio a he base year (2001/02) level. The comparison is summarized in Table The rilemma in Obsfeld and Taylor (1998) noes ha only wo ou of he following hree goals can be seleced for an open economy: (i) use exchange rae as an insrumen for price sabiliy; (ii) keep capial accoun open for efficiency and flexibiliy; and (iii) pursue an acivis moneary (ineres rae) policy. 15

16 Table 6: Susainable and Desirable Deficis In The 10h Plan Conex Susainable Defici 1/ Desirable Defici 2/ Base year (1) (2) (3) g=6.5 g=8.0 Combined Cener Saes Noes: 1/ Table 2.21 of Tenh Plan 2/ Table 2.22 of Tenh Plan Source: Tenh Five Year Plan, GoI Comparing columns (1) and (3) indicaes he sizable fiscal correcion needed a he level of he saes o achieve susainabiliy even if he assumed growh of 8 percen were o maerialize. Comparing columns (2) and (3) shows ha an even bigger fiscal correcion is needed o generae he needed public savings for invesmen compaible wih 8 percen growh in he firs place. I follows ha a fiscal adjusmen is needed o generae he invesmen needed for growh which will hen lead o deb susainabiliy: India is no going o grow ou of is deb problem wihou a fiscal adjusmen. 4.2 Impac on Privae Invesmen Nowihsanding he boom in privae invesmen afer he 1991 crisis and he pick up in growh during he Eighh Plan period, growh slowed o 5.8 percen over o Overall, growh in he 1990s was similar o ha in he 1980s, he aainmen of a rend rae of growh of abou 6 percen. Wha will push he rend rae of growh o 8 percen? The Tenh Plan documen noes ha more public saving and invesmen and hence a fiscal adjusmen are par of he answer, because of he direc benefis (more infrasrucure) and he spin-offs (banks having more incenive o lend o he privae secor, a less uncerain macroeconomic environmen). Srucural reforms will also help, especially addressing impedimens such as small-scale indusry reservaions and labor regulaions. 42 Similarly, creaing a more conducive environmen for privae invesmen will also be helpful in he more compeiive environmen in which Indian firms find hemselves oday. In his conex, one mus ask if privae invesmen has been crowded ou by he macro policy sance. 43 This is difficul o prove conclusively because he rise in he privae savingsinvesmen gap noed above (Table 4) may also have come abou as he resul of a needed period of consolidaion following he privae invesmen boom during he Eighh Plan period. Bu here is subsanial evidence o sugges ha he macro policy sance has no been he mos conducive o promoing privae invesmen. 41 There is some conroversy abou how much even of his lower growh was arificially boosed by he Fifh Pay Commission (FPC) Award. Acharya (2001) conends ha suiably adjusing real value-added in he caegory Public Adminisraion and Defence would lower growh o 5.4 percen. RBI (2003a) finds a minimal impac of he FPC award. 42 For a deailed discussion of srucural reform issues, see World Bank (2003a). 43 World Bank (2000) and Reynolds (2001) documen a negaive associaion beween fiscal deficis and privae invesmen, alhough esablishing causaliy from high deficis o low privae invesmen is a differen maer. 16