82686 v1 Report No. 82686-IN Stabilization and Fiscal Empowerment: The Twin Challenges Facing India's States Volume I (Executive Summary and Main Report) May 10,2004 Poverty Reduction and Economic Management Unit South Asia Region .--·-· CURRENCY EQUIVALENT Currency Unit Indian Rupee (RS) US$1 = Rs 44.66 FISCAL YEAR April 1 - March 31 ACRONYMS AED Additional Excise Duties MAY Multi-Axle Vehicle APDRP Accelerated Power Development & MODVAT Modified Value-added Tax Reform Programme BoP Balance ofPayment MoU Memorandum of Understanding en Confederation of Indian Industry MTFP Medium Term Fiscal Policy CAG Controller of Audit General MTFRP Medium Term Fiscal Reform Policy CENVAT Central Excise Value-added Tax NGO Non-Government Organization CST Central Sales Tax OECD Organization for Economic Co-operation & Development DC Defined Contribution OED Operations Evaluation Department GST General Sales Tax PE Public Enterprises EGS Education Guarantee Scheme PEM Public Expenditure Management EU European Union PSU Public Sector Undertaking FAO Food and Agriculture Organization RBI Reserve Bank of India FC Finance Commission RD Revenue Deficit FCI Food Corporation oflndia RR Revenue Receipt FRA Fiscal Responsibility Act RTS Representative Tax System FRF Fiscal Reforms Facility SAL Structural Adjustment Loan GDP Gross Domestic Product SEB State Electricity Board Go I Government of India SLPE State Level Public Expenditure GSDP Gross State Domestic Product STU State Transport Undertaking HCV Heavy Commercial Vehicle T&D Transmission & Distribution HIPC Highly Indebted Poor Countries URIF Urban Reforms Incentive Fund HIPS Highly Indebted Poor States VAT Value-added Tax IMFL Indian Made Foreign Liquor VRS Voluntary Retirement Scheme LCV Light Commercial Vehicle WUA Water Users Association Vice President: Praful Patel Country Director: Michael F. Carter Sector Director: Sadiq Ahmed Sector Manager Kapil Kapoor Task Managers: V.J. Ravishankar and Marina Wes Explanatory Note This report is written in three different versions to suit a variety of audiences: o The Executive Summary (in Volume I) is 4 pages long and gives a quick overview. o The Main Report (also in Volume I) is 26 pages long, and conveys the main fmdings. o The Detailed Report (in Volume II) is 90 pages long, and provides the detailed analysis. The executive summary and main report are brief on attributions and sourcing to save space. We have in fact drawn heavily on work by Indian and international experts. Full referencing and attribution can be found in the detailed report. This report was prepared by a Team consisting of Stephen Howes, Rinku Murgai, Smita Kuriakose (expenditure issues), Marina Wes, Farah Zahir, Bala Bhaskar Kalimili (fiscal issues), Ronnie Das-Gupta (revenue issues) and V.J. Ravishankar, Bill McCarten and Anwar Shah (fiscal federalism), Vidya Kamath, Shahnaz Rana, Rita Soni, and Jyoti Sriram (report management). The report has benefited greatly from inputs by the whole India PREM Team, and from inputs by the many sectoral colleagues. Box 1: Definitions and Data )> If not indicated, the state-level data is aggregated for all states. )> Much of the report focuses on 14 major states, which make up 91% of the population oflndia. )> These 14 major states are further divided as poor states (those with a per capita income in 1999/00 ofRs. 10,000 or less at 1993/94 prices) and other states. )> The poor states are Bihar, Madhya Pradesh, Orissa, Uttar Pradesh and Rajasthan while the other states are Maharashtra, Punjab, Haryana, Gujarat, Kerala, Karnataka, Tamil Nadu, West Bengal and Andhra Pradesh. )> The data for the three newly created states (created in 2000/01) of Jharkhand, Chattisgarh and Uttaranchal have been aggregated with that of the states of Bihar, Madhya Pradesh and Uttar Pradesh respectively in order to maintain consistency over time. However, data for Jharkhand is not available for 2000/01 and so is not included in the analysis for 2000/01. )> New- state debt as published by the RBI is included in the old -state debt for 2000/01 (actuals) and 2001/02 (revised estimates). )> India's Gross Domestic Product (GDP) at market prices is used while referring to All States. )> Gross State Domestic Product (GSDP) at factor cost is used for ratios involving the 14 major states, or individual states. )> Revenues receipts and expenditure are expressed net of lottery expenditures; other indicators are as reported by the Reserve Bank of India. )> The latest available actuals are used. This is 2001/02 for all states combined and 2000/01 for individual states (only revised estimates for 2001/02 are available for individual states; since we know at the aggregate level there is a large deviation for 2001/02 between revised estimates and actuals, we do not use the revised estimates). The revised estimates for 2002/03, which are available for all states, are used sparingly, and as indicated. )> For Bihar and Nagaland, 2000/01 figures are revised estimates. )> 1990 and 1990/91 both refer to the Indian fiscal year 1 April 1990-31 March 1991. Debt-stock is measured at the end of the fiscal year. )> The source for tables and figures, unless mentioned in the text is the World Bank State Fiscal Database, which is built up from RBI published state-finance data. Table of Contents Executive Summary ....................................................................... . i I. The State-Level Fiscal Crisis of the Late Nineties: Genesis and Impact .... . 1 II. Expenditure Reforms ............................................................... . 8 III. Revenue Reforms .................................................................. . 15 IV. Strengthening The Fiscal Federal Framework ................................. . 18 V. Fiscal Stabilization and Empowerment at the State Level: Can it be Done? 24 TABLES Table 1 Average real per capita growth rates of India's states ................... . 2 Table 2 Fiscal indicators for poor and other states (2000/01) .................... . 4 Table 3 Starting basic salary of primary school teachers .......................... . 9 FIGURES Figure 1 State deficits and debt levels.................................................... 3 Figure 2 State-level debt/GSDP........................................................ 3 Figure 3 Aggregate expenditure trends for India's states as a percentage of GDP.............................................................................. 4 Figure 4 Non-salary operating recurrent spending per rupee of salary spending (14 major states)............................................................... 5 Figure 5 Improvements in customer satisfaction with various services in Bangalore over a 10-year period . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 13 Figure 6 Trends in States' Own Revenues as a Percentage ofGDP... ... ... ... ... 17 Figure 7 Resources other than own revenue used by states to finance expenditure 19 Figure 8 Per-capita revenues (normalized to Punjab=1), 2000/01.................. 20 Figure 9 State revenues before and after transfers (formal and FCI)............... 21 Figure 10 Trends in Goi revenues and transfers to states.............................. 21 Figure 11 All state debt-to-GDP ratio under three different fiscal targets........... 25 BOXES Box 1 Definitions and data ........................................................... . Box2 International experience with fiscal adjustment............... . . . . . . . . . . . . . 7 Box3 Mid-day meals: an example of productive, additional expenditure...... 15 Box4 International perspective on India's fiscal federalism..................... 18 Box5 Fiscal responsibility legislation at the state-level in India: the story so far. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23 It is ironic that higher and higher deficits over time have not resulted in increasing the government's ability to spend where higher expenditure is required, for example, in the maintenance or expansion ofpublic services. Most ofthe government expenditure is now committed to servicing past debt or meeting salary and other past commitments. We now have a high fiscal deficit without fiscal empowerment. A wholesale change in the government's fiscal policy and making it more responsive to changing requirements are now essential. The process of liberalisation and economic reforms, launched in 1991, and pursued actively in recent years, has yielded positive results, removed some of the structural rigidities, and created potential for higher growth. At the same time, it will be a mistake to be complacent about our recent successes. These gains can disappear very quickly unless a stronger programme is launched in the next few years to further improve our economic decision-making processes, remove scope for political discretion, reduce un-productive expenditure, and improve the quality of governance at all levels. - Bimal Jalan, Former Governor of the Reserve Bank of India, Convocation Address IIM-Ahmedebad, April 3, 2004. EXECUTIVE SUMMARY India, home to more than one billion people, has experienced rapid growth over the past decade, averaging about 6% per year between 1992/93 and 2003/04. Growth has been supported by a wide- ranging reform program- launched after the balance-of-payments crisis in 1991 -to open and deregulate the economy. Progress has also been made in reducing income poverty and other social indicators. The consolidated (center and states) fiscal deficit at 10% is the main challenge for macro policy management. In recent years much has been said about the toll that the aggregate fiscal deficit is taking on sustained growth. With states accounting for nearly half the consolidated fiscal deficit, a fiscal turn- around at the state-level is an essential component of any meaningful reform effort. More than that, with India's states having more expenditure responsibilities than states in any other developing-country federation bar China, fiscal performance at the state level is also an important determinant of state-level development. Already, the increasing divergence in performance among states is widely recognized as a drag on India's development, and on the country's ability to achieve the ambitious targets it has set for itself under the Tenth Plan. A slow secular deterioration in fiscal performance of Indian states over the eighties and nineties was catalyzed into a fiscal crisis in the late 1990s. We use the term "fiscal crisis" not to be alarmist, or to suggest faltering macroeconomic stability, but to convey the sharp deterioration in state-level fiscal performance witnessed in the late nineties - deficits rose, the state-level debt-stock, which had been declining, started rising rapidly, and off-budget liabilities also grew quickly. India's states now appear to be the most leverage sub-national governments in the world. State governments also faced a liquidity crunch, and started to find it much more difficult to pay bills, and even salaries. State governments themselves certainly perceived that they were facing a crisis, as witnessed by the language used in many of the fiscal "White Papers" they produced. The fiscal deterioration at the state level not only called into question India's overall fiscal sustainability, but weakened the developmental and poverty impact of state governments, especially in the poorer states by squeezing productive spending, and reducing their quality. Real expenditure growth in critical sectors such as health and education slowed or ground to a halt, while non-salary inputs were crowded out by growing salary, pension and interest payments. The poorest states were hit the hardest. More positively, the fiscal stress of the late nineties also gave rise to an intense state-level reform effort. Six years on, this report documents the many initiatives undertaken by the states to restore fiscal sustainability, and become more effective agents of development, but also asks what more needs to be done. It is widely agreed that significant challenges remain. States continue to borrow to finance current spending. If reforms falter, or are reversed, fiscal stress would intensify and the development effort would weaken: the quality and quantity of productive expenditures would fall and debt-levels would continue to rise. Clearly, further reforms are needed to consolidate and continue the momentum of what is already underway. The Fiscal Reforms Agenda: Stabilization and Fiscal Empowerment The agenda backed in this report is one that receives widespread support from both the central and state governments in India. Of course, states must reduce deficits to sustainable levels. But simply averting bankruptcy will not be enough. Striving for fiscal empowerment - that is, shifting to a fiscal stance that makes the states more effective agents of development - rather than merely stabilization requires not just aggregate fiscal adjustment, but expenditure restructuring, far-reaching expenditure management reforms, and comprehensive revenue reforms and mobilization. While the ultimate responsibility for fiscal adjustment, and indeed for development, at the state level lies with India's states, this reform project cannot be carried out by the states alone. The Government oflndia (Goi) has also a critical role to play in promoting expenditure and tax reforms, and above all in strengthening further the federal fiscal framework (the "rules of the game") so that states have both the resources to develop and the incentives to perform. This report considers implications of the state-level fiscal empowerment reform agenda for both levels of government. Our findings are summarized below. (The main report also considers implications for external funding agencies- see para. 74.) Implications for State Governments Expenditure reforms must address the salary bill since this is such a large part of state government spending (30%). A moderate reform program, based only on real wage and hiring restraint (rather than pay cuts and forced retrenchment), can deliver significant fiscal savings: more than 2 percentage points of GDP over the coming decade. Pay restraint is justified by the large public-private wage differential (more than 100%, and growing); hiring restraint by the opportunities attrition offers to right-size government, and by the need to reverse the crowding out of non-salary inputs. A bolder program involving retrenchment would free up more fiscal space, but no state has yet shown the political courage to pursue such an approach. Pensions are a rapidly-mounting liability, but states have shown that they can be contained by both immediate (parametric) reforms and longer-term (structural) reforms. Better data on pensions, and a model for forecasting liabilities and simulating reforms are also needed. Subsidies have proved difficult to cut, largely for reasons of political economy. The power subsidy is the largest state-government subsidy, and provides a good example of the difficulties involved in reforming subsidy regimes: reform of the power sector in agriculture has been a case-study of repeated failure. There are no easy or guaranteed solutions - ifthere were, they would surely have been tried -but to make further progress in power sector reforms, tackling the lack of commercial discipline in the power sector has to be the top priority. Without promoting commercial discipline, by distancing service providers from government, through mechanisms such as corporatization and privatization, no other remedy will work. The same prescriptions, and implementation difficulties, will apply to reforms of subsidies other than power. Indeed, one ofthe most important lessons from the power sector is that subsidies are here to stay. Just as hiring and real wage restraint is a more feasible strategy for controlling salaries than wage cuts or downsizing so too, in many cases, maintaining subsidies at their current levels will be a major success. State governments should focus more on improving subsidy management and targeting. This, together with improving commercial discipline in the subsidy-receiving sectors, will have major efficiency gains that will themselves result in substantial subsidy savings. Public enterprise reforms - closure and privatisation - will have a limited immediate fiscal impact, but will stop the need for fiscal transfusions to keep loss-makers afloat, and prevent the build-up of future liabilities. Interest savings will largely follow from reduced borrowing, but states can also, with central government facilitation, take advantage of a low interest rate regime by aggressive debt-restructuring. The quality of spending must be improved, and it can be, as numerous success stories from the states show. These cases need to be replicated and scaled-up: o Agency-specific reforms including an increased role for the private sector can improve service delivery. o The broader enabling environment is key to improving service delivery and the quality of spending; reforms are required to: promote effective oversight; improve commercial discipline and cost-recovery where possible; encourage civil society to monitor government performance; promote transparency, and punish corruption. 11 o Improving the financial environment within which the state governments operate will be critical to improve the effectiveness of government spending; and o Improving the planning process to focus on prioritisation of limited resources to maximise outcomes rather than merely to maximise the aggregate size of plan expenditure. Spending in priority areas needs to be increased as the quality of spending is improved. Though the priority areas will vary from state to state, most states regard basic infrastructure, maintenance, and social sector spending as key. The states contribute more than half of India's general government capital spending, which is at about 3% of GDP, against the 4-5% target regarded as necessary to support sustained rapid growth. Maintenance spending on state highways is at about 40% of what it should be. Social sector spending, overwhelmingly financed by the states, is also low relative to need: spending on non-salary inputs is abysmal, a fact which is reflected in deteriorating facilities, and is one determinant of low-quality service provision. State revenue performance has been better than performance at the centre over the past ten years. Nevertheless, there is scope to boost revenues further through revenue policy and administration reforms to mobilize additional resources for India's pressing development needs. As a guiding principle, the revenue productivity of Indian states can be increased by broadening the tax base coupled with reforms to ensure few low rates and limited exemptions. This will also help to improve compliance. The sales tax is the most important state-level tax. A key revenue reform would be to replace the sales tax and other minor commercial levies by a consumption-type VAT on goods. The move to an integrated VAT on goods (and, if possible, services) is likely to have appreciable growth benefits, provided tax administration is also streamlined and tax payer compliance improves. Encouraging progress in sales-tax reform - which led to a large boost in revenue - has been stalled by an impasse over VAT introduction. This impasse might be broken by states moving individually towards VAT, without compensation, but allowing them to choose their rates (subject only to a floor) rather than insisting on uniform rates across states since this gives rise to demands for compensation. Stamp duty collections could be improved by lowering the very high rates on property sales, and improving compliance by re-engineering. There is also scope to increase non-tax revenues, including user charges and mineral royalties. Tax administration reforms and improved inter-state coordination are probably even more important than tax policy reforms. Of all pending reforms, improved enforcement technology and procedures coupled with staff incentives, management flexibility and effective anticorruption institutions have the greatest potential to lead to a significant and sustained increase in state revenue. Implications for the Central Government. Central leadership will be crucial to the sub-national· fiscal reform effort. On the expenditure side, the establishment of a new central pay commission could derail the reform effort and undo much of the progress that has been made so far. There are a number of critical state-level tax reforms which require central leadership. Allowing the states to move forward individually with VAT introduction but without compensation is perhaps the most important. This is closely followed by phasing out of the distorting and inequitable CST, which will likely require compensation. Transfer to states of the right to tax services is a very positive initiative; quadrupling the constitutional limit on the professions tax (or replacing the nominal limit by a formula based one) would lay the groundwork for the states to raise almost 1% of GDP in tax. A shift to a rules- based setting for royalty rates on major minerals would help the states in relation to this important source of non-tax revenue. Most importantly, Gol's own improved tax performance will also help reverse the decline in the central transfers reaching the states. 111 India's federal fiscal system shows many strengths, prime among them stability. However, the fiscal crisis has also brought to light a number of weaknesses, which undermine incentives for performance at the state level, and have hurt the poorer states. Over the nineties, states have received fewer grants and more loans; this trend needs to be reversed. There is unanimous agreement on the need to harden budget constraints at the state level, something well within Gol's powers to act on. Progress has already been made in this direction through preventing the build-up of arrears by state power utilities. A major next step would be to give states more flexibility to borrow, especially from the market, but under a centrally-imposed global borrowing cap, and facing interest rates which reflect creditworthiness. States which have difficulty accessing the market can be given a central government guarantee - at a price. Go I is better placed to act as a regulator of state borrowings, enforcing the global cap, rather than as a creditor itself. Grants and loans should be de-linked so that each can be given the very separate treatment it requires. While grants can be distributed on the basis of need and performance, loans need to be provided on the basis of creditworthiness. The state planning process, as currently constituted, puts fiscal policy in a permanently expansionary mode and requires rethinking. Debt-restructuring on commercial principles is in line with allowing states more flexibility under a global cap. Performance-based debt-relief will also help poor, highly-indebted states, but has the risk of, however well-designed, undermining hard budget constraints. More progressive grants, and ones which reward fiscal reformers may be a better alternative. The federal grant regime is complex, and views on how it should be reformed divergent. The current situation in which the poorest states, like Bihar and UP, receive fewer grants than middle-income and faster-growing states such as AP and Karnataka, cannot be optimal. Grants need to be made more progressive and at the same time contain stronger incentives for performance. Various options for tackling both these goals can be considered. Finally, the central government can back the states' own reform efforts. Five states have already passed fiscal responsibility acts which provide time-bound deficit reduction targets. Goi can encourage other states to follow suit, and can base their own assessments of the states' reform efforts in terms of compliance with the state's own legally-mandated targets. Are fiscal stabilization and empowerment achievable? Scenario analysis suggest that a joint program of state and central reforms, as per the outlines of the preceding summary, will enable states to stabilize their debt and achieve a reasonable fiscal deficit target (3% of GDP). However, they cannot eliminate the revenue deficit until several years after the Goi target date of 2007/08, unless they sacrifice productive areas of expenditure. This suggests that debt stabilization and fiscal deficit control should be the first priority of state-level fiscal reform: too tight a revenue balance target will achieve stabilization but at the cost of empowerment. The poor states can achieve only debt-stabilization, and neither fiscal nor revenue deficit targets by 2007/08, without compression of productive expenditures. They will require additional central support, through debt-restructuring and additional central transfers, if they are to achieve satisfactory fiscal adjustment and at the same time avoid compression of productive expenditures. In summary, the ambitious national agenda of fiscal adjustment and empowerment for state governments is feasible but will require, for all states, appropriate choice of fiscal targets, and for the poor states additional central government support. Above all, what is needed is a joint central-state government reform program aimed not just at fiscal adjustment but at strengthening the development effectiveness of India's states. Anything less will jeopardize India's prospects for sustained rapid growth and poverty redm:tion. lV I. THE STATE-LEVEL FISCAL CRISIS OF THE LATE NINETIES: EVOLUTION AND IMPACT This report is written to help share the lessons and success-stories of state fiscal reform to date, and to assist states and the central government in implementing the national agenda of state-level fiscal stabilization and empowerment. 1 Following two decades of high growth, and a decade of liberalization, there is growing confidence within India, as well as internationally, about the state of the economy and India's potential to become a developed country. A recent report by Goldman Sachs suggested that India could become one of the world's three largest economies in less than 30 years, and that income per capita in 2050 could be 35 times current levels. There is a sense that India stands at the threshold, but also that crossing that threshold will require a quantum improvement in government effectiveness, particularly at the state level, given the extensive responsibilities of India's state governments for infrastructure and human development. 2 Especially since the late nineties, when states experienced a sharp fiscal deterioration, states have been undertaking reforms aimed not only at fiscal stabilization through the reduction of deficits - though this is of course necessary - but also at fiscal empowerment: more and better spending in priority areas, backed up by an enhanced revenue mobilization effort. This stock-taking report aims both to share the lessons of reform to date, and to anticipate what more needs to be done. States in India play an increasingly important role in devising and implementing policies to stimulate growth and promote human development. 3 India, home to one billion people, is a federation of28 States and 7 Union Territories. The largest (Uttar Pradesh) is populous enough to be the fifth largest country in the world; the smallest (Sikkim) has a population of less than half a million. Among the 14 major states, the richest has a per-capita income that is five times that of the poorest state. Infant mortality ranges from a low of 14 per 1000 in Kerala, on par with east European countries, to 96 for Orissa, more typical of sub-Saharan African nations. In an era of liberalization, a state's own policies and quality of government is increasingly critical for a state's development. 4 Under the Indian constitution, state governments are assigned significant responsibilities in sectors such as agriculture, industry, infrastructure, education, health, and social welfare. India's state governments are key financiers of a number of areas critical for enhancing growth and reducing poverty: in 2000/01, 57% of India's total government capital expenditure was financed by the states, as was 97% of irrigation maintenance, 39% of road maintenance, 90% of public health expenditures, and 86% of public education expenditures. In fact, India's states are responsible for a higher proportion of general government spending than in any other developing country, except China. Moreover, the states' increasingly large fiscal deficits mean their fiscal policy is not only an important determinant of their own developmental performance but also oflndia's overall macroeconomic and fiscal sustainability. 5 One of the striking features of Indian States prior to the 1990s was the relative uniformity of policies across states. The environment within which states operate changed significantly after 1991 with the central government's liberalization of the trade and investment regime, and the growth of regional political parties. These developments allowed the states a larger role in determining their development paths and attracting investment. 1 The performance of India's states is increasingly divergent. 6 Balanced regional growth has always been an objective of India's development strategy since independence; and is supported by a system of redistributive transfers to the states. Of the 14 major states, the 5 poorest have a per capita income ofRs 10,000 (in 1993/94 constant prices) or less: Bihar, Madhya Pradesh, Orissa, Rajasthan and UP. These are classified throughout this report as the poor or low-income states. (See Box 1 at the front of the report for a brief Table 1: Average real per capita growth rates of discussion of definitions and data used in the report.) India's states These states are home to over 40% of India's 1980/81-1989/90 1990/91-1999/00 population, and nearly 50% of India's poor. These states are also slower-growing than the other states Poor states 2.4 2.5 (Table 1). In particular, following the liberalization of Other states 3.1 4.8 1991, most of the middle and high income states were able to take greater advantage of the new conditions, because of better initial conditions, governance, infrastructure and human resources than the poorer states. Some of the poorer states failed to improve their state-level policies to offset their initial disadvantage in attracting new investment. As a result, the average per capita income in the poor states relative to the others has decreased from 71% in 1980/81 to 54% in 1999/2000. 7 Although the share of population below the poverty line declined and human development indicators improved in all states, progress on these broader indicators was also in some cases faster in the fast-growing and better-off states. In some cases human development indicators actually worsened in the poor states. For instance, the number of underweight children increased over the nineties for all poor states, whereas it declined for most of the other states. Unless the poor states improve their performance, it will become increasingly difficult to accelerate poverty reduction and development in India. A slow secular deterioration in fiscal performance over the eighties and nineties was catalyzed into a state-level fiscal crisis by the Fifth Central Pay Commission pay awards in the late 1990s. 8 We use the term "fiscal crisis" not to be alarmist, or to suggest faltering macroeconomic performance but to convey the sharp deterioration in state-level fiscal performance witnessed in the late nineties - deficits rose, the state-level debt-stock, which had been declining, started rising rapidly, and off-budget liabilities also grew quickly. There was also a liquidity crunch, not on the economy, but on state--governments themselves, who started to find it much more difficult to pay bills, and even salaries. State governments themselves certainly perceived that they were facing a crisis. Many wrote White Papers, such as the Government of Tamil Nadu did in August 2001, to apprise legislators and the public "of the extent and causes of the serious financial crisis confronting the state". Finally, there is a close parallel with the balance of payments crisis of the early nineties, not in the nature of the crisis, but in its impact: just as the BoP crisis of 1991 gave rise to a decade of central government reforms, so the state- level fiscal crisis gave enormous impetus to reforms at the state level. 9 While the large pay awards were the immediate cause for the sharp fiscal deterioration, the secullar worsening in the revenue (current) balance of the state governments can be traced back as far as the past two decades, and was related to the growth of populist policies, symbolized by rapid growth in publiic employment and the introduction in many states of free power to farmers in the eighties. The growing revenue deficit was prevented from translating into a higher fiscal deficit until the second half of the nineties only because capital expenditures were compressed. On the expenditure side, the interest burden grew over the nineties initially due to a hardening of interest rates as the old regime of financial repression and subsidized rates for government borrowing was brought to an end in the nineties, and subs1~quently due to a rising debt-stock. By contrast, revenues fell over the nineties. The state-level 2 revenue/GDP ratio fell from around 12% in the late eighties and early nineties to 10.1% in the worst year of the fiscal crisis, 1998/99. Both own-revenues and Go I transfers fell, though the latter more so, as its share in total revenue declined from slightly above 40% up to 1993/94 to 36% in the late nineties. Deficits and debt levels rose sharply in the late nineties. 10 The sharp increase in expenditures in the latter half of the nineties alongside declining revenues could only be supported by much higher borrowing by state governments. The aggregate fiscal deficit of the states jumped from 2.9% of GDP in 1997/98 to 4.3% in Figure 1: State deficits and debt levels 1998/99, and has remained above 30 5.0 4% of GDP since. The state-level 4.5 25 debt stock which was actually 4.0 3.5 falling in the first half of the 20 nineties reversed its downward 5 3.0 ~ ~ 15 2.5 ~ ;, course in 1997/98 and has been "' 10 2.0 rising ever since. (Figure l) 1.5 1.0 Despite central government 5 0.5 control over state government 0.0 borrowings being enshrined in the qS5~:£~ m~ ~ ~ ~ constitution, state governments !illiill!!lllDebt/GDP - F i s c a l deficit/G D P were able to increase their borrowing largely by drawing on sources over which Gol exercises no active control: there was a boom in "small savings" loans, a relatively expensive form of debt with high administratively-determined interest rates; and state governments borrowed from their own employees by impounding promised salary increases into their provident funds. Off-budget liabilities also increased rapidly. 11 The most significant off-budget liability is in the power sector. Over the Figure 2: State- level debt I GSDP course of the nineties, cash-strapped governments were unable to meet their 40% rapidly rising subsidy obligations to their electricity companies. Thus an 35% increasing proportion of power sector Poor losses, which reached over 1.4% of GDP 30% by 2000/01, were met off budget, in large part through arrears to central 25% utilities. The rise in the state-level deficits is even sharper (by more than 20% one percentage point) once this is taken account of. State-government guarantees have also risen sharply since 1996, and their quality has declined because an increasing proportion of government guarantees go to back borrowing by government-owned special purpose vehicles whose debt servicing will revert entirely to the budget. Such liabilities, albeit off budget, are thus in the nature of actual rather than contingent state government liabilities. 3 The poorer states saw the most severe fiscal deterioration. 12 The degree of fiscal deterioration has been worse in the poorer states (Figure 2, Table 2). They are more reliant on central transfers and so have suffered more from the reduction in these transfers; they have also experienced greater variability in revenues. On the expenditure side, with higher initial debt stocks and salary bills (relative to total spending), they also suffered more from the shocks of the nineties: spending increased by twice as much in the poor states as it did in other states. Fiscal indicators deteriorated more in poorer than in richer states, and have reached alarming levels. Interest payments as a share of own revenues are nearly twice as high in poor states than in other states, and are still increasing. In Bihar and Orissa, debt service preempts more than 90% of own revenues. Table 2: Fiscal indicators for poor and other states (1999/00) Own Fiscal Go I revenues Deficit/ RD I Interest payments Transfers/GSDP /GSDP GSDP GSDP I own revenues Poor states 6.1 7.0 6.3 4.1 46.5 Other states 2.9 8.8 5.5 3.3 29.4 Although associated with an increase in public spending, the fiscal cns1s weakened the developmental and poverty impact of state governments, especially in the poorer states; it also called into question India's overall fiscal sustainability. 13 Various studies have showed the importance of both the quantity and quality of state government expenditures for poverty reduction (See Box 1.1 of Volume II). Although the fiscal stance was expansionary in the late nineties, the impact on expenditure quantity and quality was largely negative. Once one adjusts for the large increases in interest and pension payments, and for the large real salary increase in the second half of the nineties, aggregate expenditure has continued the downward trend observed through the first half of the nineties (see the lower, declining line in Figure 3). The only positive fiscal development in the post- 1996/97 period is an increase in real growth in state-level capital expenditure. In other respects, the c 13.0 t"'"'ir~~~=--~~~~~~~-~~~-~~~~~~ fiscal crisis weakened the i:: ~ developmental and poverty impact ~ 12.0 P:-~--'--'~i,~~~~-~~--'~~~~~--'~~~- of state governments by, for example, slowing down real growth of e:xpenditure in education, and halting real growth in health co expenditure. The poorer states in ql "' "' .... "' .;, CD "' ,.!. "' "' 0 0 9 N 9 particular suffered. They show lower "' N "' "' "' cb "' "' "' 0 0 real expenditure growth in all the -+-Total expenditure __.....Potentially productive expenditure, adjusted for real salary increases expenditure categories than the middle-income and rich states; less than a t e growt education, health, and irrigation maintenance. 4 14 Quality of spending also suffered, as expenditures became much more salary-intensive. By 2000/01, the ratio of non-salary operating recurrent expenditure to salary expenditure was 1.1 for the better-off states, marginally better than at the start of the nineties, but for the poor states it had fallen precipitously to 0. 73, a roughly 20% fall from the early nineties (Figure 4). Given that there was virtually no hiring in the second half of the nineties, it is ironic that all states, but especially the poorer ones, should, at the end of the nineties, have at their disposal significantly fewer non-salary resources in relation to their salary bill than they had midway through the Figure 4: Non salary operating recurrent spending per rupee of salary nineties. Accounts of teachers 1.4 without rooms to teach in, and 1.3 doctors lacking drugs to 1.2 dispense have not surprisingly become very common. 1.1 . ... 15 The deterioration m .. : c: 1.0 0.9 state finances also significantly weakened India's 0.8 overall macroeconomic 0.7 0.73 performance, which, with the 0.6 exception of the burgeoning ...... (J) S! fiscal deficit was strong in the nineties. Almost half of the ~ 8i ...... - Low Income States .....t.- Others States consolidated fiscal deficit is now made up of state-level deficits. The sharp fiscal deterioration gave rise to an intense state-level reform effort. 16 As mentioned, the sharp deterioration in fiscal performance generated a sense that "business-as- usual" was not an option, and gave enormous momentum to reform efforts at the state level. Many states have now adopted reform programs, albeit to differing degrees of strength and credibility, to return deficits to sustainable levels, and more broadly to improve government efficiency and effectiveness. 17 Many states have adopted medium-term fiscal plans or frameworks. Five states have provides statutory backing to these plans through the passage of fiscal responsibility legislation, which mandates the achievement of prudent fiscal targets within a fixed period. On the expenditure side, a number of states have contained spending by restricting recruitment, increasing wages at less than the rate of inflation, hiring new employees on a contract basis at much lower than regular rates, and curbing growth in administrative expenditures. Some states have cut the cost of their existing pension scheme, and are preparing for a shift to a new, cheaper contributory pension scheme. Power sector reforms, and other measures to cut subsidies, have also assumed critical importance in recent years. The closure and privatization of public sector enterprises has been accelerated. On the revenue side, reform measures have broadly aimed at enhancing revenue receipts through revision of tax rates, broadening the tax base, and improving tax compliance. Other important initiatives relate to the preparatory work for the introduction of a VAT, and rationalization of user charges mainly relating to power, water and transport. 18 The Government of India also moved swiftly to help states undertake fiscal and sectoral reforms. The fiscal deterioration also gave rise to the birth of MoUs between Goi and some state governments in 1999/00, outlining medium-term strategies towards fiscal consolidation. This practice developed further into the Fiscal Reforms Facility starting in 2000/01, following the recommendations of the Eleventh Finance Commission to link a portion of untied central grants to the fiscal correction performance of 5 individual states. A number of other reform facilities have since sprung up, and external funding agencies have also tied a portion of fund transfers to fiscal and sectoral reforms. Recent years have shown some signs of improved fiscal performance. 19 The intensified revenue effort appears to be paying off. States' own-revenues have bounced back to pre-crisis levels as a percentage of GDP, though central government transfers are yet to follow suit. Economic growth of 6%-plus, and recovery from the industrial recession of the late-nineties, has also helped to boost revenues. The shock of the wage and pension increases associated with the Fifth Pay Commission is settling, and states are benefiting fiscally from tight hiring restraint. States are also starting to benefit from lower interest rates - not just on fresh borrowing, but also through debt-swaps. Committed spending as a share of total revenues has begun to fall. The primary deficit of the states has been on a declining path since 2000, though the debt stock continues to rise. Liquidity has also improved, and states are less likely to turn to the Reserve Bank for overdrafts. But concerns about the level and composition of the fiscal deficit remain. States continue to borrow to finance current spending. Poor states show far fewer signs of recovery than richer states. 20 For all this progress, it has to be noted that the combined fiscal deficit of the states has stayed between 4-5% of GDP, and the revenue deficit between 2.5-3%. This implies, among other things, that more than half of the states' borrowing is to finance current expenditures, and that the nation will fall far short of the Fiscal Reforms Facility goal of eliminating the states' combined revenue deficit by 2004-05. 1 While some states have shown significant deficit reduction results, few have been able to sustain an improvement, and several have shown better results only by running up arrears. Many fiscal targets agreed to by the states have been missed, even by those adopting fiscal responsibility acts. The debt and debt service burden of the states continue to increase and preempts a large share of available resources, especially in the poor states. The primary deficits of poor and rich states have converged, but the former continue to show higher revenue and fiscal deficits, as well as very high levels of salary-intensity in their expenditure. Off-budget liabilities of states also continue to be a serious concern. Large volumes of off- budget debt services are now becoming due and some states have already shown their inability to service this debt. A halt in reforms would be bad for all states as the quality and quantity of productive expenditures woulld fall, and debt levels would steadily build. 21 Whatever improvements have been made to date would be lost without further reforms. We consider a scenario in which the combined states' fiscal deficit stabilizes roughly at its current level of 4.8%, due to revenue/GDP being flat, a resumption in salary bill growth, perhaps because of a new pay commission, and a failure to reduce subsidies. In such a scenario, debt levels continue to increase, and the composition and quality of spending deteriorates. Committed spending (debt service, wages, pensions and subsiidies) increases from 89% of total revenues in 2003/04 (10.5% ofGDP) to 106% oftotal revenues in 2007/08 (12.5% of GDP). How rapidly debt would accumulate would depend on how permissive the central government is. But under a "no reform" scenario, the choice would be between more borrowing and Jless non-salary productive spending, since less borrowing would simply result in a crowding out of capital expenditure and O&M expenditure. 1 It should also be admitted that there is a dearth of up-to-date data on state fiscal performance, with the most recent actuals available for only 2001/02 (see Box 1 at the start of the report). We base this analysis in part on more recent revis1ed and budget estimates, and in part on our knowledge of the situation in several states. The need for better and more timely state data is stressed at the end of the report (see para. 75). 6 There is a broad consensus on the fiscal challenges facing India's states. 22 As already mentioned, states must of course reduce deficits to sustainable levels. But simply averting bankruptcy will not be enough. In addition to fiscal stabilization, what is sometimes referred to in India as "fiscal empowerment" also needs to be pursued, that is, the states need to shift to a fiscal stance that makes them more effective agents of development. To become effective agents of developments, states must also increase the quantity of spending in priority areas, and improve the quality of government expenditure. 23 This positive agenda for fiscal reform is in line with international experience with fiscal adjustment. As Box 2 summarizes, the empirical literature finds that more successful fiscal adjustments rest at least in part on restructuring of recurrent expenditures. They find that fiscal consolidations achieved through cutting selected current expenditures tend to be more successful and trigger higher growth rates than adjustments based solely on revenue increases and cuts in more productive spending. According to this analysis, protecting capital expenditures during a fiscal adjustment leads to higher growth, as does an increase in the share of current spending on non-wage goods and services. Reductions in the public sector wage bill are not harmful for growth for the sample as a whole. Reallocating government expenditures to more productive uses is also correlated with more persistent fiscal consolidation episodes. Box 2: International experience with fiscal adjustment Over the past decade a number of studies of fiscal adjustment have been undertaken. Following Alesina and Perotti (1995), fiscal adjustment strategies are sometimes broadly divided into two categories: 'Type 1', primarily relying on cuts in recurrent spending, and 'Type 2 ', relying primarily on tax increases with spending cuts mostly limited to public investment. In a study of 20 OECD countries for the period 1960-1994, 60 episodes of fiscal consolidation were identified. Of these episodes, only 16 were lasting, and among these successful cases 73% were based at least in part on recurrent spending cuts. Although most fiscal adjustment efforts rely on tax increases to lower the deficit and the debt burden, those successful in addressing fiscal imbalances rely more heavily on cuts in current expenditures than tax increases. McDermott and Wescott (1996) find similar results for 74 episodes of fiscal adjustment in 20 countries during 1970- 995. Whereas a little less than half of the 'Type 1' adjustment cases were successful, only 1 out of the 6 'Type 2' adjustment cases was successful. They also found that 'Type 1' adjustments are more likely to reduce the public debt ratio. Complementary to this, research by Alesina and Ardagna (1998) finds that the success of fiscal stabilization is not just determined by the size of the adjustment, but also by the composition. Adjustments that focus on cuts in transfers and wages are more likely to succeed in reducing the primary structural balance. Finally, Fiscal adjustments that rely primarily on cuts in transfers and the wage bill not only last longer, but can also be expansionary; on the other hand adjustments relying on tax increases and cuts in public investment tend to be contractionary and unsustainable (von Hagen, Hallett and Strauch, 2001). This work is admittedly based largely on OECD experience, though there is one study which confirms the findings for low-income countries (Gupta, Clements, Baldacci, Mulas-Granados, 2002). 24 Addressing these challenges requires much of the states: expenditure restructuring to generate fiscal savings on the expenditure side, far-reaching expenditure management reforms, and comprehensive revenue reforms and mobilization. This daunting agenda, which is explored in Sections 2 (on expenditure) and 3 (on revenue), cannot be carried out by the states alone. It also requires Go I actions to strengthen the framework for reforms and development. These issues relating to fiscal federalism (the "rules of the game") are explored in Section IV. Section V. looks at whether simultaneous pursuit of both fiscal stabilization and empowerment is possible. 7 25 The precise sequencing and appropriate packaging of reforms will vary from state to state. While political economy considerations lean towards emphasizing revenue enhancing reforms in the immediate term, given the resistance to expenditure restructuring measures, there is economic merit in following the exactly opposite sequence -- that is, to first improve the composition and quality of public spending before raising more resources from the public. But this can take time, and improving quality may also require additional resources (e.g. for more non-salary inputs, since salary costs are fixed). While it is difficult to draw generic lessons on how reforms should be sequenced and packaged, there is already a lot of experience at the state level on different reforms, what works and what doesn't. One aim of this report is to help share these lessons more broadly. In this spirit, the report draws on the interactions over many years now of Bank staff with various state governments, whether through collaboration over lending, analytical work, or simply information-sharing. Of course, to keep it manageable we have to focus on fiscal reforms; exclusion or inadequate treatment of sectoral or investment climate reforms should not be taken to indicate lack of importance. II. EXPENDITURE REFORMS 26 Given the low levels and the worrying recent trends in both quantity and quality of expenditure, there is an urgent need for expenditure restructuring to free up fiscal resources, for reforms to improve the quality of spending, and for increased spending in priority areas. We consider these in turn, starting with expenditure restructuring, examining where savings can be found in relation to the salary and pension bill, subsidies, public enterprises, and interest payments. Salaries are such a large part of government spending that they must be at the core of any expenditure reform program. A feasible reform program, based on wage and hiring restraint, can deliver significant fiscal savings: more than 2 percentage points of GDP over the coming decade. 27 Salaries make up 30% of state government spending. States can make significant savings in the salary bill in the coming years by ensuring that a policy package of real wage and hiring restraint is in place. A policy of public-sector wage restraint is justified by the fact that most public-sector employees in India are greatly overpaid relative to their private sector counterparts - a long-standing trend and one exacerbated by the Fifth Central Pay Commission. India's public-private wage differentials are in fact among the highest in the world, and now stand at over 100%. Studies for a large number of countries using similar methodologies find similarly large differentials to those observed in India only in two African countries (Ghana and Cote d'Ivoire) and in some regions of Brazil. Of course, this hides the fact that wage compression implies under-payment for senior civil servants relative to market, but this is the exception rather than the rule. About 40% of state civil servants are teachers. In rural areas, where most of them live, they belong to the top decile of the income scale. It is difficult to justify further real wage increases for them given the pressing needs to improve school infrastructure, and bring down pupil- teacher ratios in many states. 28 Critical for maintaining a policy of wage restraint will be avoidance of establishment of a pay commission for as long as possible, since such a commission would in all likelihood lead again to a significant increase in real wages. Pressure for a real wage increase is mounting, but giving into this will sacrifice many of the hard-won gains of the last few years. If and when base salaries are adjusted, it will be important to put more emphasis on local market comparators in determining salary levels: this would limit the scope for real public-sector wage increases to areas in which it is needed. Since the last pay commission shows the influence of the central government on pay-related matters, the Government of India has a special obligation of leadership in this area. 8 29 However, this should not be interpreted to mean that states are passive implementers of a central pay policy. Far from it. Since the pay commission increases, several states have refused to pass on full cost-of-living allowances. And, contrary to popular belief, salaries across states are far from uniform. Table 3 shows the starting, base salary for teachers in a number of states. Salaries of teachers in some states are 25% less than in other states Table 3: Starting basic salary of a primary school teacher - these states save up to 10% on their State Basic Salary( Rs/month) total salary bill as a result. Some states 1995 2003 have also saved large amounts by Andhra Pradesh 1,010 3,750 hiring new teachers on a contract- Karnataka 1,040 3,300 basis, and paying them much less than Orissa l,080 3,600 Punjab 1,200 4,550 a newly-hired regular teacher gets: Tamil Nadu 1,200 4,500 perhaps half or sometimes even one- Uttar Pradesh 1,100 4,500 fifth. The "para-teacher" phenomenon Note: These are less than half of average, all-in salaries; they is most famous and extensive in exclude all allowances. Madhya Pradesh but has now spread - - - - - - - - - - - - - - - - - - - - - - - to most states. Many para-teachers have exactly the same qualifications and responsibilities as regular teachers. Evaluations suggest that these fiscal savings do not result in any loss of quality. Para-teachers seem to deliver a quality of service that is not necessarily high, but not any lower than that provided by regular teachers. Thus the para-teacher phenomenon appears to be a rational response by state governments to the excessive premium attached to public-sector wages. Governments should recognize that para-teachers might naturally graduate to become regular teachers, or that the distinction between the two might become blurred, but should also consider further extending the para-teacher principles to all new hires: lowering entry wages, and making employment contract-based, at least for an initial period of time. (See Box 2.2 of Volume II for a discussion of para-teachers in India). 30 Hiring restraint is justified, even though India's civil service is small by international standards (and, therefore, likely to grow in the long term), because there are large areas of over-staffing as well as under-staffing. Even in areas where more hiring is required, such has been the crowding out of non-salary by salary spending, it is unclear that the marginal rupee should be spent on salaries, especially given low civil service productivity, as evidenced, for example, by high levels of absence of service delivery providers. Targeted retrenchment programs have not been a success in India (civil servants are reluctant to leave; and state governments reluctant to push them to do so), but much can be achieved through attrition. Several states show attrition rates in excess of 3%, which, under a zero net hiring policy, allow significant scope for expansion through new hires in priority areas such as health and education. 31 Hiring restraint has been in evidence since the late nineties. Many states imposed hiring bans since the onset of the fiscal crisis, and since 1997, state employee strength shows an overall, marginal reduction of 0.7%. Although the salary bill is now starting to fall again- relative to GDP- due to this strong hiring restraint, state governments continue to pay the price of the Fifth Pay Commission: the salary bill is about 1 percentage point of GDP higher than it would have been without the pay-rises, and, as we have seen, the salary-intensity of expenditure is higher than it was in the mid-nineties, despite no net hiring. But if the policy of no net hiring can be sustained, and if salary increases can be restricted to the rate of inflation, then, under reasonable assumptions, the salary bill could fall by as much as 2 percentage points of GDP over the coming decade, freeing up valuable resources for deficit reduction and non-salary expenditures. 32 The emphasis we give to salaries is not because we are anti-employee, but because of the importance of salaries to state government. The reforms we endorse here are mild ones. They envisage neither real wage cuts nor for forced retrenchment, both common responses in most countries to fiscal deterioration. They are the minimum required for India's state governments to recover their fiscal health, and improve expenditure productivity. 9 Pensiions are a rapidly-mounting liability, but states have shown that they can be contained by both immediate (parametric) reforms and longer-term (structural) reforms. Better data on pensions, and a model for forecasting liabilities and simulating reforms are needed. 33 Pensions are another major expenditure item (about 7.5% of total expenditure) and source of fiscal vulnerability. The annual average increase in pension spending was 30% between 1995/96 and 2000/01 making pensions the fastest growing expenditure item in state budgets. Two types of reforms can be considered: parametric reforms to contain the cost of the current pay-as-you-go system; and structural reforms to enable shifting to a cheaper and less risky defined-contribution scheme. 34 The most important parametric reform is to index pensions only to prices and not to real wages. This implies that any pay commission, if constituted, should confine itself to real adjustments in the wages of civil servants currently employed. States have shown that additional savings can be found in the short-term by reforms to the parameters governing retirement benefits. This can be done by, for example, by reducing the amount of pension that can be commuted at retirement, by using less-biased discount rates for these commutations, by calculating the base pension not on the last month's salary but on the last year's or last 3 years, or by reducing the amount of accrued leave that can be encashed at retirement. 35 In the longer term, a switchover to the proposed Gol defined contribution scheme will also generate savings. Rough calculations suggest that the net present value of savings to government of each new employee who switches to the new scheme is more than 30% of the net present cost of putting the new employee under the existing defined benefit scheme. States have been invited to join this scheme, and several are planning to. To maximize and bring forward the fiscal costs of such a shift, it should be extended not only to new recruits, but to employees who have not yet been in service for long enough to qualify for the existing pension scheme (in many states, employees who have had less than 10 years of service). 36 Most states are completely in the dark as to the future course of their pension expenditures, since they have no data on the demographic profile of either pensioners or civil-servants. Some states, notably Karnataka, AP and Tamil Nadu, have developed pension projection models, and Gol is now encouraging other states to follow suit. Results from these models show that with parametric reforms, the growth in pensilons can be contained as a percentage of GDP to its current levels. There is also evidence of considerable abuse of the pension system (for example, pensions being drawn after death): better data on pensiloners would enable a crackdown on abuse which would likely result in substantial savings. Subsidies have proved difficult to cut, largely for reasons of political economy. The power subsidy is the largest state-government subsidy, and provides a good example of the difficulties involved in reforming subsidy regimes. 37 Subsidies are typically the prime candidate for expenditure cuts, but reducing subsidies has been a surprisingly difficult task - surprisingly difficult against the hopes and expectations of many governments and analysts, but not against the backdrop of international experience. The largest state subsildy is for the power sector: this reached about 1.4% of GDP in 2001/02. Large implicit subsidies are provided for the irrigation and higher education sectors, while explicit subsidies are provided by many states for public transport, housing and food. Subsidies are typically highly inefficient: they might lower prices, but they also lower quality. Because the subsidy is so expensive for government, quantity rationing is often introduced, with all the attendant inefficiencies that produces. A recent survey of farmers ranked satisfaction with money-lenders far in excess of satisfaction with irrigation and electricity services. Subsidies are also often regressive. This is particularly true of the power subsidies to farmers, sine~: only better-off farmers can afford to invest in a pump and bore-well. The size of the subsidy is 10 typically only one problem in the sector to which the subsidy is provided. Often there is a break-down in commercial discipline both between the service-provider and the customer - so that even the subsidized user-charge is not paid - and between the government and the service-provider - so that, amidst counter- charges of operational inefficiency and political interference, the subsidy owed to the service-provider is often not paid. 38 There has been some progress over the late nineties and into the new millennium in reducing the losses involved in the sale of electricity to households, and, more generally, the non-agricultural sector. A number of states have established independent regulators (21 at last count), imposed significant tariff increases, enforced collection of bills, cracked down on power theft, and ensured universal metering of domestic connections. There are some signs that the pattern of rising losses was reversed in 2002-03. However, whether is a reversal that will be sustained remains to be seen. Reform of the power sector in agriculture has been a case-study of repeated failure (see Box 2.3 of Volume II on the experience with attempts to increase the agricultural tariff in a number of states). Until this problem is solved, there can be no lasting solution to the problems of the power sector, since agriculture will continue to threaten to absorb whatever savings are made in the non-agricultural segment of the power sector. Yet this history of repeated reform failure itself implies that there are no easy or guaranteed solutions: if there were, they would have surely been tried. To make further progress in power sector reforms, tackling the lack of commercial discipline in the power sector has to be the top priority; without commercial discipline, no other remedy will work. 39 To take an extreme case, there is little point increasing tariffs if they are not paid. From this perspective, privatization is an attractive option. Private sector suppliers are likely to be less tolerant of non-payment by customers, and more ready to disconnect. Yet, the two privatization experiences so far have shown that this privatization is no panacea. Orissa privatized its distribution sector in 2000, but the benefits from this have been limited and delayed. Delhi privatized in 2002, and looks more promising, but more time is needed before a definitive assessment can be made. Neither Orissa nor Delhi have significant agricultural loads, so whether privatization of agricultural loads can succeed remains an untested possibility. Privatization carries with it its own risks, but given how politicized the power sector is in rural areas, it is hard to see commercial discipline being introduced into the rural segment of the power sector without it. There are many different forms of privatization which could be attempted, ranging from contracting out of metering and billing at the local level, to introduction of bulk-supply arrangements to groups of farmers or cooperatives, to part or full privatization of existing public-sector utilities. There is no guaranteed recipe for success and more experiments are needed to see what works. 40 There are many other reforms required beside privatization, including the reduction of theft and losses, reduction in industrial tariffs, increases in household and agricultural tariffs, improvements in the quality of electricity supply, and the promotion of competition in generation, now enabled by the Electricity Act 2003. Metering of farmers is critical if agricultural tariffs are to be linked to consumption levels. Metering would also enable subsidies to be better targeted, by use of a tariff schedule which increases the per unit tariff with consumption levels. Farmers resist metering, but would perhaps start to demand meters if governments could credibly commit that only metered tariffs would be subsidized. Some states have had at least some success with meter installation, but only to find that the meters are not read (Karnataka), or that mass cheating results in the introduction of metering leading actually to a drop in revenue (Rajasthan). This underlines the futility of adopting reforms if the fundamental issue of increasing the level of commercial discipline in the sector is not resolved. The same prescriptions, and implementation difficulties, will apply to reforms of subsidies other than power. 11 41 The most important of these is to promote commercial discipline by distancing the service- providers from government, through mechanisms such as corporatization and privatisation. Service providers must be empowered to take action against non-paying customers. Rules governing the allocation of subsidies should be clearly specified, and mechanisms put in place to ensure adherence by all parties, both government and service-providers. Perhaps the most important lesson from the power sector for the issue of subsidy reduction is that subsidies are here to stay. This suggests more attention be given to subsidy management. 42 Just as hiring and real wage restraint is a more feasible strategy for controlling salaries than wage cuts or downsizing so too, in many cases, maintaining subsidies at their current levels will be a major success. There has been some success in reducing food subsidies, partly through better targeting, and partly through Goi making more cheap grains available. But many subsidies will remain large claimants on government spending. State governments should focus more on improving subsidy management and targeting; this together with improving commercial discipline in subsidy-receiving sectors will have major efficiency gains which will themselves result in substantial subsidy savings. Public enterprise reforms - closure and privatisation - will not provide large, immediate fiscal gains, but will prevent the future build up of utilities, and will prevent the need for budgetary support to keep loss-making public enterprises afloat. 43 Beyond the fiscal imperative, public enterprise restructuring is also warranted from the perspective of removing governments from activities where there is no basis in public policy for their involvement. If this rationale is accepted, profit-making enterprises should also be put on the block. Most state governments have launched some sort of restructuring program for their public enterprises AP, Karnataka, Kerala and West Bengal account for nearly half of all PSUs, and all are actively pursuing reforms in this area. There have been more cases of closure than privatisation. This is partly because it is less political, but also partly because states have few viable privatisation candidates. Many lessons have been learnt, including the importance of political commitment, since every PSU will have a reason not to be included in the restructuring exercise. Institutional arrangements do not substitute for political commitment, but are also important. Establishing a dedicated unit with the mandate and capacity to implement the state's restructuring policy helps. It is no co-incidence that the two leaders in PSU reform- Orissa and AP -have both benefited from extensive consultancy support in this area. 44 The failure of retrenchment in the core civil service stands in contrast to its widespread use for public enterprise employees. The difference is perhaps that enterprises, unlike government, can be closed or sold, and thus their employees have far less job security than core civil servants. States have displayed widely varying degrees of generosity in their severance payments, though there is no evidence that states with lower severance payments have found it more difficult to attract VRS candidates. Interest savings will largely follow from reduced borrowing, but states can also take advantage of a low iinterest rate regime by aggressive debt-restructuring. 45 Some of the negotiated loans the states have taken allow for pre-payment. Put options on off- budget borrowing can be exercised. Goi has opened a window to allow for swapping high cost Goi and small saving debt. Goi could also help by allowing states to fund debt-restructuring by allowing additional market borrowing- but on the basis that the states finance the debt-swaps by approaching the market on their own, and on the basis of a credit rating. Alternatively, the states could finance debt-swaps through reform-based adjustment loans available from external agencies. The quality of spending must and can be improved. 12 46 The strength of the case for increasing spending levels in health, education and infrastructure depends on the extent to which quality can be improved. Quality is undermined by various problems, including: skewed composition of spending towards salaries, a regressive distribution of benefits; low civil-service productivity, high levels of corruption, an ineffective spread of funds over too many projects, and duplication of services provided by the private sector. What can be done to improve the quality of spending? Although the problems are legion, Indian states have also made some remarkable achievements in this area: these are detailed in a special annex to Chapter Two of Volume II. We highlight some common themes and examples of ways in which expenditure efficiency can be improved across sectors. These are essentially ways in which both citizens and policy makers can provide service-providers with stronger incentives to deliver effective services. Agency-specific reforms including an increased role for the private sector can improve service delivery. 47 Most government products and services are delivered by government agencies, most of whose performance levels leave a lot to be desired. Governments face two choices: either to get these agencies to work better under state-government ownership, or to transfer responsibility for the service - either by sale, liberalization, or decentralization - to the private sector or local government. Several states have shown that the performance of state government agencies can be improved, at least in some contexts, and that such tools as computerization, training and citizens' charters can deliver results, and provide employees with better incentives and resources to deliver. Similarly, experiences with privatization of solid waste management in some cities have been successful, as has deregulation of long-distance public transport. Even if governments are not ready to privatize entire services, individual service components can be outsourced. Several states now outsource the maintenance and cleaning of public hospitals, for example. At the same time success from private-sector participation is not guaranteed. Much depends on the enabling environment. The broader enabling environment is key to improving service delivery and the quality of spending regardless of agency ownership. 48 Whether services are delivered by the private sector, local government, or state government, success or failure will very much depend on the broader environment within which such services are provided. Especially where the private sector is already a predominant supplier (as in the case of the health sector), providing effective oversight and information can be as or more important than the government's own funding. Improving the enabling environment requires that commercial discipline and cost recovery should be promoted to improve accountability. Unless there are compelling efficiency and equity arguments to the contrary, governments should try to make Figure 5: Improvements in customer satisfaction with various services self-financing, should services in Bangalore over a 10-year periodg6 100 94 92 increase charges for better services, 90 85 78 and should refuse to provide services 80 73 to those who do not pay. Promoting "0 70 t3 60 commercial discipline is also key in ·~ 50 government-provider relations. V> 40 ~ 30 Outputs need to be clearly specified, 20 and full payments made in return. The 10 rapid improvement in Bangalore's 0 service delivery illustrates that "() § accountability, and through it, actual 8 service delivery can be significantly .£ u .:d :g 0... 111111 1994 1111111999 CJ 2003 improved by encouraging civil society to actively monitor government performance and promises (Figure 5). 49 Some states have adopted legislation to make procurement processes more transparent (Karnataka, TN) and to provide a legislative basis for the public's right to information (Delhi, Rajasthan, Tamil Nadu, Karnataka, Maharashtra and Goa). But passing legislation is only one part of a strategy for transparency. Governments need to make much more information available on a regular basis, following agreed rules. They also need to overhaul institutional structures to reduce the scope for discretion and corruption. Service delivery improvements can also be facilitated by decreasing the discretionary nature and the volume of mass transfers and in particular leaving reform champions in place. Strong and independent anti-corruption commissions can play an important role in launching high publicity raids and thereby putting pressure on the Government and its agencies to reform. Improving the financial environment within which the state governments operate will be critical to improve the effectiveness of government spending. Important experiments to improve financial accountability are now underway. 50 A series of State Financial Accountability Assessments suggests that the most important reform that states can undertake in the area of public expenditure management would be to base the budget on realistic revenue forecasts, to restrict new policy initiatives to the budget period, and then to relax post- budget central controls on spending. The quality and efficiency of spending would also be improved by enhancing departmental accountability and flexibility in the budgetary process. Departments could be given more flexibility to spend money as they best see fit to achieve agreed targets within an agreed envelope of resources. Lastly, accounting and auditing arrangements need to be tightened. India's accounting and audit arrangements are fine on paper, but neglected in practice: often, audit observations are not responded to and many local governments do not even produce accounts, let alone audits. A number of states have started to introduce far-reaching financial management reforms; much more learning across states is required to accelerate change in this area. India's states are currently stuck in a low-quality low-quantity expenditure equilibrium. Any fiscal empowerment strategy has to attack this problem from both sides. Thus, side by side with improvements in quality, need to come increases in expenditures in priority areas. 51 In many cases, improving quality will imply more expenditure. For example, with salaries fixed, spending more on non-salary inputs will require an increase in aggregate expenditure. Teachers may be more likely to be motivated in a well-equipped classroom; doctors perhaps more likely to attend to their work if they have drugs to dispense. And human development expenditures are likely to become more progressive as they increase in size. The mid-day school meal scheme, which has recently been instituted in a large number of states, provides a good example of how good quality additional expenditures can really make a difference in the social sectors (Box 3). Expenditure in priority areas can also, in theory, compensate for cuts elsewhere, e.g. on subsidies, though such effective packaging of reforms by state governments is yet to be seen. 52 Which particular spending areas should be increased, and by how much, will likely vary from state to state, depending on initial conditions, and identified priorities. Most states regard the broad areas of basic infrastructure, maintenance, and social sector spending as their priority, but it is important to bear in mind that sectors influence each other: thus the best way to spend money to improve health may be on water supply, for example. The states contribute more than half of India's general government capital spending, which is at about 4% of GDP, against the 6% target widely regarded as required to support sustained rapid growth; maintenance spending on state highways is at about 40% of what it should be; social sector spending, overwhelmingly financed by the states, is also low relative to need. Spending on non-salary inputs in the social sector is abysmal, a fact which is reflected in deteriorating facilities, and is 14 one determinant of low-quality service provision; in the poorest states, even salary spending is too lower: Bihar, for example, has pupil: teacher ratios of 1:90. Box 3: Mid-day meals- an example of productive, additional expenditure In mid-1995, the Government of India launched a centrally sponsored scheme, the National Programme of Nutritional support to Primary Education, under which cooked mid-days meals are to be served to children in all government and government-aided primary school: the central government provides the grains for free, and the state government provides cooking and logistical facilities. Orders from the Supreme Court in 2001 on a "right to food" litigation led to the widespread introduction of this scheme across most oflndia., though less so in the poorer states of Bihar, UP and Orissa. An evaluation of the scheme in three states (Rajasthan, Chattisgarh, and Karnataka) shows that the scheme has helped improve school enrolment as well as attendance, in addition to eliminating classroom hunger and breaking caste and class barriers. In particular, the study finds that enrollment increased on average by 15% in the three states covered in the year in which the mid-day mean was introduced, and notes that "qualitative data point firmly in the direction of a significant improvement in daily attendance. The study points to some shortcomings that need to be urgently addressed - inadequate infrastructure in some areas, insufficient monitoring of quality standards, need for more varied and nutritious menus etc. However, the experience so far also shows that public expenditure on the mid-day meals program has the potential to contribute significantly to both educational and social outcomes. Source: Dreze and Goyal, 2003 III. REVENUE REFORMS Strong growth in state revenues are essential to ensure fiscal imbalances are sustainable and developmental spending is sufficient to achieve the desired developmental outcomes. 53 Abstracting from cyclical factors - especially the down-turn in the late-nineties - state governments have done well to keep their own-revenue/GDP ratio almost constant, but this has not been enough to offset a decline in central transfers to the states. An increase in the own-revenue/GDP ratio of the states will be critical to reduce deficits, and fund much needed spending increases in priority areas. 54 India combines a very low tax take with very high rates. Stamp duties on property transactions are among the highest in the world (sometimes above 10%, compared to the 1-2% in many countries), as are combined center and state indirect taxes (often 30% compared to half that in many Asian countries). High rates and a low tax take imply a narrow base, reflective in particular ofthe inability of India's states to tax agriculture and services. Thus the great bulk of taxes are raised from industry which only constitutes 25% of the economy. One of the key challenges facing India's states is thus to broaden the tax base. Another is to simplify India's tax system and reduce corruption. India's indirect tax system is probably the most complex in the world, and surveys have shown state tax offices to be among the most corrupt government agencies in the country. The sales tax is the most important state-level tax. Encouraging progress in sales-tax reform - which led to a large boost in revenue - has been stalled by an impasse over VAT introduction. Progress might be made by allowing states to move individually towards VAT, without compensation, but on the basis of floor rather than uniform rates. 15 55 Significant progress has been made in strengthening revenue performance since the late 1990s. The Gol-led joint decision of state governments to introduce floor rates and discontinue industrial tax incentives from January 2000 gave an enormous boost to sales tax growth in the next two years. The key reform going forward is to replace the sales tax and other minor commercial levies by a destination-based consumption-type VAT on goods. 56 The centre has to take the lead ifthe decade-longjourney of the states' VAT is to resume. The co- ordinated one-time shift of all states to a VAT regime has come undone, with repeated postponements of the target date, the latest one being indefinite. Progress to VAT was stalled by political controversy as well as complications over the formula for compensation to be provided by Gol to states which would lose from VAT due to the introduction of uniform rates. Given the readiness and desire of several states to introduce VAT, a way out of the current impasse might be to allow states to graduate individually into VAT, without compensation, with the choice of rates left to the individual states (so that they can choose a revenue-neutral rate, and thus not require compensation), restricted only by a cross-state agreement on floor rates. Haryana is the one state which has not been deterred by GoI' s postponement of VAT. It in fact introduced a VAT on its own, effective April 1, 2003. It has since experienced reasonable revenue growth, and has no plans to roll back to stay the course. 57 The move to an integrated VAT on goods (and if possible services) is likely to have appreciable growth benefits, provided tax administration is also streamlined and tax payer non-compliance is held in check, but it will not be a panacea. Even after VAT is introduced at the state level, India will have one of the most complex indirect tax systems in the world. International experience speaks clearly to the difficulty of running an un-coordinated dual VAT; the long-term goal has to be to merge central excise (CENV AT) and state sales taxes (VATs) into a single, harmonized VAT system. Phasing out of the distorting and inequitable CST is critical, and will likely require compensation. 58 The Central Sales Tax (CST) is particularly distortionary and there is widespread agreement that it should be eliminated. Because the CST is levied on the basis of the origin of goods, it has enabled the relatively advanced states to export the burden of their own taxes to the less developed states, which tend to be net consumers. Origin-based taxation of interstate sales also leads to tax evasion via branch transfers and consignment sales, and is a serious impediment to the achievement of economic efficiency and a common market. There is no reason for the phase out of CST to be linked to VAT, but it will require compensation to the losers. It will also demand increased surveillance over interstate trade since shifting to zero-rating of inter-state exports will increase the incentive to evade the state sales taxes by claiming goods are being exported. Transfer to states of the right to tax services is a very positive initiative. 59 This constitutional amendment will give states access to the faster growing sector of the economy, and reduce India's high level of vertical imbalance. Over time, states should be allowed to integrate the taxation of services into their sales tax!V ATs. There is a rich agenda for reforms in other taxes, requiring wide-ranging actions by the center and the s.tates. 60 Currently, the professions tax- the base of which is essentially income or presumed income in the case of the self-employed - is not levied in five major states at all. Where it is levied, enforcement is weak. A key problem with the professions tax is that the constitutionally-imposed ceiling is very low, and needs to be raised. This requires action at the centre to amend the constitution and quadruple the floor rate. We estimate that potential collection from the professions tax is possibly 10 times the current level 16 of Rs 2, 200 crore (that is, 0.9% of GDP as against 0.1% currently collected) with ceiling revision and improved collection. There is also much scope to reduce evasion in state excise duties. The reform of stamp duties and registration fees requires a multi-pronged approach, including cutting stamp duties on property transactions from their current very high levels, through improving compliance and reducing corruption by a number of business re-engineering reforms (computerization, banning of stamp paper, etc). The transport tax can be reformed by raising the tax on private two and four-wheelers relative to buses. Liberalizing the public transport sector would help to grow the revenue-base, and reduce reliance on public-sector utilities with poor tax-payment records. Many other distortionary and non-transparent levies, mostly of minor revenue importance, need to be merged with the VAT (entry taxes, special levies). Increases in the electricity tax on households are warranted to compensate for low tariffs. Non-tax revenues have stagnated, and need more policy attention from state governments. 61 In contrast to own tax revenues, there has been a marked deterioration in the performance of non- tax revenues (Figure 6), reflecting in part the low 8 .0% Figure 6: Trends in States' Own Revenues as a Percentage of .0% 3 attention they have received c;np 2.8% 7.0% ., ~. . ::::::;> 6 . •Own from policy makers. Unlike Revenues (LHS) a ---..- ~ ~ 2.6% tax revenues, where many 2.4% problems are common across 6.0% Own Tax Revenues (LHS) 2.2% taxes and their 5.0% 2.0% administrations, different non- tax revenues sources have very 1.8% 4.0% distinct problems, though 1.6% across-the- board institutional 3.0% 1.4% strengthening 1s also of 1.2% importance. 2.0% +---+---+-+--+----t-+--+---tc--+--+-t---+--+-+---+--t--+ 1.0% N !2 62 Revenue from mineral royalties is already the fifth most important source of own revenue in some states (after the four main taxes discussed above) and there is potential for further increase. Measures that may further increase the importance of this source of revenue include: rule-based setting of royalty rates by the Centre, and also states, including revision at least once in three years with reference to market prices; streamlining of the clearance process for grant of mineral prospecting licenses with the help of automation; strengthening of administration, particularly where minor minerals are of importance; and introduction of self-assessment and risk-based scrutiny of royalty returns to reduce litigation. 63 The performance of user charges has been poor. A key reason for this is inadequate, and in many cases, deteriorating cost recovery. As already discussed, some user charges are politically very sensitive, but not all. More regular monitoring and frequent adjustments would help the state governments. Of course, in areas where external benefits or merit good characteristics are unimportant, rather than attempting recovery of user charges, the government should ideally withdraw, and let market forces prevail. Tax administration reforms are probably more important than tax policy reforms but have received relatively less attention to date. 64 Sustained improvement in the revenue performance of the states cannot be achieved without thorough institutional reforms in tax administration and inter-state coordination facilitated by the induction of modern technology. The Centre could take the lead in institutionalizing inter-state 17 coordination, not only in tax policy but, perhaps more important, tax administration and information sharing. Institutions for coordination and information sharing between central and state tax departments also need urgent attention. Most importantly, strengthening weak tax administration institutions and related e-governance reforms will enable effective use of enforcement information, increased managerial control, better incentives for tax department staff, and improve taxpayer services. (See Box 3.4 ofVolume II on the positive experience with performance measurement in the commercial tax department in AP). Of all pending reforms, improved enforcement technology and procedures coupled with staff incentives, management flexibility and effective anticorruption institutions have the greatest potential to lead to a significant and sustained increase in state revenue. IV. STRENGTHENING THE FISCAL FEDERAL FRAMEWORK While India's federal fiscal system shows many strengths, prime among them stability, the fiscal crisis has also brought to light a number of weaknesses, which undermine incentives for performance, and hurt the poorer states. 65 India's federal system stacks up well internationally by several criteria (see Box 4 for an international perspective). It is a very stable system, with a high degree of legitimacy provided by the constitutionally-mandated and respected Finance Commissions that are appointed every five years. It avoids some of the glaring defects which several other federations have suffered from, such as allowing state governments to "print money" or to regularly default on their debt or borrow overseas. Box 4: International perspective on India's fiscal federalism Expenditure responsibilities India is very decentralized in Sub-national as a terms of expenditure responsibilities. The share of state to total Eercentage of total ... {111{21 expenditure is about 57%, this is the same as Canada, and below ExEenditure Revenue China, but above most other countries. The average for OECD {11 {21 countries is 35% and for Latin America 15%. China 81.5 59.7 1.37 Canada 58.8 53 1.11 Revenue transfers and vertical gap. The share of state in total India 56.7 39.1 1.45 revenues in India is about 39%. This is about average, but low Denmark 53.9 31.7 1.70 given the extensive expenditure responsibilities oflndia's states. Australia 49 31.7 1.55 The ratio of expenditure to revenue for India's states is 1.45, Argentina 45.6 39.1 1.17 which is below Australia and Denmark in the table, but above U.S.A. 44.4 42.1 1.05 everyone else. This reflects a high degree on central transfers as Germany 38.8 33.8 1.15 well as borrowing. Equalization. Most federations have some sort of commitment to equalization built into their federation. Some, such as Canada and Germany, have the goal of horizontal equalization explicitly built into its constitution.; India's constitution, in many respects detailed, is quite vague on the principles which should guide central transfers, leaving the onus of interpretation on the Finance Commission. Similarly, in terms of outcomes, central transfers are progressive in practically all federations, but very equalizing only in some (Canada, Germany, Mexico, Australia), and partially equalizing in others (Brazil). As we discuss in detail in the text, India fits in the latter category, no doubt in part because the constitutional and institutional commitment to equalization is India is much less explicit. Borrowing rules India's rules for sub-national borrowing, especially as they are put into practice, are relatively liberal. An IMF study of 53 countries found that all but 6 had some restrictions on sub-national borrowing; in 16 sub-national governments were barred altogether from borrowing; in 19 overseas borrowing was banned. (see Ter- Minassian and Craig, 1997; Anand, Bagchi and Sen, 2003) Overseas borrowing by states is banned in India, but, unlike in many other countries, there are no aggregate limits on domestic borrowing: by contrast, in Italy, for 18 example, regions can borrow only for capital projects and subject to a limit that debt service not exceed 25% of a particular definition of revenue; Brazil's states, under the country's Fiscal Responsibility Act, face a debt limit equivalent to two times the states' net current revenue to be achieved within a 15-year period, with annual eductions equal to 1/15th of the original excess over the limit. Two restrictive features of India's borrowing arrangements, which have not always been present in some others, and have caused trouble when they have not (e.g. in Argentina, Brazil), are that states are not able to borrow from central or provincial banks ("print money") and that there is a limited, though not absent, history of debt write-offs. India is unusual in acting as a creditor to state governments. Only a few other developing countries do this, for example, Pakistan. Debt levels: Reflecting the relatively liberal borrowing regime, India's states seem to be the most highly-leveraged in the world. In 2000, for all of India's states combined, the ratio of debt to revenues stood at 203%. Canada was next with 189%, Brazil with 170%. In Argentina the ratio was 69%, and the US 44%. 66 For all its strengths, the fiscal crisis has put the weaknesses of India's fiscal federalism in stark relief. The regime of revenue transfers is not very progressive, and contains few incentives for good performance. Central transfers have been falling over time, with no offsetting increase being provided to the states through an increased tax base. India's borrowing regime is on the one hand too strict where more t1exibility could be allowed (the detailed rules governing access to the six different borrowing sources states have could be relaxed) and on the other not strict enough where control is needed (over the total borrowing a state undertakes). Credit markets play a very small role in the allocation and pricing of loans, and all states face the same interest rate for borrowing. Thus, with no penalties for profligate behavior, states can borrow now, and worry later. Both the grant and the loan regimes are complex, and have significant amounts of discretion built in to them. Finally, grants and loans are linked together, preventing these two types of transfers from being given the different treatment they require. Over the nineties, states have received fewer grants and more loans; this trend Figure 7: Resources other than owu revenue used by states to finance expenditure needs to be reversed. 67 The nature oflndia's fiscal federal 6% transfers changed greatly over the course 5% of the nineties. States have suffered from a "scissors effect" whereby, since 1998/99, 4% borrowed resources have come to 3% supplement a state's own revenues to a greater extent than do central government 2% grants (Figure 7). India's states are now the most leveraged sub-national 1% governments in the world. Central grants 0% to the states need to recover their lost ~ II II " ~ ~ i!l 11; ~ Sl !l ground, while borrowing needs to be ~ ~ ~ ~ ~ ~ ~ ~ ~ Ii reduced to sustainable levels. This is --+--Borrowed Funds ....,.._Transferred Funds (from Gol) particularly true for the poorer states, which are highly dependent on both loans and revenue transfers to supplement their own meager resources. Revenue transfers need to be made more progressive, loans less so. States need to be given more flexibility to borrow, especially from the market, but under a centrally-imposed global borrowing cap. 68 There is widespread agreement in India that the budget constraints facing states need to be hardened. Making it much harder for states to run up arrears in the power sector has been a very positive 19 development in this regard. From the perspective of fiscal sustainability, no reform is now more important than the centre exercising greater control and restraint over the aggregate borrowings of states, as it Figure 8: Per Capita revenues (normalized to Punjab=l}, 2000/01 is entitled to do so under the constitution, by introducing global caps on aggregate borrowing. (See 0.9 +lllllf-111111-11111---liiii--W----=-------~ Box 4.6 ofVolume II on the international experience 0.8 +lllllf-111111-11111---liiii--W--~------------~ with global caps) These global caps would be 0. 7 +lllllf-111111-11111-liiii--W--ID---Dil---lillll-----....-------~ 0.6 +lllllf-111111-111-11111--W--ID---Dil---lillll----lllll---1..--~~ decided by the central government, but it would be 0. 5 +111111--.1--1111-liiii--W--ID---Dil---lillllf-lllll--llll-l\111----- useful if they could be aligned with the state's own 0.4 +lllllf-liiii--III--11111--W-----Dil---lillllf-11111--1111-l\III-W---~ reform objectives, for example, the targets listed by 0.3 +lllllf-lllll--llll-liiii--W--ID---Dil---lillll-1.1!11--1111-l\III-W-----1.1!11-- the state in its Fiscal Responsibility Act. The use of 0.2 +111-111-1~----------f---111-1~------------ global borrowing caps, common in many federations 0.1 +IIIII-III--II-I\III-W--~-----Iillll-l---ll-l·---c------- around the world, will not only force fiscal deficits 0+-~"-r-"....--.-.---~,-,..-..,..,_,--,-,--,,--, .0 ., down, but will put an end to lobbying for loans. For ., c ·c ~ efficiency reasons, and to force governments to face " :;; 0.. I the costs of fiscally imprudent behavior, not in the future but today, states need to be given more flexibility in who they borrow from. Rather than • Om! revenues IIIII Transfers having six different borrowing sources, each Note: States arranged from left to right in with their own rules, as at present, states should descending order ofper capita income have much more access to credit markets, at an interest-rate which reflects their creditworthiness. Tighter control over central government financial institutions lending to state governments would also help achieve this objective. States which have difficulty accessing the market can be given a central government guarantee - at a price. Gal could act as a regulator of state borrowings, enforcing the global cap, rather than as a creditor itself. 69 Debt-restructuring along commercial principles is in line with allowing states more flexibility under a global cap. Performance-based debt-relief will also help poor, highly-indebted states. At the same time, however well-designed the scheme, it runs the risk of undermining hard-budget constraints. More progressive grants, and ones which reward fiscal reformers may be a better alternative. The federal grant regime is complex, and views on how it should be reformed divergent. Based on international comparisons, and looking at past trends, we conclude that there is a case for expanding the tax base of the states, and increasing central grants and making them more progressive and less discretionary. Conditions on special-purpose grants can be used to strengthen incentives for performance, but the number of such grants needs to be cut. 70 The grant regime is institutionally more complex than the borrowing regime and any recommendations concerning it must face difficult issues concerning trade-offs between the need of poorer states for more funds, and doubts concerning their ability to effectively use them: a subject of much debate in India. There are three main types of grants. The largest are the Finance Commission transfers: they account for two-thirds of total central government transfers. The other two are special- purpose grants for particular schemes provided by the central government, and block grants also from the central government which help finance state plans. Finance Commission grants are the most progressive of the three, but as Figure 8 shows the current distribution of all grants together is only mildly progressive. Middle-income states often receive more in central transfers than poor states, and the high- income states enjoy total per capita revenues twice or three times as much as the poorer states (Figure 8). Indeed, once one includes resources transferred through the Food Corporation of India subsidized grain procurement practices (which principally benefit Punjab and Haryana), the federal transfer system is 20 virtually lump-sum, i.e. distributed across states on the basis of population, a remarkable outcome for such a complex system (Figure 9). 71 While the other extreme of extensive equalization is neither feasible nor desirable, given India's sharp and growing regional divergences (and absent from many other federations), the very mild progressivity of the current distribution does not seem optimal. One suggestion is to introduce a "Representative Tax System" so that poor states are assured of at least average revenue, provided they put in average revenue-collection effort. This was first introduced by Canada and is now used by a number of other countries, including Russia (see Box 4.3 ofVolume II). To ensure incentives for performance, there is widespread agreement that the reliance of the Finance Commission on "gap filling" as a basis, albeit limited, for the distribution of resources needs to be dropped. Many would also argue that conditions on special-purpose grants should be used to encourage good performance, but the number of such grants needs Figure 9: State revenues before and after transfers (formal and FCI) to be reduced and their "performance-conditional" 100~------------------------~~- distribution could be made more progressive, i.e. poor 90+-------------------------- states be enabled to access more provided they meet the performance standards. Block grants, which fund state plans, themselves should either be discontinued, or, if continued, should be delinked from loans so that each can be provided the different treatment it deserves: grants distributed on the basis of performance and need; loans on the basis of credit- worthiness. 72 As noted in para. 18, one important response by the central government to the deterioration in state finances has been the creation of a number of reform incentive funds. The Fiscal Reforms Facility rewards 0+---~----~--~----~----~--~ fiscal adjustment, and there are now also funds to 0 5000 10000 15000 20000 25000 30000 incentive power and urban reforms, among others. GSDP per capita These funds are controversial, and seen by some as too intrusive. However, they also counter-balance the incentives in the system for irresponsible behavior. While of course adverse incentives to good performance should be removed where possible, some will likely be permanent features, such as a high degree of vertical imbalance. Thus there is a good case for strengthening rather than eliminating these reform facilities. Ways to strengthen them would include: (i) increasing transparency by, for example, making all transactions and agreements under such funds public; (ii) setting the adjustment path according to the initial conditions of each state, so that targets are not unrealistic for some and inadequate for others; (iii) defining indicators in terms of targets the states themselves use (e.g. their own Fiscal Responsibility Act fiscal targets) and/or can control; and (iv) strengthening reporting requirements, so as to Figure 10: Trends in Gol Revenues and Transfers to States restrict the scope for creative 45.0% 15% Go! Revenues (RllS) accounting practices, such as :g 14% postponing payments and/or ::J 42.5% &:: 13% distorting expenditure ~ 12% ~ through transfers to public ~ 40.0% 11% ~ 10% ~ accounts. 73 The fall in central - - 0 &:: Q) 37.5% 35.0% &:: 9% 8% 7% Q) I:! ~ I:! transfers to the states is Q) ll. 6% 32.5% 5% (!) 1'- 00 CJ) 0 ..... N (') "