The Indian Constitution (1950) originally provided a two-tier federal system of government, namely the Centre (or national government), and the state governments (sub-national).1 It assigned separate tax powers and expenditure responsibilities to them. Since all mobile and buoyant taxes were assigned to the Centre and most expenditure functions were awarded to the states, this led to the vertical fiscal imbalances (Rangarajan and Srivastava, 2008). Realizing this vertical gap, the Constitution has provided for the intergovernmental transfer mechanism to transfer the resources from the Centre to the states in the form of tax devolution, grants-in-aid and centrally sponsored schemes (Rao, 2005).
Despite this Constitutional arrangement, both central and state governments borrow from internal as well as external sources when their expenditure exceeds the revenue. There are two reasons why the governments use fiscal stimuli, often financed by excess borrowing, to expand their activities above the trend level. The first one is to play a counter-cyclical role to minimize the negative impacts of growth cycles, while the second one has political motives and is due to the governments' fiscal expansions. The former reason is a response to economic cycles while the latter causes political cycles mostly driven by the timing of elections (Srivastava, 2012). The trends in public debt relative to GDP (Gross Domestic Product) since independence indicate the cyclical nature of the former and the secular upward trend of the latter (Rangarajan and Srivastava, 2005).
Although there has been a sharp deterioration in the debt/deficit situation of both governments since the second half of the 1990s, the central government adopted a rule-based framework called Fiscal Responsibility and Budget Management (FRBM) Act in 2003 and subsequently, most states adopted their own FRBM rules. These efforts improved the debt positions of both governments in the initial years but worsened after the global slowdown in 2007-08. The combined total liabilities of the Centre and the states together amounted to Rs.93869.5 billion ([approximately equal to] US $1439.3 billion)2 as on March 2016. Of these, the liabilities of all states together amounted to Rs.31749.4 billion. The combined public debt-GDP ratio in 2015-16 amounted to 69.5%. These figures raise the question of public debt sustainability in India.
In the literature, there are three major approaches for empirically testing public debt sustainability. These include (1) unit root approach (see Trehan and Walsh, 1991 and Uctum et al., 2006), (2) cointegration approach (see Hakkio and Rush, 1991; Jha and Sharma, 2004) and (3) Bohn's model (see Bohn, 1998; Abiad and Ostry, 2005; Greiner and Fincke, 2009).
Of these three approaches, the Bohn model has become popular for its statistical properties.3 The original Bohn model was linear and was estimated using the Ordinary Least Square (OLS) method. It was applied to test the debt sustainability of the US government (See Bohn, 1998). Of late, it has been widely used to test for debt sustainability in different nations (see for instances, Haber and Neck (2006) for Austria, Kia (2008) for Iran and Turkey and Kaur and Mukherjee (2012) for India). It has been further extended to include non-linearities, time-varying coefficients by using the penalized spline technique (Greiner and Kauermann, 2008), multi-country panel data set with regular fixed effects and random effects model and not the spline (Abiad and Ostry, 2005; Adams et al., 2010) and testing for the sustainability of sub-national governments (Fincke and Greiner, 2011a; Mahdavi, 2012).
Since India has a federal system of government, the debt sustainability is relevant not only for the Centre but also for the states. A few earlier studies, for example, Dholakia et al. (2004), Misra and Khundrakpam (2008) and Maurya (2015) tested for debt sustainability in Indian states using the traditional indicator approach (Domar condition and its extensions). Kaur et al. (2014) use the panel data estimation procedure for 20 major Indian states during 1980-81 to 2012-13 and find the evidence of sustainability in all the states together. However, the debt level may differ among states and the debt sustainability status may also vary in different states. For example, the debt-GSDP (Gross State Domestic Product) ratio was 54.9% in Jammu and Kashmir while it was only 14.8% in Chhattisgarh. Therefore, an aggregate level analysis may not reveal the true picture of individual states.
The main contribution of this study is that it utilizes innovatively the panel data version of the Bohn model and a penalized spline (p-spline) estimation procedure for the time series to test the public debt sustainability in each of the 20 Indian states during 2003-04 to 2014-15. Further, it shows how time-varying coefficients, also called the reaction coefficients associated with the debt-GSDP ratio for each Indian state has evolved over time.
The rest of this study proceeds as follows. Section 2 briefly describes the debt scenarios of Indian states. Section 3 reviews the relevant literature. Section 4 explains the methodology used in the study and section 5 discusses the empirical results. The study concludes with section 6 which provides various policy implications.
(2.) PUBLIC DEBT SCENARIO IN INDIA
Three different types of deficit measures are used in India for both the Centre and the states: (i) revenue deficit (excess of revenue expenditures over revenue receipts), (ii) fiscal deficit (total expenditure [revenue + capital] minus the sum of revenue receipts and non-debt capital receipts = net borrowing) and (iii) primary deficit (fiscal deficit minus interest payments). Among these broad measures, the primary deficit/ surplus (primary balance) shows the dynamic effects of the governments' policy initiatives as well as the attainment of governments' fiscal performance. As a subset of fiscal (deficit) balance, it indicates the amount of government borrowings that are required to meet the expenses other than the interest payments (primary deficit) or the amount of borrowings required to meet the netted interest commitments on previous borrowings (primary surplus). Therefore, the primary balance is the root cause of all type of deficits, which further reflects in the total debt requirements.
The government borrows money from various sources to finance its fiscal deficit and accumulation of these borrowings leads to public debt. Public debt in India consists of central and states debt (domestic and external). While calculating the combined debt of the Centre and the states governments, the Centre's lending to the states requires to be netted out (Srivastava, 2005). While the Centre in general borrows from the internal and the external sources, the states' debt is restricted to market loans and bonds, ways and means advances from RBI, loans from banks and other institutions, provident funds etc. The states can only borrow from external sources subject to ceiling and approval by the Centre. In recent years, Indian states have borrowed more loans from international agencies for the purposes of development. High and growing proportion of external debt would obviously be problematic as per theory.
Looking into the long-term profiles of debt accumulation in India, the first three decades since independence marked a period of surplus revenue at the Centre. Around 1979-80 deficit was noticed and gradually increased in all deficit measures for the next 10 years. However, after the macroeconomic crisis in 1990-91, the central government initiated various reform programs (tax and non-tax reforms, expenditure management and institutional reforms) and this resulted in a budget deficit and huge public debt. This trend continued till the implementation of the FRBM Act in 2003 (Misra and Khundrakpam, 2008).4
On the other hand, the fiscal position of the state governments remained comfortable in the first three decades since independence. The state finances exhibited signs of fiscal stress from the mid-1980s. However, the debt position of states generally remained under control during 1980-81 to 1996-97. The period from 1997-98 to 2003-04 was marked by a sharp deterioration in key deficit indicators of the state governments. Particularly, the average debt-GDP increased to 26.9% during this period. At the same time, the actual debt-GDP ratio increased to 31.8% level in 2003-04 (Kaur et al., 2014).
Table 1 presents the combined data for all the states and the Centre, reflecting the trends in primary balance and public debt, both relative to GDP. Initially, there was a minimal combined primary deficit, which later became a surplus in 2006-07 and 2007-08 for both the Centre and the states, mainly because of the introduction of the FRBM Act.5 The situation further worsened after the global crisis in 2007-08. The combined primary deficit reached an alarming 4.53% level in 2009-10. Of these, the Centre's primary deficit alone was 3.17 %. In order to bring down the deficit, the fiscal consolidation measures were adopted at all levels of government. While the Centre's deficit started declining at a very slow pace, the states' deficit declined rapidly to 0.36% in 2011-12, but again started increasing to 0.79% in 2015-16.
The aggregate debt profile of the states had been impressive up to the late 1990's, after which it started increasing from Rs.3996 billion in 1998-99 to Rs.10141billion in 2004-05 (26.3% of GDP) and then reached Rs. 27853 billion (20.8%) in 2014-15. Chart 1 shows the trends in debt-GSDP ratios and primary balance-GSDP ratios of 20 major Indian states during 2003-04 to 2014-15. In 2003-04, Himachal Pradesh (-4.21), Uttarakhand (-3.78) and Uttar Pradesh (-2.74) had a huge primary deficit-GSDP ratio, while Madhya Pradesh (3.79), Maharashtra (2.62) and Jharkhand (0.62) had a primary surplus. In 2014-15, the situation...