Macroeconomics and Finance in Emerging Market

Economies

ISSN: (Print) (Online) Journal homepage: https://www.tandfonline.com/loi/reme20

Sub-national government debt sustainability in

India: an empirical analysis

Sangita Misra, Kirti Gupta & Pushpa Trivedi

To cite this article: Sangita Misra, Kirti Gupta & Pushpa Trivedi (2021): Sub-national government
debt sustainability in India: an empirical analysis, Macroeconomics and Finance in Emerging
Market Economies, DOI: 10.1080/17520843.2021.1948171

To link to this article: https://doi.org/10.1080/17520843.2021.1948171

Published online: 06 Jul 2021.

Submit your article to this journal

Article views: 43

View related articles

View Crossmark data

Full Terms & Conditions of access and use can be found at

https://www.tandfonline.com/action/journalInformation?journalCode=reme20
MACROECONOMICS AND FINANCE IN EMERGING MARKET ECONOMIES
https://doi.org/10.1080/17520843.2021.1948171

Sub-national government debt sustainability in India: an

empirical analysis
Sangita Misraa, Kirti Guptaa and Pushpa Trivedib
a
Department of Economic Policy and Research, Reserve Bank of India, Mumbai, India; bHumanities and Social
Sciences, Indian Institute of Technology Bombay, Mumbai, India

ABSTRACT ARTICLE HISTORY

Recognizing the increasing precedence of fiscal shocks leading to a Received 7 July 2020
deterioration in states’ debt due to the realization of contingent Accepted 23 June 2021
liabilities, this study assesses the debt sustainability of Indian states KEYWORDS
by employing both conventional and augmented debts, obtained Sustainable debt; primary
by incorporating information on states’ guarantees. Results indicate balance; Indian states; Uday;
that states’ debt is just sustainable with potential signs of unsus­ panel data; random effects
tainability. Guarantees given by states, if invoked, could certainly
pose a potential risk to debt sustainability for Indian states. The JEL classification
H63; H62; H72; H740; C230
study suggests revisiting and reviewing states’ FRLs with the inclu­
sion of debt as a medium-term anchor, and greater transparency
with regard to contingent liabilities.

Introduction
The fiscal policy at the state level interacts with people and seeks to maximize people’s
welfare. Indian states today share a larger responsibility in the implementation of govern­
ment programmes and in influencing all social and economic parameters necessary to
achieve at least three-fourths of the Sustainable Development Goals (SDGs) by 2030.
Reflecting on these growing responsibilities, state budgets in India have ballooned in size
during the current decade. In terms of total receipts and expenditures together, the size of
the state governments is about three times that of the Central government, highlighting
the importance of state finances in the overall public finances of the country. The share of
state governments in general government expenditure (Centre plus states), usually
termed decentralization ratio in international parlance, hovers around 60% compared
with the global average (advanced and emerging market economies) of around 30% (IMF,
2014). While the focus of the states has been to meet the growing demand for public
services through higher capital and developmental expenditure, they have had to operate
within the available budgetary space and manage trade-offs leading to higher
borrowings1 and an increase in debt, thus underscoring the need to focus on states’
finances for medium-term macroeconomic stability and orderly financial market
conditions.

CONTACT Sangita Misra sangitamisra@rbi.org.in

The earlier version of this paper is published under RBI Working Paper Series, July 2020.
© 2021 Informa UK Limited, trading as Taylor & Francis Group
2 S. MISRA ET AL.

At an aggregate level, fiscal deficits of states taken together have broadly remained in
line with states’ Fiscal Responsibility Legislation (FRL) targets, except during the Ujwal
DISCOM Assurance Yojana (UDAY) years (2015–16 and 2016–17). There is, however,
a tendency in the post-FRL period, on the part of the states, to meet deficit targets by
cutting down expenditure, with developmental expenditure being the usual soft com­
promise, thus raising questions on the quality of fiscal consolidation with implications for
long-term sustainability (Chakraborty and Dash 2017). States have also had to adjust to
various fiscal shocks at periodic intervals in the form of schemes like farm loan waivers, the
Financial Restructuring Plan (FRP), and UDAY, which in turn exacerbate expenditure
pressures and increase debt. Tax reforms such as the Goods and Services Tax (GST),
though expected to benefit the economy, have also limited the revenue-raising auton­
omy of states. In the medium term, the state governments may have to further increase
their spending to meet the need for growing urbanization. It is, therefore, important to
analyse and understand states’ finances from a sustainability perspective and then assess
their capabilities to meet evolving demands.
Analysing the debt sustainability of state governments in India is important for three
reasons: first, unlike the Centre, states’ debt shows an upward trend despite consolidation
in terms of conventional gross fiscal deficit to gross domestic product (GFD-GDP) ratio
thus hinting at the increasing influence of off-budget borrowings on which available data
are limited. Second, most state public sector enterprises (SPSEs) are running in losses with
weak cost recovery mechanisms. Guarantees given by state governments, which repre­
sent off-budget borrowings for the states, have been on a rising trend. The majority of
these guarantees are given to the power sector, with the precedence of these liabilities
falling on states’ budgets, as has happened thrice in the 13 years from 2002 to 2015, under
One Time Settlement (OTS), FRP and UDAY.2 Third, no holistic analysis of states’ debt,
incorporating both conventional debt and off-budget debt, has been attempted,
although the existing literature suggests the need to do so.
Looking at the literature on the subject in the Indian context, the latest attempt to
analyse states’ debt was made by Kaur, Mukherjee, and Ekka (2018), a study which
concludes that while conventional debt may be sustainable for states, a comprehensive
debt sustainability analysis incorporating off-budget items/guarantees into states’ debt is
needed. Accordingly, this paper incorporates contingent liabilities like state government
guarantees in the analysis to assess the debt sustainability of states. The existing literature
across the globe has also started focusing on this aspect. Large debt spikes globally over
the last decade have been more due to stock-flow adjustments via contingent liabilities
realizations, under-reporting of deficits, valuation effects, etc., rather than large primary
deficits (Jaramillo, Mulas-Granados, and Kimani 2017). Studies also emphasize the need to
improve fiscal transparency regarding off-budget operations which have become
a conduit to circumvent the fiscal rules (Von Hagen and Wolff 2006). The paper also
adds to the literature in terms of improved temporal and spatial coverage. This is the first
state-level study to use post-UDAY period data up to 2017–18.3 Also, it covers all the
states, marking an improvement over earlier studies which covered select states.4 It may
be noted, however, that the empirical analysis in the paper is for period, 2004–05 to 2017–
18, thus excluding the current COVID-19 pandemic period.
The rest of the paper is organized as follows: Section II examines in depth the trends in
subnational debt,5 with a focus on guarantees. Section III outlines a review of the select
 MACROECONOMICS AND FINANCE IN EMERGING MARKET ECONOMIES 3

literature. Data and methodology adopted for the analysis are provided in Section IV.
Results, based on both indicator-based approach and empirical approach, are presented
in Section V. Section VI sets out the concluding observations.

Subnational debt: trends

Debt
States’ debt is increasing at a double-digit rate, at a pace higher than nominal GDP growth
(Table 1). States’ debt rose for about a decade prior to 2003–04, but witnessed
a significant consolidation after the enactment of FRLs by the states. The adherence to
the fiscal rule legislations was also supported by the debt relief provided by the two debt
restructuring schemes, viz. Debt Swap Scheme which was in operation from 2002–03 to
2004–05 and Debt Consolidation and Relief Facility which was functional during the
Twelfth Finance Commission (FC-XII) award period 2005–06 to 2009–10. As a result, the
ratio of interest payment to revenue receipts declined sharply during the period 2003–04
to 2014–15. However, after the implementation of the UDAY scheme, states’ debt
recorded a sharp increase in 2015–16 and 2016–17 to around 25% of GDP. The expendi­
ture pressures on account of large committed expenditures and schemes like farm loan
waivers also contributed to the recent rise in states’ debt (Table 1 and Chart 1). Another
indicator – debt to own revenue receipts ratio – of state governments reversed its
declining trend in the post-UDAY period, crossing 300% since 2015–16 (Chart 2).
A scatter plot depicts a positive relationship between fiscal deficit and debt. As per the
2018–19 Revised Estimates (RE) data, many states remain below the 3% ratio of fiscal
deficit to GDP, while the threshold debt to GSDP ratio of 25% as perceived by the
Fourteenth Finance Commission (FC-XIV) stands breached by many states. However, if
we consider a little strict measure of the implicit FRBM debt ceiling of 20% prescribed by
the FRBM Review Panel (as adopted in the Union Budget 2018–19), most states remain
above the threshold level (Chart 3).
On comparing India’s subnational debt vis-à-vis that of its BRICS6 counterparts, India’s
debt remains the highest (Chart 4) followed by China. China’s official debt has accumu­
lated largely due to rising local government debt and the subdued performance of public
sector enterprises (PSEs). In addition, if the off-budget liabilities of local government
amounting to 30% of GDP are added, then subnational debt for China would rise to

Table 1. State Governments’ Outstanding Liabilities.

(per cent)
Year (end-March) Amount (₹ lakh crore) Annual Growth Debt/GDP
2013 22.45 10.6 22.6
2014 25.10 11.8 22.3
2015 27.43 9.3 22.0
2016 32.59 18.8 23.7
2017 38.59 18.4 25.1
2018 42.92 11.2 25.1
2019 (RE) 47.15 9.8 24.8
2020 (BE) 52.58 11.5 24.9
Note: RE: Revised Estimates; BE: Budget Estimates.
Source: Budget documents of state governments and authors’ calculations.
4 S. MISRA ET AL.

35.0 (Post-FRBM period) (Post UDAY)

(Pre-FRBM period)
30.0
Per cent to GDP

25.0
20.0
15.0
10.0
5.0
1990-91
1991-92
1992-93
1993-94
1994-95
1995-96
1996-97
1997-98
1998-99
1999-00
2000-01
2001-02
2002-03

2006-07
2007-08
2008-09
2009-10
2010-11
2011-12
2012-13
2013-14
2014-15
2015-16
2016-17
2017-18
2018-19
2019-20
2003-04
2004-05
2005-06
IP/RR DEBT/GDP

Chart 1. States’ Outstanding Debt and Ratio of Interest Payments to Revenue Receipts. Note: IP/RR is
Interest Payments to Revenue Receipts. Source: RBI, State Finances: A Study of Budgets, Multiple years.

35 500
450
Per cent to GDP

30
400

Per cent
25 350
300
20
250
15 200
2018-19 (RE)
2019-20 (BE)
1999-2000
1991-92
1992-93
1993-94
1994-95
1995-96
1996-97
1997-98
1998-99

2000-01
2001-02
2002-03

2005-06
2006-07
2007-08
2008-09
2009-10
2010-11
2011-12
2012-13
2013-14
2014-15
2015-16
2016-17
2017-18
2003-04
2004-05

Debt to GDP Debt to Own Revenue Receipts (RHS)

Chart 2. States’ Debt: As a ratio to GDP and Own Revenue receipts. Source: RBI, State Finances: A Study
of Budgets, multiple years.

over 50% in 2017 (IMF, 2018). Other countries’ debt, e.g. Colombia, Argentina, and
Indonesia, featuring market borrowings by states, remained subdued and amounted to
less than 10% of their GDP.

Going beyond state governments’ conventional debt – guarantees

Apart from direct budgetary support to SPSEs, through equity and loans, state govern­
ments provide off-budget support to SPSEs by providing guarantees on repayment of
their borrowings taken from financial institutions and share of such off-budget borrow­
ings has gone up in recent years. This has linkages both with the institutional structure of
the fiscal framework for the Centre–state relationship and the need for strict adherence to
the fiscal rules on the part of states amidst the rising responsibilities entrusted on them.
 MACROECONOMICS AND FINANCE IN EMERGING MARKET ECONOMIES 5

Chart 3. Debt Dynamics of States.

Source: RBI, State Finances: A Study of Budgets, multiple years.

24.8
25.0 20.6
per cent to GDP

20.0
15.0 12.5
11.1
10.0 5.6 4.8 4.3 4.2
5.0 0.3
0.0
India

Colombia

Argentina

Russia

Korea
China

Brazil

Indonesia
South Africa

Chart 4. Cross-Country Subnational Government Debt 2018.

Note: Data pertains to 2016/2017 for South Africa, Russia, and China. Source: (OECD-UCLG 2019)

During the past decade, in order to meet fiscal targets amidst the rise in revenue
expenditures states very often took recourse to curtail capital expenditure. There are also
instances of pushing the capital expenditures to SPSEs by giving them guarantees with
a concomitant rise in off-balance-sheet expenditures (CAG, 2019). Accordingly, while
budgetary support to the SPSEs to meet their capital requirements saw a gradual decline,
issuance of guarantees by state governments helped these entities to borrow from the
market.7 There were thus, several positive gains from these guarantees. First, it has helped
states to undertake capital expenditure in key infrastructure sectors like utilities, trans­
portation, housing, urban development and other welfare projects of state governments
(RBI, 2002). Second, operating through SPSEs helps states achieve some of their macro­
economic goals, particularly in sectors characterized by market failures. Third, it helps
states get over their hard budget constraint. However, conducting capital expenditure
through SPSEs with the use of guarantees, while desirable, has been fraught with
6 S. MISRA ET AL.

challenges for state entities. The weak cost recovery mechanisms very often make them
a source of fiscal risk for state budgets (ECB, 2011; FC XII, 2005; FC XIV, 2015), either
implicitly through repayments or explicitly through the takeover of outstanding debt. The
poor financial performance and weak balance sheets of the SPSEs increase the vulner­
ability of state governments towards the invocation of guarantees, a frequent occurrence
in the past under schemes like OTS, FRP, and UDAY. Also, given the lack of transparency
with regard to guarantees, concerns have been raised on its usage and quality of
expenditure.
Certain institutional mechanisms are put in place to safeguard against the excessive
reliance of SPSEs on guarantees and any likely invocations. States impose a guarantee fee,
which varies from 0.5% to 2.0% of guarantees issued, depending upon the quality of the
borrower and states’ fiscal stance, though very often these are waived off. The optimal or
sustainable level of such guarantees is based on annual incremental guarantees as a ratio to
GSDP or total revenue receipts specified in the various states’ Government Guarantees Acts or
fiscal responsibility legislations (FRLs) of states; otherwise administrative limits are fixed, albeit
not all states have such caps/limits. States are required to set up the Guarantee Redemption
Fund (GRF) with the Reserve Bank using the guarantee fees. It is supposed to be used to make
payments emanating from guarantees issued on behalf of state-level bodies as per the
Twelfth Finance Commission (FC-XII) recommendation (RBI, 2019). As of 1 May 2018, this
rule is adhered to only by 12 states, which have established their GRF. Moreover, some states
have inadequate funds. On an aggregate basis, corpus in GRF hovers at less than 1.5% of
outstanding guarantees indicating small buffers if crystallization of guarantees takes place.
Needless to say, there are differences amongst states in terms of awareness about fiscal risk
associated with the issuance of these guarantees and putting an appropriate mechanism in
place to minimize the spill over effect onto state budgets from guarantees.
Nevertheless, the information base on state governments’ guarantees is very weak as
data on guarantees are generally not reported explicitly by state governments in their
budget documents. The Comptroller and Auditor General (CAG) of India, through its state-
wise report of state government finances, releases data on guarantees, though there is
a lag of two years in the reporting of the same. Against this backdrop, the following
analysis relies on historical data from these reports and is supplemented with data
obtained from RBI’s annual publication of state finances.8
The outstanding guarantees of states witnessed a declining trend in the post-FRBM
period, plummeting from 5.4% of GDP at end-March 2006 to 2.0% of GDP at end-March
2017 (Table 2). The decline in guarantees is primarily driven by the guarantees of the
power sector being subsumed in state government liabilities under UDAY, which is
typically reflected in the corresponding rise in debt levels of state governments.
However, in 2017–18, guarantees witnessed a significant jump to 2.5% of GDP (year-on-
year growth of about 38%), indicating early signs of fiscal risks (Chart 5).9 Likewise, the
augmented debt, which is the sum of conventional debt and guarantees, has recorded
a rise and stood close to 28% at end-March 2018.
In terms of sectoral distribution, guarantees provided to power utilities remain domi­
nant – greater than 60% (on average) of total outstanding guarantees. For a few states like
Rajasthan, Uttar Pradesh, Tamil Nadu, Meghalaya, Manipur, and Himachal Pradesh, guar­
antees taken by the power sector account for over 80% of total guarantees followed by
the transport sector (Chart 6).
 MACROECONOMICS AND FINANCE IN EMERGING MARKET ECONOMIES 7

Table 2. Outstanding Guarantees and Augmented Debt of State Governments.

Outstanding Guarantees
Year end- Amount (₹ lakh Annual (as % of Augmented Debt (outstanding guarantees + outstanding
March crore) Growth GDP) liabilities) (as % of GDP)
2006 1.9 −4.9 5.4 37.6
2007 1.9 −1.4 4.5 34.3
2008 1.8 −4.0 3.8 31.5
2009 2.0 7.7 3.6 30.8
2010 2.1 5.5 3.3 29.7
2011 2.1 2.2 2.7 26.7
2012 2.2 5.1 2.6 25.8
2013 3.0 32.8 3.0 25.6
2014 3.8 26.9 3.4 25.7
2015 4.3 13.0 3.4 25.4
2016 3.6 −15.0 2.6 26.3
2017 3.1 −14.4 2.0 27.1
2018 4.3 37.9 2.5 27.6
Source: RBI, State Finances: A Study of Budgets, multiple years.

Chart 5: Guarantees vis-a-vis Debt

2.0 4.0

1.0 2.0

Per cent to GDP

Per cent to GDP

0.0 0.0

-1.0 -2.0
UDAY years
-2.0 -4.0

-3.0 -6.0

-4.0 -8.0
2009-10

2010-11

2011-12

2012-13

2013-14

2014-15

2015-16

2016-17

2017-18

Y-o-Y change in Guarantees-GDP ratio

Y-o-Y change in Debt-GDP ratio
Outstanding Guarantees as per cent of GDP (RHS)

Chart 5. Guarantees vis-a-vis Debt.

Source: RBI, State Finances: A Study of Budgets, multiple years.

On close examination of UDAY, which led to a significant increase in states’ outstand­

ing debt, if a part of the guarantees is invoked in the future, the consequent financial
burden may fall on states’ budgets. As a result, states may have to resort to its financing
via borrowings like UDAY bonds with macro implications. Accordingly, an alternate
variable has been considered in the empirical analysis along with debt, which is the
augmented debt (debt plus guarantees), to explore whether states’ debt is sustainable or
not if the guarantees are invoked.10
8 S. MISRA ET AL.

Sectoral Distribution of Outstanding State-wise Sectoral Distribution of Outstanding

Guarantees Guarantees at end-March 2017
7

6 All States
100.0 Average =
2.3 % of

Per cent to GSDP

90.0 5
All States Power GSDP
80.0 Average = 1.2 % of
70.0
4 GSDP
Share in per cent

60.0 3
50.0
40.0
2
30.0 1
20.0
10.0 0

J&K

Odisha
Karnataka
Assam
Chhattisgarh

Manipur
Mizoram

Tamil Nadu
Tripura
Andhra Pradesh

Haryana

Rajasthan

Uttarakhand
Madhya Pradesh
-

Power Sector Guarantee Transport Sector Guarantee Others Transport

Other Sectors Power All States Average for Power
All States Average

Chart 6. Sectoral Distribution of Outstanding Guarantees.

Note: 1. Sectoral distribution of guarantees is available from 2009 to 2010 from the CAG reports. 2. Other
Sectors include Agriculture, Infrastructure, Service, Financial Corporations, among others. Source: (RBI, 2019).

Theoretical underpinnings and a review of the literature

Theoretical literature review
Debt, by definition, is considered to be sustainable if a country is expected to service its
debt without any large future corrections in the primary balances of the government. It
rules out a situation where the borrower accumulates debt faster than its capacity to
service the existing debt, known as Ponzi finance (Blanchard et al. 1990; Buiter, Persson,
and Minford 1985). Other definitions, particularly post-crisis, by the International
Monetary Fund (IMF) and others focus on the serviceability of debt, i.e., the sustainable
debt of a country suggests its ability to service all accumulated debt at any point in time
(IMF, 2011). Typically, debt sustainability can be explained using an accounting-based
approach associated with an inter-temporal budget constraint as shown below:
Bt ¼ Bt 1 ð1þrt Þ PBt or
(1)
Bt 1 ¼ PBt =ð1þrt ÞþBt =ð1þrt Þ

where states initial public debt Bt-1 equals the present discounted sum of primary balance
(PB) and end of period debt (Bt) with rt being the rate of interest.
Expressing as a ratio to GDP (Yt) and after rearranging we get,
bt ¼ ð1 þ rt Þ=ð1 þ gt Þ bt 1 pbt (2);
where bt = Bt/Yt; bt-1 = Bt-1/Yt-1; gt = (Yt – Yt-1)/Yt-1 and pbt = PBt/Yt
bt bt 1 ¼ ½bt 1 ð1 þ rt Þ= ð1 þ gt Þ bt 1 � pbt (3)
Rearranging we get,
Δ bt ¼ bt 1 ðrt gt Þ= ð1 þ gt Þ pbt (4)
 MACROECONOMICS AND FINANCE IN EMERGING MARKET ECONOMIES 9

The condition for debt sustainability is to place the debt on a stable or declining trajectory
which specifies Δ bt ≤ 0, as per equation 4.11 Fiscal policy is thus unsustainable when gt
= rt and gt < rt as debt grow linearly in the case of the former and explosively in the case
of the latter. Debt is sustainable when gt > rt. This is considered to be one of the necessary
conditions for sustainability (in line with Domar 1944). The other condition is that the
primary balance (pb) has to be positive and large. However, Δ bt ≤ 0, this condition can be
achieved even with primary deficits (pb<0) if the same is offset by a sufficiently large
negative interest-growth differential (rt – gt). Since the government in emerging econo­
mies, including India, usually run primary deficits, they are dependent on a negative
interest-growth differential to contain the debt-GDP ratio. It follows that if India were not
able to maintain its growth rate and if interest rates start rising, then the country’s debt-
GDP ratio could begin to spiral upwards.
For advanced economies, the interest-growth differential is usually close to zero;
further, this differential should not exist in the long run, as is advocated by growth
theories. Thus, if rt – gt is zero, then the primary balance must also be zero or in surplus
to satisfy the condition for debt to be sustainable. These are some of the theoretical
underpinnings which are utilized to assess debt sustainability in the following standard
frameworks.

Empirical literature review

In the empirical framework, the concept of debt sustainability is primarily measured by
two approaches: (i) using the indicators of debt sustainability (Blanchard 1990; Buiter,
Persson, and Minford 1985, 1987; Buiter, Corsetti, and Rubini 1993; Miller 1983), and (ii)
through the model-based approach using econometric methods of government solvency
(Bohn 1998; Hamilton and Glavin 1986; Trehan and Walsh 1988). The existing literature
does not consider one approach to be better than the other; however, it does suggest an
integration of the results from these two approaches in order to provide additional
information on the aspect of debt sustainability (Marini and Piergallini 2008).
The indicator-based approach evaluates the creditworthiness of the borrower with the
help of indicators as derived from the inter-temporal budget constraint (equation 4), viz.
(a) whether the real rate of interest minus real GDP growth is negative and/or (b) whether
there is a primary surplus or deficit. This approach has been further extended by adding
more indicators that reflect growth, liquidity, and serviceability of debt such as the ratio of
interest payment to current revenue, the difference between the rate of growth of debt,
and the rate of growth of nominal GDP, among others. While some of the early papers on
debt sustainability for India have used this approach exclusively, the latest ones have used
it along with other approaches (Dholakia, Mohan, and Karan 2004; Kaur, Mukherjee, and
Ekka 2018; Misra and Khundrakpam 2009; Mourya 2015; Pattnaik, Misra, and Prakash 2003;
Rajaraman, Bhide, and Pattnaik 2005).
While the indicator-based approach is largely forward-looking, albeit static, the empiri­
cal technique using an econometric model based on historical time-series data is con­
sidered largely backward-looking, though dynamic by nature. The model-based approach
has also different variants that have evolved over time. The unit root test was used
extensively to check debt sustainability in the United States (US), first by Hamilton and
Flavin (1986) and later by Trehan and Walsh (1988), who used it along with the
10 S. MISRA ET AL.

cointegration approach. In the Indian case, studies by Buiter and Patel (1992) and Pradhan
(2014) used the unit root approach, while Jha and Sharma (2004) and Tronzano (2013)
relied on the cointegration approach.
Given the limitations of the aforementioned model-based approaches, the Bohn
approach, which estimates the fiscal policy response function, emerged as an alternative
and has been extensively used to assess the sustainability of public debt policies amongst
different countries (Ostry and Abiad 2005; Adams, Ferrarini, and Park 2010; Greiner and
Kauermann 2008; Haber and Neck 2006; Kaur, Mukherjee, and Ekka 2018; Renjith and
Shanmugam 2018). The economic intuition behind the Bohn approach is If governments
run debt today, how will primary balances be impacted to assess whether public debt is
sustainable or not? Many studies have used this approach to explore the issue of debt
sustainability either in a panel framework or time series; however, results are subject to
the underlying assumptions and analytical techniques used to perform the analysis.
A panel estimation study by Adams, Ferrarini, and Park (2010), which combined countries
of Central Asia, East Asia, the Pacific, Southeast Asia, and South Asia for a relatively small
period, indicated public finances of developing Asia to be in good shape.
In the context of India, while studies have conventionally analysed debt sustainability
for the Centre and general government (Centre plus states) (Hakhu 2015; Mohanty and
Panda 2019; Rajaraman and Mukhopadhaya 2004; Rangarajan, Basu, and Jadhav 1989;
Seshan 1987) and both for domestic and external debt (Cashin and Olekalns 2000), the
existing empirical literature on states’ debt sustainability remains sparse with mixed
results. Early studies by Tiwari (2012) and Misra and Khundrakpam (2009) which cover
the initial years of the post-FRBM period when the debt was at a relatively higher level
found evidence of debt unsustainability in the long run, while studies such as Renjith and
Shanmugam (2018) and Kaur, Mukherjee, and Ekka (2018) covering the later time period
observed debt to be sustainable. Regardless of the time period, results amongst studies
vary on based on the estimation method adopted. Summary results of select empirical
studies dealing with this research question are presented in Table 3.
A review of the Indian literature available on public debt sustainability clearly reveals
that most studies on the subject have limited their analyses to conventional debt
sustainability and have not assessed the potential risks from contingent liability realiza­
tions. Given the fact that a significant portion of the guarantees accrues to the financially
underperforming state power-sector utilities, the prospects of their crystallization are
higher and therefore the consequent financial burden may fall on states’ budgets. It
should be noted that this issue has been highlighted by other studies as a potential
research gap (European Central Bank (ECB) 2011; Kaur, Mukherjee, and Ekka 2018).12
Contingent liability realizations have been ascribed to be a more important cause than
an addition to the primary deficit behind debt spikes across countries (Jaramillo, Mulas-
Granados, and Kimani 2017).
Against this backdrop, this study attempts to address this research gap specifically in
the Indian context by including the post-UDAY period (2015–17) when debt swelled up
sizably; incorporating a larger number of states, thereby improving spatial coverage vis-à-
vis existing studies; and, lastly, by adding the contingent liabilities in the form of state
government guarantees into the conventional debt analysis.
 MACROECONOMICS AND FINANCE IN EMERGING MARKET ECONOMIES 11

Table 3. Empirical Studies on Subnational Debt: India and EMEs.

Time Period Is Debt
Study Country (Annual) Methodology Sustainable?
Adams, Ferrarini, and Developing 1990–-2008 FGLS and SGMM Sustainable
Park (2010) Asia (30)
Ostry and Abiad (2005) EMEs (31) 1990–2002 FGLS and FE Sustainable
Bohn (1998) US 1916–1995 OLS Sustainable
Magazzino and Italy 1862–2013 Wavelet analysis Sustainable
Mutascu (2019)
Mahdavi (2014) 48 US states 1961–2008 Panel Fixed Effects Sustainable
Renjith and India 2004–05 to FE Sustainable
Shanmugam (2018) 2014–15
Kaur, Mukherjee, and India 1980–1981 to GLS SUR model Sustainable
Ekka (2018) 2015–2016
Shastri and Sehrawat India 1980–2013 OLS Not Sustainable
(2015)
Tiwari (2012) India 1970–2009 p-spline Not Sustainable
Misra and India 1991–1992 to Inter-temporal budget constraint Not Sustainable
Khundrakpam 2007–2008 approach (unit root test)
(2009)
Note: FGLS is feasible generalized least squares; SGMM is structural generalized methods of moments; FE is fixed effects;
OLS is ordinary least squares; GLS SUR is generalized least squares seemingly unrelated regression.
Source: Authors’ compilation.

Methodology and data

Methodology
On the basis of the foregoing literature review, the objective of this study is to analyse the
debt sustainability of state governments. Towards this endeavour, the empirical analysis
in this paper is based on three stages. First, a forward-looking indicators-based approach
is used to provide a quick assessment of the ability of state governments to service as well
as to repay their debts through current and regular sources of income.13 Second, the
order of integration of all the variables is checked using panel unit root tests. Third, the
fiscal policy response function for state governments is estimated in a panel framework
using a standardized Bohn approach (Bohn 1998).
The indicator-based approach considers the creditworthiness of the borrower by
analysing ratios. As per the necessary Domar stability condition, the real growth rate of
the economy should be higher than the real interest rate, while the sufficient condition
states that primary balances should be zero or in surplus to service the future debt
(drawing from equation 4). Along with these debt servicing indicators, the ratio of interest
payment to revenue receipts should be declining and less than 10% (Fourteenth Finance
Commission [FC-XIV]) and debt should be growing at a rate lower than GDP growth for it
to be sustainable.

Fiscal policy response function (Bohn approach)

A basic fiscal policy response function adopted from the Bohn approach can be illustrated
as follows:

Pt ¼ α þ βDt 1 þγ`Xt þεt (5)

12 S. MISRA ET AL.

where P is the primary balance-to-GDP in year t; D t-1 is debt stock in t-114 and X denotes
control variables, viz. real GDP gap, revenue receipts, primary expenditure, etc. ‘β’ is the
principal coefficient which measures the response of the primary balance to variations in
debt. If a rising debt-to-GDP ratio leads to a rise in primary balance, then debt tends to be
sustainable.
Equation (5) has been estimated using a fixed-effect model and/or random-effects
model. The former allows for all unobserved factors (varying across states though remain­
ing constant over time) to be correlated with the explanatory variables, while the latter
assumes that they are uncorrelated with the explanatory variables and so are included in
the error term (Gupta and Ahmed 2018). The Hausman specification test has been applied
to decide upon the fixed effects/random effects model wherein the null says there is no
significant difference between these two estimations.15
In addition, regressions are carried out with Feasible Generalized Least Squares (FGLS)
to further check the robustness of empirical results (Ostry and Abiad 2005; Adams,
Ferrarini, and Park 2010). The classic OLS regression assumes independent identically
distributed (i.i.d) error and independent autocorrelation structure. But in many cross-
sectional data sets, the variance for each of the panels differs. Recognizing that this study
uses data from heterogeneous states that may have variations of scale, ruling out
heteroscedasticity could be important. This problem of heteroscedasticity can be coun­
tered if the variance-covariance matrix of the error term is known. There are typically two
ways to address this issue. One can either (a) assume the structure of the variance-
covariance matrix which is known as weighted least squares (WLS) estimation or/and
(b) estimate it empirically from ordinary least squares (OLS) rather than assuming, referred
to as FGLS estimation. Furthermore, FGLS estimation has an advantage that it can also
correct for cross-sectional correlation and autocorrelation, if any occurs (Greene 2012).

Data
The study uses annual time series from 1992–92 to 2018–19 to perform the indicator-
based analysis for all states, while a panel analysis is conducted for the period 2004–05 to
2017–18 for 26 states. The choice of states and time period in the panel data has been
determined by the availability of guarantees data, though we have tried to remain as
comprehensive as possible. Since data points for the variable outstanding guarantees
were not available for two states, viz. Goa, and Jharkhand, augmented debt could not be
obtained. Also, with the Telangana region being merged with Andhra Pradesh, few data
points were available for Telangana’s guarantees data. Consequently, 26 states are
considered over a period of 14 years. As debt stock is taken with a lag of one year, the
number of observations in this analysis is 338 (26 × 13).
Data on various fiscal variables are extracted from State Finances: A Study of Budgets of
2019–20, RBI, while state-wise GSDP (real and nominal) is from the Ministry of Statistics
and Programme Implementation. We have used two specifications of the standard fiscal
policy response function, discussed below:
 MACROECONOMICS AND FINANCE IN EMERGING MARKET ECONOMIES 13

Specification 1

PBi;t ¼ αi þβDi;t 1 þλOutputgapi;t þδRRi;t þεi;t (6)

Specification 2

PBi;t ¼ αi þβADi;t 1 þλOutputgapi;t þδRRi;t þεi;t (7)

where PB represents primary balance to GSDP ratio; D is debt to GSDP ratio; Output gap
signifies the difference between the real output and the potential real output; RR
represents revenue receipts to GSDP ratio; ε is error term; and AD is augmented debt to
GSDP ratio.
This study analyses two variants of the principal explanatory variable debt wherein
Specification 1 (equation 6) assesses the sustainability of conventional debt ‘D’, while
Specification 2 (equation 7) is beyond estimating conventional debt sustainability and so
incorporates off-budget liabilities in the form of outstanding guarantees issued to SPSEs.
This new principal variable is referred to as augmented debt ‘AD’ in this study. Further,
under Specification 2, two scenarios are considered based on how much of guarantees
could be risky. As pointed out in the 2004 RBI paper, a realistic scenario could be to take
a certain proportion. However, the lack of transparency makes assessing fiscal risk due to
guarantees extremely challenging necessitating the use of certain proportion based on
reasonable parameters (Reserve Bank of India (RBI) 2002). Accordingly, as a rule of thumb,
converting guarantees into debt using a factor of 1/3 was adopted in some earlier papers
(Thorat and Roy 2004).
The RBI committee Report (2004) clearly mentions that the highest amount of default is
in power sector, necessitating a need to monitor power-sector guarantees closely. Since
then, however, a lot of structural changes have happened. Share of power-sector guar­
antees has risen from about 40% to 70%, on an average, with some states at 80%. Under
such a situation, adding the share of power-sector guarantees (70%) to conventional debt
to get the augmented debt is the most apt to highlight the fiscal risks coming from this
angle. The likelihood of power-sector guarantees getting crystallized is probable from
both precedence and performance point of view. During the last 13 years, on three
occasions, the Government has tried to support power-sector SPSEs. Under UDAY, almost
the majority of power-sector debt was added to states’ debt and the impact is still being
felt by many (Reserve Bank of India (RBI) 2019a). In terms of performance, most of them
are still running losses, with talk of another restructuring required. In such a situation, it
would be prudent to add the power-sector guarantees to conventional debt in order to
arrive at the augmented debt and to highlight fiscal risks. While this remain one scenario,
the alternate scenario which assumes the proportion at 100% represents an extreme
situation – as not all guarantees will be invoked at the same time – keeping this scenario
helps to understand the maximum fiscal risks posed by these off-budget liabilities. This is
important in the context of the global debate on adopting a public sector balance sheet
approach (International Monetary Fund (IMF) 2018) and also considering the fact that
some Latin American countries have started including state-owned enterprises in their
fiscal targets.
14 S. MISRA ET AL.

The coefficient ‘β’ is the key to estimating the response of primary balance to debt as
per Specification 1 and, likewise, augmented debt as per Specification 2. A positive
coefficient lying between zero and one depicts the sustainability of public debt. While
a negative β coefficient indicates the unsustainability of public debt. Two control vari­
ables, namely the real output gap and revenue receipts (RR), are employed in line with
Ostry and Abiad (2005).
The real output gap is measured as the difference between the actual real GSDP and
the potential real GSDP wherein the latter is estimated using the Hodrick–Prescott (HP)
filter. A positive coefficient of the output gap (actual output is above the potential output)
indicates that primary balance improves when real GSDP is above the potential and
therefore surplus rises in response to cyclical effects on revenues and expenditures.
The other control variable RR allows us to distinguish fiscal structures amongst states.
Given the weak revenue-raising capacity amongst the majority of Indian states, more so
after the implementation of GST, the use of revenue receipts to GSDP can be considered
as a direct proxy for states’ surplus generating capacity, particularly considering that few
states have higher revenue-generating capacities than others. Hence, RR is illustrative of
robust fiscal capability.
Furthermore, to underpin the empirical results, we conducted a robustness check by
using the revenue receipts gap and the average of RR to GSDP ratio over the previous two
years, rather than contemporaneous revenue receipts to GSDP ratio. Further, the other
control variable – primary expenditure gap – which is the difference between the actual
primary expenditure16 and the trend primary expenditure capturing deviations in primary
expenditure from long-term trend, is also considered. Higher primary expenditure from
the trend signifies deterioration in primary balance.

Empirical results
Indicator-based approach
A quick assessment of the ability of state governments to service and repay the debt
through current sources of income is conducted using an indicator-based approach. By
dividing the entire post-reform period 1992–93 to 2018–19 into different phases based on
the fiscal performance of states and by analysing the debt sustainability indicators of all
the states together, we found that the real rate of interest is lower than the real GDP
growth rate in all phases, thus fulfiling the necessary condition for debt sustainability
(Table 4).17 However, the GDP growth-interest rate differential has narrowed down in the
last five years, necessitating the need for larger and faster corrections in primary deficits
for debt to be in a sustainable path (Chart 7). Furthermore, the primary balance has
remained consistently negative (implying primary deficits) throughout all phases (except
Phase III), consequently violating the sufficient condition of debt sustainability (Table 4).
The last phase (2015–16 to 2018–19), coinciding with the issuance of UDAY bonds,
witnessed the highest primary deficit in the post-FRBM period.
The declining (r-g) coupled with a rising primary deficit hint at the fast-shrinking
window for debt sustainability. The next indicator, i.e., interest payments to revenue
receipts ratio, is essentially a debt servicing indicator reflecting how much of the revenue
receipts go towards financing interest payments. Notwithstanding a decline in this ratio
 MACROECONOMICS AND FINANCE IN EMERGING MARKET ECONOMIES 15

Table 4. Debt Sustainability of State Governments: Indicator-based Analysis.

Phase I Phase II Phase III Phase IV Phase V Phase VI
(1992–93 to (1997–98 to 2004–05 to 2008–09 to 2012–13 to 2015–16 to
Indicator 1996–97) 2003–04) 2007–08) 2011–12 2014–15 2018–19
1 2 3 4 5 6 7
r*- g < 0 −6.1 −1.0 −5.1 −10.1 −7.6 −4.4
PB/GDP ≥ 0 −0.8 −1.6 0.0 −0.6 −0.7 −1.3
IP/RR ↓↓ 15.6 22.4 19.1 13.8 12.1 11.9
D-G < 0 −1.7 7.6 −4.8 −5.0 −2.0 3.4
Notes: 1. r is the real rate of interest; g is the real output growth; PB is primary balance; IP is interest payments; RR is
revenue receipts; D stands for the growth rate of public debt; and G pertains to the growth rate of nominal GDP.
2. *Nominal interest rate is calculated as a ratio of interest payment at t to debt at t-1. CPI (IW) is used to derive real
interest rates from the nominal interest rates.
Source: (Reserve Bank of India (RBI) 2019a).

Rising debt to
GDP and
narrowing gap
16.0 between interest 35.0
14.0 rate and growth 30.0
12.0 rate indicating

Per cent to GDP

10.0 early signs of 25.0
fiscal risk.
Per cent

8.0 20.0
6.0
4.0 15.0
2.0 10.0
0.0
5.0
1992-93
1993-94
1994-95
1995-96
1996-97
1997-98
1998-99
1999-2000
2000-01
2001-02
2002-03

2005-06
2006-07
2007-08
2008-09
2009-10
2010-11
2011-12
2012-13
2013-14
2014-15
2015-16
2016-17
2017-18
2003-04
2004-05

-2.0
-4.0 0.0

Primary Deficit to GDP ratio g-r Debt to GDP ratio (RHS)

Chart 7. Sustainability of Debt.

Source: e-STATES Database (Reserve Bank of India (RBI) 2019a).

over the phases, the ratio is still higher than the tolerable limit of 10% as prescribed by the
Fourteenth Finance Commission (FC-XIV). Lastly, while debt was growing at a rate lower
than the nominal GDP growth in the post-FRL phases, in the latest post-UDAY phase, the
rate of growth of debt stood higher than the growth rate of nominal output signalling
potential debt sustainability risks. The ambivalence in this indicator-based approach of
debt sustainability demands further investigation of the solvency of state governments.
A quick comparison with previous such analysis (Kaur, Mukherjee, and Ekka 2018) reveals
deterioration in most of the indicators in the latest phase (Phase VI), which primarily looks
to be a fallout of realizations of contingent liabilities under UDAY.

Model-based econometric approach

Before opting for the panel regression approach, we investigated the time-series proper­
ties of all the variables, viz. primary balance to GDP ratio, debt to GDP ratio, augmented
16 S. MISRA ET AL.

debt to GDP ratio, real output gap, and revenue receipts. Different methods of panel unit
root tests were applied in our investigation. The Levin, Lin and Chu (LLC) panel unit root
test assumes a common unit root across cross-sections, while the Im–Pesaran–Shin (IPS)
and Fisher–ADF assume individual unit root processes. Here, the null signifies the pre­
sence of a unit root at level, while its rejection ascertains that the data series is stationary.
The results of all the panel unit root tests are presented in Table 5, and these three tests
unanimously suggest that all the variables are stationary at level.
After ensuring the stationary properties of data, the fiscal policy response function is
measured using the panel estimation methods. The Hausman specification test validates
the random-effects model. A diagnostic check revealed the presence of heteroscedasti­
city, though the model did not suffer from autocorrelation and cross-section dependence
of residuals, something typical of macro panels with long time series. Consequently, the
FGLS model is also applied as it deals with heteroscedasticity. Empirical results of alternate
models described above are summarized in Table 6.18
Model 1 (that uses random effects) gives a negative and statistically significant β
coefficient at 1% level, while Model 2 (that corrects for heteroscedasticity using FGLS)
offers a statistically insignificant β coefficient, thus rejecting the null of unsustainability of
states’ debt. These results may signify weak evidence with regard to the sustainability of
states’ debt. However, with the inclusion of all guarantees (referred to as augmented debt
under Scenario 1), debt clearly moves into the unsustainable path as shown under both
Model 3 and Model 4. This suggests that the primary balances in India respond negatively
to increases in augmented debt, indicating that the debt sustainability condition for an
intertemporal budget constraint is not satisfied. We have also considered a slightly more
likely Scenario 2 (see Models 5 and 6), when a part of guarantees, particularly those in the
power sector, is invoked, which also further confirms the likely unsustainability of states’
debt. Thus, this paper supports the argument that primary fiscal balances of subnational
governments in India respond in a destabilizing manner if the debt is considered in the
augmented form (inclusive of guarantees).19
The control variable (revenue receipts) is correctly positively signed and statistically
significant at 1% level, signifying that higher revenue generation helps to improve
primary balances. This also allows to distinguish fiscal structures amongst states, as
some states have higher revenue-generating capacities compared with others. Further,
a positive and significant coefficient of output gap indicates that primary balances

Table 5. Panel Unit Root Tests.

Variables (Levels) LLC t-Statistics IPS W-Statistics Fisher–ADF Z-Statistics
Primary Balance/GSDP −7.08*** (0.00) −5.62*** (0.00) −5.23*** (0.00)
Debt/GSDP −6.59*** (0.00) −1.89** (0.02) −2.01** (0.02)
Augmented Debt/ GSDP −9.34*** (0.00) −4.43*** (0.00) −4.23*** (0.00)
Real GSDP Gap −3.90*** (0.00) −4.29*** (0.00) −3.90*** (0.00)
Revenue Receipts −6.25*** (0.00) −2.16*** (0.01) −1.73** (0.03)
Notes: 1. P-values are mentioned in the parentheses. ***The rejection of the null hypothesis of non-stationary at 1%
level of significance; ** rejection of null of non-stationary at 5% level of significance.
2. Probabilities for Fisher tests are computed using an asymptotic Chi-square distribution. All other tests assume
asymptotic normality. Probability 0.00 implies that probability tends to zero.
3. Automatic selection of lags through Schwarz Information Criterion (SIC). All panel unit root tests are defined by
the Bartlett kernel.
Source: Authors’ calculations.
Table 6. Econometric Estimation Results, 2004–05 to 2017–18 Dependent Variable: Primary Balance as a Proportion to GSDP.
Scenario 1 Scenario 2
Model 1 Model 2 Model 3 Model 4 Model 5 Model 6
Regressor RE FGLS RE FGLS RE FGLS
Dt-1 (β coefficient) −0.05*** −0.02
(0.00) (0.12)
ADt-1 (Scenario 1)3 −0.06*** −0.05***
(0.00) (0.00)
ADt-1 (Scenario 2)3 −0.05*** −0.04*** (0.00)
(0.00)
RR 0.06*** 0.06*** 0.06*** 0.07*** 0.06*** 0.07***
(0.01) (0.00) (0.00) (0.00) (0.00) (0.00)
Outputgap6 0.05* 0.04* 0.05* (0.07) 0.05** 0.06* (0.07) 0.05**
(0.10) (0.06) (0.03) (0.03)
Constant −0.04 −0.58 0.10 −1.25 0.14 (0.87) 0.56
(0.9) (0.6) (0.9) (0.5) (0.63)
Wald chi2 107*** 302*** 108*** (0.00) 320*** 108*** 313*** (0.00)
(0.00) (0.00) (0.00) (0.00)
Hausman (chi2) 1.11 1.63 1.96 (0.58)
(1.00) (1.00)
No. of Observations 338 338 338 338 338 338
Notes: 1. Figures in parentheses are p-values; ***, **,* indicate significance at 1%, 5% and 10% levels, respectively.
2. Augmented debt is obtained after adding outstanding guarantees to the outstanding liabilities of state governments. One-year lag of debt and augmented debt is taken to surmount the
problem of endogeneity.
3. Scenario 1 assumes all guarantees are invoked, while Scenario 2 assumes financially stressed power sector guarantees are crystallized.
4. In the RE model, cross-section and time-effects are taken into account.
5. FGLS allows for heteroscedastic correlated error structure and using an AR (1) autocorrelation structure. Regressors include state dummies to allow for panel-level fixed effects (coefficients not
reported here). Time dummies are also included.
6. Real Output Gap = Actual Output – Potential Output
Source: Authors’ calculations.
MACROECONOMICS AND FINANCE IN EMERGING MARKET ECONOMIES
17
18 S. MISRA ET AL.

improve in response to cyclical effects, particularly due to the impact on states’ own
revenues and central transfers.
These empirical results are broadly robust to various alternative specifications, as
provided in Appendix Table A1. Alternate specifications are based on different control
variables such as revenue receipts gap, primary expenditure gap and average revenue
receipts with debt and augmented debt (obtained by adding only power sector out­
standing guarantees, about 70% of the total, to outstanding liabilities, which is a more
likely scenario) as prime independent variables under both RE and FGLS model specifica­
tions. It is observed that broadly the sign and significance of debt and augmented debt
are retained. In general, as we move from debt to augmented debt liabilities, the β
coefficient indicates that debt moves from sustainable to unsustainable ranges or, if
already unsustainable, the strength and significance of the coefficient increases. In the
case of a control variable, the output gap loses its significance. These results are robust to
the inclusion of time and fixed effects in the regression. These robustness checks further
buttress the empirical results of this study and hence reconfirm the augmented debt
unsustainability concerns of the states in the long run.

Conclusion and policy implications

The debt position of state governments, after remaining in the comfort zone for about
a decade post-FRL, has started showing early signs of unsustainability in the post-UDAY
period. Apart from the conventional liabilities of state governments, fiscal costs can also
stem from state governments’ risk exposure to their public sector in the form of guaran­
tees, off-budget borrowings, and accumulated losses of financially distressed PSEs. The
empirical analysis of debt sustainability in this paper, using both indicators as well as
Bohn’s (1998) panel data approach, indicates that guarantees given by states, if invoked,
can turn out to be a potential risk to debt sustainability for Indian states.
With regard to policy implications on the fiscal side, it may be pertinent to review the
FRL adopted by the states in the 2000s, much the same way the Centre revised their FRBM
in the 2018–19 Union Budget whereby debt was explicitly added as a target variable
along with fiscal deficit. This is particularly desirable considering the fact that while the
fiscal deficit of states remains well within their FRL threshold of 3% for GFD to GDP ratio,
as adopted during the early 2000s, the debt is rising at a high pace, crossing correspond­
ing implicit thresholds. Besides, no fiscal rule is static and the cross-country evidence
suggests that fiscal rules may entail improvisation contingent upon experience. Laying
down a transparent institutional mechanism may ensure that states meet their debt-
deficit targets.
In the context of the SPSEs, it may be important to (i) periodically review the perfor­
mance of SPSEs to examine whether they can deliver value for the taxpayer’s money
(International Monetary Fund (IMF) 2018); (ii) get their pricing policy right so that they are
able to ensure cost recovery through appropriate user charges/tariffs, etc. (Reserve Bank
of India (RBI) 2019a); and (iii) ensure professional and transparent operation of these units.
Measures must be adopted to ensure a comprehensive framework for guarantees man­
agement which would include elements like adherence to caps/limits based on sustain­
ability and assessing its effectiveness for the states who have adopted them; providing
transparent data on guarantees extended by states and the consequent invoking/waiving
 MACROECONOMICS AND FINANCE IN EMERGING MARKET ECONOMIES 19

off of unpaid subsidies, if any, by the state governments; and, charging guarantee fees to
avoid any kind of moral hazard on the one hand, and also to add to revenues of the states,
on the other.

Notes
1. States’ market borrowings have more than doubled in the last 5 years.
2. Under the OTS of 2003, the outstanding dues of the State Electricity Boards to the central
public sector undertakings were taken over by the state governments through the issuance
of power bonds (with SLR status) amounting to ₹29,606 crores. The FRP 2012 restructured the
debt of 7 state power distribution companies (DISCOMs) by taking over 50% of their out­
standing short-term debt obligations up to 31 March 2012. Under UDAY, the states took over
75% of the DISCOMs’ debt, through the issuance of non-SLR UDAY bonds and transferred the
proceeds to DISCOMs. A total of 16 states issued bonds under UDAY in 2015–16 and 2016–17,
leading to a sharp rise in these states’ debt, interest payments and redemption pressures.
3. Due to data constraints on guarantees, the analysis in the paper has been restricted until 2017–18.
4. Since the Fourteenth Finance Commission (FC-XIV) found the differentiation between special
category state and non-special category state redundant, we have considered all states in our
study.
5. In this study, subnational debt refers to the debt of the state governments only, local bodies’
debt is not included due to its unavailability.
6. BRICS is the acronym coined for an association of five major emerging national economies:
Brazil, Russia, India, China, and South Africa.
7. Article 293 of the Indian Constitution puts a restriction on overall borrowings by states that
are sanctioned by the Centre till the time the previous loan provided by the central govern­
ment is outstanding. Additionally, SPSEs are allowed to raise off-budget borrowings along
with guarantees which are provided by states under Article 293 (1) of the Constitution.
8. A state-wise time series on guarantees outstanding can be culled out from RBI’s annual
publication State Finances: A Study of Budgets of 2019–20 (Reserve Bank of India (RBI) 2019a)
under Statement 28.
9. At end-March 2019, guarantees’ data are available for only 11 states and amounted to 0.8% of
GDP. Using the same states’ data for the previous year, outstanding guarantees amounted to
0.6% of GDP, clearly exhibiting a rise in guarantees issued in 2018–19.
10. It is clarified that once guarantees fall on states debt, to that extent guarantees get reduced.
For example, state guarantees which increased significantly prior to 2014, fell sharply there­
after, primarily driven by subsuming of 75% of powers sector guarantees into state govern­
ment liabilities as part of UDAY scheme. After a decline for few years, again guarantees, more
so in power sector, started rising from 2017–18 reflecting fresh net accretion of guarantees
with additional fiscal risks.
11. In practise, equation (4) can also be modified toΔ bt = bt-1(rt – gt)/ (1 + gt) – pbt + ddatwhere
ddat refers to deficit-debt adjustment as a share of GDP comprising factors that affect debt but are
not included in budget balance (Bouabdullah et al. 2017; Jaramillo, Mulas-Granados, and Kimani
2017). With ddat = 0, stability conditions as stated above remain the same.
12. It may be noted that Buiter and Patel (1992) is the only previous study that has assessed the
debt sustainability for Centre, states, and PSUs together. It finds the discounted debt series –
using various alternative measures of interest rates – to be non-stationary and hence
unsustainable. The study though remains significantly dated.
13. The current sources of revenue should not include any temporary revenues/grants or capital
receipts emanating from the sale of assets.
14. To avoid the endogeneity issue, Dt is replaced by Dt-1.
15. The P-value of chi-square statistic being greater than 5%, any deviation is due to chance
alone 5% of the time or less and hence the random effect model is appropriate.
16. Primary Expenditure = Total Expenditure – Interest Payments.
20 S. MISRA ET AL.

17. The phase-wise break-up is based on trends in key fiscal parameters.

18. While FGLS is asymptotically more efficient than OLS for a large sample, for a small/medium-
sized sample it can be less efficient than OLS, although coefficients are unbiased allowing for
heteroscedasticity and autocorrelation. So, it is better to present both OLS and FGLS. FGLS
might also be inconsistent if there are individual specific fixed effects, although, in this case,
that is ruled out through the Hausman test.
19. Though different from Kaur, Mukherjee, and Ekka (2018), who observe states’ debt to be
sustainable, the results are not strictly comparable given the broader debt definition used in
the paper with expanded spatial and temporal coverage for states.

Acknowledgements
Authors are thankful to the anonymous referees and the editor(s) of the Journal of Macroeconomics
and Finance in Emerging Markets Economies for their constructive and insightful comments in
improving the quality of the paper. The views expressed in the paper are those of the authors and
not the institutions to which they belong. The usual disclaimer applies.

Disclosure statement
No potential conflict of interest was reported by the author(s).

Notes on contributors
Sangita Misra is Director in the State Finances Division of Department of Economic Policy and
Research of the Reserve Bank of India. She has been working with the Reserve Bank since 2000.
During this tenure, she has worked in various departments of the Reserve Bank including Monetary
Policy and International Department and has wide experience in the areas of fiscal policy, monetary
policy making, international diplomacy and finance, external sector, real sector and structural issues.
She has done her Masters in Economics from Delhi School of Economics and Phd from IIT, Bombay.
Prior to joining the Reserve Bank, she was a lecturer in Delhi University in Economics.
Kirti Gupta is a Manager (Research) in the Department of Economic and Policy Research (DEPR),
Reserve Bank of India (RBI). She received her PhD from the Department of Economics, Jamia Milia
Islamia, New Delhi. She received her postgraduate in Economics from the Centre for Economic
Studies and Planning (CESP), Jawaharlal Nehru University (JNU), New Delhi. She was previously a
Guest Lecturer at the Delhi University. She has worked as a researcher at Indian Council for Research
on International Economic Relations (ICRIER), a premier policy think tank. She has published her
research in reputed national and international journals.
Pushpa Trivedi is professor at IIT Mumbai in the Humanities and Social Sciences Department for last
three decades and has guided several PhD students in the field of macroeconomics.

References
Adams, C., B. Ferrarini, and D. Park. 2010. Fiscal sustainability in developing Asia. ADB Economics
Working Paper Series, No. 205.
Blanchard, O., J.C. Chouraqui, R.P. Hagemann, and N. Sartor. 1990. The sustainability of fiscal policy:
New answers to old questions. OECD Economic Studies, no. 15
Blanchard, O.J. 1990. Suggestions for a new set of fiscal indicators. OECD Economics Department
Working Paper No. 79.
Bohn, H. 1998. The behavior of U.S. public debt and deficits. Quarterly Journal of Economics 113:
949–63. doi:10.1162/003355398555793.
 MACROECONOMICS AND FINANCE IN EMERGING MARKET ECONOMIES 21

Bouabdullah, O., C.C. Westphal, T. Warmedinger, R. de Stefani, F. Drudi, R. Setzer, and A. Westphal
2017. Debt sustainability analysis for Euro Area Sovereigns: A methodological Framework. ECB
Occasional Paper No. 185, April.
Buiter, W.H., G. Corsetti, and N. Rubini. 1993. Excessive deficits: Sense and nonsense in the Treaty of
Maastricht. Economic Policy 8, no. 16: 57–100. doi:10.2307/1344568.
Buiter, W.H., T. Persson, and P. Minford. 1985. A Guide to public sector debt and deficits. Economic
Policy 1, no. 1: 13–79. doi:10.2307/1344612.
Buiter, W.H., T. Persson, and P. Minford. 1987. The current global economic situation, outlook and
policy options, with special emphasis on fiscal policy issues. CEPR Discussion Papers No. 210, CEPR
Discussion Papers.
Buiter, W.H., and U.R. Patel. 1992. Debt, deficits and inflation: An application to the public finances of
India. Journal of Public Economics 47, no. 2: 171–205. doi:10.1016/0047-2727(92)90047-J.
Cashin, P., and N. Olekalns. 2000. An examination of the sustainability of Indian fiscal policy.
Department of Economics - Working Papers Series No. 748, The University of Melbourne.
Chakraborty, P., and B.B. Dash. 2017. Fiscal reforms, fiscal rule, and development spending: How Indian
states have performed. Public Budgeting & Finance 37, no. 4: 111–33. doi:10.1111/pbaf.12161.
Comptroller and Auditor General (CAG). 2019. Report of the Comptroller and Auditor General of India
on compliance of the Fiscal Responsibility and Budget Management Act, 2003 for the year 2016–17,
Report No. 20 of 2018, January, Union Government (Civil), Department of Economic Affairs
(Ministry of Finance). https://cag.gov.in/sites/default/files/audit_report_files/Report_No_20_of_
2018_Compliance_of_the_Fiscal_Responsibility_and_Budget_Management_Act_2003_
Department_of_Economic_Affairs_Minis.pdf .
Dholakia, R.H., T.T.R. Mohan, and N. Karan. 2004. Fiscal sustainability of debt of states. New Delhi:
Study sponsored by the Twelfth Finance Commission.
Domar, E.D. 1944. The “burden of the debt” and the national income. American Economic Review 34,
no. 4: 798–827.
European Central Bank (ECB). 2011. Ensuring fiscal sustainability in the euro area. Monthly Bulletin,
April, ECB, Frankfurt am Main. https://www.ecb.europa.eu/pub/pdf/mobu/mb201104en.pdf .
Fourteenth Finance Commission. 2015. Report of the Fourteenth Finance Commission. Ministry of
Finance.
Greene, W.H. 2012. Econometric analysis. 7th ed. Upper Saddle River, NJ: Prentice Hall.
Greiner, A., and G. Kauermann. 2008. Debt policy in euro area countries: Evidence for Germany and
Italy using penalized spline smoothing. Economic Modelling 25, no. 6: 1144–54. doi:10.1016/j.
econmod.2008.02.006.
Gupta, K., and S. Ahmed. 2018. Determinants of FDI in South Asia: Does corruption matter? International
Journal of Economics and Business Research 16, no. 2: 137–61. doi:10.1504/IJEBR.2018.094009.
Haber, G., and R. Neck. 2006. Sustainability of Austrian public debt: A political economy perspective.
Empirica 33, no. 2–3: 141–54. doi:10.1007/s10663-006-9012-1.
Hakhu, A. 2015. Productive public expenditure and debt dynamics: An error correction representation using
Indian data. Working Papers No. 15/149, National Institute of Public Finance and Policy, New Delhi.
Hamilton, J., and M.A. Flavin. 1986. On the limitations of government borrowing: A framework for
empirical testing. American Economic Review 76, no. 4: 808–19.
International Monetary Fund (IMF). 2011. Modernizing the framework for fiscal policy and public
debt sustainability analysis. Public Information Notice No. 11/118 September 12.
International Monetary Fund (IMF). 2014. Fiscal Monitor, April 2014: Public Expenditure Reform:
Making Difficult Choices, IMF Fiscal Affair Department. https://www.imf.org/en/Publications/FM/
Issues/2016/12/31/Public-Expenditure-Reform-Making-Difficult-Choices-41121 .
International Monetary Fund (IMF). 2018. Fiscal Monitor, October 2018: Managing Public Wealth.
https://www.imf.org/en/Publications/FM/Issues/2018/10/04/fiscal-monitor-october-2018 .
Jaramillo, L., C. Mulas-Granados, and E. Kimani. 2017. Debt spikes and stock flow adjustments: Emerging
economies in perspective. Journal of Economics and Business 94: 1–14. doi:10.1016/j.jeconbus.2017.08.003.
Jha, R., and A. Sharma. 2004. Structural break and unit roots: A further test of the sustainability of the
Indian fiscal deficits. Public Finance Review 32, no. 2: 220–31. doi:10.1177/1091142103260858.
22 S. MISRA ET AL.

Kaur, B., A. Mukherjee, and A.P. Ekka. 2018. Debt sustainability of states in India: An assessment.
Indian Economic Review 53, no. 1–2: 93–129. doi:10.1007/s41775-018-0018-y.
Magazzino, C., and M. Mutascu. 2019. A wavelet analysis of Italian fiscal sustainability. Economic
Structures 8: 19. doi:10.1186/s40008-019-0151-5.
Mahdavi, S. 2014. Bohn’s test of fiscal sustainability of the American state governments. Southern
Economic Journal 80, no. 4: 1028–54. doi:10.4284/0038-4038-2012.223.
Marini, G., and A. Piergallini. 2008. Indicators and tests of fiscal sustainability: An integrated
approach. CEIS Working Paper No. 111.
Miller, M. 1983. Inflation adjusting the public sector financial deficit. In The 1982 Budget, ed. J. Kay.
London: Basil Blackwell.
Misra, B., and J.K. Khundrakpam. 2009. Fiscal consolidation by central and state governments: The
Medium Term Outlook. RBI Staff Studies. May.
Mohanty, R., and S. Panda. 2019. How does public debt affect the Indian macro-economy?
A structural VAR approach. National Institute of Public Finance and Policy Working Paper Series
No. 250, 22 January. New Delhi.
Mourya, N.K. 2015. Debt sustainability of a sub-national government: A case study of Uttar Pradesh
in India. Journal of Economic Policy and Research 11, no. 1: 126–46.
OECD-UCLG. 2019.. Report on World Observatory on Subnational Government Finance and Investment.
Ostry, J.D., and A.D. Abiad. 2005. Primary Surpluses and Sustainable Debt Levels in Emerging Market
Countries. IMF Policy Discussion Papers 5, no. 6: 1–19.
Pattnaik, R.K., B.S. Misra, and A. Prakash. 2003. Sustainability of public debt in India: An assessment in
the context of fiscal rules. In 6th Workshop on Public Finance, 679–735. Italy: Bank of Italy.
Pradhan, K. 2014. Is India’s public debt sustainable? South Asian Journal of Macroeconomics and
Public Finance 3, no. 2: 241–66. doi:10.1177/2277978714548637.
Rajaraman, I., and A. Mukhopadhaya. 2004. Univariate Time-Series Analysis of Public Debt. Journal of
Quantitative Economics 2: 122–34. doi:10.1007/BF03404612.
Rajaraman, I., S. Bhide, and R.K. Pattnaik. 2005. A study of debt sustainability at state level in India.
Mumbai: Reserve Bank of India.
Rangarajan, C., A. Basu, and N. Jadhav. 1989. Dynamics of Interaction between government deficit
and domestic debt in India. RBI Occasional Papers, September, 163–205.
Renjith, P., and K. Shanmugam. 2018. Sustainable debt policies of Indian state governments. Margin:
The Journal of Applied Economic Research 12, no. 2: 224–43. doi:10.1177/0973801017753283.
Reserve Bank of India (RBI). 2002. Report of the Group to Assess the Fiscal Risk of State Government
Guarantees, July, Mumbai.
Reserve Bank of India (RBI). 2019a. State finances: A study of budgets of 2019–20, September.
Reserve Bank of India (RBI) 2019b. Handbook of Statistics on Indian Economy. Mumbai: Reserve Bank
of India.
Reserve Bank of India (RBI). multiple years. State Finances: A Study of Budgets. Reserve Bank of India.
Seshan, A. 1987. The burden of domestic public debt in India. RBI Occasional Papers, June, 45–77.
Shastri, S., and M. Sehrawat. 2015. Fiscal policy sustainability in India: An empirical assessment.
Journal of Economic Policy and Research 10: 97–112.
Thorat, U., and S. Roy 2004. Contingent Liabilities at the state level – The Indian Experience. Paper
presented at the World Bank Conference, May 21–22.
Tiwari, A.K. 2012. Debt sustainability in India: Empirical evidence estimating time-varying
parameters. Economics Bulletin 32, no. 2: 1133–41.
Trehan, B., and C. Walsh. 1988. Common trends, the government’s budget constraint, and revenue smoothing.
Journal of Economic Dynamics and Control 12, no. 2–3: 425–44. doi:10.1016/0165-1889(88)90048-6.
Tronzano, M. 2013. The sustainability of Indian fiscal policy: A reassessment of the empirical evidence.
Emerging Markets Finance and Trade 49, no. Suppl. 1: 63–76. doi:10.2753/REE1540-496X4901S105.
Twelfth Finance Commission. 2005. Main Recommendations of the Twelfth Finance Commission.
Ministry of Finance, Government of India.
Von Hagen, J., and G.B. Wolff. 2006. What do deficits tell us about debt? Empirical evidence on
creative accounting with fiscal rules in the EU. Journal of Banking and Finance 30, no. 12: 3259–79.
doi:10.1016/j.jbankfin.2006.05.011.
 MACROECONOMICS AND FINANCE IN EMERGING MARKET ECONOMIES 23

Appendix

Table A1. Robustness Checks: Estimation Results 2004–05 to 2017–18. Dependent Variable: Primary
Balance as a proportion to GSDP.
Regressor Model 1 Model 2 Model 3 Model 4 Model 5
Model: Random Effect
Dt-1 (β coefficient) −0.045*** −0.037** −0.04** (0.02)
(0.00) (0.03)
ADt-1 −0.067*** −0.039**
(0.00) (0.02)
RRgap6 0.40*** 0.41***
(0.00) (0.00)
Average RR8 0.007* 0.006*
(0.09) (0.08)
PEgap7 −0.33***
(0.00)
Outputgap5 −0.027 −0.025 −0.045 −0.261 0.03
(0.28) (0.32) (0.80) (0.90) (0.22)
Constant 0.99 1.28 0.05
(0.20) (0.11) (0.54)
Hausman (chi2) 4.3 2.8 6.06
(0.99) (0.99) (0.97)
No. of Observations 338 338 338 338 338
Model: FGLS
Dt-1 (β coefficient) −0.021*** −0.018 −0.01
(0.01) (0.354) (0.46)
ADt-1 −0.037*** −0.027*
(0.00) (0.09)
RRgap6 0.44*** 0.43***
(0.00) (0.00)
Average RR8 0.01* 0.004*
(0.07) (0.07)
PEgap7 −0.39***
(0.00)
5
Outputgap −0.01 −0.01 0.717 0.75 0.03
(0.18) (0.11) (0.70) (0.78) (0.31)
Constant 0.99 3.04** 0.92
(0.20) (0.03) (0.53)
No. of Observations 338 338 338 338 338
Notes: 1. Figures in parentheses are p-values; ***, **,* significant at 1%, 5% and 10% levels, respectively.
2. Augmented debt is obtained by adding outstanding guarantees to the outstanding liabilities of state governments.
One-year lag of debt/augmented debt is taken to surmount the problem of endogeneity.
3. In the RE model, cross-section and time-effects are taken into account.
4. FGLS allows for heteroscedastic correlated error structure and an AR (1) autocorrelation structure.
5. Real Output Gap = Actual Output – Potential Output.
6. RRgap is revenue receipts gap.
7. PEgap is primary expenditure gap.
8. Average RR is calculated by taking the average of two years of revenue receipts (RR).
Source: Authors’ calculations.