Economic Analysis and Policy ( ) –

Contents lists available at ScienceDirect

Economic Analysis and Policy

journal homepage: www.elsevier.com/locate/eap

Full length article

Public debt and economic growth in India: A reassessment

Debi Prasad Bal, Badri Narayan Rath ∗
Indian Institute of Technology Hyderabad, India

article info abstract

Article history: This paper examines the effect of public debt on economic growth in India between 1980
Available online xxxx and 2011. Using the autoregressive distributed lag ARDL model, the paper traces a long-run
equilibrium relationship between public debt and economic growth. The error correction
Keywords: model (ECM) results show that central government debt, total factor productivity (TFP)
Domestic debt
growth, and debt-services are affecting the economic growth in the short-run, and that the
External debt
results are consistent with our a priori expectation. It is recommended that the government
ARDL
TFP growth should follow the objective of inter-generational equity in fiscal management over the long
Economic growth term in order to stabilize debt-GDP ratio, particularly, after the global financial crisis.
Debt service © 2014 Published by Elsevier B.V. on behalf of Economic Society of Australia, Queensland.

1. Introduction

The continuous rise in government spending widens the gap of fiscal deficit, and thereby forces the government to depend
on public debt from both internal and external sources. Though the Indian government tries hard to reduce the fiscal deficit
by promoting an inflow of foreign investment and disinvestment, sustaining a lower fiscal deficit becomes challenging,
mainly due to high subsidies on food and fertilizer. The economic consequences of high fiscal deficits result in heavy public
debt, which is likely to affect the economic growth of the nation.
Public debt in India can be classified into external and internal debt. The internal public debt of India has increased in
terms of gross domestic product (GDP) from 36.8% in 1960 to 41.2% in 1970, and from 41.6% in 1980 to 55.3% in 1990.
Although it registered a declining trend from 1995 to 1998–1999, it rose further to 53.4% in 1999, and even more in 2010
at 66%. However, if other liabilities are taken into consideration along with the internal debt, the figures are much higher,
both in pre-reform and post-reform periods. The trends in India’s external debt also increased from USD 261 billion at the
end of March 2010 to USD 305.9 billion at the end of March 2011. The increase can be directly linked to the higher external
commercial borrowing and short-term flows. The share of commercial borrowing in total external debt increased from 19.7
% by the end of March 2005 to 28.9% by the end of March 2011. The long-term external debt accounted for 78.8% of the total
external debt and the remaining 21.2% was short-term debt.
It can be seen from these trends that the total debt (sum of debt and other liabilities) of the central government has
increased significantly. The central government debt amounted to 45.3% of GDP during 1980–1981, and increased to 69.7%
during the fiscal year of 1991–1992. Similarly, the combined debt of central and state governments was 52.4% of GDP in
1980–1981, but it increased to 80.1% in 1991–1992. The average central government debt was 58.2% of GDP during the
pre-reform period from 1980 to 1991 (Handbook of Statistics on Indian Economy, 2012).
The public debt scenario of the Indian government in the post-reform period was worse than in the pre-reform period.
Post-reform, the central government debt was 68.3% of GDP in 1992–1993, increasing to 72.3% in 2002–2003, and then

∗ Correspondence to: Department of Liberal Arts, Indian Institute of Technology Hyderabad, India. Tel.: +91 40 23016052.
E-mail address: badri@iith.ac.in (B.N. Rath).

http://dx.doi.org/10.1016/j.eap.2014.05.007
0313-5926/© 2014 Published by Elsevier B.V. on behalf of Economic Society of Australia, Queensland.
2 D.P. Bal, B.N. Rath / Economic Analysis and Policy ( ) –

slightly declining in consecutive years till 2010–2011 (Handbook of Statistics on Indian Economy, 2012). However, it remains
an alarming fact that the average public debt of the central government during the post-reform period was 65%, which is
higher than the debt of the pre-reform period.
Further, the combined central and state governments’ average debt (public debt plus other liabilities) during the post-
reform period was 79% (Handbook of Statistics on Indian Economy, 2012). Although there has been marked improvement
in repaying external debt (principal and interest payment) by the central government, particularly after 2000, the internal
debt and other liabilities, such as the National Small Saving Fund (NSSF), the Provident fund, and the Deposit and Reserve
funds are on the rise, particularly after the global financial meltdown in 2007–2008.
The present study examines the impact of public debt on economic growth in India by addressing key macroeconomics
variables.
In the process, it differs from existing literature in three novel ways. Firstly, unlike other studies (Kannan and Singh, 2007;
Ghosh, 2006; Rangarajan and Srivastava, 2005; Singh, 1999), this study takes additional key macroeconomics indicators,
such as internal and external debt, debt service payment, total factor productivity and export into the growth equation,
and examines the relationship between debt and economic growth in an extended growth accounting framework for India.
Secondly, this study uses more sophisticated statistical methodology to explore both short-run and long-run effects of public
debt on economic growth using the autoregressive distributed lag (ARDL) model (Pesaran et al., 2001). Thirdly, most of the
previous studies are almost a decade old. Therefore, it has become imperative to revisit the public debt–economic growth
linkage when the Indian economy is facing serious problems for maintaining GDP even 6% of growth. Most of the credit-
rating agencies have downgraded India’s sovereign rating, which has created considerable fiscal pressure on the current
government.
The remainder of this paper is organized as follows: Section 2 reviews the literature; Section 3 details the methodology in
terms of the analytical framework and data sources; Section 4 presents the results; and Section 5 discusses the conclusions
and implications for policy.

2. Review of literature

Classical economists, such as Smith (1776, p. 878), Ricardo (1951, p. 247), and Mill (1845, p. 230) viewed the economic
effect of public debt on a nation as destructive. The Ricardian Equivalence theory noted that the financing of public
expenditure via taxation and borrowing is equal. Its argument is that the repayment of debt will take place through
future taxation, which means individuals will increase their savings by buying the bonds that have been issued by
the government. So, according to Ricardo there is neutral effect of public debt on economic growth. Whereas, in the
Investment Saving–Liquidity Preference Money Supply (IS–LM) model, Keynesian economists pointed out that an increase
in government debt induced by deficit-financed fiscal policy will increase the level of income, the transaction demand
for money and prices. This will cause the rate of interest on bonds to rise with a fixed money supply. According to
Keynesian theory, if the private sector perceives government securities as net wealth, the deficit will further amplify private
consumption expenditures, transaction demand, interest rates, and prices. The effects of expansionary fiscal policy on capital
formation may be strengthened through the accelerator effects and thereby raise economic growth. On the other hand, the
monetarists have argued that the macroeconomic effect of debt financed is crowding out the private investment through
increasing levels of interest rates. Hence, public debt will affect economic growth in a negative manner. Further, the debt
overhang theory suggested that if future debt gets larger than the country’s repayment ability, the expected debt-service
costs will discourage further domestic and foreign investment, and thus harm economic growth.
Several studies have found an inverse linear relationship between total debt and economic growth both across countries
and at a single country-level analysis. The empirical work by Mitchell (1988), Baro (1989, p. 238), and Camen and Rogoff
(2011) used UK data to show that public debt has a significant impact on economic growth. According to a study by Forslund
et al. (2011), results indicate a negative correlation between domestic debt and inflation in developing countries. Ismihan
and Ozkan (2012) found that public debt can harm countries when the financial movement is underdeveloped. Westphal-
Checherita and Rother (2012) argued from their analysis of twelve European countries that there exists a non-linear impact
of public debt on economic growth, which was mainly due to private savings, public investment and total factor productivity.
As viewed by Schclarek (2004), 59 developing countries from 1970 to 2002 were examined and found that a significant
inverse relationship existed between external debt and economic growth.
Similarly, Kumar and Woo (2010) came to the conclusion from their empirical study of 38 advanced and emerging
economies between 1970 and 2007 that there was an inverse relationship on initial debt and economic growth across.
Qureshi and Ali (2010) found that the high level of public debt had negatively affected the economy of Pakistan between 1981
and 2008. Some of the important studies, which particularly address the issues in Indian context, are as follows: Singh (1999)
has examined the long run relationship between domestic debt and economic growth using the Johansen cointegration
technique. His study supported the hypothesis of Ricardian equivalence in India. Kannan and Singh (2007) showed that
public debt and a high level of fiscal deficit had an adverse effect on interest rates, output, inflation and the trade balance
in the long run in India. Similarly, Rangarajan and Srivastava (2005) argued that a large fiscal deficit and interest payments
to GDP adversely affected economic growth. They also pointed out that public debt negatively affected the growth of the
Indian economy.
 D.P. Bal, B.N. Rath / Economic Analysis and Policy ( ) – 3

3. Analytical framework and data sources

The analytical framework used in the study involves the measurement of India’s total factor productivity (TFP) growth
using the Malmquist approach, and the autoregressive distributed lag (ARDL) model, and the data sources for determining
economic growth and public debt in India are outlined.

3.1. Analytical framework

The relationship between economic growth and public debt in developing countries has largely been based on a standard
production function model:

Y = f (K , L). (1)

Where Y , L and K are measures of output, capital and labor respectively. According to debt overhang theory if the debt of
a country is more than its repayment capacity, this discrepancy will negatively affect investment and the ability to work,
and therefore affect the growth of the economy. Pattillo and Poirson (2002) argued that debt has an inverted U-shaped
relationship on economic growth.
The present study includes domestic debt, external debt, exports and debt service ratio in the growth function. The
inclusion of exports as one of the key inputs in the production function is consistent with Cunningham (1993) who argued
that large-scale exports from a country lead to raised productivity, and hence positive economic growth. Theoretically, the
total factor productivity growth has a positive impact on economic growth in the long run. Studies by Checherita and Rother
(2010), and Pattillo and Poirson (2004) found that the total factor productivity (TFP) affects the growth of the economy.
Therefore, it would be appropriate to include TFP as one of the key determinants for economic growth. The present study
extended the function from Eq. (1) to Eq. (2) as:

Y = f (Debt, TFP, Exports). (2)

In Eq. (2), the variable debt is considered as domestic debt, external debt and debt service as a percentage of exports in order
to determine the individual effects on economic growth. This study uses the Cobb–Douglas production function to establish
the output equation for India for the 30-year period, 1980 to 2010.

β γ
Yt = At DLαt ELt DSt TFPtλ EXPtθ , (3)

where α, β, γ , λ and θ are treated as the elasticity coefficients of the domestic debt (DL), the external debt (EL), debt
service (DS) payment of the country, total factor productivity (TFP), and exports (EXP) respectively. By undertaking a simple
manipulation of taking the natural logarithms on both sides, we re-wrote Eq. (3) as:

ln Yt = At + α ln DLt + β ln ELt + γ ln DSt + λ ln TFPt + θ ln EXPt + εt . (4)

The reasons for applying the logarithm conversion were mainly to seasonally adjust all the variables and for the magnitude
change of the variables. It also helped us to interpret the results in elasticity term. Theoretically, we expected that the
coefficients of TFP and EXP would be positively associated with economic growth. A country’s high level of productivity
boosts the aggregate output in the long-run, and as a result, it has a positive impact on economic growth. An increase in the
exports of a nation usually leads to a positive impact on income and therefore on economic growth. However, domestic and
external debt may show an ambiguous effect on economic growth. As we mentioned in the literature review, public debt
(both internal and external) can have a negative impact on economic growth, but the Keynesian economists pointed out
that an increase in government debt induced by deficit-financed fiscal policy will increase economic growth. The empirical
literature on different countries show mixed results. Some of the studies find the public debt has positive impact on economic
growth up to threshold level, but if the debt crosses beyond 60 percent of GDP, it will have a negative effect on growth.

3.2. Measurement of TFP using the Malmquist approach

This paper measures the total factor productivity (TFP) growth for the aggregate economy of India using the non-
parametric method of Data Envelopment Analysis (DEA). DEA was originally designed to study the relative efficiencies of
different firms or managerial units assumed to have a common best practice production technology available to them. The
method enables a comparison among firms on the basis of the extent to which inputs are used efficiently in the production of
output, given the technology (see Farrell, 1957; Charnes et al., 1978). However, there are studies, where the DEA technique
has been used to measure the overall productivity of a nation as whole (See, Fare et al., 1994; Maudos et al., 1999; Tian
and Xiaohua, 2012). The description below draws primarily upon the work of Fare et al. (1994). The Malmquist productivity
index is explained using the distance function. We adopt an output-oriented approach of computing TFP growth in this
paper.
4 D.P. Bal, B.N. Rath / Economic Analysis and Policy ( ) –

DEA involves the use of linear programming methods to construct a non-parametric piece-wise frontier technology
expressed by the following production possibility set:
S = {(x, y) : y can be produced by x}
 
J
 J

= (x, y) : λj ymj ≥ ym , xn ≥ λj xnj ∀n , λj ≥ 0∀j , (5)
j =1 j=1

where x is the input vector and y is the output vector, and in the last expression we have introduced J points and index m for
type of output, index n for type of input. The variables λj (j = 1, . . . , J ) are non-negative weights and we define the constant
returns to scale. Basic properties are that the production set is convex, includes all points, and envelopment is done with
minimum extrapolation. The output-oriented Farrell radial efficiency measure, Φi , for each unit of i, of a set of J observations,
is calculated by solving the following linear program set up:
Max Φi (6)
s.t.
J

λj ymj − φi ymi ≥ 0, m = 1, . . . , M (7)
j =1

J

xnj − λj xnj ≥ 0, n = 1, . . . , N
j =1

λj ≥ 0, j = 1, . . . , J .
Each type of output is scaled up with the same factor, Φi , until the frontier is reached to the definition of the Farrell efficiency
measure.
The Malmquist TFP index measures the TFP growth between two data points by calculating the ratio of the distances
of each data point relative to a common technology. If the period ‘t’ technology is used as the reference technology, the
Malmquist (output-oriented) TFP growth between period ‘s’ (the base period) and period ‘t’ can be written as:
dt0 (qt , xt )
mt0 (qs , xs , qt , xt ) = . (8)
dt0 (qs , xs )
It may be noted here that in the above equations the notation (q, x) represents the input and output respectively. Similarly,
the notation ds0 (qt , xt ) represents the distance from the period ‘t’ observation to the period ‘s’ technology. A value of m0
greater than one indicates positive TFP growth from period ‘s’ to period ‘t’, while a value less than one indicates a TFP decline.
To avoid the necessity to either impose the restriction or to arbitrarily choose one of the technologies, the Malmquist TFP
index is often defined as the geometric mean of these two indices. That is
1/2
ds0 (qt , xt ) dt0 (qt , xt )

m0 (qs , xs , qt , xt ) = × . (9)
ds0 (qs , xs ) dt0 (qs , xs )
In this study, while measuring the TFP growth, the GDP of the country is treated as the single output and both labor force
and capital stocks are taken as two inputs.

3.3. ARDL model specification

We implement the autoregressive distributed lag (ARDL) model, or bound testing approach (Pesaran et al., 2001) here
to check the existence of the short and long-run relationship between per capita economic growth, domestic debt, external
debt, debt services ratio, TFP growth and exports in India. We employ the ARDL model for two reasons. First, there are
combinations of I (0) and I (1) order among the variables, which are included in the model. Second, this approach is more
suitable for the small and finite sample data period (Pesaran et al., 2001).
Before presenting the Pesaran et al. (2001) ARDL test, it is essential to focus on vector autoregression (VAR) of order p for
the growth function
k

xt = η + βj xt −j + εt , (10)
j =1

where xt = [DLt , ELt , DSt , TFPt , EXPt ]′ , η is a vector associated with the constant term, and βj is a matrix of VAR parameters
for lag j and εt is the white noise disturbance term. The Vector Error Correction (VEC) model can be written as:
k−1

∆ x t = η + x t −1 + λ λj ∆xt −j + εt , (11)
j =1
 D.P. Bal, B.N. Rath / Economic Analysis and Policy ( ) – 5

where ∆ is first difference and λ is the long-run multiplier matrix as:

λYY λYX
 
λ= . (12)
λXY λXX
The Wald test (F -statistics) is also important for the ARDL test. This test identifies the existence of a long run relationship
among the variables. The null and alternative hypotheses are as follows:
H0 = β1 = β2 = β3 = β4 = 0 and H1 ̸= β1 ̸= β2 ̸= β3 ̸= β4 ̸= 0. (13)
The computed F-test will be compared with the critical tabulated value of Narayan (2005). According to Pesaran et al. (2001)
and Narayan (2005), the lower bound critical values assumed that the explanatory variables are integrated of order zero,
while the upper bound critical values assumed that the explanatory variables are integrated of order one. Therefore, if the
computed F -statistic is smaller than the lower bound value, the null hypothesis is not rejected and it can be concluded that
there is no long-run relationship between the variables. Alternatively, if the computed F -statistic is greater than the upper
bound value, there is the existence of a long-run relationship among the variables. If the computed F -statistic is in between
the lower and upper bound, the result is considered inconclusive.
Once we identify the long-run relationship, the next step of the ARDL model is to estimate the long-run coefficient from
the equation, which is as follows:
ρ1
 q1
 q2
 q3

ln Yt = β0 + β1 ln Yt −i + β2 ln DLt −i + β3 ln ELt −i + β4 ln DS t −i
i =1 i =0 i =0 i =0

q4
 q5

+ β5 ln TFP t −i + β6 ln EXP t −i . (14)
i=0 i =0

Before estimating the long-run specification, we need to determine the lag order through VAR. In the final step, we obtain
the short-run dynamic parameter by estimating an error correction model (ECM). This is as follows:
ρ
 q
 q

∆ ln Yt = α0 + βi ∆ ln Yt −i + γj ∆ ln DLt −j + δk ∆ ln ELt −k
i=1 j =1 k=1

q
 q
 q

+ εl ∆ ln DS t −l + θm ∆ ln TFP t −m + σn ∆ ln EXP t −n + ϑ ECM t −1 + ϵt , (15)
l=1 m=1 n=1

where β, γ , δ, ρ, θ , σ , and µ are the short-run dynamic coefficients and ϑ is the coefficients of speed of adjustment which
is expected to have a negative sign.

3.4. Data sources

The per capita gross national product (GNP) is defined as economic growth. The domestic liabilities (DL) and external
liabilities (EL) are treated as the public debt and the data were collected from the Handbook of Statistics on Indian Economy
published by RBI (2012-130 citation). The debt service is defined as a percentage of export (DS) and it was collected from
World Development Indicators (WDI) published by World Bank. The data on exports (EXP) was collected from International
Financial Statistics published by IMF. Finally, the estimation of TFP relied on the data on Real GDP, labor force and gross
fixed capital formation collected from the World Bank data catalog. The capital stock variable has estimated using Perpetual
Inventory Method (PIM) and detailed procedures are explained in the Appendix.

4. Results

A range of commonly used summary statistics for the key variables is presented in Table 1. After presenting the summary
statistics, the first step is to know the stationary or non-stationary property of the variables in order to avoid the spurious
regression in time series analysis. Hence, the ADF tests are used and the results are presented in Table 2.
The ADF (Dicky and Fuller, 1979) test is used for knowing the order of variables. The above results clearly reveal that the
model follows a mixture of I (0) and I (1). The results of the unit root test motivate to employ the autoregressive distributed
lags model (ARDL) cointegration technique proposed by Pesaran et al. (2001). The results are presented in Table 3.
The results in Table 3 show that the calculated F-statistic is greater than the Narayan (2005) and Pesaran et al. (2001)
critical value at 5%level and 10% level respectively. Hence, we reject the null hypothesis of no long-run relationship between
these variables, which implies that there exists a long-run relationship between public debt and economic growth. The
study chooses the optimum of lag 3 through VAR model by following AIC, SC and HQ criteria. Having found a long-run
relationship amongst the variables, the next step is to estimate the long-run effects of public debt, debt-service ratio, total
factor productivity and exports on economic growth. The long-run estimation results are presented in Table 4.
6 D.P. Bal, B.N. Rath / Economic Analysis and Policy ( ) –

Table 1
Descriptive statistics.
Variables Mean Max. Min. S.D. Skewness Kurtosis J.B. test

LYt 9.88 10.61 9.37 0.36 0.42 2.05 2.09

LDLt 8.54 10.54 6.18 1.31 −0.18 1.86 1.84
LELt 6.88 7.90 4.90 0.90 −0.81 2.25 4.13
LDSt 3.02 3.59 1.73 0.50 −0.83 2.96 3.61
LEXPt 6.61 9.16 4.19 1.59 −0.03 1.70 2.16
LTFPt 5.01 5.30 4.62 0.22 −0.27 1.68 2.63
Note: LY = Economic Growth, LDL = Domestic Debt, LEL = External Debt, LDS = Debt Service, LEXP = Export, LTFP = Total Factor Productivity.

Table 2
Results of unit root test.
Variables Level 1st order difference Inference on integration

LYt 0.0798 (0.99) −5.4975 (0.00) I(1)

LDLt −4.7919 (0.00) – I(0)
LELt −1.4643 (0.81) −6.6160 (0.00) I(1)
LDSt −1.6988 (0.72) −4.6566 (0.00) I(1)
LTFPt −1.7672 (0.38) −5.4590 (0.00) I(1)
LEXPt −3.2768 (0.09) – I(0)
Note: Figures in the parentheses are p-value.

Table 3
Bound F -test results.
Country F -statistic value Lag length Significance level Bound critical values by Bound critical values by
Narayan (2005) Pesaran et al. (2001)
I (0) I (1) I (0) I (1)

1% 4.483 6.320 3.74 5.05

India 4.269299 3 5% 3.120 4.560 2.45 3.61
10 % 2.560 3.828 2.12 3.23
Note: Critical values are borrowed from Narayan (2005).

Table 4
Long run elasticities.
Variables Coefficients t-test statistics

LDLt −0.4078** −2.10

LELt −0.2384* −5.34
LDSt 0.1032** 1.98
LTFPt −0.3035 −1.51
*
LEXPt 0.2617 3.50
Constant 12.64* 14.19
Trend 0.0919* 3.77
Note: Dependent variable LPGNP.
*
p < 1%.
**
p < 5%.

The above results illustrate that both domestic debt and external debt have a significant negative impact on economic
growth of India. That implies that higher public debt, irrespective of its source, is reducing economic growth in India in the
long-run. This finding is not surprising; particularly in the context of India where most of the government borrowings are
utilized in consumption expenditure and very few portions go towards forming productive capital. Similarly, the coefficient
of debt service payment has a significant and positive impact on economic growth in the long-run, which corroborates with
the a priori expectation. Although theoretically we expect a positive relationship between TFP and economic growth, the
result in Table 4 indicates that the coefficient is not statistically significant. Finally, the coefficient of exports has a positive
and significant impact on economic growth. As India has constantly changed trade policies in last two decades, it has helped
the country by increasing the exports of goods and services in the world market, which might positively affect the economic
growth.
The final step of the ARDL model is the error correction for estimating the short-run parameter with speed of adjustment.
The results of the error correction model are presented in Table 5.
The results confirm that there exists a stable long-run relationship among the variables which is confirmed by the
significance of the error correction term (Bannerjee and Mestre, 1998). The coefficient on the lagged error correction term
measures the speed of adjustment. The lagged error correction term is negative and significant, which implies that the
series is non-explosive and that a long-run equilibrium is attainable. Because the ECMt−1 measures the speed at which
the endogenous variable adjusts to changes in the explanatory variables before converging to its equilibrium level. The
 D.P. Bal, B.N. Rath / Economic Analysis and Policy ( ) – 7

Table 5
Results of error correction representation.
Variables Coefficients t-test statistics

∆LDLt 0.164 1.05

∆LDLt (−1) 0.526** 2.94
∆LDLt (−2) 0.435** 2.96
∆LELt 0.001 0.04
∆LELt (−1) 0.149** 3.49
∆LELt (−2) 0.069** 2.19
∆LDSt 0.024 1.22
∆LDSt (−1) −0.112** −5.42
∆LTFPt 0.720** 4.71
∆LTFPt (−1) 0.377* 2.07
∆LTFPt (−2) 0.319** 2.22
∆LEXPt 0.196** 3.88
∆Ct 9.505** 4.72
∆ Tt 0.069** 3.86
ECMt (−1) −0.751** −4.65
R-squared 0.953
Adjusted R-squared 0.861
Prob (F -statistics) 0.000
χ(2Auto) (2) 5.04 (0.025)
χ 2(Norm)(1) 4.44 (0.10)
χ(2Het) (1) 0.057 (0.81)
Note: Dependent variable ∆ LPGNP.
*
p < 1%.
**
p < 5%.

Fig. 1. CUSUM test.

Fig. 2. CUSUM SQ test.

coefficient of ECM term suggests an adjustment of approximately 75% towards the long-run equilibrium after one year. The
result also reveals that in lag periods one and two, both domestic and external debt have significantly and positively affected
the growth in the short-run. The debt service ratio has negatively affected the growth of the economy in the first lag. The
total factor productivity positively affects the growth in the short-run, and its lag periods also contribute significantly to the
growth of the economy. Finally, exports positively affected the economic growth of India between 1980 and 2010.
To make the ARDL results more robust, in the final stage we checked the stability of the regression coefficients using
cumulative sum (CUSUM) test and Figs. 1 and 2 clearly show that the critical values did not exceed the 5 % level of
significance.
8 D.P. Bal, B.N. Rath / Economic Analysis and Policy ( ) –

5. Conclusion and policy implications

The paper examined the impact of India’s public debt during 1980 to 2010 on its economic growth. Our approach
extended the production function by including the variables of TFP growth and exports. This study estimated the TFP
growth using the DEA analysis and examined its short- and long-run impact on economic growth. The impact of public
debt on economic growth using the ARDL model appears to be a methodological addition to existing debt literature in the
Indian context. The results derived from the ARDL model show that there is a long-run equilibrium relationship among the
variables. However, TFP growth had no long-run impact on economic growth. Through the ECM model, it was observed that
all the variables affected economic growth in the short-run and these results were consistent with our a priori expectation.
From the policy perspective, the government must try to identify the solution for enhancing the economic growth, which will
ultimately reduce the public debt of the country. Even though there was minimal risk for India at one stage for its refinancing
needs of existing debt, the government liabilities as percentage of GDP have substantially increased, particularly since the
global financial crisis. Therefore, it would be advisable for the government to follow the objective of inter-generational
equity in fiscal management over the long-term.

Acknowledgments

The authors gratefully acknowledge the suggestions of the editor and anonymous referee on an earlier draft of this
paper. We thank Paresh Kumar Narayan, Bibhudutta Panda, and Prabheesh K.P. for providing useful suggestions. The usual
disclaimer applies.

Appendix

Measuring capital stock

The physical capital stock data are not readily available for India. Thus, following Easterly and Levine (2002), we used a
Perpetual Inventory Method (PIM) to compute capital stocks. Specifically, let K (t ) equal the real capital stock in period ‘t’. Let
I (t ) equal the real investment rate in period ‘t’. The real investment is defined in this paper as gross fixed capital formation
at constant 2000 US$. Let ‘d’ equal the depreciation rate, which we assumed equals 0.07. Thus, the capital accumulation
equations stated as: K (t ) = (1 − d)K (t − 1) + I (t ). To make an initial estimate of the capital stock, we made the assumption
that the country is at its steady-state capital–output ratio. Thus, in terms of steady-state value, let k = K /Y , let g = the
growth rate of real GDP, Y is the real GDP and let i = I /Y . Then, from the capital accumulation equation, plus the assumption
that the country is at its steady-state, we knew that k = i/[g + d]. Thus, if we can obtain a reasonable estimate of the steady-
state values of ‘i’, ‘g’ and ‘d’, then we can compute a reasonable estimate of ‘k’. Then, using the calculated value of ‘k’, an initial
estimate of capital stock (k) multiplied with initial GDP (Y ) can be obtained. In order to calculate the initial estimate of ‘k’,
we assumed the steady state capital–output ratio (d) = 0.07. We have constructed the steady-state growth rate (g ): a
weighted average of the countries average growth rate during the first ten years for which we have output and investment
data and the world growth rate. The world growth rate was computed as 0.0234. Based on Easterly et al. (1993), we gave a
weight of 0.75 to the world growth rate and 0.25 to the country growth rate in computing an estimate of the steady-state
growth rate for each individual country. We then computed ‘i’ as the average investment rate during the first ten years for
which there are data. Thus, with values for ‘d’, ‘g’, and ‘i’ for each country, we have estimated ‘k’ for each countries. To reduce
the influence of business-cycles on estimates of Y , this article used the average real GDP value between 1979 and 1981 as
an estimate of initial output. Thus, the capital stock, for example, in 1980 is given as: Y ∗ k.

References

Bannerjee, A.D., Mestre, R., 1998. Error-correction mechanism test for cointegration in single equation frame work. J. Time Ser. Anal. 19, 267–283.
Baro, R., 1989. The Ricardian approach to budget deficit. J. Econ. Perspect. 3, 37–54.
Camen, R., Rogoff, K., 2011. From financial crash to debt crisis. Amer. Econom. Rev. 101, 1676–1706.
Charnes, A., Cooper, W.W., Rhodes, E., 1978. Measuring the efficiency of decision making units. European J. Oper. Res. 2, 429–444.
Checherita, C., Rother, P., 2010. The impact of high and growing government debt on economic growth: an empirical investigation for the Euro area. Euro.
Central Bank Working Paper 1237.
Cunningham, R.T., 1993. The effects of debt burden on economic growth in heavily indebted nations. J. Econ. Dev. 18, 115–126.
Dicky, D.A, Fuller, W.A., 1979. Distribution of the estimation for autoregressive time series with a unit root. J. Amer. Statist. Assoc. 74, 427–431.
Easterly, W., Kremer, M., Lant, P., Summers, L.H., 1993. Good policy or good luck? Country growth performance and temporary shocks. J. Monetary Econ.
32, 459–483.
Easterly, W., Levine, R., 2002. It’s not factor accumulation: stylised facts and growth models. In: Loayza, N., Soto, R. (Eds.), Econ. Growth: Sources, Trends
and Cycles. Central Bank of Chile, Santiago, Chile, Mimeo.
Fare, R., Grosskopf, S., Morris, M., Zhongyang, Z., 1994. Productivity growth, technical progress, and efficiency in industrialized countries. Amer. Econom.
Rev. 84, 66–82.
Farrell, M.J., 1957. The measurement of productivity efficiency. J. Roy. Statist. Soc. 120, 253–290.
Forslund, K., Lima, L., Panizza, U., 2011. The determinants of composition of public debt in developing and emerging market countries. Rev. Dev. Fin. 1,
207–222.
Ghosh, S., 2006. Can higher debt lead to higher welfare? A theoretical and numerical analysis. Appl. Econ. Lett. 5, 111–116.
Handbook of Statistics on Indian Economy, 2012-13. Reserve Bank of India Database: http://www.rbi.org.in/scripts/AnnualPublications.aspx?head=
Handbook%20of%20Statistics%20on%20Indian%20Economy.
 D.P. Bal, B.N. Rath / Economic Analysis and Policy ( ) – 9

Ismihan, M., Ozkan, F.G., 2012. Public debt and financial development: a theoretical exploration. Econ. Lett. 115, 348–351.
Kannan, R., Singh, B., 2007. Debt-deficit dynamics in india and macroeconomic effects: a structural approach. Munich Personal Repec Arch. 16480. Available
on the web: http://mpra.ub.uni-muenchen.de/16480/.
Kumar, M.S., Woo, J., 2010. Public debt and growth. IMF Working Paper WP/10/174.
Maudos, J., Pastor, J.S., Serrano, L., 1999. Total factor productivity measurement and human capital in OECD countries. Econ. Lett. 63, 39–44.
Mill, J.S., Principle of Political Economy, http://www.gutenberg.org/licens.
Mitchell, B., 1988. British Historical Statistics. Cambridge University Press, Cambridge.
Narayan, P.K., 2005. The saving and investment nexus for China: evidence from cointegration tests. Appl. Econ. 37, 1979–1990.
Pattillo, C.H., Poirson, Ricci, R., 2002. External debt and growth. IMF Working Paper 02/69.
Pattillo, C.H., Poirson, Ricci, R., 2004. What are the channels through which external debt affects growth? IMF Working Paper 04/15.
Pesaran, H.M., Shin, Y., Smith, R.J., 2001. Bounds test approaches to the analysis of level relationships. J. Appl. Econometrics 16, 289–326.
Qureshi, M.N., Ali, K., 2010. Public debt burden and economic growth: evidence from Pakistan. Int. Res. J. Fin. Econ. 53, 100–108.
Rangarajan, C., Srivastava, D.K., 2005. Fiscal deficits and government debt: implications for growth and stabilisation. Econ. Pol. Weekly 40, 2919–2934.
Ricardo, D., 1951. On the principles of political economy and taxation. In: Sraffa, P. (Ed.), The Works and Correspondence of David Ricardo 1. Cambridge
University Press, Cambridge, with the collaboration of M. Dobb.
Schclarek, A., 2004. Debt and economic growth in developing and industrial countries. Available on the web: www.nek.lu.se/nekasc.
Singh, C., 1999. Domestic debt and economic growth in India. Econ. Pol. Weekly 34, 1445–1453.
Smith, A., 1937. In: Cannan, E. (Ed.), The Wealth of Nations. Random House, New Work.
Tian, Xu, Xiaohua, Y.U., 2012. The enigmas of TFP in China: a meta-analysis. China Econ. Rev. 23, 396–414.
Westphal-Checherita, C., Rother, P., 2012. The impact of high government debt on economic growth and its channels: an empirical investigation for the
Euro area. Euro. Econ. Rev. 56, 1392–1405.