External debt accumulation and its impact on economic growth in Pakistan.
Ali, Rifaqat ; Mustafa, Usman
1. INTRODUCTION
The accumulation of external debt is common phenomenon of the
developing countries and it has become a common feature of the fiscal
sectors of most of the economies. A country with lower saving rate needs
to borrow more to finance the given rate of economic growth. So external
debt is obtained to sustain the growth rate of the economy, which is
otherwise not feasible with the given domestic resources. Pakistan is
one of the developing countries and faces serious debt problems,
according to World Bank Report 2000-2001, Pakistan is among the Highly
Indebted Countries (HICs); because Pakistan's present and future
debt situation is very grim.
According to the World Bank total external debt may be defined as
debt owed to nonresident repayable in terms of foreign currency, goods
or services. External debt is the composition of long term debt (public
and publicly guaranteed debt plus private non guaranteed debt), short
term commercial debt and International Monetary Fund (IMF) loans. Prior
to early 1970s the external debt of developing countries was primarily
small and official phenomenon, the majority of creditors being foreign
governments and international financial institutions offer loan for
development project [Todaro (1988)]. At the same time current account
deficit was common which increased the external indebtedness of the
developing countries, until when Mexico, despite an oil exporter,
declared in august, 1992 that it could not services its debt ever since,
the issue of external debt and its servicing has assumed critical
importance and introduced the debt crises debate [Were (2001)].
Several factors have contributed to high rate of debt accumulation
in developing countries. These factors are wide-ranging and
interconnected. The major factor was the 1973-74 oil price increased by
Organisation of Petroleum Exporting Countries (OPEC) led to general
deterioration in the external payments position of the oil importing
developing countries and forced many of them to borrow heavily. Like
other oil importing countries, Pakistan also suffered from these
international events of exceptional nature. These events which imposed
severe strains on its Balance of Payments (BOP) position hampered its
development efforts and led to a marked increase in the volume of
international indebtedness as well as its debt servicing liabilities.
While improper implementation of macro economic policies, political
instability, corruption and poor law and order situation are the main
internal factors for rapid growth of external debt.
Effects of external debt accumulation on investment and economic
growth of the country are always remaining questionable for
policy-makers and academicians alike. There is no consensus on the role
of external debt on growth. It has both positive and negative aspect,
different experts are in view that external debt will have favourable
effect on economic growth because external debt will increase capital
inflow and when used for growth related expenditures can accelerates the
pace of economic growth. It will not only provide foreign capital for
industrial development but will also give managerial know how,
technology, technical expertise as well as access to foreign markets for
the mobilisation of a nation's human and material resources for
economic growth. On the other hand when external debt accumulated beyond
a certain limit, it will contract the economic growth by hampering
investment. A leading explanation for this negative relationship is the
so-called debt overhang hypothesis, which states that high level of
indebtedness discourage investment and negatively affect growth as
future higher taxes are expected to repay the debt.
Pakistan faces serious debt problem, which threaten the economic
future of the country. Burden of external debt and debt servicing have
continued to grow over time. According to the World Bank report
2000-2001, Pakistan is among the HICs; because Pakistan's present
and future debt situation is quite dismal. In 1970 the value of external
debt in absolute term was $ 3.4 billion which went to $ 9.93 billion in
1980. The external debt approximately doubled over from 1981 to 1990 and
reached to $ 20.66 billion. External debt showed rising trend during
1990-99 as it increased from $ 20.66 billion to $ 33.89 billion. It
declined to $ 32.78 billion in 2000 due to debt rescheduling. Then
external debt was $ 35.74 billion in 2003, in the last few years
external debt increased at an unprecedented rate and reached to $ 54.60
billion in 2010 [Pakistan (2010) and World Bank (2007)].
Comparison of indicators of indebtedness to geographically and
income related countries signify that Pakistan is severely indebted as
compare to South Asian and Low Income Countries (LICs) during the last
four decades. External debt as a percentage of Gross National Product
(GNP) was 45.20 percent in Pakistan as compared to 24.17 percent of
South Asian countries and 36.78 percent in LICs. Total reserves as
percentage of external debt was 13.93 percent in Pakistan as compared to
30.94 percent of South Asian countries and 24.67 percent in LICs.
Foreign debt has been a major disbursement item in Pakistan's
exports earnings budget. External debt to export of goods and services
was 356.83 percent in Pakistan as compared to 256.80 percent of South
Asian countries and 243.72 percent in LICs (these figures are the
average of study period i.e. 1970-2010). All these indicators signify
the severity of debt crises that Pakistan is facing.
The study is organised as follows, chapter one is introduction of
the study. Chapter two gives a review of the theoretical and empirical
literatures related to the study. Chapter three presents methodology
where methods and techniques to test the hypothesis have been discussed,
chapter four is econometric analysis of the study, last chapter is
conclusions. References are given at the end of the study.
2. LITERATURE REVIEW
Traditional studies on the external debt problem have focused
mainly on the development of the magnitude and trends of the external
debt in the LICs and then followed by other studies which have examined
the debt burden indicators and severity of the debt problem [Ahmed
(2008)]. Academic research on external debt and its impact on economic
growth have only exploded after the debt crises that hit many developing
countries in the early 1980's. However, recently many empirical
studies have been conducted to assess the impact of external debt on
economic growth but the results are ambiguous.
Oleksandr (2003), divided the existing literature on the related
topic into three groups. A first group of theories suggest that because
poor countries are far away from steady states any investment injection
in form of foreign debt could lead them to have accelerated economic
growth through capital accumulation and productivity growth [Pattillo,
et al. (2004)]. Therefore foreign debt has a positive impact on growth
up to certain threshold level. Second group of theories, stress that
high accumulated debt stock have negative impact on growth. A leading
explanation for this negative relationship is the so called debt
overhang hypothesis of Krugman (1988), and Sach (1989), then advocated
by Cohen (1993). Third group of theories combines these two effects and
argued that the impact of debt on growth is nonlinear.
The relationship between foreign debt and economic growth has
mainly focused on the negative effect of "debt overhang".
Krguman (1988), defined the debt overhang as a situation in which the
expected repayment on foreign debt falls short of the contractual value
of the debt. Likewise, Borensztein (1990), defined the debt overhang as
a situation in which the debtor country benefits very little from the
return to any additional investment because of the debt service
obligations.
The review of existing empirical studies of external debt and
economic growth relationship indicated that it an inadequate to make any
generalisation of the relationship between economic growth and external
debt. Therefore, it is necessary to consider the case of each country or
group separately.
Shabbir (2009) investigated the impact of external debt on economic
growth in 24 developing countries from 1976 to 2003. The study applied
random effect and fixed effect estimation. The results showed that debt
servicing to GDP negatively affect the economic growth and may leave
less funds available to finance private investment in these countries
leading to a crowding out effect.
Adosla (2009) examined the effect of external debt service payments
on the economic growth in Nigeria by using ordinary least square
multiple regression method for his analysis. It was found out that debt
service payments have negative impact on economic growth.
Abu Baker and Hassan (2008), focused to analyse the impact of
external debt on economic growth in Malaysia. The analysis was conducted
both at aggregate and disaggregate level. The empirical results
indicated that total external debt positively affect the economic growth
at aggregate and disaggregate level. In the short run, total external
debt had positive effects on economic growth. It also revealed that
Malaysia had not suffered from debt overhang problem.
On a Similar line Cholifihani (2008), analysed the short run and
long run relationship between external debt and income in Indonesia from
1980 to 2005. The findings showed that GDP, DSR, capital stock, labour
force and human capital inputs have a long run equilibrium relationship.
External debt servicing showed a significant negative relationship with
GDP, which indicated that debt overhang phenomenon, has occurred in
Indonesia in the long run. While labour force and human capital was main
supporting variables of GDP in the long run; however capital stock is
significant variable in boosting economic growth.
Hasan and Butt (2008) explored the association between external
debt and economic growth in Pakistan for the period of 1975-2005 using
Auto Regressive Distributed Lag (ARDL) approach to cointegration.
Results indicated that labour force and trade both in the long run and
the short run mainly determined economic growth in Pakistan. Total debt
was not to be an important determinant of economic growth either in the
short-run or the long run mainly due to inefficient use of external
debt.
Boopen, et al. (2007), investigated the relationship between
external public debt and the economic performance for state of Mauritius
over the period 1960-2004. The results suggested that external debt have
been negatively associated with the output level of the economy in both
short and long run. Bicausality between public debt and economic
development was also reported. Moreover, there were also evidences that
public debt have negative impact on both private and public capital
stock of the country thus confirming the debt overhang and crowding out
hypotheses.
Patenio and Tan-Curz (2007), studied the relationship between
external debt servicing payments and economic growth in Philippines for
period 1981 to 2005. Results showed that economic growth was not very
much affected by external debt servicing. This was probably because
external debt servicing in Philippines was not yet a threat in economic
growth and thus, Philippines should not fear of experiencing debt
overhang in the near future.
Clements, et al. (2003), examined the channels through which
external debt affect economic growth in 55 LICs over the time 1970-1999.
The study suggested that beyond a certain threshold, higher external
debt is associated with lower rates of growth of per capita income. The
results indicated a threshold level of around 30-37 percent of GDP or
around 115-120 percent of exports. The study observed that the negative
effect of debt on growth works not only through its impact on the stock
of debt, but also through the flow of service payments on debt, which
are likely to 'crowd out' public investment. This is so
because service payments and repayments on external debt soak up
resources and reduce public investments. The damaging impact of debt
servicing on economic growth is attributable to the reduction of
government expenditure resulting from debt induced liquidity
constraints.
It is worth mentioning that the majority of existing empirical
literature report that external debt adversely affects economic growth.
Cunningham (1993), Afxentiou (1993), Deshpande (1997), Were (2001),
Karagol (2002), Cholifihani (2008), Hameed, et al. (2008), reported that
the external debt negatively affect the economic growth. Whereas Warner
(1992), Cohen (1993), Afxentiou and Serletis (1996) and Patenio and
Tan-Curz (2007), concluded that external debt did not affect the
economic growth. While Omet and Kalaji (2003), and Abu Baker (2008),
report the positive impact of external debt on economic growth. The
theoretical literature has summarised the following channels namely debt
overhang, liquidity constraint, fiscal effect, productivity suppression
and reduction in human capital accumulation along which external debts
affects negatively growth [see Krugman (1988) and Savvides (1992)].
3. MODEL SPECIFICATION AND EMPIRICAL STRATEGY
This study employed the extended model of production function
originally applied by Cunningham (1993), to investigate the effect of
debt burden on economic growth in sixteen heavily indebted nations.
Cunningham (1993), presumed that the production function only consist of
physical capital, labour and debt service.
The model assumed that there is no human capital. Romer (1986),
investigated that physical capital is important for the production
function but the human capital is vital. Therefore, Karagol (2002),
extended the Cunningham model to incorporate Romer's
conceptualisation of human capital. Human capital consists of skill,
abilities and knowledge of particular workers therefore, to investigate
the relationship between external debt burden and economic growth the
study insert variable of human capital that can be proxied by annual
government education expenditures.
Karagol (2002), covered data of Turkey and Wijeweera, et al.
(2005), used data of Sri Lanka employed education expenditures
representing human capital in the model. Karagol suggested that
education expenditures may not be a suitable proxy for human capital in
case of Turkey. In contrast, in case of Sri Lanka, the results suggested
that education expenditures may have been an appropriate proxy for human
capital.
This study used external debt as a percentage of GDP to capture the
effect of external debt because external debt as a percentage of GDP
signifies the indebtness relative to economic strength of the country.
The model of this study was:
Y = f (HK, K, L, EDY)
The production function used the following specification:
[Y.sub.t] = [[beta].sub.0] + [[beta].sub.1] HK + [[beta].sub.2] K+
[[beta].sub.3] LF + [[beta].sub.4] EDY + [[epsilon].sub.0] ... ... ...
... (1)
[Y.sub.t] = Gross National Product (GNP)
HK = Human capital, It consists of the skills and knowledge of
particular workers (Annual education expenditures of Pakistan used as a
proxy of human capital)
K = Capital stock (Capital formation)
LF = Total labour force
EDY = External debt as a percentage of GDP
[[epsilon].sub.0] = White noise error term
By applying natural logs, the model was
L[Y.sub.t] [[beta].sub.0] + [[beta].sub.1] LHK + [[beta].sub.2] LK
+ [[beta].sub.3] LLF + [[beta].sub.4] LEDY + [[epsilon].sub.0] ... ...
... (2)
It has now become a standard practice to check the univariate time
series of variable by using a unit root test in each series before
estimating any equation. If there is a unit root, then the particular
series is considered to be non-stationary. Moreover, estimation based on
non stationary variables may lead to spurious results which produce high
[R.sup.2] and t-statistics, but without any coherent economic meaning
[Granger and Newbold (1974)]. In accordance with standard practice it
was checked whether the variables are stationary or not.
In this study Augmented Dicky Fuller (ADF) test was carried out for
checking unit roots. ADF has three different specifications, the first
excludes both the trend and the intercept, second specification includes
the intercept but excludes the trend term and the third specification
includes both the trend and the constant term. The study used the third
specification. The purpose to use the ADF to testing the null hypothesis
that a series does contain a unit root (i.e., it is non stationary)
against the alternative hypothesis of stationarity.
[DELTA][Y.sub.t] = [[beta].sub.1] + [[beta].sub.2]t +
[delta][Y.sub.t-1] + [alpha] [[summation].sub.i = 1] [alpha] [Y.sub.t-1]
+ [[epsilon].sub.t] (3)
Where [Y.sub.t] is relevant time series, t is time trend and e, is
white noise error term.
It is also important to select an appropriate lag length; too few
lags may result in rejecting the null hypothesis when it is true (i.e.,
adversely affecting the size of the test), while too many lags may
reduce the power of test [Harris and Sollis (2003)]. The study used the
Schwarz criterion (SC) and Hannan-Quinn information criterion (HQ) to
choose the appropriate lag length.
3.1. Cointegration Analysis
After checking univariate of all time series variables, the study
test cointegration among the variables of the model (GNP, human capital,
capital, labour force and external debt). The reason of the
cointegration test was to determine whether a group of non stationary
series is cointegrated or not.
With the aim of determining long run relationship between variables
cointegration technique is adopted. Two main cointegration techniques
are generally used; Engle and Granger (1987), technique and Johansen
(1988), approach. In order to test cointegration among variable the
study applied the Johansen cointegration technique. This technique
depends on direct investigation of cointegrating Vector Auto Regressive
(VAR) representation.
[Y.sub.t] = [[alpha].sub.1] [Y.sub.t - 1] + [[alpha].sub.2]
[Y.sub.t-2] + ... + [[alpha].sub.k] [Y.sub.t-k] + [[epsilon].sub.t],
Where, [Y.sub.t] is n x 1 vector of I (1) endogenous variables (GNP
and its determinants) in the VAR system [e.sub.t] is a vector of white
noise error terms.
The Johansen procedure is designed to statistically determine the
number of cointegrating vectors in the VAR. In order to determine the
number of cointegrating vector Johansen (1988), provides two different
likelihood ratios tests to determine the value of cointegrating vector.
These are the Trace test:
LR = T[summation] ni = r + 1 ln(1 - [lambda]i)
And the Maximum Eigenvalue test statistics:
LR= T In(1 - [[lambda].sub.r+1])
Trace statistic is a joint test where the null is that the number
of cointegrating vectors is less than or equal to r against an
alternative that there are more than r. Maximum Eigenvalue test conducts
separate tests on each Eigenvalue and has its null that the number of
cointegrating vectors is r against an alternative of r+1. The null
hypothesis was tested sequentially from low to high values of r. The
testing procedure ends when a null hypothesis fails to be rejected for
the first time [Rusike (2007)].
3.2. Short Run Dynamics
The final step of the analysis involved the estimation of short run
relationship between external debt and GNP. The short run model was used
to identify whether the effect of external debt is permanent or
transitory. If the responses are significant in the short run only, then
the impacts of change in external debt is transitory. On the other hand,
if the impacts are significant in both short and long run, then there
will be temporary and permanent effect. If there is an equilibrium or
cointegration relationship among non stationary variables there has to
be an error correction representation [Engle and Granger (1987)].
Relationship between [Y.sub.t] and [X.sub.t] with an error correction
specification as;
[DELTA][Y.sub.t] = [[beta].sub.0] + [[beta].sub.1] [DELTA][X.sub.t]
- [pi] [[??].sub.t-1] + [[epsilon].sub.t]
[[beta].sub.i] will have the short run effect, that measure the
immediate impact that a change in [X.sub.t] will have on change in
[Y.sub.t]. On the other hand is the adjustment effect and shows how much
of disequilibrium is being corrected, i.e. the extent to which any
disequilibrium in the previous period effects any adjustment in the
[Y.sub.t] period.
The error correction mechanism integrates the short run dynamics
with the long run equilibrium without losing long run information. This
term captures the short run relationship. It attempts to correct
deviations from the long run equilibrium path and its coefficient can be
interpreted as the speed of adjustment or the amount of disequilibrium
transmitted each period to economic growth [Ndung'u (1993)].
3.3. Data Collection and Data Definitions
The empirical analysis of this study used the time series data
coverd the period from 1970 to 2010. All the data obtained from
Government publication, Annual Economic Surveys of Pakistan (various
issues), World Development Indicators [CD (2007)], Federal Bureau of
Statistics and State Bank of Pakistan. Besides this, International
Financial Statistics (1FS) of IMF has also been used to supplement the
information.
4. ESTIMATIONS AND ANALYSIS OF RESULTS
This section provides graphical analysis and test the existence of
unit roots of each series using the Augmented Dickey Fuller (ADF) test.
The optimal lag length for the unit root and Johansen's
cointegration tests are decided by the Schwartz Criterion (SC) and
Hannan-Quinn information criterion (HQ). Then the study detects the
number of cointegrating vectors by employing Trace statistics and Max
Eigenvalue test. After that the cointegration analysis was employed
using the Johansen (1988), cointegration technique and calculate the
normalised long run equilibrium equations. Finally the study estimated
the Vector Error Correction Modeling (VECM) for short run dynamics.
4.1. Results for Unit Root Test
Non-stationarity of time series data has often been considered as a
problem in empirical analysis. Working with non-stationary variables
leads to spurious regression results, from which further inference is
meaningless. Therefore, it is important to test the stationarity of all
series entering in the model. The ADF test was used to test the
stationarity of the series. The null hypothesis was that the variable
under investigation has a unit root, against the alternative that it
does not. The results of the test for the variables are presented in
Table 3. In addition to the ADF test, the study also attempted to
examine the trend of the variables graphically. The graphical
illustration of the variables demonstrates the similar characteristic of
the variables as the ADF test.
[FIGURE 1 OMITTED]
The results reported in Table 3 are carried out with trend and
intercept. Results indicated that all series exhibit non stationary in
levels. In other words, the null hypothesis that each of the time series
has a unit root cannot be rejected. However, there is no evidence of a
unit root when the series are first differenced. The no stationary
hypothesis was dismissed in all cases. It means that all the variables
under investigation are stationary at first difference at 1 percent
level of significance except LK which was stationary at 5 percent level
of significance, as can be inferred from Table 3.
4.2. Optimal Lag Selection
After analysing the result of unit root test next step is to find
out the lag order for cointegration. One must determine the optimal lag
structure of the model, i.e. the number of lags that will capture the
dynamics of the series. Results of two different criterions for optimal
lag selection are presented in Table 4. Both SC and HQ statistics
suggested one lag as optimal lag.
4.3. Results from the Conitegration Analysis
The results from the unit root test indicated that the entire
variables entered in the model are non-stationary at level and became
stationary at first difference. While the optimal lag length criteria
suggested one as optimal lag.
The next procedure was to test the existence of long run
relationship among the variables in the model. This study applied
Johansen (1988), cointegration test to examine whether there is more
than one single cointegration relationship.
Johansen's cointegration procedure mainly focused to find out
the number of cointegrating vectors in the system. If the number of
cointegrating vector (0 [less than or equal to] r [less than or equal
to] n) is zero, it would imply, that there is no long run relationship
among the variables. On the other hand, if there are r cointegrating
vectors, it suggests that there are (n-r) common stochastic trends among
the variables that link them together.
Tables 5 and 6 revealed the results of Johansen cointegration test
based on Trace statistics and Max Eigenvalue respectively. These tests
statistics help to evaluate whether there exist a long run relationship
exist among LY, LHK, LK, LLF and LEDY. Both of these tests showed the
long run equilibrium relationship among non stationary variables
entering in the model. The null hypothesis of no cointegration was
rejected therefore, the alternative hypothesis that at least one
cointegrating vector was accepted by both test at 5 percent level of
significance. According to the results of Johansen's test, it can
be argued that a long run relationship exist among LY, LHK, LK, LLF and
LEDY and there exist precisely one cointegrating vector in the estimated
model.
Variables considered in the determination of economic growth have
expected signs except labour force. Human capital and capital positively
affect the economic growth where as external debt and labour force
affects it negatively.
In the context of LDCs, economic theory suggests that human capital
is an important determinant of economic growth. Various theoretical
models include human capital as a factor of production and consider the
accumulation of human capital as an element of the growth process.
Empirical evidence for number of countries also confirmed this
relationship. Lucas (1993), argued that accumulation of human capital
serve as an engine of economic growth. Mankiw (1992), further extended
the theory and consider human capital as an additional accumulatable
factor. He provided evidence that changes in human capital ultimately
translates into significant changes of growth rates. Barro and Lee
(1993) and Benhabib and Spiegel (1994), provided evidence that human
capital accumulation promotes economic growth.
Empirical findings indicated that human capital has a positive
effect on economic growth and have the second most substantial effect on
GNP i.e. 0.31. This means that 1 percent increase in annual education
expenditure (used as proxy of human capital) leads to increase GNP by
0.31 percent. This relationship was significant at 5 percent level of
significance. This indicates the low level of government expenditure on
education in Pakistan.
The results also indicated a positive relationship between capital
and economic growth. This in line with the general assertion that the
capital is a key factor of production hence it is positively associated
to economic growth. Since capital is one of the major determinants of
GNP therefore, according to estimation it reports the positive effect on
economic growth. On this estimates 1 percent increase in capital leads
to increase GNP by 0.53 percent. This association was significant at 5
percent level of significance. The relationship was consistent with
economic theory. This indicates the scarcity of capital in Pakistan.
Labour force showed the negative impact on economic growth; where
as the study hypothesised the positive effect of labour force. Firstly,
it can be argued that Pakistan is labour abundant country. More
unskilled labour having low productivity is unlikely to increase the
level of output in the country. Secondly, agriculture is the largest
sector of the economy and 45 percent of total employed labour force is
working in this sector [Pakistan (2010-11)]. Agriculture sector suffer
from disguised unemployment, majority of the people belong to this
sector seem to be actively participated in economic activities, but
having zero marginal productivity. Therefore, labour was negatively
related to economic growth. These results were conflict with Hameed, et
al. (2008), who found the positive impact of labour force and economic
growth in Pakistan. While Wijeweera, et al. (2005), found the same
result for Sri Lanka. Positive impact of education expenditures also
indicated that there may be scope of improving labour efficiency by
increasing education expenditure in Pakistan.
Results reported in Table 7 indicated that external debt has
negative long run relationship with economic growth. The rationale is
that high ratio of external debt as percentage of GDP leads to lower the
rate of economic growth i.e. an increase in 1 percent in external debt
as percentage of GDP will reduce the GNP by 0.42 percent. These results
confirmed the existence of debt overhang problem. This hypothesis
hypothesised that having heavy debt burden the government will have to
increase taxes in the future to finance the high debt service payments.
That increase in taxes means a lower after tax return on capital and a
reduced incentive to invest. Lower investment leads to slower growth
[Krugman (1987 and 1985); Sachs (1984 and 1986)].
However in the long run repayments of principal and interest
payment absorb the significant portion of foreign reserves making it
difficult to launch new investment projects. This implies that rising
external debt deter economic growth. The findings were consistent with
the literature with Geiger (1990), Cunningham (1993), Afexientue (1993),
Sawada (1994), Deshpande (1997), Karagol (2002), and Hameed, et al.
(2008), (in case of Pakistan) found negative relationship between debt
burden economic growth.
Among the variable capital stock and human capital contributed to
boost the economic growth in the country during the period of the study.
While being a labour abundant nation labour contributed negatively.
Heavy external debt act like a future tax therefore, it verified the
occurrence of debt overhang situation in Pakistan during the period of
the study. All these associations are statistically significant at 5
percent level of significance.
4.4. Short Run Dynamics
Short run dynamic equation has two important objectives. Firstly,
it can be used to investigate whether the impact of any external debt
burden is stable or temporary. If the responses are significant both in
long run and short run, it can be said that changes are permanent as
well as transitory. Finally, the Error Correction Term (ECT) provides
information about the speed of adjustment in response to a deviation
from the long run equilibrium. The short run results of the model are
depicted in Table 8.
The results reported in Table 8 indicate that the short run results
of first difference of log-level variables are similar to long run.
Short run dynamics indicated that the short run impact of human capital
and capital is positive and statistically significant. External debt
exerts negative effect on economic growth in short run and the size of
negative ' impact is stronger than long run. GNP will increase 1.23
and 1.50 percent as a result of 1 percent increase in capital and human
capital respectively. While the short run estimates show the
insignificant negative association between labour force and economic
growth.
Relationship between external debt and economic growth is found to
be negative and significant in short run. This indicated that external
debt effect in Pakistan during the period of study is permanent as well
as transitory and debt overhang occurs both in short and long run.
Negative impact of external debt in short run is stronger than long run,
1 percent increase in external debt as a percentage of GDP will cause
0.48 percent decrease in GNP in short run. Mismanagement of external
debt is the main contributing factor of this negative effect in the
short run.
After determining the existence of cointegrating relationships,
disequilibrium may exist in the short run. If a long run relationship
between different variable exists then an error correction process is
also taking place. The coefficients of the ECT provide information about
the speed of adjustment toward the long run equilibrium after a short
run shock. The speed of adjustment coefficient is correctly signed. The
ECT is significantly different from zero, indicating the existence of
error correction mechanism and implying that the D(LY), D(LHK), D(LK),
D(LLF) and D(LED/Y) converge to the long run equilibrium relationship.
The speed of adjustment of the equilibrium error term suggest that if a
shock inserted into the model 33 percent deviation is rather corrected
with in the first year. The ECT is negative and significant with high
t-values of 3.21, confirms that findings of the study are regarding the
cointegration relationship.
5. CONCLUSIONS
The study attempted to examine the long run and short run impact of
external debt on economic growth in Pakistan over the period of
1970-2010, considering GNP as a function of annual education
expenditures (proxy of human capital), capital, labour force and the
external debt. Then long run equilibrium equation was obtained by
applying Johansen cointegration test while short run results were
obtained through Vector Error Correction Modeling. Finally Error
Correction Term was measured to capture the speed of adjustment.
Empirical evidence revealed that external debt exerts a negative
impact on economic growth; clearly indicate that higher external debt
discourages economic growth. Therefore it verified the occurrence of
debt overhang situation in Pakistan during the period of the study.
Capital as a key factor of production, positively affects the economic
growth. This indicates that capital investment has a lot of potential to
accelerate the pace of economic growth. Human capital has positive
impact on economic growth signified that an educated and highly
productive labour force can lead to speed up the growth process. Labour
force showed the negative impact on economic growth indicated that more
unskilled labour having low productivity is unlikely to increase the
level of output in the country.
Short run results also confirmed the significance of capital
formation and human capital to generate national income. Short run
results showed the similar sign of variable entering in the model as in
the long run but significant negative association of labour' force
and economic growth exist only in the long run.
A significant adjustment parameter obtained from the cointegration
equation confirmed the long run relationship. An estimation of
adjustment parameter suggested that 33 percent of any deviation to the
long run equilibrium corrected in one year.
From the policy prospective it is recommended that increased
domestic saving and export earnings could also raise the estimated
growth rate and reduce the reliance of the economy on external debt. It
is very important to create conducive environment for investment and
much focus of the policies should be on the inflow of Foreign Direct
Investment (FDI), while the inflow of debts should be minimised. There
is severe need of close monitoring and consistent debt management
strategies to avoid the misutilisation of external debt.
Comments
Two papers on external debt (this one and the next one) with
different model, methodologies and scope was limelight of the session.
It would be good if we can have a debate on the issue and come up with
one sound policy since external debt is among the bigger problems of
Pakistan and growth is our first priority. The association among the two
leads to interesting policy implication. The paper finds negative
association among the two variables and argued that this confirms
existence of debt overhang. Few comments on the paper may improve it if
incorporated in the revised version.
(1) Every growth equation should have human capital variable and
the paper used expenditure on education as percentage of GDP. As a
discussant I am not comfortable in using education expenditure as
percentage of GDP as a proxy to human capital. There are lots of other
indicators available in the literature which should be explored.
(2) All the variables are in log-level form thus external debt
should be taken as log-level form instead of long of external debt as
percentage of GDP.
(3) It is not mentioned but 1 assumed that GNP and other variables
are taken in real terms.
(4) I believe short run error correction estimates are obtained by
taking the first difference of the variables, since Table 4.6 does not
mention that whether they are log-level variables or first difference of
log-level variables.
(5) May I suggest using external debt servicing as well with
external debt.
(6) An important result which may not be the scope of the study but
attracts my attention at least is negative association between growth
and labour. Several studies has come up with negative association and
blame "unskilled" labour but I think it is a modelling error.
We are taking labour as exogenous, however, it should be endogenously
treated in our model. There is a chance that negative association tells
us over-sizing of the firms which is true for public sector but not for
private sector.
(7) The paper does not touch the FRDL issue. If it is relevant to
the study I may suggest to include it in the analysis.
M. Ali Kemal
Pakistan Institute of Development Economics, Islamabad.
REFERENCES
Abu Bakar, N. A. and S. Hassan (2008) Empirical Evaluation on
External Debt of Malaysia. International Business and Economics Research
Journal 7:2, 95-108.
Adosla, W. A. (2009) Debt Servicing and Economic Growth in Nigeria:
An Empirical Investigation. Global Journal of Social Sciences 8:2, 1-11.
Afxentiou, P. C. (1993) GNP Growth and Foreign Indebtedness in
Middle-Income Developing Countries. International Economic Journal 7:3,
81-92.
Afxentiou, P. C. and A. Serletis (1996) Growth and Foreign
Indebtedness in Developing Countries: An Empirical Study Using Long-term
Cross Country Data. The Journal of Developing Areas 31, 25-40.
Ahmed, M. M. (2008) External Debts, Growth and Peace in the Sudan:
Some Serious Challenges Facing the Country in the Post-Conflict Era. CHR
Michelson Institute SR 2008: 1, Sudan.
Ayadi, F. S. and F. O. Ayadi (2008) The Impact of External Debt on
Economic Growth: A Comparative Study of Nigeria and South Africa.
Journal of Sustainable Development in Africa 10:3, 234-264.
Barro, R. and J. W. Lee (1993) International Comparisons of
Educational Attainment. Journal of Monetary Economics 32, 363-394.
Benhabib, J. and M. Spiegel (1994) The Role of Human Capital in
Economic Development: Evidence from Cross-Country Data. Journal of
Monetary Economics 34, 143-173.
Boopen, S., P. Kesseven, and D. Ramesh (2007) External Debt and
Economic Growth: A Vector Error Correction Approach. International
Journal of Business Research, 211-233.
Borensztein, E. (1990) Debt Overhang, Debt Reduction and
Investment: The Case of Philippines. (IMF Working Paper, No. WP/90/7).
Cholifihani, M. (2008) A Co-integration Analysis of Public Debt
Service and GDP in Indonesia. Journal of Management and Social Sciences
4:2.
Clements, B., R. Bhattacharya, and T. Q. Nguyen (2003) External
Debt, Public Investment, and Growth in Low-Income Countries. (IMF
Working Paper. 03/249).
Cohen, D. (1993) Low Investment and Large LDC Debt in the 1980s.
American Economic Review 83:3, 437-449.
Cunningham, R. T. (1993) The Effect of Debt Burden on Economic
Growth in Heavily Indebted Nation. Journal of Economic Development 18:1.
Deshpande, A. (1997) The Debt Overhang and the Disincentive to
Invest. Journal of Development Economics 52:1, 169-187.
Engle, R. F. and C. W. J. Granger (1987) Co-integration and Error
Correction: Representation, Estimation and Testing. Econometrica 55,
251-278.
Geiger, L. T. (1990) Debt and Economic Development in Latin
America. The Journal of Developing Areas 24, 181-194.
Granger, C. W. J. and P. Newbold (1974) Spurious Regression in
Econometrics. Journal of Econometrics 2:2, 111-120.
Hameed, A., H. Ashraf, and M. A. Chaudhry (2008) External Debt and
its Impact on Economic Growth in Pakistan. International Research
Journal of Finance and Economics 20.
Harris, R. and R. Sollis (2003) Applied Time Series Modelling and
Forecasting. Chichester, England: Jhon Willey and Sons, Ltd.
Hasan, A. and S. Butt (2008) Role of Trade, External Debt, Labour
Force and Education in Economic Growth Empirical Evidence from Pakistan
by using ARDL Approach. European Journal of Scientific Research 20:4,
852-862.
Johansen, S. (1988) Statistical Analysis of Cointegration Vectos.
Journal of Economic Dynamics and Control 12:2/3, 231-254.
Karagol, E. (2002) The Causality Analysis of External Debt Service
and GNP: The Case of Turkey. Central Bank Review 1, 39-64.
Krugman, P. (1985) International Debt Strategies in an Uncertain
World. In G. W. Smith and J. T. Cuddington (eds.) International Debt and
the Developing Countries. Washington, DC: World Bank.
Krugman, P. (1987) Prospects for International Debt Reform. In
UNCTAD, International Financial Issues for the Developing Countries,
Geneva.
Krugman, P. (1988) Financing vs. Forgiving a Debt Overhang: Some
Analytical Issues. Cambridge, Massachusetts: National Bureau of Economic
Research. (NBER Working Paper No. 2486).
Lucas, R. E. (1993) On the Determinants of Direct Foreign
Investment: Evidence from East and South East Asian. World Development!
1:03, 391-406.
Mohamed, M. A. A. (2005) The Impacts of External Debt on Economic
Growth: An Empirical Assessment of the Sudan: 1978-2001. EASSRR 21:2,
Sudan.
Ndung'u. N. (1993) Dynamics of the Inflationary Process in
Kenya. Unpublished PhD Thesis, University of Gothenburg, Gothenburg.
Oleksandr, D. (2003) Non-linear Impact of External Debt on Economic
Growth: The Case of Post Soviet Countries. Unpublished M.A. Thesis,
National University of Kyiv-Mohyla Academy.
Omet, A. M. G. and F. Kalaji (2003) External Debt and Economic
Growth in Jordan: The Threshold Effect. International Economics 256:3,
337-355.
Pakistan, Government of (Various Issues) Pakistan Economic Survey.
Islamabad: Ministry of Finance.
Patenio, J. A. S. and A. Tan-Cruz (2007) Economic Growth and
External Debt Servicing of the Philippines: 1981-2005. 10th National
Convention on Statistics (NCS).
Patillo, C., H. Poirson, and L. Ricci (2002) External Debt and
Growth. (IMF Working Paper).
Patillo, C., H. Poirson, and L. Ricci (2004) What Are the Channels
Through Which External Debt Affects Growth. (IMF Working Paper).
Romer, P. (1986) Increasing Returns and Long Run Growth. Journal of
Political Economy 94, 1002-1037.
Rusike, T. G. (2007) Trends and Determinants of Inward Foreign
Direct Investment to South Africa. Unpublished M.A. Thesis, Rhodes
University, South Africa.
Sachs, J. D. (1984) Theoretical Issues in International Borrowing.
Princeton Studies in International Finance 54. New Jersey: Princeton
University.
Sachs, J. D. (1986) The Debt Overhang Problem of Developing
Countries. Paper presented in memorial to Carlos Diaz-Alejandro,
Helsinki.
Sachs, J. D. (1989) Introduction to "Developing Country Debt
and the World Economy" NBER Chapters. In Developing Country Debt
and the World Economy 1-34. National Bureau of Economic Research, Inc.
Savvides, A. (1992) Investment Slowdown in Developing Countries
during the 1980s: Debt Overhang or Foreign Capital Inflows? Kyklos 45,
363-378.
Sawada, Y. (1994) Are the Heavily Indebted Countries Solvent? Tests
of Inter Temporal Borrowing Constraints. Journal of Development
Economics 45, 325-337.
Shabbir, Safia (2009) Does External Debt Affect Economic Growth:
Evidence from Developing Countries. Downloadable at:
http://aysps.gsu.edu/sites/default/files/
documents/ECON_MA_shabbirS.pdf.
Todaro, M. P. (1988) Economic Development in the Third World.
(Fourth Edition), Longman, New York and London, p. 411.
Warner, A. M. (1992) Did the Debt Crisis Cause the Investment
Crisis? Quarterly Journal of Economics 107:4, 1161-1186.
Were, M. (2001) The Impact of External Debt on Economic Growth in
Kenya. United Nation University, World Institute for Development
Economics Research. (Paper No. 2001/116).
Wijeweera, A., B. Dollery, and P. Patberiya (2005) Economic Growth
and External Debt Servicing: A Cointegration Analysis of Sri Lanka, 1952
to 2002. (Working Paper Series in Economics 2005-8).
World Bank (2007) World Development Indicators (CD 2007).
Washington, DC: World Bank.
Rifaqat Ali <
[email protected]> is Subject Specialist
(Economics), Education Department, Punjab. Usman Mustafa
<
[email protected]> is Chief Training Programme/Head, Department
of Business Studies, Pakistan Institute of Development Economics,
Islamabad.
Table 1
Summary of Literature Review of External
Debt and Economic Growth Relationship
Time
Date Author Period Sample Findings
2008 Abu Baker 1970-2005 Malaysia External debt
positively affect
economic growth
2008 Ayadi and 1970-2007 Nigeria and Confirm the negative
Ayadi South Africa impact of external
debt on economic
growth.
2008 Hameed, 1970-2003 Pakistan Debt service burden
et al. inversely affect
economic growth.
2008 Colifihani 1980-2005 Indonesia External debt payment
has significant
negative relationship
to GDP.
2007 Patenio and 1981-2005 Philippines Economic growth was
Tan-Curz not affected by
external debt
servicing.
2005 Mohamad 1978-2001 Sudan External debt works
against economic
growth
2003 Clements 1970-1999 55 low income Beyond certain
countries threshold levels
external debt
negatively affect
economic growth.
2003 Omet and 1970-2000 Jordan External debt
Kalaji positively affect
economic growth below
optimal debt level
i.e. 53 percent of
GDP
2002 Wijeweera, 1952-2000 Sri Lanka Debt overhang had not
et al. exist in Sri Lanka.
2002 Karagol 1956-1996 Turkey Debt service is
negatively correlated
to economic growth.
1997 Deshpande 1971-1991 13 Severely The relationship
Indebted between external debt
Countries and investment is
negative.
1992 Warner 1960-1981 13 Less External debt does
and 1982- Developed not reduce
1989 Countries investment.
Table 2
Variables Names and Description
Variable Name Variable Description
LY Log of GNP
LHK Log of human capital
LK Log of capital
LLF Log of labour force
LEDY Log of external debt as a percentage of GDP
Table 3
Results of ADF Test for Non Stationarity
Variables ADF Test in Level ADF Test in First Difference
Calculated Lags Calculated Lags
LY -2.50 1 -4.52 (**) 1
LHK -2.29 1 -4.46 (**) 1
LK -2.73 1 -4.20 (*) 1
LLF -2.06 1 -6.92 (**) 1
LEDY -2.75 1 -5.67 (**) 1
Note: The asterisks (*) and (**) indicates statistical
significance at the 5 percent and 1 percent significance
level.
Table 4
Optimal Lag Selection
Lag SC HQ
0 -5.149482 -5.29265
1 -11.64190 * -12.50092 *
2 -10.59315 -12.16803
3 -9.409976 -11.70071
* Indicates lag order selected by the criterion
SC: Schwarz criterion
HQ: Hannan-Quinn infonnation criterion
Table 5
Unrestricted Cointegration Rank Test (Trace Statistics)
Hypothesised No of Eigenvalue Trace 5 percent
Cointegration Equation Statistics Value Critical Value Prob **
None * 0.62436 75.461 69.818 0.0165
At most 1 0.48028 39.233 47.856 0.2511
At most 2 0.22607 15.017 29.797 0.7790
At most 3 0.13767 5.5350 15.494 0.7497
At most 4 0.00147 0.0545 3.8414 0.8154
Table 6
Unrestricted Cointegration Rank Test (Maximum Eigenvalue)
Hypothesised No. of Eigenvalue Trace 5 percent
Cointegration Equation Statistics Value Critical Value Prob **
None * 0.62436 36.228 33.876 0.0257
At most 1 0.48028 24.215 27.584 0.1273
At most 2 0.22607 9.4822 21.131 0.7917
At most 3 0.13767 5.4805 14.264 0.6803
At most 4 0.00147 0.0545 3.8414 0.8153
Trace test and Max-eigenvalue test indicate 1 cointegrating
eqn(s) at the 0.05 level
* denotes rejection of the hypothesis at the 0.05 level
** MacKinnon-Haug-Michelis (1999) p-values
Table 7
Long Run Equilibrium Equation Dependent Variable (Log GNP)
Independent Variable Coefficient t-statistics
Constant -1.3227 4.2378
Log (Human Capital) (*) 0.31277 4.6239
Log (Capital Stock) (*) 0.52918 6.4832
Log (Labour Force) (*) -0.16823 -10.3901
Log (External Debt as -0.42394 -8.24052
Percentage of GDP) (*)
Note: The asterisks (*) indicates the statistical
significance at 5 percent level of significance.
Table 8
Short Run Results of the Model
Independent Variable Coefficient t-statistics
Log (Human Capital) (*) 1.500092 3.65554
Log (Capital Stock) (*) 1.231827 3.15180
Log (Labour Force) (*) -0.116589 -1.63426
Log (External Debt as -0.482802 -4.92390
Percentage of GDP) (*)
Error Correction Term (*) -0.328 -3.21314
Note: The asterisks (*) indicates the statistical
significance at 5 percent level of significance.
