Sign up

External debt and economic growth: a vector error correction approach.
Subject:
Economic development (United States)
Economic development (Analysis)
External debts (Analysis)
Authors:
Seetanah, Boopen
Padachi, Kesseven
Durbarry, Ramesh
Pub Date:
09/01/2007
Publication:
Name: International Journal of Business Research Publisher: International Academy of Business and Economics Audience: Academic Format: Magazine/Journal Subject: Business, international Copyright: COPYRIGHT 2007 International Academy of Business and Economics ISSN: 1555-1296
Issue:
Date: Sept, 2007 Source Volume: 7 Source Issue: 5
Product:
Product Code: 9008000 Economic Programs-Total Govt; 8515300 Development NAICS Code: 926 Administration of Economic Programs; 5417 Scientific Research and Development Services
Geographic:
Geographic Scope: United States Geographic Code: 1USA United States

Accession Number:
178945825
Full Text:
ABSTRACT

The paper investigates the link between external debt and economic performance for the case of the small island developing state of Mauritius over the period 1960-2004. Given the possibility of endogeneity and dynamics in the relationship, a Vector Error Correction (VECM) Framework was employed to explain the short and long variations of the country's output level. Result from the analysis shows that external debt has been negatively associated with the output level of the economy in both short and long run. Interestingly bi-causality between public debt and economic development is also reported. Moreover there are evidences of the debt overhang and crowding out hypotheses as public external debt have also negative impact on both private and public capital stock of the country.

Keywords: External Debt, Growth, VECM

1. INTRODUCTION

Growing public debt is a worldwide phenomenon and it has become a common feature of the fiscal sectors of most of the economies. The inadequate debt management and a permanent growth of debt to Gross Domestic Product ratio may result in negative macroeconomic performance, like crowding out of investment, financial system instability, inflationary pressures, exchange rate fluctuations and more importantly adverse effects on economic growth. In fact the theoretical literature has summarized 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. There are also certain social and political implications of unsustainable debt burden. Persistent and high public debt calls for a large piece of budgetary resources for debt servicing. Consequently, the government is forced to cut allocations for other public services and it faces serious difficulties in executing its electoral manifesto, if it has.

Although there is a substantial literature on the impact of external debt on growth, relatively few studies have focused on low-income countries, particularly small island developing estates per se. This paper thus aims at analyzing the impact of external debt on the economic growth of Mauritius over the period 1960-2004. This study is based on the small island state of Mauritius and the latter provide a good case study because as most low income countries, it does not have access to international capital markets and thus the impact of external debt on growth on these economies can be different as compared to emerging market countries. Thus island has indeed been facing some troubling public debt problems since the last decade. In fact the island public debt ratio average is around 55% and its debt servicing as a percentage of recurrent revenue to about 25%, thus indicating some concerns for the government.

External debt does not only affect growth a priori, but countries with better economic performance may also better deal with the external debt phenomenon. In fact higher economic growth in turn increases a country's creditworthiness and this may attract more capital inflows. If the capital inflow is long term or Foreign Direct Investment (FDI), the need to borrow may decrease. Moreover external debt may have indirect effects through private and public investment through the debt overhang and crowding out effects. Moreover one should also not ignore the indirect effects of debt accumulation and service through private investment (debt overhang) and public spending (crowding out). Thus given the possibility of endogeneity and important feedback effects, we use dynamic time series analysis, namely a Vector Autoregressive framework, to analyse the hypothesized link. Such a framework will allow important insights on the role of external debt on not only economic growth but ultimately on private and public investment as well. The remainder of the paper is divided in three sections: section 2 provides a brief theoretical and empirical survey of the literature; section 3 describes the data employed and outlines the estimation methods and reports the results; and section 4 draws conclusions.

2. LITERATURE REVIEW

Debt theoretically affects growth through the following channels namely debt overhang, liquidity constraint, fiscal effect, productivity suppression and reduction in human capital accumulation (see Krugman, 1988; Savvides 1992; Agenor and Montiel, 1996; Serven, 1997 and Moss et al, 2003).

2.1 Related Work

There have been several attempts to empirically assess the external debt-economic growth link, the debt overhang and crowding out effects, mainly by using OLS. Most of the earlier empirical studies include a fairly standard set of domestic, debt, policy and other exogenous explanatory variables and the majority found one or more debt variables to be significantly and negatively correlated with investment or growth (see Krugman, 1988; Borensztein, 1990; Greene and Villanueva, 1991; Deshpande, 1997 and more recently Pattillo, Poirson, and Ricci ,2004. Among developing countries evidences supporting the debt overhang hypothesis features research from Iyoha (1996), Fosu (1999), Mbanga and Sikod (2001), Maureen (2001) and Clements, Bhattacharya, and Nguyen (2003). There exist a scant amount of research investigating the link between external debt and growth taking into account the causality and endogeneity issues. Among the few are those of Levy and Chowdhury (1993) who reported that level of indebtedness affect the GNP negatively and also indirectly through discouraging domestic investment. Metwally and Tamaschke (1994) also showed that there is a statistically significant two way relationship between the debt stock and debt service. Recently, Schclarek (2004) used GMM dynamic panel econometric technique to confirm the negative and a significant relationship for developing countries (but not for industrial countries).

In contrast, there also exist few studies could also not established negative relationship between debt and growth, for instance Cohen's (1993) for developing countries (LDCs). Djikstra and Hermes (2001) review a number of studies on the 'debt overhang' hypothesis and concluded that the empirical evidence might be inconclusive.

A summary of literature would tend to suggest that the majority of existing empirical literature report a negative effect of external debt on growth. However studies on developing countries case studies, particularly for small island developing states, have been limited and an even more scant number of works have been taking into account the endogeniety and dynamic issues in the modeling of the hypothesis.

3. THE ECONOMETRIC SPECIFICATION

3.1 The economic model

We follow the literature (see Fosu, 1999, Iyoha, 2000 among others) in specifying the following reduced-form growth model augmented with a debt variable to assess the impact of external debt on growth.

y = f (PRISTOCK,PUBSTOCK,XMGDP,SER,PDGDP) (1)

where PRISTOCK and PUBSTOCK are the country's private and public capital stock respectively and has been computed using the Perpetual Inventory Method (PIM) as recommended by the OECD (2001a), XMGDP is total of export and imports divided by the GDP of the country and is measure of openness, and SER is the secondary enrolment ratio and proxies for the quality of human capital and PDGDP is the external public debt ratio. The latter has been a widely employed indicator of the external debt stock burden and is defined as the value of the stock of external debt as a share of GDP. Two other indicators, namely the net present value (NPV) of the stock of external debt as a share of GDP (NPVEXGDP) and the debt servicing as a share of recurrent revenue (DEBTSER), were also used to consolidate the results. The capital stock of the country has been segregated into its two components namely private and public capital stock for further analysis and more insights on the effect of debt on each type of capital.

The main sources of our independent variables are from the World Bank's the Central Statistical Office and the International Financial Statistics'(IFS) except for the case of SER, where the country's Central Statistical Office's biannual digest of Statistics has been consulted. The dependent variable output was proxied by the real Gross Domestic Product per capita at constant price (Y) and was generated from the IFS. The time period of the study is over the years 1960-2004. The study importantly uses dynamic econometrics namely a Vector Autoregressive model (VAR), which takes into account the possibility of dynamic feedbacks among the variable in the model.

3.2 The Econometric Model and preliminary tests

Recall equation 1 above and taking logs on both sides of the equation and denoting the lowercase variables as the natural log of the respective uppercase variable results in the following:

y = [[beta].sub.0] + [[beta].sub.1]pristock + [[beta].sub.2]pubstock + [[beta].sub.3]xmgdp + [[beta].sub.4]ser + [[beta].sub.5]pdgdp + [epsilon] (2)

Before considering the appropriate framework of the econometric model, it is important to investigate the univariate properties of all data series and to determine the degree to which they are integrated. Both the augmented Dickey-Fuller (ADF) and Phillips-Perron (PP) unit-roots tests have been employed for that purpose and the results shows that all the variables are integrated of order 1 (I(1)) and thus stationary in difference. Further analysis in term of cointegration using the Johansen Maximum Likelihood approach has been undertaken. The Schwarz Bayesian criterion (SBC) suggested a VAR of order 2 and the tests based on Maximum Eigenvalue and the Trace test for cointegration both revealed that there was one cointegrating vector for the specification (same was found when substituting the other public debt variables namely npvexgdp and debtser).

3.3 Reverse Causality and Endogeneity Issues: The VAR framework

An often overlooked fact in the literature is that the relationship between foreign debt and economic growth may not be a one-way relationship. It is assumed that excessive debt affects a country's economic development in several ways as discussed above, but the relationship may be endogeneous encompassing reverse causality as well.

There is a consensus that in its early development stage, a country needs a substantial inflow of external foreign financing in order to fill the savings and foreign exchange gap. If the country attracts capital or is able to borrow from abroad, it can ease the foreign exchange shortage and provide a source for necessary imported goods for investment. When investment increases, economic growth also increases. Higher economic growth in turn increases a country's creditworthiness and this may attract more capital inflows. If the capital inflow is long term or Foreign Direct Investment, the need to borrow may decease. When the need to borrow decreases the growth rate of the debt stock will decline in the following period. Given that the debt service directly depends on the debt stock, this implies potentially higher domestic investment, accordingly higher growth and hence higher creditworthiness and more capital inflow. This interrelationship implies that the possible link between debt economic growth should be analyzed with a system of equation that accounts for endogeneity and feedback effects.

Moreover it has also been argued that that growth in investment facilitates economic growth through the accelerator principle (Fosu, 1996). In fact, debt can additionally influence economic growth via its effect on the productivity of investment. And even if debt service payments do not reduce savings and investments significantly, 'they could still decrease output growth directly by diminishing productivity as a result of the adverse change in investment mix' (Fosu, 1996). The impact of external indebtedness, therefore, simultaneously affects investment and economic growth. In other words, estimating only the growth equation would underestimate the effect of external indebtedness on economic growth since investment has been found to be a key determinant of growth.

To take into account the above possibility, we use a VAR framework for the econometric modeling. We follow Johansen (1988) and Enders (1995) by writing the following specification

[Z.sub.t] = [[PSI].sub.1][Z.sub.t-1] + [[PSI].sub.2][Z.sub.t-2] + .... [[PSI].sub.k][Z.sub.t-k] + [mu] + [[eta].sub.t] t=1.... t (3)

Where [Z.sub.t] = vector of (n x k) dimension

[[PSI].sub.k] = vector of (n x n) dimension

[[eta].sub.t] = Vector of unanticipated impulses (movements in [X.sub.t]) [??] niid(0,[SIGMA]), where n is the number of variables in the VAR, k is the dimension of the VAR, and t is time.

For the present analysis, the VAR consists of 6 endogeneous variables (n=6), [Z.sub.t] = [[y.sub.t], [pristock.sub.t], pubstock, [xmgdp.sub.t], [ser.sub.t], [pdgdp.sub.t]] and a constant term. [z.sub.t] is thus a 6 x 1 vector containing the above variables. Small letters denote that the variables are in natural logarithmic terms. The system features 2 lags (k=2) that were chosen using SBC. Since the variables are I (0) after applying the difference filter once and the specification is cointegrated with one cointegrating vector, we impose this constraint upon our unrestricted VAR to enable a Vector Error Correction Model (VECM) formulation. The short run dynamics can be studied using the following general VCEM:

[DELTA][Z.sub.t] = [[GAMMA].sub.1][Z.sub.t-1] + [[GAMMA].sub.2][DELTA][Z.sub.t-2].... + [[GAMMA].sub.k-1][DELTA] [Z.sub.t-k-1] + [PI][Z.sub.t-k] + [mu] + [[eta].sub.t] t=1.... t (4)

Where [DELTA] [Z.sub.t] is the vector of growth rates of the above five variables, and the [GAMMA] s are estimable parameters, [DELTA] is a difference operator, [[eta].sub.t] is as defined above. [PI] is the long run parameter matrix with rank equal to r), the number of cointegrating vectors (in our case it is one). With r cointegrating vectors (r=1), [PI] can be decomposed as [PI] = [alpha]'[beta], with [alpha] and [beta] both being 6 x 1 matrices. [alpha] has been defined as the adjustment or loading coefficients which measure the strength of the cointegrating vectors in the VECM or in other words the speed of adjustment. The [beta]'s are parameters in the cointegrating relationship and represent the long run coefficients.

3.4 Empirics and Analysis

The estimates of [alpha] and [beta] are presented in table 1. The long run parameter of public debt is reported to be significantly and negatively associated to the output of the country with an output elasticity of -0.168. This result may suggest that a 1% increase in public debt ratio is accompanied with a 0.17% reduction in the output level of the country (the two other external debt indicators also yielded a negative relationship (-0.11 and -0.23 respectively) between external debt and growth thus reinforcing the hypothesised link.

The findings are consistent with the literature with Iyoha (1996) and Fosu (1999) for Sub Saharan African countries. Clements, Bhattacharya, and Nguyen (2003) also reported elasticities in the same range for developing countries. As expected both capital stocks have positive effects on output level with private stock being more productive. Openness and human capital are associated with positive and significant output coefficient and are consistent with recent empirical evidences for developing economies.

3.4 Estimates of the Error-Correction Model

Engle and Granger (1987) showed by the error-representation theorem that cointegrated variables implies in effect an error correction model (ECM). He argued that regression of the first difference of cointegrated variables would result in misspecification error. Accordingly, the VAR was accordingly formulated in a Vector Error Correction model (VECM) to analyse the dynamics of the relationship. This involves the inclusion of the lagged errors of the cointegrating regression as one of the independent variables in the regression equation.

The Error Correction Model (see Engle and Granger, 1987) was then considered and estimated. The estimated error-correction equations are not subject to residual autocorrelation at the 5% significance level and the regression results appear in Table 2. We also tested for weak exogeneity of the variables using a Wald test and confirmed that the variables are not weakly exogenous at the 5% critical value and thus the same model specification is used. The variables in the system are also all endogenous, given that the lagged error-correction terms in all the equations of the VECM are significant.

The results from the table (column 1) confirms that public debt do explain short term variations in the output level of the country. The short run output elasticity of 0.103 and suggest that a 1 percentage-point increase in the growth rate of external public debt leads to a 0.103 percentage-point decrease in the growth rate of output after one year. This is lower as compared to the long run parameter and suggests that it might take some time for increasing debt to have its full in an economy. Private and public capital, openness and quality of labour are all significant and have the expected signs. It is also reported that 41% of the disequilibrium is corrected in the next period which shows an average adjustment speed.

The estimate shown above gives thus the direct effect of public debt on output in the short-run. There are some indirect effects as well. From the table it can be interpreted that a 1 percentage-point increase in the growth rate of public debt of the country leads to around 0.14 percentage-point reduction in the growth rate of private capital after one year, thus confirming the debt overhang hypothesis (Greene and Villanueva, 1991; Serven, 1997; Fosu, 1999 and Chowdhury, 2001 reported similar results). A 1 percentage-point increase in the growth rate of private capital is also seen to lead to around 0.65 percentage-point increase in the growth rate of output after one year. The latter two pieces of information taken together imply that a 1 percentage-point increase in the growth rate of public debt leads to a 0.1 percentage-point decrease in the growth rate of output after two years. This is an estimate of the indirect effect of transport capital on output in the short-run via the private capital channel. Moreover, indirect effect through the public capital accumulation link is also observed. In fact 1 percentage-point increase in the growth rate of public debt of the country leads to around 0.1 percentage-point reduction in the growth rate of public capital after one year. This is an indication of important crowding out effects and is consistent with Savvides (1992). On the other hand a 1 percentage-point increase in the growth rate of public capital is also seen to lead leads to around 0.18 percentage-point increase in the growth rate of output after one year and the total indirect effect through the crowding out effect may be estimated to be 0.18 percentage point.

Furthermore, education, as proxied by the secondary enrolment ratio, is also negatively influenced by the debt level of the country with a reported coefficient of -0.07. This is also in line with the study from Stephens (2001) who, for 24 African HIPCs, found that each additional US$1 in debt service results in US$ .33 decrease in education spending. Interestingly, from the column 7 of the above table, it can be observed that the public debt level of the country is also affected by the economy's output implying important feedback and reverse causality effects in the modeling of debt growth relationship. This is consistent with recent studies from Metwally and Tamaschke (1994), Olgun et al. (1998) and Schclarek (2004) amongst others who accounted for the possible endogeneity issues.

Using the ECM based causality approach, test on various pairs of causal effects were performed. Results from the ECM based causality are in general in line with those from the VECM approach. Moreover impulse response analysis performed also confirmed the above findings (Detailed analysis of impulse response functions is available from the authors upon request).

4. SUMMARY OF RESULTS

The paper investigated the relationship between external public debt and the economic performance for the case of the small island developing state of Mauritius over the period 1960-2004. Given the existence of endogeneity and feedback issues in the relationship, a VECM was specified to investigate if public debt explains both the long run and short run variation in output level. The results suggest that external debt have been negatively associated with the output level of the economy in both short and long run. Interestingly bi-causality between public debt and economic development is also reported. Moreover there are 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. Error correction modeling confirmed the existence of a stable long-run relationship and moreover determined that a deviation from the long-run equilibrium following a short-run shock is corrected by about 50 per cent after each year. The above results highlight the adverse effect of external public debt on economic performance. They provide new evidences for the case of island economies using recent cointegration approach in a dynamic framework.

REFERENCES

Agenor, P. and Montiel, P., Development Macroeconomics, Princeton, New Jersey: Princeton University Press, 1996.

Borenzstein, E., "Debt Overhang, Credit Rationing and Investment", Journal of Development Economics, 32, 1990, 315-335.

Clements, B., Bhattacharya, R. and Nguyen, T.Q., "External Debt, Public Investment, and Growth in Low-Income Countries", Working Paper WP/03/249, International Monetary Fund 2003.

Cohen, D., "Low Investment and Large LDC Debt in the 1980s", American Economic Review, 83 (3), 1993, 437-449.

Desphande, A., "The Debt Overhang and the Disincentive to Invest", Journal of Development Economics, 52, 1997, 169-187.

Dijkstra, G and Hermes, N., "The Uncertainty of Debt Service Payments and Economic Growth of Highly Indebted Poor Countries: Is There a Case for Debt Relief?" unpublished manuscript (Helsinki: United Nations University), 2001.

Enders, W., Applied econometric time series, John Wiley & Sons, Inc, 1995.

Engle, R. E., and Granger, C.W.J., "Co-integration and error correction: representation, estimation, and testing", Econometrica, Vol. 55, No 2, 1987, 251-276.

Fosu, A. K., "The Impact of External Debt on Economic Growth in Sub-Saharan Africa", Journal of Economic Development, Vol.12, No. 1, 1996, 34-46

Fosu, A. K., "The External Debt Burden and Economic Growth in the 1980s: Evidence from Sub-Saharan Africa", Canadian Journal of Development Studies, Vol. XX, No. 2, 1999, 307-18.

Greene, J., and Villanueva, D., "Private Investment in Developing Countries" IMF Staff Papers, Vol. 38, No. 1 (March 1991), 33-58 (Washington: International Monetary Fund).

Iyoha, M.A., "Macroeconomic management of Nigeria's external sector in the post-SAP period", Nigerian Journal of Economic and Social Studies, 38(1), March 1996.

Johansen, S., "Statistical Analysis of cointegration vectors", Journal of Economic Dynamics and Control, Vol. 12, 1988, 231-54

Krugman, P., Financing vs. Forgiving a Debt Overhang, Journal of Development Economics, Vol. 29, 1988, 253-268.

Levy, A and Chowdhury, K., "An integrative analysis of external debt, capital accumulation and production in Latin America, Asia-Pacific and Sub-Saharan Africa", Journal of Economics and Finance, 17(3), 1993, 105-119.

Melese, G. D, "External Debt and Economic Growth in Ethiopia", IDEP e-Newsletter, United Nations African Institute for Economic Development and Planning (IDEP), Volume 2. No4, October-December 2004

Metwally, M.M and Tamaschke, R., "The interaction among foreign debt, capital flows and growth; case studies", Journal of Policy Modelling, 16 (6), 1994, 597-608.

Pattillo, C., Poirson, H., & Ricci, L., "What Are the Channels Through Which External Debt Affects Growth?" IMF Working Paper, WP/04/15, 2004

Schclarek, A., "Debt and Economic Growth in Developing and Industrial Countries", Unpublished, 2004

Serieux, J and Samy Y., "The Debt Service Burden and Growth: Evidence from Low-Income Countries", The North-South Institute Working Paper, August 2001.

Serven, L., "Uncertainty, Instability, and Irreversible Investment: Theory, Evidence and Lessons for Africa", World Bank Policy Research Working Paper No. 1722, 1997. (Washington: World Bank).

Savvides, A., "Investment Slowdown in Developing Countries during the 1980s: Debt Overhang or Foreign Capital Inflows", Kyklos, Vol. 45, No. 3, 1992, 363-78.

Stephens, M., "External Debt, Government Spending and Growth in Heavily Indebted Poor Countries", unpublished Ph.D. thesis (New York University), 2001.

Boopen Seetanah, University of Technology, MAURITIUS

Kesseven Padachi, University of Technology, MAURITIUS

Ramesh Durbarry, University of Technology, MAURITIUS

AUTHOR PROFILE

Boopen Seetanah is a lecturer in Economics and Finance at the University of Technology, Mauritius. His research interests are in transport and development economics, and in the modelling of tourism demand.

Kesseven Padachi is a fellow of the Chartered Association of Certified Accountants, UK and has more than 10 years practical experience from both the private and public sector. Currently he is a lecturer in Accounting and Finance at the University of Technology, Mauritius. His research interests are in corporate finance and short-term financial management, with emphasis on the SMEs.

Ramesh Durbarry earned his Ph.D at the University of Nottingham and is currently an Associate Professor and Head of School at the School of Public Sector Policy and Management at the University of Technology, Mauritius.
TABLE 1: ESTIMATES OF LONG RUN PARAMETERS ([alpha] and [beta] vectors)

Variables   [beta]       t-ratios   [alpha]      t-ratios

y               1                   -0.345 ***    -3.51
pristock    -0.89 ***    -3.78      -0.464 ***    -3.67
pubstock    -0.32 **     -2.23      -0.243 *      -1.93
xmgdp       -0.743 **    -2.33      -0.325 **     -2.32
ser         -0.523 ***   -4.23      -0.353        -1.48
Pdgdp        0.168 **     2.15      -0.634 **     -2.24

* significant at 10%, ** significant
at 5%, *** significant at 1%

TABLE 2: ESTIMATES OF THE ERROR CORRECTION MODEL.

                                 [DELTA]      [DELTA]
Variables           [DELTA]y     pristock     pubstock

Constant            -1.11 *      1.36 **      -0.72 *

[DELTA]             -0.121       0.164 *       0.174
[y.sub.t-1]

[DELTA][pristock     0.653 ***   0.753 ***     0.023
.sub.t-1]

[DELTA][pubstock     0.178 **    0.122 **      0.675 **
.sub.t-1]

[DELTA][xmgdp        0.346 **    0.122 **      0.134
.sub.t-1]

[DELTA][ser          0.268 **    0.142 **      0.138
.sub.t-1]

[DELTA][pdgdp       -0.103 **    -0.142 *     -0.096 *
.sub.t-1]

[V.sub.t-1]         -0.411 ***   -0.464 ***   -0.243 *

[R.sup.2]            0.635       0.723         0.523

                    [DELTA]      [DELTA]      [DELTA]
Variables            xmgdp         ser         pdgdp

Constant             1.832        1.36 *       1.45 **

[DELTA]              0.145        0.246 *     -0.172 **
[y.sub.t-1]

[DELTA][pristock     0.114 *      0.175 *     -0.072 *
.sub.t-1]

[DELTA][pubstock     0.112        0.334 *     -0.026
.sub.t-1]

[DELTA][xmgdp        0.745 *      0.121       -0.112 *
.sub.t-1]

[DELTA][ser          0.156 *      0.754 *      0.024
.sub.t-1]

[DELTA][pdgdp        0.067       -0.071 *     -0.763 **
.sub.t-1]

[V.sub.t-1]         -0.325 **    -0.353       -0.634 **

[R.sup.2]            0.532        0.332        0.562

* significant at 10%, ** significant at 5%, *** significant at 1%
Gale Copyright:
Copyright 2007 Gale, Cengage Learning. All rights reserved.