Determining the Optimal Level of External Debt and Debt-Growth Relation: A Case Study of Malaysia

Proceedings of the Annual Vietnam Academic Research Conference on Global Business, Economics, Finance & Social Sciences (AP16Vietnam Conference) ISBN ...
Author: Gloria Cobb
0 downloads 1 Views 716KB Size
Proceedings of the Annual Vietnam Academic Research Conference on Global Business, Economics, Finance & Social Sciences (AP16Vietnam Conference) ISBN 978-1-943579-92-1 Vietnam, 7-9 August, 2016 Paper ID: V662

Determining the Optimal Level of External Debt and Debt-Growth Relation: A Case Study of Malaysia Qaiser Munir, Faculty of Business, Economics and Accountancy, University Malaysia Sabah. Email: [email protected] Winnie Abdul Nasir, Faculty of Business, Economics and Accountancy, University Malaysia Sabah. Kok Sook Ching, Faculty of Business, Economics and Accountancy, University Malaysia Sabah.

___________________________________________________________________________

Abstract This paper attempts to empirically determine the optimal level of external debt and investigate the relationship between external debt and economic growth in Malaysia over the period 1970-2013. The main finding in this paper is that the estimated threshold value of the country’s external debt ranged from 50 percent to 60 percent of GDP. Cointegration test is used to ascertain the long-run equilibrium relationship between the variables which confirms that there is cointegration between external debt and economic growth in the long-run. Granger (1969) causality test suggests a unidirectional causality running from economic growth to external debt. There is no evidence of debt-overhang in the country but crowding out effect does exist. In this contribution, the sustainable level of external debt in Malaysia is determined. In addition, we offer insight of the impact of external debt to the country’s economic growth. ___________________________________________________________________________ Key Words: Optimal level of external debt, Economic growth, Developing countries JEL Classification: C22, F34, F43; O10; O40

1

www.globalbizresearch.org

Proceedings of the Annual Vietnam Academic Research Conference on Global Business, Economics, Finance & Social Sciences (AP16Vietnam Conference) ISBN 978-1-943579-92-1 Vietnam, 7-9 August, 2016 Paper ID: V662

1. Introduction A major macroeconomic issue that has raised much attention of policy makers and researchers is related to external debt and its impact on economic growth. Like many other developing countries, Malaysia is facing the challenges of global economic slowdown and increase in global competition. External borrowing is considerable important to support the growth of domestic economy. In this study, we are motivated to identify Malaysia’s external debt sustainable level and recognize the long-run relation and short-run causalities that may exist between external debt and economic growth. The remaining of this paper is organised as follows: Section 2 provides a brief review of related literature. Section 3 discusses data and methodology. Section 4 explains the empirical findings. Section 5 concludes.

2. Literature Review It is important for a country to keep its external debt within sustainable level. Beyond the sustainable level, external debt may adversely affect economic growth. At the optimal level of external debt, a country is expected to effectively transform external borrowings into investment which in turn, stimulate a more rapid economic growth. Pattillo et al. (2002) investigate the nonlinear impact of external debt on economic growth using panel data for 93 developing countries, and find that the average impact of external debt becomes negative at around 160-170 percent of exports and 35-40 percent of GDP. Clements et al. (2003) study a sample of 55 low income countries throughout the period 1970-1999 in examining the effect of external debt on economic growth. It is found that beyond certain threshold level at around 20-25 percent of external debt to GDP, higher external debt levels are associated with lower economic growth rates. Reinhart and Rogoff (2010) find that emerging markets have low external debt sustainable level which is around 60 percent, while in advanced economies is at around 90 percent. Jayaraman and Lau (2009) examine whether external debt contributes to economic growth in Pacific Island Countries (PICs) and obtain results showing the coefficients of external debt are positive and significant. In the long-run, there is no evidence of causality between real GDP and external debt. In the short run, there is significant causal link from external debt to real GDP. Butts et al. (2012) suggest that real short-term external debt and real GDP are positively correlated as well as co-integrated. There is also evidence of Granger causality running from real GDP to real short-term external debt. Abdul Rahman (2012) examines the impact of external debt on economic growth over a short time span from the first quarter of 2000 until the fourth quarter of 2011, and confirms no evidence found in supporting the significance impact of external debt on economic growth either in the short-run or long-run.

2

www.globalbizresearch.org

Proceedings of the Annual Vietnam Academic Research Conference on Global Business, Economics, Finance & Social Sciences (AP16Vietnam Conference) ISBN 978-1-943579-92-1 Vietnam, 7-9 August, 2016 Paper ID: V662

3. Data and Methodology 3.1 Data We apply annual time series data covering a period of 44 years, from 1970 to 2013. The data are sourced from World Development Indicator (WDI) of World Bank 2015. The variables used are gross domestic product growth (LGDP); external debt to GDP (LEXTDGDP); gross fixed capital formation (LGFCF); general government final consumption expenditure (LGGFCE), export-to-GDP (EXPGDP), and debt service-to-GDP (LDSGDP). 3.2 Econometric Methodology 3.2.1 Stationarity Tests

The Kwiatkowski et al. (1992) (KPSS) test is usable to test whether the null of stationarity can be rejected. In other words, KPSS is the test for the null that

is I(0),

different from DF-GLS unit root test that test for the null hypothesis that a time series

is

I(1). KPSS can be used to confirm results from other unit root tests. To allow for the possibility of structural breaks, we employ the Lumsdaine and Papell (1997) two breaks test. Lumsdaine and Papell unit root test extended the Zivot and Andrews (1992) to address the flaws in the one-structural break unit root test. In particular, Model A and Model C in Zivot and Andrews (1992) are extended and called Model AA and CC respectively, where, model AA allows for two breaks in the intercept of trend and model CC allows for two breaks in the intercept and slope of the trend. Below show the models AA and CC: Model AA: k

yt  k  yt 1  t   DU 1t  DU 2t   d j yt  j   t

(1)

j 1

Model CC: k

yt  k  yt 1  t  1 DU 1t   1 DT 1t  DU 2t DT 2t   d j yt  j   t

(2)

j 1

3.2.2 Cointegration Tests

For the series that are integrated of same order, the next step is to perform cointegration test. The conventional cointegration test employed in this paper is ARDL Bound Testing Approach to Cointegration. The most critical aspect of ARDL approach is that it does not assume all variables are integrated of the same order. The variables analyzed can be integrated of order one I(1), or they can be stationary I(0) or the mixture of both. Past studies have shown that the outcomes of traditional cointegration tests are sensitive to structural breaks. For this reason, the Hatemi-J (2008) cointegration test is performed which accounts for two endogenously-determined breaks.

3

www.globalbizresearch.org

Proceedings of the Annual Vietnam Academic Research Conference on Global Business, Economics, Finance & Social Sciences (AP16Vietnam Conference) ISBN 978-1-943579-92-1 Vietnam, 7-9 August, 2016 Paper ID: V662

3.2.3 Causality Tests

Cointegration relationship usually implies the existence of causality, even though it does not point out the direction of the causality. Therefore, we employ the Granger (1969) causality test to identify the direction of possible causalities. Granger causality analysis can identify whether two variables move one after the other or contemporaneously. The causal relationship between economic growth and external debt can be evaluated by estimating the following regressions;

Where m and n represent the lag length and should be set equal to the longest time over which one series could be reasonable help to predict the other. represent economic growth and external debt, respectively.

and and

are uncorrelated

stationary random processes, and subscript t denotes the time period. 3.2.4 Threshold Robust Analysis

Threshold robust analysis is demonstrated in this present study. We follow Chudik et al. (2013) in determining the threshold value. To explore the importance of heterogeneity and potential nonlinearity in the external debt-economic growth nexus, firstly, we commence with the following baseline homogeneous time series data model;

Where

is a threshold dummy represented by the indicator variable

which takes the value of 1 if debt to GDP is above the given threshold value of , and zero otherwise.

4. Empirical Findings 4.1 Stationarity Tests In order to avoid misleading results, econometric theory requires that variables must be stationary before the application of standard econometric techniques. For this, all the series used are tested using the DF-GLS and KPSS unit root tests. The unit root test results of DFGLS and KPSS are presented in Table 1. The null hypothesis of DF-GLS is a time series contains a unit root, while KPSS has null hypothesis that time series is stationary. The rationale of using more than one traditional unit root tests is to ensure robustness of our results. Test statistics in the KPSS test are all significant indicating that the series are

4

www.globalbizresearch.org

Proceedings of the Annual Vietnam Academic Research Conference on Global Business, Economics, Finance & Social Sciences (AP16Vietnam Conference) ISBN 978-1-943579-92-1 Vietnam, 7-9 August, 2016 Paper ID: V662

nonstationary. In the first difference, all the series are significant in the DF-GLS test, but not significant in the KPSS test, implying that the series are integrated of order one, that is I(1). Table 1: Results of Traditional Unit Root Tests Test Statistics DF-GLS

KPSS

Conclusion

DF-GLS

A: Level

KPSS

Conclusion

B: First Difference

Model Specification: Intercept

Model Specification: Intercept

LGDP

0.597

0.842*

I(0)

∆LGDP

-5.501*

0.298

I(1)

LEXTDGDP

-0.641

0.529*

I(0)

∆LEXTDGDP

-3.953*

0.179

I(1)

LGFCF

-1.846

0.158*

I(0)

∆LGFCF

-3.902*

0.143

I(1)

LGGFCE

-1.863***

0.617*

I(0)

∆LGGFCE

-6.905*

0.176

I(1)

LEXPGDP

-0.944

0.699*

I(0)

∆LEXPGDP

-4.434*

0.304

I(1)

LDSGDP

-1.440

0.218*

I(0)

∆LDSGDP

-6.642*

0.277

I(1)

Model Specification: Trend and intercept

Model Specification: Trend and intercept

LGDP

-1.289

0.155**

I(0)

∆LGDP

-5.840*

0.051

I(1)

LEXTDGDP

-2.039

0.141***

I(0)

∆LEXTDGDP

-4.756*

0.092

I(1)

LGFCF

-2.225

0.154**

I(0)

∆LGFCF

-4.553*

0.090

I(1)

LGGFCE

-2.770

0.141***

I(0)

∆LGGFCE

-7.560*

0.142

I(1)

LEXPGDP

-1.278

0.129***

I(0)

∆LEXPGDP

-4.881*

0.177***

I(1)

LDSGDP -1.539 0.200** I(0) ∆LDSGDP -7.039* 0.123 I(1) Notes: ∆ denotes the first difference operator. Asterisks (*), (**), (***) denote the statistically significant at 1%, 5% and 10% significance levels respectively.

4.2 Unit Root Test with Structural Breaks We perform unit root test with two structural breaks proposed by Lumsdaine and Papell (1997) in order to verify the order of integration of each series. In Model CC, the null hypothesis of unit root is rejected in favour of stationarity for LGDP and LGFCFGDP, both at 1% level of significance. In the first difference, all the series in each model show rejections of unit root in favour of stationarity, at 1% level of significance in most of the series, and 10% level of significance in LGFCFGDP. In other words, in first the difference, all the series are stationary either at 1% or 10% levels of significance. In terms of break dates, the breaks in the series locate around the 1980’s and 1990’s which may correspond to the external shocks such as, Global Recession in 1985-1986 due to commodity price shock, the Asian financial crisis in 1997-1998. In 1970’s, the breaks observed may correspond to the oil price shock in 1979. Table 2: Results of Lumsdaine and Papell Unit Root Test Model AA: Break in intercept

Model CC: Break in trend and intercept

t-statistic

k

TB1

TB2

t-statistic

k

TB1

TB2

Conclusion

-3.482

0

1975

1989

-7.301*

0

1984

1997

I(1)

Level LGDP

5

www.globalbizresearch.org

Proceedings of the Annual Vietnam Academic Research Conference on Global Business, Economics, Finance & Social Sciences (AP16Vietnam Conference) ISBN 978-1-943579-92-1 Vietnam, 7-9 August, 2016 Paper ID: V662

LEXTDGDP

-6.275*

5

1980

1993

-6.449

5

1984

1996

I(1)

LPDEBT

-4.655

3

1980

1992

-4.764

3

1989

1998

I(0)

LGFCFGDP

-6.034

1

1988

1997

-6.776*

1

1985

1997

I(1)

LEXPGDP

-6.740

2

1988

2006

-5.893

2

1980

1997

I(0)

LGGFCE

-4.750

0

2000

2008

-6.245

0

1979

2000

I(0)

LDSGDP

-4.105

1

1975

1983

-4.535

1

1984

2001

I(0)

∆LGDP

-7.059*

0

1980

1997

-8.219*

0

1987

1997

I(1)

∆LEXTDGDP

-8.057*

1

1986

1992

-8.894*

1

1984

1990

I(1)

∆LPDEBT

0

1987

1997

-7.092*

0

1987

1997

I(1)

∆LGFCFGDP

-6.456* 6.068***

0

1996

1999

-6.594***

0

1996

1999

I(1)

∆LEXPGDP

-8.332*

1

1986

2000

-8.573*

1

1984

2000

I(1)

∆LGGFCE

-8.613*

0

1979

1982

-8.843*

0

1979

1984

I(1)

1st Difference

∆LDSGDP -9.152* 0 1988 1991 -10.515* 0 1978 1985 I(1) Notes: The 1%, 5% and 10% critical values obtained from estimating model AA are -6.74, -6.16 and 5.89 respectively. The 1%, 5% and 10% critical values obtained from estimating model CC are -7.19, 6.75, and -6.48 respectively. *, **, and *** imply rejection of the null hypothesis of non-stationary at 1%, 5% and 10% significance level, respectively. TB is the estimated break year and k stands for the endogenously selected lag order for the minADF test. The lag is selected using Akaike Information (AIC) criteria.

4.3 Cointegration Test We use ARDL approach due to its simplicity and it is recommended for a small sample size (Pesaran et al. 2001, and Ghatak and Siddiki, 2001). The results reported in Table 3 show that there is evidence of cointegration when both LEXTDGDP and LGDP are taken as dependent variables. This is supported by the computed F-statistics which are statistically significant at the 1% level. The computed values are higher than critical values. Thus, the results strongly reject the null hypothesis of no long-run relationship between the variables in both models. Having found the long-run relationship among the variables, the next step is to estimate the long-run effects. The long-run estimation results are shown in Table 4. Table 3: Result of ARDL Bound Testing Cointegration Approach Regressor

F-statistic value

Optimal Lag length

Conclusion

LEXTDGDP

5.694*

3

There is a long-run relationship

LGDP

6.832*

3

There is a long-run relationship

Bound critical values by Pesaran et al. (2001) I (0)

I (1)

1%

4.3

5.23

5%

3.38

4.23

10%

2.97

3.74

Notes: * Rejection of null hypothesis of no cointegration at 1%. The optimal lag length for both models chosen based on Akaike Information Criteria (AIC)

6

www.globalbizresearch.org

Proceedings of the Annual Vietnam Academic Research Conference on Global Business, Economics, Finance & Social Sciences (AP16Vietnam Conference) ISBN 978-1-943579-92-1 Vietnam, 7-9 August, 2016 Paper ID: V662

Table 4: Results of ARDL Long-run Estimation Coefficient St. Error t-Statistic Regressor: LEXTDGDP LGDP LGFCFGDP

Prob.

*6.9977 0.3840

2.4678 0.4593

2.8356 0.8361

0.0094 0.4117

-2.6153 **-0.3691

0.7959 0.1394

-3.2858 -2.6472

0.0032 0.0144

Regressor: LGDP LEXTDGDP *-0.1133 0.0196 -5.7729 LGFCFGDP *0.3712 0.0363 10.2417 LDSGDP *0.2463 0.0589 4.1774 LGGFCE *0.0561 0.0017 32.6831 Note: Asterisk *, and ** denotes significant level at 1% and 5% respectively.

0.0000 0.0000 0.0002 0.0000

LDSGDP LGGFCE

Based on Table 4, all coefficients in both models are significant except for LGFCFGDP, and all have mixed signs. In a model where LGDP is used as regressor, LEXTDGDP has a negative effect on LGDP and is significant at the 1% level. On the other hand, LGFCFGDP and LDSGDP are found to have a positive effect on LGDP. Similarly, LGGFCE also has positive effect on LGDP. The final step of the ARDL approach is the error correction to estimate the short-run parameter with the speed of adjustment. The short-run estimation results are shown in Table 5. As can be seen from the results, the error correction

coefficient is negative as expected, and is very significant

indicating that the models will converge to the long-run equilibrium regardless of the regressor used. When external debt is used as dependent variable in the long-run estimation, the coefficient of the variable LGDP is positive and significant means that economic growth positively affects external debt. If economic growth increase, external debt also increases. This finding is similar with the findings of Abdelhadi (2013) and Mohd Daud et al. (2013) in which the accumulation of external debt is associated with an increase in economic growth up to the optimal level. Table 5: Results of ARDL Short-run Estimation Coefficient

St. Error

t-Statistic

Prob

Regressor: LEXTDGDP ∆ LGDP

*-2.1001

0.4077

-5.1508

0.0000

∆ LGFCFGDP

**0.3079

0.1485

2.0731

0.0496

∆ LDSGDP

*-0.7057

0.1781

-3.9619

0.0006

∆ LGGFCE

*-41.1081

6.1044

-6.7342

0.0000

*-0.2729

0.0405

-6.7422

0.0000

*-0.1228

0.0191

-6.4428

0.0000

Regressor: LGDP ∆ LEXTDGDP ∆ LGFCFGDP

*0.2026

0.0293

6.9126

0.0000

∆ LDSGDP

0.0223

0.0438

0.5089

0.6145

∆ LGGFCE

*8.1595

1.3022

6.2661

0.0000 7

www.globalbizresearch.org

Proceedings of the Annual Vietnam Academic Research Conference on Global Business, Economics, Finance & Social Sciences (AP16Vietnam Conference) ISBN 978-1-943579-92-1 Vietnam, 7-9 August, 2016 Paper ID: V662

*-0.3714

0.0597

-6.2223

0.0000

Note: Asterisk *, and ** denotes significant level at 1% and 5% respectively. On the other hand, the signs of the coefficients of external debt when GDP is used as dependent variable are consistent with priori expectation that external debt and economic growth have a negative relationship in the short-run and long-run. The negative relationship between external debt and economic growth is consistent with the findings of Borensztein (1990a), Borensztein (1990b), Elbadawi et al. (1997), Butts et al. (2009), Iqbal and Zahid (1998) and Ramzan and Ahmad (2014). In addition, when LGDP is used as regressor, debt service-to-GDP is found to have a positively significant impact on GDP growth in the longrun, but positive and insignificant impact on GDP in the short-run. The results of the ARDL bounds test may be unreliable once the series has break point. The lack in ARDL bounds test motivates us to apply the structural break cointegration tests by Hatemi-J (2008). It has been accepted extensively in the econometric literature that traditional cointegration tests that do not take account of potential structural breaks will have low power. The rejection of null hypothesis of no cointegration is on modified models while modified

statistics in all

statistics confirm the cointegration existence in Model C/T and

Model C/S. The existence of a cointegration relationship between external debt and economic growth indicates that the series are moving together in the presence of structural breaks. Table 6: Results of Hatemi-J (2008) Test for Cointegration between External Debt and Economic Growth

Test statistic

Estimated test value

Break Dates

1% Critical value

5% Critical Value

10% Critical value

(i) Model: Change in Level, c *-7.816

1989, 1997

-8.353

-7.903

-7.705

-5.724

1986, 1993

-8.354

-7.903

-7.705

-39.177 1986, 1993 (ii) Model: Change in Level and Trend, c/T

-140.135

-123.870

-116.169

**-7.912

1986, 1986

-8.353

-7.903

-7.705

*-7.799

1985, 1986

-8.354

-7.903

-7.705

1985, 1986

-140.135

-123.870

-116.169

***-9.199

1987, 1997

-8.353

-7.903

-7.705

*-7.824

1985, 1994

-8.354

-7.903

-7.705

-50.135

1980, 1994

-140.135

-123.870

-116.169

-45.348 (iii) Model: Intercept and Slope, c/s

Notes: The approximate asymptotic critical values for tests of cointegration with two regime shifts are taken from Hatemi-J (2008) with 4 number of independent variables, m =4. The lag

8

www.globalbizresearch.org

Proceedings of the Annual Vietnam Academic Research Conference on Global Business, Economics, Finance & Social Sciences (AP16Vietnam Conference) ISBN 978-1-943579-92-1 Vietnam, 7-9 August, 2016 Paper ID: V662

length was selected using downward t-statistics with a maximum lags of 8. *, **, *** denote the rejection of the null hypoth0esis at 10%, 5% and 1% level respectively. The break dates located from the Hatemi-J (2008) cointegration test may correspond to shocks arose from Global Recession and the Asian financial crisis. These global events may have caused the shifting of the long-run relationship between the variables. 4.4 Causality Test Results Following Engle and Granger (1987), if there is cointegration exist between two variables, there must be Granger causality between them either a unidirectional or bidirectional relationship. Since there exists at least one cointegrating vector in the previous ARDL bound cointegration test, it is best to estimate causal links between external debt and economic growth. Table 7 summarizes the results of pairwise Granger causality test between GDP and external debt. The main result indicates that in the short-run, LGDP is causing LEXTDGDP at the 10% significance level, and the LEXTDGDP has no impact on LGDP in the short-run. Thus, we find that there exists a short-run unidirectional linkage between Malaysia’s economic growth and external debt, running from GDP to external debt. This finding is consistent with Butts et al. (2012). Table 7: Results of Pairwise Granger Causality between External Debt and GDP Pairwise Null Hypothesis

Obs

LGDP does not Granger Cause LEXTDGDP 37 LEXTDGDP does not Granger Cause LGDP

F-statistics

P-value

***2.347

0.0596

1.965

0.1069

Inference Reject H0 (Unidirectional Causality) No Granger causality

Note: *** indicates the rejection of the null hypothesis at 10% significance level. 4.5 Threshold Robust Analysis The results of robustness analysis are shown in Table 8. Following Chudik et al. (2013), we estimate the threshold validity on the basis of splitting the threshold values into arbitrary threshold brackets of 20% until 60% for external debt. The estimates of of across different values of

are obtained. The estimates of

and

for values

that are quite stable

, in line with the rather small estimates obtained for

particular, all the coefficients of

. In

are significant at the 1% significance level. The

differences between average GDP growth above a certain debt to GDP ratio and below the same threshold level are relatively flat over a range of values for . The estimates of

show

that the whole average GDP growth declines when external debt to GDP ratio increases, as shown by the negative signs of the coefficients. There is a significant tipping point that lies between 50% and 60% of external debt to GDP ratio, a 1% and 5% significance levels respectively for both 50% and 60%, where long-term growth is cut down substantially.

9

www.globalbizresearch.org

Proceedings of the Annual Vietnam Academic Research Conference on Global Business, Economics, Finance & Social Sciences (AP16Vietnam Conference) ISBN 978-1-943579-92-1 Vietnam, 7-9 August, 2016 Paper ID: V662

Table 8: Estimates of the Average External Debt Threshold Effects on Output Growth Ƭ

20% 30% OLS Estimates with ), where -0.020 (0.017) 0.079* (0.016)

-0.017 (0.012) 0.074* (0.011)

40%

50%

60%

-0.003 (0.010) 0.079* (0.008)

-0.035* (0.011) 0.071* (0.006)

-0.034** (0.016) 0.065* (0.006)

Notes: The estimates are based on the following specifications: where

is the log of GDP growth. The heteroscedasticity-robust

standard errors are reported in the parentheses. *, ** denotes significant level at 1% and 5% respectively. Table 9 presents the estimates of the average external debt threshold effects on economic growth with other variables in the model. The results show that the threshold estimates only found to be significant from 50% until 60% of external debt, implying that the threshold turning point lies between 50%-60% of debt ratio, in which the negative impact of debt on growth starts taking effect. Table 9: Estimates of the Average External Debt Threshold Effects on Output Growth with Other Variables Ƭ OLS Estimates

GFCFGR EXPGR GGFCEGR DSGDP R2

20%

30%

40%

50%

60%

-0.008 (0.015) 0.065* (0.016) 0.195* (0.062) 0.053 (0.068) -0.158** (0.596) -0.012 (0.016)

-0.011 (0.011) 0.067* (0.010) 0.187* (0.061) 0.034 (0.071) -0.168* (0.062) -0.014 (0.015)

-0.018 (0.008) 0.070* (0.007) 0.165** (0.067) 0.024 (0.063) -0.164* (0.054) -0.016 (0.014)

-0.018** (0.009) 0.064* (0.005) 0.173* (0.061) 0.057 (0.062) -0.151** (0.059) -0.009* (0.001)

-0.021** (0.010) 0.061* (0.005) 0.194* (0.064) 0.057 (0.064) -0.151** (0.060) -0.008* (0.001)

0.53

0.54

0.58

0.57

0.57

*, ** denotes significant level at 1% and 5% respectively. The heteroscedasticity-robust standard errors are reported in the parentheses.

5. Conclusion This paper is an attempt to determine the optimal level of external debt in Malaysia using annual data spanning from 1970 to 2013. This paper also aims to assess empirically the relationship between external debt and economic growth, with other growth determinants. After confirming that all the variables are stationary, ARDL bound cointegration test is carried out and reveals that there is a long-run relationship between external debt and 10

www.globalbizresearch.org

Proceedings of the Annual Vietnam Academic Research Conference on Global Business, Economics, Finance & Social Sciences (AP16Vietnam Conference) ISBN 978-1-943579-92-1 Vietnam, 7-9 August, 2016 Paper ID: V662

economic growth. The cointegration test of Hatemi-J (2008) with two structural breaks is employed to capture the effect of possible external shocks in the long-run. The test results suggest that there exists a long-run relationship between external debt and economic growth, and significant break points are found which may correspond to the Global Recession 1985/86, and the Asian Financial crisis in 1997-1998. The results of Pairwise Granger causality analysis suggest a unidirectional causality from economic growth to external debt. The estimated threshold value of Malaysia’s external debt using robustness test is in the range between 50 and 60 percent. In other words, external debt is likely to start rendering negative impact on economic growth around the threshold level in the range 50-60 percent of external debt-to-GDP.

References Abd Rahman, M. H., 2012, How Federal Government’s Debt Affect the Level of Economic Growth? International journal of Trades, Economics and Finance 3(4), 323-326. Abdelhadi, S. A., 2013, External Debt and Economic Growth: Case of Jordan (1990-2011). Journal of Economics and Sustainable Development 4(18), 26-33. Borensztein, E., 1990a, Debt Overhang, Credit Rationing and Investment. J. Dev. Econ 32(2), 315-335. Borensztein, E., 1990b, Debt Overhang, Debt Reduction and Investment: The Case of Philippine. IMF Working Paper No. 569252. Butts, H. C., 2009, Short Term External Debt and Economic Growth-Granger Causality: Evidence from Latin America and the Caribbean. Rev Black Polit Econ 36, 93-111. Butts, H. C., Mitchell, I., and Berkoh, A., (2012), Economic Growth Dynamics and Short-Term External Debt in Thailand. The Journal of Developing Area, 46(1), 91-111. Central Bank of Malaysia. Quarterly Bulletin, Second Quarter Report 2013. Kuala Lumpur: Central Bank of Malaysia, 2013. Chudik, A., Mohaddes, K., Pesaran, M. H., and Raissi, M., 2013, Debt, Inflation and Growth. International Monetary Fund, IMF. Clements B., Bhattacharya R., and Nguyen, T. Q., 2003, External Debt, Public Investment, and Growth in Low-Income Countries. IMF Working Papers, 03(249). Cohen, B. D., 1993, Low Investment and Large LDC Debt in the 1980’s. The American Economic Review 83(3), 437–449. Ghatak S., and Siddiki, J., 2001, The use of ARDL Approach in Estimating virtual exchange rates in India. J Appl Stat 28(5), 573-583. Granger, C. W. J., 1969, Investigating Causal Relations by Econometric Models and Across Pectral Methods. Econometrica 37, 424-439. Gregory, A. W., and Hansen, B. E., 1996a, Tests for Cointegration in Models with Regime and Trend Shifts. Oxford Bulletin of Economics and Statistics 58, 555-560. Hatemi-J, A., 2008, Tests for Cointegration with Two Unknown Regime Shifts with An Application to Financial Market Integration. Empirical Economic, 497–505. doi:10.1007/s00181-007-0175-9 Iqbal, Z., and Zahid, G. M., 1998, Macroeconomic Determinants of Economic Growth in Pakistan. Pak. Dev. Rev 37(2), 125-148. Jayaraman, T. K., and Lau, E., 2009, Does external debt lead to economic growth in Pacific island countries. Journal of Policy Modeling 31(2), 272–288. doi:10.1016/j.jpolmod.2008.05.001 11

www.globalbizresearch.org

Proceedings of the Annual Vietnam Academic Research Conference on Global Business, Economics, Finance & Social Sciences (AP16Vietnam Conference) ISBN 978-1-943579-92-1 Vietnam, 7-9 August, 2016 Paper ID: V662

Lumsdaine, R. L., and Papell, D. H., 1997, Multiple Trend Breaks and the Unit Root Hypothesis. Review of Economics and Statistics 79(2), 212-218. Ministry of Finance. Economic Report, Kuala Lumpur: Ministry of Finance, various issues. Pattillo, C., Poirson, H., and Ricci, L., 2002, External Debt and Growth. Internatioanl Monetary Fund, IMF. Perron, P., 1989, The Great Crash, the Oil Price Shock, and the Unit Root Hypothesis. Econometrica 57, 1361-1401. Pesaran, H. M., Shin, Y., and Smith, R. J., 2001. Bounds Test Approaches to the Analysis of Level Relationships. J. Appl. Econometrics 16, 289–326. Ramzan, M., and Ahmad, E., 2014, External Debt Growth Nexus: Role of Macroeconomic Policies. Economic Modelling 38, 204–210. doi:10.1016/j.econmod.2013.12.014 Reinhart, B. C. M., and Rogoff, K. S., 2010, Growth in a Time of Debt. American Economic Review 100, 573–578. Reinhart, C. M., Reinhart, V. R., and Rogoff, K. S., 2012, Public Debt Overhangs: Advanced-Economy Episodes Since 1800. Journal of Economic Perspectives 26(3), 69–86. doi:10.1257/jep.26.3.69 Sachs, J. D., 1989, “The Debt Overhang of the Developing Countries” in: G. Calvo, R. Findlay, P. Kouri and J.B. de Macedo, eds., Debt, Stabilization, and Development, Basil Blackwell, Cambridge, MA:80-102. Sen, A., 2003, On Unit Root Tests Whtn The Alternative is a Trend Break Stationary Process. Journal of Business and Economic Statistics 21(1), 174-184. Zivot, E. and Andrews, K., 1992, Further Evidence On The Great Crash, The Oil Price Shock, and The Unit Root Hypothesis. Journal of Business and Economic Statistics 10(10), 251–70.

12

www.globalbizresearch.org

Suggest Documents