This article was downloaded by: [University of Leeds] On: 22 September 2008 Access details: Access Details: [subscription

number 773557620] Publisher Routledge Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK

Applied Economics Letters

Publication details, including instructions for authors and subscription information: http://www.informaworld.com/smpp/title~content=t713684190

Financial development and economic growth: an ARDL approach for the case of the small island state of Mauritius
Boopen Seetanah a a University of Technology, Mauritius, Pointes aux Sables, Mauritius First Published:August2008

To cite this Article Seetanah, Boopen(2008)'Financial development and economic growth: an ARDL approach for the case of the small

island state of Mauritius',Applied Economics Letters,15:10,809 813

To link to this Article: DOI: 10.1080/13504850600770889 URL: http://dx.doi.org/10.1080/13504850600770889

PLEASE SCROLL DOWN FOR ARTICLE

Full terms and conditions of use: http://www.informaworld.com/terms-and-conditions-of-access.pdf This article may be used for research, teaching and private study purposes. Any substantial or systematic reproduction, re-distribution, re-selling, loan or sub-licensing, systematic supply or distribution in any form to anyone is expressly forbidden. The publisher does not give any warranty express or implied or make any representation that the contents will be complete or accurate or up to date. The accuracy of any instructions, formulae and drug doses should be independently verified with primary sources. The publisher shall not be liable for any loss, actions, claims, proceedings, demand or costs or damages whatsoever or howsoever caused arising directly or indirectly in connection with or arising out of the use of this material.

Applied Economics Letters, 2008, 15, 809813

Financial development and economic growth: an ARDL approach for the case of the small island state of Mauritius
Boopen Seetanah
University of Technology, Mauritius, Pointes aux Sables, Mauritius E-mail: [email protected]

Downloaded By: [University of Leeds] At: 15:35 22 September 2008

The article investigates the dynamic empirical link between financial development and economic performance for the case of the developing small island state of Mauritius using a unique time-series data set over the period 1952 to 2004. The analysis was performed using two different proxies for financial development in an ARDL framework. The results suggest that financial development have been contributing to the output level of the economy in both short and long run. It thus highlights the economic importance of financial development and provides new evidence for the case of island economies using recent cointegration approach.

I. Introduction The importance of the financial sector in the economic development has received much attention in the recent literature.1 A strong consensus has emerged in the last decade that well-functioning financial intermediaries have a significant impact on economic growth,2 see King and Levine (1993), Jayaratne and Strahan (1996), Ross and Levine (1997), Chowdhury (1997), Demirgu -Kunt and c Maksimovic (1998), Rajan and Zingales (1998), Levine and Zervos (1998), Neusser and Kugler (1998), Levine et al. (2000), Wachtel (2003). However until now most studies have been focused on developed countries cases and it is only lately that scholars have been implicitly dealing with the issue of causality and dynamics in the financial development

and economic growth link (Levine et al., 2000). Studies using time-series analysis for developing country cases have been scarce and to our knowledge no studies have been performed for the case of developing small island states. Empirical findings from developed countries cases are not directly applicable and relevant to island states given their special characteristics and vulnerability. The aim of the article is thus to investigate the empirical link between financial development and economic performance for the case of the developing small island state of Mauritius using a unique time-series data set over the period 1952 to 2004 and allowing for dynamics. It is hoped that the study will add new insights to and also to fill a gap in the literature. The structure of this article is as follows: Section II describes the preferred modelling function used,

Refer to King and Levine (1993) for a comprehensive theoretical overview. There are however some empirical works which could not establish a significant link, for instance, Dawson (2003), Ram (1999) among others.
2

Applied Economics Letters ISSN 13504851 print/ISSN 14664291 online 2008 Taylor & Francis http://www.informaworld.com DOI: 10.1080/13504850600770889

809

810
elaborates on the data collection and investigates the empirical link between financial development and economic growth for the case of Mauritius. Section III concludes and deals with policy implications.

B. Seetanah
PRIGDP isolates credit issued to the private sector, as opposed to credit issued to governments, government agencies and public enterprises. Furthermore, it excludes credits issued by the central bank. PRIGDP has been used extensively as an indicator because it improves on other measures of financial development (Levine et al., 2000). Higher levels of PRIGDP are interpreted as higher levels of financing services and therefore greater financial intermediary development. This proxy shall also be utilized to investigate and consolidate the financial development and economic growth. The main sources of our independent variables are from the World Banks and the IMFs International Financial Statistics (IFS) except for the case of SER, where the countrys Central Statistical Offices biannual digest of Statistics has been consulted. The dependent variable output was proxied by the real GDP per capita at constant price (Y) and was generated from the IFS. The time period of the study is over the years 1952 to 2004. The econometric model and preliminary tests Recall models 1 and 2 aforementioned 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 0 1 ivtgdp 2 xmgdp 3 ser 4 liqgdp " 1 y 0 1 ivtgdp 2 xmgdp 3 ser 4 prigdp " 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 DickeyFuller (ADF) (1979) and PhillipsPerron (PP) (1988) unit-roots tests have been employed for that purpose and the results are summarized in Tables 1 and 2 given further. Test for stationarity (refer to Tables 1 and 2) shows that all our variables are integrated of order 1 (I(1) and thus stationary in difference) except ivtgdp, the proxy for openness which is an I(0) variable. Testing for cointegration using the ARDL In this part we test the existence of a long-term relationship (cointegration) in using testing and estimation procedure advanced in Pesaran and Shin (1999). For the specification 1 and 2 above, the error correction versions of the ARDL model in the

II. Methodology and Analysis Causality and dynamic issues and ARDL model To allow for causality and dynamics and given that not all of our time-series are stationary to the same order (some are I(0) while others are I(1), this is discussed below), the recently developed cointegration procedure by Pesaran et al. (2001), the autoregressive distributed lag model (ARDL) procedure will be applied. In fact the procedure allows for different long-run relationships and short-run dynamics and this is important for the estimation of the equilibrium conditions. In addition, the Pesaran et al. (2001) technique can be implemented regardless of whether the variables are integrated of order (1) or (0) and can be applied to small finite samples. The economic model We follow the standard literature, see King and Levine (1993), Ross and Levine (1997), Levine and Zervos (1998), Levine et al. (2000), Wachtel (2001), in specifying the economic model subsequently. y fIVTGDP, XMGDP, SER, LIQGDP model 1

Downloaded By: [University of Leeds] At: 15:35 22 September 2008

y fIVTGDP, XMGDP, SER, PRIGDP model 2 where IVTGDP is the countrys investment divided by its gross domestic product (GDP; investment ratio), 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. To measure financial development we use monetary aggregates, more specifically LIQGDP (refer to model 1) which is the ratio of liquid liabilities to the countrys GDP. Liquid liabilities include currency plus demand deposits and interest-bearing liabilities of banks and nonbank financial intermediaries. This is a typical measure of Financial Depth and has been widely used (McKinnon (1973) and King and Levine (1993)). The second widely used measure, PRIGDP (refer to model 2), has been the value of credits by financial intermediaries to the private sector divided by GDP. This measure of financial development is more than a simple measure of financial sector size.

Financial development and economic growth

Table 1. Summary results of unit-root tests in level form: DickeyFuller and PhillipsPerron test Variables y ivtgdp xmgdp ser liqgdp prigdp Lag selection 1 1 1 1 1 1 Aug. DickeyFuller 1.46 3.58 2.13 0.91 1.22 1.22 Phillips Perron 2.59 2.284 2.83 2.91 0.56 0.56 Critical value 2.924 2.924 2.924 2.924 2.924 2.924 Variable type I(1) I(0) I(1) I(1) I(1) I(1) Aug. DickeyFuller (time trend, t) 2.2 4.34 3.11 1.18 0.79 0.79 Critical value 3.51 3.51 3.51 3.51 3.51 3.51

811

Variable type I(1) I(0) I(1) I(1) I(1) I(1)

Table 2. Summary results of unit-root tests in first difference: D/F and PhillipsPerron test Variables (in log) y xmgdp ser liqgdp prigdp Lag selection 0 0 0 0 0 Aug. DickeyFuller 8.57 8.77 4.23 4.75 4.23 PhillipsPerron 8.78 5.29 3.74 2.97 3.74 Critical value 2.936 2.936 2.936 2.936 2.936 Variable type I(0) I(0) I(0) I(0) I(0) Aug. DickeyFuller (with time trend, t) 8.98 8.65 4.45 5.03 4.45 Critical value 3.508 3.508 3.508 3.508 3.508 Variable type I(0) I(0) I(0) I(0) I(0)

Downloaded By: [University of Leeds] At: 15:35 22 September 2008

variables y, invgdp, xmgdp, ser and liqgdp (model 1) and prigdp (model 2) are given respectively by y 0
n X n X i1

b i y t i

n X i1

For model 1, the hypothesis that is being tested is the null of nonexistence of the long run relationship defined by Ho : 1 2 3 4 5 0 And the alternative hypothesis is H1 : 1 6 0; 2 6 0; 3 6 0; 4 6 0; 5 6 0 The recommended statistic is the F-statistics for the joint significance of 1, 2, 3, 4 and 5. Computation of this F-statistic requires running the following regression yt 0 byt1 civtgdpt1 dxmgdpt1 esert1 fliqgdpt1 "t 5

ci ivtgdpti ei serti

di xmgdpti

n X i1

i1 n X i1

fi liqgdpti 1 yt1 2 ivtgdpt1

3 xmgdpt1 4 sert1 5 liqgdpt1 "t 3 y 0

n X i1 n X i1 n X i1

mi yti

n X

pi ivtgdpti ri serti

qi xmgdpti

i1 n X i1

and a variable addition test is subsequently made by including the following. 1 yt1 , 2 ivtgdpt1 , 3 xmgdpt1 , 4 sert1 , 5 liqgdpt1 It should be however be noted that the distribution of the F-statistic is nonstandard, irrespective whether regressors are I(0) or I(1). Pesaran et al. (1996) have tabulated the appropriate critical values for different number of regressors and whether the regressors contain an intercept or a time trend. The F-Statistics F(y/ivtgdp, xmgdp, ser, liqgdp) turned out to be 6.34 (This is also confirmed by the maximum eigen values and trace values of the Johansen test for cointegration) and exceeds the upper bound of the critical value band. We thus reject

si liqgdpti 1 yt1 2 ivtgdpt1

3 xmgdpt1 4 sert1 5 liqgdpt1 "t 4 Since we have annual observations, we chose n 1 for the maximum order of lags in the ARDL model in both cases and carry out the estimation over the period of study. In fact the same lag length was chosen when using the final prediction error due to SBC.

812
Table 3. Estimated long run coefficients based on ARDL approach Regressor invtgdp xmgdp ser liqgdp prigdp Constant Coefficient SBC (1, 0, 1, 0, 1) Model 1 0.919*** 0.69*** 0.401*** 0.131* 7.38*** t-ratio 14.35 2.97 2.96 1.93 1.84 5.83 Coefficient SBC (1, 0, 1, 0, 1) Model 2 0.861*** 0.731*** 0.361*** 0.100** 0.622***

B. Seetanah

t-ratio 10.23 4.18 3.43 2.12 3.57

Dependent variable is y. * significant at 10%, ** significant at 5%, *** significant at 1%.

Table 4. Error correction representation for the selected ARDL model Regressor ivtgdp xmgdp ser liqgdp ECM(1) R-square DW Coefficient AIC (1, 0, 1, 0, 1) Model 1 0.596*** 0.187** 0.162** 0.154** 0.403*** 0.526 1.81 t-ratio 4.12 1.6 2.75 2.00 4.21 Coefficient SBC (1, 0, 1, 0, 1) Model 2 0.584*** 0.187* 0.152** 0.065* 0.415*** 0.58 1.76 t-ratio 4.06 1.69 3.23 1.66 4.43

Downloaded By: [University of Leeds] At: 15:35 22 September 2008

Dependent variable is y. * significant at 10%, ** significant at 5%, *** significant at 1%.

the null hypothesis of no long-run relationship between the variables irrespective of their order. The test results thus suggest that there is a long-run relationship between the variable. Substituting liqgdp by prigdp as the alternative proxy for financial development and replicating the same procedure as above yielded an F-Statistics F(y/ ivtgdp, xmgdp, ser, prigdp) of 4.54 which implied the existence of cointegration in this alternative specification as well. Estimation results. Given that both specifications are cointegrated, the unrestricted error correction representation of the ARDL model is given by Equations 3 (for the case of liqgdp) and 4 (for the case of prigdp) respectively. The next stage of the procedure would be to estimate the coefficients of the long-run relations and the associated error correction model (ECM) using the ARDL approach. The order of the distributed lag on the dependent variable was selected by the Schwartz Bayesian Criterion (SBC)3 and turned out to be one. The SBC criteria select the ARDL (1, 0, 1, 0, 1) for both models respectively. The long-run estimated coefficients are shown in the Table 3.
3

The point estimates are not too different under both specifications. It is observed that financial development may have contributed positively to the output of the country in the long run. In fact a 1% increase in the liquid liabilities to GDP ratio (liqgdp) is associated with a 1.3% increase in output level. The link is also confirmed by using prigdp as the proxy and it was estimated to be slightly lower at 0.1. As expected, for the case of Mauritius, the investment level and the extent of openness of the economy have been the main ingredients for economic development. The quality of human capital is also reported to have been an important factor. From Table 4 both models suggest that the impact of financial development on the output of Mauritius has been positive and significant. The coefficients for the other explanatory variables are well-behaved and have the expected sign and significance. Moreover, the coefficient of the ECM of the selected ARDL (1, 0, 1, 0, 1) is negative and highly significant at 1% level. This confirms the existence of a stable long-run relationship and points to a long-run co-integration relationship between variables. The ECM represents the speed of adjustment to restore equilibrium in the dynamic model following a disturbance. The coefficient of the ECM is around 0.4 in both cases and

Pesaran and Smith (1998) found that SBC is preferable to AIC, as it is a parsimonious model that selects the smallest possible lag length, while AIC selects the maximum relevant lag length.

Financial development and economic growth

this implies that a deviation from the long-run equilibrium following a short-run shock is corrected by about 40% after each year.

813
Jayaratne, J. and Strahan, P. (1996) The finance-growth nexus: evidence from bank branch deregulation, Quarterly Journal of Economics, 111, 63970. King, R. and Levine, R. (1992) Financial indicators and growth in a cross section of countries, Working Paper No. 819, World Bank, Washington DC. King, R. G. and Levine, R. (1993) Finance and growth: schumpeter might be right, Quarterly Journal of Economics, 108, 71737. Levine, R. and Zervos, A. (1998) Stock markets, banks, and economic growth, American Economic Review, 88, 53758. Levine, R., Loayza, N. and Beck, T. (2000) Financial intermediation and growth: causality and causes, Journal of Monetary Economics, 46, 3177. Mc Kinnon, R. (1973) Money and Capital in Economic Development, Brookings Institution, Washington DC. Neusser, K. and Kugler, M. (1998) Manufacturing growth and financial development: evidence from OECD countries, Review of Economics and Statistics, 80, 63848. Pesaran, M. and Smith, P. (1998) Structural analysis of cointegrating VARs, Journal of Economic Survey, 12, 471505. Pesaran, M. H. and Shin, Y. (1999) An autoregressive distributed lag modelling approach to cointegration analysis, Chapter 11, in Econometrics and Economic Theory in the 20th Century (Ed.) S. Storm, The Ranger Firsch Centermial Symposium, Cambrigde University Press, Cambridge. Pesaran, M. H., Shin, Y. and Smith, R. (1996) Testing for the existence of a long-run relationship, DAE Working Papers 9622, Department of Applied Economics, University of Cambridge. Pesaran, M. H., Shin, Y. and Smith, R. (2001) Bounds testing approaches to the analysis of level relationships, Journal of Applied Econometrics, 16, 289326. Phillips, P. C. B. and Perron, P. (1988) Testing for a unit root in time series regression, Biometrica, 75, 33546. Rajan, R. G. and Zingales, L. (1998) Financial dependence and growth, American Economic Review, 88, 55986. Rajan, R. and Zingales, L. (2000) The great reversals: the politics of financial development in the 20th century, OECD Economics Department Working Papers, No 265, OECD Publishing. Ram, R. (1999) Financial development and economic growth: additional evidence, Journal of Development Studies, 35, 16474. Wachtel, P. (2003) How much do we really know about growth and finance? Federal Reserve Bank of Atlanta Economic Review, 88, 3347.

IV. Summary of Results The article investigated the relationship between financial development and the economic performance for the case of the developing small island state of Mauritius over the period 1952 to 2004. To account for the fact that not all of the series were integrated to the same extent and also to incorporate dynamics in the analysis, an ARDL approach was used. The analysis was also performed using two different proxies for financial development and the results suggest that financial development have been contributing to the output level of the economy in both short and long run. Error correction modelling was used to confirm the existence of a stable long-run relationship and moreover determined a deviation from the long-run equilibrium following a short-run shock, which is corrected by about 40% after each year. The above results highlight the importance of financial development and moreover provide new evidences for the case of island economies using recent innovative cointegration approach.

Downloaded By: [University of Leeds] At: 15:35 22 September 2008

References
Chowdhury, A. (1997) The financial structure and the demand for money in Thailand, Applied Economics, 29, 4019. Dawson, P. J. (2003) Financial development and growth in economies in transition, Applied Economics Letters, 10, 8336. Demirgu -Kunt, A. and Maksimovic, V. (1998) Law, c finance, and firm growth, Journal of Finance, 53, 210739. Dickey, D. A. and Fuller, W. A. (1979) Distributions of the estimators for autoregressive time-series with a unit root, Journal of the American Statistical Association, 75, 427831.