Islamic finance and economic growth: The Malaysian experience

Salina Kassim

PII: S1044-0283(16)30042-4
DOI: doi: 10.1016/j.gfj.2015.11.007
Reference: GLOFIN 338

To appear in: Global Finance Journal

Received date: 17 March 2015

Revised date: 25 May 2015
Accepted date: 9 November 2015

Please cite this article as: Kassim, S., Islamic finance and economic growth: The
Malaysian experience, Global Finance Journal (2016), doi: 10.1016/j.gfj.2015.11.007

This is a PDF file of an unedited manuscript that has been accepted for publication.
As a service to our customers we are providing this early version of the manuscript.
The manuscript will undergo copyediting, typesetting, and review of the resulting proof
before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that
apply to the journal pertain.
 ACCEPTED MANUSCRIPT

Islamic Finance and Economic Growth: the Malaysian Experience

Salina Kassim
Associate Professor
IIUM Institute of Islamic Banking and Finance
International Islamic University Malaysia

T
50728 Kuala Lumpur
Malaysia

R IP
Abstract
The increasing presence of Islamic banking and finance in Malaysia’s

SC
financial sector and the country’s exemplary role in the global Islamic finance
industry has called for an evaluation of the contribution of Islamic finance to
the real economic activity. In order to empirically investigate the impact of

NU
Islamic finance on performances of major macroeconomic indicators, this
study applies the ARDL approach on quarterly data set for Malaysia covering
the period from 1998 to 2013. The results suggest that Islamic finance has
MA
started to make important contributions to the real economy by effectively
carrying-out the financial intermediation role of pooling and channeling funds
to the investment activities. In view of the important contributions of Islamic
finance to the Malaysian economy, continuous efforts need to be undertaken
to further expand the industry. This includes refinement of the legal and
D

regulatory framework to enable healthier growth of the industry, thus further

TE

strengthens Malaysia’s position as the leader in the Islamic finance industry at

the global level.
P

Keywords: finance-growth nexus, Islamic banking and finance, real

economy, Malaysia, ARDL
CE
AC

1. INTRODUCTION

Islamic finance was considered as a new phenomenon in the global financial scenario in the

1970s, starting with the establishment of the first Islamic commercial bank, namely Dubai

Islamic Bank in 1975. Since then, the Islamic finance industry continues to record remarkable

progress. Today, total Islamic banking asset is estimated to reach US$1.7 trillion, recording an

annual growth of 17.6% since 2000 (Bank Negara Malaysia, 2014). Amid the rapid growth of

the industry globally, Malaysia has been at the forefront with total Islamic asset accounting

for nearly 13% of the global Islamic finance industry and 8% of the global Islamic banking

assets (SESRIC, 2012; Ernst and Young, 2013). At the domestic front, the Islamic banking

1
 ACCEPTED MANUSCRIPT

asset stood at US$125 billion in 2012, equivalent to 20% of the country’s total banking sector

asset. The sector also continued to stage impressive growth of 24% per annum in the period

2008-2012 despite the extremely difficult financial and economic scenario in view of the

T
global financial crisis which started in the US in 2007/2008. The sector is equally active in

IP
terms of financial intermediation as the Islamic banks’ financing and deposits grew at an

R
average rate of 23% and 32%, respectively over the period of 2008 to 2012 (Ernst and Young,

SC
2013).

Given the high level of the Islamic banking sector assets, its increasing market share

NU
as well as its active financial intermediation role, both in Malaysia and globally, it is timely
MA
and highly relevant to assess the contribution of Islamic finance to the real economy. This

need is particularly important in the case of Malaysia, where Islamic finance is being

positioned to play an increasingly important role in the economy and currently accounts for a
D
TE

substantial share of the banking sector assets. More importantly, the country has frequently

been set as a benchmark for many other countries aspiring to develop their Islamic banking
P

and finance sectors.

CE

Despite the strategic role of Islamic finance to the Malaysian economy and the
AC

important position it occupies in the global Islamic finance industry, only a few studies have

been focusing on the Islamic financial sector in relation to the real economy. Most of these

studies either adopted a bi-variate approach or tend to omit other important macroeconomic

variables, which may influence the relationship between Islamic finance and the real

economy. Consequently, this study seeks to fill the literature gap by investigating the

relationship in a multi-variate context by including not just the Islamic banking indicators and

economic output variable, but other variables that potentially have significant influence on the

development of the Islamic finance industry, particularly in the Malaysian economic setting.

Additionally, the study hopes to shed light on the nature of short- and long-run relationships

2
 ACCEPTED MANUSCRIPT

between Islamic finance and the economic activity, thus contributing towards better policy

design and practice of Islamic banking. Findings from this study would enable policymakers

and industry-players to evaluate their financial policy options and business strategies with the

T
objective of increasing the contribution of Islamic finance to the real economic activities.

R IP
2. LITERATURE REVIEW

SC
The relationship between finance and economic growth has been one of the intensely

researched topics, particularly in the context of conventional finance. As early as the seminal

NU
work of Bagehot (1873), finance has been shown to have positive impact on the real
MA
economic activity as money market development enabled the flow of capital to productive

trades in England during the time. Schumpeter (1911, 1934) argued that financial

intermediaries (banks, in particular) occupy the center-stage in economic development. These

D
TE

views were solidified theoretically by the financial repression hypotheses advanced by

McKinnon (1973) and Shaw (1973). Later, the advent of endogenous growth models provided
P

further insights and basis for the empirical investigation on the role of finance in economic
CE

growth process with some of the notable empirical studies which supported the finance-
AC

growth nexus include King and Levine (1993), Levine and Zervos (1998), Rajan and Zingles

(1998), and Hassan, Sanchez and Yu (2011).

Despite the rich literature on finance-growth nexus, economists remained divided on

the nature of the relationship between the two. Consequently, four major hypotheses have

emerged on this issue, namely the supply-leading hypothesis, demand-following hypothesis,

mutual dependence hypothesis and neutrality hypothesis. The supply-leading hypothesis,

which is rooted in the Schumpeter views and championed by McKinnon (1973) and Shaw

(1973) in their financial repression theory which propose that financial development leads to

economic growth. On the other hand, the demand-following hypothesis was pioneered by

3
 ACCEPTED MANUSCRIPT

Robinson (1952) who asserts that “finance follows growth leads”, implying that growth in the

real economy leads to financial development. The mutual dependence hypothesis was

advanced by Patrick (1966) which postulates that the causality between financial development

T
and economic growth is reciprocal. It argued that in under-developed countries, finance leads

IP
growth whereas in highly developed economies, finance becomes increasingly demand-

R
following. Finally, Lucas (1988) advocated the neutrality hypothesis, which contends that the

SC
finance-growth relationship is unnecessarily over-stressed as it is not necessarily an important

factor in the growth process.

NU
While Islamic banking and finance has started to make in-roads in more than 75
MA
countries globally, studies on its role in contributing to the real economy in these countries

have remained relatively few. With the increasing presence and significance of Islamic

finance, several studies focus on the unique role that it may play in the economic growth
D
TE

process. An ideal Islamic financial system is predominantly equity-based, suggesting a close

link between the financial sector and real economic sector, by this the causal relationship
P

between financial development and economic growth is supposed to be bi-directional. A

CE

developed Islamic financial system will mobilise more investment fund and allocate same to
AC

firms, thereby enhancing investment and real sector growth. A healthy and profitable real

sector will yield more return to the Islamic banks, thus enhancing their development.

Moreover, fostering economic growth (social benefit) shall be a core role of all Islamic

financial institutions as is the case of other institutions of an Islamic economy, where the

socio-economic harmony among agents is one of the economic goals (Anonym, 2010).

A major characteristic of an Islamic financial system that enables it to uniquely

contribute to economic growth process is the profit-sharing principle, which foster equity in

income distribution leading to social justice and long-term economic growth. It also improves

the efficiency of capital allocation since return to capital depends on its productivity.

4
 ACCEPTED MANUSCRIPT

Moreover, Islamic financial institutions encourage investment as investment depositors

receive a share of the banks’ profit. Overall, due to less emphasis on debt financing, the

Islamic financial system is supposed to be more stable than its conventional counterpart,

T
therefore, the loss of capital and wealth due to frequent financial crises and their negative

IP
impacts on the growth process are avoided, leading to more sustainable economic growth

R
(Goaied and Sassi, 2010). Despite the perceived roles of Islamic financial institutions on

SC
economic growth, studies in the area have been mostly theoretical and conceptual. There were

few empirical studies due to data scarcity and the fact that there are handful of countries with

NU
Islamic financial system and in most of them, the Islamic financial system is small relative to
MA
the dominant conventional system.

One of the theoretical studies was conducted by Nagaoka (2011) who explores the

potentialities of Islamic finance to contribute to economic growth of modern economy and the
D
TE

Islamic world. By theoretically examining the distinctive features of Islamic financial

products compared to those of the conventional finance, Islamic finance is found to embed the
P

monetary sector into the real sector (embedded financial system), which is described as
CE

universal financial system characterised by a great potential of sustainable development. It

AC

can be inferred from this that the close relationship between Islamic financial institutions and

the real economy will enable them contribute more effectively to economic development than

their conventional counterpart. In the same vein, Khoutem and Nedra (2012) show that

Islamic participative financial intermediation can enhance economic growth, by resolving ex-

ante and ex-post asymmetric information problems on the basis of profit and loss sharing

principle. It also reduces transaction costs and permits risk sharing and all these lead to an

optimization of the saving-investment process.

Empirically, Goaied and Sassi (2010) used an unbalanced panel data from 1962-2006

for 16 countries in the Middle East and North Africa (MENA) region and employed the

5
 ACCEPTED MANUSCRIPT

generalised method of moment (GMM) to investigate the effect of banking sector on

economic growth. Using credit advanced to private sector by Islamic banks to represent

financial intermediation, the overall results revealed that there is no significant relationship

T
between Islamic banking development and economic growth, in some instances the

IP
relationship was even significantly negative (specifically in the case of oil exporting

R
countries). The Islamic banks, in particular, exhibited weak association with economic

SC
growth. This result might have reflected the problem of “Dutch disease” commonly associated

with resource-rich countries.

NU
Similar conclusion was also arrived at by Barajas, Chami and Yousefi (2010) who
MA
found that the beneficial effects of financial deepening on economic growth differs between

oil exporting and non-oil exporting countries. The financial deepening process was found to

be considerably smaller in the oil-exporting countries compared to the rest of the world.
D
TE

However, combining Islamic and conventional banks in the same model as was done in the

study may not clearly isolate the impact of Islamic banks on growth, and there may be some
P

feedback effects between the two.

CE

Comparing the case of selected Asian countries and Gulf Cooperation Council (GCC)
AC

countries, Yusof and Bahlous (2013) assess the contribution of Islamic finance to economic

growth in Malaysia, Indonesia and selected GCC countries, which are among the earlier

countries that adopted the Islamic banking practices. Using annual data over the period 2000

to 2009 and employing panel cointegration, variance decompositions and impulse response

functions, the study finds that Islamic banking contributes to the growth process of these

countries, both in the short and long runs. However, in the short run, this contribution is

stronger in Malaysia and Indonesia than in GCC countries.

It is important to highlight that the above cross-sectional studies tend to give all

countries under consideration equal weight, in that they are assumed to be homogeneous,

6
 ACCEPTED MANUSCRIPT

likewise the estimated coefficients represent only an average relationship, which is not likely

to apply to individual countries in the sample (Aziakpono, 2003).

In a time series study, Abduh, Brahim and Omar (2012), investigated the link between

T
both Islamic and conventional financial development with the economic growth in

IP
Bahrain. Using quarterly data for the period 2000Q1-2010Q4 and employing the Johanson

R
and Juselius cointegration test and Vector Error Correction Model, the results show a

SC
significant positive relationship between Islamic finance and economic growth in the long

run. However, in the short run, no evidence was found for a relationship between them. Over

NU
all, Islamic finance was found to have a bi-directional relationship with economic growth in
MA
Bahrain. In contrast, the conventional finance was found to have a significant relationship

with economic growth, both in the long run and short run.

In the context of Indonesia, Abduh and Omar (2012) examine the relationship between
D
TE

Islamic banking development and economic growth using quarterly data covering the period

2003-2010 and employed the bound-testing approach of cointegration and error-correction

P

modeling.The results revealed that a significant relationship exists betweeen Islamic finance
CE

and economic growth, both in short-run and long-run. However, both studies adopt a direct
AC

method of relationship, i.e regressing Islamic bank financing and deposits on economic

growth without involving other growth determining factors, such as gross fixed capital

formation, openness, inflation, this may affect the robustness of their results.

In specific case of Malaysia, Furqani and Mulyany (2009) examined the relationship

between Islamic banking and economic growth in Malaysia; one of the leading countries in

Islamic finance. They employed co-integration test and vector error correction modeling

(VECM) on a time-series data covering the period from 1997 to 2004. Using total Islamic

bank financing as indicator of Islamic banking intermediation, the results revealed that there

is long-run bidirectional relationship between Islamic and fixed investment, with GDP causes

7
 ACCEPTED MANUSCRIPT

the growth of Islamic banks (consistent with the demand-following hypothesis). The results

support the theoretical postulations that Islamic banks enhance investment which if

productive, will also lead to their development.

T
Using an extended quarterly data from 1998 to 2012, Abdul Manap, Abduh and Omar

IP
(2012) examined the causal relationship between Islamic banking development and economic

R
growth as well as between Islamic banking development and capital formation in

SC
Malaysia. Using the Toda-Yamamoto and bootstrap Granger non-causality tests, the results

show that Islamic financial development causes economic growth but not the other way

NU
round. This implies that further development of Islamic finance in Malaysia will contribute to
MA
the economic growth.

The above studies, however, may suffer from statistical bias due to omitted variables

issue. The assumption that economic growth (GDP) to be determined either by Islamic
D
TE

financing only and or gross fixed capital formation is rather constraining in the statistical

sense. In the context of standard income determination model, the external sector plays a
P

significant role particularly for a small opened economy such as Malaysia where external
CE

trade plays a significant role in influencing the real economic activity. Moreover, the stability
AC

of the domestic macroeconomic environment (often measured by inflation) is likely to

influence both the development of the Islamic financial sector and the real economic

activities. Neglecting these realities is a limitation on the part of these studies. Therefore, the

current study aims to address this limitation by investigating the effects of Islamic finance on

the real sector of the Malaysian economy, while controlling for other growth determining

variables such as trade openness, government expenditure and inflation.

3. METHODOLOGY

3.1 Variables and Data Sources

8
 ACCEPTED MANUSCRIPT

In order to represent the real sector of the economy (which is the dependent variable in the

finance-growth nexus analysis), the industrial production index is used as a proxy. This is

appropriate given the theoretically anticipated link between Islamic finance and real economic

T
activities. As for the independent variables, two Islamic financial development indicators and

IP
several major macroeconomic indicators are considered. First, total deposit of the Islamic

R
banks is used to capture the ability of the Islamic banks to mobilise funds from the surplus

SC
units of the economy and second, total financing by Islamic banks is used as an indicator to

signify the ability of the Islamic banks to finance real economic activities. These indicators

NU
are adopted because they captured all the shariah-compliant financial instruments used by the
MA
Islamic banks to mobilize and allocate funds.

Gross fixed capital formation is included as a proxy for capital accumulation or

investment, which is one of the major channels through which financial intermediaries may
D
TE

influence growth. In order to avoid the problem of biasness due to omitted variables, other

variables are included in the model to control for the possible effects of other growth
P

determining factors. These are general government expenditure (representing the public
CE

sector), trade openness (representing the external sector) and inflation (as an indicator for
AC

macroeconomic stability). Government expenditure may lead to budget deficit, which if

financed by borrowing from the financial system has the potential of crowding-out private

investment, hence negatively affects growth. Alternatively, government spending if

effectively carried-out, may contribute positively to growth. Another macroeconomic

indicator, trade openness, is hypothesised to contribute positively to economic growth by

providing domestic entrepreneurs access to foreign markets. Inflation indicates the stability in

the economy, which in turn gives impact on consumption, savings and investment decisions,

hence the levels of deposits in and financing by the Islamic banks as well as economic

growth.

9
 ACCEPTED MANUSCRIPT

The selection of the variables is consistent with the existing studies in this area such as

Yusof and Bahlous (2013); Abdul Manap, Abduh and Omar (2012); Abduh, Brahim and

Omar (2012); Abduh and Omar (2012) and Furqani and Mulyany (2009), among other

T
studies. In terms of nature of data, this study utilises quarterly data for the period 1998Q1 to

IP
2013Q4, measured in local currency and at constant prices. The data for Islamic bank

R
development is sourced from the Monthly Statistical Bulletin of Central Bank of Malaysia -

SC
Bank Negara Malaysia, and the data for the remaining variables is obtained from the

International Monetary Fund’s International Financial Statistics database.

NU
MA
3.2 Estimating Model

This study adopts the cointegration technique to examine the relationship between Islamic

banking developments on economic growth. The method is chosen because of the tendency of
D
TE

the presence of unit root in most macroeconomic and financial time series data which can

result in spurious regression. Two principal approaches to cointegration have been widely
P

used in the literature, namely the two-step residual-based procedure for testing the null of no
CE

cointegration by Engle and Granger (1987) and the system-based reduced rank regression
AC

approach based on Johansen and Juselius (1990). However, an important condition for the

application of these methods is that the variables must be integrated of the same order, that is

integrated of order one I(1). This condition introduces a degree of uncertainty into the analysis

of level relationship.

Due to the constraints imposed by the Engle and Granger and Johansen and Juselius’s

approaches to co-integration, this research adopts the bounds testing approach to cointegration

based on Autoregressive Distributed Lag (ARDL) model framework as presented by Pesaran,

Shin and Smith (2001). Unlike the previous approaches, the ARDL model does not impose a

restricting assumption that all the variables must be integrated of the same order. Therefore,

10
 ACCEPTED MANUSCRIPT

the ARDL can be applied regardless of whether the variables are all I(0), I(1) or mutually co-

integrated (Pesaran et al., 2001). It also addresses the problems associated with omitted

variables and autocorrelations, in addition to providing unbiased and efficient estimates as

T
well as valid t-statistics even in the presence of endogenous regressors (Narayan 2004;

IP
Odhiambo, 2010).

R
The ARDL approach to cointegration involves the estimation of the restricted error

SC
correction (EC) version of the ARDL model, for this research, the models involving both

Islamic bank deposits and financing are respectively presented as follows:

NU
 ln( IPI ) t   0  1 ln( IBD ) t 1  2 ln(GFCF ) t 1  3 ln(OPN ) t 1  4 ln(GCE ) t 1
p p p
 5 INFt 1  6 ln( IPI ) t 1   1  ln( IBD ) t i    2  ln(GFCF ) t i    3  ln(OPN ) t i
MA
i 1 i 0 i 0
p p p
   4  ln(GCE ) t i    5 ( INF ) t i    6  ln( IPI ) t i   t (1)
i 0 i 0 i 0
D

 ln( IPI ) t   0  1 ln( IBF ) t 1  2 ln(GFCF ) t 1  3 ln(OPN ) t 1  4 ln(GCE ) t 1

TE

p p p
 5 INFt 1  6 ln( IPI ) t 1   1  ln( IBF ) t i    2  ln(GFCF ) t i    3  ln(OPN ) t i
i 1 i 0 i 0
P

p p p
   4  ln(GCE ) t i    5 ( INF ) t i    6  ln( IPI ) t i   t (2)
CE

i 0 i 0 i 0

Where p is the optimal lag length, IPI is the industrial production index, IBD is Islamic
AC

banks deposits, IBF is Islamic bank financing, GFCF is gross fixed capital formation, OPN is

trade openness, GCE is general government expenditure, INF is the rate of inflation and µt is

the disturbance term.

All the variables are in natural logarithm form, except for INF, which is a rate and

contains negative values, hence cannot be logged. F-test is conducted to detect if the variables

are cointegrated, that is if they have long-run relationship. The null hypothesis for both

models is H0: λ1=λ2=λ3=λ4=λ5=λ6=0, which is tested against the alternative hypothesis H1:

λ1≠λ2≠λ3≠λ4≠λ5≠λ6≠0.

11
 ACCEPTED MANUSCRIPT

However, the asymptotic distribution of the F-statistic is not standard for bounds test,

therefore, the computed F-statistic is evaluated using the critical values presented by Pesaran

et al. (2001). These critical values are of two sets, the lower bound critical values which

T
assume all the variables to be I(0) and the upper bound critical values, which assume all the

IP
variables to be I(1). The decision rule therefore is, if the computed F-statistic is less than the

R
lower bound value, then H0 cannot be rejected, hence we conclude that there is no long run

SC
relationship between the variables. On the other hand, when the computed F-statistic is greater

than the upper bound value, then we reject H0 and conclude that long run relationship exist

NU
between the variables being considered. If instead, the computed value falls in between the
MA
two bounds, then the results is inconclusive.

Upon establising long-run relationship among the variables, then a long run model is

estimated for both deposit and financing models respectively as follows:

D
TE

p p p
ln( IPI ) t   0   1 ln( IPI ) t i    2 ln( IBD ) t i    3 ln(GFCF ) t i
i 1 i 0 i 0
p p p
   4 ln(OPN ) t i    5 ln(GCE) t i    6 INFt i   t
P

(3)
i 0 i 0 i 0
CE

p p p
ln( IPI ) t   0   1 ln( IPI ) t i    2 ln( IBF ) t i    3 ln(GFCF ) t i
AC

i 1 i 0 i 0
p p p
   4 ln(OPN ) t i    5 ln(GCE ) t i    6 INFt i   t (4)
i 0 i 0 i 0

In order to get the short run coefficients, an error correction model (ECM) is estimated,

the ARDL specification of the ECM for the two models of this research is as follows:

p p p
 ln( IPI ) t   0   1  ln( IPI ) t i    2  ln( IBD ) t i    3  ln(GFCF ) t i
i 1 i 0 i 0
p p p
   4  ln(OPN ) t i    5  ln(GCE ) t i    6  ln INFt i  ECTt 1   t (5)
i 0 i 0 i 0

12
 ACCEPTED MANUSCRIPT

p p p
 ln( IPI ) t   0   1  ln( IPI ) t i    2  ln( IBF ) t i    3  ln(GFCF ) t i
i 1 i 0 i 0
p p p
   4  ln(OPN ) t i    5  ln(GCE ) t i    6 INFt i  ECTt 1   t (6)
i 0 i 0 i 0

T
Where, ECM is the error correction term. However, cointegration only implies causality

IP
but does not show its direction. According to Engle and Granger (1987) representation

R
theorem, if cointegration is found to exist among variables, causality relationship should be

SC
investigated within a dynamic error correction framework as represented in equations (5) and

(6). An error correction model provides the opportunity to separate the long run and short run

NU
Granger causality. The short run dynamics are captured in the specific coefficients of the
MA
lagged terms, while the error correction term contains information of the long run causality.

Hence, if the coefficient of each lag independent variable is significant, it signifies short term

causation and a negative and statistically significant error correction term signifies long run
D
TE

causality (Adebola, Yusoff and Dahalan, 2011).

At this juncture, it is worthy of note that the selection of optimal lag length is critical at
P

all the stages of analysis outlined above, usually, two major lag length selection criteria are
CE

employed. These are the Schwarz Bayesian Criterion (SBC) and Akaike Information Criterion
AC

(AIC). The SBC always select parsimonious model, this means it chooses the smallest

possible lag length, on the other hand, the AIC chooses the relevant maximum lag length. In

addition to this, the SBC-based models have lower prediction error than the AIC in all cases

(Jalil and Ma, 2008). Thus, in order to satisfy the principle of parsimony and avoid large

prediction error, lag length selection in this study is based on the SBC criterion.

Moreover, diagnostic tests are conducted to test the goodness of fit and structural

stability of the models and to enable the results be relevant for policy recommendation. In this

regard, serial correlation test, normality test, stability test and heteroskedasticity test are

conducted.

13
 ACCEPTED MANUSCRIPT

4. RESULTS AND DISCUSSIONS

This section progresses in five steps: first the stationarity status of the variables is checked

T
through unit root tests; second, the existence of long run relationship among the variables is

IP
established through bounds F-test; third and fourth steps involve estimations of the long run

R
and short run coefficients, together with the error correction term of the models; and finally,

SC
fifth, the goodness of fit and structural stability of the models were investigated through

diagnostic tests.

NU
MA
4.1 Results of Unit Root Tests

While the ARDL approach to cointegration is applicable whether the variables are integrated

of order zero or one, i.e. I(0) or I(1), it is still necessary to conduct stationarity tests in order to
D
TE

verify that none of the variables is integrated of order two, i.e. I(2). This is because as noted

by Ouattara (2004), the presence of I(2) variables renders the computed F-statistics provided
P

by Pesaran et al., (2001) invalid. For this reason, the Augmented Dickey Fuller (ADF) and
CE

Phillips-Perron (PP) tests for stationarity are employed. A plot of all the variables has
AC

indicated the presence of intercept and trend, thus in all the tests, a constant and linear trend

were included.

<Insert Table 1>

The unit root test results from Table 1 has revealed that the logs of Islamic bank

deposit, gross fixed capital formation, government consumption expenditure and inflation

have no unit roots at their levels. This means the variables are stationary, hence integrated of

order zero, I(0). On the other hand, the null hypothesis of having unit root could not be

rejected in the cases of industrial production index, Islamic bank financing and trade

14
 ACCEPTED MANUSCRIPT

openness, which means these variables are not stationary at levels. However, upon taking the

first difference, they became stationary, implying that they are I(1). This mixture of I(0) and

I(1) variables has justified the selection of the ARDL approach to cointegration, which

T
accommodate such scenario.

IP
4.2 Results of the Cointegration Test: Bounds F-test

R
As stated earlier, the bounds tests approach to co-integration is adopted and the results are

SC
presented in Table 2 below. It revealed that for both models, the calculated F-statistics are

NU
higher than the upper critical bounds at 1% level of significance (5.964 > 4.68 and 12.533 >

4.68). Consequently, there is strong statistical evidence of the existence of long run
MA
relationship between the variables in this study. This result suggests that Islamic banks

deposits and financing in Malaysia have long run equilibrium relationships with the real
D

economy.
TE

<Insert Table 2>

P

4.3 Estimates of Long Run Relationship

CE

Two models were estimated in this study, the first involves Islamic banks deposits as an

indicator of Islamic financial development and the second includes Islamic banks financing.
AC

The results for both models are presented in Table 3 below. Islamic banks deposit was found

to have a weak impact on the real economy in the long run, where at 10% level of

significance, a 1% change in Islamic bank deposits leads to only 0.07% increase in industrial

production. This is expected since not all the deposits in Islamic banks that are meant for

financing real economic activities. Although, Islamic banking principles ensure that saving is

matched with investment, some savers may not be interested to invest, but rather safe-keep

their money for liquidity purpose under the contract of wadiah (or safe-keeping). In such case,

the amount of saving channelled to investment is reduced, because unlike in conventional

15
 ACCEPTED MANUSCRIPT

banks, in Islamic banks channelling savings for investment purposes is not automatic, it is

rather at the discretion of the saver.

<Insert Table 3>

T
IP
On the other hand, Islamic bank financing has a strong and statistically significant

R
impact on the real economic activities in the long run with a 1% change in Islamic banks

SC
financing leads to 0.152% increase in industrial production. Financial intermediation by the

Islamic banks in Malaysia is shown to contribute positively to the economic growth of the

NU
country. The strong contribution of Islamic financing to the real sector is consistent with the

principle of Islamic finance, which provide for strong linkage between the financial and the
MA
real sectors of the economy. This finding is consistent with that of Abduh, Brahim and Omar

(2012) in Bahrain, Yusof and Bahlous (2013) in Malaysia, Indonesia and GCC countries, and
D

Abduh and Omar (2012) in Indonesia.

TE

In both models, trade openness appeared to have the strongest effect on industrial
P

production. This is indicative of the role that external trade plays in the economy of Malaysia.
CE

Conversely, inflation has consistent negative effect on industrial output in both models,

although the magnitude is very small. Investment in the form of gross fixed capital formation
AC

is also found to be important in influencing industrial production.

The results of the diagnostic tests are contained in the lower segment of Table 3. It

shows that there is no serial correlation among the residuals of both models, the residuals are

also normally distributed and homoscedastic. The models were also correctly specified, thus

the coefficients estimated in these models are stable, efficient and unbiased, hence can be

used for policy recommendation purposes.

4.4 Estimates of Short Run Relationship and the ECM

16
 ACCEPTED MANUSCRIPT

The results of the short run effect of Islamic bank development on real economic activities are

presented in Table 4. It shows that Islamic bank deposit is not having any effect on industrial

production in the short run, suggesting a lag between the time of receiving deposits and

T
channelling of funds for real economic activities. This time lag is taken by the Islamic banks

IP
to perform other financial intermediation-related activities, which help to mitigate and

R
diversify risks and ultimately, ensure that funds are allocated to deserving investment

SC
opportunities. The results also show that in the short run, Islamic banks deposit does not

Granger cause industrial production. The error correction term for the Islamic bank deposit

NU
model is -0.283 and is statistically significant, implying that 28% of deviations from the
MA
equilibrium path are adjusted in one quarter, suggesting the existence of long run Granger

causality from Islamic bank deposit to industrial production.

D

<Insert Table 4>

TE

On the other hand, Islamic bank financing has a positive and significant effect on
P

industrial production in the short run, suggesting the existence of short run Granger causality
CE

from Islamic banks financing to real economic activity. This is contrary to the findings of

Abduh, Brahim and Omar (2012) in Bahrain, but reinforced the findings of Yusof and
AC

Bahlous (2013) in Malaysia, Indonesia and GCC countries and that of Abduh and Omar

(2012) in Indonesia. The error correction is -0.417 and is statistically significant at 1%, this

means 41.7% of all deviations from the long run equilibrium path are adjusted in one quarter.

It also signifies the existence of long run Granger causality from Islamic bank financing in

Malaysia to the real economy. This result is consistent with the findings of Abdul Manap,

Abduh and Omar (2012). The R-squared values of 0.641 and 0.654 for the Islamic bank

deposit and financing models respectively have shown that the models are good fit.

Similarly, as suggested by Pesaran and Pesaran (1997), the cumulative sum of

recursive residuals (CUSUM) and the cumulative sum squares of recursive residuals

17
 ACCEPTED MANUSCRIPT

(CUSUMSQ) tests are carried out to test for structural stability. In addition to various specific

serial correlations, model stability, normality and heteroscedasticity tests conducted the

CUSUM and CUSUMSQ for overall stability of the estimated coefficients in all the

T
regressions is carried out. The results for both models show that all the plots of the CUSUM

IP
and CUSUMSQ statistics are within the critical bounds of the 5% significance level 1.

R
Therefore, the null hypothesis, which states that all the coefficients in the regressions are

SC
stable, cannot be rejected, thus, the results are suitable for policy recommendation.

5. CONCLUSIONS
NU
MA
The high proportion of Islamic banking assets and their increasing presence and significance

in the Malaysian financial sector has called for the evaluation of the role of the sector in
D

enhancing the real economy of the country. This study employed the ARDL framework and
TE

bounds testing approach to co-integration to empirically investigate the influence of Islamic

banks deposits and financing on industrial output in Malaysia. The results revealed that while
P

the Islamic deposits are not having significant effect on the real economy in the short run, in
CE

the long run, the Islamic deposits are having a positive effect on the real economy. In this
AC

regards, the study suggests that the Islamic banks should promote shari’ah-compliant

investment deposits in order to attract long-term savers and avoid the time lag in pooling

funds for investment purposes. This finding concurs well with the role of the Islamic banks as

investee and as such, the major source of funds should be long term in nature to enable the

banks to invest in productive investment projects.

The results also show that Islamic banks financing activities are making significant

contribution to the real economic activities both in the short and long runs, with the long run

contribution being stronger. This finding suggests that the Islamic banks in Malaysia are

1
The plots of the CUSUM and CUSUMSQ statistics are not included here in order to conserve space. However,
they are readily available upon request from the author.

18
 ACCEPTED MANUSCRIPT

effectively carrying-out the financial intermediation role of pooling and channeling funds to

productive investment activities.

The above contributions of Islamic finance to the real economic activities are made

T
possible by the principle of equity participation embedded in it. The investor-investee relation

IP
between the depositors and the Islamic banks based on the concept of risk sharing should lead

R
to better monitoring of investments, hence higher productivity, more stable financial sector

SC
due to absence of interest rate risk, among others and hence, more sustainable economic

growth. These features are evident in the various Islamic finance instruments, such as

NU
mudharaba (profit sharing), musharakah (joint venture) and ijarah (leasing) among other
MA
instruments. For instance, mudharaba and musharakah are both long term financing

arrangements, devoid of interest, in which profit/loss is shared among the saver and

entrepreneur based on appropriate ratio. They also involved the participation of both parties
D
TE

right from deciding the type of project to be undertaken to its implementation, hence proper

monitoring is ensured. The implications of these findings to the practice of Islamic banking in
P

Malaysia are that, given the lack of and weak contributions of Islamic banks’ deposits to the
CE

real economy in the short and long runs respectively, the banks need to design more longer
AC

term instruments of deposits, such as the counterpart of the negotiable certificate of deposit in

the conventional banking.

For the financing aspect, although it contributes to industrial production in both short

and long runs, the long run effect is shown to be stronger than the short run, thus the banks

need to balance their funds allocation for take up and business expansion purposes. Although,

deposit and financing are supposed to be closely linked in Islamic banking, these findings

suggested otherwise, therefore, the banks should take measures to bridge the gap.

In view of the important contribution of Islamic finance to the Malaysian economy,

continuous efforts need to be undertaken to further expand the industry. On top of the list is

19
 ACCEPTED MANUSCRIPT

the importance of a conducive legal and regulatory framework that needs to be further refined

to support the transformation of the Malaysian Islamic finance industry to become a global

industry, thus further strengthening the country’s position as an exemplary role and a leader in

T
promoting Islamic finance the global level.

R IP
REFERENCES

SC
Abduh, M. and Omar, M. A. (2012). Islamic banking and economic growth: the Indonesian
experience. International Journal of Islamic and Middle Eastern Finance and
Management, 5(1): 35-47.

NU
Abduh, M., Brahim, S. and Omar, M. A. (2012). A study on finance-growth nexus in dual
financial system countries: evidence from Bahrain. World Applied Sciences Journal,
20(8): 1166-1174.
MA
Abdul Manap, T. A., Abduh, M. and Omar, M. A. (2012). Islamic banking-growth nexus:
evidence from Toda-Yamamoto and bootstrap Granger causality test. Journal of
Islamic Finance, 1(1): 59-66.
Adebola, S. S., Yusoff, U. S., and Dahalan, J. (2011). An ARDL approach to the determinants
D

of non-performing loans in Islamic banking system in Malaysia. Arabian Journal of

Business and Management Review,1(2).
TE

Anonym. (2010). Islamic banking’s role in economic development: Future outlook.

Aziakpono, M. (2003). Financial intermediation and economic growth in economic and
monetary union: the case of SACU and CMA. Paper presented at the ESSA
P

Conference, Lord Charles Hotel, Somerset West, 19 September 2003.

Bagehot, W. (1873/1999). Lombard Street: A Description of the Money Market, New York:
CE

John Wiley & Sons, Inc.

Barajas, A., Chami, R. and Yousefi, S. R. (2010). The finance-growth nexus re-examine: are
there cross-region differences? IMF working paper.
AC

Engle, R. F. and Granger, C. (1987). Co-integration and error correction: representation,

estimation, and testing. Econometrica, 55: 251-276.
Ernst and Young (2013). World Islamic Banking Competitiveness Report 2013-14: the
Transition Begins, Ernst and Young 2013.
Furqani, H. and Mulyany, R. (2009). Islamic banking and economic growth: empirical
evidence from Malaysia. Journal of Economic Cooperation and Development, 30(2):
59-74.
Goaied, M. and Sassi, S. (2010). Financial development and economic growth in the MENA
region: what about Islamic banking development? Institut des Hautes Etudes.
Hassan, M. K., Sanchez, B. & Yu, J. (2011). Financial development and economic growth in
the Organisation of Islamic Conference Countries. JKAU: Islamic Economics, 24(1):
145-172.
Jalil, A. and Ma, Y. (2008). Financial development and economic growth: time series
evidence from Pakistan and China. Journal of Economic Cooperation, 29(2): 29-68.
Johansen, S. and Juselius, K. (1990). Maximum likelihood estimation and inference on
cointegration - with applications to the demand for money. Oxford Bulletin of
Economics and Statistics, 52(2): 169-210.

20
 ACCEPTED MANUSCRIPT

Khoutem, B. J. and Nedra, B. A. (2012). Islamic participative financial intermediation and

economic growth. Journal of Islamic Economics, Banking and Finance, 8(3): 44-59.
King, R.G. and Levine, R. (1993). Finance and growth: Schumpeter might be right. The
Quarterly Journal of Economics, 108(3): 717-737.
Levine, R. and Zervos, S. (1998). Stock markets, banks, and economic growth. American
Economic Review, 88: 537-558.

T
Lucas, R. E. (1988). On the mechanics of economic development. Journal of Monetary
Economics, 22: 3-42.

IP
McKinnon, R. (1973). Money and Capital in Economic Development.Washington: The
Brookings Institute.

R
Nagaoka, S. (2011). Islamic finance for sustainable development: its historical background
and potentialities in the modern world. 8th International Conference on Islamic

SC
Economics and Finance, Qatar National Convention Centre, Doha, 21 December,
2011.
Narayan, P. K. (2004). Fiji’s tourism demand: the ARDL approach to co-integration. Tourism

NU
Economics, 10(2): 193-206.
Odhiambo, N. M. (2010). Finance-investment-growth nexus in South Africa: an ARDL-
bounds testing procedure. Econs Change Restruct, 43: 205-219.
MA
Ouattara, B. (2004). The impact of project aid and programme aid on domestic savings: A
case study of Côte d’Ivoire, paper presented at the conference “Growth, poverty
reduction and human development in Africa”, Centre for the Study of African
Economies, April.
Patrick, H. T. (1966). Financial development and economic growth in underdeveloped
D

countries. Economic Development and Cultural Change, XIV(2): 174 – 189.

TE

Pesaran, M. H. and Pesaran, B. (1997). Working with Microfit 4.0: Interactive Econometric
Analysis. Oxford University Press, Oxford.
Pesaran, M. H. Shin, Y. and Smith, R.J. (2001). Bounds testing approaches to the analysis of
P

level relationships. Journal of Applied Econometrics, 16: 289-326.

Rajan, R.G. and Zingales, L. (1998). Financial dependence and growth. The American
CE

Economic Review, 88(3): 559-586.

Robinson, J. (1952). The Generalisation of the General Theory, in the Rate of Interest, and
Other Essays (2nd ed.). London: Macmillan.
AC

Schumpeter, J. A. ([1911] 1934). Theory of Economic Development: An inquiry into Profits,

Capital, Credit, Interest and the Business Cycle, New Brunswick, USA: Transaction
books.
SESRIC (2012). Islamic Finance in OIC Member Countries. OIC Outlook Series, May 2012,
Statistical, Economic and Social Research and Training Centre for OIC Countries
(SESRIC).
Shaw, E. S. (1973). Financial Deepening in Economic Development, New York: Oxford
University Press.
Yusof, R. M. and Bahlous, M. (2013). Islamic banking and economic growth in GCC & East
Asia countries, Journal of Islamic Accounting and Business Research, 4(2): 151 – 172.

21
 ACCEPTED MANUSCRIPT

Table 1. Results of Unit Root Tests

ADF Test PP Test

T
Stationarity
Variables Level First Difference Level First Difference

IP
Status
LIPI -1.671 -6.198*** -1.796 -6.105*** I(1)

R
LIBD -3.32* -3.566*** -3.279* -6.198*** I(0)

SC
LIBF -3.078 -3.107** -1.595 -4.665*** I(1)
LGFCF -2.345 -4.494*** -3.99** -10.507*** I(0)

NU
LOPN -2.403 -4.172*** -3.027 -12.854*** I(1)
LGCE -2.140 -3.859*** -14.79*** -39.455*** I(0)
INF -4.72*** -6.955*** -3.243* -5.723*** I(0)
MA
Notes: Lag lengths are selected based on Schwarz Bayesian Criterion; the test statistics are compared with critical
values from Mckinnon (1996); ***, ** and * denote significance at 1%, 5% and 10%, respectively.
D
TE

Table 2. Results of Bounds F-Test for Long Run Relationship

P

IBD Model, F =5.964

Computed F-Statistic
CE

IBF Model, F = 12.533

Critical Bounds (k = 5)
Levels of Significance I(0) I(1)
AC

1% 3.41 4.68
5% 2.62 3.79
10% 2.26 3.35
Note: The critical values are based on Pesaran et. al. (2001), case table C(iii) case III:
unrestricted intercept and no trend.

22
 ACCEPTED MANUSCRIPT

Table 3. ARDL Estimate of Long Run Relationship

Model With IBD Model With IBF
Regressors Coefficients t-statistics Coefficients t-statistics
LIBD 0.067* 1.741 - -

T
LIBF - - 0.152*** 5.382

IP
LGFCF 0.266** -2.163 0.215*** -3.124
LOPN 1.224*** 7.492 0.870*** 8.047

R
LGCE 0.021 0.373 -0.142** -2.365

SC
INF -0.019** -2.188 -0.012** -2.084
Intercept -8.938*** -5.437 -4.519*** -3.517

NU
Diagnostic Test Statistics
Serial correlationX2(4) 1.740 [0.783] 1.369 [0.860]
2
Functional formX (1) 1.964 [0.161] 2.051 [0.152]
MA
2
NormalityX (2) 2.930 [0.231] 0.880 [0.644]
2
HeteroscedasticityX (1) 1.461[0.227] 0.067 [0.795]
D

ARDL (2,0,0,1,0,0) and ARDL (1,0,0,0,1,0) for IBD and IBF models respectively, selected based on SBC
Note: ***, ** and * represent statistical significance at 1%, 5% and 10% respectively, p-values in [ ].
P TE
CE
AC

23
 ACCEPTED MANUSCRIPT

Table 4. Results of ARDL Short Run Estimate and ECM

Model With IBD Model With IBF
Regressors Coefficients t-statistics Coefficients t-statistics
∆LIPI(-1) -0.217* -1.905 - -
∆LIBD

T
0.019 1.454 - -
∆LIBF

IP
- - 0.063*** 4.097
∆LGFCF 0.075* -1.963 0.090*** -2.953

R
∆LOPN 0.499*** 7.537 0.363*** 6.255

SC
∆LGCE 0.006 0.377 -0.016 -1.047
∆INF -0.005** -2.446 -0.005** -2.470

NU
Intercept -2.533*** -3.396 -1.885*** -3.318
ECM(-1) -0.283*** -3.891 -0.417*** -6.190
Diagnostic Test Statistics
MA
R-squared 0.641 0.654
Adjusted R-squared 0.585 0.607
F (7, 52), F(6, 53) 13.017 16.370
D

SE-regression 0.021 0.020

Residual sum of squared 0.022 0.022
TE

DW-statistic 2.042 1.928

ARDL (2,0,0,1,0,0) and ARDL (1,0,0,0,1,0) for IBD and IBF models respectively, selected based on SBC
P

Note: ***, ** and * represent statistical significance at 1%, 5% and 10% respectively.
CE
AC

24