Financial development and economic growth in Australia: An empirical analysis

Transcription

1 Nanyang Technological University From the SelectedWorks of James B Ang 2004 Financial development and economic growth in Australia: An empirical analysis James B Ang, Nanyang Technological University Available at:

2 Empirical Economics (2004) 29: DOI /s Financial development and economic growth in Australia: An empirical analysis Shandre M. Thangavelu, Ang Beng Jiunn, James Department of Economics, National University of Singapore, Faculty of Arts & Social Sciences, AS2 Level 6, 1 Arts Link, Singapore , Republic of Singapore ( ecssmt@nus.edu.sg) First version received: October 2001/Final version received: October 2002 Abstract. The paper empirically examines the dynamic relationship between financial development and economic growth in Australia in terms of bankbased and market-based financial structure. A time-series approach using the VAR Model is used to provide evidence for the dynamic relationship. The paper provides empirical evidence on the causal impact of the financial market on the economic growth of the Australian economy. The results suggest that financial intermediaries and financial markets have different impacts on economic growth given their diverse roles in the domestic economy. In particular there is evidence of causality from economic growth to the development of the financial intermediaries. On the other hand, development in the financial markets causes economic growth but there is no evidence of any causality from economic growth to financial markets. The sensitivity test using different interest rates does not change the results. Key words: Financial markets, financial intermediaries, economic growth, vector autoregression JEL classification: O16, G18, G28 1. Introduction Two schools of thought highlight the importance of financial development for economic growth and their causal relationship but hold completely contrasting perspectives. Schumpeter (1911) argues that well-functioning banks are able to identify innovative entrepreneurs that allow funds being channeled We would like to thank Tilak Abeysinghe and Rajagurn Gunasekaran for their helpful comments on the first draft. Also, we would like to thank the Editor, Prof. Baldev Raj, and two anonymous referees for their helpful comments.

3 248 S. M. Thangavelu and A. B. Jiunn, James to the most promising investment projects. In contrast, Robinson (1952) argues that economic growth creates demand for more financial services and thereby leads to financial development. However, only in recent years that attempts have been made to study the relationship between these two variables in both a theoretical and empirical framework (Beck et al. 2000; Levine and Zervos 1996, 1998; Levine 1997; Levine et al. 2000; Rajan and Zingales 1998; Demetriades and Hussein 1996; Luintel and Khan 1999). 1 The current study re-examines the above causal relationship between financial development (henceforth FD) and economic growth (henceforth EG) in terms of bank-based and market-based system using a time-series framework for the Australian economy. Although previous studies have examined the causal relationship between financial development and economic growth in a timeseries framework (Demtriades and Hussein 1996; Luintel and Khan 1999), the departing feature of the current study is to re-examine the relationship in terms of bank-based and market-based financial structure. As opposed to the view that suggests the relative importance of market-based and bankbased financial structure for economic growth (Levine 2001; Beck and Levine 2002), the paper suggests that the bank-based and market-based system has different impact on the economic growth of the domestic economy. Given their diverse roles, it is possible for the financial intermediaries and financial markets to have mutually reinforcing role in the overall development of the financial system and hence economic growth. The implications of bank-based and market-based financial systems for resource allocation and economic growth are highlighted by Allen and Gale (2000) and Levine (2001). 2 The importance of the financial markets is to allow borrowers borrow funds directly from the lenders by selling financial instruments such as securities to the borrowers. Although financial markets are not likely to attract small savers, a well-developed stock market allows risk diversification and better resource allocation. As opposed to financial markets, the bank-based models suggest the indirect financing role of the financial intermediation that allows funds to be mobilized from large number of small savers to large number of borrowers. Therefore the role of financial intermediaries increases the overall propensity to save, provide liquidity and mobilize fund to most productive investments in the domestic economy. Recently, Levine (2001), Demirguc-Kunt and Levine (2001), Beck and Levine (2002) and Demirgue-Kunt and Maksimovic (2002) studied the relative importance of bank-based and market-based financial system in promoting growth using regression analysis. In a panel cross-country study, Levine (2001) and Demirguc-Kunt and Levine (2001) found positive evidence of overall development of the financial system on output growth; however, they did not find any evidence of bank-based or market-based financial systems in promoting real per capita GDP growth. Based on these results, they conclude that the cross-country comparisons do not suggest that distinguishing between bank-based and market-based is analytically useful for understanding the 1 See Levine (1997) for an excellent survey of the literature. 2 Levine (2001) also suggests financial arrangements and legal system in promoting the financial sector as important factors for financial market development and economic growth. The current study does not address these issues as it only focuses on the impact of bank-based and marketbased financial structure on economic growth.

4 Financial development and economic growth in Australia 249 process of economic growth. Using a similar panel study, Beck and Levine (2002) examine the importance of the financial structures on industry growth, new establishment formation, and efficient capital allocation. Their results give no evidence of the market-based or bank-based financial structure having any impact on industrial growth. Furtherrmore, Demirguc-Kunt and Maksimovic (2002) using firm level data found no evidence for the financial structure as a robust predictor of economic growth. Although these regression studies make significant contribution in studying the above relationship, the results are seriously flawed in the methodology used since there could be significant simultaneity bias between the two variables. 3 Granger (1980) points out that it is possible for two variables to be highly correlated but not necessary causally linked. Therefore these findings cannot be interpreted in a causal sense. The cross-section nature of the technique cannot allow different countries to exhibit different patterns of causality. This is because it is likely that in some countries financial sector is leading while in others it lags behind the real sector. This means that the causality result obtained is only valid on the average, but not all. The above growth-regression methodology has been severely criticized in recent years. The testing results are fragile and difficult to interpret (Lee et al. 1996). The failure to effectively control for cross-country heterogeneity and for the endogeneity of the explanatory variables creates large biases. Also, cross-country regressions do not resolve the issues of causality. Several studies have also found that the relationship between financial intermediation and economic growth depend largely on the sample of countries chosen in the study (Fernandez and Galetovic 1994; De Gregorio and Guidotti 1995). 4 This therefore highlights the danger of grouping all countries as a single entity and the importance of being country specific in this area of study. The objective of our study is to construct an empirical model to examine the causal relationship between financial development and economic growth in the time-series framework for the Australia economy. The paper employs time series methodology of vector autoregressive (VAR) model and Granger causality test to estimate the above relationship. Sufficient number of variables has also been worked out to measure the level of financial development. On top of that, the time span of this study cover from 1960 to 1999, and with the use of quarterly data, we sufficiently solve the asymptotic problems. To study the different roles of bank-based system and market-based system in the Australian economy, we used different financial indicators related to financial intermediaries and financial markets. In contrast to the results that neither the bank-based or market-based financial systems are particularly effective in promoting growth, the results of 3 More recent studies attempted to address the causality between financial development and economic growth using a panel cross-country analysis (Beck et al. (2000), Levine et al. (2000)). These panel studies, after controlling for simultaneity effects through instrumental variables, found positive relation between financial intermediation and long-run economic growth. 4 For instance, Fernandez and Galetovic (1994) split King and Levine s sample into OECD and non-oecd countries and show that the correlations fall and become insignificant for OECD countries. By adding more countries to King and Levine s sample, De Gregorio and Guidotti (1995) divide the sample into three groups based on per capita income at the beginning sample period. It was shown that the correlations rise and become more significant as initial incomes fall.

5 250 S. M. Thangavelu and A. B. Jiunn, James our time-series study on the Australian economy suggest that financial intermediaries (bank-based system) and financial markets (market-based system) tend to have different role in promoting growth in the economy. The results provide strong support for the Robinson s hypothesis that economic growth promotes financial development. In particular, economic growth tends to cause financial sector development when financial variables related to financial intermediaries are used. However, we do not find any causal evidence from financial intermediaries to real per capita GDP growth. On the other hand, we found strong evidence that market-based system promotes economic growth in the Australian economy. Financial indicators related to financial markets, i.e. stock market, support the hypothesis that financial sector development causes output growth. The results clearly highlight the different, but mutually reinforcing, role of direct financing of the financial markets and the role of indirect financing by the financial intermediaries in promoting economic activity in the domestic market. We organize the remainder of this paper as follow. Section 2 discusses the importance of banking sector and financial markets in Australia. Section 3 explains the variables used and time series methodology employed in this paper. In Sect. 4, we provide the results of our model. Finally, we conclude in Sect The Australian banking sector and financial market Comprising only 0.3% of the world s population, Australia is the fourteenth largest economy in the world and the fourth largest economy in the Asian region after China, Japan and Indonesia. It has a GDP twice that of Hong Kong, and is larger than the combined GDPs of Singapore, Malaysia, Thailand, and the Philippines. Australia is an advanced, industrialized economy with a sophisticated service and manufacturing base. Australia has the thirteenth largest stock exchange in the world and the second largest in the Asian region after Japan. The Sydney Futures exchange is the largest financial futures exchange in the Asia Pacific region. Australia s sound financial environment is built on a strong, reliable prudential regulatory system and strong linkages to global markets. This combination provides the necessary foundation for the operation of an efficient economy The Banking sector Taking total deposit money bank assets divided by the nominal GDP, Fig. 1 clearly shows that there is an increasing trend of this ratio over the years. In 1960, 44% was recorded but after 40 years it has increased to 101%. Such increasing trend in the ratio observed indicates that the growth in bank assets surpasses growth in nominal GDP. Figure 2, on the other hand, shows the trend of GDP contributed by different sectors over the fourteen years. We observe that the finance industry has an obvious upward trend followed by the wholesale and retail trade, restaurants and hotels industry. Other industries are either experiencing stagnant or declining trend. Between 1980 and 1994, we observe that the share of GDP contributed by the finance industry has increased from 17% to 26%

6 Financial development and economic growth in Australia 251 % Year Fig. 1. Deposit money bank assets (as a percentage of GDP) (Source: International Financial Statistics) % 40 Agriculture Manufacturing Wholesale & Retail Trade, Restaurants Finance Industry Community, Social & Personal services Others Year Fig. 2. Share of GDP contributed by different sector (Source: OECD Compendium 1999 CD-ROM) over the fourteen years. 5 This indicates that the finance industry has gained more weights in terms of its GDP contribution. Hence, such observation tends to suggest that finance sector has a great impact on the economic growth. Manufacturing sector and others tend to be less important The financial market The functions of financial market are to provide funding for companies listed on the stock exchange through the issuance of securities, provide investment 5 Six categories of industry are studied for comparison purpose, namely: 1) agriculture, hunting, forestry, and fishing; 2) manufacturing; 3) wholesale and retail trade, restaurants and hotels; 4) finance industry; 5) community, social and personal services; and 6) others. The finance industry also includes real estate and business services. Others refer to mining and quarrying, electricity, gas, and water, construction, transport, storage, communication and those activities that are not adequately defined.

7 252 S. M. Thangavelu and A. B. Jiunn, James opportunities for both institutional and retail investors and offer liquidity through the secondary market operation. The role of ASX (Australian Stock Exchange) is crucial in the Australian economy because it facilitates the flow of funds from investors to productive enterprises, thereby generating wealth, employment and economic growth. ASX provides the markets for trading equities, debt securities and equity and index derivatives. More than one fifth of the total trading on the equities market is from foreign investors. As such, ASX has important supervisory responsibilities over market participants and over listed companies. To examine the share turnover, Table 1 provides information on turnover volumes from 1994 to It is clear that the number of shares traded, turnover volume, and number or trades have all increased over the years. 3. Data and methodology 3.1. Variables and data sources The selection of key variables to indicate the level of financial services produced in the economy and how to measure the extent of financial intermediation are the major problem in most empirical studies. Construction of financial development indicators is an extremely difficult task due to the diversity of services involved. Furthermore, there is a diverse array of agents and institutions involved in the financial intermediation activities. In our study, we use real GDP per capita (Y) as the measurement for economic growth. In addition, we used bank claims on private sectors to nominal GDP (C), domestic bank deposit liabilities to nominal GDP (D) and equities turnover to nominal GDP (E) to measure the level of financial development in Australia. The measure of bank claims on private sectors to nominal GDP (C) provides direct information on financial intermediation. The amount of funding supplied to entrepreneurs largely account for the quality and quantity of investments, and in turn for economic growth. Hence, it is expected that this variable may have causal influence on the real GDP per capita. It is common to see in the literature that broad money M2 to nominal GDP has been used as a financial indicator. However, we exclude currency in circulation (M1) from the broad money stock (M2) as this better reflects how extensive bank deposits are being used. Thus, we use ratio of domestic bank deposit liabilities to nominal GDP (D) as one of the proxies to measure for financial development. Finally, the variable on the equities turnover to nominal GDP (E) is used to reflect the level of liquidity in the stock markets. The functioning of equity markets affects liquidity, risk diversification, acquisition of information about firms, corporate control, and savings Table 1. Share turnover in Australian stock exchange Year Number of shares traded (million) Turnover volume (AUS $ million) Number of Trades ( 000) Source: Australian Stock Exchange, 1999.

8 Financial development and economic growth in Australia 253 mobilization. Hence, the functioning of stock markets can alter the rate of economic growth by improving the quality of these services (Levine and Zervos 1996). Since Australian stock market is very large in terms of its size, it is important to take this into account in our analysis. Nevertheless, it is important to handle the causality testing with care because both financial development and economic growth can be driven by a common variable, such as propensity of households to save (Rajan and Zingales 1998; Luintel and Khan 1999). Two types of interest rates, the Reserve Bank discount rate (R) and Money Market rate (M), are used to account for this common variable. R is the rate at which the central bank lends to banks, shown on an end-of-period basis. M is the rate on short term lending between financial institutions, also known as Interbank Offer Rate. We used quarterly data for all the variables involved from 1960 to Data for equities turnover to nominal GDP (E) is obtained from the Reserve Bank of Australia. The rest of the data is obtained from International Financial Statistics (IFS) data set ranging from 1960 to All the data is transferred into natural logarithms, except the interest rates, for the usual statistical reasons Time series methodology Vector autoregressive model The vector autoregression (VAR) model is commonly used for forecasting a system of interrelated time series and for analyzing the dynamic impact of random disturbances on the system of variables. In matrix notation, the VAR model for k variables can be written as y t ¼ A 0 þ A 1 y t 1 þþa p y t p þ e t ð1þ where y t is a n vector of endogenous variables, A 1,..., A p are matrices of coefficients to be estimated, and e t is a vector of innovations that may be contemporaneously correlated with each other but are uncorrelated with their own lagged values. The conventional Granger causality test involves specifying a bivariate pth order VAR as follows: x 1t ¼ a 10 þ P 11 ðlþx 1t 1 þ P 12 ðlþx 2t 1 þ e 1t ð2þ x 2t ¼ a 20 þ P 21 ðlþx 1t 1 þ P 22 ðlþx 2t 1 þ e 2t ð3þ where a 10 and a 20 are constants and P ij (L) are polynomials of order k 1in the lag operator L. The VAR model can be used to test the hypothesis that some variables do not Granger cause some other variables. The null hypothesis is that x 1t does not Granger cause x 2t amounts to a zero polynomial P 21 (L) and can be tested by a standard F-test. It has been argued that causality tests are sensitive to the lag length in the VAR. We acknowledge this by using the Akaike Information Criterion (AIC) in determining the optimal number of lags Vector error-correction model One problem with the standard Granger causality tests, which Granger (1988) noted is that it is possible to find no causal relationship between two variables

9 254 S. M. Thangavelu and A. B. Jiunn, James that are cointegrated. Hence, if two or more variables are found to be cointegrated, an error correction term (ECT) has to be incorporated into the equation in estimating causality (Engle and Granger 1987). This is because by doing so, the ECT reintroduces the information lost in the differencing process, thereby allowing for long run equilibrium as well as short run dynamics. Therefore if two variables are not cointegrated than the conventional Granger causality tests are valid in a levels VAR framework. However, if the two variables are cointegrated than the VECM framework is introduced to study the dynamic relationship between the two variables. Assuming a bivariate model for simplicity, the error correction model takes of the following form: DX ¼ a 10 þ Xp a 1i DX t i þ Xp b 1i DY t i þ c 1 ECT t i þ e 1t i¼1 i¼1 DY ¼ a 20 þ Xp a 2i DY t i þ Xp b 2i DX t i þ c 2 ECT t i þ e 2t i¼1 i¼1 ð4þ ð5þ where D is the difference operator, X and Y are two endogenous variables, e 1t and e 2t are zero-mean, serially uncorrelated disturbance terms, and ECT t ) i is the error correction term derived from the long-run cointegrating relationship. From the first equation, we say that DY Granger causes DX if all b 1i is not zero. Likewise Eq. (5), DX Granger causes DY if all b 2i is not zero. A Wald statistic that follows a standard v 2 with degrees of freedom equal to the number of restrictions can be used to test for non-causality. This is known as the augmented Granger causality test. However, the link between cointegration and causation becomes explicit in the vector error-correction model (ECM) where there are two sources of causation, through the ECT or through the lagged difference terms. Thus, ECT measures the long run equilibrium relationship while the coefficients on lagged difference terms measure the short run causal relation. Normally, we perform the augmented Granger causality tests on the first difference of the variables since they are all I(1). However, in our case, we modify the Wald test by setting both b ij (L)=0 (for example the coefficients of the third variable in Eq. (5)) and c i = 0 (where c i = error correction/speed of adjustment coefficients). This is because in a cointegrated system, variable j does not Granger cause variable i if all lagged values of variable j do not enter the i equation and if variable i does not respond to the deviation from long-run equilibrium (see Enders, p. 371). Also, it must be noted that this test of joint significance of lagged dynamic and error correction terms is also a test for strong exogeneity (Charemza and Deadman 1992). 4. Estimation results 4.1. The model Three variables are used to construct the model, namely economic growth, financial development and interest rates. For our estimation purpose, we construct six models whereby each model contains only three variables (see Table 2). We use Y as the proxy for economic growth while C, D, and E are used as proxies for financial development. In order to separately study

10 Financial development and economic growth in Australia 255 Table 2. The six models Model Variable 1 Y, C, R 2 Y, C, M 3 Y, D, R 4 Y, D, M 5 Y, E, R 6 Y, E, M the effects of financial intermediaries on output growth, we used bank deposits liability to nominal GDP ratio (D) and bank claims on private sector to nominal GDP ratio (C). On the other hand, we used equities turnover to nominal GDP ratio (E) to study the effects of direct financing that is the growth of financial markets such as stock markets on output growth. We apply two different interest rates, namely reserve bank discount rate (R) and money market rate (M) to check whether the results are stable Unit root test The first step of our testing procedure is to study the unit root properties of the variables. Therefore we tested the stationarity of variables by using the Augmented Dickey Fuller (ADF) test and Phillips-Perron (PP) Test. A necessary but not sufficient condition for cointegration is that each of the variables should be integrated of the same order, and the order must be greater than or equal to one. The test results presented in Table 3 unanimously suggest that all variables are integrated at order 1, I(1), i.e., they become stationary after first differencing Cointegration test The next step, therefore, is to proceed with cointegration test. In conducting cointegration analysis, we employ the Johansen (1988) methodology because of its superiority over other alternatives. Since Johansen cointegration test is sensitive to the lag length lag used, prior to performing the cointegration test, we use the Akaike Information Criterion (AIC) to determine the appropriate number of lag. The results of maximal eigenvalue test and trace test are presented in Table 4. We find one cointegrating vector in Model 2 based on the k max test. The null hypothesis r=0 can be rejected at 5% significance level. This indicates the testing results for cointegration are sensitive to the types of interest 6 We also conducted the KPSS stationarity test (Kwiatkowski et al. 1992) and unit roots test with structural breaks as suggested by Perron (1997). Both test results provides strong evidence that all the variables have unit roots at the levels and support the above ADF and PP test that they are all of order 1, I(1). We would like to thank the anonymous referee for highlighting the above tests.

11 256 S. M. Thangavelu and A. B. Jiunn, James Table 3. Unit root test xvariables ADF test PP test 5% Critical values* I(0) I(1) I(0) I(1) Y )0.543 )6.243 )0.663 ) )2.880 C ) ) )2.880 D ) ) )2.880 E )0.317 )5.288 )0.404 ) )2.887 R )2.123 )4.298 )1.724 )7.496 )2.890 M )2.502 )4.382 )2.026 )8.880 )2.887 * Critical values follow McKinnon (1991). Table 4. Johansen cointegration test Model Eigenvalue Maximal eigenvalue test Trace test H 0 H 1 Stat. 5% H 0 H 1 Stat. 5%* 1 k 1 = R = 0 R = r=0 r (p=5)** k 2 = R = 1 R = r 1 r k 3 = R = 2 R = r 2 r= k 1 = R = 0 R = r=0 r (p=3)*** k 2 = R = 1 R = r 1 r k 3 = R = 2 R = r 2 r= k 1 = R = 0 R = r=0 r (p=5) k 2 = R = 1 R = r 1 r k 3 = R = 2 R = r 2 r= k 1 = R = 0 R = r=0 r (p=5) k 2 = R = 1 R = r 1 r k 3 = R = 2 R = r 2 r= k 1 = R = 0 R = r=0 r (p=4) k 2 = R = 1 R = r 1 r k 3 = R = 2 R = r 2 r= k 1 = R = 0 R = r=0 r (p=4) k 2 = R = 1 R = r 1 r k 3 = R = 2 R = r 2 r= * Critical values follow Johansen and Juselius (1990). ** p = optimal lag length. *** Model 2 has one cointegrating relationship. rates used. It is argued that the k max test is preferred to the trace test to pin down the number of cointegrating vectors (Enders, p. 393). Johansen and Juselius (1990, p. 9) pointed out that one would, however, expect the power of this procedure (the trace test) to be low, since it does not use the information that the last three eigenvalues have been found not to differ significantly from zero. Thus, one would expect the maximum eigenvalue test to produce more clear cut results. Hence, we conclude that there is one cointegrated relationship and formulate a VEC model to estimate the causal relationship in Model 2. The evidence of such cointegration appears in Model 2 suggests that the three variables are bound together by a long run equilibrium relationship. This is consistent with economic theory as it indicates that economic growth and financial development have a long run equilibrium relationship. Hence,

12 Financial development and economic growth in Australia 257 this suggests that causality between the variables must exist at least in one direction Causality testing Our testing objective is to examine the causal relationships between economic growth and financial development. The optimal lag length, F-statistics, and P-values for each model are reported in Table 5. The summary of all Granger causality test results is reported in Table 6. 7 The test results for both Model 1 and 2 show that Y Granger-causes C at significance level of 10% despite different interest rates are used. Comparing Model 3 and 4, we find bi-directional causality between Y and D in Model 3, but only unidirectional causality is found from Y to D in Model 4 when M is used as the interest rate. In Model 5 and 6, we find causality running from E to Y irrespective of the types of interest rates used. To provide more insights to the above results, we divided the results into two parts: (i) if financial development indicators related to the banking sector are used, we have causality running from EG to FD (Model 1, 2, 3 and 4); and (ii) if financial development indicator related to the financial sector is used, we obtain causality from FD to EG (Model 5 and 6) Bank-based system: Banking sector and economic growth The evidence that Y (real GDP per capita) Granger-causes C (logarithm of claims on private sector divided by nominal GDP) reported in Model 1 and 2 suggests that with higher economic growth may encourage more saving and lending in the economy. This may be due to better business prospects projected by entrepreneurs when high economic growth is achieved. Such evidence supports the view that high economic growth leads to rapid financial development. On the other hand, the evidence obtained from testing results of Model 3 and 4 that Y (real GDP per capita) Granger-causes D (logarithm of bank deposits liability divided by nominal GDP) suggest that higher economic growth leads to more demand for deposits. This is in line with Robinson s (1952) argument that where enterprise leads finance follows, strong economic growth results in higher demand for various types of financial services. Hence, more financial institutions, products and services emerge in the markets to meet the increased demand. Therefore, bank may require larger deposit base to finance the increasing demand for borrowings from the private sectors as a result of better economic prospects. It is clear that the results for Model 3 and 4 are not stable as we find bi-directional causality when Reserve Bank discount rate is used as interest rate while uni-directional causality when Money Market rate is used as interest rate. This indicates that the Reserve Bank discount rate, which is under the control of the central bank, has impact on economic growth while the short-term 7 Results of the causal relationship between interest rate and economic growth, as well as interest rate and financial development are not the main focus of this study and hence excluded in Table 5.

13 258 S. M. Thangavelu and A. B. Jiunn, James Table 5. Granger causality test Model Lag Null hypothesis F-statistics P-value Causal inference 1. (Y, C, R) 2 C does not Granger-cause Y Accept H 0 Y does not Granger-cause C * Reject H 0 2. (Y, C, M) 2 C does not Granger-cause Y Accept H 0 Y does not Granger-cause C * Reject H 0 3. (Y, D, R) 5 D does not Granger-cause Y * Reject H 0 Y does not Granger-cause D *** Reject H 0 4. (Y, D, M) 4 D does not Granger-cause Y Accept H 0 Y does not Granger-cause D ** Reject H 0 5. (Y, E, R) 3 E does not Granger-cause Y ** Reject H 0 Y does not Granger-cause E Accept H 0 6. (Y, E, M) 3 E does not Granger-cause Y ** Reject H 0 Y does not Granger-cause E Accept H 0 Note: *, **, and *** indicate significance level of 10%, 5% and 1% respectively. Table 6. Summary of testing results Model Variable used Testing result Causal relationship 1 Y, C, R YÞC EGÞFD 2 Y, C, M YÞC EGÞFD 3 Y, D, R YÛD EGÛFD 4 Y, D, M YÞD EGÞFD 5 Y, E, R EÞY FDÞEG 6 Y, E, M EÞY FDÞEG Money Market rate does not. It is not surprising to find Reserve Bank discount rate having an important role on the banking sector and economic growth, since the discount rate is used by the Reserve Bank of Australia as one of monetary tools to stabilize the financial sector and domestic economy. In addition, as Fry (1997) argued that higher interest rates encourage entrepreneurs to invest in high return projects and thereby increase the productivity of capital used. Higher interest rates are likely to attract larger deposit base available for financing innovative investments that may lead to higher output growth. In our study the Reserve Bank discount rate is consistently higher than Money Market rate, therefore such outcome is not surprising. The evidence presented in Model 3 supports our hypothesis that bank deposits are being used extensively to finance loans, and thereby lead to higher output growth. One of the important characteristics in the Australian financial system is that the banking system is very well developed. Large number of branches, ATM, EFTPOS outlets and the use of sophisticated technology have facilitated many transactions. The existence of such establishments has strengthened the confidence of households and encouraged them to save in bank thereby creating a large deposit base for bank. It is believed that larger deposit base indicates that the banks are able to provide more financial products and services that may be conducive to the development of economy. The function of pooling savings by the financial intermediaries may enable investors to access more funds. This therefore allows large-scaled and profitable investments to be launched.

14 Financial development and economic growth in Australia Market-based system: Financial markets and economic growth It is found in Model 5 and 6 that causality runs from E (logarithm of equities turnover divided by nominal GDP) to Y (real GDP per capita). As Levine and Zervos (1996) argued, a well-developed stock market is able to provide different kinds of financial services as compared to the banking sectors. Hence, it may provide different forces to investment and growth than the development of the banking system. In addition, development of stock market might have resulted in more mobilized capital, diversified risks and availability of useful information required for investment. This is consistent with Obsfeld s (1994) view that international risk sharing through internationally integrated stock markets improves resource allocation and hence accelerates economic growth. The existence of regulatory bodies to correct financial market failures could have also contributed to such rapid development in economy. Hence, we may conclude that better functioning stock markets results in better informed investment decisions and thereby accelerate economic growth. 5. Conclusion The impact of the financial sector on the economic growth of the domestic economy depends on both the role of direct financing of the financial markets and indirect financing of the financial intermediaries. The empirical test results using financial indicators related to financial intermediaries suggest that the financial sector is reactive to the demand generated from the economic development i.e., economic growth causes financial development in Granger s sense. On the other hand, the results of using financial market indicators is also consistent with Schumpeter s (1911) view that the incremental flow of services by the financial sector are essential for funding investment in research and development and thus for economic growth (shown in Models 5 and 6). Financial deregulation seems to have worked positively on the economic growth of Australia. During the 1970s and 1980s the controls in financial sectors were removed and the Federal Reserve Bank adopted a more openmarket-oriented approach in managing the financial sector. In Model 5 and 6, we conclude that better developed stock market leads to higher economic growth. This is because development of stock markets allows risk diversification and better resource allocation. Since financial deregulation improves efficiency in terms of competition and greater accessibility to financial products, there are positive impacts on economic growth. Therefore the financial deregulation initiative in Australia in the early 1980s is consistent with our results. References Franklin A, Douglas G (2000) Comparing financial systems. MIT Press, Cambridge, MA Beck T, Levine R, Loayza N (2000) Finance and the sources of growth. Journal of Financial Economics 58: Beck T, Levine R (2002) Industry growth and capital allocation: Does having a market- or bankbased system matter? Journal of Financial Economics 64:

15 260 S. M. Thangavelu and A. B. Jiunn, James Caselli F, Esquivel G, Lefort F (1996) Reopening the convergence debate: A new look at crosscountry growth empirics. Journal of Economic Growth 1: Charemza WW, Deadman DF (1992) New directions in econometric practice. Edward Elgar, Cambridge Demetriades PO, Khaled AH (1996) Does financial development cause economic growth? Timeseries evidence from 16 countries. Journal of Development Economics 51: Demirguc-Kunt A, Maksimovic V (2002) Funding growth in bank-based and market-based financial systems: Evidence from firm level data. Journal of Financial Economics (forthcoming) Demirguc-Kunt A, Levine R (2001) Bank-based and market-based financial system: a crosscountry comparison. In: Demirguc-Kunt A, Levine R (eds) Financial structure and economic growth: Cross-country comparisons of banks, markets and development. MIT Press, Cambridge, MA, pp Enders W (1995) Applied econometric time series. Wiley, New York Engle RF, Granger CWJ (1987) Cointegration and error correction: Representation, estimation, and testing. Econometrica 55: Fernandez D, Galetovic A (1995) Schumpeter might be right but why? Explaining the relation between finance, development, and growth. Mimeo, John Hopkins University, SAIS, Baltimore, MD Fry MJ (1997) In favour of financial liberalization. The Economic Journal 107: Granger CWJ (1980) Testing for causality: A personal viewpoint. Journal of Economic Dynamics and Control 2: Granger CWJ (1988) Some recent developments in a concept of causality. Journal of Econometrics 39: Johansen S, Juselius K (1990) Maximum likelihood estimation and inference on cointegration with application for demand for money. Oxford Bulletin of Economics and Statistics 52: Johansen S (1988) Statistical analysis of cointegrating vectors. Journal of Economic Dynamics and Control 12: King RG, Levine R (1993) Finance and growth: Schumpeter might be right. Quarterly Journal of Economics 108: Kwiatkowski D, Phillips P, Schmidt P, Shin Y (1992) Testing the null hypothesis of stationarity against the alternative of a unit root. Journal of Econometrics 54: Lee K, Pesaran MH, Smith RP (1996) Growth and convergence: A multi-country empirical analysis of the solow growth model. Department of Applied Economics Working Papers, Amalgamated Series No. 9531, University of Cambridge Levine R (1997) Financial development and economic growth: Views and agenda. Journal of Economic Literature 35: Levine R (2001) Bank-based or market-based financial systems: Which is better? Carlson School of Management, University of Minnesota Levine R, Loayza N, Beck T (2000) Financial intermediation and growth: Causality and causes. Journal of Monetary Economics 46:31 77 Levine R, Zervos S (1996) Stock market development and long-run growth. The World Bank Economic Review 10: Levine R, Zervos S (1998) Stock markets, banks, and economic growth. American Economic Review 88: Luintel KB, Khan M (1999) A quantitative reassessment of the finance-growth nexus: Evidence from a multivariate VAR. Journal of Development Economics 60: Obsfeld M (1994) Risk-taking, global diversification, and growth. American Economic Review 84: Perron P (1997) Further evidence on breaking trend functions in macroeconomic variables. Journal of Econometrics 80: Rajan Raghuram G, Zingales L (1998) Financial dependence and growth. American Economic Review 88: Robinson Joan (1952) The generalization of the general theory. In: The rate of interest, and other essays. Macmillan, London, pp Schumpeter JA (1911) The theory of economic development. Oxford University Press, Oxford (Reprinted 1969)