Financial Development,Stock Market And Economic Growth:Comparative Study

For many years now economists have argued that well-functioning and integrated banks and stock markets can have positive and accelerating effect on economic growth. The development of financial intermediaries can positively affect economic growth through various channels of transmission, such as savings mobilization and the efficient allocation of capital.In the last three decades, the availability of capital markets data and the increase in theoretical literature have suggested that well-functioning stock markets can spur long-run economic growth by changing levels of technical progress, savings rates and the overall efficiency of economy. However, there is an opposing view that the development of financial intermediaries, particularly capital markets, can have a detrimental effect on economic growth, especially in developing countries. This study explores causality and co-integration between banking, stock markets development and economic growth. It presents a comprehensive review of the theoretical and empirical literature to establish the framework that connects this vast nexus. We have considered four case-countries for empirical analysis: Croatia, Serbia, Slovenia and China to inspect and compare the banks-stocks-growth nexus in countries that vary in size, geographical location and stage of economic development but are all former socialist countries, with exception of China, which still presently defines itself as a socialist country with respect to economic and social system.This research aims to conduct empirical analysis of bank and stock market development in the economic growth process by examining the causality inferences in different Vector Autoregressive environments with a focus on omitted variable condition. Excluding and important variable like a stock market in analyzing the causality between banking development and economic growth might have produced spurious results in previous research, as suggested by Coporale and Pitis (1997). This assumption also draws from Beck and Levine (2002) who claim that leaving out capital markets from models of financial development makes for spurious results. Co-integration of selected variables was also explored and serves as a cross-check for Granger Causality results.In addition to causality and co-integration estimations this study also employed the Impulse Response Function (IRF) technique to examine which part of the banking and stock market sectors has the strongest and most enduring effect on economic growth when external,,shock" is introduced in the system. Varying econometrical approaches were used to provide clear and comprehensive results. First, an Augmented Dickey Fuller (ADF) unit root test was applied to test the variables of each country for the existence of unit roots. Second, bivariate and trivariate VAR models were developed for each of the four case-countries to inspect the separate and joint effects of credit institutions and stock market development on economic growth. The VAR model was used to capture the linear interdependencies among multiple time series. Third, the causality between paired variables of both bivariate and trivariate VARs was inspected trough the approach of Toda and Yamamoto (1995) to Granger non-causality. Fourth, the inference regarding causality direction was cross-checked by the Johansen Co-integration method. Finally, Impulse Response Function (IRF) analysis was applied to the trivariate VARs for each of the countries except Serbia where a bivariate banking-growth VAR was used to examine the interrelationships between variables in the VAR system.To stress out the difference in causality results that might follow from incorporating a previously omitted variable three types of Vectro Autoregressive (VAR) models were developed for three of the case countries in this study while for Serbia only a bivariate VAR was developed due to missing stock markets development data. First, a bivariate VAR examined the relationship between banking development and economic growth. Granger causality results, when inspected in a bivariate banking-growth model, confirm the "demand-following"hypothesis for the two case countries; Croatia and Serbia. No significant causality was established for Slovenia, while a ,,supply-leading" hypothesis was confirmed for China. Second, a bivariate VAR explored the relationship between stock market development and economic growth. Causality inferences from this model confirm the ,,supply-leading" hypothesis for two case-countries; Croatia and Slovenia. No significant causality was established for China. Due to missing data a stock-growth bivariate VAR model could not be developed for Serbia. Third, a trivariate VAR exploring the banking-stock markets enter the VAR model simultaneously with banking development. Strong empirical evidence from the trivariate VAR model confirms the,,supply-leading"hypothesis for two case-countries; Croatia and Slovenia. For China ,,bi-directional" causality inference was established between banking development and economic growth, while a ,,demand-following" hypothesis was confirmed in the case of stock markets and real output growth. A trivariate VAR model could not be developed for Serbia.The empirical results provide strong support for theories according to which well-developed and integrated banks and stock markets can promote and foster long-term economic growth and promote economic development by fuelling the engine of growth through faster and more efficient capital accumulation and distribution.Comparing causality in bivariate and trivariate models gave a new view to the dynamic properties of the system. When Granger Causality was inspected in a trivariate VAR model most of the variables that proxy for banking and stock market development entered regression significantly unlike in the separate bivariate models. Comparing the bivariate and trivariate models causality results revealed that the addition of the third variable, stock markets, to the banking-growth VAR model was relevant for all three case-countries and that the direction of causality changed because of this addition.