In the aftermath of the Asian financial crisis of 1997-1998, development of the domestic bond market particularly the corporate bond market, has been high on the policy agenda of many economies. This is especially the case in developing countries. Amongst other lessons learnt, the Asian crisis highlighted the problems of over-dependence on the banking sector and the underdevelopment of the bond markets.
There are many studies outlining the rationale and importance of developing government and corporate bonds markets. However, there are very limited empirical studies on the direct linkage of development between the two markets. Qualitatively, some literature suggests that the government bond market must be in place first in order for the corporate bond market to develop. The government bond market development can have two opposite effects on corporate bond market. The first effect is the positive, liquidity or market creation effect. The other is the negative crowding out effect.
The objective of this thesis is to investigate the impact of government bond market on the development of corporate bond market based on a sample of 45 countries over the time period 1989 to 2007. The analysis uses the market size to GOP ratio as an indicator of bond market development for both the government and corporate bond markets, and also includes a number of other determinants of the corporate bond market as control variables.
This thesis makes use of the Generalised Methods of Moments (GMM) approach proposed by Arellano and Bond (1991), to estimate a dynamic panel model. The purpose is to determine which of the two competing effects of the government bond market dominates. The econometric results show that in general, if government bond market size is below a certain critical level, the crowding out effect dominates. If the size is above that critical level, the liquidity effect will dominate. The finding holds across the two subsamples of developed countries and emerging markets economies. Based on the sample data obtained, the sizes of most government bond markets are found to be below the estimated critical level and thus the crowding out effect found to be dominant.