Financial liberalisation and the capital structure of firms listed on the Johannesburg stock exchange

Abstract

This thesis examines the impact of financial liberalisation on the capital structure of non-financial firms listed on the Johannesburg Stock Exchange (JSE). The research hypotheses have been formulated and tested by using various econometric estimation procedures. The empirical assessment begins by resolving some methodological issues relating to the dating of financial liberalisation and the measurement of leverage. Next, a panel of 100 non-financial firms is constructed for the period 1989 to 2007. The pooled Ordinary Least Squares, fixed (within), and random effects (Generalised Least Squares) models are used to estimate the impact of various aspects of financial liberalisation on firm capital structure. Robustness checks are performed by using the instrumental variable technique and the Arellano and Bond (1991: 277) two-step Generalised Method of Moments (GMM) procedure. Irrespective of the procedure used, the lifting of international sanctions and stock market liberalisation is associated with a reduction in leverage for all sets of firms. Capital account liberalisation causes firms to access more debt, and there is mild support for domestic financial sector liberalisation. Stock market liberalisation provides avenues for firms to increase their debt maturity structure. The size of the stock market is positively associated with leverage, and banking sector development is negatively correlated to leverage. The effect of financial liberalisation is more pronounced for larger firms, suggesting that the process of financial liberalisation causes a wedge between small and large firms. Furthermore, regime dummies are interacted with firm specific determinants of capital structure. The notable facts arising out of the interactive dummy variable exercise reveal that the lifting of international sanctions and stock market liberalisation cause significant structural shifts in the profitability, growth and tax rate variables. This is evident for the book and market value of the debt to equity ratio. The results of the dynamic model of capital structure document evidence of transaction costs for both the pre and post liberalisation regime. However, it appears that transaction costs reduce dramatically in the post liberalisation regime. The associated coefficient of adjustment increases accordingly. This suggests that a developed financial system promotes competition amongst the issuers of capital, thereby lowering borrowing and transaction costs. The empirical significance of firm specific determinants in a closed economy appears to be weaker than that of the post liberalised regime. Furthermore, firm size speeds up the adjustment to the target level of leverage for firms operating in both the pre and post liberalisation regimes. Finally, the results confirm most of the theoretical predictions of capital structure theories.