Structural Changes and Dating of Stock Market Liberalisation: Evidence from the Jse Securities Exchange South Africa

Abstract

In this paper the authors examine the problem of dating stock market liberalisation using time series of South African stock market data. Defining the date of stock market liberalisation as that on which there is a structural change in time series data, they test for structural breaks. They perform tests on time series of monthly dividend yield, volume of shares traded and aggregate stock price of the JSE Securities Exchange. From the time series of the aggregate dividend yield and stock market liquidity, they report two structural breaks -the first of which occurred in February 1990, and the second in December 1992. From time series of the real aggregate stock price one structural break that occurred in December 1992. Noteworthy is that the structural breaks occurred earlier than the JSE's official liberalisation date of March 1995, thus suggesting political and economic risks were the more binding constraints to foreign investment than legal barriers.

Notes

1 It is essential to note the following difference regarding the way the JSE and the SARB compute the share price index. Whereas the JSE use the free float of companies that are actively traded to compute the share price index (which is not available in the public domain), the SARB computes its index by including all listed companies to overcome possible daily volatility.

2 Other authors such as Banerjee et al. (1992), and Christiano (1992) among others, have proposed alternative tests for unknown breaking points. However, the search for break points of this paper has solely limited to three known dates.

3 See Thompson and Ward (1995) for a review of studies on the efficiency of the JSE and the related random walk theory.