Enhanced indexation of South African equities: a co-integration analysis

Abstract

This study presents an analysis of cointegration based tracking strategies; a classical index tracking strategy and a long-short equity market neutral tracking strategy. The first strategy attempted to track a reconstructed index, whilst the second strategy attempted to track an enhanced index. Quantitative portfolio managers have traditionally used the concept of correlation to analyse co-movements, which over years evolved into conditional correlation. In contrast, this study applied cointegration rather than correlation in portfolio optimisation. The concept of cointegration relies on the long-run relationship between time series i.e. stock prices and an index. The data that was used was the price history of the JSE Top 40 market index and its constituent stocks for the period 02/11/2009 - 31/12/2015. The study found that it was only the designed tracking portfolio of 10 stocks that was sufficiently cointegrated with the index (JSE Top 40), albeit at higher volatility at various times in the out of sample period and performance that slightly lagged the index even before accounting for transaction costs. As far as the enhanced index tracking strategy based on cointegration, the study found no convincing empirical evidence for the South African equities market. Further research should be done to test the robustness of the cointegration based index tracking strategies under different market conditions and its applicability within sophisticated trading strategies and stock selection methodologies.