Stock return synchronicity, earnings informativeness, and institutional development : evidence from African markets

Abstract

This thesis contains the outcome of three separate but interrelated empirical analyses on stock return synchronicity, earnings informativeness and institutional development in a sample of African markets. The first analysis investigates the level and determinants of stock return synchronicity. Some recent studies have provided a theoretical argument that contrary to conventional wisdom, stock return synchronicity can be high in a strong information environment as market participants are less surprised about the occurrence of future events arising out of having more available information. This may therefore imply that stock return synchronicity can be conversely low in a relatively weaker information environment. The first empirical analysis of the thesis tests this conjecture using a total of 616 firms across five African countries (Botswana, Ghana, Kenya, Nigeria and South Africa) over the 2005-2015 period. The main measure of stock return synchronicity used is the R2 from a market model regression of individual stock returns on the returns of a corresponding market index. The findings show that on average, firms in African markets do not exhibit high levels of stock return synchronicity, providing support for the view that stock return synchronicity can be low in markets with relatively weak transparency and conversely high in strong information environments. In regression analysis, the main driver of stock return synchronicity, however, is firm size, whilst contrary to some previous studies, ownership structure has no impact. These results are robust to different measures of stock return synchronicity that include both a lagged market index and a world market index. They are also robust to different estimation techniques including Fama-Macbeth regressions and ordered probit regressions. The second empirical analysis of this thesis investigates the informativeness of earnings announcements in African stock markets and examines whether conditional on the level of stock return synchronicity, market reactions to earnings announcements are influenced by firm fundamentals or trading frequency. This chapter uses a set of 1762 annual earnings announcements across 369 firms from three countries (Kenya, Nigeria and South Africa) over the 2005-2015 period. In univariate analysis, the main measure of earnings informativeness is Normalised Volatility, which divides volatility during a 21-day event window by volatility in a period of 120 days outside of the event window. Normalised volatility indicates that earnings announcements are informative across the sample. The results are driven by less frequently-traded stocks (stocks which experience price changes of between 50% to 74% of trading days in the previous year), although informativeness is also present for highly traded stocks (stocks which experience price changes in at least 75% of trading days in the previous year). Informativeness manifests more clearly at announcement and in the postannouncement window, and there is little evidence of leakage. Cross-sectional tests, using regression analysis, provide evidence of an effect of both earnings fundamentals and investor behaviour on stock returns around earnings announcements. The third and final empirical analysis examines the impact of two institutional factors— the mandatory adoption of IFRS and the perceptions of corruption, on the market reactions earnings informativeness within the same period of 2005-2015. The first part of the analysis tests whether earnings became more informative following the mandatory adoption of IFRS. This analysis is restricted to only Nigeria and South Africa as Kenya adopted the use of IFRS prior to the start of the sample period of this study. The second part of this analysis tests the impact of the perception of corruption on earnings informativeness in a sample made of firms from Kenya, Nigeria and South Africa. Both univariate and regression results show that the mandatory adoption of IFRS did not lead to significant improvement in earnings informativeness. This finding is consistent with the view that the improvement in accounting standards must be accompanied by effective mechanisms of enforcement in order to realise their capital market benefits. However, with respect to corruption, there is a significant negative impact on earnings informativeness in terms of abnormal trading volume. Overall the findings in this chapter point to the growing importance of how the institutional environment can have capital market implications for firms. Therefore, more work needs to be done to strengthen the institutional framework in order to further enhance the price-discovery process in these markets.