Equity super sectors connectedness and its determinants: evidence from the Johannesburg Stock Exchange.

Babatunde, Samuel Lawrence.

Equity super sectors connectedness and its determinants: evidence from the Johannesburg Stock Exchange.

Files

Babatunde_Samuel_Lawrence_2023.pdf (9.4 MB)

Date

2023

Authors

Babatunde, Samuel Lawrence.

Abstract

Everything depends on everything else. More importantly, macroeconomic and financial connections have proved to be more fundamental compared with others. The reality of dynamic connectedness and time varying correlation as precursors to contagion and systemic risk are proven through the super sectors, namely the Automobile and Parts, Chemical, Telecommunication, Technology, Energy, Health, Finance, Insurance and General Industrial super sectors of the Johannesburg Stock Exchange, with daily sample period from 1 January 2006 to 31 December 2021. The first objective is to determine the systematically important super sectors in the different extreme periods. The second objective is to determine the return linkages of the equity super sector, while the third objective is to examine the dynamic connectedness and the shock propagation among the super sectors during the extreme risk events. Finally, the fourth objective is to evaluate the determinants of volatility connectedness of the JSE equity super sectors. The different extreme events considered alongside the full sample periods for this study are the 2007/2008 Global financial crisis (GFC), the 2009-2011 European Debt Crisis (EDC), the 2017-2018 U.S-China trade war (U.S-China TWR) and the late 2019-2021 COVID-19 pandemic. This study employs the Page et al., (1999) model with the Granger causality model of Billio et al., (2012) to accomplish objective one. While in objective two, the DECO-GARCH model of Engle and Kelly (2012) was employed to establish the time varying equicorrelations status of the super sectors through the rolling window analysis. For objective three, the realised volatilities of the super sectors were obtained through the Garman and Klass (1980) model and thereafter, the dynamic connectedness and direction of propagation were determined through the Diebold and Yilmaz (2009, 2012 and 2014) model alongside the TVP-VAR of Antonakakis et al., (2020). The study further employed the nonlinear autoregressive distributed lag (NARDL) model to determine the asymmetrically significant determinants of total sectorial volatility connectedness of the JSE market in the fourth objective. Findings from this study revealed the Telecommunication super sector is the most systematically important super sector during the full sample size analysis. It was revealed that the equicorrelation of the super sectors is positive and high, this was also the case for the rolling window results except for the years not within the extreme period, yet the least equicorrelation was 0.1491 for the year 2012-2013, while the highest was 0.7022 for the COVID-19 pandemic period. It was also established that the total connectedness of the sample period and the different extreme periods were high, suggesting a high interconnectedness of the super sectors. Lastly, the determinant estimation results show LSAVI, LDMR and LEPU as the asymmetrically significant drivers of total sectorial volatility connectedness on the JSE market. This study is the first to investigate sectorial connectedness, equicorrelation and the determinants of volatility connectedness in South Africa and in Africa at large. This study contributes to the limited literature on systemically important equity super sectors and sectorial dynamic connectedness and dynamic equicorrelation in the emerging market. First the result shows that the Telecommunication sector is the most important node for the EDC, the U.SChina trade war and the COVID-19 pandemic periods. While the Insurance and the Energy are the highest ranked super sectors amongst the network of super sectors for the full sample period and for the GFC period, hence making these super sectors the most systemically important nodes during these selected periods. It also shows that the sectorial common equicorrelation on the JSE is high and time varying with higher values for the year where extreme events occurred such as the GFC, EDC, and the COVID-19 pandemic period. This result is also a revelation that during the period of financial or economic crisis correlation of sectors are high compared to non-crisis periods. Third, the dynamic connectedness results show that the sectors on the JSE are interconnected and a shock to one sector can have a spillover effect on another close sector in the value-chain. Fourth, the South African volatility index, the Economic Policy Uncertainty and the Domestic Market Return are symmetrically and asymmetrically significant determinants of the sectorial volatility connectedness of JSE market. These findings from this study have implications for economic policy makers, portfolio and fund managers, foreign and local investors, sector regulators and researchers/academics in the field of finance.