Predicting Regime Shifts in Johannesburg Stock Exchange Allshare Index (Jse-Alsi): A Markov-Switching Approach

Abstract

It has been proven several times that linear models are unable to encapsulate nonlinear dynamics of macroeconomic and financial data such as inflation rates, exchange rates and stock prices to mention fewer. As a result, to overcome this problem, this current study adopted the nonlinear models due to the fact that they have required qualities to apprehend nonlinearity in a dataset. In order to predict a regime shifts, a five-day Johannesburg stock exchange allshare index (JSE-ALSI) spanning period from 02 January 2003 to 28 June 2019 was used as an experimental unit. This current study firstly employed Teräsvirta neural network test to detect the presence of nonlinearity and proceeded to estimate a two regime Markov-Switching autoregressive (MS-AR). The results of Teräsvirta neural network test revealed a highly significant nonlinearity with permanent seasonality as demonstrated by Kruskal-Wallis test. The predicted regime shifts by a latent dynamic allowed the autoregressive and variance parameters to promptly react to vital systemic shocks. As a result, this current study allowed volatility to oscillate between high and low volatility regimes that produced an expected duration of high volatility of two year and two months. This was a clear indication that there is a regime shifts in JSE-ALI which are modeled using Markov-Chain (MC) stochastic process. These findings may be used to inform robust policy making with the aim of safeguarding both the JSE and other global stock markets from the episodes of stock market crash. Moreover, other researchers can utilize the results of this study to calculate the risk associated with structural breaks and high volatility periods.

Financial modelling of cryptocurrency: a case study of Bitcoin, Ethereum, and Dogecoin in comparison with JSE stock returns.

“…However, this was not the case with the JSE stock data, where the single regime model outperformed the two regime models. This contradicts the findings of Oseifuah and Korkpoe (2019) and Makatjane and Molefe (2020) who found the single regime less effective. However, we are quick to note that we used a period that starts in 2017, whilst the studies mentioned above-considered sample periods that cover the global financial crisis of 2008.…”

Several explorations on how to construct an early warning system for local government debt risk in China

“…Regarding the contagion risk, Bianchi et al [38] take the network structure perspective and use the standard eigenvector centrality to model contagion in financial market; Anagnostou et al [39] incorporate contagion in portfolio credit risk model by using network theory; Berloco et al [40] use the network model to capture firms' fragility to shocks. Regarding the dynamic risk, Jutasompakorn et al [41] identify the banking crisis dates via the MS-AR model; Xaba et al [42] explore the performance of MS-AR model to forecast the quarterly exchange rate of South Africa; Makatjane and Kagiso [43] realize a dynamic early warning of the Johannesburg stock exchange all-share index through a two regime MS-AR model. Referring to the above practices, this paper will apply the network model and MS-AR model to construct a comprehensive EWS local government debt risk in China.…”

Identifying structural changes in the exchange rates of South Africa as a regime-switching process

“…The first breakpoint occurred in June 2003. According to Makatjane and Molefe (2020), this was due to the invasion of Iraq, after which the price of oil weakened, causing a depreciation in exchange rates, especially in SA. However, Bonga-Bonga and Kabundi (2015) further indicate that the SARB had been going ahead with a contractionary policy.…”