A comparative analysis of the volatility nature of cryptocurrency and JSE market


Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 International License

Despite the rapid growth of developing markets, aided by globalization, comparative studies of cryptocurrency and stock market volatility have focused on the developed markets and neglected developing ones. In this regard, this study compares cryptocurrency volatility with that of the Johannesburg Stock Exchange (JSE), a developing market. GARCH-type models are applied to daily log returns of Bitcoin, Ethereum, and the FTSE/JSE 4O in two ways. Firstly, the models are applied directly; secondly, structural breaks are tested and accounted for in the models. The sample period was from September 18, 2017, to May 27, 2021. The results show higher volatility and higher volatility persistence in cryptocurrency than in the JSE market. They also show that persistence is overestimated for cryptocurrencies when structural breaks are not accounted for. The opposite was true for the JSE.
Moreover, the two cryptocurrencies were found to have close to identical volatility plots that differ from that of the JSE. High volatility periods of cryptocurrency also did not coincide with that of JSE and those of JSE did not coincide with the cryptocurrency ones. There is also evidence of an inverse leverage effect in cryptocurrency, which opposes the normal leverage effect of the JSE market.

view full abstract hide full abstract
    • Figure 1. Daily log returns
    • Figure 2. ACF plots of returns and squared returns
    • Figure 3. Daily log returns
    • Figure 4. QQ plots and ACF plots for the standardized residuals
    • Figure 5. Volatility plots with Bitcoin atop followed by Ethereum and lastly JSE
    • Figure 6. Structural breakpoints as identified by the PELT method
    • Figure 7. QQ plots and ACF plots for residuals for models with structural breaks
    • Table 1. Descriptive statistics of daily log-returns Bitcoin, Ethereum, Dogecoin, and JSE
    • Table 2. Stationarity tests for the returns
    • Table 3. Model selection
    • Table 4. Parameter estimates for the selected models
    • Table 5. Breakpoints identified in the return series
    • Table 6. Parameter estimates for models with structural breaks
    • Conceptualization
      Forbes Kaseke
    • Data curation
      Forbes Kaseke
    • Formal Analysis
      Forbes Kaseke
    • Investigation
      Forbes Kaseke, Shaun Ramroop, Henry Mwambi
    • Methodology
      Forbes Kaseke, Shaun Ramroop, Henry Mwambi
    • Project administration
      Forbes Kaseke, Shaun Ramroop, Henry Mwambi
    • Software
      Forbes Kaseke
    • Validation
      Forbes Kaseke, Shaun Ramroop, Henry Mwambi
    • Visualization
      Forbes Kaseke, Shaun Ramroop
    • Writing – original draft
      Forbes Kaseke
    • Supervision
      Shaun Ramroop, Henry Mwambi
    • Writing – review & editing
      Shaun Ramroop, Henry Mwambi