Structural instability and volatility shifts in emerging-market banking equities: evidence from South Africa's big five banks
Abstract
Banking-equity returns in emerging markets may change markedly between calm and stressed periods. This study examines structural instability in the daily returns of South Africa's five largest listed banks: Standard Bank, FirstRand, Absa, Nedbank, and Capitec. Daily closing prices from the IRESS Research Domain dataset cover 27 January 2016 to 23 January 2026. Absolute and squared log returns measure changes in typical return magnitude and volatility intensity. The empirical framework combines the Zivot–Andrews unit-root test, Bai–Perron multiple-break estimation, the robust ICSS κ 2 variance-stability test, common banking-factor analysis, post-estimation regime characterization, break clustering, and placebo and no-break comparisons. All empirical analyses were implemented using Python 3.12.4 (Python Software Foundation, Wilmington, Delaware, USA), distributed through Anaconda, on a 64-bit Windows 11 operating system. Every bank exhibits breaks in absolute and squared returns. The strongest clustering occurs in February and March 2020, with further clustering in September and November 2020. FirstRand records the most absolute-return breaks, whereas Standard Bank and Nedbank record the most squared-return breaks. The robust ICSS κ 2 test rejects unconditional variance stability for Standard Bank, FirstRand, Absa and Nedbank, but not for Capitec. Several breaks remain after common banking movements are removed, indicating heterogeneous bank-level adjustment. Additional transformed-return breaks occur in 2016–2018 and 2022. Structural instability is multidimensional and depends on the risk proxy examined. Breaks in return magnitude and volatility intensity are accompanied by statistically supported unconditional-variance shifts for four of the five banks, while Capitec shows transformed-return breaks without a confirmed robust variance shift. The findings support bank-specific, break-aware risk monitoring and caution against treating all forms of instability as equivalent.