Non-parallel investment strategies in south african interest rate markets

• Quantifying SA non-parallel risk premia, rising during global risk-off events. • Developing butterfly strategies using novel weight schemes for risk-adjusted returns. • Scenario analysis showing slope duration's role in curve flattening/steepening. • Identifying optimal weight schemes and 3-6 month rebalancing to cut trade costs. • Finding SA excess returns, supporting semi-strong EMH and Market Segmentation Theory. This article analyses the viability of butterfly strategies in South African interest rate markets, challenging the efficient market hypothesis (EMH). It explores various weighting schemes, adapted from Martellini et al.'s (2002) and includes a high convexity weight scheme, which are applied to South African bond and swap curves, assessing their profitability and risk-adjusted returns. The study investigates both passive and tactical approaches, considering the impact of interest rate cycles and slope durations and examines the effects of trade costs and rebalancing frequencies on strategy performance. Mean reversion tests are conducted to evaluate the potential for tactical strategies, suggesting the existence of exploitable risk premia. The article aims to provide insights into optimal investment strategies for extracting non-parallel risk premia in the South African fixed-income market, considering market idiosyncrasies such as Market Segmentation Theory (MST). Overall, we find that the high convexity and duration and cash neutral weight schemes provide positive net yields that allow for balanced upside and downside investment risk.