Browsing by Subject "Signal Processing"

Now showing 1 - 1 of 1
  • No Thumbnail Available
    Item
    Open Access
    Factor investing for South African pension funds
    (2026) Letlaka, Thabo; Botha, Pieter
    This study investigates optimal factor portfolio strategies in South African, developed, and emerging equity markets. Leveraging Fama-French factor models, we construct single- and multi-factor portfolios based on the factors: size; value; momentum; quality; and low volatility. Single-factor portfolio construction follows an approach that begins with processing market data into factor signals and ends with portfolio construction based on transformed factor signals. Our results show that the return on assets and return on equity signals are the top performing signals on a risk-adjusted basis across all three markets, followed by the momentum signals. Multi-factor portfolios are constructed using factor equal weighting (EW), minimum variance (MV), and equal risk contribution (ERC) methods. Our results show that EW portfolios consistently outperform traditional indices on a risk-adjusted basis, and thus can be viewed as giving an investor optimal strategic factor exposure. ERC portfolios outperform EW portfolios and can be considered as giving investors either strategic or tactical factor exposure. On tactical factor allocation strategies, we construct multi-factor portfolios based on tactical factor momentum and volatility indicators and find that the tactical factor allocation method based on the factor momentum indicator is robust across all three markets, especially in South Africa and developed markets.