Abstract
study adopted multiple hypothesis testing framework to address a crucial question within the emerging discourse of rules-based investing of whether the way investors impound securities factor information has a material impact on the performance of multifactor portfolios. To answer the question, the study created and compared 24 multifactor portfolios using the integrated approach which aggregates factor characteristics at the security level and the combination approach which combines sleeves of standalone single-factor portfolios. Simulation portfolios were created using 10-year financial data of eligible Johannesburg Stock Exchange companies. The study found that while the integrated portfolios delivered higher absolute and risk-adjusted returns across most of the 12 pairs of simulated multifactor portfolios, several information and Sharpe ratios differentials appeared insignificant when subjected to multiple hypothesis frameworks which account for the number of trials. Instead, the study found that style blending is a more powerful driver of the profitability of multifactor portfolios than construction methodologies. On implementation, the results showed that when neutralising the impact of trade netting, the combination and integrated approaches produce statistically comparable outcomes. On diversification, the study found in favour of the combination approach, but the differentials were marginal and not meaningful for investors to pursue
Key words
Integrated approach, combination approach, multifactor portfolio, multiple hypothesis testing, smart beta, factor investing, risk parity, low volatility, momentum, value, factor intensity.