Vincent Ateka
Abstract
This study explores the impact of incorporating financial and sustainability constraints into portfolio optimization using Sequential Least Squares Quadratic Programming (SLSQP). Building on the traditional 60:40 stock-bond allocation, four models are evaluated: a benchmark portfolio, a return-constrained model, a model integrating Minimum Spanning Tree (MST) asset selection, and a final model that adds an ESG threshold. Using South African stocks and bonds from January 2022 to December 2023, the findings show that while none of the portfolios outperformed the risk-free rate, constrained models achieved improved expected returns and marginally better Sharpe ratios. The ESG-integrated model demonstrated the highest risk-adjusted performance, supporting the view that sustainability considerations can enhance portfolio stability. However, the limited asset universe constrained diversification benefits. Overall, the results underscore the value of constraint layering in portfolio design and affirm the utility of SLSQP for solving complex, real-world optimization problems in emerging markets. The study highlights both the financial and ethical benefits of investing in ESG stocks for South African investors.