Portfolio Optimisation for Nigerian Pension Funds Using Mean-Variance Analysis

Nigerian Pension Fund Administrators face the dual mandate of preserving capital and generating competitive real returns for an expanding contributor base approaching 10 million registered members, within an asset allocation framework regulated by the National Pension Commission investment guidelines. This study applies Markowitz mean-variance portfolio optimisation to identify efficient frontier asset mixes for the Nigerian pension fund universe using monthly return data for the period January 2019 to December 2023. Asset classes included in the analysis are Nigerian Stock Exchange All Share Index equities, FGN Eurobonds, domestic FGN bonds across three tenor buckets, state government bonds, commercial real estate investment trusts, and money market instruments. Covariance matrix estimation is performed using the Ledoit-Wolf shrinkage method to address estimation error in high-dimensional settings. The Capital Market Line is derived assuming the three-month treasury bill rate as the risk-free rate. Efficient portfolios at five target return levels are characterised and compared against the average PFA portfolio allocation published in National Pension Commission quarterly reports. Results reveal that current average PFA allocations are inefficient, lying below the efficient frontier, with suboptimal overweighting of domestic government bonds. A rebalanced portfolio on the efficient frontier achieves equivalent expected return with 23 percent lower portfolio volatility. Keywords: mean-variance optimisation, pension fund, efficient frontier, Nigerian capital market, asset allocation.