Portfolio optimization and asset dependency analysis across investment regimes and select asset classes in Kenya

Abstract

This study investigated the performance, dependency structures, volatility dynamics, and portfolio optimization strategies for the Nairobi Securities Exchange (NSE) 20 Index, gold, Bitcoin, and government bonds over the period from 2018 to 2023, within Kenya’s emerging market context. Employing a quantitative approach, the research utilizes daily price data from the NSE, London Bullion Market Association, Coin Market Cap, and Central Bank of Kenya, analyzed through the Markov Regime-Switching Model (MSM), copula models, GARCH volatility modeling, and mean-variance optimization. The objectives were to assess asset performance, examine interdependencies, evaluate volatility impacts, and develop optimal portfolio strategies across distinct economic regimes. Findings revealed three regimes: pre-crisis stability (2018–February 2020), crisis turmoil (March 2020–December 2021), and post-crisis recovery (2022–2023). The NSE exhibits a negative return and high volatility in crisis, while gold and bonds provide stability, and Bitcoin shows speculative gains. Dependency shifts indicate stronger crisis co-movement with volatility peaking in turmoil. Optimized portfolios adapt, favoring Bitcoin in stability, bonds and gold in crisis, and a balanced mix, Bitcoin in recovery, outperforming a static benchmark. The study concludes that regime-specific strategies enhance risk-adjusted returns, recommending dynamic allocations for investors and regulatory enhancements for market resilience. These insights contribute to financial decision-making in Kenya’s volatile market environment.