Efficient strategies for optimising portfolios

For every investor and trader, the goal is to achieve maximum returns with minimal risk. This objective remains consistent regardless of the chosen investment strategy. In this context, various portfolio optimisation techniques exist to help investors extract the highest possible returns per unit of risk taken.

What is portfolio optimisation?

Portfolio optimisation refers to the process of selecting the optimal portfolio from a range of feasible portfolios to achieve specific objectives, such as maximising returns or minimising risks. Introduced by Harry Markowitz in the 1950s, modern portfolio optimisation techniques have significantly evolved from their foundational concepts.

What is Modern Portfolio Theory?

Modern Portfolio Theory posits that investors should not evaluate an investment’s risk and return characteristics in isolation. Instead, they should assess how it impacts the overall portfolio’s risk and returns. This approach allows for the construction of portfolios consisting of multiple assets that offer greater returns without necessarily increasing risk.

What is efficient frontier analysis?

An efficient frontier comprises investment portfolios expected to deliver the highest returns for a given level of risk. A portfolio is deemed efficient if no other portfolio offers higher returns for the same or lower level of risk.

Here are some widely acknowledged portfolio optimisation techniques that assist in identifying portfolios lying on the efficient frontier, thereby maximising returns:

Mean-Variance Optimisation (MVO)

MVO, pioneered by Harry Markowitz, balances expected risk and returns. It uses programming to determine optimal portfolio weights that either maximise expected returns for a given risk level or minimise risk for a given expected return. However, this method assumes a normal distribution of returns, which may lead to concentrated portfolios due to unstable estimates of expected returns and covariance.

Black-Litterman Model

This advanced model enhances MVO by incorporating investor views alongside market equilibrium, resulting in a more stable and diversified portfolio. It provides realistic estimates of expected returns.

Risk Parity

Risk Parity focuses on balancing the risk contribution of each asset rather than capital allocation. This approach creates more balanced and diversified portfolios, though it overlooks asset class returns.

Factor Investing

This strategy builds portfolios based on factors like value, momentum, size, quality, and low volatility, which are believed to drive asset returns. However, factor effectiveness varies across market conditions, necessitating factor timing.

Robust Optimisation

Robust Optimisation addresses uncertainty and estimation errors in inputs like expected returns and covariance. It optimises portfolios by considering worst-case scenarios within defined uncertainty sets.

Machine Learning-Based Optimisation

This approach uses algorithms such as reinforcement learning and neural networks to predict asset returns and risk and optimise portfolio construction. It adapts to market changes and enhances predictive accuracy.

Conditional Value at Risk (CVaR) Optimisation

CVaR optimisation minimises potential losses beyond a specified threshold (tail risk). It effectively captures extreme risks and tail events.

Stochastic Programming

This approach uses probabilistic models to account for uncertainty in future returns and scenarios, improving decision-making over multiple periods.

Bayesian Portfolio Optimisation

Bayesian methods refine estimates of expected returns and risk using historical data and update beliefs with new information, making them adaptive to changing market conditions.

These portfolio optimisation techniques offer distinct strengths suited to various investment scenarios and objectives. Combining insights from multiple techniques enhances portfolio management effectiveness and resilience.

(The writer is Vice President - Research, Teji Mandi)