Investment portfolio optimization is a complex and often contentious subject, posing challenges and sparking debates among financial experts. The predominant philosophy taught in academic finance programs revolves around two key frameworks: the Mean-Variance framework, pioneered by Harry Markowitz, and the Capital Asset Pricing Model (CAPM), developed by William F. Sharpe, John Lintner, and Jan Mossin in the early 1960s. The Mean-Variance framework, developed by Harry Markowitz, also plays a role in understanding how to optimize investment portfolios. However, both CAPM and the Mean-Variance model have their limitations when applied to real-world situations. These models face criticism because they make certain assumptions that might not always hold true. As finance evolves and we gain new insights, researchers are continuously working to find better ways to optimize investment portfolios, leading to ongoing innovation in the field.
The Capital Asset Pricing Model (CAPM) builds upon Markowitz’s mean-variance model, but it inherits some limitations and adds its own assumptions, which can make it challenging to use for alternative investments, like hedge funds, and non-linear assets. These key assumptions can be simplified as follows:
- All investors are assumed to have the same expectations about how assets will perform. This means they all agree on mean and variance as the only way of market assessment, making everyone see the market opportunities in the same way.
- Investors are assumed to be logical and risk averse. They aim only to maximize expected utility by end of the time horizon.
- The markets are perfect with no taxes, inflation, transaction costs, and short selling limitations.
- Investors can borrow and lend any amounts at the risk-free rate.
- All assets are perfectly liquid.
- Asset distribution of returns are normal, i.e., perfectly bell-shaped.
- The markets are in equilibrium, and prices cannot be manipulated.
- The total number of assets on the market and their quantities stay the same during the defined time frame.
While the Capital Asset Pricing Model (CAPM) and Mean-Variance models hold significant academic value, their practical application in today’s dynamic markets and with complex financial instruments is increasingly questionable. Let’s narrow our discussion to portfolio optimization only. There is a strong reason why the mean-variance portfolio optimization dominating in academic finance programs is rarely used in practice by investment practitioners and asset managers – it doesn’t work as expected. When performing portfolio optimization, our goal is to identify asset allocations corresponding to highest return and lowest risk. However, if risk is measured solely by variance, we tend to either overestimate risks for positively skewed distributions or underestimate them for negatively skewed ones. Keep in mind that real-life asset return distributions are rarely normal.
In recent years, several advanced portfolio optimization frameworks have emerged to overcome the limitations of traditional mean-variance optimization methods. One such example is the Risk Shell Portfolio Optimization system developed by ABC Quant. This portfolio optimization framework was initially created to optimize hedge fund portfolios (specifically hedge fund of funds) and has undergone continuous refinement and enhancement over the past two decades. Currently, it can be effectively applied to any multi-asset portfolios including equities, funds, private equities and so on. Below are some noteworthy unique features of this framework:
- Advanced risk optimization statistics that overcome the limitations of the mean-variance model, including CVaR, Maximum Drawdown, Conditional Drawdown, Omega Ratio, and more.
- Macroeconomic factor constraints, which allow constructing market-neutral portfolios virtually immune to severe market events. This feature is particularly crucial given the current market uncertainties.
- Integrated stress testing models to mitigate risks related to extreme market conditions when optimizing portfolios.
While it might seem excessive to combine macroeconomic factors and stress testing with portfolio optimization, these features actually address one of the major shortcomings of conventional portfolio optimization. Even a perfectly optimized portfolio can pose significant risks during extreme market conditions and stressful events, such as the Credit Crunch of 2008 or the COVID-19 pandemic.
Another crucial aspect of portfolio optimization, which is often underestimated, is optimization backtesting. The optimization process is complex, involving various settings and options. With numerous objective functions and constraint models to consider, it’s easy to end up with hundreds of potential optimization scenarios. The challenge lies in determining which scenario would produce the best results for your specific portfolio in advance. To streamline and expedite the process of selecting the optimal optimization scenario, Risk Shell Portfolio Construction platform provides a dedicated Optimization Backtesting feature. This component can efficiently test hundreds of scenarios simultaneously, saving time and effort.
Going beyond conventional max-min optimization constraints the Risk Shell platform offers hundreds of user-defined criteria which can be used as constraints, for example liquidity, manager rating, geographical or strategy exposures, stress test drawdowns and many more.
In summary, the need for portfolio optimization in today’s complex markets depends on the type of optimization used. Traditional models like CAPM and Mean-Variance may struggle to handle real-life scenarios effectively and might not be as useful. However, emerging alternative optimization solutions offer promise in constructing robust portfolios with customizable risk-return profiles, providing potential benefits in navigating today’s challenging investment landscape.