As we all grapple with the potential changes artificial intelligence and machine learning will bring to the industry, at some point in the future, there is one seminal change in the investment industry that we feel has not been explored enough, locally.

That change, is the evolution of portfolio construction beyond traditional asset classes. That asset pricing theory has evolved over the past couple of decades is beyond question. It has moved from the restrictive Capital Asset Pricing Model (CAPM) and efficient markets hypothesis to Fama and French models and behavioural finance models. Portfolio construction has not evolved much from Markowitz’s Modern Portfolio Theory to incorporate some of the newer contributions to finance.

For a long time, the success of alternative investments mainly hinged on two promises: the promise of diversification and the promise of alpha. The financial crisis of 2008 has shown the limitations of this framework. Firstly, that the hedge fund industry had confused volatility with diversification. Secondly, there had been little emphasis on bad times, since in reality markets could also underperform. If there was a good lesson we all learnt from the Global Financial Crisis of 2008 (GFC), it’s that in stress times correlations tend to 1, that is, asset classes exposed to a common risk factor, will behave similarly. Traditional
diversification failed when it was needed the most.