Several leading asset management firms claim they can improve outcomes for investors by combining quant investment with traditional active management. This seems a logical step given the growing availability of data and the pace of progress in the field of information technology.
But make no mistake, our understanding of investment is not changing fundamentally. The application of ‘big data’ and artificial intelligence (AI) to portfolio management should therefore not be hailed as a revolution. Investors must remember this limitation to avoid being swayed by marketing messages.
This is not to deny that the rising power of quants is an exciting development. Proper use of non-financial information, such as news and social media sentiment, can give managers an edge. AI can help identify investment opportunities or improve risk-management processes. By leveraging on these opportunities, managers can deliver better returns for clients. It is essential for managers to stay on top of these developments and not be left behind.
At the same time, traditional portfolio management is not going to be wiped out by the rise of quants. Skilled high-conviction managers will always aspire to deliver outperformance to their clients. It is likely, however, that skilled managers will become fewer and further between.
Nevertheless, discretionary management can work effectively with systematic managers to deliver higher returns for investors. The winners in this area will be those who overcome the organisational and cultural challenge that comes with combining the two approaches.
These are exciting prospects but there is no investment management revolution in sight. Despite the significant amount of academic research, particularly in fields such as behavioural finance, the theoretical concepts underpinning portfolio management remain the same. By and large, portfolio managers apply the capital asset pricing model (CAPM), which shows significant weakness in empirical tests.
Behavioural finance has introduced new useful perspectives but portfolio managers rely on models that produce unreliable predictions of economic and financial behaviour. There are no widely accepted alternatives so far.
AI techniques could even help researchers improve on classical models but the basic problems of investment remain the same.
Until a new modern theory that better describes how financial markets work is developed the computers will work on the basis of these classical, and imperfect, models. The results produced by AI-enhanced models might be slightly better but investors should not expect anything radically new.
Carlo Svaluto Moreolo, Senior Staff Writer