A lot has been said about the future perspective for portfolio managers with the increasing possibilities of artificial intelligence (AI). Sometimes I get the impression that AI will replace more or less everyone in asset management and especially in portfolio management. But can such an assumption pass a reality check?
It is true that a lot of work in portfolio management can be taken over by AI, but I don’t think that AI will replace portfolio managers at all. In fact, the fund management industry has been using algorithms for decades to determine valuations of markets and securities. These algorithms often use data which have been generated by tools using natural language processing (NLP) to extract the data from company reports and media reports. In addition to this, many fund management companies have developed quantitative tools to evaluate the attractiveness of markets and asset classes, which are used to steer the asset allocation and risk profile or the respective portfolios.
That said, most of these portfolios still have a human interface (portfolio manager) who is judging the outcome from the quantitative models, as some results, especially when so-called black swan events appear, may not make sense and can lead to misinterpretations and therefore wrong decisions by the model. Therefore, it is no surprise that asset managers employ data scientist to secure a high data quality by cleaning the data generated from the NLP tools and to keep the models up to date. Especially the latter is another task which currently can’t be done by AI itself, since the evaluation of the eligibility of single data points within the selection process is somewhat depending on the wider understanding of what is going on in a given economy. Even as AI has currently some flaws, one needs to bear in mind that we are currently at the starting point of a wave of AI-related innovations.
With regard to this, I do believe that AI will become much better in identifying malicious data and might be able to identify the best measures to identify the securities which fit the approach of the fund manager best. This means that fund managers and analysts will appreciate AI-based solutions, as these new tools will perform basic, as well as somewhat advanced tasks in the areas of quantitative research and portfolio construction for them. The reduced quantitative workload may enable them to perform more in-depth qualitative work, which hopefully will lead to an outperformance of the respective fund compared to its benchmark.
To answer the question raised in the headline, I believe that AI will be a game changer in portfolio management as the vast majority of quantitative tasks will be performed by AI in the future, but I don’t think that AI will replace analysts and portfolio managers, since active management is still considered rather an art than a science.
The views expressed are the views of the author, not necessarily those of LSEG.
