The application of artificial technology to investment management suffers from unmet expectations, at least at this time. We think that some voices in the industry sold investors a resort weekend, only to deliver a highway hotel experience. Capital sources consequently have become more circumspect. Is there still momentum to be found in the artificial-intelligence story? Or are investors leaving it behind?
Hedge-fund managers who use artificial-intelligence approaches have not delivered the sort of performance that early-stage investors expected. The Eurekahedge AI Hedge Fund Index only beat the S&P 500 Index in two of eight years since 2011. This year, the artificial-intelligence benchmark underperformed in the first quarter.
The frustration felt by both managers and investors is worth exploring. If anything, enthusiasm for AI approaches to investment management were mitigated by the reality of challenging markets. Broader participation in proven AI strategies will strengthen the field over time.
Why Did AI-Based Strategies Have Such Allure?
One of the early casualties of the consolidation period that we are seeing was San Francisco-based Sentient Investment Management. As a privately-held firm, there are limits on available information, but we know that they managed close to $100 million. By industry standards, that volume of capital is disproportionately large for a new-to-market hedge fund.
We see three reasons for the wave of capital that was deployed early in the life cycle of many artificial intelligence-driven strategies:
► Managers Sold Access to “Bright, Shiny Toys”
Cloud computing and big data opened new asset-management vistas. Suddenly algorithmic traders could compress thousands of machine-learning computations into a narrow band of time. The potential to exploit these breakthroughs seemed endless.
► Managers Exploited the Potential of Their “Secret Sauce”
Heresies abound. Faster, smarter, deeper data sets have a strong correlation with the ability of a manager to generate outsized returns, right? And all the better to prove that point using a model portfolio that incorporates track-record bias.
► Managers Were Misled by the Quality of Alternative Data
New data sets, such as mobile-based location data, have not been as forthright as data scientists expected. Time series can be wildly inconsistent or randomly skewed. There is the potential for regulatory standards to constrain recurring access to this data.
Investment managers of course have a commercial interest to learn from their missteps. We are now seeing a healthy pragmatism injected into the AI hedge-fund business, setting the stage for renewed industry growth over the cycle ahead.
Are Investors Suspect of Artificial Intelligence?
As with any unproven asset class, potential investors in AI hedge funds carry unique behavioral biases. We acknowledge these traits so that we are effective in our dialogue.
► Investors Are Uncomfortable with Black-Box Approaches
Qualified names may look and sound intrigued when you speak to them in a meeting. But most institutional investors are weary of products that are difficult to understand and have an unproven track record. We see parallels here with selected quantitative strategies.
► Investors Are Lulled by Proven Asset Strategies
Past performance may not indicate future results, but it matters in the capital-raising business. At a time when some AI-based managers were selling aspiration, selected mainstream strategies were delivering tangible results.
► Investors Soured on New-Era Investment Ideas
AI-based trading strategies are an altogether different asset class than cryptocurrencies or initial coin offerings, but they may be discolored by association. The high degree of volatility we saw in other new-era assets classes, including outright fraud in the case of initial coin offerings, may have kept some investors at the sidelines.
Is There a Future for AI Hedge Funds?
Investors will ultimately gyrate to those AI hedge funds that hit the mark on risk-return criteria, but that may be a narrow group within the alternatives universe. The days of institutional investors being intrigued by the idiosyncrasies of structured and unstructured data are over. Successful players will reference the strength of their Sharpe ratio, aligning their work with known performance metrics.
More broadly, a sign of a maturing industry is that AI hedge funds are looking beyond their initial investor base in Silicon Valley. Managers are competing in the open market for global capital; they are pitching skeptics at sovereign wealth funds and endowments worldwide. Importantly, those players who carry their own against other asset classes echo a common theme in providing thorough due-diligence transparency. ■
© 2019 Plotinus Asset Management LLC. All rights reserved.
Unauthorized use and/or duplication of any material on this site without written permission is prohibited.
Image Credit: Peshkova at Can Stock Photo Inc.