Rise of the Stock Picking Machines: A Prelude to a Crash?
These two news items that occurred in March may seem unrelated, but they actually share a common thread.
An Uber self-driving car crashed in Arizona and reports surfaced that BlackRock (NYSE: BLK) had dismissed some if its stock-picking fund managers in favor of quantitative investment strategies.
What ties these two events together is that whether you’re talking cars or stock markets, there is a clear trend for delegating significant oversight to machines when these technologies in many cases are far from proven. The consequences could be deadly to people and their pocketbooks.
BlackRock chief executive Larry Fink called his move a “pivot” away from areas of active management that have fallen out of favor. To be fair to BlackRock, they are one of many fund managers that are adopting quantitative and computer programs to improve returns.
But there is a long history of Wall Street managers embracing technologies and quantitative strategies where they don’t fully understand the risks.
Just a few years ago, there was the 2010 Flash Crash as a result of algorithms and automated programs that manipulated the market. During the 2008 global financial crises, regulators were not prepared for the role of derivatives and other quant strategies in worsening the crises. Warren Buffett has called derivatives “financial weapons of mass destruction.”
You can even go way back to the first big quantitative blow up in the late 1990s, when the hedge fund Long-Term Capital Management (LTC) lost billions on failed absolute-return strategies.
The LTC debacle was a failure so big that 16 financial institutions had to bail it out to avert a market collapse, which turned out to be a dress rehearsal for the 2008 financial collapse when mortgage bets soured.
Investors should be reminded that August 2007 was known on Wall Street as the “quant meltdown,” as these funds had to sell in a falling market.
We had the opportunity to talk to the CEO of a new financial tech start-up that aims to use computer algorithms to purely make selections for income investors. We found this path somewhat startling, given that income investors are the most risk-averse and the most in need of protection.
We argued to the CEO that many algorithms and computer programs in the last few years completely missed the development of negative rates on bonds, the early 2016 market collapse and Brexit, and caused significant losses, even as they did predict the mid-2016 energy rally. But we were assured that that the income algorithm was “stable.”
What we think is happening here is that like the run-up to the 2008 global financial crises, given the seemingly never-ending increases in the stock market propelled by stimulus (monetary and now potentially fiscal), investors are getting complacent over the very real risks that exist in the market.
The Case for Active Management
It may sound self-serving that human investment analysts would make a case for active management in the face of emerging automation. But the evidence clearly favors human oversight, at least today.
As some technologists will tell you, not all industries will be easily automated. The body of evidence of the last three decades shows why investing at this point in history has a level of complexity and unpredictability that cannot be easily quantified or computerized.
Some of the people arguing for automation don’t understand or don’t want to acknowledge the risks. Like Uber with its robot car, they are in favor of driving first and asking questions later.
However, the only thing that has made markets predictable in the last few years has been central bank stimulus, which has been a gift to passive index fund managers that are no doubt on perpetual vacations.
But as monetary policies normalize around the world, the markets will become more volatile and difficult to predict. The answer is that combining human and computer oversight in active management is the best of all worlds for investors.
If chess is any guide to the future, players have moved on from trying to beat computer chess players to playing in teams with chess programs.
What they found has been that the combination of humans and computers can make them more unbeatable as “certain human skills were still unmatched by machines when it came to chess and using those skills cleverly and cooperatively could make a team unbeatable,” according to a recent BBC story on computer-human chess matches. “Humans playing alongside machines are thought of as the strongest chess-playing entities possible.”
Meanwhile, as BlackRock automates, there is evidence that human active managers are beginning to beat the market now that monetary policies around the world are tightening and investors can’t just sit back and buy and hold indexes.
Roughly 45% of all U.S.-based actively managed stock, bond and other mutual funds were outperforming their benchmark indexes in February, according to the Wall Street Journal, citing Morningstar data.
So, chalk up a win for the active human investment managers and be careful which robot you trust, as the battle between humans and machines continues to rage.