Forget AI. The Real Risk Is the Dumbing Down of Markets.
(Bloomberg Opinion) — Many on Wall Street worry about the impact of injecting artificial intelligence into financial markets. Perhaps we should pay equal attention to subtracting human intelligence. The long-term trend toward indexing —the dumb strategy of buying everything at whatever price the current owners ask just because it’s included in some benchmark — has greatly benefitted investors by reducing asset management costs and protecting from human behavioral biases to chase hot stocks and ideas and from panicking over short-term reverses.
As fewer ordinary investors and traditional money managers paid attention to security valuations over the years, the task of keeping market prices in line with economic reality has fallen to highly levered hedge funds using sophisticated and aggressive strategies. Although these funds represent only a small fraction of total assets, leverage and aggression gives them outsized influence on prices, trading and liquidity, especially with increasing amounts of assets in passive strategies, which remain on the sidelines for pricing purposes.
But today some hedge fund money is moving into dumb versions of these strategies. One of the fastest growing products in investment management is quantitative investment strategies, or QIS for short, implemented by securities dealers as swaps or structured products. QIS volume rose to $370 billion by mid-2022, according to Bloomberg News, citing an estimate by consulting firm Albourne Partners. This is large enough to have significant effects on markets and prices. Moreover, there are indications that the business will continue to grow rapidly.
QIS simply means that investment decisions are made by quantitative rules rather than human judgment. A simple example is a fund that buys stocks if their average price over the last 50 trading days is higher than their average price over the last 200 trading days, and shorts stocks whose 50-day average price is lower than their 200-day average price (this is a popular version of a “moving average” strategy, which is one type of momentum strategy).
Many hedge funds use QIS, although generally in far more complicated manners than the example above. Moreover, most hedge funds are constantly reviewing and adjusting strategies, and typically combine many QIS in a single fund. Mutual and exchange-traded funds are available that use QIS of intermediate complexity and adjustment. Another way for investors to access QIS is in structured products or swaps with dealers such as Goldman Sachs Group Inc. or JPMorgan Chase & Co.
QIS structured products and swaps have been around for well over a decade, but the early versions got a bad reputation among investors. It’s easy for dealers to write contracts that favor the firm rather than the investor. A fund manager has a fiduciary duty to act in the best interests of the investor, and while that duty has not always been honored in full by all managers, it’s still a powerful protection. When investors sign contracts with dealers, the dealer is legally entitled to think only of its own interest. Of course, most dealers care about reputation and repeat business, which sometimes limits the amount of customer abuse, but it’s still the case the investors are more wary of a dealer than a fund manager.
Around 2020 or so, the QIS structured product and swap business cleaned up quite a bit, and it has lost much of its bad reputation. Investors got smarter about contracts, and dealers saw more advantage in customer satisfaction than maximizing profit on each transaction. The more satisfied customers were, the larger the transaction volumes, and the more interest dealers had in offering good products at fair prices.
But the dealer versions of QIS are dumber than the hedge fund versions. That has pros and cons for investors, but mainly dangers for market stability and efficiency. On the good side for investors, performance will not be eroded by managers without the discipline to stick to strategies. On the bad side, no one is monitoring or adjusting when necessary.
Another two-sided point is that it costs something to offer a fund. A manager must underwrite research and sales efforts to get initial investors, and generally has to seed the investment and pay for set-up costs. Most funds run at a loss for years until they compile a track record long and good enough to attract enough subscriptions to break even. While those costs are low enough that there are lots of “me too” funds and funds chasing the latest hot idea, there is zero investment for a dealer to offer a new QIS product. Therefore, every “me too” and hot idea – and not-so-hot idea for that matter – is being pitched by at least some dealers.
Switching the perspective from investors who might use QIS to the market as a whole, removing assets from the smart, sophisticated hedge fund managers who dominate the price-setting process is a clear danger. If losses or increasing leverage costs sideline the human managers, while the dumb automated strategies churn along without oversight, no one will be flying the plane.
Some people dream of replacing smart humans with even smarter artificial intelligence that lacks human behavioral biases and sometimes perverse incentives. That may or may not be a good idea. But replacing smart humans with dumb computers clearly has more downside than upside for markets. If QIS, like index funds, are net positives for investors we need to accommodate them — but we must also be alert to the danger of too much dumbing down of markets.