Exchange-traded funds were at the center of the ‘flash crash’ last week, but it wasn’t because ETFs are flawed.
A lot of people out there are trying to make that connection. As we reported in our coverage of the crisis, around two-thirds of all securities that had their trades canceled by Nasdaq and NYSE Arca last Thursday were ETFs. It’s easy to confer blame.
The data, however, don’t back that up. We don’t yet know with 100 percent certainty why ETFs played such a central role in the debacle, but we have a number of good theories, and none of them points to a “flaw” in the ETF system.
Still, it’s not a coincidence that ETFs featured so prominently. If there’s somehow a repeat of last Thursday’s craziness—if market reforms don’t materialize and high-frequency traders don’t adjust—ETFs will likely be right in the middle of it all again. They are powerful tools, but the way they function in the market needs to be understood.
The most popular explanation I’ve heard is that ETFs are exposed to mistaken prices in underlying stocks. Computers are constantly monitoring the share price of ETFs and comparing those to the fair value of their underlying components. When prices get out of whack, computers will arbitrage the difference away, selling an ETF and buying its underlying securities, or vice versa.
That’s a wonderful thing. It’s what keeps ETFs trading in line with fair value, and is the key to how ETFs function.
A Few Explanations
But if you get bad prices in a few securities—if, say, a few large-cap stocks drop $20—you get a cascading effect throughout the system. Low stock prices cause ETFs to appear overvalued, causing computers to sell, which in turn drives prices lower.
A related explanation is that ETFs simply are disproportionately impacted by program trading, because they are monitored by so many algorithmic traders on a daily basis. As a result, if something goes wrong with such high-frequency trading, the impact of any errors will be amplified in ETFs.
A third explanation is that many trading systems benchmark ETFs against the S&P 500 “e-mini” futures contract on the CME in Chicago. These systems monitor the futures contract—the most liquid security in the world—as a proxy for market movements, and watch for a divergence between futures and ETFs. One of the rumors of what started the flash crash is that e-mini contracts had a bad price print. If that indeed happened, the trading algorithms might have kicked into high gear and started the computers selling ETFs.
A fourth explanation is that, when the markets get crazy, people turn to ETFs for quick market exposure. In this case, as the market fell apart, investors likely turned to ETFs to gain quick exposure, whether short or and long. After all, if you’re trying to sell right now, you’re not going to bother with whether to sell short GE, GM, AAPL or CSCO. You want to sell the market, and that means ETFs. As a result, you saw more ETFs affected.
A final explanation is that, unlike with individual securities, investors do not set up deep out-of-the-money limit orders in ETFs. A fundamental analyst might decide that Proctor & Gamble is worth $40/share, even if it’s trading at $60/share, and might put in a limit order to buy 100,000 shares of PG when it hits $40. These forgotten limit orders then “catch” stocks when they are caught in a wave of selling.