Computer Traders Aren't Auto-Programmed to Trade Well

With all the talk of program trading and other computerized forces contributing to last week’s S&P 500 crashes, there is one question we have heard a lot: How can we normal investors even hope to do well in a machine-dominated market? The answer is simple but counterintuitive: by being human, having independent thought and not being forced by programming to overreact to volatility in the blink of an eye.

Those who are paid to overthink volatility on a daily basis universally agree trading algorithms bear at least some blame for big down days on Monday and Thursday. Specifically, they fingered program trading—algorithms forcing big blocks of trades once indexes hit certain levels. As Zachary Karabell explained it at The Washington Post last week:

JPMorgan recently estimated that only 10 percent of stocks are now traded daily by individuals making active choices about what to buy or sell. That figure may be low, but we do know that a large portion of all stocks are owned in relatively new investment vehicles called Exchange-Traded Funds (ETFs) that represent a passive basket of stocks, and that a number of hedge funds are comprised of tens of billions of dollars investing not just in those ETFs, but in derivatives (instruments such as options or futures) that are structured to generate double or triple the returns of those ETFs, or the inverse. Most days, all those baskets and derivatives cancel each other out. But they sit there waiting for days when activity accelerates, pegged to programs that buy and sell based on how quickly prices are moving. The result is a market that is often placid but can take a slight wind and turn it into a selling vortex and then reverse direction and become a buying hurricane.

Here’s how most people interpret the above: The machines are driving the market, and the humans don’t get a chance to trade until the machines have made all the money. But we have a different take: Whether or not machines are in the driver’s seat, they aren’t allowed to think about things like whether their designated selling point will save them from a deeper decline or sideline them ahead of a recovery. They aren’t allowed to think at all! They have to react to what just happened in markets, with no thought of the future. They are not the winners here. They are the suckers.

Think about it. Last Monday, the S&P 500 dropped 4.1% for the day.^[i] Algorithms mandating trades when daily volatility hit -4.0% would have kicked in near day’s end, locking in that loss. The next day, the index rebounded a bit, closing up 1.7% after a wild day.^[ii] But unless the “buy” algorithms triggered at the right time, the machines would have missed a good chunk of it. It is all too easy to envision a bot chasing its little robot tail last week: Selling late on Monday, buying late on Tuesday—just in time for two days of red ink—and selling on Wednesday or, worse, late Thursday, the day the S&P 500 lost -3.8%.^[iii] And the day before it rose 1.5%.^[iv]

Who is the real winner in this scenario? The machine that executed trades in microseconds? Or the human who sat on their hands and waited it out, remembering markets usually bounce quickly from sharp selloffs during bull markets? Yes, the hand-sitter would still be in the red, but they at least captured the good along with the bad and didn’t cause heaps of friction in their portfolio.

2010’s infamous Flash Crash, when US stocks plunged nearly -10% in about half an hour, is another striking example. At the time, people were frustrated that the machines got out and the humans didn’t.^[v] But unless you had to raise cash for a home purchase or what have you on that very day, even the biggest intraday drop didn’t matter. When your time horizon isn’t two minutes long, you can take your time and think critically. The machines don’t have that luxury.

We like the way a gentleman named Peter Tuchman explained it in a Washington Post interview on Friday. If you don’t know Tuchman’s name, you probably know his face: He is the NYSE trader with the wild silver hair, the one photographers always seek out because his facial expressions tell the day’s story perfectly. Here is what he said about last week’s wild swings:

On Tuesday, we had a flash crash that lasted eight minutes. The market went from down 500 points to down 1,500 and rallied back 900 points in four minutes. …

The best thing to do is step away for a moment, watch it happen and then get back involved. It’s like hitting black ice. You can drive on a sunny day. Then you drive in 65 inches of snow, you know the potential for slipping, sliding, skidding. But once you hit black ice, you start sliding sideways and you have no control. Nothing you can do to stop it. That’s a flash crash. It’s electronically catalyzed. It has a mind of its own and usually recovers.

To complete his analogy (and sorry if it doesn’t work, as we are not analogy people), when you start to spin out, the secret isn’t to try to steer yourself out of it. You will end up fishtailing and potentially crashing. Program trading amounts to robots trying to steer out of a skid. And hopefully that is the last analogy we will ever attempt.

Trading algorithms are slaves to past market movement and their programmers’ biases. Humans have the luxuries of free will and patience. Learning from the robots’ foibles and seeing firsthand the futility of immediately reacting to volatility is how we win. So the next time stocks have a big, sudden down day, spare a thought for the poor bots, step away from the computer, and take a walk.