BATS Glitch Shows Stock Trading Is Too Complex, Not Crooked
Depending what’s being measured, $400,000 can seem like a lot of money or a trivial sum.
For a month’s rent at a beach house – even a really nice beach house – $400 grand is pretty steep. As an amount of money lost over four years in slightly mispriced electronic stock orders by clients of the BATS Global Markets exchange due to apparent programming errors, $400,000 doesn’t even rise to the level of a rounding error.
BATS on Thursday disclosed that, since October 2008, its systems allowed a small percentage of trading orders (an average of 410 per day) to be executed at prices inferior to the regulator-mandated “national best bid or offer,” costing customers $420,360 in total.
The admission is getting more attention and eliciting louder commentary than the magnitude of the breach would seem to merit, mostly because it is easily piled atop previous trading glitches and failures to reinforce the general public sentiment that today’s hyper-fast stock market is somehow rigged against the average investor.
But the proper way to view this revelation is not as further evidence of plugged-in middlemen routinely picking the pockets of unassuming investors. Rather, it’s yet another reminder that the overall market apparatus has gotten far too complex, too subject to the tortured interweaving of rules governing order-flow minutiae and too technologically diffuse to assure long-term system stability or effective regulatory supervision.
The real effect
By a quick estimate, the errant trades at BATS affected 0.004% of all trades over the relevant period, and reportedly affected sophisticated electronic market makers rather than everyday investors.
Justin Schack, a market-structure expert and managing director at Rosenblatt Securities, commented via Twitter that it’s “interesting that customers did not detect or complain about the problem the entire time,” suggesting the amounts weren’t material and the errors not obvious.
Still, he notes, it raises questions about how many undetected issues await discovery, and is sure to raise further questions about the way exchanges pursue high-velocity order flow by offering arcane, highly contingent order types to be programmed into trading engines and set free.
As dozens of venues for equity orders have emerged to compete for active traders’ business, commissions have collapsed and volume capacity exploded. Yet as the perpetual game of “spy versus spy” between highly engineered trading shops has raged, regulators have allowed many new order types geared to create tiny, fleeting eddies in the flow of orders that one small subset of trading firms are geared to exploit.
Fostering instability
This sort of complexity layered across so many venues with insufficient real-time transparency can foster the sort of instability that has erupted occasionally, such as in the “flash crash” of 2010.
Amy Butte, former chief financial officer of NYSE Euronext (NYX), wrote in a recent Bloomberg View op-ed that our markets “no longer work” in part because “today, you need a super computer or a doctorate to understand the rules of the stock market.” She reports that there are more than 100 order types, which can then multiply into further variation based on time-of-day and other specifications.
Some potential solutions seem to be gaining support, including limiting order types. Another possibility is requiring high-volume traders to pay not only for executed trades but for the system capacity they use with their millions of electronically floated, long-shot orders that never are executed and are withdrawn in milliseconds.
The quicker some measures are taken to rationalize the system and set simpler rules, the greater the chances for preserving the considerable benefits to investors of competition and technological investment that arose through deregulation – while winning back some of the lost faith of average investors, who seem not to realize or appreciate how much cheaper and easier trading has become for them.
If done carefully, industry authorities can get us away from the tired story line in which the disclosure of an honest error that added a tiny sliver of frictional costs to professional order books is treated as an example of the little guy being picked off.
