How Technology (Usually) Prevents Market Crashes
If Thursday’s sudden fall was caused by human error, it should’ve been caught by safety systems.
While there has yet to be any official confirmation, CNBC reports that Thursday’s precipitous 1000 point drop in the stock market was a product of a trader hitting “b” for billion instead of “m” for million when punching in a stock trade.
The Dow Jones Industrial Average fell nearly 1000 points yesterday. |
You would
think that the high-speed electronic trading systems brokers rely on to execute
trades–the NYSE Arca,
which is the world’s second-largest electronic exchange and processes trades in
as few as 600 microseconds–would be idiot-proofed against this sort of
mistake.
As I
discovered in an interview conducted last October with Paul Adcock, Executive
Vice President of Operations at the New York Stock Exchange, trading gateways do
have built-in controls designed to prevent precisely the error that may have
occurred Thursday–but that it’s an optional setting.
This story is only available to subscribers.
Don’t settle for half the story.
Get paywall-free access to technology news for the here and now.
Subscribe now
Already a subscriber?
Sign in
You’ve read all your free stories.
MIT Technology Review provides an
intelligent and independent filter for the
flood of information about technology.
Subscribe now
Already a subscriber?
Sign in
“One
we thing do on all electronic trading systems are […] share checks,” Adcock said. “They’re for those times when someone does a ‘fat-finger’ and sends a million [shares] in, not a
thousand.”
Similar
checks can also be set up to make sure that a trader doesn’t attempt to buy or
sell a stock at a price that is more than 10% smaller or larger than its
current price. “I may instruct Arca to set up risk parameters on the
gateway,” says Adcock. “For example, for orders to be not larger than
1000 shares; We try to keep people from hurting themselves in a market that’s
moving in microseconds.”
Adcock
also explained why a large selloff by a single player could rapidly precipitate
a wave of panic selling by other traders: An aggressive buyer (or seller) can
easily become “the ax” in a stock, pushing its price either up or
down as the volume of buy or sell orders they are sending to an exchange
outstrips the supply of either sellers or buyers who can take the other side of
the trade.
Because
the stock reported to have precipitated yesterday’s sudden fall via a large
sell-off, Procter & Gamble, is traded on open exchanges, every trade on
that stock immediately “goes to tape,” which is the public record of
what’s going on in the exchange at any given moment.
That’s
when computers running automated trading algorithms swoop in, driving a stock
further up or down. “If computers see tapes moving in a certain
direction,” says Adcock, “the momentum players come in and say ‘I
want to get in.’”
Because
speed is a competitive advantage in the world of high speed trading, there is
an arms race going on between hybrid human/electronic trading floors like NYSE
and upstart electronic-only trading floors like Direct Edge. Meanwhile, the complexity of
such trades is only increasing, further enhancing the potential for human
error: the NYSE Arca platform accepts more than 75 different order types,
allowing for elaborate trades triggered over different time scales with a
variety of customizable conditions that must be satisfied before a trade goes
through. For more on this, see Trading Shares in
Milliseconds (subscription required).
Given
these complexities, it isn’t surprising to see the NYSE advertise
the fact that it still employs human traders who can step in and slow down the
trading of a stock whose price seems to be wildly out of line. While the old
system is far from infallible, goes the argument, at least humans have the
power to stop and think before they decide to engage in panic selling.
If the CNBC report is accurate, it’s a story as old as invention itself–of human error and technology
run amok.