BATS IPO: Accidental Death or Murder?
04/06/12 - 06:00 AM EDT
NEW YORK (
) -- In the board game
, death could come from Colonel Mustard with a candlestick in the parlor.
On Wall Street, the culprit could be a competitor with a $2,000 sell order over a computer network.
The failure of
BATS Global Markets'
initial public offering two weeks ago has brought out conspiracy theories among a community of Wall Street "algos," or algorithm-based traders, who use computer science and small, rapid-fire trading to make millions of dollars.
The theory, first cited on the industry blog Zero Hedge and later explored by industry publication Advance Trading, says small groups of buy-and-sell orders were sent over the vast network that connects U.S. stock exchanges with the intent to push prices down to harrowing levels as BATS tried to tap the public markets for the first time.
The deadly order was intentionally programmed to send out low-ball offers that would hit a so-called automated quoter running on Nasdaq to nudge the IPO into a death spiral, according to theorists.
"It was an algorithm from somebody connected through Nasdaq and the intent was to bring the price down to zero," says Eric Scott Hunsader, a software developer and president of market data-feed firm Nanex. "And it wasn't a stupid algorithm. It was something that was intent on working its magic."
Trading glitches at the high-speed electronic exchange on March 23 forced BATS to withdraw its much-anticipated IPO. The same glitch also caused a brief, but dramatic, drop in
According to published reports and BATS' own timeline, the company priced an initial public offering of 6.3 million shares at $16 on Thursday, March 22, with shares set to open on the company's own BZX Exchange the following day under the ticker "BATS." Nasdaq would also quote and trade prices on BATS shares.
On the day of the IPO, BATS shares began trading at 10:30 a.m. But at 10:45 a.m., BATS reported that it was having trouble with its IPO auction system. By 12:45 p.m., BATS had pulled its IPO.
The IPO's collapse was a major setback for Lenexa, Kansas- based BATS, which was looking to raise about $100 million from the offering. "Although our affected market has reopened, in the wake of today's technical issues, which affected the trading of certain stocks, including that of BATS, we believe withdrawing the IPO is the appropriate action to take for our company and our shareholders," BATS Chief Executive Joe Ratterman said in a statement the day of the collapse.
A spokesman for BATS and Nasdaq declined to comment for this story.
Hunsader argues that the nature of the trades hitting the BATS shares are evidence that the exchange was the target of a Wall Street dirty trick from a computer algorithm.
"If you look at
, it was perfectly logarithmic, so the bids must have been generated by an automated quoter. This algorithm went to work hitting those bids more than 500 times -- all the way to 0.0002 in a perfect mathematical relationship in the space of 900 milliseconds," Hunsader says. "Plus, It wasn't reacting to anything like rival bids. It was the only thing executing at the time."
Key to the BATS conspiracy theory is the collapse in its IPO price -- from $16 down to pennies -- during the first hours of trading on the shares.
Traditionally, the Wall Street banks running the IPO determine the price and traders set up so-called electronic "layer quotes," which are orders that will execute within an expected range to gobble up the shares. If the BATS IPO was priced at $16, an algorithm would be programmed to buy shares at different layers "away" from that price.
For example, an algorithm would execute an order to buy 100,000 shares of BATS if its price hits $13, and then 50,000 if it becomes more expensive at $15.
But there would be no logical reason for a buyer to send a bid at the rock-bottom level seen in the BATS IPO, says Howard Tai, a senior analyst with market consulting firm AITE Group.
"I am going to take the high road and say that anybody in his right mind -- even a programmer -- would not have written logic like that. And if they did, that programmer should lose his job," Tai says. "You have to use everyday common sense and market experience in writing code."
Tai declines to say whether he believes the BATS IPO was intentionally brought down by a bad algo, and says there are less conspiratorial reasons for the price to fall so quickly. Tai explains a coding error which creates a "loop effect" that would match a buyer and seller quickly could force prices to the bottom.
But in order for those scenarios to work, Tai explains, there would need to be someone on the same side of the trade with the same intent of hitting any price. "The other seller would have had the same glitch that says 'size or price doesn't matter, make it an immediate or cancel order and then fire!'" Tai says.
Hunsader also argues that the nature of the trades reveal that it wasn't a rogue retail trader, but a firm tied to the market's framework.
"They were intermarket sweep orders hitting the bids, so that narrows it down to an audience of professionals" Hunsader says. "You have to be a broker/dealer or somebody connected to the exchange in order to tag the order ISO. This algorithm was easy to spot, because the other exchanges, for whatever reason, were not posting quotes."
Hunsader declines to name any possible trigger men for the BATS hit. "All that is for certain is that it was someone executing at Nasdaq," he says.
And there are culprits that may never be caught. There is no way for regulators or Wall Street to prove there was a trade sent out to intentionally bring down the BATS IPO, says AITE's Tai.
"These machines are flashing out orders in rapid-fire fashion in tiny 100-share orders. So it would take a very fast computer to catch it. And if you try and look at it with your own eye on the screen, you will never see it. They will be gone before you even blink. Regulators don't yet have the experience, the resources or technology to follow this."