OUT OF ORDER:
Are Controls Needed on Choice, Complexity and Competition?
January 10, 2013
Over the past two years, BATS Global Markets conducted a study of how many “unique order type combinations” can be used to set buy or sell instructions on its two national exchanges in the United States.
There are at least 2,000, according to the seven-and-a-half-year-old challenger to the New York Stock Exchange. There are likely similar amounts of permutations now on the Nasdaq Stock Market, the four-decade-old pioneer of the all-electronic market, and exchanges operated by Direct Edge, which started out 14 years ago as an “electronic communications network.”
Orders that allow investors to publicly display blocks of shares for sale, but keep in “reserve” a portion of the order, hidden from view.
Orders that stay hidden, altogether, until the value of a stock that has been frozen or “locked” by identical bid and offer prices—which is prohibited—has been resolved.
Orders that “sweep” all markets for the best prices.
Orders that look for prices only in venues that have costs of transaction or offer rebates.
Orders that expire after a given length of time.
Orders that are set to execute at one tick inside the best bid available anywhere.
And so on and on and on.
Can there be too many order types? Maybe, maybe not.
But the number of order types and the almost limitless permutations in the way they can be used to buy or sell shares add undue complexity to trading, according to a wide swath of buyside and sellside traders. Their sheer volume and the multiplicity of ways they can be used across not just one, but every, national exchange makes keeping up with how markets work a serious challenge. And it makes many institutional traders suspect that, somehow, they’re being taken advantage of by high-speed trading and market-making firms that propose the greatest number of new ways to get to the front of the line in markets, based on displaying the best price before anyone else does.
“There’s an order type a minute being created,” said Jennifer Setzenfand, outgoing chairman of the Security Traders Association, which represents 4,200 individuals involved in trading equities and equity options.
“There are perfectly good order types that add liquidity to our markets,” said Setzenfand. “And there are predatory order types that some may argue also add liquidity, but get in the way of institutional orders.”
The most straightforward problem is simply keeping up. “It’s hard for me to understand the need for 2,000 order types,” said John M. Donahue, senior vice president and head of equity at Fidelity Capital Markets, a unit of the mutual fund giant.
Indeed, the head of one of the industry’s largest electronic brokerages, who declined to be quoted on the record because of the size of his business, says it is extremely difficult to get a handle on how each order type works, even if you try to keep up with all the filings made on the subject by the exchanges.
But the exchanges say they are simply responding to customer demand: There has to be a base of interest in each mechanism for placing a buy or sell order, or it wouldn’t be introduced.
“This allows people to trade in different ways in different time frames with different objectives,” said BATS chief operating officer Chris Isaacson.
The proliferation comes as market participants already struggle to keep up with a modern reality where orders are sent not just across 13 national exchanges but also into scores of “dark pools” that operate out of public view, and another hundred more internal pools of orders managed by brokers.
And the competition to keep orders on exchanges, instead of being handled by alternative trading systems and internally by brokers, is intense. Overall trading volume on the nation’s 10 largest exchanges dropped 28 percent in the first 10 months of 2012 and is running at about 2007 levels. That means approval from the Securities and Exchange Commission will be sought for almost any order type that likely will appeal to a broad segment of trading firms, brokers or investors.
CUSTOMIZATION AND COMPETITION
The creation of customized orders is to keep the major players in the industry—those generating the most trading volume—coming to play in an exchange’s given arena.
One example of a new highly specialized order type is the NBBO Offset Pegged Order, from Direct Edge.
That allows a market maker, in particular, to automatically redefine the price of an order by its distance from the national best bid or offer.
Creating this sort of new order type helps Direct Edge differentiate itself. But it typically also occurs in response to a need or desire expressed by an exchange’s membership. Those memberships now include high-speed market makers such as GETCO, Tradebot and Knight Capital, as well as brokers representing institutional investors and handling orders for retail investors.
“A natural thing that starts occurring is, you start talking to your clients, and you start saying to them, ‘How can I help you?’” said David Polen, head of business development for Fidessa Group, which supplies trading systems and market data services to both buyside and sellside customers. “And they say, ‘Well, if you tweak the infield fly rule, it would help my team.’ You say, ‘Oh, OK, you know what? I’ll tweak the infield fly rule. I want you to play in my stadium.’”
But the process doesn’t stop with one tweak of the infield fly rule. If you have eight different clients, you can end up with eight different versions of the rules, since, Polen said, “each one of them has a slightly different idea of what they want.
“And you know what?” he said. “It’s easy for you to do,” using the flexible trading systems now in use.
In this case, Direct Edge made it clear this order type was aimed at one type of customer: market makers.
Nothing necessarily nefarious there. This, the exchange operator said, is a one-sided order.
“As a result, a member acting as a market maker would need to submit and maintain continuously both a bid and an offer using the order type,” it said. That is to comply with the exchange’s quotation requirements for market makers.
And the operator even added branding to the lure of the order type. Market makers who place “attributed quotes” with this order type could identify themselves, noted Bryan Christian, head of sales at Direct Edge. That means a market maker can use attribution as a tool for advertising that a particular order is its quote.
This helps members “make more educated trading decisions,” the exchange says.
The New York Stock Exchange makes the explicit point that none of its order types are aimed at a specific type of trading firm and all are marketed to its membership as a whole. Nasdaq takes much the same stance, and BATS says it won’t pursue any order type if there isn’t wide interest in it.
“We take ideas for functionality from the broad base of our membership,” Isaacson told Traders Magazine, “and we filter those through our own view of the market and what is fair and orderly. Clearly, if an order type comes through purely for the self-interest of (one) customer, we’re not going to allow it.”
In 2012, BATS introduced a pricing incentive for orders, called NBBO Setter, that was aimed at market makers.
But in this case, it says, the beneficiaries were all investors.
BATS gives “a slightly higher rebate to someone who’s making markets on BATS, and they set the NBBO,” Isaacson said. This encourages market makers to (1) improve the price that’s available to the entire public, and (2) do it on a BATS exchange.
The self-interest of exchanges can come into play. The Nasdaq Stock Market, for instance, has a “price to comply” order type that makes sure any order complies with the SEC’s National Market System rules. But the order type is designed to keep the order on Nasdaq and not end up elsewhere.
That can seem at odds with the rules. “Shouldn’t it really just route to the best bid and offer that is out there? Shouldn’t that be the default?” said the head of trading at a major institutional trading firm, with nearly a trillion dollars under management.
“If we’re going to have a national market system, shouldn’t all these venues be routing to the best bid and offer for the customer’s execution? Why are they pricing it to keep it at that venue?” said the trader, who insisted on anonymity, to keep trading on all venues with all participants.
Nasdaq also has a Supplemental Order type that “only interacts with small orders to avoid interacting with the price-clearing orders,” the trader contended.
“What the hell is that about? If I go in there with an order to take out the offer, why aren’t I getting everything that is actually offered there? That doesn’t do anybody any good.”
That order type, by Nasdaq’s wording, “provides a final chance for routable orders to execute on Nasdaq, before being routed away” and for suppliers of liquidity “to interact with attractive order flow.” The orders “execute only against orders smaller than the size of the supplemental order interest.”
Much of the problem of proliferation centers around the fact that, even if the process of getting new order types is “onerous,” as Isaacson put it, the number of order types keeps rising inexorably. Figuring out how each works defies the ability of the human brain to absorb, evaluate and adapt.
That’s why the STA’s Setzenfand says, “We would like to see a better notification process for order type filings and inclusion of examples of how order types impact time and price priority.”
To that end, Nasdaq OMX Group has tried to turn the proliferation of order types from an inexplicable morass into a learning opportunity for its members.
And a potential competitive advantage.
While BATS, the NYSE and Direct Edge kept introducing new order types, Nasdaq OMX in November started to explain how each of its 10 main ways of putting in buy and sell orders worked.
The exchange operator began circulating to its members an interactive rundown on how its order types work, assuring them that they can’t be used by any market participant to jump ahead in the queue for striking a transaction.
The combination of audio descriptions with summaries and examples on slides, in synchronized chapters, was sent by email to head traders, technology contacts, compliance officers and executives of member firms. Traders Magazine obtained a copy.
The automated presentation was an attempt to educate members and investors on the workings of order types and how public bids and offers interact with hidden orders not shown on lit or “displayed” markets.
The presentation itself, narrated by Michael Blaugrund, a vice president of Nasdaq OMX transaction services in the United States, made no recommendations on how to best use order types.
But he and Nasdaq specifically note that “there are no order types that offer queue spot privilege or queue jumping,” a reference to concerns that some “hidden” orders can jump to the head of the line, when a trigger turns them into publicly displayed quotes.
That exemplifies the pressure being felt by exchanges to show they are not playing favorites, when they seek rule changes when filing for approval of new order types or functions.
“We’ve seen in the past few weeks an increased interest among our member firms, but really also in popular culture, if you will, in exchange operation,” Blaugrund said in an interview with Traders Magazine. “And I think it’s our view that the more transparency around how things operate, the better.”
The interactive breakout of its order types, now widely disseminated via email and on Nasdaq OMX’s Web sites, means institutional investors and retail investors can see “there are no secret order types, every member has access to every order type, and that every order is going to interact based on price-time priority,” Blaugrund said. “There are no orders that give anyone some sort of special advantage.”
SLANTING THE PLAYING FIELD
But brokers and institutional traders suspect—or fear—that many of the new order types are set up largely to slant the playing field toward the trading firms with the shortest time horizons.
Enter a permutation known as the post-only order. “The ultra-high-frequency trading firms are doing everything in their power to come up with enhanced ways for them to (engage in) post-only order flow,” said Sal Arnuk, co-founder of Themis Trading in Chatham, N.J., and a critic of electronic mechanisms that allow such “predators” to get their orders to the front of a queue of orders in exchanges that are almost exclusively based on giving priority to the first order to show up at a given price.
The reason: rebates. Use of a “post-only” order allows a trader to “post” an order but “only” have it come into play if there is “adequate price improvement,” in BATS’s terms. That means, in turn, it’s able to earn a rebate for adding liquidity to the market, because it does not automatically take out a displayed order that might be resting on an electronic order book.
The post-only order is not particularly innovative, said Jamie Selway, head of liquidity management at agency broker Investment Technology Group. Plus, the focus on rebates “is not a good thing always. If you’re a broker and you’re concerned more about the fee you pay than best execution, that’s sort of a conflict of interest.”
As of Nov. 1, BATS charged 29 hundredths of a penny per share for removing liquidity from its primary exchange, known as the BZX market. Conversely, a firm got somewhere between 25 hundredths and 29 hundredths of a penny for adding liquidity.
That’s “only” 29 cents on each 100 shares being traded. But if you “add” 10 million shares of liquidity, it adds up. At $29,000 a day, that would be roughly $7.25 million a year.
But the type of order that has drawn the most scrutiny in the past year is something industry consultant and former UBS managing director Haim Bodek calls a “hide and light” order.
This is an order that is allowed to “hide” out of sight on an order book when prices are locked or crossed. Such an order can “transform from a hidden order into a Protected Quotation at precisely the time a market (is) permitted to display an aggressive price,” Bodek argued this fall in a series of articles on the Tabb Forum, an online discussion site.
The most well-known version of this came to be the Hide Not Slide order, from Direct Edge. This was used in a Wall Street Journal article featuring Bodek that attempted to show how such an order could allow a hidden order placed by a high-speed trading firm to jump ahead in line of an institutional rival for the same order.
Such jumping allegedly occurs when the price of a given company’s shares “lock” at an identical price on two different markets. The order at Direct Edge hides until the price slides. Then it becomes lit, at the original higher price. And the investor whose price slid, even a penny, gets a lower priority in getting what otherwise would have been the same order filled.
The charge that the hidden order gets to unfairly jump to the front of the line “has been at times misinformed,” said William O’Brien, chief executive of Direct Edge.
“You don’t want to reject a customer’s order unless you absolutely have to,” he told Trader Magazine.
The prohibition on locking the market at identical bid and offers on different exchanges or crossing markets, where the bid gets higher than the offer, means an exchange can’t take process orders under those conditions.
“So when you get one that would” lock or cross a market, he said, “you have to decide what to do with it. Now, rejecting it is something you can do, and that’s one option that we make available. But we try to give our customers other options,” besides having it rejected.
But jumping to the head of the queue is not one of those options, he said. “It’s a beautiful theory slayed by ugly facts,” he said.
The first order has to be canceled. And the second order has to be at a higher price.
Then, the normal priorities of first to market with the highest price take effect.
“I mean, look, an exchange, by definition, has queue jumping all the time,” he said. “I want to buy at 20, you want to buy at 20.01—you jumped the queue. Right? A Hide Not Slide order is simply someone willing to pay a better price. And so in certain circumstances, they have priority.”
Even so, permutations that keep orders hidden until they suddenly light up can alter the basic purpose of markets to maintain a continuous market in stocks, said Selway. “The notion of price-time priority gets a little squirrelly,” he maintained. “Do we really have a continuous price-time market when we have a large amount of hidden orders piled up, and then the price changes, boom, then we go?”
Over the long haul, the question boils down to this: Should there be more order types or fewer?
“If everyone wants to stop competing, we can,” said Isaacson, on the idea of putting a cap on the number of ways that orders can be placed. “But that doesn’t seem like it’s in the best interest of our markets.”
Should there be a moratorium on order types?
Isaacson thinks not. “I think the train has left the station,” he told the Security Traders Association.
The nation’s oldest exchange even avers that it’s false to lay the proliferation of order types at the root of the national exchanges’ operating problems.
“To kind of pick order types and say that’s the problem, we should fix that, is sort of a microscopic view of an ecosystem that is far more complex than any” single problem, said Joseph Mecane, executive vice president and co-head of the U.S. Listing and Cash Execution business of NYSE Euronext.
Even so, the SEC has never denied an order type submitted by an exchange, even if its review process can be “onerous,” as Isaacson puts it.
The proliferation, Setzenfand contends, means that there should be some sort of centralized review body with explicit review criteria that determines whether a particular addition to the thousands of existing permutations gets approved.
“No rule should disadvantage a particular business model or give advantage to a particular business model,” she said. “It needs to be a level playing field when it comes to creating order types.”
The problem may resolve itself. “I don’t think there are too many order types,” said Ian Winer, director of equities trading at brokerage Wedbush Securities. “A lot of the dynamic, fast-type trading that was taking place a couple years ago has sort of been arbitraged out of the market by the law of diminishing returns as more players entered and drove down profits.”
Profitability of high-frequency trading is plummeting, noted industry consultant Larry Tabb of the Tabb Group. In 2008, profits were $8 billion a year. For 2012, they were $1.8 billion.
And the volumes reflect this, Winer said. The volatility of markets in 2008 as the credit crisis erupted were ripe for exploitation by high-speed firms trading on small and large fluctuations in prices. Average volume on the top exchanges? It was 6.6 billion shares, according to the Securities Industry and Financial Markets Association. In 2012? Just 4.4 billion.
Perhaps so, but brokerages of all sizes and equities trading teams of all kinds at mutual funds and other institutional investing firms have it on their accounts to know and keep up with all the order types that do get approved.
The Nasdaq Stock Market, for one, provides what Blaugrund calls “a full test facility” where all existing or new order types can be tried out and checked, one weekend a month.
Tested or untested, “the ownership is on each trader and each institution to understand what is out there,” said Setzenfand.
And deal with the thousands of order types that are out there. And still arriving.