With nine U.S. options exchanges competing for order flow and liquidity, technology and the race to zero latency have become the battleground for supremacy.
As volatility spiked in August and early September, U.S. options exchanges handled record trading volumes -- surpassing 500 million contracts -- while investors and traders turned to equity options, exchange traded funds (ETFs) and options on ETFs to hedge their exposures or take on risk and bet on the volatility. But while the heavy volumes tested their electronic trading platforms and infrastructures, the options venues were well prepared, and a few even increased their market share.
"Everyone's systems performed well, so the customer should have had a decent experience," asserts Tom Wittman, president of Nasdaq OMX's PHLX options exchange, which has held more than a 20 percent market share in single-name options for 13 consecutive months.
Prior to the spike in volatility, several of the big options exchanges upgraded to faster matching engines that run on more powerful servers. In addition, three options exchanges moved their matching engines to third-party data centers in New Jersey to get closer to (and thus reduce the trading latency of) their liquidity providers, which typically make markets on multiple options venues.
In April, the International Securities Exchange began a phased migration to Optimise, a new global trading platform built by transatlantic teams from the ISE and Eurex, the derivatives arm of the ISE's parent, Deutsche Borse. Around the same time, ISE moved its matching engine to two new data centers -- the Equinix NY4 Internet Business Exchange (IBX) facility in Secaucus, N.J., will serve as the exchange's primary data center, and the Telx facility in Clifton, N.J., will be the backup data center. Meanwhile, this spring NYSE Euronext's Amex Options exchange moved its electronic platform into the exchange operator's Mahwah, N.J., data center, and last year the CBOE's C2 platform and the BOX Options Exchange both migrated to Equinix's Secaucus data center.
According to Andy Nybo, head of derivatives research at Tabb Group, options exchanges continually upgrade their networks and matching engines to stay competitive in terms of speed and capacity. "Almost any aspect of the options trading industry has become an arms race in terms of technology," he observes. "Having a robust trading platform is a key part of a strategy for options exchanges, particularly with trading volumes and market data rates and investor interest surging to record levels."
But low latency is critically important, Nybo stresses. "The exchange's ability to attract not only the buy side but market makers to participate in a marketplace depends on fast access, including both the ability to publish quotes and to remove and update quotes," he says. "Market share shifts for a number of reasons -- pricing scheme, market models, as well as technology platforms and the amount of liquidity and order flow in a platform."
Nasdaq Options Market 2.O
To increase its appeal to options market makers and traders and boost the market's liquidity, Nasdaq OMX in September 2010 migrated the Nasdaq Options Market (NOM) -- its secondary U.S. options market, which is aimed at high-frequency traders -- to the same version of the INET platform on which its PHLX options exchange runs. This means that both of Nasdaq OMX's options markets now share the same interface for market data and posting quotes, the same risk protection mechanisms and the same application programming interface (API). "It becomes easier for the institutional firms, the market making firms and the retail firms to understand the systems architecture, and it should cut down on their time and expenses [for connecting to the markets] because they [no longer] have to code to each one," according to Nasdaq's Wittman.
Wittman explains that the original NOM platform essentially used a copy of the Nasdaq cash equity system to trade options; the latest iteration launched on the new architecture, known as NOM 2.0, utilizes the software developed by PHLX to run multiple matching engines in the U.S. options space, which reduces latency by dividing up the number of symbols that runs on each matching engine. "We rewrote the NOM platform on NOM. 2.0 using PHLX XL technology," he relates, noting that PHLX currently runs on four matching engines while NOM features two.
"We like to keep our hardware and networking costs lower, so we don't need to run as many match engines [as some of our competitors]," says Wittman, who indicates that one of NOM's rivals runs 14 or more matching engines. "We use some technology in the matching engines that allows us to run more efficiently and with less boxes."
In addition to the architectural improvements intended to speed the options platform, Wittman reports, Nasdaq will offer bulk quoting on NOM 2.0, allowing market makers to update their options quotes in bulk. "So market makers could send 200 messages at once, as opposed to sending quote updates one at a time," he says. "When they run their quoting models, they want an iteration in two microseconds. ... With bulk quoting, ... it's just more efficient."
Colocating Options Engines
Latency considerations also are a major driver behind the trend of relocating matching engines in third-party data centers. Options markets are taking advantage of space within these data centers to be closer to the options and equity venues whose matching engines are hosted nearby (or in the same facilities), and thus to be closer to market makers, which are colocated in the same data centers.
Most U.S. options exchanges rely on automated market makers (which use algorithms and computerized models to update their bids and offers and submit orders) for liquidity. By hosting its matching engine in a data center where its liquidity providers' servers are colocated, an exchange can cut down the time it takes to transmit market data, enticing the algos to trade more on the exchange.
When the Chicago Board Options Exchange (CBOE) launched the all-electronic C2 Options Exchange in 2010, it located the matching engine in an Equinix data center. In an interview earlier this summer, CBOE CIO Gerald O'Connell told WS&T that he decided to put C2's engine in the same data center with other options and equity markets, including NYSE Arca, to be closer to the firms "that already are there."
The fact that options exchanges depend on the underlying cash equity markets for prices, which originate on the east coast, also was a factor in the CBOE's decision. By locating C2 in New Jersey, CBOE was able to eliminate the 13 milliseconds of latency involved in transmitting the equity feed to its proprietary data center in Chicago, O'Connell explained.
According to Nasdaq's Wittman, both PHLX and NOM are colocated in Nasdaq's data center in Carteret, N.J., alongside Nasdaq's equity markets. PHLX moved into Nasdaq's data center from its previous site in Philadelphia in 2008, one year after Nasdaq completed the acquisition of PHLX. The move "gives our customers who are connected in that facility greater bandwidth and speed," Wittman relates, adding, "Obviously the firms that are colocated get the faster turnaround time since there is less network latency." But Wittman is quick to point out that anyone can take advantage of colocation and that Nasdaq's colocation customers include market makers, banks and retail firms.
NYSE Amex Options: Been There, Done That
While many of the options exchanges are turning to a combination of faster trading platforms and colocation to grab market share, one exchange operator already is enjoying the benefits of those investments. NYSE Euronext's two platforms -- NYSE Arca and NYSE Amex -- accounted for almost 25 percent of the options volume in August 2011. The big gain was in the Amex options business, which raised its share to 14 percent in August, up from 12 percent the prior year and from a little more than 6 percent in 2009, according to Steve Crutchfield, CEO of NYSE Euronext's NYSE Amex Options exchange
According to Crutchfield, there are a number of factors behind the Amex Options exchange's volume growth, starting with Amex's acquisition by NYSE Euronext and the exchange's migration to the Universal Trading Platform (UTP). The move to UTP was intended to create a state-of-the art electronic matching engine, says Crutchfield. "That brought much better performance than existed on the legacy options platform."
More recently, the Amex matching engine was moved into NYSE Euronext's 100,000-square-foot data center in Mahwah, N.J., where it shares space with the NYSE Arca options exchange as well as all of the NYSE cash equity markets. But it wasn't a case of simply moving the equipment, Crutchfield notes; all of the infrastructure was replaced and all of the servers were upgraded, he reports.
One statistic Crutchfield offers to quantify the performance improvements is the time it takes to accept a quote from a liquidity provider. The turnaround in the acknowledgment of quotes is now in the sub-millisecond range for outliers, he offers. "This is what enables our participants to scale up their volume on the exchange," Crutchfield says. "Since liquidity providers are exposed to a large number of resting bids and offers, there is more incentive to trade aggressively on exchanges where that is as brief as possible."
Pouring On the Liquidity Plays
Beyond latency reduction, exchanges also are tinkering with fees and ownership structures to attract liquidity. For example, in June Amex sold a majority stake in its platform to a coalition of broker-dealers and order flow providers. The new ownership structure has given the dealers a greater incentive to boost the volume of business they contribute to the exchange, according to Crutchfield.
Amex also slashed fees on its complex order book for firms that send orders electronically, and as of May 1 stopped charging customers altogether for electronic delivery of complex orders. "It's always well-received when you cut fees to zero," notes Crutchfield.
Ivy is Editor-at-Large for Advanced Trading and Wall Street & Technology. Ivy is responsible for writing in-depth feature articles, daily blogs and news articles with a focus on automated trading in the capital markets. As an industry expert, Ivy has reported on a myriad ... View Full Bio