In the past month, NYSE and Nasdaq prevailed in a decade-long litigation defending their ability to charge premium prices for depth-of-book market data. At the same time, IEX is planning to launch without selling depth-of-book market data at all. What’s behind these seemingly opposite models of the same business?
The SIFMA lawsuit—which started out as a challenge to NYSE’s first application to charge fees for depth-of-book data after it acquired Arca in 2006, and evolved into an antitrust case against both NYSE and Nasdaq that was not decided until 2016—revealed some interesting facts about who actually uses depth-of-book data.
NYSE and Nasdaq data sales reps testified that fewer than 200 firms subscribe to depth-of-book data feeds, and only around 60 subscribe to both NYSE’s and Nasdaq’s feeds. These subscribers are the largest players in the equity markets, responsible for upwards of 90% of all market volume on Nasdaq. A Nasdaq rep testified:
“[T]he only customers who absolutely need depth-of-book data from all the exchanges are customers such as algorithmic and high-frequency traders, who are a small proportion of the overall total of customers, but who exert enormous power over the Exchanges because they represent a gigantic percentage of the trading market.”
Despite SIFMA’s expert testimony that the audience for depth-of-book data was broader—“without depth-of-book data from all the exchanges, a trader seeking to execute a particularly large order would not know how to make his routing and trading decisions”—the substantial order flow controlled by the largest depth-of-book customers ultimately lost them the SIFMA case. An unnamed firm retaliated against certain Nasdaq data fee increases by routing their order flow to other venues. The administrative law judge hearing the case found that the undisputedly vibrant competition for order flow spilled into the same parties’ market data fee negotiations, making those competitive too.
SIFMA has not yet announced whether it will appeal this ruling to the DC Circuit, which has already shown some skepticism towards the exchanges’ order-flow-competition-constrains-market-data-prices theory in an earlier ruling in the same case.
Meanwhile, the newest exchange, IEX, is planning not to offer depth-of-book feeds at all, at any price. Only top-of-book quotes on IEX will be published, and those will be available for free.
Why are NYSE and Nasdaq willing to pay a decade’s worth of litigation costs to defend their right to charge a relatively short list of customers for depth-of-book data, while IEX is opting out of that revenue stream completely? Here are a few possibilities:
1. IEX’s position may be temporary. There is a well-worn path of new venues not charging for market data in their early days as they build market share. Island and Archipelago, the first ECNs, gave away depth-of-book data for free until they were acquired by Nasdaq and NYSE. BATS did not charge for depth-of-book data until July 2013. (Days after the SIFMA decision was released last month, BATS, which was not a party to that case, announced plans to raise its depth-of-book data subscription fees.) Mature exchanges, particularly publicly traded ones, are more interested in building stable recurring revenue streams that insulate their earnings from inevitable market volume fluctuations. Selling data products fits that mature business model nicely.
2. Publishing depth-of-book quotes requires expensive infrastructure. The bandwidth needed to distribute that data in real time is significantly greater than what is needed to distribute top of book only. IEX may have chosen to defer that overhead investment.
3. IEX may be making a political statement. As revealed in the SIFMA case, the biggest users of depth-of-book data are the HFT firms that many IEX supporters dislike. IEX itself has thus far been friendly to these players in private, needing their market-making order flow to grow market share. But depth-of-book data, while a separate product, is usually sold alongside colocation services to HFT firms that are latency sensitive. Because IEX’s story has been its opposition to firms who use colocation to outrun those who don’t, it could be perceived as off-message to start profiting from a product that is perceived to be associated with colocation.
4. Publishing depth-of-book quotes could expose a technical inconsistency between intermarket sweep orders (ISOs) and IEX’s 350 microsecond speed bump. ISOs enable traders to execute against displayed quotes inferior to the NBBO if orders are simultaneously routed to execute against protected BBO quotes at other venues. If IEX displayed inferior quotes, those quotes would occasionally attract ISOs, some of which would not get filled if the inferior quote was stale due to the speed bump (e.g., if it was cancelled fewer than 350 microseconds before the ISO was entered). This “bait and switch” result would frustrate users of ISOs whose objective is to access larger blocks of stock at inferior prices. To be sure, there is always a risk of missing stale quotes on IEX due to the speed bump, even at the top of book, but the downside of missing large blocks at inferior prices in ISOs can be worse because the other sweep orders are likely to print and impact prices. To avoid frustrating customers, IEX may be choosing not to publish depth-of-book ISO bait.
As long as those responsible for routing orders have access to all available depth-of-book quote data, it shouldn’t matter whether one venue chooses to keep all its depth-of-book quotes hidden.
But what about the latest use case for depth-of-book market data: trade surveillance? Are the exchanges’ data business plans ready for that demand?
The Market Access Rule requires all registered broker-dealers to monitor their customers’ order flow for potential manipulation, including spoofing and layering. To effectively detect these types of behavior, it is necessary for broker-dealers to compare their own firm’s local order message traffic with the visible order book.
As prosecutors have alleged in the much-publicized ongoing “Flash Crash Spoofer” case against Navinder Sarao, in a layering scheme a trader seeks to deceive other investors by “creating a false appearance of market depth.” See U.S. v. Navinder Sarao, No. 15 Cr. 75 (N.D. Ill. Feb. 11, 2015), Complaint at 6. Sarao’s alleged layering entailed “typically plac[ing] his orders in the middle of the order book on the sell side, such as at levels 4 through 8” of the order book. Id. at 9. Sarao allegedly developed automated trading software to reprice his orders as the market moved, so that they would always remain several ticks above the best offer. Id. at 10. The only way to identify behavior of this sort – orders placed at levels 4 through 8 of the order book – is to know where levels 4 through 8 of the order book are. The only way to know that is to consume depth-of-book market data.
The Sarao case is hardly the only example of regulators referencing depth-of-book data to identify spoofing and layering. See, e.g., U.S. v. Michael Coscia, No. 14 Cr. 551 (N.D. Ill. Oct. 1, 2014) Indictment at 5 (“The quote orders would typically be the largest orders in the market within three ticks of the best bid or offer price, usually doubling or tripling the total quantity of contracts within the best bid or offer price”); Proposed FINRA Rule 4554, 81 FR 11851, Mar. 7, 2016 (requiring ATSs to report depth-of-book quote changes in order to “greatly enhance FINRA’s ability to perform certain order-based surveillance, including layering, quote spoofing and mid-point pricing manipulation surveillance, by enabling FINRA to more fully reconstruct an ATS’s order book”);Remarks of CFTC Commissioner Mark Wetjen, May 21, 2015 (“In order to detect other types of manipulation like spoofing, layering, and flipping, plus new types of gaming strategies, receiving order book and message data is necessary”).
My firm’s work in developing and bringing to market our Surveyor platform has shown us first hand the necessity of using depth-of-book quote data to effectively detect spoofing and layering. For an example of our depth-of-book-based layering analysis, click here.
Because regulators now require market participants to monitor their trading for spoofing and layering, and because detection of spoofing and layering requires analysis of depth-of-book quotes, not consuming all available depth-of-book quotes as part of their compliance program puts firms at their own regulatory peril.
The trade surveillance use case for depth-of-book market data could change the analysis in the SIFMA case. The universe of firms required to monitor for spoofing and layering is much broader than the core group with sufficient order flow to exert price pressure over the exchanges. The logic used to justify the exchanges’ prices to HFT firms does not apply to smaller firms who still need depth-of-book data for trade surveillance. The exchanges could avoid this problem by offering reduced prices for surveillance use. Some exchanges already do allow free use of delayed depth-of-book data, which is sufficient to monitor for spoofing and layering.
The absence of depth-of-book quotes on IEX is not a problem for broker trade surveillance. The harm of spoofing and layering is deceiving the market using visible non-bona fide quotes. Quotes that are not displayed cannot deceive.
The post Exchanges Take Opposite Approaches On Depth-Of-Book Market Data appeared first on Trillium Management, LLC.