Re: File Number SR-NYSE-2015-57, Notice of Filing and Immediate Effectiveness of a Proposed Rule Change Establishing Fees for the NYSE Integrated Feed
Date: December 10, 2015
To: Robert W. Errett, Deputy Secretary, Securities and Exchange Commission
From: Alec Hanson, founder of AHSAT LLC, a proprietary trading firm
The proposed pricing of New York Stock Exchange LLC (hereafter, "NYSE") of its proprietary data products is not consistent with its legal responsibilities as a national securities exchange with a protected quote. It would impose an inappropriate burden on competition and is unfairly discriminatory, including against smaller computerized traders such as ourselves. Just for example, we have seen our costs for the same usage of NYSE's previous best data feed, Open Book Ultra ("OBU"), go from $1800 per year three years ago to $25,600 last year, $58,800 this year, $132,000 next year, and $330,000 if we switch to NYSE Integrated, as may be necessary to stay competitive. Such an inflation rate can best be explained by viewing NYSE as a monopolist with captive customers, a status awarded to it by regulation, which it is now abusing to the detriment of the markets. The SEC should stop this, deny NYSE's proposed rule change, and deny the termination of managed data programs like Managed Non-Display.
Per 15 U.S.C. 78f(b)(5) and (8), NYSE as an SRO has a responsibility to set rules that do not promote "unfair discrimination among customers...
and brokers," that "remove impediments to and perfect the mechanism of a free and open market and a national market system, and, in general, to protect investors and the public interest," and that "do not impose any burden on competition not necessary or appropriate." $330,000 per year is an impediment and a burden. NYSE itself has recently called on these responsibilities in response to IEX's application to become an SRO, arguing that speed should be a responsibility of a protected quote. I am arguing that fair access to the data determining that protected quote is also a responsibility, where fair access means affordable, broad in practice, and without a speed or content disadvantage. Such fair access promotes breadth of competition, which in turn leads to more liquid and robust markets.
So why exactly is the proposed pricing anti-competitive and discriminatory? First, ending Managed Non-Display is bad for smaller customers that use it, as well as for vendors who support it. Customers who relied upon this program and unit of count before it will see large price increases, in some cases beyond what their businesses can sustain.
Those and other customers who did not use Managed Non-Display will often still be compelled to upgrade from OBU to Integrated for its informational and speed advantages explained below, or risk falling behind the competition. When fractions of a microsecond *relative to the
competition* count for certain traders, the 100-microsecond advantage of NYSE Integrated is a requirement to be in business. For such firms, it is not voluntary or optional as NYSE states. The result is that some small firms may either reduce or end their business from lack of profit in light of better informed competition or from high fees payable to NYSE. This advantages larger customers, and will only promote the trend of consolidation that we have seen among market makers and short-term traders already engaged in a latency arms race. Such consolidation is unhealthy for the markets, because these missing firms present more competition to provide liquidity and stabilize markets in times of stress. If only a few market makers are left and they happen to come under stress, everyone will suffer. Simply look at the consolidation in the banking industry before the financial crisis and the subsequent too big to fail problem as evidence. Fortunately, some of NYSE's competitors have recognized this, and in particular, Nasdaq and BX offer the Managed Data Services ("MDS") program which charges fees of $100 to $375 per month per server, making it exceedingly more affordable for a firm with only a few servers. NYSE's proposed pricing is also anti-competitive versus other ATS's, brokers, and exchanges. Their customers push them to have the fastest direct feeds because they fear being "picked off" if only using slower consolidated feeds in their routing and matching. In this sense these faster proprietary feeds can be a must-have for them as well, despite the exorbitant costs. We should expect to see the same unfair advantaging of larger firms among such intermediaries. Finally, since this pricing hurts smaller and more computerized customers, and I suspect NYSE's competitors have relatively more of those types of customers, it hurts NYSE's exchange competitors by targeting their customer base, which is anti-competitive.
Having outlined the policy goals, legal requirements, effects, and winners & losers, I would like to provide some context against the half-truths in NYSE's filing. Remember that each exchange is the sole possessor of information about all of its orders, potentially indicating supply and demand among its customers, and no other exchange can offer that data. It has a monopoly on that information. As for particular feeds, NYSE's OBU feed provides depth-based information per price level.
For example, it would indicate a certain number of orders and the sum of the quantities at a given price level. Most flagship feeds from other exchanges for the past several years, including NYSE Integrated and sibling Arca Integrated, provide information per order, meaning they detail the coming and going of each order. In receiving them, we know not only the number and aggregate size of orders at a price, but whether the orders are generally the same size or dominated by one or two large orders, and whether the order leaving has been there for a second or an hour. This is richer information about NYSE than is available outside Integrated. Furthermore, the exchange claims it would be possible to integrate several data sources, but due to inaccuracy in timestamps from different sources and multiple simultaneous events, it is sometimes impossible to know certain orderings and associations other than from the Integrated feed. Were the data content actually equivalent, it would not be difficult for someone capable of handling this volume to integrate separate feeds. Finally, the Integrated feed is over 3 times or 100 microseconds faster to publish data from the matching engine
(source: private communication with a NYSE product manager). OBU is already faster than the consolidated feed. In its filings NYSE argues for the availability of close substitutes, but clearly they are selling unique data with a speed advantage, and at these prices, to a privileged few.
Further I would like to clarify some pricing policies. NYSE used to offer unit of count, which offered savings to smaller firms similar to the MDS program. This was ended, and Non-Display Use fees and Managed Non-Display were introduced last year. As of Jan 1, 2016 and hardly mentioned in any filings I can find, Managed Non-Display will end, and the pricing for the same data and usage for a previously eligible customer will jump as noted in the first paragraph, contrary to NYSE's claim in their filing that OBU "would continue to be available... at the same prices at which they are currently available." I also note currently the access fee for OBU includes the fees for BBO and Imbalances, and BBO and Trades fees had been combined, but all will be separated come Jan 1 (source: notice from NYSE, Aug 20, 2015). So when NYSE claims they are only raising the price from the sum of $7000 to $7500, really it is up from $5000 now, plus the comparable jump in Non-Display Use fees will be from $11,000 to $20,000 per month. Arca Integrated is slated for a similar elimination of Managed Non-Display, taking its price from $4000 per month to $10,000 (access fee + single category non-display use). Even NYSE MKT, currently charging $1250 per month for comparable OBU access, will be $7500 per month for Integrated starting Jan 1, 2016. For comparison, come Jan 1 and allowing MDS, the next highest non-NYSE-Group exchange will cost only $1500 per month.
While other exchanges have also raised prices at abnormal inflation rates, those under NYSE Group have been the worst offenders and are disproportionate to the amount of data or market share they include.
NYSE argues that data and transaction fees are joint products. But actually if we think carefully about the burdens of each, charges for data are a higher portion of costs for those customers who trade less with NYSE. In this sense, market data fees are charges for taking the information but choosing to transact elsewhere. If NYSE said they were going to charge their members *transaction fees* for trading on other venues, we would find this patently anti-competitive. But why is it okay if they charge excessive market data fees for trading away? I think NYSE MKT Integrated's pricing of $7500 per month is good evidence. Given its 0.06% market share (source: notional value summary from batstrading.com/market_summary for Dec 1 to Dec 9), virtually no one is actually using this data to trade on MKT. But some people have to have it. NYSE MKT is abusing Reg NMS and extracting an unreasonable rent via its SRO and protected quote status. In the case of NYSE Integrated, I would argue that the use of market share in listings and order flow to levy high prices for market data is similar to Microsoft's use of market share in Windows to promote its other products such as Internet Explorer, which was found to be anti-competitive. The NYSE filing also argues that data and transaction pricing may be complimentary, implying that they may charge more than the competition for market data, but their transaction fees may offset this. They don't. This is obfuscated by the tier-based pricing most exchanges employ, but using a representative top 5 broker and looking at what the exchange keeps between their take fees and provide rebates, NYSE is the highest among the top 9 SRO's (ranked according to batstrading.com/market_summary).
NYSE simply charges the most for both products. Yet the Order Protection Rule mandates we do business with them.
The irony of these price increases is that the technology changes probably save NYSE money, rather than being some great cost they incurred to offer novel technology. Their sister SRO, Arca, has been using this same technology for years, and NYSE Group is simply consolidating onto one technology platform, which likely saves them support costs. Other exchanges have long had comparable technology as well, and are able to offer their data at prices an order of magnitude or two lower. So to the extent it matters under current law, the pricing probably does not reflect reasonable costs to the exchange.
NYSE also argues that non-display pricing is justified because it affords the user greater revenue and cost saving opportunities.
Effectively this says to market participants that if they invest in better technology, make themselves more efficient, and do a good job, NYSE should receive a portion of their profits just because of NYSE's privileged position as an SRO. This discourages market participants from innovating, and in the long-term leads to less healthy markets. It discriminates against those traders who work via computer rather than eyeball. Furthermore, since most of the exchanges and ATS's that "compete directly with the Exchange for order flow" do so through automation and may accumulate multiple hefty Non-Display Use fees, and the portion of fees that are for Non-Display Use is higher for NYSE Integrated, this pricing policy can also be seen as targeting NYSE's competitors.
NYSE's argument against per server pricing, or in favor of ending Managed Non-Display is basically that they can't count. Yet Nasdaq has successfully done so in its MDS program for many years. NYSE continues to be able to count orders of magnitude more Users than servers with enough accuracy to continue all User fees. We have no reason to believe that it is any harder to track and permission servers than humans.
Automatic software tracking of entitlements has been widespread for many years. In short, we have good technical solutions for this supposed problem, many of which NYSE already uses. In my opinion per server is actually a very fair pricing scheme that encourages competition; it is also a fairer reflection of the amount of usage and required resources, i.e. both the benefits and marginal costs of the data. Also keep in mind that these managed pricing programs are optional to the customer. I ask the SEC to prevent NYSE from discontinuing this means of access and to push lower pricing thereof, in line with Nasdaq MDS. While NYSE Integrated and NYSE Group are the most egregious examples, rising prices in the past few years, especially for smaller or non-display customers, has been a problem across the industry.
In closing, one of the most effective counterarguments to critics of high frequency trading is that anyone can do it. Fair access means anyone can sign up for the same data as the largest firms on Wall Street
-- same speed, same content, same price. But when that price is $330,000 per year for just one of almost a dozen SRO's, not everyone can. That speed and information advantage is only available to the richest, biggest players. That's unfair. And an overly consolidated ecosystem of market makers and their service providers is unhealthy to the liquidity and stability of the markets. NYSE is using its market share and its legal position as a protected quote to disadvantage certain market participants and probably its competitors while enriching itself. The SEC should deny this proposed rule change, limit or undo future and past market data fee increases, and require fair access programs such as Nasdaq's MDS.
Thank you for your consideration,