Generating and disseminating market data is a much more difficult process than it seems, and the SEC’s proposal to overhaul the consolidated tape threatens the current balance of power – as well as the exchanges’ bottom lines. TABB Group founder and research chairman Larry Tabb breaks down the challenges with the current US consolidated market data-creation process, the SEC’s efforts to create a less-conflicted governance structure, how the proposal could impact market participants, and what’s next.
The process of generating and disseminating market data seems as if it should be straightforward. A quote is posted: pre-trade market data is created. A buyer and seller are matched: post-trade market data is created. How difficult can it be?
The SEC just last week proposed that the industry propose a new consolidated market data process. While the SEC requested that any proposal include a number of contentious aspects, the mere fact that the regulator’s proposal is just a proposal for a proposal actually is one of the more contentious aspects of the plan. In fact, it’s so contentious that two of the five SEC commissioners dissented.
But I digress. To understand the new proposal, one needs to understand the challenges of the existing consolidated tape.
While the notion of a consolidated public market data tape is one of the strongest cornerstones of our competitive and fragmented exchange environment, there are inherent latencies and governance challenges that push firms not only to buy the content-rich and lower-latency direct exchange feeds, which can cost firms hundreds of thousands of dollars a month, but also to bypass the official consolidated tape and acquire lower-cost top-of-book information provided by the exchanges.
The Convoluted US Consolidated Market Data-Creation Process
To understand why firms are choosing among the SIP, direct feeds, and exchange top-of-book feeds, one must understand the challenges of the current and convoluted US consolidated market data-creation process.
First, there are three SIPs and three plans, operated by two groups, run by two exchanges. The SIPs are represented by the three major tapes: Tape A, Tape B and Tape C. Tape A is comprised of NYSE-listed stocks; Tape C contains Nasdaq-listed stocks; and Tape B, which historically listed stocks on the American Exchange, now includes all corporate stocks and ETFs listed outside of the NYSE and Nasdaq. In addition, the three tapes are run under three plans, which are managed by two exchanges. Tape A and Tape B are managed by the Consolidated Tape Association (CTA), and Tape C is managed under the Unlisted Trading Privileges plan (UTP). The ICE/NYSE operates Tape A and Tape B SIPs out of its Mahwah, NJ, data center, while the Tape C SIP is operated and run out of Nasdaq’s Carteret, NJ, data center. While this made more sense when the NYSE and Amex traded only their own listed stocks and Nasdaq stocks were traded over the counter, today all exchanges trade all listed names.
The three plans, two tapes, and two consolidated tape broadcast locations create a problem. If a Nasdaq-listed stock is quoted or traded at the NYSE (or any of the other ICE/NYSE exchanges), the quote posted or the trade consummated in Mahwah needs to be shipped down to the Nasdaq data center in Carteret (approximately 35 miles away), to be aggregated into the Tape C data feed, and then shipped back to Mahwah to be consumed by anyone co-located at the NYSE data center. Just the transport of this data causes the SIPs to be between 350 microseconds (via microwave) and 850 microseconds (via fiber) slower than consuming the data directly from the NYSE or Nasdaq data centers, depending on the transport mechanism. That delay also doesn’t count the time it takes for ICE/NYSE or Nasdaq to aggregate, normalize, and redistribute the data.
So SIP data, as it is architected today, is inherently slow compared to consuming exchange-produced direct feeds in the co-located exchange data centers.
Second, SIP data is only top of book. This means that only the best-priced (round lot – 100 shares or more) trade at the top of the queue is displayed by the SIP. While this may have been fine back in 2004, when the average trade size was more than 500 shares per trade and when the average NYSE execution took approximately 9 seconds, today the average trade size is less than 200 shares per trade, more than 50% of the trades are odd lots (smaller than 100 shares), and quotes are produced and trades are executed in microseconds (or faster). So top-of-book information is less relevant than the full depth of book or some hybrid depth of quotes that consolidates volume at each price point.
Between the inherent delays and the top-of-book nature of the consolidated feeds, the SIPs have been all but rendered useless for trading. The SIP today is good only for human-consumable visual-display use, and while the NYSE and Nasdaq have significantly invested in improving the timeliness of the SIPs, there is no amount of investment and/or technology resourcefulness that can be applied to the existing architecture that can render these products useful for trading in this modern and algorithmic age.
Creating a Less-Conflicted Governance Structure
This leads me to the last and most significant challenge with the SIPs: the governance structure of the two organizations that run the consolidated tapes – the CTA and UTP – and the three plans that govern the SIPs – the Consolidated Tape Association Plan (CTA); the Consolidated Quotation Plan (CQ); and the Joint Self-Regulatory Organization Plan Governing the Collection, Consolidation, and Dissemination of Quotation and Transaction Information for Nasdaq-Listed Securities Traded on Exchanges on an Unlisted Trading Privileges Basis (UTP).
Given both the CTA and UTP are comprised of SROs, which themselves are comprised of the exchanges that sell both high-speed/low-latency direct feed data as well as top-of-book SIP-competitive data, there is little pressure for the exchanges to radically change the consolidated tape architecture.
Which brings me back to last week’s proposal.
Last week the SEC proposed that the exchanges (and FINRA) work together to come up with a new consolidated market data plan. The SEC “order[ed] the Participants in the Equity Data Plans to jointly develop and file with the Commission, as an NMS … a New Consolidated Data Plan that consolidates the three current Equity Data Plans.” The SEC further outlined a number of terms and conditions that must be included in the new plan, the most significant of which are:
- Consolidate the three SIP plans and two entities in charge of the SIPs into a single plan and governing body.
- Create a new governance structure that will assign each exchange group one vote; and any exchange with greater than 15% market share will receive a second vote.
- The exchanges and FINRA will comprise two-thirds of the SIP governing structure, while one-third of the plan will be represented by independent entities that have no vested interest in selling market data (which includes exchanges).
- An exchange will no longer be allowed to administer the SIP, which must be administered by an independent third-party entity not “controlled by a corporate entity that offers for sale its own proprietary market data product, either directly or via another subsidiary.” That said, the SIP can continue to be operated by an exchange entity.
While many of the SEC proposals do not get into the detailed mechanics of how the SIP will be operated and how data will be delivered, the SEC proposal is a good first step in determining a less-conflicted governance structure. This new governing body hopefully would then be able to tackle the practicalities of developing a new, more robust and more usable consolidated tape.
Now, with a new governance structure, a new and better consolidated tape could tackle some major issues:
- Reducing the latency associated with how data is aggregated, normalized, and redistributed. This hopefully will enable the SIP processor to aggregate and distribute consolidated data at each major exchange data center (ICE/NYSE, Nasdaq and Cboe), instead of data first being shipped to Mahwah and/or Carteret to be aggregated, normalized, and redistributed. This is part of Nasdaq’s Total Market proposal.
- Increase the depth-of-book levels incorporated in the consolidated tape – whether this is something akin to the addition of a new feed that aggregates all liquidity at a price point per exchange for the top five or 10 book levels (something similar to an idea I believe was proposed by the NYSE), or some other method of providing more depth-of-book information into the consolidated tape.
- Fix or revise the professional versus non-professional definitions that determine how much a firm needs to pay the SIP per user. Current rules categorize a janitor working for a broker as a professional while labeling a full-time but independent day-trader as a non-professional. This also is addressed in Nasdaq’s Total Market proposal.
- Putting odd-lot quotes into the consolidated quote. This may sound simpler than it is. A one-share quote for BRK-A at $340,500 represents real value, while a one-share quote for SIRI at $7.00 does not. A discussion is needed – and it is happening –on how to capture real liquidity for high-priced stocks on the tape without cluttering the tape with de minimis quotes that represent little if any serious liquidity.
While these suggestions are not revolutionary, they should improve price discovery and provide firms with greater choice in the selection of the type of data they need for their trading purposes. In turn, this should increase competition and reduce the cost of market data.
Now, will this reduce the need for the fastest and most algorithmic firms to use exchange-based direct fees? Probably not. The firms providing liquidity day in and day out need data by exchange, by price point and by lot as fast as it is produced. However, the majority of brokers, buy-side firms, ATSs and even exchanges do not need data at this level of detail for their trading and/or routing purposes.
Will these changes impact exchanges’ bottom lines? Probably. Exchanges have had a good run for the past decade as their exchange market data revenues have increased at a compound annual growth rate of 4.4%-10% if all exchange and non-exchange data revenues, technology and connectivity revenues are included. While we don’t think exchanges should be penalized for maximizing their revenue opportunities, given the buy-side fee compression and commission compaction on the sell side, a bit more parity between data producers and consumers would not be bad.
So, is this SEC proposal a panacea? Probably not. But it is a good start. With two of the five commissioners dissenting, however, the question is: Could this proposal for a proposal be an attempt by the commission to put the redefinition of a consolidated tape off for years?
Personally, I don’t think so. Being part of the SEC Fixed Income Market Structure Committee, reading and listening to the chairman and commissioners, as well as knowing the director of trading and markets for years, I believe the commission is serious about change. That said, I am not a student of SEC process; if the commissioners believe this is the best way forward, who am I to disagree?
This SEC proposed order is just that – a proposal out for industry comment. And while the proposal calls for the exchanges/SROs to develop a new consolidated market data plan, this is not an SEC order for the SROs to develop a plan. This proposal is just a strawman for developing the minimum guidelines to be included in an SEC order for the SROs to develop the plan.
This is the time for brokers, investors, exchanges, ATSs, and interested parties to comment. Then it is up to the SEC to incorporate these comments into an SEC order that will then be given to the exchanges/SROs to develop a plan.
While we may think that the exchanges/SROs will develop a single plan, however, I am not so sure the exchanges will coalesce around a unanimous opinion. Personally, I think there will be at least two exchange plans submitted and maybe more.
The three major exchange groups (ICE/NYSE, Nasdaq and Cboe), which have strongly influenced the CTA and UTP committees for decades, likely will share a similar view on a future consolidated tape, and we are probably to see them publish one plan. That said, given Cboe is less involved with the current consolidated tape, it may develop its own plan.
While Cboe may be an outlier, it would be hard to believe that upstart exchanges such as IEX and the future MEMX, LTSE, or even MIAX would fully concur with their larger and more established exchange brethren in a single view of the consolidated tape. Depending on the speed at which MEMX, LTSE and/or MIAX receive their equity exchange medallions, and whether one or more of these upstarts links with Cboe, I would expect that we will see at least two, if not three, SRO plans submitted.
These plans will be digested by the SEC, which will create a very specific and definitive order defining not only the governance structure of the new consolidated tape plan, but details on how this plan and the underlying technology should operate. This future plan not only will incorporate some facets of the plans received from the exchanges, it also will include ideas generated by the industry.
The final plan should be all-encompassing, neither ignoring the exchanges’ needs or turning its back on the industry. Exchanges, their broker clients, and investors all are integral parts of the production and consumption of the consolidated tape, and the SEC should take into consideration all of their opinions.
While hopefully this proposal will be a “Kumbaya moment” for the industry, driving development of a better consolidated tape, most likely it will not be. If the SEC alters the governance structure away from the exchanges, the locus of power over the consolidated tape will tip toward data consumers, and exchanges will not be happy. But will the exchanges be upset enough with the final order to attempt to stop the process? That is a good question.
Since the exchanges are now for-profit, and the SEC proposal seeks to reduce the exchanges’ governance/influence over the consolidated tape, any final order most likely would increase market data competition and impact revenues. This will push the exchanges to strongly exert their influence in guiding the final order to their benefit. But if the exchanges fail, I assume they would seek relief in the courts, as they did with the access fee pilot.
But, given both the NYSE and Nasdaq have offered their own proposals to make the consolidated tape more competitive, hopefully the process of developing a new consolidated tape plan will be harmonious and stay out of the courts.
The consolidated tape is an integral and beneficial aspect of the US equity markets. It provides the world with a single view of the market that is available to most everyone at a reasonable cost. While historically this technology has created opportunities for various constituents to monetize market data at the expense of others, I hope the SEC can, with the cooperation of the exchanges and the industry, redraw the US market data framework to reposition the consolidated tape as the efficient data distribution platform we need.
Market data is truly at the heart of American markets and finance. Traders, investors, issuers and many citizens need the data broadcast by the consolidated tape not only to value their holdings, but to know how to prepare for the day and week ahead, their kids’ educations, their retirements, and their financial lives. This makes restructuring the tape critical.
While I expect the process to redefine the consolidated tape to be contentious, I do hope that we can work together to restructure the consolidated tape to be more effective, efficient and competitive. Getting this right not only is critical for the SEC and the Industry, it is essential for the economy and the country as well, as most of our financial lives rely on it.