For decades, the vibrant and competitive patchwork quilt of US equities markets functioned smoothly thanks in no small part to the fact that the information flow from these fragmented markets was controlled via two official consolidated tapes of equities market data—the Consolidated Tape Association administered by the New York Stock Exchange, and the Unlisted Trading Privileges (UTP) Plan administered by Nasdaq, collectively referred to as the Securities Information Processor (SIP) feeds. 

Each comprised data from its primary market (i.e., NYSE or Nasdaq), its main competitor, and the network of regional exchanges throughout the US. However, with the consolidation of those regional markets under centralized mega-exchange groups, and the proliferation of proprietary market data products from those exchanges, market participants are questioning the relevance of the SIP feeds, and whether they still serve a purpose in today’s modern trading environment.

The creation of the SIP feeds dates back to the 1970s and the need to consolidate best bid and offer data from all US equity exchanges and to create “official” national best bid and offer (NBBO) reference prices—a valuable piece of information because nearly 97 percent of US equity trades are executed at or within the NBBO, according to Nasdaq.

But the disco days are over, and just as bell-bottoms and platform shoes have (mostly) disappeared, so have many of the exchanges that made the SIPs a necessity. For example, NYSE has absorbed the American Stock Exchange and the Pacific Stock Exchange—and itself been acquired by Intercontinental Exchange, which also bought out the Chicago Stock Exchange—while Nasdaq acquired the Boston and Philadelphia exchanges, and Bats Global Markets bought Direct Edge and was itself acquired by the Chicago Board Options Exchange. And while some of those marketplaces retained their exchange licenses under their new ownership to offer different market models, connecting to data from all stock markets at once is no longer the technical burden it once was.

At the same time, investors—both retail and institutional—have become more sophisticated. As technology has become more commoditized and affordable—such as ticker plant technologies once reserved for data vendors in the business of consolidating data streams—trading firms now have the capabilities and capacity (as well as the appetite, given the emergence of high-performance strategies, such as high-frequency trading, which rely on low-latency direct feeds from exchanges) to consolidate direct feeds themselves.

As a result, some in the industry who now spend a small fortune on direct exchange feeds are asking why we still have SIP feeds at all.

“The SIPs may be good enough for retail investors looking for quotes … but if you’re talking about algorithmic and automated market-making systems that require a certain level of timeliness … the SIPs don’t work,” says Joe Wald, CEO and co-founder of New York-based Clearpool Group, a provider of algorithmic trading technologies. “The SIPs do and will have their place for screen-based, retail trading and for academics researching the data. The SIPs are great for certain things. They’re an important utility to have … but they don’t come close to meeting the demands of sophisticated investors.” 

Not all share this view. Speaking at a two-day roundtable on market data on October 25 and 26 last year organized by the SEC, Bryan Harkins, executive vice president and co-head of the markets division at Cboe Global Markets, said the majority of its trading participants appear to be content with the top-of-book SIP data.

“We have about 200 members … and only less than half of them take our depth-of-book product. So I don’t know what they’re doing, but clearly half of our customers can do business on our platform by using this SIP,” Harkins said.

But most broadly agree with the sentiment that SIP data plays an important role in retail brokerage, but not in the cutting-edge, bulge-bracket trading—though some say it could become more widely used if the SIPs made some changes to the content they carry and how they capture it. For example, speaking at the same SEC event, Paul O’Donnell, managing director at Morgan Stanley, said one step forward could be to plug the same wireless-delivered direct feeds used by trading firms into the SIPs’ already low-latency—if not ultra-low latency—consolidation process.

“We could use that for a lot of our business. Maybe not all of it. … But certainly it starts to become a real choice at that point, and we’d be thinking hard about where we use that, versus where we don’t,” O’Donnell said.

Room for Improvement

Though often cited as a criticism of the SIP feeds by trading firms with high-performance strategies, most observers say latency is a red herring: Yes, the SIP feeds are slower than direct feeds, but the latency is acceptable, and only a minority of firms care about the difference, they say. And for these observers, the benefits of proposed technology initiatives—such as a cloud-based “distributed SIP” with data collection infrastructure replicated within key datacenters that host exchange matching engines—have more to do with improving resiliency than the accompanying reductions in latency.

But while technology is one area of focus for modernizing the SIPs, the industry is also looking at what additional content could make the feeds more relevant, with sources saying the SIPs are hamstrung by only offering limited top-of-book data while the exchanges can offer other valuable datasets.

For example, industry participants have proposed adding market depth data to the SIPs, which would show much more liquidity, reflecting the greater volume that exists outside the BBO. Other suggestions include adding imbalance data from opening and closing auctions, since more trading now takes place during these periods, as well as odd lot-sized data, though this would require trading venues to implement changes around order sizes.

“If I were a bank or other institution, I would want a lot of the controls that currently exist for Level 1 [best bid and offer] data also placed over Level 2 [market depth] data, so that exchanges can’t set their own prices and rules. The SIP feeds are currently much more regulated than proprietary exchange feeds,” says one veteran industry observer.

But change—no matter how desired—can be a slow and tedious process. “We work within the constraints of the Plan, which is overseen by the SEC, so we can’t just make changes.  We collaborate with the advisors, get buy-in from the exchanges, and obtain SEC approval,” says Emily Kasparov, chair of the SIP operating committees. 

“Any significant change is a very formal process—just as it would be for any of the exchanges—involving a proposal and a plan amendment with a description of what would be done, the timeframe … then the SEC would publish the proposal on its website. Depending upon the change, there may be a comment period … and then the SEC would make a decision. It may take a few months,” Kasparov adds. 

In addition, implementing any such improvements requires the exchange operators that participate in the SIP feeds to overlook some inherent conflicts of interest between running a for-profit exchange data business and contributing to an industry standard SIP feed.

These conflicts can be as simple as withholding some of the content types described above, operating two-tier technology stacks for running their own data infrastructures versus the SIP infrastructures, or competing products that directly undercut the SIP feeds, such as Nasdaq Basic, which was sold with the mantra of replacing SIP feeds at a lower cost. Bats Global Markets introduced a similar product, Bats One. But while both provide SIP-like data for their own markets, they don’t provide a full, comprehensive view of all US equity markets. And though they tout a cheaper alternative to the SIP feeds, assembling that comprehensive view from multiple providers would prove a more costly proposition.

“As a private investor, I’d love to see market depth data on the SIP feeds, and to see trade sizes of 2,000 shares … but then you’d be undercutting the value of exchanges’ proprietary data products,” the industry observer says.

“There needs to be a combination of legal and commercial change. If exchanges as for-profit entities are allowed to create products that make the SIPs obsolete … there needs to be change,” Wald says.

Not everyone sees the conflict in the same light, though. Some exchanges downplay the issue of conflict, while others claim no such conflict exists for them. At the SEC event, Cboe’s Harkins noted that the exchange makes “a lot more money” from its participation in the SIP feeds than from proprietary data products, and that “having a strong SIP … is core to our business.”

Obligation Complication

It should be noted that there isn’t a specific obligation on firms to subscribe to the SIP feeds. In the past, firms took the SIP data because it was the industry standard, and there wasn’t a better alternative. Now that more performant direct feeds exist, though many firms still grudgingly subscribe to the SIPs, they aren’t strictly required to. Their obligation is to provide best execution, and to do so with suitable data to support that function, wherever that data is sourced from. But in today’s high-performance ecosystem, the notion of best execution has become blurred.

“The legal obligation is on best execution … and that is basically dictated by what the buy side considers to be best execution,” says Clearpool’s Wald. “But you cannot fulfill best execution requirements and service the buy side with just the SIP … so it’s a non-starter to use a broker who can’t demonstrate that they are using proprietary feeds for best execution.”

Indeed, according to a study conducted last year by Greenwich Associates, 25 percent of firms polled said they would refuse to trade with a broker that only used SIP data, compared to 2 percent that said yes—though another 34 percent said they would trade with a broker that only used SIP data “only if there was regulatory clarity” that by doing so the firm would be fulfilling its best execution obligation.

“It’s clearly more economical to take the SIPs than it is to go to all the exchanges—from a feed perspective as well as in terms of the technical infrastructure required to support it. But I think lots of market participants feel they need to know the best price for best execution on any venue, including understanding the implications of latency … and if the SIPs are slower, there certainly is the potential for a different result at that microsecond,” says Dan Connell, managing director of Greenwich Associates’ market structure and technology practice. “If the SIPs met the conditions of best execution, maybe people would use them for trading as well.”

The NBBO-All and End-All

The best measure of best execution is comparing trade prices to the NBBO. However, many note that the current trend of firms capturing and consolidating low-latency direct feeds in-house can produce differing BBOs, depending on what feeds they consume, and their latency from the source—i.e., that since the BBO may change from one microsecond to the next as data updates, a one-microsecond latency difference in one of the feeds could skew a firm’s BBO.

“The complaints of the SIPs being slow are overblown,” the industry observer says. “Only high-frequency traders need the absolutely fastest feeds … and the benefit of that extra hop is that you get an NBBO. Yes, you can create that yourself, but you might get a different result for a firm in Chicago versus for a firm in New York.”

And herein lies one of the ironies of efforts to update the SIP feeds: the suggested alternatives may increase flexibility and reliability, but also potentially increase complexity and the potential for fragmented versions of an official record—such as different versions of the NBBO.

Speaking to Inside Data Management before the SEC’s roundtables last year, Oliver Albers, senior vice president and head of strategic partnerships for Nasdaq’s Global Information Services business—which, as administrator of the UTP Plan, preempted the SEC event with its own manifesto of suggestions to modernize and improve the SIP feeds, warned against over-complicating the current setup.

“A lot of people are advocating for a distributed SIP … which would change our market structure. Nasdaq’s position is that we think the structure of the SIP today works pretty well,” Albers said. “Multiple or distributed SIPs could be confusing and benefit some stakeholders at the expense of the investing public … and may mean a significant increase in costs—who would bear those costs?”

The cost of market data remains a major source of animosity between financial firms that require data and the exchanges that produce and sell it—for example, low cost is one of the founding principles of the recently announced Members Exchange (MEMX), formed by a consortium of nine sell-side and buy-side firms, which aims to disrupt the exchange status quo. Over recent years, with the advent of “effective upon filing” status for fee-liable data products, most exchange data revenue-raising initiatives have gone unchecked by the SEC. When the regulator rejected two long-disputed filings from NYSE and Nasdaq last year—albeit for not supplying sufficient information to support the fee levels, rather than for explicitly being unfair (the products and fees in the filings had long-since been superseded anyway)—it angered exchanges, who argued that the SEC was overstepping its bounds, while giving hope to data consumers for an “evolutionary leap” in the SEC’s thinking towards the provision of core market data.

The Inevitability of Evolution

Though there is still fierce debate over what form the SIPs should ultimately take, and what user base they should focus on addressing, there is consensus that evolution is necessary. 

“The SIPs are an important part of the US market structure … and a critical part of what makes our markets so transparent and liquid,” says Clearpool’s Wald. “The SIPs’ mandate should be A; the exchanges’ mandate should be B. They serve different constituents and different use cases, and they need to be fairly priced. But they should coexist.”

But that evolution depends on the exchange participants being willing to accept and agree on change, and if necessary, the SEC being willing to drive it. Though at time of writing, the SEC has yet to announce any decisions based on its two-day roundtable on market data last year, observers say the fact that the regulator held such a wide-ranging discussion solely devoted to market data is a sign that it realizes the topic needs to be addressed, and that the purpose of the event is to move the industry as a whole closer to concrete action.

“Following the market data roundtables, we’ve created a governance sub-committee to address peoples’ questions about structure and to consider potential changes … and whether we need plan amendments or SEC rulemaking to make changes,” Kasparov says. 

One frequently suggested change is removing the requirement that all exchange participants in the SIPs must agree unanimously on any changes, and allowing the end-user advisors to also vote on proposals. Critics say the unanimity rule allows exchanges to oppose anything that would threaten their proprietary business, but—speaking at the SEC roundtable event—Kasparov said the requirement protects smaller exchanges against the whims of larger rivals.

“I think the current construct works,” Kasparov said. “But I think it’s important to remember when we do consider changing the allocation of votes, that we do protect minority rights, right? I’ve been a member participant at CHX [the Chicago Exchange, now owned by Intercontinental Exchange], which had one vote. That vote was important to us. It made a big deal, you know. And it made a difference in the plans in some circumstances where unanimity was required. And we were able to make some good changes and good strides to come to agreement with our vote. So I think that’s important.”

Arguably, any changes to such an important part of the US capital markets infrastructure are “important.” After all, the SIP feeds aren’t just an alternative to exchange offerings; they’re the official backup quote feed of record, and—though intertwined with the exchanges—operate independently from them. Without a mandate from the SEC that firms take the feeds, the SIPs are also subject to the budgetary whims of consumer firms, so to maintain these positions, the SIPs must inevitably modernize and find balance between their obligations and the exchanges’ commercial needs.