Not so Fast: SEC’s SIP Rule Speeds Ahead, But Faces Bumpy Road

The Securities and Exchange Commission (SEC) says that market data consumers operate in a two-tier system, where the haves can afford direct feeds from exchanges, and the have-nots make do with the public, consolidated feeds from the securities information processors (SIPs), which contain less information and are slower.

As SEC commissioner Allison Herren Lee put it last year while announcing the agency’s finalization of its market data infrastructure modernization rule: “In today’s markets, while some can afford a pricey trip along a freshly paved, proprietary high-speed toll lane, others are relegated to a cheaper ride on a public highway with cracked pavement and potholes.”

The commission’s rule, which it finalized in December, will upgrade the road surface for these less well-funded customers, ensuring that the “trading experience on the SIP data highways is closer to that of the trading experience on the proprietary autobahns. The commission expects this will enhance the utility of SIP data especially for the market participants that do not compete in the trading space where participants seek to slice milliseconds off trading times,” Lee said.

Lee’s is not a metaphor that, once you stop to think about it, accurately represents how market data is consumed in real life. The SIPs aren’t just for Main Street investors, humans looking at prices on a screen: The association of the direct feeds with more sophistication is less and less true, as the exchanges have upgraded the SIPs’ speed, and many institutional users have come to rely heavily on them.

Trading technology provider BestEx Research looked at regulatory filings on alternative trading systems (ATSs), including large broker-dealers, and found that they use SIP feeds a lot: either as a single source of pricing information, as a supplement in lieu of select direct feeds, or as a backstop for certain data. “More than one third of ATSs use the SIP—exclusively—for pricing trades. In some cases, ATSs use direct feeds for some markets but use one SIP for the primary markets, perhaps due to the high cost associated with those feeds in particular,” BestEx CEO Hitesh Mittal wrote.

But clumsy metaphors aside, it’s the regulator’s job to worry about leveling the playing field for all investors, and since there are all kinds of users using the SIPs, there is an argument to be made for reducing the differentials between them and exchanges’ direct feeds. The way the commission has looked to do this in the final rule is to mandate the exchanges to send market data to the SIPs in the same manner as it does to its proprietary data customers, while providing for the emergence of multiple SIPs, called competing consolidators. These measures—and the hope that competing consolidators could be located in the same datacenter as their subscribers, who would no longer have to wait for the data to come from Mahwah or Carteret—is how the SEC hopes to reduce latency in the SIPs.

Like everything related to market data regulation, there is little agreement about whether the SEC should try to reduce the latency of the SIPs. Some market participants say the SIPs are fast enough. And for, say, large broker-dealers, which couldn’t compete in algorithmic executions if they were to rely on the SIP feeds alone, the rule is unlikely to reduce latency in the SIPs to the extent that it would benefit them, since factors like the consolidation of the data itself adds latency.

Others, like BestEx’s Mittal, say that if indeed there are a variety of users reliant on the SIPs, then there should be a real attempt to reduce latency. But the question at this point is no longer should the SIPs be faster, because the SEC has finalized the rule that says they should be, but rather, will a system of multiple SIPs reduce latency in the consolidated tapes?

At this point, the regulator hasn’t provided enough of an economic analysis to satisfy potential competing consolidators that it’s worth taking the risk in registering as one. And if they did emerge in enough numbers to justify the changes made to the SIPs (the SEC wants up to 12), it doesn’t necessarily follow that their emergence will reduce these geographic latency differentials. The SEC itself acknowledges that in this new world, broker-dealers that performed their own data consolidation would have a speed advantage over market participants that had to consume data from a competing consolidator. Not to mention the fact that the competing consolidators themselves would wind up competing over speed, since, while they might receive the data from the exchanges at the same moment, their own switches and hardware and how long they take to consolidate the data will produce a speed differential among them.

In other words, in trying to mitigate the expense to investors of a two-tiered system by introducing more competition, the regulator could be in danger of introducing…a two-tiered system.