Firms need error rates approaching perfection when they are engaged in CAT reporting — a key takeaway, among others, from an FTF webinar, “Are You Ready for Interfirm CAT Linkage?”
Despite the pandemic, sell-side firms in 2020 met a very challenging first-round of compliance with the Consolidated Audit Trail (CAT) trade data surveillance regulatory initiative, learning firsthand that gathering, cleansing, and reporting interfirm transaction data to regulators is much easier said than done.
Yet firms are applying lessons learned and more to the CAT deadlines in 2021 as they adjust to the unprecedented scale of a big data trade surveillance initiative.
These are the major takeaways from a webinar on the subject, “Are You Ready for Interfirm CAT Linkage?” The event was sponsored by Gresham Technologies and produced by FTF.
Overall, adjusting to the demands of CAT reporting is likely to be a new challenge for small-to-medium-sized firms, so the key to success is education and training as exemplified by the complexity of the interfirm CAT linkage reporting requirements, says Harshad Pitkar, founder and CEO of RegEdge.
“When we have to report the audit trail of an order, it is extremely challenging. It is not like a position reporting or transaction reporting where you’re just reporting trades or reporting positions at the end of the day,” says Pitkar, who was a webinar panelist. “For CAT you need to report each and every change to the order, all the hops, slices, aggregations … The smaller firms are not used to that level of complexity and storing that level of granularity.”
An Ops Tightrope
Like the felines that its acronym refers to, this CAT may have at least nine lives as it has had a controversial path since the SEC officially launched it on July 11, 2012, when it adopted Rule 613 under Regulation National Market System (NMS).
The SEC requires national securities exchanges and national securities associations “to create, implement, and maintain” CAT, an unprecedented big data undertaking. FINRA CAT LLC, a subsidiary of Financial Industry Regulatory Authority (FINRA) serves as the plan processor for the effort.
The CAT plan processor must ultimately create an access method for the SEC to access CAT data for regulatory and oversight purposes and must provide capabilities for firms to run complex searches and generate reports.
But, before the plan processor can do that, firms must meet an ongoing set of CAT deadlines, which resume in January 2021 after the CAT interfirm linkage deadline in the fall of 2020. (More information about the deadlines can be found here https://catnmsplan.com/timeline.)
As broker-dealers revamped systems and operations to meet the CAT interfirm linkages reporting deadline of Oct. 26, they found themselves walking an operational tightrope, according to webinar panelists. They had to balance the necessity of processing and reporting huge volumes of transaction data against intense demands to squeeze out nearly all errors.
In the run-up to the initial CAT deadline, firms were reporting that they were experiencing high error rates.
Shelly Bohlin
In fact, long before the Oct. 26 deadline, when the interfirm linkage code was released for usage in production environments in August, there were high error rates at the start of the reporting effort, says Shelly Bohlin, chief operating officer (COO) of FINRA CAT LLC. Bohlin was a webinar panelist.
However, as firms got closer to the Oct. 26 deadline, they were able to push down their error rates dramatically to less than three percent, Bohlin says.
“Initially, the error rates were very high — in excess of 40 percent on the sender and receiver side,” Bohlin says. “Since that time, firms have been working extremely hard. I think getting coordination with their counterparties has been a huge lift.”
In fact, during the FTF webinar, only slightly more than a third of online participants, when polled, reported error rates of less than one percent. Half reported error rates ranging from two percent to 10 percent, and nearly 14 percent reported that their error rate levels were above 10 percent.
The snapshot results were:
- 0 to 1 percent error rates: 36.4 percent
- 2 percent to 5 percent error rates: 27.3 percent
- 5 percent to 10 percent error rates: 22.7 percent
- Greater than 10 percent: 13.6 percent
FINRA CAT officials and others are pushing firms to have extremely low error rates because of the very large volumes that firms will have to process.
“Let me just do some basic math here,” says Joshua Beaton, head of Americas trade and transaction reporting at Morgan Stanley.
Joshua Beaton
“If your firm is submitting 100 million records per day … and you have a 1 percent error rate, that’s still 1 million records a day that need to get fixed by T+3. If you’re doing a tenth of that, that’s 100,000 records a day that need to get corrected,” says Beaton who also was a webinar panelist.
“These numbers are so large that you need to have a very low error rate in order to keep up with it, in order to stand a chance to fix things,” Beaton says. “We’re not talking hundreds of exceptions here — we’re talking tens of thousands, hundreds of thousands unless your quality is very good right off the bat.”
Given the huge volumes, the high levels of exceptions, and other problems to deal with, cutting-edge tools such as machine learning (ML) and artificial intelligence (A.I.) could be seen as likely candidates for finding a shortcut through the data maze.
But, despite the media hype, ML and A.I. are not likely to be immediately useful although they may one day become very desirable for data lineage checks to find the root cause of problem areas, according to the webinar panel. Systems that leverage A.I. and ML are out of reach for most firms now and tend to be pricey, causing most firms to not see compelling use cases yet.
Citing anti-money laundering (AML) and its embrace of A.I. as an example, Philip Flood, the chief commercial officer for Inforalgo, which is owned by Gresham Technologies, says that it’s early days for cutting-edge IT and CAT reporting.
Philip Flood
“A.I. is now in use with AML technology and has some application in regulatory reporting but it’s still in its infancy,” Flood says. “With the volumes and data quality issues for CAT, there is a good case for its use in both unsupervised learning to make reporting decisions and supervised learning to optimize through reconciliation.”
As for the major pain point of interfirm linkage errors, “it’s difficult to use A.I. where manual collaboration with a counterparty is required especially when different reporting systems are being used,” Flood says. “At present, the rush to comply and meet the deadline stifles innovation for firms developing inhouse solutions.”
Looking ahead, RegTech vendors “will be essential in driving” A.I. innovations for CAT reporting, Flood says.
So, instead of the cutting-edge solutions, firms are finding that more practical approaches are meeting their needs such as the effective parsing of data, the establishment and refining of workflows, and the adaptation of generalized tools for CAT reporting, according to the webinar panelists.
Buy, Build or Both?
While firms and exchanges have found that they can quickly shrink their high error rates to 1 percent, getting below that point takes a long time. To cope with the scale and granularity of CAT reporting, the FTF webinar participants report that firms are exploring third-party reporting solutions and internally developed systems.
A quick poll of webinar participants revealed that a majority of them favored internally developed systems and/or customization of third-party platforms to facilitate CAT reporting.
Harshad Oitkar
Actually, firms that adopt third-party, hosted solutions are more likely to need customization because a majority of the work pertaining to data sourcing and normalization has to be done internally, RegEdge’s Pitkar says. This is especially true for mid-sized and large broker-dealers that have complex trade flows and multiple order management and routing systems
“Even if you use a vendor, I would say 50 percent of the work is inhouse,” Pitkar says. “It’s more like a hybrid where firms are doing an in-house build for consolidating data, normalizing it, and sending it to the vendor.”
Beaton says that he was not surprised by the polling question results because many firms reviewed that what they had in place for the incumbent Order Audit Trail System (OATS) reporting to discern whether it could be adjusted for CAT reporting. They found that the legacy systems were “not fit for purpose,” Beaton says.
So, what should firms do with their legacy OATS reporting platforms?
“The legacy OATS reporting solutions will need to continue to be maintained in parallel to submissions to CAT,” Flood says.
“FINRA proposed in August 2020 to remove the OATS series rules when firms are able to demonstrate accurate and reliable reporting over a 180-day period. Certainly, most market participants and SRO’s [self-regulating organizations] would like to retire OATS as soon as possible to avoid the operational cost of maintaining duplicate reporting. This is likely to begin in Q2 after the go-live of Phase 2c equities, but the retirement could take an extended period of time. For the short-to-medium term, firms should plan to continue with OATS support in 2021.”
The full, free webinar, which also included Jeff Wells, technical project manager for CBOE Global Markets, happened Oct. 22, 2020, can be seen via FTF’s archives here: http://bit.ly/38rjFSB
Need a Reprint?