The most bedeviling aspect of exchanging regulatory margin for non-cleared derivatives is the calculation of portfolio risk sensitivities, which are used as inputs for the industry’s standard initial margin model, or Simm. 

Buy-side firms caught in the regime are relying on a small group of vendors—including AcadiaSoft, Bloomberg, IHS Markit and TriOptima—to perform this task. But first, the trade data needed to run the calculations must be standardized.

This is where the problems start. 

While recognized data standards for describing over-the-counter derivatives do exist—notably, financial products mark-up language, or FpML—these have not been widely adopted on the buy side.

“Many firms do not have their trades in this format [FpML], so the data-mapping work and quality of this data for the purposes of calculating risk sensitivities is a challenge,” says Amir Khwaja, chief executive of Clarus FT, which offers a Simm calculation service.

Some vendors include data standardization as part of their Simm calculation services. AcadiaSoft, the dominant vendor, does not. Clients that sign up to use its IM Risk Generator service must standardize and deliver their trade data in a specific format, called ORE XML—or have one of its partners do this for them, at an additional cost.       

“What might be a little unique about AcadiaSoft is that they do have this ORE XML requirement,” says a collateral specialist at a custody bank. “For firms that don’t have the capability to match that, you’re paying for time and materials.”

Most buy-side firms are unfamiliar with ORE XML and are said to be completely baffled by the standardization requirement, which calls for trades to be broken down into their constituent parts—running to hundreds of data fields for the most complex instruments – and then translated into the vendor’s code. 

“With all the permutations of the OTC derivatives that you can have, it’s not an easy feat even for those that are very sophisticated with derivatives systems and data capabilities,” says Judson Baker, derivatives product manager at Northern Trust.

With all the permutations of OTC derivatives that you can have, it’s not an easy feat even for those that are very sophisticated

Judson Baker, Northern Trust

AcadiaSoft has amassed the lion’s share of buy-side margin calculation business, in part due to its bargain-bin pricing. The firm claims its service will cost most clients far less than $50,000 a year. Competitors are charging at least double that.

Nearly half the 100 or so firms that will start exchanging margin early in the next phase of the regime are expected to use AcadiaSoft for Simm calculations.  

But that leaves clients with the data standardization problem.

Fred Dassori, head of risk products at AcadiaSsoft, makes no bones about the scale of the challenge. “There’s no question that there is some amount of work required for data standardization. It’s not a snap-your-fingers issue,” he says. “If bilateral derivatives data were easily standardizable, the market would have already gravitated to a single standard.”

When the industry was still facing a September 2020 deadline for the next phase of compliance, there were real concerns about whether firms would be able to complete this work in time.

“One of the reasons timing came under threat is that, suddenly, everyone due to be in phase five is understanding the work involved to become up and running,” says a rival margin analytics vendor. 

Then the coronavirus outbreak hit and regulators agreed to postpone the deadlines by a year. That should give AcadiaSoft’s clients enough time to get their data in order—and the OTC derivatives market may finally end up with the common data standard it has been lacking.

Loyal to the ORE

The non-cleared margin rules require counterparties to post regulatory initial margin in segregated custody accounts to back their bilateral trades. The regime has been rolled out in phases since 2016 and an estimated 1,000 firms are due to be caught in the next waves. On April 3, regulators agreed to temporarily pause implementation, delaying the fifth and sixth phases until September 2021 and 2022, respectively. 

Firms subject to the rules must calculate margin amounts using a standard grid or a regulator-approved internal model, such as the International Swaps and Derivatives Association’s Simm.

Risk sensitivities are a critical input for Simm and must be provided in a standard template, called a common risk interchange file. Generating this file is considered to be one of the biggest challenges with using the model. 

“In terms of where the challenge lies, the complexity of Simm sensitivities is high up on the list,” says the rival margin analytics vendor.

AcadiaSoft realized this and called in outside help. In August 2018, it tapped Quaternion, a risk analytics and software firm, to provide the models needed to calculate Simm sensitivities.

To run the calculations, Quaternion needs trade data to be in ORE XML, an open-source version of Extensible Markup Language that it calls Open Source Risk Engine.

Converting trade data to ORE XML involves mapping each bit of information about a trade—for example, the maturity, notional and cash flows—to a series of data fields, known as components. A vanilla interest rate swap may have 10 to 15 such components. An amortizing swap may require another 50 to 60 data fields to capture individual cash flows on each side. For a collateralized debt obligation, hundreds of data fields may be needed.

There are institutions out there who probably underestimated the complexity or the nature of the data associated with OTC derivatives

Donal Gallagher, Quaternion

The whole process can take three weeks for vanilla portfolios and around six weeks for the most complex. Some of AcadiaSoft’s clients were overwhelmed by the enormity of the task.

“This plumbing at the heart of the middleware of the financial system is not a new thing,” says Donal Gallagher, chief executive of Quaternion. “What might be true is that it’s newer to the phase five and phase six firms, so there are institutions out there who probably underestimated the complexity or the nature of the data associated with OTC derivatives.”

Around half the firms that signed up to use AcadiaSoft to calculate Simm sensitivities have punted this work back to Quaternion.

AcadiaSoft’s Dassori says these firms have made a conscious decision to outsource the project. “The reason the clients take us up on it is because we can likely do it more efficiently than they can,” he says. “It depends on what kind of resources the client has in house and how much effort they want to dedicate to this transformation.”

AcadiaSoft declined to disclose the cost of the data standardization service. 

Some of AcadiaSoft’s other partners are also stepping in to assist. Northern Trust, which has an arrangement with AcadiaSoft to provide margin calculations for its fund administration clients, will also offer a data standardization service in an effort to keep costs down.

“For years we’ve been sending our clients’ OTC derivatives positions over to pricing vendors, which includes pretty much the same level of information required for the risk sensitivities and Simm. So we’re pretty accustomed to having to maintain and extract all this information in an organized fashion,” says Northern Trust’s Baker.

At least one other custodian bank is understood to be developing a similar service for its clients.

A new standard?

Other vendors have different models. Cassini Systems says it will convert the trade data received from clients into its preferred format—which is derived from FpML—free of charge.

Bloomberg accepts trade data in any format, including Excel CSV. Three-quarters of clients that use Bloomberg for Simm calculations also use its risk management service—meaning trade data is already in the vendor’s preferred format. For the remaining clients that use third-party systems for derivatives, “we can easily integrate those valuations within our Simm infrastructure,” says Eduardo Pereira, product manager for Simm at Bloomberg.

Buy-side firms that already use Calypso or IHS Markit to process and manage their derivatives trades do not need to provide any additional data to calculate risk sensitivities. Those that use the vendors only for Simm calculations must input the data in a standard format, with some flexibility on how this is done.     

TriOptima did not respond to requests for information about its services.

But these alternative offerings cost more than double what AcadiaSoft charges for IM Risk Generator, with annual fees in the $100,000 to $200,000 range, two vendors tell Risk.net.

A second rival margin analytics vendor says AcadiaSoft’s price tag—comprising a subscription fee and a volume charge—is far too low considering the overheads and complexity involved. 

We could end up with as much as 50% or more of phase five firms using this service, which means having their data in this standard format

Fred Dassori, AcadiaSoft

But Dassori makes no apologies for offering the service as cheaply as possible. “We feel like our mission here is to make it easy for these firms to meet their compliance needs. So we have been very aggressive in the way we price the service. And we’ve tried to make it so that it’s not a barrier to meeting their compliance needs,” he says.

While competitors gripe about its pricing and service model, AcadiaSoft may be poised pull off a major victory. The move to delay the fifth phase of the rules until September 2021 should give the vendor and its clients more than enough time to complete the data transformation work. And with the phase five cohort estimated to include around 300 buy-side firms – with roughly 100 expected to begin exchanging margin soon after coming into scope—Dassori sees an opportunity to cement ORE XML as the data standard for OTC derivatives. 

“If things work out as we hope, we could end up with as much as 50% or more of phase five firms using this service, which means having a sizable portion of the market’s bilateral derivatives data in this standard format,” he says. “Because AcadiaSoft provides access to other vendors, that also opens up significant opportunities to those firms in terms of being able to take advantage of new services with lower up-front costs.” 

That could throw a spanner in the works on another major standardization effort, led by Isda—the architect of Simm. The industry body is developing new trade data standards as part of its common domain model (CDM)—an ambitious project aimed at slashing the cost and complexity of the pre- and post-trade lifecycle for OTC derivatives. But the CDM will come too late for phase five and six firms seeking a silver bullet for Simm sensitivity calculations.

A margin source at a software firm says that if previous experience is anything to go by, AcadiaSoft’s ORE XML format should carry the day. In the absence of common data standards, “vendors end up conforming to some other vendor’s spec, and whoever the bigger vendor is, you end up conforming to them.”

Others remain to be convinced. “I’m not sure if the market is ready to go into one single trade format that will be dictated by AcadiaSoft,” says a director at a third margin vendor. “It sounds like a big step.”