A Conversation on the New MBS Central Counterparty
by James Conmy
Murray Pozmanter, DTCC managing director, Clearance and Settlement Product Management
DTCC's Fixed Income Clearing Corporation (FICC) subsidiary continues to move forward on its multiyear initiative to build a central counterparty (CCP) for the mortgage-backed securities (MBS) market. The project, which will reduce cost and risk to the industry, represents the biggest change to the processing of these securities in more than 20 years. It will allow FICC to become a CCP to eligible transactions,bringing the MBS market under the umbrella of the CCP services that DTCC already provides for U.S. government securities, equities and the corporate and municipal debt market.
To gain an understanding of the MBS CCP initiative, how far along it has come and how it will impact the industry, @dtcc spoke with Murray Pozmanter, DTCC managing director, Clearance and Settlement Product Management.
FICC has been working to build the MBS CCP for several years now. How far along is it?
Our goal is to have the pool netting component of the new CCP fully up and running before the end of 2008, and we are rolling it out in phases.
The first step - matching specified pool trades through FICC's Real-Time Trade Matching (RTTM) system - underwent customer testing earlier this year and began operation on June 22. The second step is currently in customer testing. Its aim is to make it easier to cancel an existing mortgage pool that a firm chooses not to make delivery of - and substitute a different pool - using only a single message in FICC's Electronic Pool Notification service. The third step, mortgage pool netting, is scheduled to begin testing in the first quarter of 2008.
The mortgage-backed market has some unique characteristics that have influenced the design of the new CCP. Could you talk about this?
The mortgage pass-through product differs from the other products for which DTCC serves as CCP. In the government market, in the equity market, even in the corporate market, it's relatively simple. The trade is executed. It's netted exactly as it's traded. It's settled exactly as it's traded. And it's one process. It's a T+1 environment for governments, a T+3 environment for equities.
Our goal is to have the pool netting component of the new CCP fully up and running before the end of 2008, and we are rolling it out in phases.
In the mortgage-backed market, you trade much further forward and you don't trade exactly the same instrument you will settle. Rather, you trade "to-be-announced," or TBA, contracts, and you don't exchange information on what's going to be delivered until 48 hours prior to settlement. As a result, mortgage trading and settlement is a two-step process. It consists of TBA trading followed by the actual allocation of pools that are going to be delivered.
The existing mortgage-backed Settlement Balance Order netting model takes place at the TBA level, reducing the value of TBAs to be settled by 95%. The resulting obligations from the TBA net - the remaining 5% - then have pools assigned to them that are used for delivery. The CCP will provide multilateral netting and novation for the eligible pools, so in effect, it will provide "double netting."
How will the addition of pool netting benefit the firms that trade in this market?
As I said, the existing netting process reduces the number of TBAs by 95%. Still, on a "class A" settlement day, when all the current coupons for Fannie [Federal National Mortgage Association] and Freddie [Federal Home Loan Mortgage Corporation] 30-year products settle, approximately 300,000 mortgage pools pour through our systems. So that leaves ample room for improvement. The new central counterparty will enable us to capture a good part of that remaining 5%, which will further reduce operational risk and cost for the industry.
How much of that 5% will you capture?
Based on our analysis of the numbers, we expect to reduce the overall remaining settlements by around two-thirds. So combined with the initial netting of 95% of TBAs, FICC will be netting more than 98% of mortgage-backed trades submitted. That's on par with the netting factor achieved by NSCC [National Securities Clearing Corporation], DTCC's clearing subsidiary for equities.
Still, the Mortgage-Backed Securities Division [MBSD] will not net all the pools. Is this a change from earlier plans?
Yes. It's a change prompted by customer feedback and careful analysis of the numbers. In the early planning stages, we anticipated stepping in as the CCP at the pool level to take delivery of all pools, whether or not they had a large netting factor.
But when we analyzed the costs and talked to member firms, we realized it didn't make sense for FICC to step into the middle of pool deliveries for which the netting impact was minimal, if at all. Remember, the clearing costs FICC will incur as a counterparty to the trades will be passed back to members in the form of fees. So if we step into a trade that does not have a netting factor, the clearing cost will double without a corresponding benefit. In other words, we'd be taking a two-sided settlement between two of our participants, and turning it into a four-sided settlement with FICC in the middle. So we recently took a step back to re-think this issue and arrived at the concept of using algorithms to determine which pools would benefit from netting.
How will that work?
Every time we run the pool net, we'll use an algorithm to calculate the potential netting factor and the volume in the system for a particular pool so that we can see which pools will benefit the most from netting. The current thinking is that we'll only net those pools that have at least $20 million in par value, and at least a 50% netting factor.
The remaining, less active pools will not be netted. Instead, they will continue to be settled between the parties assigned through the Settlement Balance Order process, which is how they are handled today. [See related article for more on pool netting.]
What are customers doing to prepare for the implementation of the MBS CCP?
We've been working closely with customers to conduct testing for each stage of the rollout.
Throughout this project, we have made every effort to minimize the requirements for our customers. One way we've done that is to take a back-to-front approach to build the CCP.
Rather than change the existing Settlement Balance Order model for TBA netting, we started at the point in the process where the resulting pools are actually settled, and are working forward from there. In other words, we've started with the pool obligations that support the securities being traded, taking the actual information that's going to be delivered and netting it.
It may seem counterintuitive, but doing it this way enables firms to make the fewest possible changes to the piece of the model that already is up and running, while receiving the maximum benefit at the back end of the process.
What kind of cost savings can firms expect from the MBS CCP?
Overall settlement expenses will decline. We expect a reduction in deliveries and in fail rates, and there's also the intangible benefit associated with lower operational risk.
In addition, the proposed common margining for customers that are members of both MBSD and the Government Securities Division (GSD) of FICC will result in savings because these customers will gain the ability to offset positions in one division against their positions in the other, and thus have a lower margining requirement. As you know, firms often pay for financing on margin posted to FICC, so to the extent they can lower their margin requirements by offsetting the GSD positions against their MBSD positions, they will realize savings. @