Throughout the history of the secondary mortgage market, the “To Be Announced” (TBA) mortgage-backed securities (MBS) forward contract has enjoyed a very high level of liquidity and, as a result, has required relatively few data inputs for efficient market pricing. However, this dynamic has changed for a variety of reasons, and there is now far more granularity in the pricing of the mortgage asset. Loan-level data is now essential to have, but data-driven bid tape execution will continue to dominate the loan sale process into the future. The resulting increased level of complexity in pricing of the asset absolutely requires a robust and educated best execution process.
An essential characteristic providing the foundation for this liquidity has been the generic nature of the TBA vehicle, and as such, the market pricing has been determined, with relatively few data fields and average market prepayment speeds, given the coupon and product. There have historically been some exceptions to this rule in which bond investors will pay up for securities that have high concentrations of loans that can be expected to prepay at a far slower rate than the overall market average.
An example and the most well-known type of mortgage concentration pay-up is that of low loan balances. The simple rationale here is that, given the fixed costs that exist in refinancing a mortgage, a borrower with a small loan balance will be less likely to prepay his or her loan, given a drop in market rates, compared with a borrower with a larger loan balance.
As such, bond investors will pay more for this bond due to the lengthened expected enhanced cashflow versus a non-callable bond. And due to the relatively large unit count of loans needed to securitize a liquid pool of loans, these bond market pay-ups were generally reserved for very large lenders and aggregators. Because these specified pay-ups do not have an associated TBA-forward contract, the value is unhedgeable. As such, in general, these pricing enhancements were not propagated to the respective institution’s front-end pricing and, instead, simply resulted in back-end profits for the lender or aggregator.
However, over the last two years, the rapid proliferation of data-driven execution of closed loan collateral has enabled aggregators to price loans far more efficiently than ever before. The short term and certainty of delivery aspects of bidding on closed loans have allowed aggregators to safely pass through these specified pickups directly to the lender.
Additionally, Fannie and Freddie are now pushing a variety of spec pay-ups through to their respective cash window executions. This transmission of the securitization-level specified pay-up, all the way down to the loan level, has dramatically leveled the playing field with regard to these pricing enhancements.
More importantly, smaller lenders are now put in a position to price some of this enhanced pricing into their front-end pricing. The natural result of this is that there will be an increase in the optimization of loan collateral placement within securities, and as a result, the market prepayment speeds will increase and their resultant value will further deteriorate relative to the specified pay-up security. At our firm, we believe that this increased pricing efficiency, as well as some of the correspondent market dynamics that will be explained, will only accelerate the move to data-driven bid tape execution.
Around the same time that the specified security pay-up field was being leveled, dramatic changes were occurring in the correspondent mortgage market. Over the last three years, the mortgage industry has seen an enormous influx of aggregator whole loan buyers. This was initially caused by a quick rise in rates in summer 2013, which many refer to as the “taper tantrum.” This rise in rates and resulting slowing of prepayment speeds caused mortgage servicing rights (MSRs) to, once again, become an in-vogue asset to hold. The quickest way to buy large amounts of newly originated mortgage servicing quickly was to enter the correspondent investor space.
In an attempt to minimize any hedging complications to their secondary departments, many of these new investors initially entered the market by only buying closed loan production on a mandatory basis. Additionally, due to an initial lack of requisite broker-dealer assignment of trade approvals, in some cases, bid tape execution was the quickest and safest method to rapidly begin buying loan production on a live basis. As an incentive for sellers to negotiate these takeouts into their best execution analysis, these new buyers all provided loan-level pricing execution and did not require any large volume minimums. These buyer accommodations quickly succeeded in attracting a large number of lender-sellers to incorporate these additional executions into their investor set.
A survey of today’s correspondent investor execution environment shows that the vast majority of whole loan investors not only accept, but also recommend the bid tape process as the best method of achieving their highest execution levels. There are some market participants that continue to expect bid tape execution to fade in significance, as the number of whole loan buyers falls back to historical levels. Many also believe that in addition to cashflow simplification and trading cost benefits, the efficiency of the assignment of trade method will outweigh the whole loan granularity benefit of the bid tape.
We strongly disagree with this perspective. Due to the minimal amount of information transfer at the time of commitment, an efficient direct trade or assignment of trade program relies on a fairly generic asset. As we reviewed at the outset, from the buyer’s perspective, myriad pricing factors, such as TBA spec pay-ups, the Community Reinvestment Act and MSR valuation, are only accelerating the path to further granularity in pricing.
As previously discussed, a prime example of this is the recent agency cash window differentiation of conventional pricing based on loan balance, occupancy, FICO and even state. Additionally, it is widely believed that the Common Securitization Platform proposed for 2018 will only further perpetuate pricing granularity within the resulting TBA. The basis for this thinking is that the competing agency policies, underwriting and culture could create further variation prepay speeds. As the lender’s ability to efficiently transmit accurate and real-time pipeline data to the prospective buyer continues to evolve, we can expect bid tape execution methods to continue to dominate.
A fundamental precept of any robust best execution analysis is that it must be able to accurately compare and analyze all net flow executions against loan-level bid tape pricing results. This requires secondary marketing managers to account for more factors in their best execution.
One important additional consideration that is needed when accommodating bid tape results into a best execution analysis is collecting and adjusting for accurate TBA MBS market spots. Live executions such as direct trade or agency cash window pricing can be generated through live TBA data feeds and automated Web crawlers, respectively. However, bid tape execution will come from the whole loan buyer and generally will be based on a given market level. Many experienced aggregator trading desks will provide these TBA market levels at the time they produce their bid tape pricing levels. Market adjustments must be then incorporated to properly compare across all execution types. At the time of actual commitment, these must be reconciled against the live executions as the bid is refreshed with the desk. In a moving market, this is an extremely complicated and time-sensitive process.
Another necessary consideration when comparing a large number of investor executions is the variation in funding or administration fees. Certain new aggregator and co-issue entrants have set their funding fees substantially higher than historical norms. This is paramount in all best executions involving aggregator versus agency and, as a result, retain versus released decisions. Most MSR valuation models will have incorporated an estimate of servicing file boarding costs into the net present value, and as such, the funding fee of the released execution must be accounted for in order for the analysis to be a correct comparison.
One note about the investor funding fee analysis: Due to the fact that the majority of pricing engines cannot accurately factor funding fee differentials into their scenario best execution, net front-end margins between investors will not be consistent. To mitigate this issue, we recommend taking the investor fee differences by product, weighting by that product’s average loan amount and then adjusting the specific investor margins accordingly.
Although considerations such as market spots and file fees do not come close to exhausting the list of essential factors that need to be accounted for in a robust best execution analysis, we use them to illustrate the complexities that are created from the new data-driven world of secondary execution. Only when all necessary considerations are accounted for and bid tape pricing is properly incorporated can a true best execution actually be achieved.
Phil Rasori is chief operating officer at Mortgage Capital Trading Inc., a hedge advisory firm and developer of secondary marketing software. He can be reached at firstname.lastname@example.org.