Last week's head-fake selloff and subsequent rally have put MBS prices back in sight of their early-February (and all-time) highs.  This in turn has forced investors to look for ways to improve their investment results and carry-adjusted returns.  Research over the last decade has demonstrated that a number of attributes result in improved prepayment performance-in a market where the weighted average dollar price is in the area of 104, this means that investors want MBS with slower prepayment speeds.  This can accrue to the benefit of forward-thinking originators that can capture at least some of these payups in their execution.

There are a number of loan attributes that tend to result in slower prepayments and/or reduced responsiveness to refinancing incentives.  The best-known attributes are loans with smaller loan balances and impaired credit.  Investors will pay significantly more than TBA prices for pools backed by loans with these characteristics.  As a result, the market for agency MBS has created tiers in order to trade the variety of products being created; pools backed by these loans trade as "specified pools" (or "spec pools") away from the TBA market. 

There are a number of reasons that smaller loans tend to prepay relatively slowly.  Aside from the generally lower incomes and financial strength of the borrowers, small loans have to overcome significant hurdles in order to justify the fixed expenses involved with refinancing.   For example, a borrower with a $230,000 loan can save about $67 a month by refinancing into a loan with a 50 basis point lower rate.  If the total fixed costs of refinancing are $1,000, the borrower breaks even in roughly 15 months.  Borrowers with an $85,000 loan, by contrast, would only reduced their P&I payment by $25/month, meaning that it would take over 3 years to recoup their expenses.

Other specified pool buckets have formed around credit characteristics.  Loans with weaker credit (such as low FICO scores and/or high LTVs) have long tended to pay relatively slowly, since these borrowers have difficulty in getting new loans.  (Their prepayment advantage is even more pronounced in a tight-credit environment.)  There are also specified pool cohorts for characteristics such as "brand new loans."  Since borrowers tend not to refinance their loans immediately upon closing, these pools also tend to prepay very slowly for a number of months after issuance.  Finally, there are other new programs (such as MHA/HARP) that also trade at a premium to TBAs.

 "Generic" MBS (i.e., loans without any definable prepayment advantage) are delivered directly into TBAs.  A variety of separate segments have developed, most commonly by loan size.  The most commonly traded tiers are for loans with maximum pool balances of $85K, $110K, $125K, $150K, and $175K.  The key element is that these pools define the maximum loan balance in the pool, not the average.  This improves the performance of the pool by limiting the size dispersion within the pool. 

The different segments are generally priced by coupon and attribute, and (at current prices) are at very strong levels.  Even some of the historically muted prepayment stories, such as maximum-$150K conventionals, are trading extremely well.

The key for originators is to monetize the incremental value of their spec pool loans.  While some correspondent lenders will pay up for some loan attributes, most originators are forced to sell their production at generic levels.  In most cases, selling loans in order to garner some of the market payups for the loans means that originators need to have the authority to create their own pools and hold the servicing for the loans.  Originators cannot maximize their revenues without taking advantage of these payup opportunities.