Open Platform: AFT's Michael Bykovsky

May 1, 2006

Gone in 60 seconds

The mortgage-backed securities (MBS) market has grown considerably over the last two years, attracting large numbers of new players looking for clear-cut arbitrage opportunities. Applied Financial Technology's Michael Bykhovsky explains that, although they are fast disappearing, not all of those opportunities are gone just yet

The MBS market players have become increasingly sophisticated, especially in proprietary trading and hedge fund operations, wiping out weaker participants. This has lead to the virtual disappearance of the clear arbitrage opportunity. MBS prices reflect relatively efficiently the differences in prepayment propensities of loans with different ages and coupons. Differential pricing of other characteristics such as loan size, geography, purpose, FICO, LTV (loan to value ratio) and property type has been haphazard. This almost always leads to mis-priced variability in the intrinsic value of the mortgages. The differential propensity to refinance represents an enormous arbitrage opportunity, one of the few still remaining in the mortgage market. The variability in the economic values of otherwise similar interest-only (IO) strips can be as high as 30%.

A deal may be backed by hundreds of pools, or thousands of loans. Their detailed characteristics are generally not available through analytical systems and even if they were, evaluation of the effects of the distribution of individual tranches is extraordinarily difficult.

The answer is in prepayment scores that summarise all loan characteristics, other than age and WAC (weighted average coupon) and their effect on loans' differential propensity to refinance or to move. The scores are used as modifiers to a generic prepayment model, which is designed to accurately project aggregate behaviour of these loans, without taking into account borrower and loan characteristics other than loan type, age and WAC. The algorithm for the score is developed using historical loan prepayment experience versus the generic model projections.

For investors, the prepayment scores simplify the process of loan valuation in relation to prepayment propensity in the same way FICO credit scores simplify loan valuations in relation to their default propensities. The prepayment scores do not change with movements of interest rates and implied volatilities, whereas the values of MBS prepayment options do. Now, loans with similar coupons, ages, and scores have similar economic values.

Dramatic differences

The graphic opposite demonstrates the ability of the scoring process to discriminate for refinancing propensity. All conforming loans originated in 1997 with WAC of 7.5% (+-0.25%) were scored and then bucketed into deciles by this score. In the absence of refinancing incentive – until mid-1998 and between mid-1999 and January 2001 – these buckets prepaid at similar speeds, but when the interest rates dropped, the higher scored buckets always prepaid faster. The difference was dramatic. In 1997, these loans were priced identically, even though their economic values were clearly different. Given the interest rates and volatility environment in 1999, the calculated economic values of these loans differed by as much as 3% – the value differential was clearly realised when the interest rates did move.

For this to be successful, all agency pools must be scored based on additional pool information like geographic distribution, loan size, FICO, LTV and purpose. The score's ability to discriminate pools' prepayment behaviour is equally dramatic. When calculating differential economic values of a variety of pools, utilising a score-modified prepayment model, and comparing them to market pay-ups for low loan balance pools, little relationships can be seen between the market pay-ups and the intrinsic economic value of that additional information.

Further extensions can be made to agency and non-agency CMOs (collateralised mortgage obligation). The scores of the agency CMOs can be calculated based on the pool scores that back them and non-agency CMOs on the scores of the loans that back them. AFT maintain databases of pools backing all-agency CMOs and of loans backing 80% of non-agency CMOs. Most analytical systems do not provide pool-level or loan- level information on the collateral backing the deals. The combination of scoring and databases allows traders to access the information and efficiently use it.

Using the technology

There are a number of ways in which market participants are starting to use the technology to take advantage of the arbitrage. Several proprietary trading desks and hedge funds have already reorganised their trading activities around the scoring algorithm. Their trades now involve a long position in a low-scored IO and a high-scored payment-only (PO) strip of the same collateral type, age and coupon and a short position in TBAs (to be announced). The trade removes most of the market risk, allowing them to benefit from the desirable score-based prepayment behaviour and gradual market recognition of the differential behaviour of different legs of the trade.

This trade can take a variety of other forms involving more complex combinations of MBS derivatives. Some trading desks have been using the algorithm to deliver lower valued – on the basis of score – collateral against TBAs, or to cherry pick specified pools when the market does not price the economic value of the score accurately. Mortgage bankers have started cherry picking within their portfolios by selling off higher scored loans and looking to acquire lower scored ones.

The score makes it possible to immediately identify desirable MBSs without having to perform a loan-by-loan analysis of the collateral backing it, exploiting the last remaining large arbitrage opportunity. Without the score, analysis is almost impossible due to computational and data constraints, limitations of the analytical systems and the general unavailability of the models that can make use of all the varying data contents. The score also makes it possible to communicate desirable collateral characteristics; an otherwise nearly impossible task since it involves specifications of combinations of more than seven criteria.

Michael Bykhovsky is president and CEO of Applied Financial Technology, an analytic software firm providing mortgage banks and MBS traders with mortgage loan prepayment propensity scoring.