Raw Finance

In a webcast on Nov. 12, 2009, Fitch Ratings analysts noted that some recoveries in distressed residential mortgage-backed securities (RMBS) may be substantial.  Fitch’s webcast reviewed its new Recovery Rating system for residential mortgage-backed securities (RMBS) for investors and other interested parties.

The Recovery Rating will be issued alongside the more commonly known long-term rating.  The long-term rating system grades bonds by a letter rating system from D at the bottom to AAA at the top.  A long-term rating of “B” or better indicates no expectation of loss.  A rating below “B” generally means an expectation of at least one dollar of loss.  For those used to hearing terms like “investment grade” or “speculative grade” (sometimes derogatively, or lovingly, referred to as “junk”), these are industry-made terms.  However, to translate: “investment grade” is typically a Fitch rating of “BBB” or higher (to “AAA”); and “speculative grade” is typically a Fitch rating of “BB” or lower (to “D”).

When the credit crisis hit last year, most RMBS bonds were downgraded all the way to the CCC distressed level.  Since it is widely recognized that some loss will occur, the question for each bond became, how much?  The answer is not only important to structured finance investors, but also banks, financial regulators and economists trying to gauge the damage from the financial crisis.

The Recovery Rating is Fitch’s attempt at that answer.  In addition to the long-term rating, Fitch now displays the Recovery Rating, which is a number between 1 and 6.  Each number corresponds to a range of percentile recoveries.  For example, a Recovery Rating of 2 corresponds to an expected recovery of 71 to 90 percent of the par value of the bond; a Recovery Rating of 5 corresponds to an expected recovery of 11 to 30 percent.

There are several factors that go into calculating the Recovery Rating (and for those who would like more detailed information, I have included a link to the webcast with slides on Fitch’s website at the end of this column).  Two factors that I found particularly relevant are: (1) the recovery percentage includes future cash flow on the performing assets that comprise the bond; and (2) the baseline scenario assumes an unemployment rate of 10.3 in mid-2010 and a further price decline in housing of 10 percent (as a national average).

To determine the expected recovery percentage, Fitch calculates both the amount of principal likely to be returned and the interest likely to be collected until the maturity date.  Adding these figures together equals the total recovery.  Fitch then applies a 10-percent discount rate to the total recovery to determine the present-value of the recovery (the 10-percent rate was determined by Fitch as a simple, yet conservative standard measure).  The present value of the total recovery divided by the par value of the bond is the expected recovery rate.  Fitch provides an example of this calculation in the slides accompanying its webcast.

This seems to be a fair and transparent calculation.  I would only point out that it relies on continued performance (cash flow) of a portion of the underlying assets and so, investors and anyone else relying on this information need to keep in mind that any further deterioration in performance will affect the recovery rate.

Performance on the underlying assets (housing) relies in part on employment.  This is why the baseline assumption of a 10.3 percent unemployment rate bears watching.  While I generally agree with that assumption (even though the U.S. is currently at a 10.2 percent unemployment rate, I do not believe it will significantly worsen from here), some economists have warned of the rate reaching 12 or 13 percent next year.  Should the worse scenario come to pass, the assumption under which Fitch believes “B”- and higher-rated bonds would not likely sustain losses would seem to require adjustment, which in turn would threaten more bond classes with downgrades below the “B” rating.  Also, Recovery Ratings would likely need to be adjusted for distressed bonds.

I asked Fitch whether the unemployment rate was a concern for them.  After the webcast, Grant Bailey, who participated in the webcast, responded in an e-mail message, “Yes, if the unemployment and economic situation turned out to be materially worse than we’re currently projecting in our ‘B’ case, it would likely result in a revised ‘B’ mortgage pool loss estimate and rating downgrades.”  However, Mr. Bailey explained, “The employment situation is just one component. [Whether a revision and subsequent rating downgrade would be undertaken] would obviously depend on home prices, the effectiveness of loan modifications and mortgage behavior. But significantly higher unemployment than what is currently assumed would likely put negative pressure on ratings.”

It would seem then that, in the case of RMBS, the unemployment rate is not a lagging indicator (as it is with regard to future economic growth or contraction), but rather a possible leading indicator.

In the webcast, Fitch analysts generally noted that the average recovery rate for below “B”-rated RMBS bonds in Prime mortgage is expected to be 90 percent.  The expected average recovery rate drops to 80 percent for Alt-A RMBS.

I also inquired as to whether federal government efforts to encourage lenders to modify residential mortgages that are either in default or in danger of default have impacted Fitch’s work on RMBS.  In an e-mail response, Grant Bailey noted two effects on Fitch’s analysis, “First, [the Home Affordable Modification Program] has resulted in lower mortgage coupons for many of the loans which affects excess spread in some Alt-A structures and most subprime structures.  Second, it has significantly affected the timing of losses by delaying liquidation, so we’ve adjusted our loss-timing assumptions to try to account for that.”

Please click the following link to view the entire webcast presented by Fitch Ratings: Low Ratings, High Recoveries.