More woes for Countrywide

admin | January 30, 2019

This document is intended for institutional investors and is not subject to all of the independence and disclosure standards applicable to debt research reports prepared for retail investors.

A recent action taken by Bank of New York as trustee on 278 Countrywide trusts could materially affect valuations of both interest-only and principal and interest bonds in those deals. The action could also be a potential headwind to the affected trusts being called.

Despite some of these deals having been outstanding for more than fifteen years, a lawsuit was filed by a single plaintiff in December claiming that Bank of New York as trustee has miscalculated payments owed to interest-only classes of 156 contested trusts. The plaintiff seeks to not only change the trustee’s methodology for calculating future payments but also to recoup prior differences in the disputed calculation methodology. Changing the future cash flow could negatively impact prices on principal and interest bonds in groups with affected interest-only classes. Recouping past lost interest would curtail available cash flow to the deals, which, if large enough, could have a similar impact on the contested trusts as a ‘servicer first’ scenario, potentially shutting off cash to the rest of the deal until the cash flow due the interest-only classes are recouped in its entirety.

The issue at hand is whether the trustee should be calculating the monthly cash flow due to the interest-only bonds based on a set of loans initial or current coupon. This may sound like a fairly straightforward proposition, but like many issues in the legacy market can be maddeningly complex. Through the magic of ratio-stripping, certain loans were tagged as ‘non-discount mortgage loans’ and the interest from these loans could be directed to interest-only, principal only or principal and interest bonds either in part or whole. These ‘non discount mortgage loans’ carry a minimum coupon which varies deal to deal. In the disputed trusts, the cash flow due to the IO is equal to the balance of these loans times their weighted coupon less the designated ‘non-discount’ coupon.  Easy right?

In the words of College Gameday’s Lee Corso, “Not so fast my friend.” As collateral performance deteriorated, many of these ‘non–discount’ loans got modified well below the ‘non-discount’ coupon leaving no excess interest to pay to the IO class. The plaintiff is asserting that irrespective of modifications, the interest due the IO classes should be calculated off each loan’s initial rather than current rate. Given the lack of requisite coupon to support that payment, affected trusts would have to use principal to pay interest, under-collateralizing the trust and pushing losses higher. The plaintiff invoked precedent that Deutsche Bank as trustee has used the initial coupon to calculate interest payments on interest-only bonds on the RAST shelf, whose operative documents contain similar language. In response to the suit, Bank of New York has filed an Article 77 petition in New York State Supreme Court asking the court to decide whether the interest paid to the IO class should be based off the loans’ initial or current note rate on not only the 156 contested trusts but an additional 122 Countrywide trusts where BNY acts as trustee.

To say quantifying the impact of the suit is challenging is an understatement. Sizing and pay rules for interest-only classes vary deal to deal. IOs can be stripped off an entire deal, a single group or a single bond. Additionally, the ‘non-discount’ coupon can vary. And the proportion of ‘non-discount’ mortgage loan interest allocated to the IO class can vary as well.  As a result, the impact of changing the calculation for both past and future payments will impact each trust differently. With that said, some high level assumptions can be made to see which of the 156 contested trusts could potentially be hit the hardest if the plaintiff were to effectuate their claim.

Calculating the monthly cumulative interest shortfall between the initial and current note rate on modified loans in each period in each of the affected trusts and comparing that cumulative interest amount to the outstanding collateral balance should provide a reasonable albeit imperfect guidepost to what deals may be poised to underperform the most. The caveat being that this methodology will by definition overstate the impact for two reasons. First, this methodology assumes an interest ‘shortfall’ that is the difference between the original note rate and the modified note rate in a given period rather than the difference between the original note rate and the ratio strip rate that the excess IO coupon would be struck off of. For example if a loan was modified from 7.0% to 4.0% with a ratio strip rate of 6.0%, we would calculate the ‘shortfall’ at 3% rather than 1%. Additionally, we are looking at all modified loans in affected trusts when in point of fact many of those modified loans may not have been allocated to the ‘non-discount’ ratio strip group.

Exhibit 1: Past & future modified interest ‘shortfall’ as a percentage of UPB (top 25 affected trusts)

Source: Amherst Insight Labs, Amherst Pierpont

In terms of quantifying the total past and future magnitude we estimate that recasting the payments back all modified loans’ in affected trusts would generate 36 bps of weighted average excess coupon and nearly $1.8 billion in past interest. Fundamentally, this overstates the impact because this is the difference between the initial and current rate rather than the initial and ‘non-discount’ loan rate. Changing the methodology going forward could push an additional $1.1 billion to IO classes assuming a current 1.47% difference between initial and modified coupon rates, average model duration of five years and a collateral balance of just under $12 billion.