Sell CRT B1s and rotate into other mortgage credits
admin | June 12, 2020
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.
The recent rally in credit risk transfer securities has seen some bonds doubling in price in less than a month. While there are reasons to be optimistic that COVID-19 related defaults and losses may be contained, bonds further down in the capital structure may have limited upside at this point, even to fairly rosy scenarios, and likely have much greater downside to harsher stresses. Sell CRT B1s and rotate into other areas of mortgage credit.
A sharp rally in CRT
The recent risk on trade in mortgage credit has been most pronounced in certain corners of the GSE CRT market. Certain late vintage B1 classes have roughly doubled in price over recent weeks in spite of an absence of timely, specific performance data to support what appears to be a fairly seismic shift in investor sentiment about the potential tail of virus-related defaults and losses (Exhibit 1).
A confluence of factors has likely fueled the rally:
- Faster than anticipated prepayment speeds over the past two months
- A slowdown in forbearance uptake by borrowers
- Relatively robust housing data
- The enterprises’ announcement that debt burdens on borrowers coming out of forbearance will likely be eased as missed payments can be deferred to maturity without accruing interest, and
- Attractive carry at current valuations given the lag in defaults and losses associated with loans currently in forbearance
Exhibit 1: Month-over-month percent price change across CRT B classes
There is still significant uncertainty around the impact of COVID-19 on GSE borrowers. Terminal rates of unemployment will undoubtedly be higher, and it’s unclear what the population of those unemployed may look like. A second wave of outbreaks could drive an additional spike in forbearance. Additionally, fast prepayment speeds that serve to deleverage these pools and shorten the deals may in fact be pulling forward the tail of adversely selected borrowers still remaining in the pool as better borrowers prepay.
A look at current upside and downside
It makes sense to look at both where relative value may lie in base, bull and bear scenarios. It appears most of the B1 universe would have asymmetric downside to a material deterioration in fundamental performance.
It appears that the biggest potential unknown in CRT is the rate of borrower default after forbearance. The market has likely taken some comfort in strong borrower performance after exiting forbearance in the wake of recent natural disasters. However, borrowers affected by COVID-19 may not exhibit similar performance. The virus has longer lasting and more widespread effects.
The analysis assumes a baseline 2% monthly delinquency rate for six months for an average loan in forbearance and a 25% roll to liquidation assumption. Default rate multipliers overlays are then applied for attributes such as loan size, geography, FICO and LTV, derived from recent specified pool performance data. Based on these assumptions, later vintage high LTV transactions appear to offer the most nominal spread while more seasoned low LTV deals offer the least. Deals that trade the widest look to offer roughly 200 bp more spread than those that trade on the tighter end of the range. (Exhibit 2)
Exhibit 2: Later vintage high LTV B1s offer the widest nominal spread
Wider spreads may be a function of elevated uncertainty with regards to the potential future performance of late vintage high LTV borrowers. If these borrowers exit forbearance and are unable to resume their previously scheduled monthly payments, they will likely need a permanent or flex modification. Given the modification waterfall employed by Fannie Mae and Freddie Mac, higher WAC, high LTV 2019 vintage borrowers may be subject to rate modifications. The interest shortfalls between the original and modified rate would flow through as a principal loss to the junior-most bond in the capital structure on a monthly basis. And these higher WACs in these pools that been generally viewed as a positive to bondholders as they have been the driver of elevated prepayment rates, may turn out to be a credit liability if a large swath of these borrowers ultimately receive permanent modifications.
Holding spreads constant and shifting the roll-to-liquidation assumption to just 10% of borrowers who go into forbearance has a fairly negligible impact on most of the universe and only has an outsized impact on certain off-the-run classes, which the model sees as already trading nominally tight at current valuations. That suggests that absent forbearance and delinquency rates levelling out over the next few months, CRT B1 classes may have already traded back to their full fair value. Conversely, increasing the forbearance assumption to 3% per month for six months with a 50% roll-to-liquidation shows the entire universe has asymmetric downside to higher forbearance and liquidation assumptions. (Exhibit 3)
Exhibit 3: The CRT B1 universe appears to have limited upside at current prices
This asymmetry suggests that absent forbearance and delinquency rates levelling out over the next few months the overwhelming majority of the CRT B1 universe may be trading at or near fair value, with limited upside. Other areas of mortgage credit may offer a more favorable balance of upside return and downside risk.