Uncategorized
FICO drives delinquencies in CRT
admin | July 17, 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.
More than any other loan attribute, FICO appears to be the leading indicator of serious delinquency in loans covered by Fannie Mae and Freddie Mac’s Credit Risk Transfer programs. Pools with larger populations of lower FICO borrowers may underperform. And while risk of default and liquidation may ultimately be low, these loans potentially bear elevated modification risk where, even absent a default, investors may bear losses.
Trying to paint a picture of delinquency trends in CRT can be challenging since performance data is lagged by one to two months depending on the program. An early read on delinquencies nevertheless shows that the populations of seriously delinquent loans generally exhibit significantly lower credit scores than loans that are current or in earlier stages of delinquency. As of the most recent remittance data roughly 93.5% of loans referenced in Freddie Mac STACR deals are current pay while roughly 95% of loans referenced in Fannie Mae CAS transactions are current. The difference may be attributed to a host of factors but one material one may be the fact that Fannie Mae lags their reporting by one more month than Freddie Mac which likely does not capture an incremental month of Forbearance uptake in May in Fannie Mae loans.
Looking across the admittedly delayed picture of delinquent loans referenced in CRT transactions shows that across both the Fannie Mae CAS and Freddie Mac STACR programs, somewhat unsurprisingly, lower FICO loans make up a disproportionately large amount of late stage delinquencies, potentially suggesting that credit score may ultimately be the primary driver of borrower performance in these pools. Loans with less than 700 FICOs make up over half of Fannie Mae CAS loans that are more than 60 days delinquent and roughly 45% of STACR late stage delinquencies. Seriously delinquent loans in Fannie Mae’s CAS program have a weighted average FICO of 702, roughly 50 points lower than current pay ones across the program. Similarly, seriously delinquent loans in Freddie Mac STACR deals have a weighted average credit score of 706, also roughly 50 points lower than the current pay cohort. The somewhat obvious caveat here is that given the lag in reporting loans that are more than 60 days past due would have initially fallen past due in February in CAS deals and March in STACR deals where the impact of COVID-19 on the economy was relatively muted. But with that said, it appears the value of FICO in forecasting borrower performance may be more predictive than other loan attributes across CRT (Exhibit 1).
Exhibit 1: Serious delinquencies marked by lower FICO loans in CRT
Source: Fannie Mae, Freddie Mac, Amherst Pierpont Securities
Unsurprisingly, serious delinquency rates of loans in COVID impacted states are elevated relative to their overall exposure in the deals. New York based loans 60 or 90 days past due in STACR transactions make up nearly 10% of those serious delinquencies, more than double their representation in the broader cohort. NY based serious delinquencies in CAS transactions are less pronounced but this is likely driven by the incremental reporting las CAS deals as they make up nearly 9% of 30-day delinquent loans in those deals. Concentrated catastrophic risk is by no means a new phenomenon to CRT as investors have historically been exposed to forbearance related delinquencies associated with hurricanes, flood and wild fires. Deals with the largest exposures to NY are generally seasoned transactions or those backed by off the run HARP collateral. Historically, losses associated with natural disaster related forbearance have been negligible but one concern which may weigh on CRT investors as the seasonal impact of natural disasters rapidly approaches is how any potential catastrophic risk may further impair borrower performance that has already been compromised by the pandemic.
Other loan attributes such as the borrowers’ combined LTV, debt-to-income (DTI) ratio or the size of the borrowers loan balance appear to be more muted drivers of serious delinquency, at least to date, as the disparity between those attributes on current pay loans versus delinquent ones appear to be modest. Seriously delinquent loans in both CAS and STACR transactions have combined LTVs in the mid 80s, roughly in line with current pay ones. Seriously delinquent loans in both programs have modestly higher debt-to- income ratios than current ones in both programs but are less than 5% higher than the average DTI of current ones. And average loan balances on delinquent loans are only slightly higher than current ones. Given this, it appears that credit score may ultimately be the best predictor of serious delinquency.
Assuming this to be the case, all else equal, investors may look to concentrate their exposure to CRT in pools that have a smaller tail of lower FICO loans. Excluding already delinquent loans from the analysis, and filtering for the percentage of reference loans with FICOs less than 700 shows that deals backed by HARP collateral and certain 2019 vintage on the run deals have the largest tails of loans with lower FICO scores. The combination of outsized exposure to New York coupled with lower FICO scores may drive underperformance in deals backed by HARP collateral relative to other on the run transactions. (Exhibit 2)
Exhibit 2: Scanning the CRT universe for low FICO loans
Source: Fannie Mae, Freddie Mac, Amherst Pierpont Securities
Additionally, lower FICO loans in HARP transactions have some of the highest average incremental risk based pricing or SATO relative to other pools. Investors will likely become increasingly attune to the loan rates of seriously delinquent loans in CRT transactions. Given uncertainties about the future path of the pandemic, it is challenging to predict what the ultimate rate of both default and loss given default on CRT transactions may be. And history in all likelihood will not be a great predictor of future performance as CRT borrowers that were impacted by a natural disaster likely benefitted from a faster ‘V-shaped’ recovery in affected areas, while the economic impact of the pandemic may be much more ‘U-shaped’ in nature. Even if ultimate default rates and losses are small, investors will still bear any rate modification risk to these pools. Given this, certain HARP deals and 2019 vintage transactions with higher WACs may have outsized risk to rate modifications than other deals with comparable exposures to lower FICO loans.