By the Numbers
Improving fundamentals, rising issuance in non-QM
Chris Helwig | August 13, 2021
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 non-QM market seems to be on firm footing in the wake of pandemic. Delinquency rates on loans backing non-QM trusts are rapidly approaching pre-pandemic levels. And more recent issuance has been marked by improvement in the attributes of loans backing these trusts. Strong fundamentals along with high advance rates, a low cost of funds and apparently favorable economics for sponsors of securitizations should fuel increasing issuance.
The non-QM market has proven resilient in the wake of the pandemic. Overall non-QM delinquency rates stood at 4.5% in March of 2020, spiked to nearly 24% in June and have subsequently fallen below 7.0% as of last month’s remittance. Performance across limited documentation loans, an area of concern among investors as many self-employed borrowers with limited documentation looked disproportionately affected by the pandemic, has surpassed the performance of fully underwritten loans. Loans underwritten using a property’s expected rental income have performed better than the broader non-QM market. And delinquency rates on both alternative documentation and DSCR loans sit roughly 100 bp below those of the broader non-QM cohort (Exhibit 1).
Exhibit 1: Alternative documentation non-QM loans have outperformed post-pandemic
Source: CoreLogic, Intex, Amherst Pierpont
While broad declines in delinquency rates across most sectors of mortgage credit can be attributed to a robust housing market, large-scale fiscal and monetary stimulus and a strong consumer balance sheet, outperformance of certain sectors of the non-QM market are likely fueled by additional factors as well. Despite an overall loosening of underwriting standards, particularly regarding borrowers’ assets and income, the presence of strong compensating credit characteristics remained. For example, borrowers qualifying with fewer bank statements typically ended up with less leverage. Low origination LTVs coupled with continued home price appreciation appear to be the primary driver of outperformance across both alternative documentation and DSCR loans.
The quality of loans going into non-QM trusts continues to improve. Looking at a couple elements of layered risk in loans backing non-QM trusts—namely the percentage of lower FICO and higher LTV loans—shows the population of those loans backing non-QM trusts have strengthened somewhat substantially in the wake of the pandemic. Lower FICO loans, which comprised 36% of collateral backing non-QM trusts issued in the first quarter of 2020, made up just 26% of collateral backing deals issued in the second quarter of this year. And loans with original LTVs greater than 90 have fallen from 13% of collateral to 5% of loans over the same two observation quarters (Exhibit 2). Interestingly, despite the quality of these pools improving over time, the average risk-based pricing attached to recently securitized loans is roughly in-line with that of loans with weaker attributes issued in the first quarter of last year.
Exhibit 2: Collateral improves but risk-based pricing remains elevated
Source: CoreLogic, Intex, Amherst Pierpont
Improved credit quality, the drop in interest rates and a broad-based flattening of the credit curve have combined to increase advance rates and reduce issuers’ cost of funds. The increase in credit quality can be observed through the attachment point on ‘B’ classes of non-QM deals and the subsequently higher advance rate on par liabilities as a function of lower ‘B’ attachment points. For example, CSMC 2021-NQM5, a recently issued non-QM securitization achieved a 99.45% advance rate by issuing ‘AAA’ through single ‘B’ rated bonds. In this transaction the single ‘B’ attached at 55 bp, translating to a 99.45% advance rate on par bonds sold. At a 2.3-year weighted average life for transaction and a weighted average spread on liabilities issued of 86 bp, that translates to an all-in cost of funds of roughly 1.2% for greater than a 99% advance rate (Exhibit 3). Additionally, increased valuations on pools of collateral backing non-QM trusts should allow issuers to satisfy risk retention requirements while potentially having to hold less principal risk in the securitization as risk retention requirements are market-value based.
Exhibit 3: High advance rates, low COF should help fuel non-QM issuance
Source: Bloomberg LP, Amherst Pierpont – Based on CSMC 2021-NQM5 transaction priced on 7/27/2021
Estimating securitization economics
Low liability costs appear to be translating to favorable securitization economics for sponsors of non-QM deals. Using estimated coupons and clearing levels for pools of non-QM loan pools coupled with liability coupons, prices, yields and structural leverage of recent non-QM securitizations, economics associated with retained classes of deals held by sponsors can be estimated. A $300 million pool with just over a 4.30% net WAC trading at roughly $105-8/32 would have a market value of $315,375,000. Applying current coupons, prices and subordination levels available in the market on the ‘AAA’ through ‘B’ rated classes, a coupon for the residual excess IO can be derived. Once a multiple is assigned to that excess IO, the market value for that class can be derived as well. The difference between the market value of the loans and the priced classes of the deal produces an estimated market value for the unsold B3 class. Using that market value of the B3 and its par amount, a price can be derived for the B3 class. Based on this methodology and assuming a 3.49% coupon at just over a 1.5 times coupon multiple on the excess IO, it’s estimated that sponsors may be able to retain the B3 class at less than a $50 price, the IO at a conservative multiple and satisfy risk retention requirements by holding just those two classes of the deal which should provide a strong tailwind to continued issuance (Exhibit 4).
Exhibit 4: Estimating non-QM securitization economics
Source: Bloomberg LP, Amherst Pierpont