The Big Idea
A midterm exam in MBS
Brian Landy, CFA | July 14, 2023
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.
Agency MBS spreads drifted wider over the first half of the year in large part due to the regional bank failures in March. This boosted MBS supply as the FDIC sold the assets of the two failed banks and kept banks on the sidelines. Net supply from new origination should pickup in the 2nd half of the year, boosting the supply of higher coupon MBS and keeping the pressure on option-adjusted spreads. Although spreads have matched expectations thus far this year, turnover has been somewhat lighter than expected, although much higher than the catastrophically low speeds some people feared. The share of loans in specified pools has increased this year, driven by pools backed by purchase loans with at least 95% loan-to-value ratios.
At the start of the year, it seemed unlikely that option-adjusted spreads would tighten. Demand was light since the Fed was no longer purchasing mortgages, bank buying had dropped precipitously, and mortgages must compete with other asset classes for interest from money managers. Net supply was likely to be elevated compared to pre-pandemic norms, despite slowing home purchase activity, since home price appreciation has driven the size of an average new loan higher.
Spreads did spend much of the year wider than at the start of January and have currently returned to roughly that starting level (Exhibit 1). However, the key reasons were unexpected. First, the failures of Silicon Valley Bank and Signature Bank in March surprised the market and option-adjusted spreads jumped wider as the market prepared for their assets to be sold. Second, spreads widened as concerns mounted in May that the United States might default on its debt obligations.
Exhibit 1. Option-adjusted spreads widened throughout much of 2023.
Spreads should still have a difficult time tightening. A big reason is that net supply usually is higher in the second half of the year (Exhibit 2). The chart shows monthly net supply from 2014 through 2019. The largest months of net supply typically fall in the shaded areas, which indicate the second half of each year. Supply is primarily driven by home sales; most new home sales contribute directly to MBS supply, and existing home sales lift supply when the loans for new purchases are larger than the old loans being paid off. This seasonal pattern means that net supply will typically be higher in the second half of the year, all else equal. Furthermore, new home sales have spiked over the last couple of months and supply has increased as well, reaching $30 billion in June. That is the largest amount since last September. And the FDIC has roughly 25% of its MBS portfolio left to sell, which will also contribute to supply.
Exhibit 2. MBS net supply is typically higher in the second half of the year.
At the start of the year, two net supply forecasts based on housing projections from Fannie Mae and the Mortgage Bankers Association suggested that supply would average $25 bn to $35 bn per month this year. Reaching the midpoint of those estimates may be difficult, as it would require $45 bn/month supply average over the rest of the year. But reaching the low end of the range still feels plausible and would rise above the pre-pandemic average of roughly $20 bn/month net supply. The recent increase in mortgage rates does present a risk that origination falls more rapidly in the fall and further cuts into supply.
Turnover has also come in lighter than last year’s expectation, but has been stronger than the catastrophically low levels many were expecting. Prepayment speeds for 30-year agency MBS averaged 5.1 CPR from January through June. Turnover also tends to be a little faster in the second half of the year, however, since the slowest speeds come in January and February. The rest of the year should average a little faster than the first half, although the recent increase in mortgage rates poses a risk to that. It still seems likely, absent a recession, that turnover will climb back into the 6 CPR to 7 CPR range over the next year or two as loans season. For example, consider that 2020 2.5%s prepaid at 6.7 CPR in June and 2020 2.0%s at 5.5 CPR in June. Those borrowers are deeply out-of-the-money. Each of those cohorts prepaid roughly 0.7 CPR faster than the same-coupon 2022 vintage loans, so total speeds should move higher as the 2021 vintage seasons. Ginnie Mae loans, especially VA, have maintained faster discount speeds than conventional loans this year.
Specified pooling picks up
The share of loans placed in specified pools instead of generic pools increased in 2023 compared to 2022. Lower demand for MBS allowed dollar rolls to trade below carry, which raises the attractiveness of buying pools instead of TBA. And higher loan limits allowed originators to place more jumbo conforming loans into TBA-deliverable pools without reaching the 10% de minimis limit, which frees originators to create pools with lower pay-ups. In conventional pools, roughly 7% of issuance shifted from generic pools into specified pools. Most of those loans are high LTV loans originated by state housing finance agencies or using Fannie Mae’s and Freddie Mac’s high LTV lending programs, called HomeReady and Home Possible. In fact, a larger amount of high LTV pools were issued in the first half of 2023 than over all of 2022.
The share of Ginnie Mae loans in specified “custom” pools also increased—an extra 5% of production was placed in custom pools instead of generic multiple issuer pools compare to 2022. The largest increase was in pools backed by only FHA loans and pools with loans between $175,000 and $250,000.
FHA MIP Cut
It seemed likely that the FHA would lower insurance premiums in 2023, and that happened early in the year. However, the cut was 30 bp, larger than expected. That brings insurance premiums back to the level at the time of the 2008 financial crisis, which ultimately depleted the FHA’s insurance fund and necessitated an infusion from the United States Treasury. A roughly 15 bp cut had seemed safer. However, the large cut brings is upside to investors. It seems highly unlikely that annual premiums will ever drop below this level, removing a policy risk that investors have worried about for many years.