By the Numbers
A survey of relative value in specified pools
Brian Landy, CFA | January 26, 2024
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. This material does not constitute research.
Investing in specified pools can be a good alternative to buying TBA pass-throughs when pools offer higher carry and higher convexity. Special dollar rolls can improve TBA carry, but conventional rolls are currently only special in 30-year 3.5%s and 4.0%s. Rolls have been less likely to run special since the Fed and banks stepped away from buying agency MBS. Some specified pools are a better value than others—pools backed by 100% Texas or 100% Florida, low FICO pools and high LTV pools, which often offer better carry than TBA at low pay-ups. There are options in all coupons 5.0% and below that have better carry than TBA. However, in coupons higher than 5.5%s most specified pools are negative carry to TBA—they offer better convexity and lower spread durations, but investors pay for that by receiving less carry.
One approach for assessing the value of a specified pool is to construct a proceeds- and duration-neutral portfolio of the corresponding TBA and Treasury notes (Exhibit 1). This compares FNCL 2.0% pools from 2021 that are 100% Texas collateral to FNCL 2.0% TBA. The same market value is invested in the two MBS positions, and the positions are made duration neutral by buying equal amounts of 2- and 10-year Treasuries. The analytics were run using Yield Book’s model and specified pool cohort; the specified pool was priced using the Bloomberg’s BVAL price. The carry in the table is the average monthly carry expected over the next year.
Exhibit 1. FNCL 2.0% 2021, 100% Texas.
As of 1/24/2024. Analytics run with Yield Book’s production model. Specified pool priced using BVAL for that cohort in Bloomberg’s MBS index. Carry is the monthly average of the expected carry over the next 12 months. Dollar roll financing is not included.
Source: Yield Book, Bloomberg, Santander US Capital Markets
The Texas pools offer higher yield, OAS, and expected carry than the corresponding FNCL 2.0% TBA. The model thinks that the specified pool has more negative convexity, but the difference is small. The carry projections show that the convexity difference is negligible—the specified pool outperforms across all the rate scenarios with only a small reduction if rates fall. Dollar roll financing is not included in the analysis, but the FNCL 2.0% roll is not currently special.
Each FNCL 2.0% specified pool cohorts with at least $5 billion balance outstanding was compared to a proceeds- and duration-neutral portfolio in the same manner as the Texas cohort (Exhibit 2). This table shows the difference between the specified pool portfolio and its replicating portfolio. The two MBS positions have the same market value and the two Treasury notes have the same market value, just like in the preceding example. The pool types are sorted from highest to lowest carry. The table also includes columns to show the pay-up for the pool type and the number of months it takes to break even. Green rows highlight pool types that are in the top third of shortest breakeven times.
Exhibit 2. FNCL 2.0% specified pools.
As of 1/24/2024. Table sorted by carry from high to low. Green rows indicate specified pool breakeven is in the top 33% of specified pools analyzed. Analytics run with Yield Book’s production model. Specified pool priced using BVAL for that cohort in Bloomberg’s MBS index. Carry is the monthly average of the expected carry over the next 12 months. Dollar roll financing is not included. Cohorts with at least $5 bn UPB.
Source: Yield Book, Bloomberg, Santander US Capital Markets
Pools backed by 100% Texas and 100% Florida have the highest expected monthly carry over the next year. Low FICO and high LTV pools also score well. The various loan balance cuts also offer higher expected carry than TBA but fall a little behind the other pool types. Most of the top performing pools come from the 2020 vintage. With only two exceptions, the top performing pools come from the 2020 vintage. However, the 2020 vintage often commands a higher pay-up than the 2021 vintage, so the latter vintage tends to have shorter breakeven times. For example, the FNCL 2.0% 2021 Texas pools have a 14.6 month breakeven time compared to 23.3 months for the same pools from the 2020 vintage
Investing in specified pools does carry a risk that pay-ups will change. Some of that is caused by the duration difference between the specified pool and the TBA; this analysis mitigates that risk since the comparisons are duration-neutral. However, pay-ups can also change as investor preferences ebb and flow for a given pool type, so some pay-up risk remains. So, some investors may still prefer specified pools that have shorter breakeven times or low pay-ups to minimize the risk that pay-ups change.
The 2.5% coupon looks like the 2.0% coupon—the highest carry comes from 2020 vintage pools (Exhibit 3). Geo stories continue to fare well, followed by low FICO and high LTV, and then loan balance. However, pay-ups tend to be higher in this coupon, lengthening the breakeven times. Only a handful of pool types, mostly with lower expected carry, have a low enough breakeven time to warrant highlighting the row.
Exhibit 3. FNCL 2.5% specified pools.
As of 1/24/2024. Table sorted by carry from high to low. Green rows indicate specified pool breakeven is in the top 33% of specified pools analyzed. Analytics run with Yield Book’s production model. Specified pool priced using BVAL for that cohort in Bloomberg’s MBS index. Carry is the monthly average of the expected carry over the next 12 months. Dollar roll financing is not included. Cohorts with at least $5 bn UPB.
Source: Yield Book, Bloomberg, Santander US Capital Markets
The 3.0% adds several seasoning cohorts, and a few of them offer the highest excess carry in this coupon (Exhibit 4). For example, the highest carry comes from the 2021, 2015, and 2012 vintages. The FNCL 3.0% 2021 has the highest excess carry and a low pay-up, which combine for a short breakeven period. The excess carry tends to be a little lower than for the 2.0% and 2.5% pools, and the breakeven times are generally longer.
Exhibit 4. FNCL 3.0% specified pools.
As of 1/24/2024. Table sorted by carry from high to low. Green rows indicate specified pool breakeven is in the top 33% of specified pools analyzed. Analytics run with Yield Book’s production model. Specified pool priced using BVAL for that cohort in Bloomberg’s MBS index. Carry is the monthly average of the expected carry over the next 12 months. Dollar roll financing is not included. Cohorts with at least $5 bn UPB.
Source: Yield Book, Bloomberg, Santander US Capital Markets
Many of the 3.5% coupon pool types offer good carry and have short breakeven times (Exhibit 5). Pay-ups are low on various spec polls from 2022, like 100% Texas, 100% Florida, and low FICO. However, the 3.5% dollar roll is currently special by about 0.66/32s, which removes most of the carry advantage of the specified pools in this coupon. Investors that can dollar roll may prefer to invest in TBA instead of specified pools. But investors that cannot dollar roll have some nice specified options in this coupon.
Exhibit 5. FNCL 3.5% specified pools.
As of 1/24/2024. Table sorted by carry from high to low. Green rows indicate specified pool breakeven is in the top 33% of specified pools analyzed. Analytics run with Yield Book’s production model. Specified pool priced using BVAL for that cohort in Bloomberg’s MBS index. Carry is the monthly average of the expected carry over the next 12 months. Dollar roll financing is not included. Cohorts with at least $3 bn UPB.
Source: Yield Book, Bloomberg, Santander US Capital Markets
Like the 3.5% coupon, some 2022 vintage pool types offer nice excess carry at low pay-ups (Exhibit 6). Florida, Texas, and low FICO pools are at the top of the list and have low breakeven times. However, this coupon’s dollar roll is also special, by about 0.57/32s. Investors that can dollar roll probably prefer TBA, although that comes with worse convexity and longer spread durations. But investors that cannot roll should consider specified pools in this coupon.
Exhibit 6. FNCL 4.0% specified pools.
As of 1/24/2024. Table sorted by carry from high to low. Green rows indicate specified pool breakeven is in the top 33% of specified pools analyzed. Analytics run with Yield Book’s production model. Specified pool priced using BVAL for that cohort in Bloomberg’s MBS index. Carry is the monthly average of the expected carry over the next 12 months. Dollar roll financing is not included. Cohorts with at least $3 bn UPB.
Source: Yield Book, Bloomberg, Santander US Capital Markets
The 2022 vintage also offers good value in the 4.5% coupon (Exhibit 7). Low FICO, Florida, Texas, and high LTV all have better carry than TBA at low pay-ups. The loan balance pools have better carry than TBA, but less than the other prepay stories and at much higher pay-ups. Several pool types are negative carry vs. the TBA after hedging duration. They have higher convexity, so in larger rate moves will perform better than TBA, but an investor must pay for that convexity. The 4.0% dollar roll is not special, so investors should consider spec pools in this coupon.
Exhibit 7. FNCL 4.5% specified pools.
As of 1/24/2024. Table sorted by carry from high to low. Green rows indicate specified pool breakeven is in the top 33% of specified pools analyzed. Analytics run with Yield Book’s production model. Specified pool priced using BVAL for that cohort in Bloomberg’s MBS index. Carry is the monthly average of the expected carry over the next 12 months. Dollar roll financing is not included. Cohorts with at least $3 bn UPB.
Source: Yield Book, Bloomberg, Santander US Capital Markets
A few pools types—low FICO, Florida, and Texas—from 2022 offer better carry than TBA in the 5.0% coupon (Exhibit 8). However, the returns are a little lower than for similar pools in the 4.5% coupon. Pay-ups tend to be low so there is little pay-up risk. The list of pool types that are negative carry vs. TBA is even larger in 5.0%s than it was in the 4.5%s.
Exhibit 8. FNCL 5.0% specified pools.
As of 1/24/2024. Table sorted by carry from high to low. Green rows indicate specified pool breakeven is in the top 33% of specified pools analyzed. Analytics run with Yield Book’s production model. Specified pool priced using BVAL for that cohort in Bloomberg’s MBS index. Carry is the monthly average of the expected carry over the next 12 months. Dollar roll financing is not included. Cohorts with at least $3 bn UPB.
Source: Yield Book, Bloomberg, Santander US Capital Markets
Almost every specified pool cohort in 5.5%s and above is expected to be negative carry vs. the corresponding TBA after matching duration (Exhibits 9, 10 and 11). Dollar rolls are not special, but investors in specified pools are paying, through negative carry vs. TBA, for better convexity and lower spread durations.
Exhibit 9. FNCL 5.5% specified pools.
As of 1/24/2024. Table sorted by carry from high to low. Green rows indicate specified pool breakeven is in the top 33% of specified pools analyzed. Analytics run with Yield Book’s production model. Specified pool priced using BVAL for that cohort in Bloomberg’s MBS index. Carry is the monthly average of the expected carry over the next 12 months. Dollar roll financing is not included. Cohorts with at least $3 bn UPB.
Source: Yield Book, Bloomberg, Santander US Capital Markets
Exhibit 10. FNCL 6.0% specified pools.
As of 1/24/2024. Table sorted by carry from high to low. Green rows indicate specified pool breakeven is in the top 33% of specified pools analyzed. Analytics run with Yield Book’s production model. Specified pool priced using BVAL for that cohort in Bloomberg’s MBS index. Carry is the monthly average of the expected carry over the next 12 months. Dollar roll financing is not included. Cohorts with at least $3 bn UPB.
Source: Yield Book, Bloomberg, Santander US Capital Markets
Exhibit 11. FNCL 6.5% specified pools.
As of 1/24/2024. Table sorted by carry from high to low. Green rows indicate specified pool breakeven is in the top 33% of specified pools analyzed. Analytics run with Yield Book’s production model. Specified pool priced using BVAL for that cohort in Bloomberg’s MBS index. Carry is the monthly average of the expected carry over the next 12 months. Dollar roll financing is not included. Cohorts with at least $3 bn UPB.
Source: Yield Book, Bloomberg, Santander US Capital Markets