September 3, 2024 | Premium Service | Just what happened to the great recession of 2024, as the BLS revises GDP up to 3% in Q2 2024? "Our recession indicators are flashing red," warned famed Toronto economist David Rosenberg on CNBC last week.
Our view of the credit markets says that those screaming about the risk of recession are, well, a bit early in the inflation/deflation cycle. Four years early, to be precise. Bank consumer loan default rates are falling for now three quarters. Are we really cutting interest rates? Will the FOMC really boost residential home prices another 20-30% to gun Wall Street profits? Yes they will.
Back in 2017, The Institutional Risk Analyst wrote about the duration trap created for banks and other buyers of mortgage backed securities (MBS) by the Bernanke FOMC’s massive bond purchases (“Banks and the Fed's Duration Trap.” Nobody paid much attention. Now, however, another leg down awaits lenders and investors who rely upon the US bond market. Ben Bernanke's social engineering experiment known as “QE” is about to throw us yet another curve in the world of fixed income securities.
As before, the problem lies in the duration of variable maturity mortgage-backed securities and whether or not these orphan bonds actually follow the broad markets. Are we correlated please? After the Fed began to raise interest rates in February 2022, it was assumed that as interest rates eventually went back down, the “mark-to-model” problem of unrealized losses on low coupon, very long duration securities also would go away. Would that it were so.
The MBS are still here, just hanging out, with prepayment rates in single digits. And as it just so happens, the dealer community’s inventory of late-vintage Treasury and agency debt is at all-time highs, above March 2020 levels by golly. “Net Agency MBS holdings are now at their all-time highs ($2B above March 2020), with net MBS Pass-through holdings ($69B) now at their highest level since August 2020,” writes Scott Buchta at Brean Capital. He continues:
“During both of these periods the dealers were providing a considerable amount of liquidity to the market. The March "spike" prompted the Fed to open the floodgates with regards to balance sheet expansion while in August they were providing a lot of liquidity, in addition to the Fed, in the face of increased MBS production from surging refis. Are recent moves in anticipation on a more aggressive Fed rate cut cycle?”
Maybe. Our thought about the Street’s bulging inventories of recent issue Treasuries and MBS is this: What if the Fed does nothing in September? What if dealers have to sit on record inventories of relatively high-coupon securities until after the November election? As spreads between asset returns and funding tighten, is the conventional wisdom about bond yields and banks more wrong than ever before? The chart below shows average yields for conforming mortgages and the 10-year Treasury note c/o FRED.
An increase in the long end of the yield curve even as the federal funds rate is cut, would be “the worst possible scenario” for low-coupon MBS, according to an intriguing research note from Stifel Financial (SF) last week. Sell side firms, after all, are usually so constructive. Bloomberg reported that the forward curve suggest the long end will be little changed as the short end drops and the yield curve "disinverts," Stifel strategists said. This view suggests that banks waiting to sell low-coupon MBS into a rate rally may be disappointed.
When the 3% conventional loan we took out during COVID was sold into a 2% Fannie Mae MBS in February of 2021, the Fed was still actively buying paper for its $7 trillion SOMA hoard and spreads were tight. At the time, the loan origination market was insane, with sellers taking points of profit out of usually loss-leading loans and prepayment rates on even new pools in mid-double digits.
Between 2020 and 2022, the effective average life of many new issue MBS pools was less than 24 months. Sellers of loans with servicing, ignoring reps & warranties, started soliciting the former clients for refinance within weeks of the close. It is interesting to note that China's communist government is considering a mass refinance effort for $7.9 trillion in residential mortgages, AFR reports. Interesting to note, in China under Xi Jinping and the US under Joe Biden and Kamala Harris, policy makers resort to brute force assaults on private markets to achieve economic goals.
Today those low-coupon MBS pools that are owned predominantly by the Fed but also by some of the largest banks, have effective average lives of 10-15 years instead of 15 months. As maturity extends, the prices for the MBS fall and spreads over Treasury yields widen. Moreover, because the Fed and the banks do not hedge these vast held-to-maturity positions (something like a third of all $13 trillion in US residential loans), the paper is illiquid. MBS 2s and 3s trade, if at all, at a steep discount to current coupon MBS with coupons in the 5s and 6s. The table below shows forward rates for Ginnie Mae MBS from Bloomberg for September delivery in TBAs.
Ginnie Mae TBAs
Source: Bloomberg (08/30/24)
The Fed thinks it does us all a favor by sequestering trillions of dollars in private market duration related to mortgage finance inside the system open market account (SOMA). When the Fed buys MBS as opposed to Treasury debt, they transfer income from private mortgage loans to the Treasury (less the Fed's operating expenses and SOMA portfolio losses, of course). Treasury (via the GSEs and Ginnie Mae) was paid to guarantee these loans, but now they take the income as well. One of these days, we'd love to hear Fed Chairman Jay Powell or his successor explain why the Fed buying MBS for the SOMA, on net, is helpful to full employment or price stability.
When short-term rates fall, new vintage MBS with 6% and 7% coupons will go up a lot, but the performance of older COVID-era paper may underwhelm. The market risk created by the FOMC in the world of mortgage finance is considerable and not always obvious. Some current production MBS will underwhelm more than others, for example, because coupons for loans inside Ginnie Mae MBS average half a point lower than similar vintage conventional MBS, Bloomberg reports.
Why the disparity between conventional loans and FHA and VA assets? Because of intense competition for loans or, more specifically mortgage servicing assets, over the past two years. Those loans that were being paid-down a point or more by United Wholesale Mortgage (UWMC) and Rocket Companies (RKT) over the past year will refinance first. If hyper aggressive buyers like UWMC pay 7-8x multiples for new issue loans with servicing, then sell the servicing to Lakeview or Freedom on a 4-5x multiple, are they making money? Late vintage MSRs will be the the first assets to evaporate in a rate reduction scenario. God does have a sense of humor.
“GNMA II investors need to account for this difference in loan rates,” writes Rich Easterbrook of Oppenheimer, noting the impact on prepayment risk. Roughy 12.5% of loans in GNMA IIs are at or in the money for refinancing, Oppenheimer estimates, but it climbs to 16.5% “using the appropriate FHA and VA loan rates.” Or in other words, the fierce competition for loans over the past 18 months has created distortions in loan pricing that may also create considerable prepayment risk for lenders and investors in servicing assets.
Big picture, as the Fed does eventually begin to ease interest rates, that huge slug of long-duration MBS sitting on the books of banks may hold down prices on the long end of the yield curve, making a long-awaited normalization of interest rates possible. But this also implies that residential mortgage rates will remain elevated and MBS spreads over the Treasury curve could even widen.
Those low coupon MBS with 10+ year average lives will price against 10 or even 15 year Treasury paper and with very wide spreads to compensate dealers for the negative carry. The great chart from Ginnie Mae below was published in an earlier comment and shows the skew in the mortgage market in terms of coupons of government-insured MBS.
Source: Ginnie Mae
These technical musing about the bond market and interest rates come as we near the end of the third quarter, with funding costs down small and default rates still moving sideways for now three consecutive quarters. Was 2023 the recession that never was? Perhaps, but we think that the indigestion caused to markets and financial institutions through the creation of $6 trillion in below market mortgage securities will be a continued negative factor weighing on the credit markets. Indeed, the individual bank data from this period makes clear who was paying attention to market risks like coupons and duration, and who was not.
The Consumer Lenders: AX, AXP, ALLY, COF, SOFI, SYF
Comments