Find us
1909 K St, NW
Suite 510 Washington, DC 20006
202 204 7907
202 204 7907
Follow Us
All content Copyright 2023 Fixed Income Insights. All rights reserved.
Taxable Markets

After a Dramatic 2022,
Re-Focused MBS Investors Look for a Calmer Environment in 2023

Executive Summary

To put it mildly, 2022 has evolved with a number of surprises for MBS market participants. The view of the Fed and its actions changed course dramatically and that caused some record-breaking results in fixed income, in general, and within the MBS market, specifically. After a very interesting review of what transpired last year, this article will move quickly into a study of the current demand side of the market. It will then prescribe some strategies for the mortgage book in light of an historically inverted yield curve and still-busy FOMC.
FHN FPA (1).jpgFHN FPA (1).jpg
In 2022, the Fed abruptly exited the MBS market by the end of the first half of the market and US-domiciled banks significantly slowed their buying of MBS in the second half. Since these are, by far, the two largest holders of MBS, spreads performed predictably wider last year.
Figure 1: MBS Spreads Performed Wider than Expected in 2022
figure 1.svg
If one looks at static spreads either from the perspective of z-spreads or the current coupon vs. the 5/10 blend, the MBS market was on the order of 60-70 basis points wider relative to this time last year. To be sure, the overall index z-spread is only nine basis points wider, but that is not adjusted for dollar price. After all, the index dollar price has declined by 13 price points from just over $103 to just under $90 since this time last year. In addition, the OAS of the current coupon MBS is on the order of 20-25 basis points wider than at the beginning of 2022. Not only have MBS market participants been forced to deal with wider spreads and negative returns, the volatility around those returns has been nothing short of unprecedented. After a three-month losing streak from August through October, some “volatility” was welcome for the month, because it signaled a reversal with a very strong recovery in performance in November. But all of this “whipsawing” of performance created the highest realized excess return volatility in this history of the MBS index data.
Figure 2: The Volatility of Excess Return Performance Currently at Historical Highs
Figure 2.svg
Another major factor in 2022 was supply, which was much less than the previous year. Net supply of MBS in 2022 was right around $550 billion. A view of last year’s supply in the context of the last several years is displayed in Figure 3 below.
Figure 3: Net Supply Has Begun a Longer Down Trend
Source: FHN Financial and Bloomberg
As for projected supply going forward, somewhere in the area of $20 billion/month for the foreseeable future makes sense. That is not simply a guess based on the recent trend. There is a fairly well-defined relationship between a 3-month average of net supply and a properly-lagged reading of the MBA’s purchase index. That is shown in Figure 5 below.
Figure 4: Current Level of MBA Purchase Index Suggests $15-$20 Billion Net Monthly Supply
The vertical arrow represents the current level of the MBA purchase index. The red dot represents the most recent data point on the regression that predicts the 3-month average of the monthly net supply vs. the 6-week average, lagged by one month, of the purchase index. A place directly on the regression line would put net supply at $15 billion/month. However, the conforming balance limit has continued to increase rapidly throughout the past few years, including an increase of 12% for the 2023 production year, as Alexis Vilimas points out in the next article. Given this recent history of conforming balance limit increases, therefore, it may be better to assume that supply can come in a bit above the regression line. Therefore, $20 billion/month or approximately $250 billion for next year makes sense.
Against the backdrop of the supply picture for both the past few years and next year, attention turns to the demand side of the MBS market. Before any meaningful discussion of demand-side projections, it important to set up who are the main players on the demand side in the next two charts.
Figure 5: MBS Dominated by Two Investor Types – Approximate Book Value [7, cirlce]
As has been the case for a number of years, US banks and the Fed are the two largest holder-types of MBS securities. Together, they own around 65% of the market. However, the rate of ownership of these two entities — as a percentage of overall holders of MBS securities — is beginning to slip.
Figure 6: Dominance of Fed and Banks in MBS Slipping During Era of Tighter Money
This brings into sharp focus the prospects for potential demand from these two entities going forward. We know what the demand from the Fed will be for the foreseeable future: none. The current policy of QT to help fight current inflation means that the Fed will not be buying MBS at all throughout 2023. Rather, the Fed will let the portfolio roll off. As we have stated many times before, it is our view that the Fed will NOT sell MBS assets from the SOMA portfolio during the foreseeable future. There are many good reasons for this view, the primary one of which is that primary mortgage rates are already much higher and the housing market is slowing appreciably. The Fed does not need to sell at this point.
That leaves the question: what is going on with the banks? Simply put, they are under a great deal of pressure as a result of Fed tightening and QT. And given their size and influence on both the banking industry and the MBS market, most of the current pressure is on the top 10 banks.
Figure 7: Most Bank Investment Portfolio Assets Reside in Top-10 Banks
Due to the unprecedented move in rates, banks have booked some realized losses in 2021 and 2022. However, that pales in comparison to the AOCI (adjusted other comprehensive income) that remains on balance sheet due to fixed income losses.
Figure 8 a & b: Portfolio Realized Losses and AOCI Putting Pressure on Larger Banks
This amounts to almost $350 billion in AOCI loss adjustments as of the end of 3Q22. The share for the top-10 banks is around $130 billion. By contrast, the banking system realized only $1.5 billion in losses for the entirety of 2022. The largest single contributor to this stress has been the MBS portfolio, as it is the largest allocation within the investment portfolio. This is particularly true for banks in the top-10 asset category.
Figure 9: Largest Portfolio Allocation Is to MBS, Especially for Top-10 Category
Despite this stress, caused largely by the MBS portfolio, the outcome for banks — particularly the top-10 banks — could have been much worse. After all, the implied AOCI adjustment for the banking system is currently close to 6% of portfolio book value. However, the Bloomberg intermediate aggregate index — a decent proxy for the duration of bank portfolios — has a book loss of around 15% through the end of 3Q22 (i.e. negative 12% TRR after adjusting for about 3.00% of yield carry). To use a more conservative estimate, the entire 15yr MBS market has an implied book loss of 14%. Either way, the “market” did much worse than implied by banks AOCI adjustment. What is the source of that?
Figure 10: Early and Aggressive Allocations to HTM Saved the Day with AOCI
Large banks, in particular, began to aggressively move new purchases, particularly in MBS, into Held-to-Maturity (HTM) starting in 2H20, and that reduced a very large amount of pressure on capital by reducing AOCI adjustments in 2022. The chart in Figure 13 displays the percentage of the MBS portfolio in HTM, which is currently at an astounding 80% for the top-10 bank category. That pulls the entire sector above a 60% allocation to HTM for the MBS portfolio.
We think all of this has two very important implications for bank demand in the MBS sector going forward. And given the importance of bank demand for MBS, in general, these implications may be critical for the performance of the sector into 2023 and beyond.
  1. It is highly unlikely banks will be incentivized to sell MBS to any large degree in 2023. First, 80% of bank holdings of MBS simply cannot be sold without tainting the HTM portfolio. Second, the average losses are still much too large with an average dollar price of the MBS still more than 10 points below par and the large legacy 30yr 2.0 and 2.5 cohorts trading with even lower $80-handles.
  2. The inverted yield curve and the nature of bank portfolios to onboard close-to-par coupons will keep whatever bank demand there is in the current coupons. The higher coupons, after all, price off the now-cheaper points on the yield curve. We have heard from many bank portfolio managers that they would like to onboard securities at these now-cheaper levels, but they do not have excess cash. But if the curve becomes too inverted, it will not make sense to add relative to the cost of funds.
Therefore, we think net bank demand for MBS in 2023 will be flat to slightly net negative in the base case. Only a re-steepening of the yield curve associated with less loan demand will cause bank demand for MBS to increase. This would be, of course, the full-blown recession scenario whereby the Fed would cut rates again in the face of large recessionary pressures. That is not the base case, at least not for the first half of 2023.
If the Fed is out of the MBS market as buyers (it is) and the banks are going to be a bit constrained in terms of liquidity while nursing still-large AOCI adjustments (they are), the demand for MBS must come from somewhere else. Some of it can come from overseas IF the currency markets settle down and the cost of hedging currency risk for non-US- domiciled investors decreases. However, we think the largest marginal source of “new” demand for MBS can come from asset managers. We mentioned in Figure 2 above how volatile 2022 has been for MBS in terms of excess return performance. However, that pales, by orders of magnitude, in comparison to how volatile are corporate credit returns.
Figure 11: Corporate Credit Just a Touch More Volatile than MBS
This additional volatility for corporate credit manifests itself during periods of both strong positive and negative excess return performance. The key is to determine what is the likely outcome going forward. The chart in Figure 16 below demonstrates that when corporate credit lags, it does so around periods of weaker/uncertain economic activity and to a very large magnitude. The blue arrows in the chart point to periods of recession and during periods of positive GDP growth of less than 2.0%. The only time that corporates outperformed during one of those periods was in 2013 in response to the unnecessary QE3 of late-2012.
Figure 12: Recessionary Periods Historically Not Kind to Corporate Credit Returns
The final factor to mention in this outlook, and perhaps the most compelling, is the shape of the yield curve. We display the UST curve out to the 10yr tenor along with the one-year forward curve and the change in rates assuming the forward curve in Figure 14 below.
Figure 13: Deeply Inverted Yield Curve Could Become Less Inverted, But Not Steep
The read from the yield curve is quite straightforward, in our view. That is, the 2/10 UST yield curve in 2023 will be either less or more inverted than it is now, depending on the commitment of Fed policy-makers to tame inflation, but it is quite unlikely that investors will face a “normal,” upward sloping curve in 2023. This means that both rates product and corporate credit, which are bullet structures and therefore “roll down” a normally-shaped yield curve will now be forced to “roll up” an inverted yield curve. That is not a positive TRR factor for these competing sectors. By contrast, MBS are a wide-window, monthly-amortizing product that should perform better with a flat-to-inverted yield curve.
In conclusion, 2022 was a very tough year, historically speaking, for MBS investors, whether levered (banks) or long-only TRR (asset managers). But spreads are now wider and the current setup of an inverted yield curve and a potential recession bode well for par- and discount-priced mortgage product vs. rates and corporate credit alternatives in 2023. And the MBS sector will need the participation of asset managers and other TRR- focused “private” investors in 2023, as both the Fed and banks will likely be neutral participants on the demand side of the market.