Cross Sector P. White, L. Wash
The Fed delivered a hawkish 50bp cut. The Treasury curve bear steepened, TIPS breakevens widened, and credit spreads tightened, reflecting a higher probability of a soft landing. Given still elevated policy uncertainty we think outright duration or curve themes provide poor risk reward and favor looking for conditional trades and relative value. Spreads look snug but we do not favor shorting them.

Governments J. Barry, P. White, A. Borges, L. Wash
The curve decoupled from medium-term policy expectations this week, despite a hawkish 50bp delivery to kick off the easing cycle. This could reflect increased confidence the Fed’s front-loaded cuts will extend the expansion. If this narrative persists the curve can steepen bullishly or bearishly, but we will be patient before adding back to steepener exposure. Intermediate valuations appear modestly cheap, but yields are in the lower 40% of their recent range: stay neutral on duration for now. Initiate 3.125% Nov-41/ 2% Nov-41 flatteners for RV. We review the July TIC data. While the backdrop has turned more bullish for TIPS, valuations are not compelling. Stay neutral but hold energy-hedged 5s/10s breakeven curve steepeners.

Interest Rate Derivatives S. Ramaswamy, I. Ozil, P. Michaelides, A. Parikh
Despite this week’s policy action, policy uncertainty remains elevated, but forwards are now well below the dots in the Reds, causing us to favor asymmetric exposure to higher yields in the Reds - Initiate weighted Greens/15s flatteners and/or 1s/5s conditional bear flatteners. Stay neutral on swap spreads across the curve but initiate 2s/3s spread curve flatteners for relative value. Markets will likely enter a lull until the next Jobs report - we turn tactically bearish on short expiry vol in intermediate and longer tails. Initiate long exposure to second principal factor vol by buying 1Yx3Y straddles versus 1Yx10Y.

Short-Term Fixed Income T. Ho, P. Vohra
Despite the very front-end steepening this week, yield curves remain deeply inverted, challenging liquidity investors willingness to add duration. The yield spread between government and prime funds might narrow as we get further into the easing cycle. We provide a MMF holdings update.

MBS and CMBS J. Sim
Mortgage spreads had an excellent week and are rich in OAS terms, but wide ZV spreads should continue to draw in buyers focused on nominal spreads. This week’s Fed cut was a welcomed move to the CRE/CMBS market participants. Also, we discuss potential extension risk in the CMBS market

ABS and CLOs A. Sze, R. Ahluwalia
ABS spreads held firm with heavy supply packed in before, and some selling of short-dated paper after, the Fed’s first rate cut of the cycle.

Investment-Grade Corporates E. Beinstein, S. Doctor, N. Rosenbaum, S. Mantri
The 50bp Fed combined with light bond issuance and strong fund inflows all contributed to spreads 6bp tighter on the week. With spreads now just 6bp from recent tights, valuation is less compelling.

High Yield N. Jantzen, T. Linares
High-yield bond yields and spreads declined 31bp and 29bp over the past week to 7.08% and 349bp, respectively. The HY index is up +1.5% in September with Cable/Sat (+4.7%), Telecom (+4.0%) and CCCs (+4.2%) outperforming.

Municipals P. DeGroot, Y. Tian, R.Gargan
Next week begins a gauntlet of tax-exempt supply expected to run up to the election, at the same time as reinvestment capital dips to the lows of the year. As such, levels and spreads are expected to be progressively cheaper moving through the fall. Certain pari tax-exempt debt is now 2 sigma cheap vs pari corporate bonds.

Emerging Markets L. Oganes
In EM fixed income, we are OW GBI-EM local rates, and MW CEMBI and EMBIGD. EM bond flows were +$541mn (+0.14% of weekly AUM, up from -$705mn).

Source: J.P. Morgan

End of the beginning or beginning of the end?

• Economics: The Fed delivered a 50bp cut at this week’s FOMC meeting, in line with our expectations, bringing the policy target to 4.75-5.00%. We expect the Fed to deliver a 50bp cut in November before beginning a series of 25bp cuts in December. We revise higher our 3Q GDP growth forecast to 2.5% and 4Q24 to 1.25%
• Treasuries: The curve decoupled from medium-term policy expectations this week, despite a hawkish 50bp delivery to kick off the easing cycle. This could reflect increased confidence the Fed’s front-loaded cuts will extend the expansion. If this narrative persists, the curve can steepen bullishly or bearishly, but we will be patient before adding back to steepener exposure. Intermediate valuations appear modestly cheap, but yields are in the lower 40% of their recent range: stay neutral on duration for now. While the backdrop has turned more bullish for TIPS, valuations are not compelling. Stay neutral but hold energy-hedged 5s/10s breakeven curve steepeners.
• Interest Rate Derivatives: Policy uncertainty remains elevated and we continue to prefer conditional trades which offer asymmetric reward in either bullish or bearish states of the world. To position for rising yields in the Reds, we like weighted Greens/15s flatteners. With low event risk on the horizon for the next two weeks, we turn bearish on shorter expiry and intermediate and longer tail volatility. We are more cautious on shorter tails, and recommend a neutral stance. While we remain broadly neutral on swap spreads for now, we now recommend positioning for a 2s/3s maturity matched swap spread curve flatteners
• Short Duration: Despite the very front-end steepening this week, yield curves remain deeply inverted, challenging liquidity investors’ willingness to add duration. The yield spread between government and prime funds might narrow as we get further into the easing cycle
• Securitized Products: Mortgage spreads had an excellent week, responding favorably to the 50bp cut. That’s put mortgages at new YTD OAS tights, but wide historical ZVs will continue to draw in buyers focused on nominal spreads
• Corporates: The 50bp Fed cut, combined with light bond issuance and strong fund inflows, contributed a 7bp tightening in spreads. With spreads now just 7bp from recent tights, valuations are less compelling
• Near-term catalysts: Aug Personal income (9/27), Aug JOLTS (10/1), Sep ADP (10/2), Sep employment (10/4), Sep CPI (10/10), Sep PPI (10/11)

After weeks of anguishing anticipation, the Fed delivered a 50bp cut on Wednesday to bring the Fed funds target to 4.75 – 5.00%, in line with our expectations. During the ensuing press conference Chair Powell described the outsized move as a “recalibration” to preserve the currently strong labor market from downside risks as well as a signal of “commitment not to get behind.” To that extent, the median September FOMC dots indicate a cumulative 50bp of easing over the next two meetings, 100bp in 2025, and an additional 50bp in 2026. Meanwhile the SEP also showed the median unemployment rate projection for this year and next is 4.4%, up 0.4%-pt and 0.2%-pt, respectively, from the June projections but largely stable ( Figure 1). Powell’s characterization of the employment situation was more sanguine than expressed at the Jackson Hole Symposium, stating the labor market is “actually in solid condition” and “moving to a more sustainable level” with unemployment “in the low 4s [indicating] a really...good labor market.” Moreover the Chair recognized that the neutral policy rate is “probably significantly higher” than it was before the pandemic. In the aftermath of the meeting, price action across fixed income markets reflected a higher probability of a soft landing outcome, with a bear steepening of the yield curve, wider TIPS breakevens, and tighter spreads across corporate credit and mortgages.

These moves were also supported by economic data that have continued to surprise to the upside. Following this week’s August retail sales report, our economists revise up their tracking of 3Q GDP growth to 2.5% from 1.5% and 4Q GDP to 1.25% from 1% (see Economics). If this forecast is realized and a step down in policy rates mitigate labor risks, alongside easier financial conditions and a boost to real consumption from lower inflation, this could keep recession risks in check heading into 2025. That said, we are cognizant that labor market data and growth data have been disconnected for a number of months and are likely to remain so into the fourth quarter. We continue to expect the unemployment rate to tick up to 4.5% by year end, above the Fed’s median projection. However almost all of the recent U-rate rise can be explained by declining job finding rates, and not rising separations, which is not typical during recessions and could be consistent with a labor market normalizing around a higher NAIRU (see US: Reaching u* before r*, Murat Tasci, 9/17/24).

Figure 1: The SEP showed upward revisions to unemployment rate projections, while inflation forecasts were revised somewhat lower

Federal Reserve Summary of Economic Projections, June 2024 vs. September 2024;

Source: Federal Reserve

Figure 2: Following this week’s Fed meeting, OIS forwards priced a less dovish path for policy

Expected change in Fed funds at FOMC meetings implied by OIS forward rates; 9/17/24 vs. 9/20/24; bp

Source: J.P. Morgan

Thus, we continue to forecast the Fed will cut by another 50bp at the November meeting before stepping down to a 25bp per meeting pace in December. However we recognize this is contingent on softening in the job reports between now and then, and more benign labor market data would seal the case for a 25bp at the next FOMC meeting (see An appropriate recalibration, Michael Feroli, 9/18/24). Against this backdrop of policy uncertainty, we think outright directional trading themes in duration or curve are likely to offer poor risk-reward. Moreover spreads look snug but we still see risks they could move tighter and we do not favor shorting them.

Following the outcome of the Fed meeting, we recommended unwinding our broad curve steepener trade, which we have held in some form since last November. Importantly, this trade was largely predicated on expectations that the front end of the curve would lead the way lower as the Fed approached its first ease, recognizing also that structurally higher term premium would anchor long-end yields at higher levels, given the rapid growth of the Treasury market continues to outstrip demand from its historically price-insensitive base. It’s possible that if growth and inflation expectations continue to rise over the near term, the curve could steepen in more bearish fashion from here.

However, valuations no longer appear attractive, given that OIS forwards have more closely converged to our Fed forecast in recent weeks and as the broad curve now looks too steep after adjusting for the level of 1Yx1Y OIS rates. Figure 2 shows that OIS forwards priced a less dovish path for policy following the Fed meeting, but still imply 73bp of cuts over the balance of this year and 124bp of cuts in 2025. This is particularly important given the negative carry of steepener positions and shorter-duration Treasury supply next week could bias the curve flatter in the near-term. Turning to duration, 10-year Treasury valuations also appear somewhat cheap after controlling for the market’s medium-term Fed policy, inflation, and growth expectations, as well a the Fed’s share of the Treasury market. While this skews us towards a more bullish lean, we remain neutral with money markets closely in line with our forecast and outright yield levels still in the lower 40% of the 40bp range they have traded over the past six weeks (see Treasuries). Consistent with this view, our technical strategists note the bond is already pressed into the 4.065-4.095% tactical support zone. We see other layers at 4.235-4.26% and then 4.315-4.345%. We suspect bear steepening will have a limit given its typical correlation with commodity price trends and inflation markets. Look for that dynamic to keep the bond yield capped at 4.25% through the fall (see Technicals).

In inflation markets, the fundamental backdrop has shifted more bullishly for TIPS over the last two weeks, given the strong action from the Fed combined with resilient growth data. However, breakevens have already retraced to their widest levels since just prior to the release of the July employment report, and we believe the Fed’s data dependent stance likely leaves breakevens vulnerable to a correction lower, as we think the amount of easing currently priced in the forwards will only materialize alongside further cooling in labor markets. Additionally, our medium-term outlook on inflation remains unchanged. We believe core inflation will soften gradually, supported by core goods deflation and softening service price pressures. Fixings are priced close to our own forecasts, implying an average monthly pace of core CPI inflation near 0.27% in the next few months and near 0.25% through 1Q25. Additionally, breakevens appear fairly valued after adjusting for the repricing in the Fronts/Reds curve and broad commodity prices. Thus, we do not think breakevens are terribly compelling at current levels. For now, we recommend holding energy-hedged 5s/10s breakeven curve steepeners (see TIPS).

One way of illustrating the still elevated state of policy uncertainty is to show that weights derived from Z4 and M5 3-month SOFR futures on individual policy outcome scenarios remains largely the same after the FOMC meeting this week ( Figure 3). During this period of uncertainty, we continue to prefer conditional trades which offer asymmetric reward in either bullish or bearish states of the world. Along the swap yield curve, we note that while OIS forwards are not inconsistent with our own Fed funds forecast, the Reds are well below levels indicated in the YE25 SEP median dot. To position for rising yields in the Reds, we like weighted Greens/15s flatteners which should remain insensitive to small moves locally but tend to flatten into an accelerated rate selloff. A variation of this theme is to initiate a flatter 1s/7s curve in a selloff, using 3M expiry payer swaptions. Additionally, with low event risk on the horizon for the next two weeks, we turn bearish on shorter expiry and intermediate and longer tail volatility. We are more cautious on shorter tails, and recommend a neutral stance.

Figure 3: The passing of the FOMC meeting has not translated into better clarity on the path of the funds rate going forward

Total weights on YE24 and 1H25 policy rate scenarios representing a range of different Fed funds rates, as calculated from a decomposition of the implied probability distribution associated with Dec 2024 and June 2025 SOFR futures*; 9/17 (before FOMC) and 9/18 (After FOMC

	Dec 2024 weights		Mid 2025 weights
Funds Rate	Before FOMC	After FOMC	Before FOMC	After FOMC
4.50	0.27	0.00	0.00	0.00
4.25	0.20	0.57	0.00	0.00
4.00	0.21	0.11	0.12	0.04
3.75	0.17	0.19	0.05	0.12
3.50	0.07	0.08	0.00	0.00
3.25	0.03	0.00	0.08	0.09
3.00	0.05	0.04	0.75	0.75

* We enumerate a list of scenario-specific Normal distributions with fixed standard deviations and means that are separated by 25bp, and then require the implied distribution to be a weighted combination of these individual distributions. The weights are then solved for, by fitting to the observed prices of calls and puts at various different strikes. For more details of our approach, see What’s the rush?.

Source: J.P. Morgan., CME

Figure 4: SOFR has been steadily rising relative to the interest rate on Reserves

Rolling 2-week moving average of the O/N SOFR minus IOR differential, %; past 6 months

Source: J.P. Morgan

Turning to the Fed’s balance sheet, there was minimal discussion on QT. However, when asked about balance sheet policy during the press conference, Powell noted that “reserves are still abundant and expected to remain so for some time. The shrinkage has come out of the overnight RRP.” He also mentioned that “for a time, you can have the balance sheet shrink while also cutting rates,” as both actions can be viewed as policy normalization. Importantly, we continue to see QT as being in its endgame, with perhaps a few more months left, and project the Fed’s balance sheet to end the year at ~$7tn (see Hopscotch, Srini Ramaswamy, 8/16/24). Locally, we note that ON RRP balances fell to $239bn on September 16, marking its lowest level since May 2021, on the back of corporate tax date-related outflows and mid-month Treasury settlements. This likely contributed to the rise in SOFR, which increased by 5bp DoD to 5.38% on September 16. As we enter the GSE period and Treasury settlements work their way through the system, along with negative T-bill supply, ON RRP balances should trend back higher throughout this month, and SOFR should also move back lower (see Short-term Fixed Income). Nonetheless, beyond the next couple of weeks, we are steadily heading towards tighter liquidity conditions, with repo rates and SOFR likely to remain biased higher, even possibly necessitating an eventual reduction in IOR ( Figure 4). Given these looming possibilities, banks will likely find it preferable to deploy their spread risk budget in 2Y spread wideners rather in 3s. Therefore, while we remain broadly neutral on swap spreads for now, we now recommend positioning for a 2s/3s maturity matched swap spread curve flattener (see Interest rate derivatives).

The combination of a Fed cut with a more hawkish-than-expected press conference against the backdrop of improving economic data was a positive for corporates, with the spread on our JULI declining 7bp w/w to 105bp, the lowest since mid-July. Given that spreads are already quite tight and the upcoming easing cycle is well anticipated, we do not think Fed easing will have a meaningful impact on spreads near-term and continue to look towards a YE target of 110bp. Looking back over previous easing cycles, we note that spreads tend to decline, returns turn positive, and fund inflows increase over the 6 months following the first Fed cut. However most of these episodes were when Fed decisions were partly driven by weakness in corporate results, and today’s cycle more closely resembles the 1995 proactive easing experience in which spreads were stable over the next year (see What to expect when you’re expecting (rate cuts), Nathaniel Rosenbaum et al., 9/18/24, and Figure 5). Looking ahead, credit markets will continue to debate whether the recent payroll weakness represents the beginning of a material slowdown or simply a soft patch, with the risk to spreads that softer data raises recessionary concerns, and upcoming earnings season will be an important hurdle. On the technical front, questions remain if yield-based buyers will step away with long-end JULI yields now nearly 85bp below their YTD peaks. But we have yet to see a meaningful steepening of the spread curve which we would expect in this scenario, perhaps owing to the lack of long-end supply. Meanwhile, we would expect further rate cuts to bring stronger retail and foreign demand given the recent increase in fund inflows, the reduction in hedging costs, and more appealing relative value of front-end credit versus money market yields (see Corporates).

Figure 5: HG spreads tend to be tighter on average once the Fed cuts, though were roughly stable over the 1995 easing cycle

Change in spreads relative to the first Fed cut from 3 months prior to 12 months post, over the 1995, 1998, 2001, 2007, 2019, and 2024 easing cycles; bp

Source: J.P. Morgan, Bloomberg Finance, L.P.

Figure 6: High Yield company revenues and EBITDA growth remain subdued but are stabilizing

LQA EBITDA y/y growth; %

Source: J.P. Morgan

Similarly down the credit spectrum, a hawkish 50bp cut combined with more resilient consumer data was bullish for high yield bonds, as spreads declined 29bp over the week to 349bp. However if the Fed were to deliver on another 50bp cut in November, which is predicated on further labor market softening, we would expect spreads would widen materially as this would indicate a worsening economic outlook. Meanwhile if more benign labor data seals the deal for a series of 25bp cuts, we would expect spreads to remain near 350bp. Turning to fundamentals, we do not see much cause for concern among HY company balance sheets, with credit metrics largely stable in 2Q24. Indeed revenue and EBITDA expanded q/q following quarterly declines in 1Q and 4Q, while profit margins ticked higher off of a three-year low ( Figure 6). While leverage continues to inch upward, 2Q24 marked the second quarter of mold increases and overall leverage remains comfortably below long-term averages (see High Yield).

This week’s Fed cut is a reminder that the carry tailwind for CLOs will gradual fade as monetary policy loosens. Indeed we expect the carry/current yield for CLO AAA around 7% prior to this week’s easing will fall to 4% by YE, based on our SOFR and Treasury yield forecasts. While not unattractive, it isn’t much higher than UST yields, and not every investor is an absolute return or yield buyer. While T1 CLO AAA Primary spreads around 135bp remain below their non-recessionary averages, rising credit concerns bring focus back on CLO liability downgrades. Currently the percentage of US CLO bonds in CLOIE that have been downgraded over the past year stands at 2.3% by count and 0.4% by volume. Downgrades are concentrated in CCC bonds and among the 2012-15 vintages where the majority of bonds are currently out of reinvestment (see CLOs).

Mortgage spreads reached new year-to-date lows with the MBS market reacting positively to a 50bp rate cut and general risk-on attitude across spread products. Option-adjusted spreads look historically tight and the question remains whether technical dynamics can support further tightening. H.8 data show bank demand is starting to creep up after years spent in the doldrums ( Figure 7). Importantly however, bank buying is no longer focused on parking unhedged fixed rates in HTM to pair off sticky deposits but instead far more spread-focused, as evidenced by CMO floater issuance and portfolio layer hedging. While it is unlikely we see the same level of strong demand as during the 2010-21 period, banks could take some comfort from the initiation of the cutting cycle to add mortgages even at very tight OASs. We think $50-100bn net add could be possible over the next year. That, coupled with potential for overseas buying as the US curve steepens, means more focus on ZVs or nominal spread than before. Some money managers have been waiting to rotate out of MBS into corporates, anticipating that IGs would eventually widen, but with the JULI still near its recent tights, it’s not entirely obvious what to buy if you sell MBS. Moreover there could be some inertia in mortgage weights if managers look to keep yields high and respond to surging inflows. So, while we don’t love mortgage spreads here, we understand why it’s hard to really short them. We’d still recommend a barbell on the stack to stay away from the belly, at least for investors with flexibility on dollar price (see Mortgages).

Agency mortgage rates meanwhile are only a few basis points above 6% and the 30-year rate is poised to fall to 5.9% by year-end based on our Treasury yield estimates. Despite the 75bp reduction in mortgage rates, existing homes sales (demand) has not budged while borrowers remain less sensitive to the recent moves given just over 60% retain 4% or lower mortgage rates. Instead investors are more focused on convexity risk in jumbo collateral as 6% coupons shift more into the money. This could begin to pressure premium price spreads. Non-QM should offer a convexity advantage with a heavier investors share of loans and prepayment penalties. Meanwhile we have seen delinquencies rise in the non-QM and FHA segments while GSE and jumbo DQs remain subdued ( Figure 8). This has been driven by recent vintages and by bank statement and ‘other’ documentation borrowers. Investors are not significantly concerned as current LTVs of delinquent borrowers are below 70% and the housing market remains supported by a lack of supply. That said, if delinquencies continue to rise at current pace, AA and A classes could face extension risk in certain deals for which delinquency triggers are breached (see RMBS).

Figure 7: Bank buying appears to have picked up modestly ahead of the FOMC

2024 YTD* changes in agency MBS holdings across all U.S. banks in amortized cost terms; $bn

*The “YTD (adj. for bank trading accounts)” series adds subtracts out the $12.5bn settled in 2023 and trading account changes, spread across Class A settle increases. In Q1, we subtract $26bn of settled Class A holdings to tie out with call report increases in bank trading accounts and match H.8. with AFS + HTM call report increases. In Q2, we subtract $20bn of settled Class A holdings to match H.8. changes with AFS + HTM call report changes

Source: J.P. Morgan, Federal Reserve, FFIEC

Figure 8: Delinquencies have continued to increase in non-QM

60+ DQ (%) for loans originated since 2017

Source: J.P. Morgan, CoreLogic, Freddie Mac, FHA

While this week’s Fed cut was a welcome relief for CRE/CMBS markets, the cumulative impact of cuts to terminal will be more impactful for markets. Indeed CRE property prices seem to be bottoming, REIT stocks have rallied, and CMBS has rebounded well ahead of this cut in anticipation of the pause in rate hikes and an eventual cut. For now spreads can grind tighter on sentiment and the spread curve can bullishly flatten. We continue to believe the best relative value can be found in upper IG mezz, particularly in the AS bonds as they remain wide on both a historical basis and relative to spread comps while facing minimal extension risk (see CMBS).

ABS spreads continue to hold firm under heavy supply for another September week. In the secondary market, there was some selling of short dated ABS after Wednesday’s highly anticipated first rate cut of the cycle by the Fed. The inverted yield curve, particularly at the front end, has provided a significant boost to short ABS returns. While some adjustment to yield curve shifts is expected, we do not anticipate a significant pull back in demand for ABS as there remains an attractive spread pickup over Treasuries and comparable credits. In addition, for investors concerned about recession risks, high-quality ABS has served as a relatively stable flight-to-quality safe haven during periods of broad credit sell-offs. Meanwhile we are still seeing strong subscription and solid pricing levels across sectors in the new issue market, with compressed spread tiering for sponsor risk. While YTD supply of $17.1bn lags the $18bn we witnessed by this time last year, tight spreads and lower rates should make all-in yields more attractive to issuers for incremental supply. We think FNMNT, SYNIT, WFCIT and WFNMT offer attractive spread pickup on the top bankcard ABS programs while Canadian banks offer attractive spread pickup versus their US peers (see ABS).

Finally, we note that over the past week J.P. Morgan hosted both its Global Macro and Securitized Products Real Estate and Consumer (SPREC) conferences. Readers can find summaries of the main takeways here and here, respectively.

Figure 9: Cross sector monitor

Current levels, change since 9/13/24, 1-year average, minimum, maximum, and current z-score for various market variables; units as indicated

Source: J.P. Morgan, Bloomberg Finance L.P., ICE, IHS Markit
* 9/19/24 levels for1yr EUR-USD xccy basis, 30Y FNCL 4.5%, 10Y AAA new issue CMBS, JULI, JPM US HY Index, EMBIG, CEMBI, iBoxx Euro HG, US Financials, 10Y AA taxable muni, EUR/USD, USD/CHF,USD/JPY,Bitcoin,gold, brent,copper; 9/20/24 levels for all others

Figure 10: YTD returns on various fixed income indices; %

Source: J.P. Morgan

• As we had expected, and against consensus, the Fed began easing with a front-loaded 50bp rate cut
• We think a continued 50bp recalibration is warranted, but the next move will depend on labor market data
• A solid consumer leads us to revise up 3Q growth to 2.5%ar, ahead of annual GDP revisions
• Mixed signals on manufacturing and housing, but lower mortgage rates should offer (eventual) support

On Wednesday the FOMC delivered a forceful 50bp cut to kick off its easing cycle, with Chair Powell characterizing the move as a “recalibration” that shifted policy from bringing down inflation to safeguarding the soft landing. In addition to a greater focus on downside risks to the labor market, as presaged by Powell’s Jackson Hole speech, the July FOMC minutes, and even Governor Waller’s remarks before the blackout period, this move also reflected growing confidence in inflation returning to the Fed’s 2% target more quickly than previously thought—something that Waller confirmed in subsequent remarks he gave the Friday following the decision—even as it remains “somewhat elevated.”

While markets were about 65% priced for the move, consensus across the Street was nearly unanimously looking for a more modest 25bp cut. In his presser, Powell confirmed each of the reasons we anticipated the FOMC would choose the larger move they did. First, Powell called the move a “good, strong start” and posited that “the logic of this [action], both from an economic standpoint and from a risk management standpoint, was clear.” We agree that the economic data have evolved to suggest that policy was—and arguably still is—too restrictive if , for example, viewed through a Taylor Rule lens (Figure 1). The risk management perspective also warrants a front-loaded easing. Additionally, given the dovish tone of the July minutes, a case could be made for moving 50bp in September to make up for not going in July. Powell conceded the Fed “might well have” cut in July had it seen the jobs report released 48 hours later.

Data dependence continues

Powell also said that the Fed is “not in a rush,” in part to signal that the Fed is making decisions meeting-by-meeting, considering the totality of the incoming data and its implications for the outlook. Nonetheless, our forecast anticipates some further slowing in the labor market that compels the Fed to again cut 50bp at the November meeting. Private payroll growth at 100,000 or less, or a further move up in the unemployment rate, would make a strong case for another 50bp ease, in our view. We readily concede, however, that it could once again be a close call—and this time the October jobs report occurs during the blackout period. But should the labor market stabilize, with some improvement in job growth and a stable unemployment rate, then the path should be clear for the Fed to take a more gradual path toward neutral and deliver a 25bp cut in November, as signaled by the median dot in the latest SEP.

That said, the skew of the distribution of the dots arguably is in a more dovish direction. While we obviously don’t know which dots coincide with which FOMC participant, under the simplifying assumption that the rank order of participant’s dots within each year remains the same over time, we can back out a distribution of expected policy paths. Figure 2 shows the extent of cuts implied by each participant through the end of 2025 per this exercise. While the median dots imply 200bp of cuts in total by then (including September’s 50bp move), the histogram is clearly skewed toward greater easing.

We continue to look for the economy to avoid recession, and thus for the Fed to eventually cut to 3% by mid-2025. This would put policy in the vicinity of the neutral rate. In the SEP, the median longer-run dot again inched up slightly to now reside at just below that level. But as Powell noted in the press conference, estimates of neutral have wide uncertainty bands. Rather than looking to empirical or model-based estimates of neutral to guide policy, Powell once again suggested the Fed will “know it by its works.”

3Q GDP stronger on resilient consumers

The recalibration of policy via a front-loaded 50bp cut this week arguably was the best response to safeguard the soft-landing. The biggest support has come from the so-far resilient US consumer. This week brought another manifestation of that strength in the form of the retail sales report. The headline increased by 0.1%m/m, surpassing expectations due to a smaller-than-anticipated decline in motor vehicle sales (note however that weaker data on unit auto sales will be incorporated into the GDP report). The control category (ex-food services, autos, gas, and building materials) also rose 0.3%, which leaves it up a brisk 6.3% annualized over the last three months (Figure 3). We now expect that real consumer spending rose 0.1% in August, which while slower than the recent average still sets up a strong quarter. We have raised our forecast for 3Q real spending to 3.5%q/q saar from 3.0%.

On the back of this upgrade to consumer spending we have also raised our 3Q GDP forecast from 1.5%q/q, saar to 2.5%, and raised 4Q from 1.0% to 1.25%. We are still see labor market risks front and center, but looser financial conditions accompanied by momentum from the current quarter can bring slightly higher output growth through the remainder of the year. Preventing outright weakening in the labor market by a shift in the policy stance will, on the margin, help sustain consumer spending in coming months, albeit likely at a slower pace than the rapid clip in 3Q.

One important support for consumer spending has been the low savings rate, which was down to 2.9% as of July. Next week’s final 2Q GDP report will also bring the annual revision to the GDP and GDI data, which might increase the saving rate, as we discuss in the forecast section below for next week’s final release of 2Q GDP. The annual update will include revisions back to 2019, incorporating new data and minor methodology changes. Notable updates may affect GDI, the GDP-GDI gap (as we discuss here), and the personal saving rate, particularly through revisions in net interest payments and receipts. Additionally, the update will start using QCEW data for employee compensation estimates in 1Q, likely resulting in a downward revision to compensation due to weaker employment growth compared to CES data.

Mixed messages on housing, industry

Manufacturing output rebounded strongly in the August IP report, increasing 0.9%. This followed a 0.7% drop in July due to auto plant shutdowns and Hurricane Beryl. But in the context of weak June and July readings, the quarterly run rates look soft. Total IP is tracking a 1.2%q/q, saar decline while manufacturing in tracking a 0.7% decline in 2Q after this week’s report, leaving both total and ex-auto manufacturing essentially unchanged from a year earlier. Information from regional surveys also were mixed this week. The ISM-weighted composite rose for the Empire survey but declined in the Philadelphia Fed survey, though their average in each of the last three months is the best since 2022. This is a more positive message than contained in the manufacturing PMI or ISM surveys, and we get the next flash report for the PMI this coming week. We also get the durable goods report then, where we expect the volatile aircraft segment to weigh on total orders and shipments while core capital goods series (nondefense ex-aircraft) should see only small changes, consistent with the lack of growth in the past year.

Turning to the housing market, lower mortgage rates may improve conditions in coming months, but it is too soon to see much effect yet. Existing home sales fell 2.5%m/m to a 3.86mn pace in August, essentially retouching the cycle low from October. Existing home sales measure contract closings and so are a lagging indicator, and while we do not expect a major improvement in pending home sales next week, rising weekly mortgage applications provide a glimmer of hope. Also this week, housing starts showed a rebound in August, reversing the July decline. Residential investment will still be drag on 3Q GDP growth, tracking at -10%q/q, saar, but the pace of decline will likely moderate substantially in 4Q. We also get the August results for new home sales next week, where we expect a small pullback after a strong report in July.

All quiet on the claims front

Weekly initial jobless claims dipped to 219k in the latest report, the lowest since mid-May. We suspect some residual seasonality may have contributed to higher claims in the summer, and indeed last year there was also a similarly timed drop in claims after which they stayed low through this spring. We will be looking to see if this is repeated again in coming weeks, which would be a positive labor market sign. However, Boeing’s recently announced furloughs could create a temporary rise in claims that will muddy the signal.  Boeing announced that more than 50k workers could be furloughed, with workers off one out of every four weeks. This likely won’t have much impact on the next weekly claims report, but could show up in the one after that.

Excerpted from, United States Data Watch , Michael Feroli, September 20, 2024

The first cut is the deepest

• The broad curve hit its steepest levels since early-2022 after the Fed delivered a larger 50bp cut, though the latest move has been in more bearish fashion than the bullish steepening that has defined the last 3 months...
• ...though the delivered ease was dovish, the SEP and the Chair’s comments were more balanced, and would not necessarily support a steeper curve...
• ...and this “appropriate recalibration” is being driven by risk management concerns, front-loading easing, so as to not bring the current expansion to a premature end
• Should this action support better growth outcomes in the future it could steepen the curve. The reaction across a cross section of macro asset class this week indicates this burgeoning narrative could be gaining steam
• If this narrative persists, we unwound our steepeners prematurely, after holding them for the last 9 months, though we recommended these positions as a more bullish duration proxy. This could make the case for steepening in either direction, and is also supported by our view that term premium should rise over time
• We do not want to chase the curve given how it has decoupled from the market’s terminal Fed expectations in recent days, and given that OIS forwards have more closely converged to our Fed forecast in recent weeks. With short-duration Treasury supply to be digested next week and these trades incurring negative carry, we will be patient before adding back to steepener exposure
• Intermediate valuations have begun to flag as somewhat cheap, but with money markets converging closer to our Fed forecast and yields still in the bottom 40% of their recent range, we do not feel compelled to add duration at current levels
• Initiate 91:100 weighted 3.125% Nov-41/ 2% Nov-41 flatteners
• Foreign investors purchased $56bn of long-term Treasuries in June, the most since April 2024. Demand was concentrated among investors domiciled in the Cayman Islands and the UK while net selling pressure emanated from Japan and China

Market views

Figure 11: The curve reached its steepest levels since early-2022...

2s/30s Treasury curve; bp

Source: J.P. Morgan

Figure 12: ...but in a more bearish fashion this week, as the curve appears too steep given the level of 1y1y OIS rates

2s/30s Treasury curve regressed on 1y1y OIS (%), regression over the last 6 months; bp

Source: J.P. Morgan

The Treasury curve made new steeps for the cycle, which should be of little surprise considering the Fed delivered its first ease in more than 4 years, and a bigger cut than most market participants had expected. Indeed, long-end yields rose 9bp and the 2s/30s curve steepened 6bp by a similar magnitude this week, moving back to levels last see in early-2022 ( Figure 11). However, this week’s steepening was in a bearish fashion rather than the bullish steepening that has defined the trend for the last three months or so. Given these moves, the yield curve now appears too steep relative to its primary driver: Figure 12 shows 1y1y OIS has explained about 85% of the variation in the slope of the curve over the last 6 months, and the curve now appears 20bp too steep on this basis.

With the bulk of this curve move occurring after the FOMC decision, the natural question is whether the Fed’s actions directly drove this subtle shift that we’ve noticed. From a headline perspective, the Fed leaned dovishly, delivering a 50bp ease, in line with our forecast but larger than the 25bp consensus, but many of the other details were more hawkish: the policy statement indicated that this move was motivated by progress on inflation and evolving risks to employment and inflation, it was accompanied by a hawkish dissent from Governor Bowman, the first dissent by a governor since 2005, and though the forward guidance indicates the Committee is biased to ease further, it didn’t indicate the expected pace or timing (see An appropriate recalibration, Michael Feroli, 9/18/24).

Similarly, the SEP leaned less dovishly as well as the median dot for this year points to an additional 50bp of cumulative easing over the next two meetings, 100bp of easing in 2025, and an additional 50bp of easing in 2026 to bring policy rates back to neutral ( Figure 13). The median unemployment rate projection for this year and next is 4.4%, up 0.4%-pt and 0.2%-pt, respectively, from the June projections. The inflation projections were generally revised lower by a few tenths. Dissecting the 2024 dots in greater detail, the distribution roughly shifted down by 50bp versus the June SEP, but it’s notable now that still 9 of 19 participants see at most 25bp more easing this year ( Figure 14).

Figure 13: The SEP showed upward revision to unemployment rate projections, while inflation forecasts were revised somewhat lower

Federal Reserve Summary of Economic Projections, June 2024 vs. September 2024;

		2024	2025	2026	2027	Longer run
Real GDP	Sep 24	2.0	2.0	2.0	2.0	1.8
	Jun 24	2.1	2.0	2.0		1.8
Unemployment Rate	Sep 24	4.4	4.4	4.3	4.2	4.2
	Jun 24	4.0	4.2	4.1		4.2
Core PCE	Sep 24	2.6	2.2	2.0	2.0
	Jun 24	2.8	2.3	2.0
Fed funds rate	Sep 24	4.4	3.4	2.9	2.9	2.9
	Jun 24	5.1	4.1	3.1		2.8

Source: Federal Reserve

Figure 14: The dot plot shows 9 of 19 participants see at most 25bp more easing this year

Distribution of projected midpoint of Fed funds target range for 2024, June vs. September 2024; number of participants

Source: Federal Reserve

Finally, in contrast to what we’ve observed most FOMC days over the last year or more, yields drifted higher during Chair Powell’s press conference. The Chair’s remarks read outright dovish for the first time since 2021 according to our NLP-based Hawk-Dove Report (see Fed, Powell: Press Conference, Joseph Lupton, 9/18/24), but it’s notable that he did not repeat the comments at Jackson Hole that the Committee does “not seek or welcome further cooling in labor market conditions.” Indeed Powell offered, “The labor market is in solid condition. And our intention with our policy move today is to keep it there.” Away from this, Powell offered no clues on the magnitude of future cuts, only saying “we're recalibrating policy down over time to a more neutral level.” Finally, it’s of no surprise that QT continued apace, as the Chair said “reserves are still abundant and expected to remain for some time...We know that these two things can happen side by side in a sense they’re both forms of normalization. So for a time, you can have the balance sheet shrink, but also be cutting rates.” We continue to think the sensitivity of SOFR to changes in total liquidity balances (reserves + RRP) indicates liquidity conditions are tightening and that QT is in its end state, likely to be completed by YE24 (see Interest Rate Derivatives, US Fixed Income Markets Weekly, 8/16/24). On balance, given these comments, OIS forwards priced a less dovish path for policy, implying 73bp of cuts over the balance of this year and 124bp of cuts in 2025, versus 77bp, and 133bp, respectively, prior to the FOMC ( Figure 15). Overall, we are not surprised that the Chair took a more hawkish tone in the press conference, as this likely helped build consensus for the larger 50bp cut from more hawkish members of the Committee.

Figure 15: Following this week’s Fed meeting, OIS forwards priced a less dovish path for policy

Expected change in Fed funds at FOMC meetings implied by OIS forward rates; 9/17/24 vs. 9/20/24

Source: J.P. Morgan

Against this backdrop, it seems unclear why the curve would steepen if Chair Powell poured cold water on the notion that this 50bp ease was the beginning of a front-loaded easing cycle. However, under the surface we can see how this makes sense, for a few reasons. First, it’s clear the Fed’s balance of risks have changed. Figure 16 shows that 16 of 19 FOMC participants think the risks to inflation are broadly balanced, while 11 of 16 had seen upside risks as recently as June. Second, the Committee’s balance of risks have shifted around the unemployment rate as well: while the broad majority of the FOMC saw balanced risks around the labor markets in June, 12 of 19 now see risks skewed to the upside ( Figure 17).

Figure 16: The Committee sees balanced inflation risks...

Risks to core PCE from Summary of Economic Projections; number of participants

Source: Federal Reserve

Figure 17: ...but is more concerned about a continued rise in the unemployment rate, suggesting a risk to more dovish action

Risks to the unemployment rate from Summary of Economic Projections; number of participants

Source: Federal Reserve

Accordingly, given this backdrop, one could argue that the Fed’s “appropriate recalibration” is being driven by risk management concerns, as the Committee chooses to front-load easing, so as to prevent the premature end of the current expansion. In response to a question, on whether this 50bp cut was in order to catch up, Chair Powell argued “ we think this is timely, but I think you can take this as a sign of our commitment not to get behind. So, it’s a strong move.” Indeed, a cross-section of macro asset have all behaved similarly this week, pointing toward higher growth and inflation expectations: the market-implied terminal Fed funds rate has risen modestly, TIPS breakevens widened, HG spreads narrowed, broad equity indices, gold, and crude oil all rallied ( Figure 18).

If this new narrative persists, then it may have been inappropriate to reduce our curve steepening exposure earlier this week, after having recommended various forms of these trades for the last 9 months. However, our rationale for holding steepeners was predicated on money markets pricing a faster and deeper cutting cycle, thus driving the curve steeper in a more traditionally bullish fashion than the directionality observed in recent days. We are receptive to this new narrative, and it could mean steepeners are more of an “all-seasons” trade for now, where curves can steepen bullishly if labor markets weaken further and the Fed continues to front-load cuts, or bearishly if growth remains firm and the Fed eases less, pointing toward a higher neutral rate and better anchored inflation expectations. Moreover, in the background, the rapid growth of the Treasury market continues to outstrip demand from its historically price-insensitive base, pointing toward higher term premium and steeper curves over time (see In the eye of the beholder, 9/12/23). We do not want to chase the curve given how it has decoupled from the market’s terminal Fed expectations in recent days, and given that OIS forwards have more closely converged to our Fed forecast in recent weeks. This is particularly important given the negative carry incurred in holding broad steepeners, and knowing that there is shorter-duration Treasury supply to underwrite next week, which could bias the curve flatter over the near term.

Turning to duration, valuations now appear somewhat cheap: Figure 19 shows that 10-year Treasuries appear 16bp too high after controlling for the market’s medium-term Fed policy, inflation, and growth expectations, as well as the Fed’s share of the Treasury market, a divergence of nearly close to one standard deviation. On margin, this backdrop skews us toward a more bullish lean on duration, but with money markets more closely converging relative to our own forecast and 10-year yields still in the lower 40% of the 40bp range they have traded for the past six weeks, we are not compelled to add duration at current levels.

Figure 19: Intermediate Treasury yields are somewhat high after controlling for their fundamental drivers, but with markets converging closer to our Fed forecast and yields below their recent averages, we are not compelled to add duration at current levels

Residual from J.P. Morgan 10-year Treasury fair-value model*; bp

* Regression of 10-year Treasury yields on 5Yx5Y seasonally-adjusted TIPS breakevens (%), 3m3m OIS rates (%), Fed policy guidance (months), J.P. Morgan US Forecast Revision Index (%), and SOMA share of outstanding marketable US Treasury debt, excluding T-bills (%). Regression over the last 5-years. R-squared = 97.8%, SE = 20bp

Source: J.P. Morgan, Federal Reserve, US Treasury

Turning to relative value, we’ve noticed that originally issued 20-year bonds maturing in late-2041 and early-2042 have cheapened relative to our par curve, whereas originally issued 30-year bonds with similar maturities have outperformed. Clearly, the recent steepening of the Treasury curve supports wider yield spreads between higher-coupon OI 30s, and lower-coupon OI 20s with the same maturity, given significant duration differences. However, this dynamic is not explained by duration profiles: Figure 20 displays the securities in the 2040-2043 basket, sorted by modified duration and shows that late-41/ early-42 OI 30s appear rich relative to other securities with similar duration. In particular, the 3.125% Nov-41/ 2% Nov-41 curve appears 5.0bp too steep relative to the shape of the 15s/20s Treasury par curve ( Figure 21). Hence, we recommend initiating 91:100 weighted 3.125% Nov-41/ 2% Nov-41 flatteners (see Trade recommendations).

Figure 20: Late-41 and early-42 OI 30s appear too rich relative to surrounding securities

Yield and duration of select Treasury securities in the 2040-2043 basket plotted versus modified duration (years); %

Source: J.P. Morgan

Figure 21: The 3.125% Nov-41 / 2% Nov-41 curve appears 5bp too steep relative to our par curve

Residual of 3.125% Nov-41 / 2% Nov-41 regressed on 15s/20s Treasury par curve over the last 6 months; bp

R-squared: 71.5%; S.E.: 1.4bp

Source: J.P. Morgan

July TIC update

Turning to the monthly TIC data, foreign investors purchased $56bn long-term Treasuries over the month of July, above the six-month average, and the strongest month of net purchases since April 2024. Private investors added $88bn, the firmest month of buying since November 2023, which includes $6bn of demand from International and Regional Organizations (IROs). Meanwhile foreign official investors net sold $32bn of long-term Treasuries, the most since November 2023 ( Figure 22). Looking ahead, we expect demand from foreign private investors to moderate from the strong pace earlier this year, especially if some of the YTD demand was to position for the start of the Fed easing cycle (see Treasuries, US Fixed Income Markets Weekly, 9/13/24).

Figure 22: Foreign investors on net purchased $56bn of long-term Treasuries in July, the most since April...

Total net sales of long-term Treasuries by US residents by sector and 6-mo. moving average; $bn

Source: Treasury International Capital System

Figure 23: ...and geographically, demand was concentrated in the Cayman Islands and the United Kingdom

Net sales of long-term Treasuries by US residents from May- July 2024 by region as well as cumulative net sales year-to-date; $bn

Region	Jul-24	Jun-24	May-24	YTD 2024
Cayman Islands	50.3	-9.2	3.7	53.8
UK	8.1	-5.0	15.2	34.1
EM ex-China**	4.4	-2.3	13.1	56.0
Oil Exporters*	1.6	-2.2	2.1	11.9
Ireland	-2.6	0.2	5.7	10.6
Euro Area†	-5.0	20.7	10.0	78.1
Japan	-15.2	-30.1	-5.8	-6.4
China	-23.1	5.9	-23.9	-61.9

*Oil exporters include all OPEC members (excluding Equatorial Guinea & Angola) and also include Bahrain, Indonesia & Oman
**EM ex-China includes Russia, Brazil, African countries, Taiwan (China), South Korea and Mexico
† Euro area include Austria, Belgium, Cyprus, Estonia, Finland, France, Germany, Greece, Ireland, Italy, Latvia, Luxembourg, Malta, Netherlands, Portugal, Slovakia, Slovenia, Spain

Source: Treasury International Capital System

Turning to geographical composition, Figure 23 shows that demand in July was concentrated among investors domiciled in the Cayman Islands and the United Kingdom, who bought $50bn and $8bn of long-term US Treasuries, respectively. As a reminder, we tend to think of flows from these countries as largely reflecting hedge fund demand domiciled in these jurisdictions. Interestingly European investors net sold $5bn of Treasuries in July, though we would expect this outflow to reverse given that Treasuries remain somewhat attractive for Euro-funded investors on a FX-hedged basis. Meanwhile investors domiciled in Japan net sold $15bn of long-term US Treasury securities, their fourth consecutive month of outflows. While this is in line with the still significantly negative FX-hedged yield pickup Japanese investors face buying Treasuries, it contrasts with MoF data which indicate investors added roughly $1tn JPY of US bonds in July. Meanwhile, according to the MoF, Japanese investors appear to have bought $7.33tn JPY of foreign bonds in August, the largest monthly purchase over the past year, largely stemming from bank demand, though we do not yet know the regional composition of this net buying (see Japan Flows in Pictures, Takafumi Yamawaki, 9/9/24). Finally Chinese investors net sold $23bn Treasuries, and this comes even as gold reserves remain unchanged in recent months (see China: Gauging exporters’ “USD hoarding”, Tingting Ge, 9/8/24).

Trade recommendations

• Initiate 91:100 weighted 3.125% Nov-41/ 2% Nov-41 flatteners

-   Buy 91% risk, or $20mn notional of T 3.125% Nov-41s (yield: 4.02%; bpv: $1146/mn)
-   Buy 100% risk, or $24.7mn notional of T 2% Nov-41s (yield: 4.121%; bpv: $1021/mn)
- Weighted spread is 46.3bp. One-month weighted carry is 0bp and roll is 0.1bp

• Unwound 3s/30s steepeners

-   Unwound long 100% risk, or $180mn notional of T 3.75% Aug-27s
-   Unwound short 100% risk, or $28mn notional of T 4.25% Aug-54s
- (Unwound in US Treasury Market Daily, 9/18/2: P/L since inception: 0.2bp)

• Maintain 75%/6% weighted 5s/10s/30s belly-cheapening butterflies

-   Stay long 75% risk, or $43mn notional of T 4.625% Sep-28s
- Stay short 100% risk, or $33.3mn notional of T 3.875% Aug-33s
- Stay long 6% risk, or $1mn notional of T 4.125% Aug-53s
- (US Fixed Income Markets Weekly, 9/29/23: P/L since inception: -2.0bp)

• Maintain 23:84 weighted 2s/7s/10s belly-richening butterflies

-   Stay short 23% risk, or $28mn notional of T 4.375% Jul-26s
- Stay long 100% risk, or $40.2mn notional of T 4.125% Jul-31s
- Stay short 84% risk, or $23mn notional of T 3.875% Aug-34s
- (US Treasury Market Daily, 8/21/24: P/L since inception: -2.5bp)

• The post-FOMC price action within rates and across other asset classes reinforce the technically-derived bias we had favoring bull market consolidation going into the Sep meeting….
• ...The momentum divergence signals that triggered across multiple DM bond markets and across the respective points on the curves suggest an increased probability for backing and filling to higher yields into early-Oct and potentially more range trading through early-fall...
• ...Longer duration markets responded to those signals already as front-end yields just got sticky near key resistance levels. We suspect further risk-on trends across equity and commodity markets can help drive further bear steepening over the very near-term, but we also think those trends are self-limited via the likely TIPS breakevens widening that often correlates with all of the above. At some point, the market may rethink the Fed path to neutral if inflation markets widen too.
• We suspect 2-year note weakness will find buying interest near 3.75% and see an outside chance a backup extends as far as 3.85%. Look for the 30-year bond to find material buying interest near support surrounding 4.25%. All else equal, the recent developments reinforce the idea that 3.93-4.00% will put a floor under the long-end yield through the fall.

Longer-end US yields lead the retracement to higher levels; but the technical setup at the front-end leaves the front-end vulnerable as well

The post-FOMC price action reinforced our technically-derived bias going into the meeting— that the rally deceleration into key chart-based resistance levels favored a period of consolidation. Over the near-term, we think that setup, the associated pattern-based signals as shown in the charts point to additional backing and filling in the early-fall. We would view such price action as a consolidation within a developing bull market. For now, we are most interested in watching how the yield curves evolve, how risky markets respond to the rate decision and data, and how both of those impact inflation markets. The late-week price action saw an impulse of risk-on, bear steepening, and wider TIPS breakevens. We believe that market reaction represents one side of a very bifurcated macro-risk distribution. With the yield curve having been inverted for many quarters and the policy rate materially above what most assume to be neutral for most of that period, a sequence of policy eases has the potential to derail the inertia associated with weakening labor demand before that trend leads to a drop in consumption and/or layoffs. In that case, risky markets will eventually start to price in a rebound in the manufacturing data that has been under pressure for most of the last two years. The late-week price action gives some indication of that dynamic across markets— bear steepening, wider TIPS breakevens, rallying industrial commodities, and higher equities (specifically small cap and cyclicals). While we think that can extend over the very short-term, we do think the pass through to inflation markets and the likely resulting impact on the market-implied policy path creates a self-limiting dynamic.

At the other end of that very bimodal distribution profile, the convergence of the actual policy rate to neutral and lagged impact on the economy would trail the impact of shrinking labor demand, which in turn would lead to a sharp drop in end demand and/or layoffs. Naturally, that path would likely lead to a rapid repricing to an accommodative policy rate, a significant drop in equity markets given their high multiple and EPS growth expectations, lower commodity prices, and lower TIPS breakevens. In the absence of a sharp break in the labor data in the weeks ahead, or a rapid risk unwind for whatever reason, it is unlikely markets will get a clear answer as to which path the macro setup is heading toward in the early-fall. The overall setup bolsters the technically-derived near-term outlook for consolidation across many market trends. As such, we’ve already shrunken our 5s/30s curve trade, we suggest taking profits on the 2s/5s curve steepener, and closing out the 10-year Gilts long we entered earlier in the month. Our 10-year TIPS breakevens trade was stopped out for a wash after the post-Fed move. Lastly, we will hold the remaining 5s/30s steepener for now, as the S&P 500 attempted break to new highs is still tentative. As noted above, a continued risk-on move has the potential to help bear steepen the curve more over the very near-term. Additionally, an immediate failure to hold that breakout can damage sentiment and cause a re-focus on the risk associated with the negative tail outcome that could become self-feeding if equities start to build bearish momentum.

For key levels, we are looking for the 2-year note to trade mostly in a 3.55-3.75% range, and see an outside chance the front-end backs up as far as the 3.835-3.89% support zone in the weeks ahead ( Figure 25). At the long-end, look for the 30-year bond backup to stretch as far as 4.25%, but at that point, we think buying interest will help stabilize the market. Unless the data deteriorates materially, this week’s developments and the responses across asset classes reinforces the idea that 3.95-4.00% will likely keep a floor under the yield through the fall and potentially beyond ( Figure 28).

Figure 25: The 2-year note rally stalls near 3.525-3.555% expected resistance, but has outperformed a more substantial post-FOMC setback further out on the curve. We see tactical support at 3.71-3.75% and think a backup into the 3.80s would find a meaningful amount of buying interest. In the absence of sharp equity weakness or an immediate deterioration in the labor data, we expect the front-end to trade in a tight range in the early-fall period.

2-year note yield, daily bars with momentum divergence signals; %

Source: J.P. Morgan, CQG

Figure 26: The 5-year note hasn’t traded as sticky as the very front-end, but the belly also looks set to get very sticky and trade in a very tight range in the early-fall period. The rally stalled near the Oct 2023 channel, now at 3.435%. Look for tactical support at 3.61% with other levels at 3.73% and 3.755%. We think that zone caps yields in the near-term.

5-year note, daily bars with momentum divergence signals; %

Source: J.P. Morgan, CQG

Figure 27: The 10-year note stages a more pronounced setback from the 3.60% yield low and nearby resistance at the 3.625% Apr 2023 78.6% retrace and 3.64% Mar-May 2023 pattern breakdown. Recent bearish momentum divergence sell signals (green arrows) point to an increased probability for further backing and filling into mid-Oct. We expect the 3.91% 50-day moving average, 3.91-4.02% Aug pattern cheaps, and 4.03% Apr 38.2% retrace to put a ceiling over yields through the fall.

10-year note yield, daily bars with momentum divergence and TY premium-weighted Put/Call z-score signals; %

Source: J.P. Morgan, CQG, CME

Figure 28: The 30-year bond saw the largest response to the recent momentum divergence sell signals, rejection of the 3.94% Dec 2023 yield low, and FOMC decision. The bond is already pressed into the 4.065-4.095% tactical support zone. We see other layers at 4.235-4.26% and then 4.315-4.345%. We suspect bear steepening will have a limit given its typical correlation with commodity price trends and inflation markets. Look for that dynamic to keep the bond yield capped at 4.25% through the fall.

30-year bond yield, daily bars with momentum divergence signals; %

Source: J.P. Morgan, CQG

Figure 29: The 5s/30s curve steepening trend reaches the 59bp Mar 2021 50% retrace after the FOMC meeting. The outcome and market response bolsters our view that the curve can start to trade in a range below that level and above support at the 43bp 50-day moving average, 41.5bp early-summer 2023 peak, and 37bp Aug low. We suspect the current bear steepening dynamic will have a limit as it also pulls inflation markets with it and eventually can impact front end pricing. On the chart, we would view a fall range as a consolidation pattern after the summer 2022-2024 base pattern breakout. That formation favors an acceleration to the upside into 2025.

5s/30s curve, daily closes; bp

Source: J.P. Morgan, CQG

Figure 30: 10-year TIPS breakevens widen from the 203bp cycle low and fully mean revert to the upper end of the recent range. That stopped us out of our recent tightening trade at breakeven. We still see the 218-224bp area as a likely ceiling and key bifurcation zone. If the market happens to widen back into the 2023-2024 pattern, we think the move can accelerate quickly and quickly move back to the 235-243bp resistance zone.

10-year TIPS breakevens, daily closes; bp

Source: J.P. Morgan, CQG

This report was excerpted from Global Fixed Income Technical, Jason Hunter, September 20, 2024

Two roads diverged in a wood...

• TIPS continued to outperform, supported by strong economic data and a favorable FOMC outcome. In the wake of the Fed meeting, Thursday’s 10-year reopening auction was easily digested
• The fundamental backdrop has shifted more bullishly for TIPS over the last two weeks, given the strong action from the Fed combined with growth data that have driven upward revisions to our GDP forecasts and provided further conviction in our soft landing call
• That said, we are hesitant to chase the recent widening. The Fed’s data dependent stance likely leaves breakevens vulnerable to a correction lower, as we think the amount of easing currently priced in the forwards will only materialize alongside further cooling in labor markets...
• ...Additionally, our medium-term outlook on inflation remains unchanged. We believe core inflation will soften gradually, supported by core goods deflation and softening service price pressures
• Fixings are priced close to our own forecasts, implying an average monthly pace of core CPI inflation near 0.27% in the next few months and near 0.25% through 1Q25. Additionally, breakevens appear fairly valued after adjusting for the repricing in the Fronts/Reds curve and broad commodity prices
• Thus, we do not think breakevens are terribly compelling at current levels. For now, we recommend holding energy-hedged 5s/10s breakeven curve steepeners
   

Market views

Over the past week, TIPS continued to outperform, with 5-, 10-, and 30-year breakevens 9bp, 7bp, and 6bp wider net of carry, supported by strong economic data and a favorable FOMC outcome. Specifically, Wednesday’s 50bp cut was paired with Powell’s optimistic message that progress on inflation allowed for a forceful “recalibration” of policy to preserve the currently strong labor market. In the wake of the FOMC meeting, Thursday’s $17bn 10-year TIPS reopening auction was easily digested, clearing with just a 0.1bp tail as end-user demand rose to 93.4%, the highest since January ( Figure 31). Despite the continued decline in real yields, 5-year breakevens are now 18bp wider over the last two weeks, trading near their widest levels since just prior to the weak July employment report released on August 2^nd ( Figure 32). IOTAs were broadly stable. Outflows from TIPS-focused funds continued but moderated from the heavy pace observed in recent weeks, with these funds recording $54mn of outflows in the week through Wednesday.

Figure 31: This week’s 10-year reopening auction was easily digested with 93.4% end-user demand

10-year TIPS auction tail (LHS; bp) vs. end-user demand* (RHS; %)

* Reflects the sum of indirect and direct participation

Source: US Treasury, J.P. Morgan

Figure 32: Front-end real yields were range-bound while breakevens widened sharply

1-month seasonally-adjusted forward 5-year TIPS real yield (LHS; %) vs. breakeven (RHS; bp)

Source: J.P. Morgan

Taking a step back, the fundamental backdrop for breakevens has shifted bullishly over the last two weeks. In recent months, we have argued that TIPS performance would largely depend on whether the Fed was perceived to be cutting rates preemptively to preserve the expansion or reactively in response to a weakening labor market (see TIPS Strategy, 8/23/24). The willingness of the Fed to act aggressively to bring policy closer to neutral this week reduces the risk that the Committee is caught behind the curve. In fact, Chair Powell stated more than once during the press conference that “you can take this as a sign of our commitment not to get behind.” This strong action from the Fed also comes against a backdrop of continued upside surprises in the growth data, adding conviction to our soft landing view. Importantly, following this week’s August retail sales report, our economists have revised up their tracking of 3Q GDP growth to 2.5%, from 1.5% previously, and take their 4Q GDP growth target from 1% to 1.25% (see US Weekly Prospects, Michael Feroli, 9/20/24). Moreover, if this forecast is realized, the step down in the policy rate to mitigate risks that don’t materialize, alongside easier financial conditions and lower headline inflation, should provide a tailwind to real consumption, keeping recession risks in check heading into 2025.

That said, we are hesitant to chase the widening in breakevens for a few reasons.First, while the recalibration in policy rates arguably increases the probability of a soft landing, the Fed has made it clear that the pace and extent of future easing will depend on the evolution of the data. Thus, we think the easing priced into the forward curve through 2025 will only materialize in an environment in which labor markets continue to loosen, keeping growth risks weighted to the downside, especially given Powell’s comment that he believes the neutral policy rate is likely “significantly higher” than it was before the pandemic. Second, this is likely to keep inflation markets vulnerable to a correction lower if labor market data disappoint, as we have seen in the aftermath of each of the last two employment reports. Despite our forecast for a soft landing, we recognize that labor market data and growth data have been disconnected for a number of months and are likely to remain so into the fourth quarter. Even as our economists have revised up their growth tracking, they still look for further slowing in payroll growth in coming months, and look for the unemployment rate to tick up to 4.5% by the end of the year, above the Fed’s median projection. In a recent note, they show that almost all of the recent U-rate rise can be explained by declining job finding rates, and not rising separations, which is not typical during recessions and could be consistent with a labor market normalizing around a higher NAIRU (see US: Reaching u* before r*, Murat Tasci, 9/17/24).

Third, despite a stronger growth outlook, we also maintain a forecast of gradually cooling core inflation and our commodity strategists maintain a bearish outlook on energy prices in 2025. We recognize that core CPI inflation could stay somewhat elevated in the next couple of months supported by strength in used car prices and airfares, and we see the potential for firming early in 2025 as well, driven by residual seasonality, but this view is largely priced into the fixings curve, which implies an average monthly pace of core CPI near 0.27% through the fall and roughly 0.25% through 1Q25 ( Figure 33). Meanwhile, we expect the broader trend of core goods deflation and softening price pressures from slower wage growth and looser labor markets to remain intact. Rent inflation, which Powell noted is “the one piece that is kind of dragging a bit,” should gradually moderate, and we expect the underlying trend in rental inflation is likely around 4% annualized (see TIPS Strategy, 9/13/25). It’s also worth recognizing that there is empirical evidence that contractionary monetary policy could have had a hand in keeping rent inflation elevated, given that high mortgage rates have kept housing mobility low and demand for rental units high.¹ It is therefore reasonable to expect the initial response of less restrictive monetary policy could support further slowing in rent inflation, on the margin.

Figure 33: Front-end fixings imply core CPI averaging near a 0.27% monthly pace in coming months

CPI fixings and implied inflation rates*; units as indicated

Month	Fixing	Implied headline % m/m sa	Implied headline % oya	Implied core % m/m sa	Implied core % oya
Sep-24	314.845	0.07%	2.29%	0.25%	3.19%
Oct-24	314.890	0.13%	2.35%	0.27%	3.22%
Nov-24	314.503	0.24%	2.43%	0.29%	3.20%
Dec-24	314.322	0.27%	2.47%	0.11%	3.03%
Jan-25	315.620	0.17%	2.34%	0.25%	2.88%
Feb-25	316.765	0.19%	2.07%	0.21%	2.73%
Mar-25	318.231	0.20%	1.89%	0.41%	2.78%
Apr-25	319.128	0.21%	1.78%	0.22%	2.70%
May-25	319.923	0.09%	1.86%	0.05%	2.59%
Jun-25	320.430	0.07%	1.99%	-0.05%	2.47%
Jul-25	321.001	0.22%	2.02%	0.28%	2.58%

* To derive market-implied core inflation rates, J.P. Morgan forecasts for food and energy CPI are used

Source: J.P. Morgan

Figure 34: Breakevens appear roughly in line with our fair value framework

Residual on J.P. Morgan 5-year breakeven fair value model; bp

* 1m-forward, seasonally-adjusted breakevens are regressed on the J.P. Morgan Commodity Curve Index (JPMCCI) as well as its square and the 3mx3m/15mx3m OIS curve; regression over the last 7years; R2=94%, SE=12.7bp

Source: J.P. Morgan

Against this backdrop, we think breakevens are not terribly compelling at current valuations. Figure 34 shows that 5-year breakevens are fairly valued after adjusting for the repricing in Fronts/Reds and broad commodities. Oil prices have rebounded from cheap levels toward our commodity strategists’ year-end target, and the money market curve, which we use as a macro risk proxy, could revert flatter again if labor market data show further cooling as we expect. Thus, we recommend waiting for cheaper levels before adding widening exposure. For now, we favor structures that offer asymmetric risk/reward. Notably, we believe 5s/10s energy-hedged breakeven curve steepeners remain attractive. The front end of the forward curve has been most reactive to shifts in the macro narrative over the last couple of months, with 1Yx1Y, for example, leading the way higher over the past two weeks ( Figure 35). Figure 36 shows that 5s/10s, as proxied by the 2Yx3Y/5Yx5Y inflation swap curve is back at the low end of its YTD range. With labor market data in early October likely to show further cooling, but solidify market expectations for further easing, we think this should drive the 5-year sector to underperform, while 5Yx5Y breakevens could richen further. Meanwhile, as noted above, we think the Fed’s recalibration in policy is unlikely to drive near-term inflation higher, but further strength in risk assets and a belief that the Fed is tolerant of above-target inflation could also drive longer-run forwards to outperform along the curve. Against this backdrop, we maintain 5s/10s breakeven curve steepeners paired with a long in Brent futures.

Figure 35: The front end of the forward curve led the way higher…

1Yx1Y, 2Yx3Y, and 5Yx5Y inflation swap rates; %

Source: J.P. Morgan

Figure 36: ...and the 2Yx3Y/5Yx5Y curve is back near its flattest levels YTD

2Yx3Y/5Yx5Y inflation swap curve; bp

Source: J.P. Morgan

Trade recommendations

• Maintain energy-hedged 5s/10s breakeven curve steepeners

- Stay short 100% risk, or $110mn notional of TII 2.125% Apr-29s
- Stay long 100% risk, or $113.8mn notional of T 4.625% Apr-29s
- Stay long 100% risk, or $54mn notional of TII 1.875% Jul-34s
- Stay short 100% risk, or $59.6mn notional of T 3.875% Aug-34s
- Stay long 15 Brent futures (COZ4)
(See TIPS Strategy, 9/13/24). P/L since inception: -2.0bp

From Dovish Pause to Hawkish Easing

• The Fed delivered its first rate cut of this cycle on Wednesday, and the magnitude of the cut was on the larger side of consensus expectations, but in line with our own. However, passing of the FOMC meeting has not translated into significantly better clarity on the path of the funds rate going forward
• With policy uncertainty still elevated, given the overall bearish sentiment emanating from this week's 50bp rate cut, and given upside risk to the terminal rate relative to current forwards, we continue to favor gaining asymmetric exposure to higher yields in the Reds ...
• … initiate weighted Greens / 15s flatteners. Empirical evidence suggests that this curve will likely be locally flat, but should outperform asymmetrically in a selloff ...
• … and initiate conditional exposure to a flatter 1s/5s curve in a sell-off using 3M expiry payer swaptions. The curve is weighted to make it premium-neutral at inception, while still maintaining directionality
• Macro drivers do not point to a clear directional bias in swap spreads and we maintain our neutral stance on swap spreads across the curve ...
• … but recommend initiating exposure to 2s/3s swap spread curve flatteners. This curve appears too steep, and we believe investors such as banks will find it attractive to deploy their spread risk budget on 2s versus 3s given the attractive pick over IOR relative to risk
• Markets are likely to remain in a lull until the next employment report, causing us to turn tactically bearish on short expiry volatility in intermediate and longer tails, and turn neutral on short expiry volatility on short tails
• At a more nuanced level however, we recommend initiating long exposure to the volatility of the second principal factor, by buying 1Yx3Y straddles versus 105% of the vega-weight in 1Yx10Y straddles. This weighted spread is at year-long lows, and the package is approximately carry-neutral, which makes it attractive to initiate in our view
   

From Dovish Pause to Hawkish Easing

As widely expected, the Fed delivered its first rate cut of this cycle on Wednesday. However, the magnitude of the cut was on the larger side of consensus expectations (but in line with our own). Expectations going into the meeting were roughly evenly split between a 25bp and a 50bp cut. But despite the larger-than-forwards rate cut, the broader themes emanating from the meeting were hawkish relative to expectations. The tone of the post-meeting press conference was somewhat hawkish, causing yields to rise rather than fall as has typically been the case in recent meetings. Fed Chair Powell offered little guidance regarding the magnitude of future rate cuts, and merely stressed the data-dependent nature of future policy actions. In addition, 9 out of 19 members of the FOMC project at most 25bp of additional rate cuts by year-end 2024, as indicated in the "dots". Moreover, this week's 50bp rate cut was not unanimous and came with a hawkish dissent from Governor Bowman.

Thus, all in all, the passing of the FOMC meeting has not translated into better clarity on the path of the funds rate going forward. In Figure 1, we resolve the implied probability density functions associated with Z4 and M5 3-month SOFR futures into a set of intuitive weights (that must sum to 1) on individual policy outcome scenarios (see What’s the rush? for details). As can be seen there, except for a minor recalibration of YE24 weights in response to the realization of a 50bp rate cut, there is not much of a change in weights associated with tail outcomes. Thus, we remain very much in a market environment that is dominated by the lack of policy path clarity.

The theme of policy uncertainty is also echoed by the dot plots themselves. Taking a closer look at the distribution of the year-end dots and the long-run dots, it is clear that the bands are widening. In particular, dispersion (which we measure simply as the difference between the 75th and 25th percentile dots) has widened at the long-run and YE26 horizons, in each of the past two SEP meetings (Figure 2). This too hints at rising uncertainty, but of a different kind that is centered around the terminal rate rather than near term actions by the Fed. Indeed, this week's bear-steepening of the curve is likely a reflection of the emergence of terminal rates as a second policy factor driving rates this week.

Figure 1: The passing of the FOMC meeting has not translated into better clarity on the path of the funds rate going forward

Weights on YE24 and 1H25 policy rate scenarios representing a range of different Fed Funds rates, as calculated from a decomposition of the implied probability distribution associated with Dec 2024 and June 2025 SOFR futures*; 9/17 (before FOMC) and 9/18 (After FOMC)

* We enumerate a list of scenario-specific Normal distributions with fixed standard deviations and means that are separated by 25bp, and then require the implied distribution to be a weighted combination of these individual distributions. The weights are then solved for, by fitting to the observed prices of calls and puts at various different strikes. For more details of our approach, see What’s the rush?.

Source: J.P. Morgan., CME

Figure 2: Policy uncertainty, as measured by the Fed dots, has widened at the long-run and YE26 horizons in each of the past two SEP meetings

Dispersion* in YE24, YE25, YE26, and long run Fed Funds rate projection by FOMC members at the past 3 SEP meetings; %

* Dispserion is defined as the 75^th percentile minus 25^th percentile dot projections

Source: J.P. Morgan., Bloomberg Finance L.P.

All of this influences our approach to trading the Rates markets in the near term. First and foremost, continuing policy uncertainty means that outright directional trading themes such as duration or curve trades are likely to offer poor risk-reward in the current environment. That said, OIS forwards in the Reds appear to be well below levels indicated by the YE25 median dot (Figure 3). To be sure, forwards are at levels not inconsistent with our own Fed funds rate forecast, and FOMC median projections have often lagged the market. Nonetheless, given the overall bearish sentiment emanating from this week's 50bp rate cut, we believe it is worth exploring asymmetric ways to position for a rise in yield levels in the Reds. One attractive way to do this is via weighted Greens/15s flatteners. As seen in Figure 4, the 2Yx1Y / 3Mx15Y swap curve (0.9:1.0 weighted) is empirically likely to be insensitive to small moves in Reds locally, but is likely to flatten at an accelerating rate in a selloff. This makes it attractive as an asymmetric play on higher yield levels in the Reds and we recommend this trade (see Trade recommendations).

Figure 3: OIS forwards are at levels not inconsistent with our own Fed funds rate forecast, but are well below the Fed’s dots in the Reds

Forward 1M OIS rates at FOMC meeting dates in 2024 and 2025 as of 9/13 and 9/18, the JPM forecast for the Fed funds rate on those dates, and the YE24 and YE25 median Fed dots; %

Source: J.P. Morgan., Bloomberg Finance L.P.

Figure 4: The 0.9:1 weighted Greens/15s curve is likely to be locally stable but flatten at an accelerating pace in a selloff

3Mx15Y minus 0.9 * 2Yx1Y swap yield curve, versus 1Yx1Y swap yield; past 6 months, %

Source: J.P. Morgan.

A variation on this theme is to initiate conditional exposure to a flatter 1s/5s curve in a selloff, using 3M expiry payer swaptions. Given current implied volatility levels, doing this trade premium neutral requires selling ~2% more risk in the 3Mx5Y payer swaption leg, versus buying 3Mx1Y payer swaptions. But that additional risk is small relative to the recent beta of this curve versus front end yields; as a result, this 1:1.02 weighted 1s/5s curve is still highly directional with yields and likely to flatten in a selloff (Figure 5). Therefore, we recommend initiating exposure to a flatter 1s/5s swap curve (100:102 risk weighted) in a selloff using 3M expiry payer swaptions (see Trade recommendations).

Figure 5: The 1:1.02 weighted 3M forward 1s/5s swap curve is negatively correlated with front end yields and appears too steep relative to its recent empirical relationship

1.02 * 3Mx5Y minus 3Mx1Y swap yield curve, versus 3Mx1Y swap yield; past 3 months, %

Source: J.P. Morgan.

Swap spreads

Swap spreads are modestly wider across the curve, with the 20Y sector outperforming by the most. For instance, maturity matched (and roll-adjusted) swap spreads are roughly 0.5-1bp wider across most of the curve, but 2.5bp wider in the 20-year sector (Figure 6). The recent outperformance of the 20Y sector relative to both 10s and 30s is a reflection of relative value convergence - we had highlighted the cheapness of swap spreads in the 20-year sector two weeks ago (see Rates, unlike the economy, are not yet in “equipoise”), and these moves are in line with our expectations.

Figure 6: Swap spreads are modestly wider across the curve, with the 20Y sector outperforming

Selected statistics for maturity matched SOFR swap spreads, 9/13 - 9/20; bp

Source: J.P. Morgan.

Little has changed with respect to the drivers of swap spreads in the near term. As a brief summary, our approach to swap spreads begins with the term structure of swap spreads. As we have observed, if one examines the slope of a cross-sectional plot of maturity matched swap spreads at various benchmark maturities versus modified duration on any given day, the negative of the slope represents term funding premium (which tends to be driven by supply and demand), and the intercept (which we refer to as zero-duration spreads) is driven by factors influencing the front end. A detailed discussion of our framework can be found in Term Funding Premium and the Term Structure of SOFR Swap Spreads, and our recently revised model for zero-duration spreads can be found in Schrodinger’s Cut. Looking ahead, term funding premium appears to have become quite stable recently, and zero-duration spreads are now only modestly narrow to fair value.

Thus, macro drivers do not currently point to a convincing directional bias in swap spreads one way or the other, and we remain neutral on swap spreads for now. However, on a relative basis, 2Y swap spreads appear far too narrow relative to 3Y swap spreads (Figure 7). We believe this is unlikely to be sustained, because it represents an attractive opportunity for investors such as banks, for at least three reasons. First, with OTR maturity 2-year notes now trading near SOFR+19bp, this translates into a pick-up of ~13bp over IOR, given that the 2-week moving average of SOFR minus IOR is around -6bp. The corresponding pickup is ~3bp higher in the 3Y sector, but spread volatility in 2s and 3s is rather similar and the spread duration in the 3Y point is about 50% greater, which makes the risk-reward ratio more favorable for 2Y spread wideners at inception in comparison to 3Y spread wideners. Second, the carry-to-risk ratio improves rapidly in the case of 2Y spread wideners as the position ages and spread duration falls. Third, asset swapping USTs leaves the investor in a receive-SOFR position, which is desirable as we approach the end-game of QT and a possible scarcity of liquidity. We recently discussed a Fed staff paper that analyzed the ampleness of Reserves by examining the sensitivity of the effective funds rate to shocks in the quantity of Reserves (see When Are Central Bank Reserves Ample?, Liberty Street Economics, 8/13/2024), and generalized it to include RRP balances in the definition of liquidity and study the response of SOFR instead of the effective funds rate (see Hopscotch). As we noted there, this suggests that we are steadily heading towards tighter liquidity conditions, which is why we continue to see QT as now likely in its end game. Indeed, SOFR has been steadily rising relative to IOR in recent months (Figure 8). As RRP balances continue to fall over the remainder of the year (they fell to a low of ~$240bn intra week), repo rates and SOFR will likely remain biased higher (even though it may tactically normalize in the next two weeks as our Short Term strategists expect), even possibly necessitating an eventual reduction in IOR if repo rates were to rise enough. Given these looming possibilities, banks will likely find it preferable to deploy their spread risk budget in 2Y spread wideners rather in 3s. Therefore, we now recommend positioning for a 2s/3s maturity matched swap spread curve flattener (see Trade recommendations).   

Figure 7: The 2s/3s maturity matched swap spread curve is near its steepest levels of the year

3Y minus 2Y maturity matched swap spread differential, bp; 01/2024 - present

Source: J.P. Morgan.

Figure 8: SOFR has been steadily rising relative to the interest rate on Reserves

Rolling 2-week moving average of the O/N SOFR minus IOR differential, %; past 6 months

Source: J.P. Morgan.

Options

It was a big week for fixed income markets as the Fed delivered its first rate cut of the cycle. Implieds underperformed and declined across the surface over the week, with the exception of 30-year tails. Shorter expiry and shorter tails led the move lower, with 6Mx2Y falling 0.6 bp/day over the week, while 6Mx30Y was up by about 0.1 bp/day (Figure 9). As Figure 9 also shows, the declines in implied volatility continue to be in excess of what would be expected by our longer term empirical model estimate of fair value (for details see our 2024 Mid-Year Outlook). However, even after these declines, shorter expiry volatility continues to appear rich with respect to our fair value model to the tune of 0.5 to 0.8 bp/day in 6M expiry swaptions and 0.4 bp/day in 1Y expiry swaptions.

Figure 9: Implieds underperformed and declined across the surface over the week, with the exception of 30-year tails

Actual, Fair Value*, and Residual for various SOFR swaption structures, as of 9/13, 9/19, and changes between those dates; bp/day

* For details on our fair value framework for SOFR swaption implied volatility, see Interest Rate Derivatives 2024 Mid-Year Outlook

Source: J.P. Morgan.

As we get past the FOMC meeting this week, we turn bearish on short expiry volatility in intermediate and longer tails, and turn neutral on shorter tails. There are now reasons to be tactically bearish on gamma. As we noted above, implieds continue to appear rich to our fair value estimate. To be sure, we had maintained a long gamma bias in the past few weeks despite rich valuations, because of elevated policy uncertainty and the potential for greater jump risk. Even though this week's FOMC meeting did little to clarify the medium term path of policy rates, we still think a tactical short gamma stance is warranted, as we discuss below. 

One of the takeaways from Powell's press conference was the shift in focus from inflation to labor markets. Although Chair Powell was careful to not declare victory on inflation just yet, it seemed apparent that his focus was more on the labor markets. As our Economists have also noted (see An appropriate recalibration, M. Feroli, 9/18/2024), the size of the next cut is very much contingent on labor market data. Given this heightened importance of the Jobs report, it is very likely that markets will simply remain in a lull until then.

But given the significant event risk associated with Payrolls day, and the nonlinear nature of gamma returns, long gamma positions can remain in the black even with somewhat lower volatility in the quiet period leading up to the Jobs report. The question therefore becomes - how much is the event risk worth, and what amount of realized volatility is needed in the next two weeks for long gamma positions to breakeven. We can quantify this by noting that the event risk around Payrolls day has been worth ~12bp in the 5Y sector on average, as measured by the inter-dealer 1-day options market. Using this as an estimate of the event-day volatility, we can solve for the delivered volatility needed during the lead-up period in order for long straddle positions to break even based on current implied volatility levels. We do this for several different tails in Figure 10 - as can be seen, such an analysis suggests that markets must deliver 5.5 - 6bp/day in the pre-Payrolls period in order for long gamma positions to breakeven, assuming the event day itself proves as volatile as implied by the 1-day options market.

Can markets deliver this level of volatility in the next two weeks? We think not, at least in the case of intermediate and longer tails. Figure 11 shows the rolling 6-week realized volatility on 10Y swap yields, but calculated after excluding Payrolls day observations. In other words, it presents a reasonable sense of what non-Payrolls-day realized volatility might look like in the next two weeks as we head into the next employment report. As the chart indicates, it appears rather unlikely that realized volatility in this quiet period can reach the required level for long gamma positions to break even. Therefore, we recommend a tactical short gamma bias in intermediate and longer tails going into early October, premised upon the view that we are likely to experience a lull in the markets until then. We are more cautious on shorter tails however and prefer to avoid shorts there. Therefore, we turn neutral on short expiry volatility in the front end, but recommend bearish gamma exposure on longer tails as a tactical position for the next few weeks. 

Figure 10: Markets have to deliver ~6 bp/day on non-Payroll days to breakeven given the current level of implieds

Current (9/20) implied volatility (bp/day), 1-day FOA implied vol* on payroll days (bp/day), and the level of realized volatility (bp/day) needed in the lead up to Payrolls day for long-gamma positions to break even

* A FOA or Forward Option Agreement is an inter-dealer 1-day options market from which one can observe event day implied volatility
**Realized volatility needed on non-Payrolls days in the next two weeks is calculated as the square root of current level of implied volatility, squared, times 10, minus the square of the 1-day implied vol on Payroll days, all divided by 9.

Source: J.P. Morgan.

Figure 11: The lead-up to the next employment report will likely be a lull in markets, and a look-back at non-payrolls-day realized volatility in recent months suggests that there is a high bar to break even

Rolling 6-week realized volatility on 10Y swap yields on non-payrolls days*, bp/day, past six months

*Calculated as the rolling 6-week standard deviation of daily changes in 10Y swap yields, with the payrolls-day observations being removed.

Source: J.P. Morgan.

Although we are bearish on short-expiry volatility in most sectors, at a more nuanced level, we see value in being long second principal factor volatility. Much as yield moves across the curve are commonly decomposed into moves that stem from moves in a set of principal factors, we may analogously decompose the volatility of yields into components that arise from the volatility of principal factors. In a recent note, we described a framework that allows an investor to create long or short exposure to just the volatility of the first or second factor, for instance (see Trading Principal Factor Volatility). After this week's FOMC meeting, we believe this is an opportune time to initiate long exposure to the volatility of the second principal factor.

Why is this attractive now? To answer this, we begin by noting that another takeaway from the FOMC meeting was the upward bias in the long run neutral rates in the SEP. Indeed, the bear steepening this week was in part a response to the prospect of a higher terminal rate. Intuitively, this suggests that rate volatility will stem not just from near term policy expectations (the first principal component) but also from shifting expectations surrounding the terminal rate (or a second principal factor). This is the thematic basis for our preference for long exposure to volatility of the second principal factor.  

We will not repeat our methodology for constructing such exposure - details may be found in Trading Principal Factor Volatility. But the intuition behind the construction is as follows. If one assumes that the bulk of the variation in rates is explained by two principal components, which are uncorrelated by design, then it follows that the volatility of any given rate can be decomposed into some weighted combination of the volatility of the first and second principal factors. In general, these combination weights will be different for rates in different maturity points, and therefore we ought to be able to combine a long position in one tail with a short in another tail, to create long PC2 vol exposure while canceling out the PC1 vol exposure. Currently, one attractive way to do this is to initiate longs in 1Yx3Y tails versus 105% of the vega-risk in 1Yx10Y tails. As Figure 12 shows, this weighted volatility spread has declined substantially in the past week and is at year-long lows, making entry levels attractive. Additionally, the package is approximately carry-neutral, in another reflection of its attractiveness. Therefore, we recommend this trade (see trade recommendations).

Figure 12: The weighted 1Yx3Y minus 1Yx10Y volatility spread has declined sharply in the past week, affording attractive entry levels to initiate long exposure to volatility of the second principal factor

1Yx3Y minus -1.05 times 1Yx10Y implied volatility, past 1 year; bp/day

Source: J.P. Morgan.

Trading Recommendations

• Initiate Greens / 3Mx15Y flatteners (0.9:1.0 weighted)
  Given the bearish sentiment emanating from this week's 50bp rate cut, we look for asymmetric ways to position for a rise in yield levels in the Reds. The 2Yx1Y / 3Mx15Y swap curve (0.9:1.0 weighted) is empirically likely to be insensitive to small moves in Reds locally, but is likely to flatten at an accelerating rate in a selloff. This makes it attractive as an asymmetric play on higher yield levels in the Reds.
  -Pay-fixed in $1116.3mn notional of a 09/20/26x1Y SOFR swap at a yield of 2.989% (PVBP: $94.2/bp per mn notional). Receive-fixed in $100mn notional of a 12/20/24x15Y SOFR swap at a yield of 3.338% (PVBP: $1168.2/bp per mn notional). This trade uses risk weights of -0.9/1.0 on the 2Yx1Y/3Mx15Y swaps respectively. This trade is being initiated at a weighted yield spread of -64.8bp.
• Initiate conditional exposure to a flatter 1s/5s swap yield curve (100:102 weighted) in a selloff using 3M expiry payer swaptions
  Another way to create asymmetrical exposure to higher yields is initiate exposure to a flatter 1s/5s curve, using 3M expiry payer swaptions. Current implied volatility levels require selling ~2% more risk in the 3Mx5Y payer swaption leg, but that additional risk is small relative to the recent beta of this curve versus front end yields. Thus, this 1:1.02 weighted 1s/5s curve is still highly directional with yields and likely to flatten in a selloff.
  -Buy $200mn notional 3Mx1Y payer swaptions. (Notification date: 2024-12-20, swap tenor: 1Y, ATMF: 3.35%, strike: 3.35%, spot premium: 21.1bp per notional, forward premium: 21.4bp per notional, bpvol at inception: 6.95bp/day). Sell $43.3mn notional 3Mx5Y payer swaptions. (Notification date: 2024-12-20, swap tenor: 5Y, ATMF: 3.105%, strike: 3.105%, spot premium: 98.3bp per notional, forward premium: 99.5bp per notional, bpvol at inception: 6.85bp/day).
• Initiate 2s/3s maturity matched swap spread curve flatteners
  OTR 2-year notes are now trading near SOFR+19bp, which translates into a pick-up of ~13bp over IOR, while the corresponding pickup is ~3bp higher in the 3Y sector. However, the spread volatility in 2s and 3s is rather similar and the spread duration in the 3Y point is nearly 50% greater, which makes the risk-reward ratio more favorable for 2Y wideners. Also, the carry-to-risk ratio improves rapidly in the case of 2Y spread wideners as the position ages and spread duration falls. Lastly, as RRP balances continue to fall in coming months, SOFR could be biased higher, and banks will likely find it preferable to deploy their spread risk budget in 2Y spread wideners rather than in 3s.
  -Pay fixed in 3.75% Aug 31 2026 maturity matched SOFR swap spreads. Buy $100mn notional of the 3.75% Aug 31 2026 (yield: 3.573%, PVBP: $185.9/bp per mn notional), and pay fixed in $96.8mn notional of a maturity matched SOFR swap (coupon: 3.385%, PVBP: $192.1/bp per mn notional) at a swap spread of -18.8bp.
  -Receive fixed in 3.375% Sep 15 2027 maturity matched SOFR swap spreads. Sell $66.4mn notional of the 3.375% Sep 15 2027 (yield: 3.464%, PVBP: $280.2/bp per mn notional), and receive fixed in $63.9mn notional of a maturity matched SOFR swap (coupon: 3.248%, PVBP: $290.9/bp per mn notional) at a swap spread of -21.6bp.
• Buy 1Yx3Y versus selling 105% of the vega risk in 1Yx10Y swaption straddles
  We recommend initiating long exposure to second principal factor volatility as rate volatility will likely not stem from just near term policy expectations but also from shifting expectations surrounding the terminal rate. 1Yx3Y versus weighted 1Yx10Y is also attractive because the weighted volatility spread is at year-long lows, and the package is approximately carry-neutral
  -Buy $100mn notional 1Yx3Y ATMF swaption straddles. (Notification date: 2025-09-22, swap tenor: 3Y, ATMF: 2.991%, strike: 2.991%, spot premium: 237.8bp per notional, forward premium: 247.1bp per notional, bpvol at inception: 6.83bp/day). This trade assumes active delta hedging every business day. Sell $35.2mn notional 1Yx10Y ATMF swaption straddles. (Notification date: 2025-09-22, swap tenor: 10Y, ATMF: 3.235%, strike: 3.235%, spot premium: 625.7bp per notional, forward premium: 650.4bp per notional, bpvol at inception: 6.02bp/day). This trade assumes active delta hedging every business day.
• Unwind longs in the USZ4 factor-weighted CTD basis (with repo termed out to 12/2)
   This trade has outperformed our expectations and we recommend unwinding at a profit of 2.4 ticks. For original trade write up, see Fixed Income Markets Weekly 2024-09-13.
• Unwind longs in the WNZ4 factor-weighted CTD basis
   This trade has outperformed our expectations and we recommend unwinding at a profit of 1 tick. For original trade write up, see Fixed Income Markets Weekly 2024-09-13.
• Unwind 100:80 weighted 20s/30s maturity matched swap spread curve flatteners
  This trade has outperformed our expectations and we recommend unwinding at a profit of 2.3bp. For original trade write up, see Fixed Income Markets Weekly 2024-09-06.
• Unwind 7s/20s weighted swap spread curve steepeners
   This trade has outperformed our expectations and we recommend unwinding at a profit of 3.8bp. For original trade write up, see Fixed Income Markets Weekly 2024-08-23.
• Maintain conditional exposure to a flatter 1s/3s swap yield curve in a selloff using 3M expiry payer swaptions P/L on this trade is currently -3.5bp. For original trade write up, see Fixed Income Markets Weekly 2024-09-13.
• Maintain conditional exposure to a steeper 2s/10s curve in a rally using 3M expiry receiver swaptions, financed by selling 24% risk-weighted receiver swaptions on 7-year tails
  P/L on this trade is currently 0bp. For original trade write up, see Fixed Income Markets Weekly 2024-09-13.
• Stay long 1Yx10Y swaption straddles on a delta hedged basis coupled with a weighted long in S&P 500 futures
  P/L on this trade is currently -0.7abp. For original trade write up, see Fixed Income Markets Weekly 2024-09-13.
• Maintain 3M fwd 2s/3s (1:1.1 weighted) swap yield curve flatteners
   P/L on this trade is currently -2.7bp. For original trade write up, see Fixed Income Markets Weekly 2024-09-06.
• Stay long 3Yx3Y versus short 10Yx10Y swaption straddles
   P/L on this trade is currently -4.7abp. For original trade write up, see Fixed Income Markets Weekly 2024-09-06.
• Maintain conditional exposure to a flatter 1s/7s swap yield curve in a selloff using 6M expiry payer swaptions
  P/L on this trade is currently -5.9bp. For original trade write up, see Fixed Income Markets Weekly 2024-07-12.
• Maintain 1:0.75 risk weighted 7s/10s maturity matched swap spread curve steepeners
  P/L on this trade is currently -0.3bp. For original trade write up, see Fixed Income Markets Weekly 2024-05-31.
• Continue to overweight 1Yx10Y straddles versus a gamma-neutral amount of 1Yx15Y straddles
   P/L on this trade is currently -2.3abp. For original trade write up, see Fixed Income Markets Weekly 2024-05-03.
   
• Continue to overweight 6Mx5Y and 6Mx30Y swaption volatility (vega weights of 0.32 and 0.76, respectively) versus selling 6Mx10Y swaption volatility
  P/L on this trade is currently -0.8abp. For original trade write up, see Fixed Income Markets Weekly 2024-04-05.

Closed trades over the past 12 months

P/L reported in bp of yield for swap spread, yield curve and misc. trades, and in annualized bp of volatility for option trades, unless otherwise specified

Note: trades reflect Thursday COB levels, and unwinds reflect Friday COB levels

• After more than four years, the FOMC delivered its first rate cut, lowering the Fed funds target range by 50bp to 4.75-5.00%
• Our economists still anticipate a faster pace of rate normalization than the median dot suggests, expecting another 50bp cut at the November meeting and a 25bp reduction in December
• Both SOFR and EFFR adjusted lower, aligning with the new Fed funds target range as expected. Yields at the very front-end of the money markets curve also shifted lower, leading to a steepening of the 1m12m and 3m2y curves. Even so, these curves remain deeply inverted, challenging liquidity investor’s willingness to add duration
• Indeed, MMF WAMs moved lower YTD, particularly with respect to prime MMFs. Prime MMF reform could be a factor, though we suspect the lack of clarity on how the easing cycle will unfold and the inverted yield curves had more to do with it
• FRNs in prime MMFs rose from 15% at the start of the year to 20% as of 8/30, while that in government MMFs only increased by 2%-pts to 16% this year
• To the degree prime funds continue to position towards floaters, the yield spread between government and prime funds might narrow as we get further into the easing cycle
• Government portfolio holdings remained stable overall during August, with a $43bn decrease in Treasury, Agency, and other repo, and a $14bn decrease in Treasury coupons. However, allocations increased in RRP by $21bn, Treasury FRNs by $20bn, and T-bills by $90bn
• Prime AUMs declined slightly MoM, with reductions in holdings of Treasuries, ON RRP, and repo (excluding RRP) by $12bn, $11bn, and $7bn, respectively. They also slightly reduced their exposure to credit by $7bn during August
• Near-term catalysts: Aug Personal income (9/27), Aug JOLTS (10/1), Sep ADP (10/2), Sep employment (10/4), Sep CPI (10/10), Sep PPI (10/11)

Market commentary

After more than four years, the FOMC delivered its first rate cut, lowering the Fed funds target range by 50bp to 4.75-5.00%. This decision was made against a backdrop of progress towards the Fed’s inflation goals and aimed at preserving the current strong labor market from downside risks. While the rate reduction was widely expected, the size of the cut was much debated. Since August, our economists had anticipated a 50bp reduction at this FOMC meeting, reasoning that, in hindsight, they should have cut rates in late July and this larger move helps get them back onsides.

In the SEP, the median dot indicated two additional 25bp cuts by year-end, a total of 100bp of easing in 2025, and an additional 50bp of easing in 2026. Meanwhile, the median unemployment rate projection for this year and next is 4.4%, up 0.4%-pt and 0.2%-pt, respectively, from the June projections. During the press conference, Chair Powell offered no clues on the magnitude of future cuts, only saying “we're recalibrating policy down over time to a more neutral level.”

Our economists still anticipate a faster pace of rate normalization than the median dot suggests, expecting another 50bp cut at the November meeting and a 25bp reduction in December. However, this expectation largely depends on further softening in the next two jobs reports. If the labor data turns out to be more benign, it would support the FOMC’s “Goldilocks” scenario of 25bp cuts per meeting for the rest of the year (see An appropriate recalibration, M. Feroli, 9/19/21). Markets also seem to be in two minds whether the Fed will follow with another 50bp at the November FOMC meeting, particularly a meeting immediately after US elections. By week’s end, OIS forwards are pricing 38bp of cuts in November, another 37bp in December, and 200bp of total easing by the end of 2025 ( Figure 38).

Figure 38: By week’s end, OIS forwards are pricing 38bp of cuts in November, another 37bp in December, and 200bp of total easing by the end of 2025

OIS-implied change in fed funds effective rate by FOMC meeting, as of 9/20/24, and 9/13/24 (bp)

Source: J.P. Morgan

Turning to the Fed’s balance sheet, there was minimal discussion on QT. However, when asked about balance sheet policy during the press conference, Powell noted, “In the current situation, risks have really been stable. They haven’t come down. So reserves are still abundant and expected to remain so for some time. The shrinkage has come out of the overnight RRP.” The Chair also mentioned, “We know that these two things can happen side by side; in a sense, they’re both forms of normalization. So for a time, you can have the balance sheet shrink while also cutting rates.” Importantly, we continue to see QT as being in its endgame, with perhaps a few more months left, and project the Fed’s balance sheet to end the year at ~$7tn (see Hopscotch, S. Ramaswamy, 8/16/24).

Front-end update

Following the Fed’s consequential rate cut, both SOFR and EFFR adjusted lower, aligning with the new Fed funds target range as expected. On 9/19, EFFR declined by 50bp to 4.83%, and SOFR decreased by 51bp to 4.82%. Similarly, yields at the very front-end of the money markets curve shifted lower, leading to a steepening of the 1m1y and 3m2y curves. Even so, at current levels, these curves remain deeply inverted at -74bp and -107bp, respectively ( Figure 39). The inverted curves continue to challenge liquidity investors willingness to extend and add duration.

To be sure, despite expectations of rate cuts, MMF WAMs have moved lower this year, particularly with respect to prime MMFs. YTD, prime WAMs are down 9.13 days to 22 days. Arguably, MMF reform might have been one of the factors behind the decline as funds looked to add liquidity to meet potential redemptions, though we suspect the lack of clarity on how the easing cycle will unfold and the inverted yield curves had more to do with it. Indeed, a closer look at prime MMF holdings shows a meaningful increase of FRNs in their portfolios, rising from 15% at the start of the year to 20% as of 8/30 ( Figure 40). As we have noted previously, FRNs offer better protection in the current interest rate environment while also ensuring higher yields. Importantly, exposure to FRNs avoids an immediate drag on MMF yields, thus allowing prime funds to maintain some sort of yield advantage over government funds for the time being (see Credit liquidity investors still prefer floaters over fixed, T. Ho, 7/31/24).

Meanwhile, government MMFs have taken on a slightly less aggressive positioning with respect to their duration: government WAMs are down 6.2 days to 29 days YTD, and their FRN holdings have only increased by 2%-pts to 16% this year. In other words, they’ve been able to lock in slightly higher yields. To the degree that prime funds continue to position towards floaters, the yield spread between government and prime funds might narrow as we get further into the easing cycle.

Still, regardless of how prime funds perform relative to government funds, given the still inverted front-end yield curve and the yield advantage MMFs have over other cash alternatives, we expect MMF AUMs will continue to rise into year-end. Indeed, MMF balances have increased $77bn so far this month to a record $6.5tn, despite the $48bn outflows observed around the corporate tax date (see MMF AUMs Rise Despite Upcoming Cuts, T. Ho, 9/6/24).

Figure 39: Yields at the very front-end of the money markets curve steepened this week, but still remain deeply inverted

1m1y and 3m2y yield curve (bp)

Source: J.P. Morgan

Figure 40: Prime MMF holdings showed a meaningful increase of FRNs in their portfolios, rising from 15% at the start of the year to 20% as of 8/30

FRNs held by MMFs as a percentage of total AUMs (%)

Source: Crane Data, J.P. Morgan

August MMF Holdings

In August, taxable MMFs experienced a steady increase in balances, marking the highest net flows thus far this year. Total AUMs rose by $108bn, reaching nearly $6.5bn during the month ( Figure 41). This rise in balances is not particularly surprising, as AUMs tend to increase ahead of an easing cycle, with yield-driven investors rotating away from other cash alternatives into MMFs (see When will the front-end Treasury curve steepen?, T. Ho, 9/14/24).

Figure 41: In August, total MMF AUMs rose by $108bn, reaching nearly $6.5bn during the month

Cumulative YTD change in taxable MMF AUMs ($bn)

Source: Crane Data, iMoneyNet, J.P. Morgan

Accordingly, government MMFs experienced a relatively strong month in net flows, with balances rising by $87bn. Even so, their portfolio holdings remained relatively stable month over month. In August, there was a $43bn decrease in Treasury, Agency, and other repo, and a $14bn decrease in Treasury coupons. Conversely, they increased their allocations towards RRP by $21bn, Treasury FRNs by $20bn, and T-bills by $90bn ( Figure 42). Interestingly, government funds also increased their holdings of bills maturing within the 31–60 day range by $226bn MoM, while reducing their exposure to bills maturing in less than 30 days by $83bn and to those maturing in more than 60 days by $53bn ( Figure 43). Furthermore, government and treasury WAMs were slightly reduced by 1.4 days during August, averaging 30 days and 38 days respectively ( Figure 44).

Figure 42: In August, Treasury, Agency, and other repo decreased by $43bn, Treasury coupons by $14bn, while RRP increased by $21bn, Treasury FRNs by $20bn, and Treasury bills by $90bn

Asset allocation breakdown of government MMFs ($bn)

Source: Crane Data, J.P. Morgan

Figure 43: Government funds increased T-bill holdings the most in 31-60 day maturities by $226bn

Maturity profile of T-bills held by government MMFs, as of 7/31/24 and 8/31/24 ($bn)

Source: Crane Data, J.P. Morgan

Similarly, prime MMFs saw little change in their portfolio holdings. AUMs declined slightly MoM, with reductions in holdings of Treasuries, ON RRP, and repo (excluding RRP) by $12bn, $11bn, and $7bn, respectively ( Figure 45). Additionally, they slightly reduced their exposure to credit by $7bn during the month, with the largest decline in CDs at $14bn, partially offset by an increased allocation towards CP by $8bn ( Figure 46). Excluding the prime funds that have chosen to convert to another fund or liquidate completely, prime AUM outflows have been mostly limited throughout this year.

Figure 44: Government and Treasury WAMs were slightly reduced by 1.4 days during August, averaging 30 days and 38 days, respectively

5-day average of government and treasury WAMs (days)

Source: Crane Data, J.P. Morgan

Figure 45: Prime AUMs declined slightly MoM, with reductions in holdings of Treasuries, ON RRP, and repo (excluding RRP) by $12bn, $11bn, and $7bn, respectively

Asset allocation breakdown of prime MMFs ($bn)

Source: Crane Data, J.P. Morgan

While ON RRP balances remained mostly range-bound in August, between the low- to mid-$300bn range, the facility trended downward following the end of the month. Specifically, ON RRP balances fell to $239bn on September 16, marking its lowest level since May 2021. This decline was likely due to corporate tax outflows and mid-month Treasury settlements. This also likely contributed to the rise in SOFR, which increased by 5bp DoD to 5.38% on September 16. As we enter the GSE period and Treasury settlements work their way through the system, along with negative T-bill supply, ON RRP balances should trend back higher throughout this month. SOFR should also move back lower.

Figure 46: Prime funds slightly reduced their exposure to credit by $7bn during the month, with the largest decline in CDs at $14bn, partially offset by an increased allocation towards CP by $8bn

J.P. Morgan estimates of prime MMF exposure to banks ($bn)

Source: Crane Data, J.P. Morgan

Excerpted from Short-Term Market Outlook and Strategy, Teresa Ho, September 20, 2024

Hindsight is 20/21

• Mortgage spreads had an excellent week, with the MBS market responding quite favorably to the 50bp cut and the general risk-on trend across spread products
• That’s put mortgages at new year-to-date tights. With FN 4.5 T-OAS at just 5bp in our beta model (and only modestly wider in YB), we have to admit we’ve had some flashbacks to early 2021
• To be fair, today investors are looking at assets that are undeniably spreadier and represent a way to sell volatility (which, while off the peaks, is still historically elevated); nonetheless, OAS looks historically snug
• Bank demand has started to edge higher and foreign demand should perk up with a steeper curve and lower funding rates, reintroducing two buyers more focused on nominal spread than money managers tend to be; the market seems to be leaning into the potential for a technical regime shift
• A few reports over the past week highlighted the risk of another run at GSE privatization under a second Trump administration; while the hurdles are still significant, we do think investors should take the discussion seriously
• Last week CFPB Director Chopra highlighted efforts to help more borrowers take advantage of refinancing opportunities; while cost reduction could help reduce the incentive elbow, rewriting QM to allow for streamline refi of GSE loans could have a larger impact
• We adjust our net issuance forecast up to $215bn for the full year from our initial November 2023 call of $175bn
• The market needs to absorb roughly 40% ($93bn) of this year’s net issuance over the next four months; a sharp refi response could potentially re-arrange the exact timing of when the growth shows up
• Ginnie delinquency and buyout data still shows minimal impact from the FHA and VASP programs; on the VASP front, systems issues may be throttling servicer throughput, but with implementation required by October, we’ll be watching carefully for more signs of buyouts
• TIC data implies that foreign MBS holdings are up $51.4bn through July, with changes in holdings being increasingly driven by private investors
• TIC data also implies that Japanese investors added in July, though early September net flows indicate Japanese investors may have shed foreign bond holdings

Views

• The belly of the stack (now the current coupon) looks rich on a historic basis; prefer a barbell of the wings on the fundamentals
• Low FICO pools provide good near-term call protection for investors worried about conventional speeds
• High MDS pools offer call protection via FTHB-only LLPA waivers

Mortgage spreads had an excellent week, with the MBS market responding quite favorably to the 50bp cut and the general risk-on trend across spread products. That’s put mortgages at new year-to-date tights. With FN 4.5 T-OAS at just 5bp in our beta model (and only modestly wider in YB), we have to admit we’ve had some flashbacks to early 2021. Now, to be fair, back then we were writing about why investors shouldn’t buy 15ZV assets at -30OAS and instead should just look at a replicating basket of Treasuries. Today, investors are looking at 80-115ZV, 5-15OAS assets that are undeniably spreadier and represent a way to sell volatility (which, while off the peaks, is still historically elevated). Nonetheless, the option-adjusted spreads look historically snug.

Figure 47: Mortgage spreads had an excellent week on the back of the 50bp cut and a general risk-on trend across spread products

Current, 1m, and 6m Treasury OAS ranges across the TBA stack in our research beta model. The black dots represent the current OAS, the blue boxes represent the 1m range, and the black lines represent the 6m range (as of 9/19/2024)

Source: J.P. Morgan

The question, then, is whether the technical dynamics are set up such that these bonds can go even tighter. There are signs in the H.8 data (which must be heavily massaged to get a real sense AFS/HTM buying—see Figure 48) that bank demand is starting to creep up after years spent in the doldrums. The drivers of bank buying are clearly different today than they were back in QE4. Banks are no longer parking unhedged fixed rates in HTM as a way to pair off against a surge in sticky deposits. Instead, bank demand is far more spread focused, as evidenced by CMO floater issuance and, more anecdotally, the use of portfolio layer hedging. Still, we do get the sense that, just as back in QE4, many banks are still more focused on the nominal spreads that mortgages offer.

Figure 48: Bank buying appears to have picked up modestly ahead of the FOMC

2024 YTD changes in agency MBS holdings across all U.S. banks in amortized cost terms, where the “YTD (adj. for bank trading accounts)” series adds subtracts out the $12.5bn settled in 2023 and trading account changes, spread across Class A settle increases. In Q1, we subtract $26bn of settled Class A holdings to tie out with call report increases in bank trading accounts and match H.8. with AFS + HTM call report increases. In Q2, we subtract $20bn of settled Class A holdings to match H.8. changes with AFS + HTM call report changes.

Source: J.P. Morgan, Federal Reserve, FFIEC

Now, this is not to say that banks are going to load the boat on MBS at just any positive spread. But when we look at a long time series of current coupon ZV, it’s notable that, even having come 70bp off of the 2023 wides, we’re still 40bp above the 2010-2021 average ( Figure 49). If banks are starting to feel a little more free to add securities, taking some comfort from the initiation of the cutting cycle (which likely helps lessen the chances of significant deposit outflows, even if it doesn’t necessarily boost inflows), we can still see some finding mortgages attractive even at very tight OASs. We very much doubt that we are going back to the patterns of buying that we saw even before QE4, where deposit driven duration matching in HTM and the LCR buildup drove regular annual MBS demand of $100-200bn. But a more modest $50-100bn net add could be possible over the next year.

Figure 49: Current coupon ZV remains 40bp above the 2020-2021 average

Current coupon historical treasury ZV spread and latest level (bp)

Source: J.P. Morgan

That, coupled with the potential for more overseas buying as the US curve resteepens, means significantly more demand that’s focused on ZVs or nominal spread versus the environment of the past two years. Some money managers have been waiting to rotate out of MBS into corporates, anticipating that IGs would eventually widen. But with the JULI still near its recent tights, it’s not entirely obvious what to buy if you sell MBS. Rotating into Treasuries will give up a lot of yield, but do you want to rotate into corporates and sell credit risk instead of vol? There may be some inertia to mortgage overweights, given the spread product landscape and need to maintain a higher current yield to attract inflows. Relatedly, the recent rally should continue to support inflows into fixed income funds, which could mean that money managers can still net add MBS even without changing their weightings much (or letting them drift slightly down).

So, while we don’t love mortgage spreads here, we understand why it’s hard to really short them. We’d still recommend a barbell on the stack to stay away from the belly, at least for investors with flexibility on dollar price. This week, we look at two risks for conventionals, each coming from different ends of the political spectrum: GSE reform and streamline refis. We also revise our net issuance forecasts modestly higher to account for the rise in home prices and lower rates.

GSE privatization talk heats up again

Last Friday, the WSJ published an article titled “Trump Allies Are Working on Plans to Privatize Fannie and Freddie.” Though it was somewhat light on the precise details of how much capital needed to be raised, the report reinforced our view that the probability of a GSE privatization effort is reasonably high under a second Trump administration. Back in March, we’d highlighted that former FHFA Director Calabria’s book described how close he felt the GSEs had come to privatization at the end of the Trump administration (see our March 15 piece for more details on this, as well as the current state of the GSE capital stack). Calabria was himself quoted earlier this week by a report from HousingWire, saying that while the chance of the GSEs going private in 2025 is zero, “by [2027] I would say there's maybe 70% chance. ... Almost every decision you think you have to make, we scoped out. All those millions of dollars with my go ahead, low-key actually produced documents. So, there are plans; there are options. You can get them out. It's all feasible, doable.”

There are still clearly massive hurdles and unanswered questions regarding a privatization effort. However, investors should take seriously that there’s a huge amount that can be done administratively—and that Calabria had drawn up the plans, which presumably are still lying around. A new administration would need to appoint both a Treasury Secretary and FHFA Director interested in furthering this effort.

In our view, administrative action will never get around the central conundrum of GSE reform—that the implicit guarantee, while officially disavowed, has been made good on before. It will always seem implausible to the market that the government could allow these institutions to fail, which necessarily means there is a tension between their public mission and potential status as private entities. But that doesn’t need to be ‘solved’ for conservatorship to end; and there’s a lot of unknowns about what could happen along the way.

How serious is the CFPB about streamlining refis?

Last week, CFPB Director Chopra gave a speech focused on expanding refinancing opportunities for borrowers. He confirmed that closing costs are a key area of interest for the agency—it’s possible that there’s some money to be saved on things like FICO scores or title insurance. Savings of this sort could shrink the incentive elbow caused by fixed costs, but we thought that Director Chopra’s mention of how “lenders are required to redo some of the same steps that were completed by the borrowers when they first purchased their home” might be the most interesting part of the speech. If his comments are hinting at amending the QM definition to allow streamline refi (and we think they are), that could open the door for steeper GSE s-curves down the road.

For context, back in 2022, the CFPB put out an RFI regarding how streamline refis could benefit borrowers. At the time, they noted that “mortgage refinancing can be harder to access for borrowers with smaller loan balances. Black and Hispanic borrowers, who on average have smaller loans, have not participated in recent refinance booms at the same rate as white borrowers….New streamlined and automatic refinancing mortgage products could make sure that those buying a home now, or refinancing to cover other needs, are able to benefit from the next interest rate drop.” Chopra’s speech last week highlighted the same motivation, but doesn’t explicitly call for ‘automatic’ refinancing and only obliquely brings up streamlines. Still, we think that it’s the most meaningful focus for the agency.

For a loan to comply with the QM/ATR (Qualified Mortgage/Ability-to-Repay) rule, it needs to meet the CFPB’s standards, which include documenting the borrower’s ability-to-repay the loan. This generally requires re-underwriting the borrower at the time of origination of the new loan. There are a few exceptions—notably FHA and VA have streamline programs that are deemed QM simply because they are government promulgated rules—the ‘verification safe harbor’ loophole. The GSEs used to have a similar exclusion (the “QM patch”). However, former FHFA Director Calabria wanted the GSEs to operate under the same QM rules as the private industry, and so GSE underwriting currently requires full redocumentation on refis.

As an aside, this is the reason that the GSEs have had to shut off their High LTV Refi Offerings, which were designed to allow for >100LTV streamlined refinancing similar to the legacy HARP programs. The HLRO was built into the CRT language, and would have allowed the GSEs to maintain CRT coverage even when an underwater loan refi’d. Currently, however, the program is on ice, which also puts some limits on the GSE’s ability to allow for underwater refinancing (a clear borrower benefit) and maintain their CRT coverage (which would be lost under another, non-HLRO loan).

If the CFPB were to rewrite QM to allow for streamline refis, we suspect that there would be pressure on the GSEs to more broadly implement a streamline program (beyond just the HLRO). The rulemaking process is lengthy, so the presidential election may determine whether this is pursued. Still, given that refis are picking up and this topic seems to have new legs and focus, it’s a risk for conventional mortgages in the medium term.

We can see the argument for allowing borrowers to refi without reunderwriting, so long as their payments are strictly lower upon refinancing and their LTV ratio doesn’t dramatically increase. A lower payment will reduce default risk, which would be particularly beneficial if the rally coincides with deteriorating economic conditions. The GSEs would also clearly benefit from this, since they’re already on the hook for the credit risk. This was the rationale behind HARP; while the GSEs would never underwrite new risk at >100LTV, refinancing borrowers with extremely high MTM LTVs served to reduce default probability on loans with challenged recovery prospects.

But as always, there’s no free lunch. Introducing streamline refis would raise option costs and lower mortgage prices (and thus raise mortgage rates for new borrowers). Moreover, CRT investors may be less comfortable buying new risk with potentially stale underwriting and appraisal information.

It’s challenging to determine how much wider spreads would move if the GSEs introduced a streamline option. We used to point to speeds around the June 2009 HARP cutoff date as being a natural experiment in how much streamline refi could change speeds ( Figure 50). The ‘20/’21 refi wave was at least partly fueled by the adoption of new electronic underwriting technology and the widespread use of AVMs instead of full appraisals; the difference between modern refi underwriting and streamline is surely narrower than it would have been a decade ago. Still, given that spreads look relatively tight on conventional mortgages to us, we don’t think that the market is priced to a streamline scenario. It’s another policy risk that investors are exposed to that hinges on the outcome of the presidential election.

Figure 50: The HARP effect was clear, but may not be perfectly applicable today

3m CPRs observed in October 2016 and August 2013, for ‘09 vintage Freddie Mac 30-year loans with WAC between 5% and 6%, by loan origination month

Source: J.P. Morgan, Freddie Mac

Netting out the rest of the year

We last revisited organic net supply during our midyear outlook three months ago; at that time, mortgage rates were sitting at 7% and we had seen little evidence to move us away from our initial projection from the start of the year (+$175bn). After seeing sales activity remain relatively static from last year, and continued strength in HPA, we think it’s time to move our forecast higher to $215bn. This means another $93bn for September to December (~$23bn a month). Figure 51 shows our month-by-month expectations and how they stack up against last year. That’s in normal net terms (a month’s gross issuance minus its paydowns); it could be rearranged somewhat by rate/term refis but the overall real growth in market size isn’t dictated by recycling loans at the same size.

Putting this revised net number into context against our forecast from almost a year ago, the main change in the components making up the forecast is an increase in the purchase contribution from -$25bn to +$20bn ( Figure 52). One of the sensitivities in that forecast was HPA and its continued strength was worth $25bn based on our calculations at the time (+5.5% revision). The rest came from us revisiting how “much” each purchase loan adds to net due to re-sizing and home prices. As a reminder, we consider a closed-system of agency MBS for coming up with the net from purchase/cashout/amortization, while the securitization rate reflects that there can be movement into/out of agency MBS from other mortgage loan sources.

Figure 51: Net issuance will likely stay firmer throughout the rest of the year relative to 2023…

Monthly Agency MBS net issuance in 2023 and 2024, and our 2024 projections as of 6/28/2024, $bn

Source: J.P. Morgan, Fannie Mae, Freddie Mac, Ginnie Mae

Figure 52: …and we adjust up our full year forecast to $215bn (from $175bn) to reflect the growth in HPA and continued purchase activity

Expected change in issuance from this year to next by component, $bn

Issuance Forecast Component	2023 Total	2024 Total	2023 to 2024 Change
Purchase	504	523	20
Cashout	32	33	0
Amortization	-325	-330	-5
Securitization Rate	N/A	N/A	-35
Net total	235	215	-20

Source: J.P. Morgan

Speaking of purchase loans, the pattern of purchase buying has been abnormal since the start of 2020. Figure 53 gives a simple look at the timing of purchase loans issuance (not origination) throughout the year, with each cell reflecting the month’s activity relative to the yearly average. Pre-2020, the peak typically occurred in August, and slowly decreased through the rest of the year before hitting the lows in January-March. The past few years, particularly 2022, bucked this trend as tight inventory and rate sell offs led to purchase issuance falling off more dramatically in the fourth quarter.

This year, the situation is reversed, with rates having fallen almost a full percentage point over the past couple months. Very roughly, our model thinks that such a change is worth a 10% increase in turnover, though the exact mechanism and timing of that increase can depend on other factors. At the least, it means we look for the 2024 pattern of issuance for the rest of the year to look more like the 2017-2019 average rather than 2022/2023, which helped guide the creation of our month to month issuance forecast. A more leading indicator, purchase applications, supports this notion as levels have now drawn nearly even year over year after lagging for most of it ( Figure 54).

Figure 54: Purchase applications have caught up to last year, and might not drop as much into year-end if the rate move has any impact

MBA Purchase Index, calendar year overlay with daycount adjustments

Source: J.P. Morgan, MBA

Figure 55: Fed runoff could end the year at $193bn, close to our original expectation of $180bn

Fed runoff by month for 2023 and 2024, along with a projection for the rest of the year

Source: J.P. Morgan, Federal Reserve

On a related note, we also took a glance at our projections for Fed runoff over the last four months of the year ( Figure 55). Funnily enough, the year over year change in absolute pace has almost exactly mirrored the decrease in size of the Fed portfolio. Said another way, the Fed’s portfolio going into 2024 was 8% smaller than it was going into 2023… and our new projection for paydowns ($193bn) is 7% smaller ($208bn). This equivalence reflects how the seasoned and/or deeply out-of-the-money bonds in the Fed’s portfolio show little reactivity to rate moves. Looking ahead, it will be interesting to see if 6% (or lower) mortgage rates shakes loose more of these bonds due to loans sliding up the modestly sloped turnover s-curve.

Hints of VASP

FHA payment supplement

From May 1^st to Jan. 1^st, 2025, servicers can implement FHA’s Payment Supplement option for eligible borrowers, which allows borrowers to tap into unused partial claim funds (up to 30% of the balance of the loan at time of first default, minus any partial claim draws in the intervening period) to temporarily reduce their mortgage payments, targeting a 25% reduction.

So far, we have not seen much evidence of the program being implemented, as we have not seen an uptick in cure rates in lower and belly coupons in aggregate or across large FHA servicers ( Figure 56 & Figure 57). We will continue to monitor increases in cures as a proxy for FHA payment supplement implementation through the end of the year.

Figure 56: We have not seen an uptick in cures following the implementation of the FHA payment supplement …

FHA 90+D to current rolls as a share of all previously 90+D loans over time, 2-4.5 coupon, %

Source: J.P. Morgan, Ginnie Mae

Figure 57: ... and even at the servicer level, its hard to say if any single one has seen an increase in cures that could be tied to the new program

FHA 90+D to current rolls as a share of all previously 90+D loans over time by top 20 FHA 2-4.5 coupon servicers as of Aug. 2024, 2-4.5 coupon, %

	Jan-24	Feb-24	Mar-24	Apr-24	May-24	Jun-24	Jul-24	Aug-24
All	18%	21%	21%	19%	18%	15%	16%	16%
Lakeview	15%	22%	25%	24%	24%	18%	18%	20%
Freedom	24%	28%	19%	15%	13%	13%	14%	15%
Pennymac	26%	25%	25%	23%	22%	16%	17%	18%
Rithm	16%	19%	24%	25%	25%	20%	21%	20%
Mr. Cooper	29%	26%	28%	24%	26%	25%	25%	22%
Carrington	20%	24%	19%	17%	16%	15%	14%	12%
Rocket	18%	17%	17%	18%	13%	14%	16%	15%
Wells	5%	6%	6%	6%	6%	6%	5%	6%
US Bank	18%	20%	25%	24%	21%	15%	14%	19%
Planet	8%	10%	11%	9%	13%	16%	18%	18%
M&T	14%	30%	36%	29%	24%	16%	17%	14%
LoanDepot	9%	12%	14%	15%	13%	12%	9%	14%
Money Source	13%	15%	15%	13%	16%	12%	15%	13%
Guild	19%	19%	21%	21%	24%	21%	18%	17%
IDHH	12%	20%	23%	31%	18%	15%	15%	16%
Cross Country	6%	23%	29%	27%	28%	25%	22%	20%
New American Funding	19%	19%	19%	20%	21%	19%	24%	22%
CMG	17%	21%	22%	21%	17%	13%	17%	18%
Citizens	12%	17%	14%	9%	14%	8%	9%	19%
Truist	15%	12%	20%	17%	15%	12%	15%	12%

Source: J.P. Morgan, Ginnie Mae

VASP

The VA Servicing Program (VASP) was officially activated on May 31^st, and must be implemented no later than Oct. 1^st, 2024. Unlike the FHA payment supplement, which would cure delinquent loans in pool, VASP aims to bring down delinquencies by buying VA loans out of pool upon modification (see Agency MBS: Second Thoughts for more details).

With a month of the program implementation remaining, we do not see any meaningful upticks in aggregate VA buyouts, and there are a handful of servicers (New American Funding, PennyMac, and Truist) for which buyouts of delinquent loans have crept up over the last month or two ( Figure 58 & Figure 59). We’ve heard that some servicers have started testing the VASP pipes, but that operational difficulties may be throttling the throughput of these mods. With all servicers required to implement starting next month, we’d expect the pace of VASPing to pick up, but it remains to be seen whether the capacity exists to quickly process the backlog of delinquencies.

Figure 58: We have not seen any meaningful uptick in aggregate VA buyouts …

VA buyouts as a share of previously 90+D scheduled balances over time, %

Source: J.P. Morgan, Ginnie Mae

Figure 59: … though some servicers have started implementing the program

VA buyouts as a share of previously 90+D scheduled balances over time, by top 24 VA servicers, %

	Jun-24	Jul-24	Aug-24
All	2.7%	2.5%	2.6%
Carrington	3.2%	3.0%	1.5%
Citizens	4.3%	7.6%	3.7%
CMG	2.0%	1.3%	0.9%
Cross Country	4.9%	4.2%	1.5%
Freedom	1.6%	2.0%	1.0%
Guild	0.8%	1.1%	0.4%
Lakeview	2.3%	2.0%	1.6%
LoanDepot	3.9%	4.5%	2.3%
M&T	4.0%	1.2%	1.3%
Money Source	2.5%	2.4%	1.8%
Mortgage Research	3.9%	5.1%	2.8%
Movement	4.4%	4.3%	2.4%
Mr. Cooper	4.9%	2.9%	1.6%
NavyFed	7.3%	4.3%	6.4%
New American Funding	1.2%	1.8%	9.9%
Pennymac	3.7%	3.5%	6.4%
Pentagon	4.7%	3.4%	2.2%
Planet	0.8%	0.6%	0.7%
Rithm	1.9%	2.6%	2.8%
Rocket	3.1%	1.9%	1.0%
Sunwest	1.0%	0.2%	0.5%
Truist	5.4%	7.2%	11.8%
US Bank	3.7%	6.2%	3.2%
Wells	0.8%	1.1%	2.0%

Source: J.P. Morgan, Ginnie Mae

Foreign holdings sTICking at current levels

Foreign holdings of agency MBS are up +$51.4bn this year (through July), and only inched modestly higher month over month ( Figure 60). The evolution of foreign holdings appears to be increasingly driven by foreign private investors, especially the sizeable uptick in June ( Figure 61).

Figure 60: YTD purchases of agency MBS are implied to be +$51.4bn through July, increasing by $2.3bn month over month …

YTD purchases of MBS by foreign investors, manually price-adjusted with the index through January 2023, and adjusted using TIC-provided mark to market changes from February 2023 onwards, $bn

Source: J.P. Morgan, TIC

Figure 61: … with changes in holdings being driven by foreign private investors this year …

YTD purchases of MBS by foreign private investors, manually price-adjusted with the index through January 2023, and adjusted using TIC-provided mark to market changes from February 2023 onwards, $bn

Source: J.P. Morgan, TIC

Meanwhile, foreign official holdings of agency securities have steadily declined so far this year ( Figure 62).

Figure 62: … as foreign official holdings of agency securities have been decreasing

US Federal Agency Securities held in custody in foreign official accounts, $bn

Source: J.P. Morgan, Federal Reserve

At the country-by-country level, TIC indicates that Japanese investors beginning to add mortgages ( Figure 63). However, recent Japanese purchases and sales of foreign bonds indicate a net outflow in early September, so we will monitor whether these additions persist in the TIC data ( Figure 64).

Figure 64: … though the Ministry of Finance data indicates that Japanese flows of foreign medium-to-long-term bonds started to dip in early September

Japan weekly securities investment abroad, medium-and-long term bonds, $bn

Source: J.P. Morgan, Ministry of Finance

Week in Review

• MBA Weekly Survey: For the week ending September 13, the purchase application index rose 3.6% w/w and was 0.4% lower than year ago levels, while the refinance index rose 24.2% w/w and was 38.9% higher than the 3-month trailing level (daycount-adjusted, not seasonally-adjusted) ( Figure 65 & Figure 66).
• Freddie Enhanced Primary Survey: For the week prior to September 19, 2024, 30-year conventional conforming fixed-rate mortgages averaged 6.09%, down 11bp from the previous week ( Figure 67)
• Primary dealer specified pool positions rose to $484.2bn (+$24.7bn w/w) as-of close trading September 11. Including TBA positions of -$429.0bn, dealers were long $55.2bn (-$9.1bn w/w) pass-throughs. Other agency MBS holdings rose $3.5bn to $35.3bn.
• Fixed-rate agency gross and net issuances were $108.5bn and $27.4bn, respectively, in August.September gross supply currently stands at $69.5bn ( Figure 68).
• ICI Total Bond Long-term Mutual Fund and ETF Weekly Flows: Inflows were +$8.7bn for the week of September 11 and +$43.3bn for the month leading up to September 11 ( Figure 69).

Figure 65: MBA Purchase Index, calendar year overlay with daycount adjustments

Source: J.P. Morgan, MBA

Figure 66: MBA Refi Indices, seasonally adjusted

Source: J.P. Morgan, MBA

Figure 67: Primary mortgage rates, %

Source: J.P. Morgan, Optimal Blue, Freddie Mac, Mortgage News Daily, MBA

Figure 68: Gross and net fixed-rate MBS monthly issuance, $bn

Source: J.P. Morgan

Figure 69: ICI Total Bond Long-term Mutual Fund and ETF Weekly Flows, $bn

Source: J.P. Morgan, ICI

DQs march higher in non-QM

• The Fed gave the market what it wanted with a 50bp rate cut
• Investors are starting to focus more on convexity risk in jumbo collateral as 6% coupons shift further in-the-money. This could begin to pressure premium price spreads
• Our Securitized Products Real Estate and Consumer (SPREC) conference on Friday, September 13th, was well received by investors, with nearly 300 participants. See here for a full recap
• Delinquencies have continued to rise in the non-QM segment. FHA delinquencies are also higher, while GSE and jumbo borrowers have not experienced any significant increase
• Non-QM loans originated since 2Q22 are driving delinquencies higher
• Delinquencies have increased across all documentation types, but bank statement and ‘other’ borrowers have more significantly underperformed
• Within the ‘other’ bucket, loans underwritten using P&L statements and written verification of employment (WVOE) have the highest delinquencies, reaching almost 7% in August
• There is no significant difference in later-stage roll rates among different documentation types. The 30-to-60-day delinquency roll rates are consistently around 60%. Cure rates have also stabilized at approximately 20%
• Despite rising delinquencies, investors are not significantly concerned about losses

Figure 70: RMBS credit issuance to date...

Source: J.P. Morgan, Bloomberg Finance L.P.

Figure 71: ...and spreads

Source: J.P. Morgan
Note: Includes our on-the-run indices. Jumbo is TOAS, non-QM and SFR are spread to treasuries. CRT is SOFR DM@10CPR. HG/HY are spread to treasuries.

Market Commentary:

The Fed gave the market what it wanted with a 50bp rate cut. However, mortgage rates were already pricing in this move, so anything different would have meant a reversal of the recent trend lower in rates. We are already hearing of some borrowers getting rate locks at sub-6% levels. Headline agency mortgage rates are 6.08%. Interestingly, they were a few basis points lower before the Fed decision. Regardless, mortgage rates are poised to be below-6% by year end. Based on current 2y, 5y and 10y UST forecasts from our rates strategists, 30y mortgage rates should reach 5.9%. Of course, it’s very possible that mortgage rates start to price in beyond end of year expectations. While mortgage rates are 75bp lower over the past 3 months, existing home sales (demand) has not budged. In fact, the most recent print was down -2.5% (more than street forecasts of -1.3%). Borrowers remain less reactive to the move in rates, with just over 60% of borrowers content with a 4% or lower mortgage rate. Interestingly, investors are starting to focus more on convexity risk in jumbo collateral as 6% coupons start to shift more in-the-money. This could begin to pressure premium price spreads. Non-QM should offer a convexity advantage with a heavier investor share of loans and prepayment penalties.

Our Securitized Products Real Estate and Consumer (SPREC) conference on Friday, September 13^th, was well received by investors with nearly 300 participants. See here for a full recap.

DQs march higher in non-QM

Delinquencies have continued to rise in the non-QM segment. FHA delinquencies are also higher, while GSE and jumbo borrowers have not experienced any significant increase in delinquency rates. This is clear in Figure 72, which shows the share of loans that are 60+ days delinquent for those originated since 2017. The rapid increase in delinquencies has been driven by more recent vintages. Non-QM loans originated since Q2 2022 have particularly underperformed ( Figure 73). It is challenging to attribute this higher level of delinquencies to any specific factor. Borrowers from these periods have slightly lower FICO scores and higher LTV ratios, but these differences are not substantial enough to explain the larger increases in delinquencies. However, we also observe a similar trend in other areas of consumer lending, such as subprime auto.

Figure 72: Delinquencies have continued to increase in non-QM

60+ DQ (%) for loans originated since 2017

Source: J.P. Morgan, CoreLogic, Freddie Mac, FHA

Figure 73: Non-QM loans originated since Q2 2022 have particularly underperformed

Non-QM 60+ DQ (%) by origination quarter/year and WALA

Source: J.P. Morgan, CoreLogic

Delinquencies have increased across all documentation programs, but bank statement and ‘other’ borrowers have more significantly underperformed ( Figure 74). In 2022 vintage loans, bank statement and ‘other’ cohorts have 4.5% of loans that are 60+ days delinquent. Meanwhile, 60+ day delinquencies in full documentation and DSCR cohorts stand at 3.2% and 2.7%, respectively, as of the latest remittance. In the ‘other’ category, loans underwritten using P&L statements and written verification of employment (WVOE) represent the two largest cohorts. A significant share of loans have no documentation type reported, and these have relatively high DQs as well. Delinquencies are highest for P&L statement and foreign borrowers, reaching almost 7% in August.

Figure 74: Delinquencies have increased across all documentation programs, but bank statement and ‘other’ borrowers have more significantly underperformed

60+ DQ (%) for 2022 vintage non-QM broken out by doc type

Source: J.P. Morgan, CoreLogic

Figure 75: Delinquencies are highest for P&L statement and foreign borrowers within the ‘other’ category

60+ DQ (%) for 2022 vintage non-QM broken out by ‘other’ doc type

Source: J.P. Morgan, CoreLogic

Current-to-30 day delinquency rolls have gradually increased across all documentation programs. Approximately 1.5% of bank statement and ‘other’ loans are becoming delinquent each month ( Figure 76). A smaller share of DSCR and full documentation loans are becoming delinquent. However, there is no significant difference in later-stage roll rates among different documentation types. The 30-to-60-day delinquency roll rates are consistently around 60% across the board ( Figure 77). Cure rates have also stabilized at approximately 20% ( Figure 78). Interestingly, despite strong HPA, the delinquency-to-payoff roll rates have not changed. About 2-3% of borrowers who are 90+ days delinquent pay off their loans each month ( Figure 79), likely by selling their properties. We also note that aside from higher current-to-30-day rolls, other delinquency roll rates, such as cures, remain largely unchanged across vintages.

Figure 76: Bank statement and ‘other’ doc type borrowers have higher current-to-30 roll rates

2022 vintage current-to-30 roll rate (%)

Source: J.P. Morgan, CoreLogic

Figure 77: The 30-to-60-day delinquency roll rates are consistently around 60%

2022 vintage 30-to-60 roll rate (%)

Source: J.P. Morgan, CoreLogic

Figure 78: Cure rates have also stabilized at approximately 20%

2022 vintage DQ-to-current roll rate

Source: J.P. Morgan, CoreLogic

Figure 79: About 2-3% of borrowers who are 90+ days delinquent pay off their loans each month, likely by selling their properties

2022 vintage 90+ DQ-to-paid off roll rate

Source: J.P. Morgan, CoreLogic

Despite rising delinquencies, investors are not significantly concerned about losses. Servicers have not been offering loan modifications. Roughly 20% of loans that are 90+ days delinquent have been transitioning to foreclosure each month.Even if liquidations increase, the current LTVs of delinquent borrowers are below 70%, and the housing market remains supported by a lack of housing supply. Loss severities on the few loans that have been liquidated have been under 2%. However, investors in pro-rata structures should be mindful of rising delinquencies potentially hitting the delinquency triggers. Most deals fail the trigger when the six-month average of 60+ day delinquencies and past 12 months of modifications exceed 10-15% initially. The failure of the trigger results in the senior stack shifting from pro-rata to sequential. At current delinquency levels, we do not anticipate many deals breaching the trigger. In our analysis, only CSMC 22-NQM6 is close to failing the DQ test. However, if delinquencies continue to rise at the current pace, AA and A classes could face extension risk in certain deals.

At long last

• At long last, the FOMC cut the funds target by 50bp to 4.75-5% in response to moderating inflation but weakening labor markets. It undoubtedly was a welcome move to CRE/CMBS market participants although this cut alone isn’t too meaningful for the market in our view. The cumulative impact of cuts to terminal and how inflation/labor data evolves over the next several quarters will be more impactful on the forward path for CRE/CMBS. CRE property prices seem to be bottoming, REIT stocks have rallied, and CMBS has rebounded well ahead of this cut in anticipation of the pause in rate hikes and an eventual cut
• For now, spreads can grind tighter on sentiment, and the spread curve can bullishly flatten. Looking at the on-the-run conduit CMBS capital stack, we continue to believe the best relative value can be found in upper IG mezz, and particularly in the AS bonds
• Our securitized products research team in partnership with the businesses hosted our inaugural (and in-person only) Securitized Products Real Estate and Consumer (SPREC) conference on Friday, September 13^th. We provide some high level takeaways from our CMBS investor panel
• Over the past couple of years, concern over potential extension risk for conduit CMBS bonds has grown due to increased rate volatility and fundamental challenges in specific sectors, namely the office sector. To further understand this risk, particularly in which part of the capital stack it is most prevalent, we reviewed the timing of when 2010-2013 vintage conduit CMBS bonds have paid off
• To gauge the prevalence of bonds paying off later than initially expected, we calculated the empirical WAL of each 2010-2013 vintage conduit CMBS bond that has fully paid down and compared it to its original WAL at issuance. For all outstanding bonds, we calculated WAL under five different scenarios. It becomes apparent that WAL drift increases as you go down the capital stack, particularly at single-As and below

Weekly market snapshot

Market commentary - at long last

At long last, the FOMC cut the funds target by 50bp to 4.75-5% as our economists expected in response to moderating inflation but weakening labor markets. It undoubtedly was a welcome move to CRE/CMBS market participants although this cut alone isn’t too meaningful for the market in our view. The cumulative impact of cuts to terminal and how inflation/labor data evolves over the next several quarters will be more impactful on the forward path for CRE/CMBS. CRE property prices seem to be bottoming, REIT stocks have rallied, and CMBS has rebounded well ahead of this cut in anticipation of the pause in rate hikes and an eventual cut ( Figure 82).

Figure 82: CRE debt and equity started to rebound well ahead of this past week’s rate cut. The cumulative impact of cuts and how inflation/labor data evolves over the next several quarters will be more impactful on the forward path of CRE/CMBS

IG CMBS price, CRE property price, and REIT equity price levels versus Fed Funds Target (upper bound) (%)

Note: BBG IG Non-Agency CMBS Index Price, MSCI Real Assets CPPI, RMZ all indexed to 100 as of September 2019. RMZ drawdown in 2020 is cut off to better visualize recent trends.

Source: J.P. Morgan, MSCI Real Assets, Bloomberg Finance L.P.

For now, spreads can grind tighter on sentiment, and the spread curve can bullishly flatten. Looking at the on-the-run conduit CMBS capital stack, we continue to believe the best relative value can be found in upper IG mezz, and particularly in the AS bonds. Figure 83 shows CMBS spreads as of Wednesday’s close (spreads are flat to tighter post-FOMC) and their changes versus comparable spread products. The spread pair differences are levels at all periods. We also include 10yr averages and the current levels versus 10yr averages. This shows that AS and BBB-s stand out as bonds that remain wide to their 10yr averages. Further, they remain wide to their spread comps versus their 10yr average relationships. As we discuss below, the historical evidence of paydown timing for AS bonds for the earliest 2.0 vintages show there is minimal extension risk. From this perspective, we think these remain much too wide to LCF AAAs. BBB-s have been a great momentum trade this year and while higher WACs/debt yields on newer issue provide some comfort around default/extensions risks, these bonds have demonstrated that they, not only trade more like High Yield, but can trade very idiosyncratically. Class sizes are very small and while they can continue to tighten, they will be more susceptible to macro vol events.

Our securitized products research team in partnership with the businesses hosted our inaugural (and in-person only) Securitized Products Real Estate and Consumer (SPREC) conference on Friday, September 13^th. I moderated a CMBS investor panel of real estate credit experts and here are some high-level takeaways.

CRE fundamentals. CRE property prices have bottomed or are close to bottoming for the most part with rate cuts ahead. Rent growth has held up for most of the major property sectors. Office remains a secular demand problem and credit issues here will take years to play out. But even for office, there are winners and losers. Green shoots are appearing in submarkets like the Manhattan Grand Central corridor, but areas are like downtown LA are still deeply challenged.

Multifamily will be on better footing in the coming quarters as construction deliveries are peaking and are set to decline rapidly. Meanwhile demand is likely to remain firm given housing affordability challenges.

Data centers are a fast-emerging sector with favorable near-term fundamentals given infrastructure-related constraints (power/cooling) that will limit construction supply relative to rapidly growing demand. Panelists favor debt backed by data centers leased to hyperscalers in primary markets. At least one panelist was cognizant of obsolescence risks for older stock that back longer duration debt given the exponential growth in the tech and costly buildouts.

CMBS credit fundamentals. Panelists agreed that the relatively benign macroeconomic/rates setup is allowing the CMBS credit cycle to be a slow burn. Lacking a proper recession that causes significant cashflow impairment to CRE (other than office), the market continues to have incentive to modify/extend loans backed by assets that are not suffering from fundamental problems, delaying ultimate defaults and foreclosure liquidations. CRE CLOs have been one of the more stressed CMBS sectors, but the panelists viewed issues in this sector as largely a borrower equity impairment issue rather than a debt impairment problem.

The path forward for CRE financing. CMBS issuance has had a resurgence this year with year-to-date issuance up ~250%, led by large private equity-backed single asset/borrower refinancing deals. Benign macroeconomic conditions combined with lower rates (‘soft landing’) should benefit CMBS issuance going forward as acquisition financing makes a comeback in addition to refinancing deals. Regional banks remaining backfooted on new CRE lending also presents an opportunity for non-bank lenders/private credit. CMBS and non-bank lenders/private credit can grab a larger slice of the CRE financing pie over the coming years.

Underwriting trends. The new issue market slowed down into QT/Fed hikes but didn’t experience a hard reset like it did in the aftermath of the GFC. Coming out of the GFC, the new issue machine put its best foot forward. That is, arguably, not what happened this time as a relatively slow mark-to-market of cap rates has generated new issue deals that have been met with some suspicion on quality. One panelist cautioned for investor vigilance on new issue underwriting and even among super senior buyers who drive a lot of deal execution.

Ratings Tracker

Figure 85: CMBS ratings downgrades to upgrades ratio since March 2021

Source: J.P. Morgan, Bloomberg Finance L.P.

Primary Markets

WALking the line

Over the past couple of years, concern over potential extension risk for conduit CMBS bonds has grown due to increased rate volatility and fundamental challenges in specific sectors, namely the office sector. To further understand this risk, particularly in which part of the capital stack it is most prevalent, we reviewed the timing of when 2010-2013 vintage conduit CMBS bonds have paid off. We focused our analysis on bonds originally rated LCF AAA to BBB-. In total, our analysis included 490 bonds with an original notional value of approximately $60.2 billion ( Figure 87). As of August 2024, 361 of these bonds have been fully paid down, but 129 remain outstanding. Only one LCF AAA and 8 AM/AS bonds are outstanding today. However, as we move down the capital stack, the count of outstanding bonds increases. Currently, 66, or 56%, of 2010-2013 vintage conduit CMBS bonds originally rated either BBB+, BBB, or BBB- are still outstanding.

To gauge the prevalence of bonds paying off later than initially expected, we calculated the empirical WAL of each 2010-2013 vintage conduit CMBS bond that has fully paid down and compared it to its original WAL at issuance. Our methodology for calculating empirical WALs involved finding the weighted average of all principal paydowns and writedowns over the bond's life. The difference between a bond’s original WAL and its empirical WAL is what we describe as “WAL drift.” As Figure 88 highlights, the magnitude and direction of WAL drift vary greatly by original rating. 2010-2013 vintage LCF AAAs were, on average, about 4 months shorter than the average original WAL, and over 95% of LCF AAAs were paid off in a shorter period than their original WALs indicated. In fact, the LCF AAA of the JPMCC 2011-C5 deal had an empirical WAL more than 2 years less than its original WAL at issuance, largely due to two loans, Orland Park Place and Bird Creek Crossing, prepaying 5 years early. As for AM/AS bonds, 99% of them saw WALs extend by no more than 3 months relative to their original WALs. However, as we move down the capital stack, particularly for originally A-rated and below bonds, we see the level of prepayments decline and empirical WALs extending more substantially beyond their original WALs. This dynamic is most apparent in bonds rated BBB+, BBB, or BBB-. On average, WAL drift for these bonds is about +3 months, and about 9% of these bonds have seen WAL extension of at least a year.

As mentioned earlier, there are still about 129 outstanding 2010-2013 vintage conduit CMBS bonds that were originally rated LCF AAA through BBB-. These bonds represent 58 total deals, with 32 of the deals being from the 2013 vintage. A review of the underlying loans for each of these deals shows that the outstanding loan count per deal is low, ranging between 1 and 11. While the loan count might be low, the average loan balance is relatively high at $58mn. To put that in context, the average loan balance for all 2010-2013 conduit CMBS loans was just $23.4mn. Additionally, the underlying properties backing these loans are largely concentrated in the retail sector, specifically regional malls, and the office sector, at 60% and 26%, respectively. The underlying credit for these deals, which have now been outstanding for over 10 years, highlights the adverse selection that takes place in the conduit CMBS space as deals pay down. It appears that the longer conduit CMBS deals stay outstanding, the more likely they become concentrated with relatively large loans backed by fundamentally challenged properties that are difficult to refinance.

For the outstanding bonds that have experienced some principal paydowns, we’ve calculated empirical WALs to date and WAL drift to date, which are highlighted in Figure 89. Across the capital stack, we see that the average WAL drift to date is at least 6 months, except for the LCF AAA bond. The only outstanding 2010-2013 conduit CMBS LCF AAA is from the COMM 2012-CR4 deal, but 87% of its original balance had already paid down by the time the bond was outstanding for 10 years, helping to limit its WAL drift. However, the empirical WAL to date figure we’ve calculated for these bonds doesn’t fully capture how long it will take for them pay down entirely. To better estimate what the empirical WAL for these bonds may end up being, we ran each bond through four different scenarios which we describe below. We then used the forecasted principal paydowns and writedowns generated from each scenario run to compute a forecasted WAL.

Focusing on our zero scenario, we see WAL drift increase as we move down the capital stack. For originally AA-rated or below bonds, average WAL drift reaches at least 1.7 years ( Figure 90). Under this scenario, we estimate that over 43% of the outstanding originally AA-rated or below bonds will experience WAL drift of at least 18 months. Our debt yield-based scenario is even more punitive to these bonds, particularly for bonds originally A-rated or below ( Figure 91). This scenario estimates that bonds originally A-rated will experience WAL drift of about 3.5 years, and that 34% of these bonds will experience WAL drift greater than 4 years. About half of the outstanding BBB-rated bonds will experience WAL drift of at least 4 years under this debt yield-based scenario as well.

Figure 90: 43% of originally AA-rated or below will experience WAL drift of at least +18 months in our zero scenario

Forecasted WAL drift distribution under our zero scenario (lhs) and average forecasted WAL drift (rhs)

Source: J.P. Morgan, Trepp, Bloomberg Finance L.P., Intex

Figure 91: In our debt yield based scenario, positive WAL drift is more severe, particularly for lower IG mezz

Forecasted WAL drift distribution under our debt yield based scenario (lhs) and average forecasted WAL drift (rhs)

Source: J.P. Morgan, Trepp, Bloomberg Finance L.P., Intex

This analysis further emphasizes that the higher a bond is on the capital stack, the better protected it is from extension risk as we had alluded to in our refi success rate analyses throughout last year (see here). Reviewing the empirical data for LCF AAAs and AM/AS bonds highlights that these bonds have not seen many extensions and are more likely to experience prepayments. However, extension risk should be a greater concern as you go down the conduit CMBS capital stack, particularly A-rated bonds or lower. A significant portion of bonds originally rated A or below from the 2010-2013 vintage are still outstanding. An estimate of the WAL for this population of bonds can be achieved by taking a weighted average of the empirical data from bonds that have paid down and the data from our “Zero” scenario run. For A-rated bonds, we get a “blended” WAL of about 10.7 years, and for BBB bonds, we get 11.3 years, both greater than their original WALs at issuance. This highlights that this portion of the capital stack has shown a propensity to extend in the past, and it is something investors should be aware of when making investment decisions.

Weekly Tracker

Figure 92: Delinquency rate

Conduit CMBS 30-day+ delinquency rate including FC/REO and NP matured (%)

Source: J.P. Morgan, Trepp

Figure 93: Delinquency cure rates

Conduit CMBS 30 day+ delinquency to performing transition rates (%)

Source: J.P. Morgan, Trepp

Figure 94: Specially serviced rate

Conduit CMBS percentage of loans in special servicing (%)

Source: J.P. Morgan, Trepp

Figure 95: Office RTTO indexed to pre-pandemic levels

Kastle System Back to Work Barometer, weekly

Source: Kastle Systems, Bloomberg Finance L.P.

Cross Sector Spreads

Cross Sector Spreads (continued)

September 20, 2024

• ABS spreads held firm with heavy supply packed in before, and some selling of short-dated paper after, the Fed’s first rate cut of the cycle
• The auto lease ABS segment has seen volumes, return rates, and residual value trends normalizing along with the aggregate leasing industry supply and used car values
• Auto lease ABS continues to offer an attractive spread pickup compared to prime auto loan ABS

ABS spreads holding up to September supply crunch

ABS spreads continue to hold firm under heavy supply for another September week. In the secondary market, there was some selling of short dated ABS after Wednesday’s highly anticipated first rate cut of the cycle by the Fed (of 50bp). The inverted yield curve, particularly at the front end, has provided a significant boost to short ABS returns. While some adjustment to yield curve shifts is expected, we do not anticipate significant pull back in demand for ABS as there remains an attractive spread pickup over Treasuries and comparable credits. In addition, as far as an inverted curve typically signals heightened recession risks, high quality ABS has served as a relatively stable flight-to-quality safe haven during periods of broad credit sell-offs.

In the new issue market this week, we are still seeing strong subscription and solid pricing levels across sectors, with compressed spread tiering for sponsor risk. For example, in credit card ABS this week, a new issue printed at I+50bp, versus our benchmark AAA 3-year bankcard ABS steady at +47bp. The same bank sponsor issued back in mid-May 2023 at I+72bp, with a 7bp pickup on benchmark levels at the time. The issuance pace in card ABS this year has lagged the same period last year, $17.1bn versus $18.0bn. The tight spreads and lower rates should make the all-in yields more attractive to issuers, both programmatic and opportunistic, for incremental supply. We think FNMNT, SYNIT, WFCIT and WFNMT offer attractive spread pickup on the top bankcard ABS programs ( Figure 96). Credit metrics on those master trusts are in line with overall sector trends with robust excess spread and credit enhancement to protect AAA bondholders (for details please see our Credit Card ABS Monthly Update published earlier this week). Additionally, Canadian banks also offer attractive pickup versus their US peers in card ABS. Back in July, CARD2 (CIBC) and EVGRN (TD) issued AAA 2-year floaters in the primary market, both pricing at SOFR +68bp, a pickup of roughly 10bp over top US bankcard ABS at the time. In our opinion, the concession reflects no credit concern with the Canadian bankcard ABS given their solid trust performance and strong bank sponsor ratings, though some pickup is justified for the liquidity of the smaller 144A Canadian programs.

Auto lease ABS update

Herein, we review the auto lease ABS segment focusing on supply, collateral quality, performance trends, residual values and bond pricing. First, we start with auto leasing industry trends. With improvements in inventory and supply trends, along with the normalization and stabilization of transaction prices since the pandemic, lease volumes are recovering and have increased further in recent months. The lease penetration rate, which is the percentage of retail sales that were leased as opposed to paid in cash or financed via loans, has risen sequentially, reaching 23.5% in August from 20.2% in 2023 and 17.3% throughout 2022 ( Figure 97). For context, this penetration rate averaged 30.6% pre-pandemic (2017 through 2019). According to Experian’s “State of the Automotive Finance Market” 2Q24 report, the share of new vehicles leased increased to 25% from 21% in 2Q23 and 19% in 2Q22. Additionally, lease originations remain heavily concentrated in the prime credit segment, with 47% of new leases in 2Q24 classified as super prime (credit score 781-850) and an additional 40% classified as prime (credit score 661-780). The average credit score for new vehicle leases was 751, compared to 745 for loans, up from 748 and 743, respectively, in the prior year. The report also highlighted that electric vehicles (EVs) accounted for about 8.4% of all purchases in 2Q24, with over 47% of these EVs financed by lease, 37% through loans, and the remaining 17% through cash ( Figure 98).

Figure 97: Retail lease penetration rate and volumes

Source: Autodata, Cox Automotive, Edmunds, Moodys Analytics and J.P. Morgan estimates

Figure 98: Breakdown of financing type shares for EV purchases

Source: Experian

Next, we analyzed the loan-level lease tapes for public SEC-registered auto lease ABS transactions to determine the share of terminated leases that were returned over time. Since the onset of the pandemic through the first half of 2021, there was a continuous sequential decline in vehicle return rates on terminated leases. This decline reflected supply constraints and an increase in used car prices. This trend of lower returns (and higher retentions) persisted through most of 2022 but has since started to normalize. Given that TESLA ABS deals are not public SEC-registered and do not provide loan-level data, we used Tesla Finance’s lease servicing portfolio (aggregate book) as a proxy for EV return rates. We based our analysis on return rates from actual terminated leases for each year ( Figure 99). Similar to the aggregate trend, TESLA return rates declined to 65% in 2022 from 90% in the prior year. However, these rates have since increased more rapidly, reflecting a steeper decline in used car prices for Tesla vehicles compared to the aggregate market. We expect this trend to continue normalizing in line with used vehicle prices, higher incentives on new purchases, and an improving supply and inventory backdrop.

Figure 99: Declining used car values resulting in higher lease return rates recently

Source: J.P. Morgan, ABS-EE via 1010DATA, ABS deal documents, ManheimNote: Aggregate includes public SEC registered transactions for BMWLT, FORDL, GMALT, NALT, VWALT and WOLS lease ABS transactions; return rate calculated as a % of all leases terminated in the month. For TESLA return rate is from ‘Residual Performance Experience of Tesla Finance’s Lease Servicing Portfolio in deal documents; Base year January 1997 = 100 for Manheim Used Vehicle Value Index

With the aggregate uptick in lease volumes, year-to-date auto lease ABS supply is also incrementally higher at $25.8 billion, surpassing the full-year issuance of $22.9 billion in 2023 and $16.2 billion in 2022 ( Figure 100). This accounts for about 20% of the auto-related ABS issuance this year. Lease ABS are generally issued by captive finance companies of automakers and are backed by leases of luxury vehicles to prime and high-prime consumers. So far this year, Santander Bank NA (SBNA) and Hyundai’s HALST have led the issuance tally with $4.0bn each, followed by GM Financial’s GMALT at $2.8bn. Notably, Porsche tapped the auto lease ABS market for the first time since 2015 with its PILOT 2024-1 $0.9bn transaction.

Auto lease ABS transactions are backed by very high-quality underlying collateral pools ( Figure 101). For instance, the FICO score range for pools backing year-to-date auto lease ABS transactions ranged from 747 to 798. Porsche’s PILOT 24-1 represents the upper tier of that range with a weighted average (wa) FICO score of 790, while HALST 24-B and GMALT 24-2 report (wa) FICO scores of 776 and 781, respectively. The pools for Mercedes and Porsche have average securitization amounts of $55,129 and $68,635, respectively, with the offset being lower residual value shares at 46% and 52% of the securitization value on a discounted basis. The original terms for most leases tracked around 35-38 months, except for MBALT 24-A, which has an original term of 42 months with 10 months of seasoning. AAA target credit enhancements remained largely in line with last year’s structures, ranging between 12.5% for MBALT 24-A and 23.0% for FORDL 24-B.

Given the prime to super prime collateral, default risk across lease transactions tends to be very low, with losses rounding to zero. However, the bulk of the credit risk in lease ABS stems from residual value, which makes up a significant portion of the securitization/asset base. Year-to-date, lease transactions have 46%-65% of total securitized value linked to residuals (on a discounted basis). Residual value losses arise when the actual car value at the end of the lease term is less than the base residual value securitized. In ABS structures, the base residual value is typically the lower of the stated/contractual residual value or the ALG (Automotive Leasing Guide) residual value, with the more conservative measure helping to mitigate the risk of inflated residuals.

Reflecting the broad auto industry trends (pandemic-driven inventory crunch, higher prices, lower returns, etc.), the actual residual value experience has been largely positive for most of 2021 and 2022. However, more recently, in line with declining used car prices, residual experience has drifted lower across sponsors ( Figure 102). For example, Nissan recorded a 6.84% residual value loss for the twelve months ending March 2024 for vehicles returned to Nissan, versus gains of about 25% for the prior two years. The return rate (vehicles returned to NMAC as a share of total leases scheduled to terminate) tracked 15.2% for 2024 compared to 2.2% for 2023 and 13.2% for 2022. Based on the declines in used vehicle values from peaks in 4Q21, we do expect softer residual value prints across sponsors in the upcoming months. Historically, negative residual experience has been a frequent occurrence across sponsors and there has been no negative rating impact on lease ABS. Any negative value experience on residuals does not translate directly into stress on ABS pools, as residuals account for about 50%-60% of securitized pools, and the impact is further muted by the return rate. Furthermore, there is typically 15%-20% of hard AAA enhancement/support on the auto lease ABS structures to protect bondholders.

In terms of relative value, auto lease ABS trades wider than comparable prime auto loan ABS. Historically, this differential has fluctuated from minuscule to notable concessions depending on overall credit spreads as well as used car value expectations and auto maker sponsor headlines. For instance, the spread concession for A-4 new issue lease ABS versus indicative benchmark prime auto loan ABS was 4bp in 2021, 18bp in 2022, 25bp in 2023, and 11bp year-to-date in 2024. Earlier this year, VWALT 24-A and WOLS 24-A priced with a 2bp spread pickup versus our indicative benchmark prime auto spreads. Recently, SBALT 24-C priced last week with its A-4, AAA-rated 2.37-yr tranche at I +85bp, offering a 17bp pickup over prime auto loan ( Figure 103). We believe auto lease ABS offers an attractive spread pickup compared to prime auto loan ABS, considering the mostly prime to super prime collateral, solid credit performance trends and robust structures, despite slightly lower liquidity and higher sponsor/auto maker exposure.

Figure 103: New issue auto lease ABS A-4 tranche pricing spread versus indicative prime auto loan ABS AAA 3-year spread

Spreads (bp)

Source: J.P. Morgan

Week in review

Fifteen ABS transactions totaling $11.3bn across bankcard, auto loan, auto lease, motorcycle, aircraft, solar, rehabilitated FFELP, equipment, data center and aircraft priced this week. This brings 2024 year-to-date ABS supply to $251bn versus $200bn recorded over the same period last year. September month-to-date supply stands at $28bn versus $29bn for all of September 2023. Looking to the forward calendar, we see 8 deals totaling roughly $5bn currently in pre-marketing. ABS spreads were unchanged on the week.

On September 18^th, KBRA downgraded tranches A-2 and B of HOA 2021-1, that were placed on watch negative on 27^th June 2024, from BBB- to BB- and BB to B, respectively. The rating agency cited deteriorating performance of the Hooters system, including a declining trend in same store sales (SSS) and system-wise sales (SWS), the closure of roughly forty company owned locations and a noticeable dip in transaction DSCR.

TGIF 2017-1 A-2 was placed on watch negative by S&P on 11th September 2024 following the declaration of a manager termination event by the company on 3rd September 2024. The rating agency noted the $4.7million interest reserve, which currently appears sufficient for interest coverage on the upcoming 12 months. However, S&P also noted the potential disruptions that the termination event could trigger and believes that the tranche is unlikely to withstand significant cash flow disruptions at current rating level of ‘B-’. Of note, the tranche has seen multiple negative rating actions previously, with the most recent one being in Feb 2024 wherein the notes were downgraded from B to B- due to poor sales performance, closure of 36 stores and continued deferral of company-owned restaurant royalty payments.

Data appendix