Cross Sector P. White, L. Wash, H. Cunningham
Softening inflation should allow the Fed to cut in November followed by a quarterly cadence thereafter. We remain patient to add duration and prefer carry trades. We think mortgages have the ability to outperform corporate credit in a high-for-longer environment. Balance sheet runoff should continue through year-end and we see limited funding stress implications. We revise higher our 2024 net issuance forecast to $3.275tn.

Governments J. Barry, P. White, A. Borges, L. Wash
We expect a rangebound environment for Treasury yields in the months ahead. Even shallow easing cycles support lower yields, but Fed pricing and carry implications make it challenging to be long duration. We continue to see value in 5s/30s steepeners to position for Fed easing and fiscal dynamics. We discuss the impact of different election outcomes on the fiscal outlook and the Treasury market. We update our Treasury supply and demand forecasts. Remain bullish at the front end of the inflation curve. Though valuations have moved closer to fair, we expect a positive risk premium to emerge ahead of the US election: hold 1Yx1Y inflation swaps. Stay overweight $-SSA spreads.

Interest Rate Derivatives S. Ramaswamy, I. Ozil, P. Michaelides, A. Parikh
Swap curves should drift steeper in 2H24, but this is mostly priced into forwards except for curves anchored in the back end. Option-implied Fed expectations have coalesced around a moderate cutting path, which implies lesser jump risk and favors carry trades on the curve as well as short-gamma strategies. Term funding is expected to drift higher in coming months - we look for wider swap spreads in the front end, but long end spreads are likely to remain flat to current levels. We describe a new approach to trading the volatility of yield curve principal components that opens up a new class of relative value gamma trades - selling the volatility of the second principal component is likely to be attractive in coming months.

Short-Term Fixed Income T. Ho, P. Vohra, H. Cunningham
Funding conditions should remain benign, with liquidity remaining abundant, limiting any potential impacts to spreads. MMF reform should not have any outsized impact on money market credit spreads. Treasury repo clearing is work in progress, but concerns about readiness are emerging. Total money market supply (ex-Fed) should increase by nearly $550bn in 2H24, driven largely by T-bills and dealer repo.

MBS and CMBS J. Sim
Mortgages are a bit snug, but not terribly so. If our house rate forecast of a single cut this year is realized, higher coupons should again outperform in the second half of the year. CRE transaction cap rate expansion has been muted year-to-date through May. CMBS is no longer a ‘cheap’ asset class and spreads across the stack appear full.

ABS and CLOs A. Sze, R. Ahluwalia
We expect that ABS should continue to outperform over 2H24, with tighter spreads despite brisk issuance, and structures holding up well against minimal credit deterioration. Demand for floaters is solid and CLO T1 AAAs swept through our 140bp MY target and are closer to our 130bp YE target we set back in April.

Investment-Grade CorporatesE. Beinstein, S. Doctor, N. Rosenbaum, S. Mantri
Spreads to stay tight, though we revise our YE24 HG bond spread forecast to 110bp from 95bp (current 109bp). 2H24 Demand to be less strong with slower growth and election uncertainty. We also revise gross issuance forecast up to $1.45tr.

High Yield N. Jantzen, T. Linares
We believe leveraged credit will provide carry-esque returns over the next six months and spreads will finish 2024 well inside the long-term average.

Municipals P. DeGroot, Y. Tian, R.Gargan
Finding sustained market consensus while transitioning to an easing cycle may be difficult, but we suggest investing with a longer-term perspective, particularly in periods where Treasuries sell-off. We look for 2H24 IG muni index performance of 3-3.5% and favor adding idiosyncratic municipal risk opportunistically on market weakness, and highlight potential market cheapening in the period before the election.

Emerging Markets L. Oganes
In EM fixed income, we are MW GBI-EM local rates, CEMBI and EMBIGD.

Source: J.P. Morgan

2024 Mid-Year Outlook

• We expect core PCE inflation to soften sequentially and GDP growth to step down to 1% saar in 2H24. We see the first Fed rate cut in November, followed by a quarterly cadence thereafter. We think balance sheet runoff can continue at the now-tapered pace over the remainder of the year
• We revise up our 2024 long-term net issuance forecast across US fixed income markets by $225bn to $3.275tn given heavy corporate credit supply in 1H24. We do not expect further Treasury coupon auction size increases until November 2025
• Treasuries: We expect a rangebound environment for Treasury yields in the months ahead. Even shallow easing cycles support lower yields, but Fed pricing and carry implications make it challenging to be long duration. We continue to see value in 5s/30s steepeners. Remain bullish at the front end of the inflation curve. Though valuations have moved closer to fair, we expect a positive risk premium to emerge ahead of the US election
• Derivatives: Option-implied Fed expectations have coalesced around a moderate cutting path, which implies lesser jump risk and favors carry trades on the curve as well as short-gamma strategies. Term funding is expected to drift higher in coming months. We look for wider swap spreads in the front end, but long-end spreads are likely to remain flat to current levels
• Short Duration: Funding conditions should remain benign, with liquidity remaining abundant. MMF reform should not have any outsized impact on money market credit spreads
• Securitized products: Mortgages are a bit snug, but not terribly so. If the Fed delivers one cut this year, higher coupons should again outperform in the second half of the year
• Corporates: We look for HG spreads to stay tight, though we revise our YE24 HG bond spread forecast to 110bp from 95bp. 2H24 demand to be less strong with slower growth and election uncertainty. We also revise our gross issuance forecast up to $1.45tn
• Near-term catalysts: May JOLTS (7/2), June FOMC minutes (7/3), June employment (7/5), June CPI (7/11)

The outlook for rates and spreads in 2H24

In our year-ahead outlook published last November, we predicted that the disinflationary trend that gained momentum in 2H23 would continue in early 2024, alongside slowing growth and a cooling in labor markets, driving further declines in yields and spread decompression. Instead, core inflation accelerated over the first quarter and job growth continued to surprise to the upside, driving yields higher and spreads tighter across most fixed income products. Yields have retreated from their highs as the last two months of inflation data have been more benign, and there are clear signs that activity growth has stepped down from elevated levels observed in 2H23. Still, Treasury yields are 40-70bp higher YTD, the curve has flattened, and high-yield credit has outperformed other fixed income products ( Figure 1). As we look ahead, we expect GDP growth to step down further toward a 1% pace in 2H24, but with markets already priced for roughly 140bp of easing over the next 18 months, we look for yields to remain broadly rangebound in the second half of the year and for spreads to trade sideways. Against this backdrop, seeking positive carry trade opportunities is likely to remain a theme. Moreover, the US election could be a key catalyst for fixed income markets in the second half of the year, as well as various elections that will take place globally over coming months.

Figure 1: Rising yields drove returns negative early in the year before reversing. HY credit has outperformed

Cumulative returns on the Bloomberg US Aggregate Bond Index, together with its subcomponents, YTD*; %

Source: Bloomberg Finance L.P.
*June data through 6/28/2024

Figure 2: We look for moderate economic cooling over the next year without an outright contraction

J.P. Morgan Economic Forecast; units as indicated

Source: J.P. Morgan
* Average level for the period

Economics & Rates: Consumer spending growth has moderated recently, as the post-pandemic tailwinds supporting last year’s exceptional growth have begun to normalize. While there are mixed signals from the labor market, the broad thrust of the data indicates that nominal labor income growth should continue coming back down to earth in 2H24, providing a further tempering influence on consumers. Business capital spending has been supported by IRA and CHIPS Act-related spending. The more recent data suggest that the growth in this policy-supported capex is starting to peter out. Outside of this category, business capital spending growth has been soft, and business surveys indicate that this should remain the case in coming months. Though the inflation process has been bumpy over 1H24, we look for core inflation to cool gradually on a sequential basis over the second half, though base effects should drive core PCE slightly higher again later in the year. We forecast core PCE inflation to run at 2.8% q4/q4 in 2024 ( Figure 2). Against this backdrop, we expect the Fed will cut rates at the November FOMC meeting and continue at a quarterly cadence in 2025. While a September cut remains a risk, we think it may take the five reports available between now and the November FOMC meeting to see a broader weakening in job growth (see Economics).

Meanwhile, markets are pricing roughly 140bp of easing over the next 18 months, slightly more than our own forecast ( Figure 3). Moreover, we recognize that the lack of clear economic vulnerabilities together with an uncertain inflation outlook may leave the upcoming easing cycle shallower than many prior cycles, perhaps more closely resembling the experiences of 1995 and 2019. However, even in those cycles, yields declined more than 125bp on average in the six months prior to the first cut. Thus, we believe there is still room for yields to decline ( Figure 4). Nonetheless, even though the onset of the easing cycle appears to be just a few months away, we will be patient in adding duration longs, given negative carry implications and outright valuations. Along the yield curve, we think long-end steepeners provide the best value as we approach the start of an easing cycle, especially as most curve pairs remain near their flattest levels of the last 3 years. We prefer 5s/30s steepeners: this curve displays a high sensitivity to medium-term Fed expectations, should benefit from rising term premium, and appears too flat versus our fair value framework (see Treasuries).

Figure 3: Markets are pricing in close to 140bp of easing over the next 18 months. We and the Fed are projecting 125bp of cuts

OIS forward rates by meeting date versus J.P. Morgan forecast and median SEP dot from June FOMC meeting; %

Source: J.P. Morgan, Federal Reserve Board

Figure 4: We look for yields to remain rangebound through the summer and for the front-end to lead the move to lower yields in the fall

J.P. Morgan US interest rate forecast; %

	Actual	3Q24	4Q24	1Q25	2Q25
	28-Jun	30-Sep	31-Dec	31-Mar	30-Jun
Rates (%)
Fed Funds	5.33	5.35	5.10	4.85	4.60
SOFR	5.34	5.35	5.10	4.85	4.60
2-yr UST	4.72	4.75	4.60	4.30	4.05
3-yr UST	4.51	4.55	4.40	4.15	3.95
5-yr UST	4.33	4.40	4.25	4.00	3.80
7-yr UST	4.32	4.40	4.25	4.00	3.80
10-yr UST	4.34	4.50	4.40	4.20	4.00
20-yr UST	4.60	4.70	4.60	4.40	4.20
30-yr UST	4.50	4.65	4.60	4.50	4.35

Source: J.P. Morgan

In inflation markets, we remain bullish at the front end of the curve. Even with inflation and growth projected to cool gradually, recession fears are likely to remain at bay, the disinflationary process is likely to be bumpy, and the market remains priced for weaker inflation outturns over the second half of the year relative to our own forecast. Additionally, as the US election comes into focus, we expect the front end of the inflation curve to price additional risk premium, as trade policy is one of the few issues garnering bipartisan support in a polarized US political environment. Importantly, in the event of a Trump victory, we would expect the Former President to take a harder line on foreign policy, particularly with China, through a greater use of Executive Orders. If Trump proceeds with the tariffs that have been floated, our economists estimate a 1.1% increase in price level from a 60% tariff on China and a 1.5% increase in price level coming from the 10% universal tariff (see US: Economic policy implications of the election, Michael Feroli, 6/28/24). With IOTAs trading near their tightest levels in nearly two years, we think inflation swaps can outperform cash breakevens, and we maintain longs in 1Yx1Y inflation swaps (see TIPS).

With Fed expectations coalescing around a path of moderate rate cuts and tails thinning, this implies reduced jump risk in markets going forward and makes swap yield curve carry trades attractive as a theme. We continue to advocate for well-hedged carry trades that mitigate directional curve exposure while enhancing carry-to-risk ratios. This also implies broadly supportive conditions for short gamma strategies. That said, implied volatility levels are close to our fair value estimates, suggesting gamma strategies should be actively managed in coming months. Swaps spreads have also traded in a wide range in 1H24, but not driven by a correlation with yield levels. Instead, swap spreads have reflected swings in macro factors affecting the entire term structure. Indeed we think term funding premium - the incremental compensation demanded by Treasury investors for each additional year of duration after hedging interest rate risk - has risen in part due to declining liquidity amid ongoing Fed balance sheet reduction (see Term Funding Premium and the Term Structure of SOFR Swap Spreads, Srini Ramaswamy, 4/29/24). We expect that QT can continue at the now-tapered pace for the remainder of the year, increasing term funding premium and leading the swap spread curve to twist flatter. The upcoming US Presidential elections add some upside risk to term funding premium, as concerns around fiscal expansion and longer-term risks to Treasury supply could provide incremental narrowing pressure at the long end (see Interest Rate Derivatives).

In 2H24, we believe funding conditions will continue to be relatively benign, though acknowledge there are a variety of risks that liquidity investors will have to grapple with in the coming months. Even with QT taper, the Fed’s balance sheet is still shrinking, albeit at a slower pace, which means reserves are still increasingly approaching scarcity. As of the time of writing, reserve balances registered $3.4tn. Market expectations have largely coalesced around $3.0-$3.1tn as the lowest comfortable level of reserves (LCLoR), and while our Fed balance sheet expectations call for reserve balances to remain comfortably above $3tn during 2H24, signs of reserve scarcity could start to emerge towards year-end as we come closer to approaching LCLoR ( Figure 5). The good news is that we also project ON RRP balances to remain near $400bn for at least a few months and over $300bn for the remainder of the year, which will help to offset any funding pressures arising from balance sheet normalization. Looking beyond year-end, we expect QT to come to a halt in 1Q25 when the O/N RRP balance falls to near or below $300bn--a reasonably sized RRP facility is likely necessary as a buffer against stress in funding markets, given that the marginal propensity of RRP investors to lend in response to spikes in repo rates is higher than the marginal propensity of banks to do the same. We could see increased rotation out of the credit side of money markets as the remaining institutional prime funds shore up liquidity heading into the MMF reform deadlines this October, though it is likely any cheapening in CP/CD will be minimal, as we suspect other liquidity buyers will step in. The scaling back of both the timing and depth of the upcoming Fed easing cycle has likely extended the attractiveness of the front-end, and we expect MMF AUMs will stay elevated for the remainder of the year (see Short-Term Fixed Income).

Figure 5: We estimate that Quantitative Tightening can continue over the remainder of the year at the now-tapered pace

Current* and projected** values for Fed balance sheet assets, RRP, TGA, Reserves, and Commercial bank deposits*** through 2024; $bn

Source: J.P. Morgan, FRED, Federal Reserve H.4.1, Federal Reserve H.8
* Current as of 6/27/2024 Fed H.4. release
** Deposits as of 6/21/2024 Fed H.8. release

Corporate credit: Corporate balance sheets continue to come under strain with credit metrics steadily deteriorating, but we think they are overall still in solid shape. For high-grade issuers, the recent deterioration has been driven primarily by negative EBITDA growth combined with rising debt and interest expense. Weakness in commodity prices and reduced profit margins are eating into EBITDA, leading to increasing leverage and interest coverage metrics, but they are within long-term averages ( Figure 6, see HG Credit Fundamentals: 1Q24 Review, Eric Beinstein, 5/31/24). Companies have responded to weakening earnings by lowering buybacks and dividends, which have contributed to positive rating developments. Indeed the ratings momentum in HG has been extremely robust with an upgrade/downgrade ratio at record 4.8x, and 4.5% of the HG market has been upgraded in 2024 on a net basis.

The story is similar down the credit spectrum where US HY issuer balance sheets remain strong heading into what could prove a more challenging 2H24. We have seen modest erosion across most credit metrics including declining EBITDA and rising leverage, but they have yet to reach worrying levels. Meanwhile inflation has been stickier on the cost side than the revenue side, degrading profit margins and leading coverage ratios to soften, even more so for private companies (see 1Q24 High Yield Credit Fundamentals, Nelson Jantzen, 6/12/24). That said these metrics still stand near long-term averages and ratings are historically benign. While the ratings narrative is less supportive for loan issuers, the product is already trading near a historic extreme discount versus bonds and we are seeing access to capital broadening for B3 or lower ratings amid a historic refi wave. A slowing but not contracting economy is an ideal recipe for corporate credit, especially as EPS growth is expected to increase from 3.5% in 1Q to 18% by 4Q, based on consensus equity analyst forecasts. As we look ahead, we think deteriorating credit metrics will not become a primary driver of spreads unless economic growth weakens materially and thus we think technicals will play a much larger role over the near-term.

Figure 6: Interest coverage ratio is declining but remains modestly above its pre-Covid level

Interest coverage metrics; %

Source: J.P. Morgan

Figure 7: We see default rates rising to 2.00% for bonds and 3.75% for loans by YE24

High yield and leverage loan default rates, yield, and spread forecasts for YE24; units as indicated

Source: J.P. Morgan*par-weighted default rates

Stronger-than-expected economic growth and higher-than-expected yields led to very strong demand technicals over the first half of the year. Despite this, HG spreads did not stay sustainably below 100b. In 2H24, we expect some of the demand technicals to become less strong, especially if the start of the Fed easing cycle drives yields lower. Also, global rates dynamics have made US credit less attractive to overseas investors, and domestic fund inflows have slowed in line with middling YTD total return. Election uncertainty could increase volatility, generally a market negative. Overall, we have revised our year-end spread target to 110bp (from 95bp previously)--still a tight spread historically (see Corporates).

For high yield bonds, weakening credit fundamentals, a less supportive technical backdrop, and negative agency rating trends are expected to be offset by supportive factors such as elevated yields, strong balance sheets, benign credit ratings, improving capital market conditions, resurfacing retail inflows, and a yoy decline in default activity. For leveraged loans, the negative fundamental ramifications from elevated rates and slower growth should lead to ongoing downgrade and default risk, which we expect to be offset by elevated coupons (9%), earnings delivery, a record pace of CLO origination, steady inflows, and a negligible pace of net new supply. We believe credit will provide carry-esque returns over the next six months, with high yield bond spreads widening 25bp to 380bp, comfortably inside the long-term average, while leveraged loan spreads end the year at 8.75%. We see default rates rising to 2.00% for bonds and 3.75% for loans by YE24 ( Figure 7, see High Yield).

Given our modest default forecast, we think CLO credit performance is unlikely to meaningfully turn in 2H24 unless the economy falls into a recession. That said, we think the 1H24 bullish tone in CLOs is susceptible to macro events from geopolitics to disinflation, as demand for floaters has largely been driven by yield curve inversion. While technicals could prove strong, with US CLO new origination at near-consecutive monthly records despite little net loan creation, portfolios could be at risk of future extensions as tight spreads encourage CLO resets and as the loan maturity wall extends. We look for our CLOIE to return 3.2% for the remained of the year. We like relative value in selectively selling par-plus AAA in secondary to buy short refi at par or buying deleveraging/near-deleveraging AA and A for low spread duration while earning carry (see CLOs).

In terms of sector recommendations, in HG we have overweights in Yankee Banks, Finance Cos, Consumer, Basics, Domestic Telecoms, Insurance and Media/Entertainment, while we have underweights in Technology and Retail. In HY we are overweight in Financials, Metals/Mining, Media, and Utility while we are underweight Consumer Products, Housing, Industrials, and Paper/Packaging.

Securitized products: Turning to the mortgage space, we head into 2H24 with OASs looking a bit snug, ZVs still historically wide, and the supply/demand picture looking manageable. Higher coupons have generally outperformed Treasuries thanks to a combination of carry and spread tightening while discounts have lagged on the back of technical pressures from bank portfolio reallocations. Even after portfolio repositioning in the second quarter, bank activity has surprised to the upside and we now think they can be a net buyer over the course of the year. That said, banks have generally been letting their lower yielding fixed rates burn off and replacing with Ginnie floaters, partially offsetting otherwise positive demand for mortgages. Risks to lower total capital build from a revised Basel 3 Endgame remains a medium-term risk and risk capital constrained banks will prefer Ginnies. Meanwhile money managers have been responsible for most net conventional buying to start the year. With corporate spreads moderately higher and YTD excess returns on higher coupons giving more credence to the mortgage vs. corporate trade, we think that money managers can maintain their current allocations given the yields available. If our base rates forecast is realized, higher coupons will likely outperform, but with OASs in the 20-30bp range and the marginal buyer still spread sensitive, we think excess returns will need to be driven by earned spread rather than tightening.

Given the importance of bank and money managers to mortgage demand, a key question among cross asset investors is whether mortgages look cheap on a relative basis to Treasuries and IG corporates. We have argued in the past this depends on which mortgage one buys, but we also note it depends on how one is using mortgages as either a relative value play or as a way to position for a potential recession. Broadly speaking, we think lower coupons remain compelling 5-7yr Treasury proxies with spread pickup, while higher coupons are a way to pick up carry, fade vol, and arguably shed some exposure to historically tight IG corporate spreads. At the present, current coupon mortgages pick up ZV spread vs. similar duration corporates, but OAS doesn’t look particularly cheap vs. the bond/CDS basis ( Figure 8). Indeed at the local level mortgages have displayed a positive correlation and corporates a negative relation with yield moves given technical demand factors, but this is unlikely to hold in large rate rallies predicated on worsening economic outcomes. Our mortgage analysts compute ‘break-even spreads’ which show that unless IG corporate spreads widen by more than these thresholds, they will likely outperform negatively convex current coupon mortgages in such rate moves ( Figure 9). The question remains that if shorter duration mortgages offer relatively competitive yields vs. corporates, whether recession scenarios are priced well (see Agency MBS).

Figure 8: Current coupon mortgages still look cheap to similar duration corporates on a ZV basis…

JULI 3-5y Treasury spread, FN 30yr cc ZV spread, and difference; bp

Source: J.P. Morgan

Figure 9: …but are likely to underperform IG corporates in large rate rallies unless IG spreads widen materially

Amount of spread widening on 3-5yr IGs to make their price move equivalent to an even duration blend of FN 5s/5.5s in different rate rallies; bp

Source: J.P. Morgan

Mortgage credit spreads continued their tightening trend in the first half of the year but are bouncing back and we believe a weakening economic outlook leaves room for spreads to widen more, with jumbo 2.0 pass-through spreads softening to 1-12 back, nearer to 70 TOAS. Our empirical models favor non-QM over TBA 7s, although investors could be hesitant with very few and largely untested prepayment models for the product. We see non-QM spreads widening back out to 150-175bp on slower demand. Looking ahead, we think strong fundamentals are fully priced in and we expect mortgage credit to be more rangebound. While the top of the capital structure appears to have room to widen, technicals in the form of a limited supply of mortgage credit could keep spreads a bit firmer. Calls and tenders in CRT have most investors looking at the bonds as having a 5yr (or shorter) maturity. We continue to think the extra home appreciation and call/tender potential of ‘21/‘22 vintage CRT B2 spreads make them cheap to new issue collateral which benefits from increased credit enhancement. On the regulatory front, the FHFA approved the Freddie proposal for closed-end second liens, but only allows the agency to purchase $2.5bn over the next 18 months, significantly below initial expectations. On the housing side, we revise our home price forecast higher to 5.5% for 2024, but note risks to the upside as we head into the seasonally strong summer months ( Figure 10). Supply remains constrained and while builders are already pulling back on forward builds, they will likely suffer if rates remain elevated well into 2025 (see RMBS).

We look for the ABS market to finish 2024 with record new issue supply, spreads at the tight end of their two-year ranges and solid structures against minor underlying credit deterioration. We have seen some deterioration in credit quality, with delinquencies rising in auto and credit card debt among lower FICO borrowers indicating some stress for low-end consumers ( Figure 11). That said, ABS is often much better than what sits on bank balance sheets given heavy concentration among prime borrowers with seasoned accounts. Looking ahead, we expect minimal further deterioration if our 4.1% 4Q24 unemployment rate forecast is realized. We expect ABS spreads to reach the tights of their two-year ranges, with our indicative 3-year AAA credit card ABS at Treasury + 40bp and 3-year BBB subprime auto ABS at +135bp. We could see downside risk stemming from heavy supply in 2H24 and investor fatigue. However we note that even after the 1H24 rally, ABS still offers pickup versus comparable corporates and we think AAA benchmark bankcard and prime auto loan ABS should lead this spread narrowing given their history of solid credit fundamentals. Our top pick is AAA 3-year private credit student loan ABS which offers significant spread concession, has a robust capital structure, and reflects high underlying asset collateral (see ABS).

Figure 10: We expect home prices will be up 5.5% in 2024

Historical and projected home price growth for the Case-Shiller HPI Index; %

Source: J.P. Morgan

Figure 11: Auto and credit card delinquencies are rising above pre-COVID levels, while mortgage delinquencies increase at a slower pace

% of 30+ delinquencies by consumer debt category; %

Source: J.P. Morgan, New York FED, Consumer Credit Panel/Equifax

CRE transaction cap rate expansion has been muted YTD with blended cap rates across the major property types sitting roughly 6.5%. Levered return on equity (ignoring NOI growth and exit cap rates) has remained low as borrowing rates have remained high while LTV expansion has been limited, suggesting financing remains less accretive to returns relative to the past and this, in turn, has kept transaction activity low. We think cap rates can expand to 7% driven by seller capitulation if the high for longer rate narrative persists. We no longer think CMBS is trading cheap on a cross-asset perspective. The top of the conduit CMBS stack should largely avoid meaningful extensions over the medium-term as refi success rates are high, but spreads no longer look compelling to single-A corporates. The middle of the stack is more exposed to extension risk and spreads have closed to similar duration REIT bonds. Conduit BBB-s remain optically wide to HY CRE REIT bonds as investors question their relative value given thin enhancements. However there is value to be gained in the CRE space. We like floating SASB AAAs versus BLS CLO AAAs given their attractive spreads, low probability of losses, and sound collateral. We also like FRESB A10Fs versus 10yr Freddie K A2s which suffers from worse liquidity but offers better spread pickup with similar legal hard finals (see CMBS).

Municipals: In the municipal space we continue to favor playing the long game, and buying municipal bonds with longer term perspective, in periods where Treasuries sell off. We believe that bouts of illiquidity should be viewed as an opportunity to add idiosyncratic risk, given that a cycle turn is in the offing. We highlight that prices will be particularly vulnerable in the September/October period given an expected spike in supply as issuers look to avoid election-related market volatility. Both the slowing economy and election supply technicals should drive solid performance by year-end and slightly tighter spreads, as capital continues to chase yield. While the timing remains elusive, this should drive a sustained bull steepening of the curve and lead to inflows across municipal fund complexes, thus driving continued demand for yield, suggesting adding to duration, off-the-run structures, lower BBB & HY (on dips) before the Fed eases and the bull steepening takes hold (see Municipals).

Emerging markets: EM credit has outperformed local markets YTD, as EM sovereign spreads tightened 41bp (when properly adjusted for Venezuela’s index inclusion) and EM corporates tightened 58bp, for total returns of 2.2% and 3.3% respectively. Local markets have been hurt by higher yields (+40bp), while carry has helped limit the total return loss to -2.9% on local bonds in USD terms. Looking ahead, EM does not offer attractive risk/reward at the current juncture. Given a lack of risk-premia in EM assets ahead of the US election, there is an element of complacency. On the positive side of the ledger, we still think EM has good idiosyncratic carry in places where two cycles of distress have abated: an EM default cycle and an EM inflation cycle. Both were precipitated by a generational US hiking cycle and there is still some excess risk premia we think in some high yielding local and hard currency markets to exploit. Harvesting this idiosyncratic carry while remaining hedged against market beta is how we are currently recommending to position within EM. In terms of trading themes, we have OWs in Latam and EMEA EM while UW in Asia. Within EM sovereigns, we stay MW EMBIGD. Turning to EM corporates, we also remain MW given CEMBI outperformance and compressed relative spread. We prefer EM Europe and Africa HY for further spread compression and carry trades in Latam. We prefer non-financials over financials and like commodity segments (see Emerging Markets)

Supply & Demand update

Demand for US fixed income from abroad accelerated in 1H24 after net purchases stepped down late last year. Indeed, foreign investors purchased $400bn of long dated (>1Y) US fixed income securities through April, an increase from the $370bn bought over 2H23 but a reduction from the $630bn purchased in 1H23 ( Figure 12). Demand was once more concentrated in the foreign private sector, which represented 80% of these net purchases, with the foreign official sector contributing to the remainder. Moreover these net purchases were almost exclusively concentrated in Treasuries ($260bn) and corporates ($140bn), with foreign investors net buying only $1bn of Agencies/MBS over the course of the year to date. European investors alone account for over a third of corporate purchases but we think such demand wanes as French and UK election uncertainty combined with increasing hedging costs stemming from the recent ECB cut have helped contribute to EUR credit once more appearing attractive versus USD credit (see HG Credit Foreign Demand Monitor, Nathan Rosenbaum, 6/24/24). Meanwhile demand for Treasuries was more varied to start the year, but was mostly concentrated among investors based in Europe, Japan, and emerging markets (ex-China). We think demand will moderate over 2H24 as continued EM FX reserve drawdowns should restrain foreign official demand whereas we think some of the private demand to start the year was to position for an easing Fed.

Figure 12: Foreign investors were prolific buyers of USD corporate debt in 1H24

Net sales by US residents of long-term (>1Y to maturity) US securities to foreigners by sector, investor type*, and by month, as well as 1H24, 2H23, 2023, and 2022 totals; $bn

Source: Treasury International Capital (TIC) System, J.P. Morgan
*Excludes International Regional Organizations (IROs)
Note: Corporations include corporate ABS. Agencies/MBS is calculated as price-adjusted change in holdings

Figure 13: Bond fund inflows have been robust in 1H24

Monthly* net flows into domestic US bond funds, mutual and ETF; $bn

Source: EPFR, J.P. Morgan
*Data through May

Mutual funds and ETFs continue to remain a strong source of demand for fixed income thus far in 2024 ( Figure 13). As of May, bond funds have accumulated roughly $240bn of inflows (+3.9% of YE23 AUM), compared to $300bn of inflows over the course of 2023 (+4.4% of YE22 AUM), and this represents the second largest 1H intake of inflows in dollar terms and third largest in %-age of AUM terms over the past decade. Inflows have been predominantly skewed towards Agg and IG credit funds. Much of this growth came early in the year as the combination of elevated yields, expectations of near-term Fed easing, and lagging returns from the 4Q23 rally led to a surge in retail inflows. As we look ahead, inflation uncertainty and rangebound yields are likely to keep inflows modest over the summer and will likely only pick up alongside the Fed easing cycle if the economy softens materially. Indeed we have previously shown that more prolonged and deeper easing cycles, which are often associated with worse economic recessions, see a more pronounced increase in the volume and duration of inflows as retail flows follow returns and seek safety (see Overview, U.S. Fixed Income Markets Weekly, 12/15/23).

Turning to supply, we have updated our 2024 cross-sector fixed income net issuance forecast since our last supply update in March ( Figure 14). We now see a total of $3.275tn in fixed income net issuance for 2024 (from $3.05tn previously), which would make for an increase of over $760bn in net supply from 2023. The primary driver behind this upward revision is IG corporate credit, which we revised up by $127bn to $665bn, though supply should moderate from here relative to the rapid pace of 1H24, a seasonal dynamic likely to be accentuated by the US election.  Other products with substantial upward revisions include emerging markets (+$54bn), ABS (+$40bn), and leveraged loan corporates (+$25bn), and we now expect this year’s ex-Treasury net supply to come pretty close to matching that of 2023.

Figure 14: We revise higher our total US fixed income net issuance forecast by $225bn to $3.275tn, driven primarily by an increase in IG corporate supply

Historical and J.P. Morgan forecast for long-term (>1Y) net issuance in 2024; $bn

Source: J.P. Morgan, Dealogic, Bloomberg Finance L.P., Bond Radar, Commercial Mortgage Alert, INTEX, FNMA, FHLMC, FHLB, GNMA, LoanPerformance, US Treasury, Thomson SDC, S&P, SIFMA, Trepp
* Net of Fed purchases
** The CLOs row is not included in the totals to avoid double-counting
Notes: For IG, net issuance figures exclude EM corporates. HY net issuance is calculated as gross issuance – refis, and LL net issuance is calculated as gross issuance – refis – repricings. For CLOs, figures represent US Arbitrage CLOs only. EM corporates includes IG, HY, and quasi-sovereigns and is only net USD issuance. Additionally, the issuance is adjusted for tender/buyback/call activity. EM sovereign figures reflect net USD issuance only, which is calculated using the historical average of 80% of total net issuance. Muni figures reflect taxable munis and muni corporate bonds with municipal industry codes. RMBS credit includes the following sectors: legacy, jumbo 2.0, non-QM, CRT, SFR and RPL/NPL. ABS includes all maturities. Agency debt includes FNMA, FHLMC, and FHLB. Treasuries includes TIPS and represents net issuance to the public.

While most of the changes stem from the spread side of the fixed income universe, we would be remiss to not discuss the evolving outlook for Treasury supply. In recent years Treasuries’ share of annual net US fixed income issuance has ballooned from an average 31% in the five years proceeding the pandemic to a projected 58% in 2024. Indeed we forecast $1.9tn net long-term Treasury supply in 2024, nearly unchanged from our initial estimate in November but roughly $800bn larger than in 2023. Given the increases in auction sizes over the last year, Treasury created approximately $600bn in additional borrowing capacity for FY25 relative to FY24. Accordingly, we see no need for Treasury to make any further changes to auction sizes this year. However, we estimate Treasury will have to address a financing gap of approximately $3.3tn from FY26-FY29, and we estimate auction sizes will begin increasing again in 4Q25 (see Wait till next year: Dissecting Treasury’s financing needs in the coming years, Jay Barry, 5/24/24).

Cross Sector Monitor

Figure 15: Treasury yields rose and credit and mortgage spreads tightened over the week

Current levels, change since 6/14/24, 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
* 6/27/24 levels for iBoxx Euro HG and AA taxable munis; 6/28/24 levels for all others

Figure 16: YTD returns on various fixed income indices; %

Source: J.P. Morgan

• May core PCE at 0.08% should help build some confidence for continued disinflation
• Real consumer spending looks to be downshifting in 1H
• Housing market revival has lost momentum
• We look for 200K payroll gains in June, steady 4% unemployment rate

This week attention was squarely focused on the looming presidential election. For the economy-minded, the outcome of that election will shape the macro backdrop next year and beyond. We provide a brief sketch of the relevant policy issues in an accompanying note. Meanwhile the data flow marches on, and this week’s figures point to only modest growth in current economic activity.

Perhaps more important for markets and Fed-watchers was the very soft reading on the May core PCE price index, the Fed’s preferred inflation gauge. Core PCE prices advanced only 0.08% last month, taking the year-ago reading down to 2.6%. The soft sequential increase—the lowest in over three years—suggests the disinflationary trends that were interrupted in 1Q are getting back on track. We still think this trend will continue, giving the Fed enough comfort to begin lowering the funds rate starting in 4Q.

Economy-wide spending growth appears to have moderated recently. Real consumer spending increased a respectable 0.3% last month, but this came after a modest decline in April and it looks like consumption growth this quarter is running a little shy of a 2% annual rate. This slowing comes after we also learned this week that consumption growth only expanded at a downward-revised 1.5% annual pace in 1Q.

Indicators on the interest-sensitive sectors also came in light this week. Real business investment spending on equipment in 2Q will likely benefit from a bounceback in civilian aircraft outlays (a very volatile category), as well as growth in vehicle sales and imported capital goods. But this week’s data for domestic manufacturers’ May orders and shipments for nondefense capital goods excluding aircraft dipped, and have been moving sideways over much of the past year.

After showing some hints of recovery earlier in the year, most housing figures have turned south again more recently. That trend continued this week, as both new and pending home sales fell materially in May. Inventories of unsold new homes have been creeping higher recently, and we think this will weigh on home building in the second half, something we discuss more in this week’s Focus.

The labor market news this week was mixed. Initial jobless claims eased off a little, although continuing claims grinded up to the highest since late 2021. The labor market will remain front and center in next week’s holiday-shortened calendar. We get the May JOLTS report on Tuesday and the June employment report on Friday. For the latter we expect nonfarm payrolls to rise 200k and the unemployment rate to be unchanged at 4.0%.

Cooler inflation, softer consumption growth

Core PCE price growth for May, at 0.083% to three decimals, was the softest monthly reading since November 2020 (Figure 1). Services excluding housing was also only up 0.1%m/m (3.4%oya), and headline inflation was flat (2.6%oya). These data are good news for the Fed, though FOMC members will want to see more low readings to gain confidence that inflation is moving back to 2%. Over the last three months core inflation rose 2.7% saar, down from a recent peak of 4.5% in March (Figure 2). While consistent with the story that the firm inflation we saw in 1Q was a bump amidst a gradual moderation, the May month-on-month change probably overstates how quickly core price growth will drop. Even so, we think that the FOMC will take enough comfort from this trend to begin lowering the funds rate starting in 4Q.

Real consumer spending rose 0.3%m/m (0.26%) in May, which brought the 2Q run-rate to 1.4%q/q, saar, and we now see a risk that the full 2Q growth could come in a bit below 2%. Growth for the first quarter had originally been reported at 2.5%, then 2.0% after Thursday’s final GDP report that had been revised down to 1.5%. 1H24 consumption growth around 1.5% is a notable downshift from 2H23, when growth ran a touch over 3%. That said, relatively healthy labor markets and wage gains should still provide some support for consumers. The May personal income and outlays data offered some evidence for this, with nominal personal income increasing 0.5%m/m and compensation up 0.6%m/m. With overall prices flat during the month real DPI increased a robust 0.5%m/m, the strongest gain seen within the past year. For the quarter, however, real DPI is increasing at a milder 1.4% annualized pace, similar to the 1.3% gain in 1Q.

Some weakness in core capital goods

The details of the durable goods report were somewhat soft in May. Core capital goods shipments (nondefense ex aircraft) slid 0.5%m/m and are down at a 1.9%q/q, saar pace so far in 2Q, while core capital goods orders fell 0.6%m/m. On a 12-month basis both core capital goods shipments and orders are roughly unchanged (Figure 3). Moreover, regional Fed surveys of capex intentions point to little if any growth in core shipments in coming months, although the relationship between these surveys and the shipments data is loose. Ex-transportation shipments are tracking a 0.7%q/q, saar increase in 2Q, while total shipments are up at a 6.8% pace, reflecting 48%q/q, saar growth in volatile civilian aircraft shipments. Thus we still think overall equipment spending will be a boost for 2Q output growth despite the soft details in this week’s May report. This is also consistent with the strength in capex we have been tracking globally.  

The housing revival, no more

The prospects of high-for-longer interest rates are again weighing on the housing recovery. After some glimmers of hope for the sector earlier in the year, recent readings from the housing market have been weaker. This week both new and pending home sales dropped in May.  New home sales were down 11.3%m/m to a 619k annualized pace, but the blow was cushioned by material upward revisions to the prior three months. The May reading is the lowest since November 2022, but with the upward revisions it would be too early to conclude that new home sales are breaking lower for good. Similarly, the index of pending home sales, which lead existing home sales by one to two months, declined 2.1%m/m (-6.6%oya) to a new all-time low (ex-COVID) in May, signaling ongoing weakness in the existing home market. While a new low, the most recent reading is a just a bit below other recent observations. We think these recent readings imply that residential investment will fall 4-5% during 2H. For more on the single-family side of this, see this week’s Focus.

Calmer claims ahead of solid jobs report

Ahead of a week full of labor market data, initial jobless claims fell for the second consecutive week after heading higher for a number of weeks prior, which may have been partially related to noise around the earlier holidays. Initial claims pulled back to 233k in the week ending June 22, from 239k (previously 238k) in the prior week (Figure 4). On the other hand, continuing claims rose to 1.839m in the week ending June 15, from 1.821m. This is a new post-COVID high and is something to keep an eye on, though the latest reading isn’t that far above values in 4Q23.

The highlight of the holiday-shortened week next week will be the June jobs report. While we continue to look for the labor market to gradually cool over the rest of this year, we look for a firm report next Friday with nonfarm payrolls gaining 200k and the unemployment rate holding steady at 4.0%. Pandemic-distorted services, such as education and health care, as well as government, are likely to again account for the bulk of the job growth. Wage growth should also remain fairly firm, with average hourly earnings expected to rise 0.3% on the month. But this should put year-ago growth in earnings below 4% for the first time since mid-2021, another encouraging sign for the Fed. Earlier next week the May JOLTS report is expected to show job openings slipped below 8 million for the first time since early 2021—a further easing of labor demand and an indication that the labor market is continuing to gradually move into better balance.

Excerpted from, United States Data Watch , Michael Feroli, June 28, 2024

2024 Mid-year Outlook

• We expect Treasury yields to remain rangebound through the summer, with a decline from current levels expected later this year only once the Fed begins easing
• Yields tend to decline ahead of shallow easing cycles, as seen in 1995 and 2019. However, current market pricing and carry implications make it challenging to be structurally long duration. We find the most value in the 3- to 5-year sector
• We forecast a 25bp cut at the November FOMC meeting, followed by a quarterly cadence of cuts thereafter. We project 2-year yields to decline to 4.60% and 10-year yields to sit at 4.40% at YE24
• Synthetic forwards created from equi-notional curve flatteners provide a better carry profile for bullish duration exposure than outright longs
• Given the expected Fed easing and fiscal dynamics, long-end steepeners offer value: we find the most value in 5s/30s which appears too flat after adjusting for market-based Fed policy and inflation expectations, and the Fed’s share of the Treasury market
• The US Presidential and Congressional elections could impact the fiscal outlook and Treasury market. A Trump victory with a Republican sweep could lead to higher yields and steeper curves due to elevated volatility and fiscal concerns, while a Biden victory with a blue sweep could result in modestly lower yields
• We expect nominal coupon auction sizes to remain stable through fall 2025, as Treasury has created ample borrowing capacity for the next year. The risks to this baseline stem from potential changes to Treasury’s debt management strategy, the Fed’s balance sheet reinvestment strategy for MBS paydowns, and the results of the November election
• Demand from foreign investors and US banks has been stronger than anticipated in 1H24, and we revise our forecasts higher for the year. However, asset manager demand has been weaker than expected. Given the transition to more price-sensitive investor demand, dealer inventory in Treasuries have risen to record highs

Market views

Rates market participants entered 2024 with peak optimism: core PCE had risen 1.9% at an annualized rate in 2H23, and the outcome of the December 2023 FOMC meeting was decidedly dovish. Given this backdrop, by mid-January, OIS forwards had priced a full 25bp ease by the March 2024 FOMC meeting and 175bp of easing in 2024. This optimism was relatively short-lived though, as a series of firmer-than-expected inflation prints in 1Q24 derailed the notion of an “immaculate disinflation.” This was accompanied by a series of upside surprises to growth, driving our US growth and core CPI Forecast Revision Indices higher over most of 1H24 ( Figure 17). In response, markets priced in a high for long path for the Fed, and money markets are not pricing a full 25bp ease until the November FOMC meeting ( Figure 18). Given this pendulum shift, Treasury yields rose in 1H24, though have retreated from their local peaks, as the May CPI report assuaged the worst fears that inflation was accelerating, retail sales indicated consumption is slowing, and both JOLTS as well as initial claims gave further sense the labor market continues to slowly rebalance. Net of these factors, 2-, 5-, 10-, and 30-year yields have risen 66bp, 55bp, 48bp, and 43bp, respectively, but are also 30-40bp off their highest levels seen just a month ago.

Figure 17: There were serially upward revisions to growth and core inflation forecasts in 1H24...

J.P. Morgan US Growth FRI (lhs) and US core CPI FRI (rhs); index

Source: J.P. Morgan

Figure 18: ...and the market-implied timing of the first cut has moved back to late-2024

Date of first expected Fed ease as derived from FOMC meeting date OIS forwards;

Source: J.P. Morgan

The rise in yields has certainly been justified by fundamentals: Figure 19 uses our fair-value framework in an attempt to attribute the YTD move in 10-year yields, and shows that rising inflation and Fed policy expectations explain substantially all the 48bp rise in yields in 1H24. First, 5- to 10-year ahead seasonally adjusted TIPS breakevens have risen 26bp YTD, but they are also considerably off their highs in April, and can explain 29bp of the bearish move. Second, 3- to 6-month ahead Fed policy expectations have risen 25bp as markets have pushed back on the timing of the first expected rate cut, contributing 12bp to the rise in yields. In aggregate, our fair-value framework would have implied a 44bp rise in yields over the first six months of the year, very close to the actual increase. It’s notable, however, that 10-year Treasuries continue to appear somewhat cheap to their drivers, and have not traded consistently rich on this basis since the wake of the regional banking crisis in spring 2023, and could perhaps reflect some implicit rise in term premium.

The market’s Fed expectations are likely to remain the dominant driver of Treasury yields over the 2H24. Overall, we continue to forecast a 25bp cut in November, followed by a quarterly cadence of cuts after that, and there is growing comfort the Fed will have room to ease for a number of reasons. First, consumption is slowing: after rising at a 3% annual rate in 2H23, real consumer spending is rising closer to 2% in 1H24. Second, job openings continue to decline sharply in 1H24, falling to their lowest levels since early-2021, giving an indication the labor markets continue to rebalance. Third, we see some indication that growth momentum is slowing in other labor market indicators as well: the 4-week average of initial claims has been steadily rising for the last two months and is now approaching its highest levels since last fall, indicating the labor markets are loosening somewhat ( Figure 20). Though claims have been rising, this does not give us immediate concern about a sharper rise in the unemployment rate, as we do not see a rising trend in layoffs at this point (see Focus: Claims could move up with falling job finding rates, Murat Tasci, 6/21/24).

Figure 20: Initial claims have risen to their highest levels since September, indicating some loss of growth momentum

Initial jobless claims, 4-week moving average; ‘000s

Source: Department of Labor

Figure 21: Yields have declined ahead of other shallow easing cycles in 1995 and 2019...

Cumulative change in 5-year Treasury yields in the months around the first ease in a shallow Fed easing cycle*; bp

Source: J.P. Morgan
*Dates used: 7/6/1995, 7/31/2019, 11/7/2024

Nevertheless, given the inability to identify vulnerabilities that would drive the US economy into a severe recession, it remains likely this easing cycle will be shallower, and perhaps more closely resemble the 1995 and 2019 easing cycles rather than the aggressive easing cycles during the 2001 recession and the GFC in 2007-2008. In that vein, it would seem there is still room for yields to decline, even if the Fed eases in a shallower fashion. Figure 21 tracks the changes in 5-year yields from 6 months prior to the first ease in these “mini” easing cycles to 6 months after, assuming the Fed first lowers rates at the November 7 FOMC meeting. On average, yields tend to decline more than 125bp in the six months prior to the first cut in these more muted easing cycles.

These factors make the case to be long duration as we enter 2H24, but we think market pricing is a limiting factor, as the inversion of money market curves means long duration trades incur negative carry. Indeed, OIS forwards are pricing in 140bp over the next 18 months ( Figure 22). This is roughly in line with our own forecast, and means it will be hard for yields to decline considerably from current levels, unless markets are willing to price in an earlier start to the easing cycle, a series of faster cuts, or a deeper cutting cycle. As a result, even though the onset of the easing cycle appears to be just a few months away, carry implications and outright valuations make it challenging to be structurally long duration. Therefore, we prefer to add duration when yields are near the upper end of their recent ranges, unless we see the data give us some sense the Fed needs to lower rates more aggressively than we expect, and we will be actively watching the trends in initial claims for signs of broader labor market weakness.

Figure 22: ...but money markets are already priced for such an outcome, with OIS forwards pricing 140bp in easing over the next 18 months

Amount of easing implied in 2024, 2025, and 2026 by OIS forwards; bp

Source: J.P. Morgan

Overall, while even shallower easing cycles are supportive of lower Treasury yields, given that money markets are priced for a path of Fed easing that looks similar to our own forecasts, we think it will be challenging for yields to decline further unless we price a faster or deeper easing cycle. As a result, with our own forecast looking for the first cut in November, we look for yields to remain rangebound through the summer, only declining more significantly once the Fed begin to deliver a series of quarterly cuts early in 2025 ( Figure 23). Overall, we look for 2-year yields to decline to 4.60% by YE24, while our forecast projects 10-year yields will sit at 4.40% by 4Q24.

Along the curve, given how little is priced into 2026, we think there is value in being long in the greens, or in the 3- to 5-year sector of the spot curve. Locally, with yields near the bottom end of their recent ranges, it’s tough to be long duration, even though Treasuries appear modestly cheap on a fair value basis. Furthermore, we think the ongoing rapid growth of the Treasury market will continue to outstrip demand from more price-insensitive investors, and this shift in demand will over time lead to higher term premium, which should result in a gentler decline in long-end yields even as the Fed begins easing (see In the eye of the beholder, 9/12/23). Thus, when we add duration we prefer to stay shorter in the curve to insulate against the risks of higher term premium.

Even front-end longs incur negative carry, but as we have discussed for the last few months, there are ways to thoughtfully construct long-duration proxies with better carry profiles. Equi-notional flatteners synthetically replicate exposure to Treasury forward yields and replicate bullish duration exposure with better carry profiles than outright longs (see Treasuries, US Fixed Income Markets Weekly, 3/22/24). By maintaining equal notionals, these flatteners have a smaller bearish exposure in the short-duration leg of the trade when compared to a duration-neutral flattener, resulting in a bullish duration exposure.  Figure 24 displays various synthetic forward Treasuries constructed using equi-notional flatteners, with their carry statistics, the proportion of the variation in these forwards explained by underlying spot Treasury yields, and the beta to changes in spot yields. The table shows that spot Treasuries in the short- to intermediate sector explain a high share of the variation in these equi-notional flatteners, with relatively high betas as well. Along the curve, synthetic 2y3y and 3y2y forward Treasury yields are sitting above their 6-month averages, and exhibit a high correlation with changes in 3- to 7-year Treasury yields and relatively neutral carry profiles. Thus, we would use any bearish move to add synthetic front-end forward exposure via one of these equi-notional flatteners.

Turning to the curve, Fed easing supports a steeper curve, and our work has shown that the long end tends to steepen first in the period between the last hike and the first ease, while the front end reacts only once the Fed begins easing (see Treasuries 2023 Outlook, 11/22/22). For this reason, we have preferred long-end steepeners as a more strategic representation of our view, but these steepeners have had limited success thus far. Our work has also shown that the curve still steepens into shallow easing cycles ( Figure 25). However, the magnitude of steepening varies, as 5s/30s steepened about 75bp prior to the first ease in 1995, then traded in a range in the ensuing months, while the curve steepened just 25bp in the months leading up to the first cut in 2019. This makes sense to an extent: in 1995 the Fed eased modestly into an economy that was benefitting from rising productivity and trend growth rates, supporting a rising neutral policy rate, as policy rates passively became less restrictive over time. Meanwhile in 2019, the Fed preemptively eased in order to better anchor inflation expectations in an environment where core inflation was struggling to rise to 2%.

Figure 25: Similarly, the long end has steepened ahead of shallow easing cycles, though the curve steepened more aggressively in 1995 compared to 2019

Cumulative change in 5s/30s Treasury curve in the months around the first ease in a shallow Fed easing cycle*; bp

Source: J.P. Morgan
*Dates used: 7/6/1995, 7/31/2019, 11/7/2024

Nonetheless, we need to be discerning in choosing the right pair to express this view. Figure 26 shows that the market’s medium-term Fed policy and inflation expectations, as well as the Fed’s ownership share of the Treasury market can explain more than 50% of the variation in most Treasury curve pairs, with the R-squared above 77% for every pair except for 10s/20s, which has had a unique set of drivers in recent years. A few things stand out from this table. First, many curve pairs remain near their flattest levels of the last 3 years, making steepeners attractive if the Fed does indeed begin easing later this year. Second, most front-end pairs look too steep after controlling for these 3 factors, and when combined with exceedingly negative carry profiles, make these steepeners very unattractive. Third, most long-end pairs appear somewhat flat relative to their drivers, and the carry profiles are more reasonable than their front-end brethren. Accordingly, we think long-end steepeners offer value to position for the coming easing cycle, and we like 5s/30s, given it displays relatively high sensitivity to the market’s medium-term Fed policy expectations, and appears 7bp too flat in our fair-value framework.

The 2024 election and risks to the modal view

There are two distinct sets of risks to this modal view. On one hand, if the nascent signs of slowing we observe via consumer spending, demand for labor, and housing markets become more nefarious, this could force the Fed to ease earlier and more aggressively than we have forecast, which would present downside risk to our interest rate forecast. On the other, we think that this fall’s US presidential and congressional elections could also present risks to this modal view, particularly as it pertains to the fiscal outlook.

On that note, CBO’s latest 10-year budget projections forecasted $22tn in deficits over the next decade, an approximately $2.1tn increase from its prior projections. ¹As a result, CBO projects debt held by the public rising from 97% of US GDP to 122% over the next decade, even with the baseline assumption that the individual income tax cuts from the 2017 TCJA expire after 2025. As a result, there’s been an intense amount of focus on how the elections could move the fiscal needle beyond this baseline. As a starting point, it seems there is little room for a major amount of fiscal expansion in a divided government, and it seems this is a consensus view, with the Senate likely to flip to Republican control, while the Democrats take the majority in the House of Representatives. In this case, some extension of the TCJA for low- and middle-income taxpayers seems likely, but in all likelihood this would be funded by higher corporate tax rates, thus mitigating the impact on the deficit in future years. Thus, purely from a fiscal perspective, we do not think a Biden or Trump victory, accompanied by a split Congress, should have major implications on the Treasury market.

However, we found that former President Trump’s relatively unorthodox style of addressing trading and monetary policy via social media played a statistically significant role in elevating implied volatility (see Introducing the Volfefe Index: Quantifying the impact of presidential tweets on rates volatility, 9/6/19). Should he proceed similarly in a second term, elevated volatility could also contribute to higher term premium, as we have found volatility is a key driver in academic measures of term premium (see In the eye of the beholder, 9/12/23). Thus, a Trump presidency, even with a divided government, could result in modestly higher yields and steeper curves, all else equal.

Meanwhile, sweeps in either direction matter for fiscal outcomes. On one hand, if President Biden is reelected with a blue sweep in Congress, his Administration would likely raise the corporate tax rate from 21% to 28% and also raise taxes on high income earners, and this would likely reduce the federal budget deficit by $2.2tn over the next decade. ² On margin, this would be seen as a relief from a financing perspective, but the Tax Foundation also notes that this budget would reduce GDP by 2.2% over the next decade with lower wages and fewer jobs. As a result, we think a blue sweep would likely result in modestly lower Treasury yields.

Conversely, a red sweep could have serious consequences for Treasury yields. Former President Trump has talked about extending the tax cuts from the 2017 TCJA. Already CBO has published on how this could impact the budget, finding extending the 2017 act would add more than $4tn in primary deficits to the debt over the next decade if not accompanied by new revenue.³

On that note, former President Trump has also talked about imposing a 10% tariff on all imported goods, with a 60% tariff for imported Chinese goods. The US imported $3.1tn of good in 2023, with China comprising 16.5% of that total. At face value, this would be worth about $565bn in revenue annually, more than offsetting the loss of revenue from the TCJA extension. However, the Committee for a Responsible Federal Budget (CRFB) estimates that given expected behavioral shifts, this 60% Chinese tariff would raise only $300bn over a decade on the high side, and a $50bn loss of revenue on the low side. If we take the average of these two numbers, it’s only about $125bn in new revenue over the next 10 years. CRFB also estimated the 10% global tariff would bring in $2tn in revenue over a decade, though this was before the statement was made on Chinese tariffs. Accordingly, we would need to haircut this number by about $125bn.

In sum, the net of an extended TCJA would forego over $4tn in revenue over the next decade, with an additional $600bn in interest expense, but this would be offset by approximately $2tn in tariff revenue. Still, the deficit would widen by close to $2.5tn over the next decade, with much of this move occurring in FY26 onward. We think the fiscal concerns that briefly captured market participants’ attention last fall could come back to the forefront, driving the Treasury curve bearishly steeper. There is no perfect analog for this scenario, but 10-year yields rose more than 100bp last fall.  To be fair, even though the Fed last hiked in July 2023, markets were pricing further Fed tightening through the fall, so this rise was definitely not primarily about fiscal concerns. However, the residual of our fair-value model rose more than 70bp over that period as yields rose, indicating that a substantial part of this increase was not driven by shifts in Fed policy, growth, or inflation expectations, and fiscal concerns could be responsible for part of the rise in yields ( Figure 27).

Figure 27: Long-end yields rose 100bp+ in the fall and summer of 2023, with the residual on our fair-value model rising more than 70bp over the period, indicating this move was not just about rising growth, inflation, and policy expectations...

10-year Treasury yields (lhs; %) versus residual from J.P. Morgan 10-year Treasury fair-valued model* (rhs; bp);

Source: J.P. Morgan, Federal Reserve, US Treasury
* 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 = 98.3%, SE = 17.1bp

Figure 28: ...and UK Gilt yields rose sharply following the mini-budget announcement in September 2022

10-year UK Gilt yields (lhs; %) versus 10-year UK swap spreads (rhs; bp), vertical line denotes UK mini-budget announcement on 9/23/22

Source: J.P. Morgan

Away from the US, the UK mini-budget announcement of September 2023 sparked fiscal concerns there, as 10-year Gilt yields rose 125bp in short order, and cheapened 25bp on an asset swap basis ( Figure 28). Certainly, this disorderly move was sparked by unfunded tax cuts in this announcement but was also aided by the deleveraging of the UK LDI community. The Treasury market is tenfold larger than the gilt market, is supported by a wider set of global investors, and does not have a pooled, levered LDI universe like in the UK, so it seems it would be hard for Treasuries to experience a move anywhere near the magnitude of what was seen in the fall of 2022.

Treasury supply in 2H24

As we enter 2H24, the fiscal picture has slipped a bit further: we now project a $1.915tn deficit in FY24, versus our $1.6tn prior forecast ( Figure 29). However, we would note that this overstates the deterioration in public finances: a $145bn increase in projected outlays for student loans stems mostly from revisions that the Administration made to the estimated subsidy costs of previously issued loans and from the Administration’s proposed rule to reduce many borrowers’ student loan balances. Notably, the subsidy costs need to be recognized up front, thanks to the Fair Credit Reform Act of 1990. Adjusting for this accounting, we think this equates for a $1.770tn deficit that needs to be financed in FY24, or about 6.5% share of GDP. Though the fiscal picture appears somewhat worse than it did earlier this year, and will require more marketable borrowing, we do not anticipate this will require Treasury to lift coupon auction sizes further.

Figure 29: The budget deficit is wide for a labor market near full employment...

Annual budget balance by fiscal year*, reported vs. cash flow**; $bn

Source: US Treasury, J.P. Morgan
*FY24 and FY25 are forecasts
**Adjusts for implementation of student loan forgiveness in FY22, reversal in FY23, renewed implementation in FY24

Figure 30: ...but Treasury has created significant borrowing capacity over the last year

US Treasury net coupon borrowing capacity* versus total funding needs**; $bn

Source: US Treasury, Congressional Budget Office, J.P. Morgan
*Assumes current auction scheduled is maintained through FY29 and incorporates borrowing capacity created by SOMA reinvestments. Assumes pace of Treasury QT is cut to $25bn/month in June and to zero in January 2025
**Expected change in public debt, based on CBO’s budget projections and off-balance sheet funding needs

At the May refunding, Treasury commented, “Since August 2023, Treasury has significantly increased issuance sizes for nominal coupon and FRN securities. Treasury believes these cumulative changes leave it well positioned to address potential changes to the fiscal outlook and to the pace and duration of future SOMA redemptions...Based on current projected borrowing needs, Treasury does not anticipate needing to increase nominal coupon or FRN auction sizes for at least the next several quarters.” Indeed, Figure 30 shows Treasury’s expected funding needs stemming from our own budget deficit forecasts for the next two fiscal years, and CBO’s forecasts beyond that, and compares it with the borrowing capacity created via the current coupon auction schedule. Given the increases in auction sizes over the last year, Treasury created approximately $600bn in additional borrowing capacity for FY25 relative to FY24. Accordingly, we see no need for Treasury to make any further changes to auction sizes this year.

However, the figure also shows the picture changes drastically in future years: borrowing capacity begins to decline in FY26 and beyond, as large sized 5-year notes issued during the COVID-19 pandemic begin to mature and need to be refinanced. Assuming QT draws to a close by YE24, Treasury will have to address a financing gap of approximately $3.3tn from FY26-FY29, with the bulk of this gap appearing in FY27 and beyond. Though the longer-term path will necessitate larger coupon auction sizes over time, we do not think Treasury will need to address this gap in 2024. Given that the funding gap in Figure 30 is just $350bn in FY26 and the gap doesn’t widen in earnest until FY27, we think it’s unlikely Treasury will need to increase coupon auction sizes for more than a year, and we assume Treasury keeps auction sizes unchanged until November 2025 (see also Wait till next year Dissecting Treasury’s financing needs in the coming years , 5/24/24). Accordingly, we look for nominal auction sizes to remain stable at current levels for the balance of this year and project further modest increases to 5- and 10-year TIPS auction sizes ( Figure 31).

We now forecast $2.317tn in net-privately held borrowing needs in 2024 and, even with a heavier reliance on T-bills than we previously expected, T-bills should comprise just 18% of total net privately-held issuance in 2024, versus 65% in 2023 ( Figure 32). If we are correct, the T-bill share of debt will remain above 20% through early-2026, even if Treasury begins increasing auction sizes in late-2025 as we have previously discussed ( Figure 33). This would leave the T-bill share above the top end of the 15-20% recommended range identified by TBAC and Treasury, but it’s been broadly understood this is a soft ceiling.  Moreover, there is some indication this optimal range could be revised up. In their letter to Secretary Yellen at the May refunding, TBAC leadership "supported a return to the 15-20% range in the medium term, noting this may happen without further coupon increases. There were some members who felt the T-bill share recommendation could be revisited at a later date, given market developments and continued robust demand in the years since the original recommendation." Reading in between the lines, this indicates there is a risk that the optimal share of T-bills may migrate higher than the current recommended range over time.

Figure 32: We project $2.317tn in net privately held borrowing needs in 2024, with 80% of net issuance coming in coupons...

Annual net privately-held borrowing*, Federal Reserve purchases of Treasuries, and expected change in Treasuries held by private investors; $bn

Year	Net privately-held borrowing		Fed secondary market purchases		Net change in privately-held debt
	Bills	Coupons	Bills	Coupons	Bills	Coupons
CY 2018	387	989	0	0	387	989
CY 2019	77	1133	169	77	-92	1056
CY 2020	2547	1752	157	2184	2390	-432
CY 2021	-1195	2898	0	957	-1195	1942
CY 2022	-37	1638	0	75	-37	1563
CY 2023	2047	1107	0	0	2047	1107
CY 2024	407	1909	0	0	407	1909

Source: J.P. Morgan, US Treasury, Federal Reserve Bank of New York
* 2024 is J.P. Morgan forecast

Figure 33: ...though the T-bill share of debt is likely to remain above 20% into 2026

T-bill share of marketable Treasury debt outstanding*; %

Source: US Treasury, J.P. Morgan
* Assumes issuance forecast in Figure 31, and assumes Treasury begins increasing auction sizes in November 2025, as discussed in Wait till next year, 5/24/24

Demand update

Turning to demand, we’ve been arguing for the last 18 months that Treasury demand is shifting away from price insensitive investors, like the Fed, foreign investors, and US banks, and that more price sensitive investors will comprise a larger share of Treasury market ownership over time. Indeed, this price insensitive share of demand continues to decline, falling to 51% as of 1Q24, the lowest share since 1997 ( Figure 34). Over time, we think the ongoing rapid growth of the Treasury market will continue to outstrip the demand from these investors, resulting in higher term premium.

Figure 34: The ownership of Treasuries continues to shift away from the Fed, banks, and foreign investors...

SOMA, US commercial bank, and foreign combined ownership share of marketable US Treasury debt outstanding; %

Source: Federal Reserve Z.1

Figure 35: ...though foreign demand has remained firm YTD

Annual net foreign purchases of long-term Treasuries by type*; $bn

Source: Treasury International Capital System
* Though April 2024

However, it’s notable that demand from foreign investors and US banks has proven to be stronger than we had anticipated when we published our 2024 Outlook in November: YTD through April, foreign investors have bought $267bn Treasuries, running ahead of last year’s pace ( Figure 35). Notably, foreign official demand has already exceeded last year’s demand, even though our broad USD index has rallied more than 4% YTD, and foreign private demand has remained robust as well. On the private side, it’s notable that nearly 75% of this demand occurred over January and February, only just as the markets began to price in a more hawkish path for the Fed this year. Similarly, bank demand has been somewhat stronger than we had expected, totaling $66bn YTD, compared with our $25bn full-year forecast when we published in November.

Looking ahead, we update our forecasts to reflect some of this reality ( Figure 36). First, we raise our foreign demand forecast from $100bn to $400bn to reflect the trends discussed above. Looking ahead, our colleagues in EM economics continue to forecast a drawdown in EM FX reserves this year, which should restrain official demand in the coming months, even if the Fed does lower rates later this year. Meanwhile, we expect private demand to moderate somewhat versus the pace observed earlier this year, particularly as some of that demand was to position for Fed easing this year. Second, we raise our bank demand forecast from $25bn to $125bn: anecdotally it seems much of this demand has been on spread, and given the recent narrowing in swap spreads, we think there is room for banks to continue to add on this basis in 2H24. However, it’s notable that deposit growth is expected to slow in 2H24, as the organic upward drift we would expect will be somewhat offset by a more rapid pace of reserve declines over the balance of this year (see Interest Rate Derivatives).

Third, we modestly lower our forecast for asset manager demand from $300bn to $250bn. When we published our 2024 Outlook we argued that asset manager demand would be strong, supported by all-in yields near their post-GFC highs, but inflows would likely only pickup when total returns moved into more positive territory closer to the first cut. Since then, we’ve pushed back our first expected 25bp ease into 4Q24, and broad fixed income returns have been flat on the year. Thus, we reduced our expected demand to reflect the shift in these dynamics. Fourth, we significantly raise our dealer projections: as we discussed last week, primary dealer holdings have risen to record levels and, as a share of the overall Treasury market, are approaching levels seen in 2019 (see This one’s optimistic Broad improvements in Treasury market liquidity trends in 2024, 6/21/24). Given, other price sensitive investors have increased their Treasury holdings more slowly, we expect dealer positions will remain elevated for the remainder of this year.

Trade recommendations

• Maintain tactical 5-year duration shorts

-   Stay short 100% risk, or $50mn notional of T 4.5% May-29s
-(US Fixed Income Markets Weekly, 6/14/24: P/L since inception: 10.7bp)
 

• Maintain 100:98 weighted 4.75% Feb 37s / 4.5% Aug 39s steepeners

-   Stay long 100% risk, or $25mn notional of T 4.75% Feb-37s
-   Stay short 98% risk, or $22mn notional of T 4.5% Aug-39s
-(US Fixed Income Markets Weekly, 6/14/24: P/L since inception: 0.3bp)

• Maintain 5s/30s steepeners

-   Stay long 100% risk, or $112mn notional of T 4.875% Oct-28s
-   100% risk, or $29.9mn notional of T 4.75% Nov-53s
- (US Treasury Market Daily, 11/22/23: P/L since inception: -9.4bp)

• 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: -5.2bp)

• The 5s/30s curve 2022-2024 price action looks like a clear base pattern, with key resistance near 40bp. At minimum, we think the curve will challenge that pattern resistance in the second half, and we see a meaningful chance it breaks to the upside by yearend. Within that multi-quarter pattern, the curve has been basing above the 1.5bp Jul 2023 50% retrace after the 2023 and early-2024 rejections of key pattern resistance. We think a break above 16bp-21bp short-term pattern resistance can revive the steepening momentum.
• The 2-year note coils on the key 4.75% chart inflection, which has been a bifurcation zone for the front end since mid-2023. Look for short-term weakness to find material buying interest near 4.85-4.95% support. We expect an eventual extension to lower yields and test of the 4.475-4.50% resistance zone into the fall. We think the market can see a broader retracement of the 2024 bear trend by the end of the year.
• On the low-frequency monthly time frame, the 5-year note yield triggered momentum divergence buy signals late last year and as it reversed from the 5.003% cycle cheap. Historically, those signals have done a good job identifying the end of bear cycles. We expect the market to bullishly break through the Aug 2022 trend line in 2H24, the line that defines the terminal phase of the bear market.
• Look for short-term 10-year note weakness to find buying interest near 4.50% after triggering bearish signals on our short-term pattern recognition system. We expect the developing rally to be less dynamic in the intermediate sector and even more so at the long end, but we still think yields press lower in the second half of the year.
• The 30-year bond has support in the 4.60s that we think will hold if momentum-driven selling pressure continues into July. We are looking for a relatively range bound long end relative to a more dynamic rally in the belly and front end. That said, we see the prospects that the bond can reach resistance parameters in the lower 4-handles by the end of the year.
• 10-year TIPS breakevens rebound after the impulsive tightening to the lower end of the longer-term range. Key support rests at 205bp-215bp. We think the market pressures that support by the end of the year. Short-term, overhead resistance rests at 231-232. Key resistance remains in the 240s.

US

We came into the year expecting retracement to higher yield levels and some flattening of the curve after the fourth quarter steepening took the market to key pattern resistance levels. Those moves extended further than we thought and for longer, but have not derailed the broad strokes of our outlook. The yield curve looks to be trading in a two and a half year base pattern on the heels of a more than decade long flattening trend. Additionally, the low-frequency pattern based signals present on the yield charts that triggered as the calendar rolled to 2024 historically have implied a transition from bear to bull markets. While the mostly range bound price action has delayed the speed at which we thought the bull market would develop, we still view it as part of a medium-term accumulation pattern ahead of a more dynamic rally. Away from Treasury technicals, similar late-cycle signals are now present on equity charts, where we see low-frequency rally deceleration and continued narrowing leadership from already concentrated leadership readings. Additionally, cyclical foreign exchange pairs like USD/JPY have similar pattern-based signaling on the monthly time frame, which in the past also did a good job of identifying late-stage JPY bear markets. The timing of those signals also squares nicely with the expanding month count since the UST curves first inverted. The 3m/5y curve inverted 19 months ago. Historically, the US equity bull market peaked 18 months after initial inversion outside of the two cycles during the 1970s when the market peaked ahead of curve inversion. The subsequent bear markets were all eventually associated with economic contractions. As the demand data softens and inflation data normalizes, markets will have to differentiate a path to soft landing or a precursor to something more dramatic. From our technical perspective, the signaling, pattern development, and cross-market signaling points in the direction of lower yields and steeper curves. The bifurcated data path marks a difference of amplitude for those trends, in our view. Given the setup, we continue to suggest holding core steepening trade themes, and for now tactically trading duration. As the calendar rolls forward and into the Sep-Oct seasonally weak period for risky markets, we will be much more keen to enter bullish duration trades, and won’t hesitate to chase rallies if the labor data starts to soften.

Figure 38: The 5s/30s curve 2022-2024 price action looks like a clear base pattern, with key resistance near 40bp. At minimum, we think the curve will challenge that pattern resistance in the second half, and we see a meaningful chance it breaks to the upside by yearend...

5s/30s curve, weekly closes; bp

Source: J.P. Morgan, CQG

The 5s/30s curve has spent two and a half years basing above the -38bp cycle low and below the +38bp 2021 38.2% retracement level ( Figure 38). The flattening from that key resistance zone extended further than we thought it would, but managed to stay above the 1.5bp Jul 2023 50% retrace and the stop level for our suggested steepening trade that sits just through 0bp ( Figure 39). The Feb-Jun price action now looks like a potential base pattern that has formed above the noted support. A move through the 16bp local range highs and 21bp Mar 20 high could revive the steepening momentum. Other resistance rests at the 24bp Jan 61.8% retrace and 28.5bp early-Feb tactical pattern breakdown. At minimum, we expect the curve to pressure the key resistance levels overhead that include the 41.5bp Apr 2023 high. An upside breakout would seek the 60bp 2021 50% retrace, 82bp 2021 61.8% retrace, 93bp 2010 38.2% retrace, and 118bp 2022-2023 base pattern objective. An extension to those latter targets would likely coincide with a nonlinear break in the labor data. While not a base-case view for our economics team in 2H24, the steepener provides a convenient hedge for that tail risk. On the downside, a move below 0bp would turn our attention to the -7bp Jul 2023 61.8% retrace and then -19bp Sep-Mar pattern objective.

Figure 39: ...Within that multi-quarter pattern, the curve has been basing above the 1.5bp Jul 2023 50% retrace after the 2023 and early-2024 rejections of key pattern resistance. We think a break above 16bp-21bp short-term pattern resistance can revive the steepening momentum.

5s/30s curve, daily bars; bp

Source: J.P. Morgan, CQG

The 2-year note continues to coil right on top of 4.75%, what has been a critical chart inflection for the front end since 2023 ( Figure 40). Over that period, the zone has marked either range resistance or support, which also includes the 200-day moving average since Mar of this year. On a near-term basis, the Apr-Jun rally has stalled out at the 4.65% Apr-May equal swings objective and near the 4.62% Apr channel. A backup through 4.75% could spark a more material mean reversion to higher yields, but even then, we suspect the 4.845% 50-day moving average, 4.90% late-May yield high, and 4.955% Apr channel will cap yields. More significant support is clustered just through 5.00%, levels that held in Apr. We believe the market will extend the recent rally by the end of summer to test the next zone of resistance at the 4.50% Dec-Feb yield pattern cheaps and 4.475% Jan 61.8% retrace. As noted above, the labor data will likely dictate if that marks the initial impulse in a more dynamic trend, or a slow but steady stair-step to lower yields. In any event, the 4.205% Mar 2023 61.8% retrace continues to mark the first material layer of longer-term chart resistance, the same area the market formed a yield bottom in the Dec-Feb period.

Figure 40: The 2-year note coils on the key 4.75% chart inflection, what has been a bifurcation zone for the front end since mid-2023. Look for short-term weakness to find material buying interest near 4.85-4.95% support. We expect an eventual extension to lower yields and test of the 4.475-4.50% resistance zone into the fall. We think the market can see a broader retracement of the 2024 bear trend by the end of the year.

2-year note yield, daily bars; %

Source: J.P. Morgan, CQG

The chart we included for the belly of the curve presents the most compelling evidence to support our bullish duration view ( Figure 41). There, the red arrows on the 5-year note monthly yield chart represent our systematic momentum divergence buy signals. Those signals fired in the fourth quarter of last year and as the market reversed from the 5.003% cycle cheap. Due to the low-frequency nature of the signal at that time scale, we are unable to generate a statistical study, so are left with handful of events to provide evidence supporting our view. That said, each of the signals in recent decades have done a good job identifying the ends of bear market cycles in the Treasury market. A further rally through the two year reversal pattern trend line that now sits at 4.07% and rises to 4.20% by late-summer would add conviction to our bullish bias. Again, the data will determine the speed of the developing bull market. A convincing break through pattern resistance would turn our attention first back to the 3.745-4.00% area, where the market reversed in the Dec-Feb period. Bigger picture, the first longer-term resistance for an emerging bull market rests at the 3.35% Aug 2020 38.2% retrace and then the 3.09% 2018 cycle cheap. To higher yields, a break through the nearby 4.35% moving average confluence would open the door for a backup to the 4.46-4.50% support zone. We suspect that area will roughly cap yields through the summer. Beyond that support, the 4.755% Apr cheap held a cluster of support levels at 4.725-4.76%.

Figure 41: On the low-frequency monthly time frame, the 5-year note yield triggered momentum divergence buy signals late last year and as it reversed from the 5.003% cycle cheap. Historically, those signals have done a good job identifying the end of bear cycles. We expect the market to bullishly break through the Aug 2022 trend line in 2H24, the line that defines the terminal phase of the bear market.

5-year note yield, monthly bars; %

Source: J.P. Morgan, CQG

We suspect that the intermediate sector and even more so the long end will trade more range bound as compared to the belly and front end, but even there we think the pressure will be to lower yields as the year progresses. Like the belly, a 10-year note extension through the nearby 4.355% moving averages would allow for a test of support that sits at 4.465-4.54% ( Figure 42). That includes the 50-day moving average, Apr channel, and early-Jun yield high. Look for material buying interest near that support. We think the 10-year note will eventually press through the 4.15% Dec 61.8% retrace to test next resistance at the 4.02-4.035% late-winter chart inflection and 3.975% Oct-Apr .618 swing objective. We think the intermediate sector can start to get sticky near the 4.00% threshold, especially in the absence of a break in the labor data. To higher yields and beyond the levels already mentioned near 4.50%, next support rests at the 4.64% late-May cheap and then 4.755% Oct 2023 78.6% retrace, which held in Apr.

Figure 42: Look for short-term 10-year note weakness to find buying interest near 4.50% after triggering bearish signals on our short-term pattern recognition system. We expect the developing rally to be less dynamic in the intermediate sector and even more so at the long end, but we still think yields press lower in the second half of the year.

10-year note yield, daily bars; %

Source: J.P. Morgan, CQG

At the very long end, a 30-year bond backup through the 4.50% cluster of moving averages can drive renewed trend-following selling pressure and allow for a retest of the 4.59% internal trend line, 4.59% 50-day moving average, and 4.635% May 61.8% retrace ( Figure 43). We think the market will find footing there and the Apr trend line, that now sits at 4.69%. We are looking for the bond to trade more range bound through the summer, albeit with a slight bullish bias. Key resistance rests at the 4.285% Jan 61.8% retrace. We suspect a push through that level is more likely into the fall and when risky markets enter the seasonal weak period. Even then, look for selling pressure to emerge near the 4.135% Jan 78.6% retrace and 4.08% Oct-Apr .618 swing objective. Alternatively, an unexpected re-acceleration to higher yields through noted channel support would reseek the 4.765% late-May yield high, 4.845% Apr cheap, and 4.90% Oct 78.6% retrace.

Figure 43: The 30-year bond has support in the 4.60s that we think will hold if momentum-driven selling pressure continues into July. We are looking for a relatively range bound long end relative to a more dynamic rally in the belly and front end. That said, we see the prospects that the bond can reach resistance parameters in the lower 4-handles by the end of the year.

30-year bond, daily bars; %

Source: J.P. Morgan, CQG

Figure 44: 10-year TIPS breakevens rebound after the impulsive tightening to the lower end of the longer-term range. Key support rests at 205bp-215bp. We think the market pressures that support by the end of the year. Short-term, overhead resistance rests at 231-232. Key resistance remains in the 240s.

10-year TIPS breakevens, daily closes; bp

Source: J.P. Morgan, CQG

This report was excerpted from Global Fixed Income Technical, Jason Hunter, June 28, 2024

2024 Mid-Year Outlook

• Core inflation is expected to slow gradually over coming months, with the 3-month run rate falling to 2.6% by the end of the year (from 3.3% in May), and the year-ago rate ticking down to 3.2% (from 3.4% in May). We see headline CPI inflation falling from 3.3% to 3.1% oya
• We maintain a bullish bias at the front end of the inflation curve, as we expect inflation to soften somewhat more slowly than the trajectory implied by the inflation swap curve, and election risks are likely to push near-term inflation expectations higher as we approach the fall
• The direction of travel in the fourth quarter is more uncertain, given that markets are likely to be sensitive to the outcome of the US Presidential election...
• ...Our baseline forecast assumes that breakevens revert somewhat narrower, as a further cooling in economic data allows the Fed to deliver its first rate cut and as rising non-OPEC supply and slowing demand drives a softening in energy prices. We forecast 5-, 10-, and 30-year breakevens will end the year at 230bp, 230bp, and 225bp, respectively, though we see risks skewed to the upside
• A Trump victory would raise the likelihood of sizeable additional tariffs. Trump has floated the idea of a 10% universal minimum tariff, as well as 60% tariffs on all imports from China. If both are introduced, this could have roughly a 2.4% impact on the price level
• Supply/demand technicals in the TIPS market remain mixed. Treasury has continued to leave 30-year TIPS auction sizes unchanged while focusing issuance increases in the front and intermediate sectors, where demand has been more robust
• We forecast gross TIPS issuance will total $210bn in 2024, $17bn larger than in 2023, while net TIPS issuance rises to $88bn in real principal terms

Market views

Figure 45: After adjusting for seasonality and near-term carry, breakevens are trading 9-13bp wider YTD

1-month forward seasonally-adjusted TIPS breakevens;bp

Source: J.P. Morgan

Figure 46: Our TIPS index has returned 1.2% YTD, slightly outperforming nominal Treasuries

Quarterly total return on TIPS and TIPS return outperformance*; %

Source: J.P. Morgan
* Defined as (TIPS return)-[(TIPS index duration)/(Treasury index duration)*(Treasury return)]

Over the first three months of the year, core CPI inflation reaccelerated to a 4.5% saar pace from 3.3% in 4Q23, and energy prices rose sharply, driving TIPS breakevens wider and real yields higher.  In the last two months, however, inflation news have been more benign: core CPI inflation has retreated back to a 3.3% 3m saar through May, or 3.4% oya, while headline CPI inflation is running at 3.3% oya. After adjusting for seasonality and near-term carry, benchmark breakevens have widened 9-13bp YTD and are trading near the middle of their 6-month ranges ( Figure 45). Real yields are roughly 30-40bp higher, and TIPS have accrued 1.8% of inflation return. Overall, our TIPS index has returned 1.2% YTD, slightly outperforming nominal Treasuries on a duration-adjusted basis ( Figure 46). As we look ahead, we maintain a bullish bias at the front end of the inflation curve, as we expect inflation to soften somewhat more slowly than the trajectory implied by the inflation swap curve, and election risks are likely to push near-term inflation expectations higher as we approach the fall.

2H24 Inflation outlook

Before turning to our expectations for the evolution of benchmark breakevens and real yields over the balance of the year, we first discuss our latest thoughts on the outlook for inflation over the second half of the year. As mentioned, the recent data suggests the disinflationary process that took hold in 2H23 is getting back on track, though we still see some signs of stickiness on the core services side. Within the details, the most significant reacceleration early in the year came from core services ex-rent inflation ( Figure 47). This component jumped to a 8.2% saar in the first quarter, though we have seen a sharp reversal recently, with this component running just 2.3% saar in the last two months. Where can we expect super core CPI inflation to run going forward?

Figure 47: Core services inflation accelerated over the first half of the year...

Various components of core CPI inflation; % 6m saar

Source: BLS, J.P. Morgan

Figure 48: ...though recent data suggest the disniflationary process is getting back on track

Realized inflation rates across various components of the core CPI basket as of May 2024; units as indicated

Source: BLS, J.P. Morgan

We believe much of the strength in the first quarter was idiosyncratic and unlikely to be repeated, but we are also cautious about extrapolating forward the weak readings of the last two months. Much of the recent swing lower came from travel-related components such as hotels and airfares—we have already seen jet fuel prices begin to rebound in June, and various forecasters are looking for a chart-busting summer of travel. TSA, for example, projects this summer to be the busiest in the agency’s history (see Oil Markets Weekly, Natasha Kaneva, 6/25/24).  The step-down in health insurance CPI inflation in April came with BLS’s semiannual reset in source data, implying this category should run close to 5% saar for the next four months until the next reset. Lastly, we expect motor vehicle insurance should remain firm into the early fall before softening toward a mid single-digit annualized pace in the fourth quarter (see TIPS Strategy, 6/14/24). Away from some of these idiosyncratic factors, we expect core services inflation can gradually soften alongside cooler wage inflation.

Meanwhile, the strength in rent inflation has also been surprising since the start of the year, especially given that industry data, including the Zillow rent indexes for example, have pointed toward softening rent inflation for some time. Yet progress on this front has largely stalled, with these components rising 4.7% and 5.3% saar, respectively, in the three months through May ( Figure 48). As we’ve discussed recently, we believe that the pace of increases for lease renewals continues to run above the pace of increases for new leases, with the Zillow indexes and other market measures only capturing the latter (see TIPS Strategy, 4/26/24). This is particularly important in an environment of low housing mobility, which has kept demand for rental properties elevated and given landlords increased pricing power (listen to more on this topic here). While we see rent and OER CPI softening gradually over the balance of the year, it may be difficult for rent inflation to return to its pre-pandemic pace, and hence we expect these components to prove sticky near 4%.

Lastly, turning to goods, vehicle prices have continued to fall, with used car prices in particular 4.7% lower since the start of the year and running -7.3% saar in the three months through May. A continued steady increase in new vehicle supply has seen inventories rise from ~900k in September 2021 to ~2.7mn more recently, close to levels observed prior to the chip shortage that emerged in 2021. We expect both new and used vehicle prices will continue to decline gradually, though the pace of declines should moderate over the second half. Meanwhile, core goods prices excluding used cars have been roughly stable since the start of the year, running -0.5% saar in the last three months, generally in line with our expectations. Looking ahead, we expect core goods prices (ex vehicles) to remain close to flat. We believe that excess capacity in China played an important role in the goods price deflation experienced late last year, but this impulse is fading: notably, China PPI deflation eased to -1.4% oya in May, compared to -5.4% a year ago. Meanwhile, shipping costs have continued higher again over the last couple of months, amid ongoing crisis in the Red Sea, and rising concerns around the potential introduction of additional tariffs following the US election (see Transportation & Logistics, B. Ossenbeck, 5/24/24). Overall, we forecast core CPI inflation will soften toward a 2.6% 3m saar pace in 4Q24, with the oya rate falling to 3.2%, a slower pace of disinflation than implied by the inflation swap market ( Figure 49).

Combining this with our forecasts for food and energy CPI, headline CPI inflation should fall gradually over the course of the year from 3.3% oya in May to 3.1% by year-end. Meanwhile, core PCE inflation is running 2.6% oya and 2.7% 3m saar in May, and we see core services inflation in the PCE index also cooling gradually alongside CPI. However, base effects will be more challenging later this year as the weak readings in late 2023 fall out of the 12 month window, which should push the year-ago rate of core PCE back up toward 2.8% by year-end, and drive a narrowing in the core CPI-PCE wedge.

Maintain a bullish bias on front-end breakevens

Given this outlook, we think breakevens are likely to remain reasonably well supported over the next few months, despite the richening observed over the last couple of weeks. Notably, breakevens now appear in line with our fair value estimates, even at the front end of the curve, where breakevens had appeared cheap versus this framework for much of the last four months ( Figure 50). Recall that our fair value framework regresses 1-month forward seasonally-adjusted benchmark breakevens on our J.P. Morgan commodity curve index (JPMCCI) as well as the shape of the 3m3m/15m3m OIS curve—a macro risk proxy, with breakevens covarying positively with both factors. Notably, these two factors alone explain 88-93% of the variability in benchmark breakevens over the last 7 years.

As we look ahead, our commodity strategists forecast a 10% appreciation in broad commodity prices by year-end, and they target Brent oil at $90/bbl by September. Meanwhile, the shape of the money market curve is priced for a pace of rate cuts roughly in line with our own forecasts. These factors imply front-end breakevens to lead the way wider in the third quarter, while longer-run forwards remain anchored and the breakeven curve flattens. In addition to these broad macro drivers, US elections are coming into focus, and are likely to remain an important driver of inflation markets in the coming months, as investors consider the potential effects of policy proposals that have been floated by each candidate. Importantly, we think the potential for increased tariffs in the case of a Trump victory likely poses the most significant inflationary risk, which should drive outperformance of the front end of the curve ahead of the election.

The direction of travel in the fourth quarter is more uncertain, given that markets are likely to be sensitive to the election outcome. Our baseline forecast assumes that breakevens revert somewhat narrower, as a further cooling in economic data allows the Fed to deliver its first rate cut and as rising non-OPEC supply and slowing demand drives a softening in energy prices. Five-year breakevens should optically widen as the benchmark rolls to the October maturity. After accounting for this as well as seasonally negative carry later this year, we forecast 5-, 10-, and 30-year breakevens will end the year at 230bp, 230bp, and 225bp, respectively, although we acknowledge risks around this forecast likely skew to the upside. Overall, this forecast reflects an outlook in which breakevens continue to trade in the upper end of the range they have held over the last few months.  Combined with our nominal yield targets, we think real yields will trade rangebound over 2H24, before declining more materially in 1H25 ( Figure 51).

Figure 50: 5- and 10-year breakevens appear in line with our fair value estimates after trading cheap for much of the last four months

Residual on 5- and 10-year TIPS breakeven fair value model*; bp

Source: J.P. Morgan
* Model regresses seasonally-adjusted breakevens (bp) against JPMCCI, JPMCCI-squared, and 3m3m/15m3m OIS curve (bp)
5Y model R2=93%; SE=13bp; 10y model R2=93%; SE=10bp

Figure 51: We maintain a bullish bias on front-end inflation. Look for rangebound real yields in 2H24

Breakevens and real yields, breakeven targets, and real yield and curve levels based on those targets*; units as indicated

	Actual	3Q24	4Q24	1Q25	2Q25
	27-Jun-24	30-Sep-24	31-Dec-24	31-Mar-25	30-Jun-25
Breakevens (bp)
5Y	224	235	230	230	215
10Y	227	230	230	225	220
30Y	229	230	225	225	220
Real yields (%)
5Y	2.07	2.05	1.95	1.70	1.65
10Y	2.02	2.20	2.10	1.95	1.80
30Y	2.14	2.35	2.35	2.25	2.15

Source: J.P. Morgan
* Targets for real yields are based on our nominal yield forecasts and breakeven targets

Given these forecasts, we maintain a near-term bullish bias on breakevens. With tariff risks largely concentrated in the next couple of years, we prefer to hold exposure around the 2-year sector. It is also worth noting that cash breakevens have outperformed versus inflation swaps recently, especially in the front end to intermediate sectors, with IOTAs in these sectors trading at their tightest levels in nearly two years.  Thus, we think inflation swaps are relatively more attractive than cash breakevens for investors looking to add inflation protection in the current environment, and we maintain longs in 1Yx1Y inflation swaps (see TIPS Strategy, 6/14/24).

US elections could be an important driver of inflation markets in 2H24

In considering the potential market reaction to a Trump or Biden victory in November, it is useful, as a starting point, to look back at how the inflation curve responded in the immediate aftermath of the 2016 and 2020 elections. Figure 52 shows that in the week following former President Trump’s victory in 2016, 2Yx3Y and 5Yx5Y inflation swaps rose 16-18bp, while inflation markets were virtually unchanged in the week following President Biden’s victory in 2016. This is a crude exercise as a breakthrough in the COVID vaccine was also announced the day after the election on November 9, 2020, so it’s hard to disentangle the election impact, but consistent with these observations, we believe inflation markets would respond more bullishly to a Trump victory in 2024, with the reaction most pronounced at the front end of the curve — perhaps even shorter on the curve relative to 2016.

Importantly, we think the primary risk for inflation markets is the potential for sizeable additional tariffs. We think the election is most likely to result in a divided government, with the Senate likely to flip to Republican control, while the Democrats take the majority in the House of Representatives. Thus, we see less scope for transformative fiscal policies relative to the last two election cycles. Under either administration, we would expect fiscal policy to turn more restrictive after 2025, when the 2017 tax cuts are due to expire. Meanwhile, the risk of a meaningful increase in trade tensions with China remains elevated, as trade policy is one of the few issues garnering bipartisan support in a polarized US political environment. Importantly, Biden has taken steps to restrict US investments in China and defend against China’s anticompetitive practices. Tariffs inherited from the prior administration have largely remained in place ( Figure 53), and last month, Biden announced new tariffs on Chinese EVs, chips, steel, and other imports. Still, Former President Trump is likely to take a harder line on foreign policy, particularly with China, through a greater use of Executive Orders (see 2024 US Election Watch, J. Chang, et al, 4/25/24). He has floated the idea of a blanket 10% tariff on US imports from all countries, well above the current average tariff rate of 3% on countries excluding China, and has also discussed increasing tariffs to 60% or more on all imports from China. In 2023, the US imported $3.1tn of goods, with roughly 14% coming from China.

Figure 52: Inflation swaps rose sharply in the immediate aftermath of the 2016 US Presidential election

Changes in 1Y1Y, 2Y3Y, 5y5Y, and 10Y20Y inflation swaps in the 1 and 2 weeks following the 2016 and 2020 US Presidential elections; bp

		1Y1Y	2Y3Y	5Y5Y	10Y20Y
2020	1 week	4	0	-1	1
	2 week	-6	-7	-4	-6
2016	1 week	7	16	18	11
	2 week	11	22	24	21

Source: J.P. Morgan

Figure 53: A blanket 10% global tariff and a 60% tariff on all imports from China have been floated, which could have a significant short-term impact on inflation

Effective tariff rate estimates; %

Source: PIIE

Evidence from the Trump 1.0 tariffs suggests there was less than complete pass-through to consumer prices, though there is also some evidence that domestic producers used some of the increase in market power to push through higher prices. We think 0.3% is a reasonable estimate of the price level effect of those tariffs. Additionally, the impact was relatively short-lived, as the tariffs drove substantial redirection of trade. The share of imported goods coming from China has fallen sharply over the last 5 years from above 21% in early 2018, with trade diverting toward other countries, including Mexico, which could have limited the impact on consumer price inflation.⁴ If Trump proceeds with the tariffs that have been floated, our economists estimate a 1.1% increase in the gdp price deflator from a 60% tariff on China and a 1.5% increase in price level coming from the 10% universal tariff (see US: Economic policy implications of the election, Michael Feroli, 6/28/24). If both are introduced, this could result in a 2.4% price level increase. Of course there are other second order effects that also need to be considered—as our FX colleagues discuss in detail, tariff hikes are likely to have bullish implications for the dollar. Specifically, they estimate a 10% blanket tariff could drive 4-6% USD appreciation, all else equal (see US Elections & FX, P. Locke, 11/27/23). Overall, while this channel is unlikely to drive longer-run inflation expectations materially higher, we think upside risk is currently highest at the front end of the inflation curve, around the 2-year sector.

TIPS supply and demand outlook

Consistent with our view for relatively rangebound breakevens, we think supply/demand technicals in the TIPS market remain mixed. Treasury has continued to leave 30-year TIPS auction sizes unchanged while focusing issuance increases in the short and intermediate sectors, reflecting the trends we’ve observed in demand. Figure 54 shows that on average, end-user demand at 5-year TIPS auctions YTD has risen to new all-time highs, while long-end demand has declined. More granularly, auction allotment data reveal that foreign investor demand has increased in the short- and intermediate-sectors, while declining to its lowest level since 2021 at the longend. It’s hard to know where this demand emanated from, since data on foreign holdings of TIPS are released with a substantial lag, but we suspect increased demand for front-end TIPS YTD has largely been driven by foreign private investors.

Recall that in the year through June 2023, foreign official holdings declined for the fourth consecutive year and now only comprise 16% of the total TIPS market, down from 36% a decade ago, alongside a general reduction in the foreign reserve manager footprint within the Treasury market ( Figure 55). Regionally, TIC data indicate that investors domiciled in the United Kingdom were the largest buyer of TIPS last year, while Euro area demand weakened. Looking ahead, we expectdemand from Euro area investors to remain fairly weak, as the relative attractiveness of US CPI inflation markets relative to Euro area HICP in the intermediate and long-end sectors has faded since reaching historically cheap levels late last summer. (see TIPS Strategy, 5/3/24).

Turning to retail demand, outflows from inflation-protected bond funds have continued, though we have seen these outflows moderate in recent months; TIPS-focused funds witnessed $370mn of outflows YTD through May, compared to $12.2bn of outflows by this point last year. Notably outflows have been primarily focused in all-maturity funds, whereas the AUMs of funds benchmarked to short-term TIPS have been relatively more stable ( Figure 56). However, we have found using both an empirical regression-based attribution model as well as security-level holdings data that active TIPS funds have turned further underweight TIPS relative to their benchmark, instead choosing to reach out of index to other fixed income products (see The inflation fund landscape: Reviewing trends in flows, composition, and positioning, 5/28/24). Looking ahead, demand for inflation-protected bonds tends to be momentum driven, and we would not expect inflows into TIPS funds to pickup materially until TIPS returns turn more sustainably positive.

Figure 55: Foreign official holdings of TIPS have declined for their fourth consecutive year, through June 2023

Foreign official share of TIPS and nominal Treasuries outstanding; %

Source: Treasury International Capital System, J.P. Morgan

Figure 56: Since 2021, outflows have primarily come from all-maturity funds, while short-term inflation fund AUMs have been more stable

Cumulative net inflows for short-term inflation funds (benchmarked to 0-5Y or 1-5Y TIPS Index) versus all-maturity and other benchmark funds; $bn

Source: EPFR, J.P. Morgan

Against this backdrop, Treasury has adjusted TIPS issuance only gradually. At the May refunding, Treasury increased the June 5-year TIPS reopening auction size and the July 10-year TIPS new-issue auction size each by $1bn to $21bn and $19bn, respectively. Treasury again stated that it believes “it would be prudent to continue with incremental increases to TIPS auction sizes in order to maintain a stable share of TIPS as a percentage of total marketable debt outstanding.” Thus, we expect another $1bn increase to the October 5-year TIPS auction will be announced next quarter ( Figure 57). This would leave annual gross TIPS issuance at $210bn, $17bn larger than in 2023, while net TIPS issuance rises to $88bn in real principal terms ( Figure 58).

Figure 57: We look for $210bn of gross TIPS issuance over the course of 2024, concentrated in the 5- and 10-year sectors...

Monthly gross TIPS issuance*; Reopenings in grey; $bn

	5y TIPS	10y TIPS	30y TIPS	Subtotal
Jan 24		18
Feb 24			9
Mar 24		16		43
Apr 24	23
May 24		16
Jun 24	21			60
Jul 24		19
Aug 24			8
Sep 24		17		44
Oct 24	24
Nov 24		17
Dec 24	22			63
Total	90	103	17	210
2023	82	94	17	193
Diff.	8	9	0	17

Source: US Treasury, J.P. Morgan
*August 2024 onward represent J.P. Morgan forecasts

Figure 58: ...and forecast $88bn of net TIPS issuance this year in real principal terms

Historical annual net issuance (left axis; $bn of real principal) and TIPS outstanding as a percentage of total Treasury debt outstanding (right axis; %) and our forecasts for 2024

Source: US Treasury, J.P. Morgan

Away from near-term issuance plans, one of the TBAC charge questions asked about the likely costs and benefits of “potential new Treasury products that might assist Treasury in...ways to minimize borrowing costs, better manage its liability profile, enhance market liquidity, and expand the investor base in Treasury securities.” On the short list of products Treasury highlighted as worthy of further exploration was 3-year TIPS. Importantly, adding a new TIPS maturity point could help Treasury maintain its commitment to keep TIPS in the recommended 7-9% share of total outstanding debt, without putting undue strain on increasing the size of existing tenors. Such an addition would be long overdue, as Treasury has maintained 3 TIPS benchmark tenors since the introduction of the product nearly thirty years ago while at the same time expanding the suite of nominal products. Looking ahead, we could see Treasury explore this option further. However, we think Treasury would first prioritize a 1-year SOFR FRN. Moreover, considering the amount of further study and engagement with market participants that would need to take place, with respect to considerations around demand, sizing, and the potential placement of any new benchmark within the TIPS auction schedule, we think the earliest timing of any introduction is likely at least a year away (see TIPS Strategy, 5/3/2024).

Trade recommendations

• Hold 1Y1Y inflation swap longs

-    Stay long 100% risk, or $125mn notional of 1Yx1Y inflation swap (swap start: 6/17/25, swap end: 6/17/26).
 (TIPS Strategy, 6/14/23). P/L since inception: +13.9bp.

2024 Mid-Year Outlook

• The year so far has brought wide trading ranges and a significant repricing of Fed expectations. Forward OIS rates at a YE24 horizon have risen by over 100bp and are now priced to much less easing, even as other central banks like the ECB and BoC have initiated rate cuts. Swap yield curves and even swap spreads have exhibited correspondingly wide trading ranges this year
• Looking ahead, we anticipate a continued steepening in yield curves going into year-end, but much of this is already priced into current forwards, except for curves anchored in longer maturities and forward curves. Thus, outright curve steepening positions premised upon upcoming Fed easing are unlikely to deliver much gratification in coming months
• Fed expectations are also coalescing around a path of moderate rate cuts and tails are thinning, which likely implies reduced jump risk in markets going forward. This, together with the fact that forwards are over-pricing rate cuts in coming months, makes yield curve carry trades attractive as a theme. We continue to advocate for well-hedged carry trades that mitigate directional curve exposure while enhancing carry-to-risk ratios
• Term funding premium will likely continue to drift higher by ~0.5-0.6bp/year over the course of 2H24, in large part because of QT, which we think can continue all year at the now-tapered pace. Growth in bond fund AUMs is a possible offsetting factor, but we see such growth as unlikely to emerge in coming months as it is typically back-loaded in an easing cycle. Rising term funding premium will thus likely pressure swap spread curves flatter
• At the same time, zero-duration spreads are likely biased wider from here, in part due to the upward drift in term funding premium. Taking this into account along with our outlook for term funding premium, we see maturity matched swap spreads as biased significantly wider in the front end of the curve by year-end, modestly wider in the belly and mostly flat at the long end
• The reduction of dispersion in Fed expectations brings reduced jump risk going forward, which implies broadly supportive conditions for short gamma strategies. That said, implied volatility levels are close to fair value and will likely stay stable in 2H24. This means outright short gamma exposure, although likely to be profitable on average, may not offer attractive risk-reward characteristics due to the lack of a convincing downward bias in implieds, and gamma strategies will likely need to be actively managed in coming months
• Alternatively, investors can consider finessing volatility exposure by selling the volatility of individual principal components that characterize yield curve moves. We describe a novel approach to designing such trades, which opens the door to a new class of relative value trades in the options market. In particular, selling the volatility of the second principal component appears likely to offer better returns relative to various metrics of risk

Waiting for someone or something to show you the way

The year so far has been one of wide trading ranges in US Rates markets. Forward OIS rates at a year-end 2024 horizon are sharply higher over the course of the first half, and are only recently exhibiting some stability after trading in a 100bp range in 1Q24. The dollar is ~5% higher, but in relatively steady fashion in part due to greatly lowered easing expectations in the US, even as other central banks such as the ECB and the BoC have delivered their first rate cuts. Even swap spreads have exhibited wide trading ranges, with (for instance) 2-year spreads moving a band that was over 10bp wide in 1Q as well as 2Q. Amidst all this, credit markets have fared quite well, with the JULI index spread tighter as well as more stable (Figure 1).

The main driver of such wide ranges in US rates has been ever-shifting expectations surrounding the path of policy rates. Not only are year-end OIS rates higher, implying diminished easing expectations, but the dispersion surrounding such expectations is also now lower. Forward OIS rates are averages over a market-implied distribution, and the width of that distribution and the fatness of the tails is an important determinant of jump risk in markets. Our preferred way of characterizing this has been to decompose the implied distribution of YE24 OIS rates (inferred from the Dec 2024 SOFR futures options market) into a weighed sum of scenario specific Normal distributions, each of which can be associated with a certain policy outcome (e.g., 3 cuts by year end). Details of this approach can be found elsewhere (see What’s the rush?), but what is noteworthy here is that expectations have steadily been coalescing around what one might call a shallow path - the weights associated with scenarios spanning zero to 3 cuts have risen in total, while the weights on other scenarios has fallen. Most important of all, the weight on a 1-hike-by-year-end scenario, which rose considerably in April and May, is back near zero thanks to soft inflation data as well as Chair Powell's comments signaling that additional hikes were not in anyone's baseline view (Figure 2).

Figure 1: The year so far has been one of wide trading ranges in US Rates markets

YE23, 1Q24, and current levels for selected US rates market variables, equity indices, and credit spreads; units as indicated, current values as of 6/26/2024

Source: J.P. Morgan.

Figure 2: Following softer inflation data as well as Chair Powell’s post-meeting comments, market expectations have coalesced around a moderate easing path

Total weights on year-end policy rate scenarios representing 1 hike, Shallow Cuts*, and Deep Cuts*, as calculated from a decomposition of the implied probability distribution associated with Dec 2024 SOFR futures**; Dec 2023 - Current

Source: J.P. Morgan., CME
* Shallow Cuts correspond to the total weights on the 0 Cut,1 Cut, 2 Cuts, and 3 Cuts scenarios. Deep Cuts correspond to the total weights on the >3 Cuts scenarios. Scenarios outlined in Figure 2 of What’s the rush?.
** 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?

In addition to policy rates, there is also quantitative tightening or Fed balance sheet reduction to consider, which has been ongoing and is likely to continue for the rest of the year. Our projections for the Fed's balance sheet size, as well as the major liability components, are shown in Figure 3. We expect that QT can continue at the now-tapered pace for the remainder of the year before O/N RRP balances will decline to below $300bn, at which point we expect balance sheet runoff to halt. This is because a reasonably sized RRP facility is likely necessary as a buffer against stress in funding markets, given that the marginal propensity of RRP investors to lend in response to spikes in repo rates is higher than the marginal propensity of banks to do the same. Therefore, even though Reserves are projected to remain at over $3Tn by year end, which would have been considered ample a few years ago, the Fed will likely need to halt QT around year end or thereabouts.

Figure 3: We estimate that Quantitative Tightening can continue over the remainder of the year at the now-tapered pace

Current* and projected** values for Fed balance sheet assets, RRP, TGA, Reserves, and Commercial bank deposits*** through 2024, $bn

Source: J.P. Morgan., FRED, Federal Reserve H.4.1, Federal Reserve H.8
* Current as of 6/27/2024 Fed H.4. release
** July runoff accounts for the additional portion of Treasury runoff that will occur at June month end
*** Deposits as of 6/21/2024 Fed H.8. release

What does all this mean for our views on the markets going forward? We discuss this in the sections that follow in greater detail. But to summarize, our themes in the rates markets stem from four underlying observations. First, we do expect the Fed to begin easing in November (followed by additional cuts at a 25bp/quarter pace), but forwards are currently pricing in more cuts than will likely occur in the future. Second, markets will likely remain more stable and exhibit much less jump risk in the months ahead, in comparison with the first half of this year. Taken together, these two observations lead us to favor carry trades in coming months (see Swap Yield Curve). Third, our expectation for continuing QT over 2H24 is a driver of our view that term funding premium will likely be biased higher over 2H24. This, together with our outlook for a widening in zero-duration spreads, causes us to favor swap spread wideners at the front end and in the belly, while leaving us neutral over the medium term in the long end of the curve (see The Outlook for Swap spreads in 2H24). The upcoming US Presidential elections add some upside risk to term funding premium and an additional flattening bias to spread curves. This is because baseline expectations call for divided government (with the Senate likely in Republican control and the House in Democratic hands), but either a Blue or a Red sweep is likely to result in larger deficits (albeit of different magnitudes). While this is likely to become a growing risk, particularly to positions in longer maturity sectors, it is longer term in nature and unlikely to significantly influence swap spreads over the remainder of this year. In the options market, our expectations for benign conditions and low jump risk in the months ahead is supportive of short gamma carry strategies. But given that swaption market implied volatility levels are fairly close to fair value and will likely remain rangebound in 2H24 around our year-end projections, active two-sided management of volatility risk will likely be necessary over the remainder of the year (see Options) . Finally, we discuss a new approach to improving upon the performance of outright short gamma strategies by trading the volatility of individual principal components that describe moves in yields across the curve. While Principal Component Analysis has been used for a variety of purposes in the markets, to our knowledge it has never been used to unbundle the total volatility of yields into volatility of principal components and trade them individually. This widens the relative value opportunity set by opening up a new class of trades; we discuss how to construct such trades, as well as why options market investors should consider such strategies (see Trading Principal Component Volatility).

The Outlook for Swap spreads in 2H24

As we noted earlier, the first half has seen yields trade in a considerably wide range, thanks to dramatic shifts in Fed expectations. At the start of the year, expectations for policy rates over the course of the year were pretty much concentrated in scenarios involving deep cuts (5 or more by year end). As inflation data proved stickier than expected early on, these expectations reversed course, to the extent where a not-insignificant weighting was put on scenarios involving additional hikes in April and May. But a softer inflation print in June, followed by Fed Chair Powell’s comments all helped to mitigate the tail risk of further hikes. More recently, market expectations have coalesced around a moderate easing path that involves somewhere between zero and three cuts by year end (Figure 2). Through it all, the range in yields has been significant, with front end yields trading in a ~100bp range and long end yields trading in a ~75bp range. The range in maturity matched swap spreads has also been significant, at about 10% of the range in yields (Figure 1). But the wide ranges in swap spreads are not because of correlation with yield levels - indeed, the beta between 5-year maturity matched swap spreads and 5-year Treasury yields has been essentially zero over the past three months. Rather, it reflects the swings in macro factors that impact the entire term structure of swap spreads.

Figure 4: Over the course of 1H24, zero-duration spreads and term funding premium have traded in especially wide ranges

Zero duration spreads* (bp, left) and Term funding premium* (bp/year, right), 12/27/2023 - 6/26/2024

Source: J.P. Morgan.
* For details on our methodology for calculating Term funding premium and Zero-duration swap spreads, see Term Funding Premium and the Term Structure of SOFR Swap Spreads

As we described in detail a few months ago, the term structure of swap spreads is best understood by examining the slope, which is the negative of a measure of term funding premium, and the intercept, which we call zero duration swap spreads because it can be interpreted as the swap spread of a near-zero duration UST. Over the course of 1H24, zero-duration spreads have traded in an especially wide range, varying from -15bp at the start of the year to near zero in late April - it now stands near -7bp. Term funding premium, which can be defined as the incremental compensation demanded by Treasury investors for each additional year of modified duration after hedging interest rate risk, has also varied from as low as 3.5bp/year to a high of nearly 5bp/year - this now stands near 4.5bp/year (Figure 4). 

To contemplate the outlook for swap spreads over the remainder of the year, we begin by first projecting the outlook for term funding premium and zero-duration spreads (i.e., the parameters used in our parametric representation of the spread curve). We have discussed our model for these market variables elsewhere in detail (see Term Funding Premium and the Term Structure of SOFR Swap Spreads). But Figure 5 contains a brief statistical description of our model for term funding premium (with recently refreshed coefficients), and shows a factor-based attribution of the change seen in 1H24, as well as our projections for term funding premium over 2H24. We make two observations from this exhibit. First, the 1H24 rise in term funding premium was largely due to a reduction in the Fed's balance sheet, partially offset by the decline in RRP balances. Second, supply side effects were smaller- the increase in UST duration supply accounted for a ~0.2bp/year increase in term funding premium. Looking ahead to 2H24, these are the same factors likely to be at play. QT is expected to continue over the remainder of the year albeit at a tapered pace, and is likely to produce a ~$350bn reduction in the size of the Fed's balance sheet (Figure 3). Even with a small offset from the expected ~$150bn reduction in RRP balances, we see term funding premium rising by ~0.5-0.6bp/year from current levels by year-end.

Figure 5: We look for term funding premium to rise in 2H24, largely on the back of continuing QT

Statistics from regressing* term funding premium (TFP)** versus its drivers (units as indicated), and YE23, current (6/26/2024, except RRP as of 6/19), and YE24 projected values for drivers of term funding premium, factor level attribution of change in 1H24 ***, and 2H24 projections

Source: J.P. Morgan.
* Regression period from Apr 2021 - Jun 2024
** Term Funding Premium is defined in Figure 2 of Term Funding Premium and the Term Structure of SOFR Swap Spreads
*** Factor impact is defined as the change in each driver multiplied by the drivers coefficient.

There is one possible offsetting factor that could emerge over the course of the remainder of the year, and that is a revival of AUM growth at bond funds. Rising AUM at bond funds tends to produce passive demand for USTs across the curve (because these funds typically manage to an index), which therefore results in a decrease in term funding premium. Based on the coefficients shown in Figure 5, we can see that a ~10% increase in AUMs would suffice to fully offset the projected increase in term funding premium stemming from the other drivers. But how likely is such an increase in AUMs? Bond fund AUM growth has been lackluster, in part because of the inverted yield curve, and any pickup in AUM growth will likely be backloaded and only occur when easing is underway. Therefore, in a practical sense, we think much of the next half is likely to be characterized by an upward drift in term funding premium. All else equal, this implies a flattening bias on spread curves going forward.

There is likely additional upside risk to term funding premium from the upcoming US Presidential elections in November. Our baseline projections for the remainder of the year assume current consensus election outcomes, which call for divided government with a Republican Senate and a Democratic House. In such an event, any significant fiscal expansion is unlikely. But a potential Blue or Red sweep would make deficit expansion more likely (see Treasuries). Finally, any market impacts stemming from such fiscal expansion would depend on not just budget deficits but also shifts in the growth outlook. Given the uncertainties involved, and also given that the modal outcome appears to be divided government, we expect election concerns to not materially influence swap spreads over 2H24. That said, this certainly bears watching, especially as the political landscape becomes clearer closer to the election.

What about zero-duration spreads? All else equal, this factor points to a widening bias. A brief summary of our model for zero duration spreads, a factor-level attribution of changes seen in 1H24, and our projected change over 2H24 are all shown in Figure 6. As can be seen, a likely increase in term funding premium as well as our outlook for lower short rates by year end both point to a widening in zero-duration spreads by year end. In particular, we see zero-duration spreads going from their current level of ~minus 6bp to ~plus 2bp by year's end.

Figure 6: Zero-duration spreads are likely biased wider in 2H24, mainly because of an anticipated rise in term funding premium

Statistics from regressing* zero duration swap spreads** versus its drivers (units as indicated), YE23, current (6/26/2024, except RRP as of 6/19), and YE24 projected values for drivers of zero duration swap spreads, factor level attribution of change in 1H24 ***, and 2H24 projected value for zero duration spreads

Source: J.P. Morgan.
* Regression period from Apr 2021 - Jun 2024
** Zero-duration swap spread is defined in Figure 6 of Term Funding Premium and the Term Structure of SOFR Swap Spreads
*** Factor impact is defined as the change in each driver multiplied by the drivers partial beta from the regression in the previous figure.

Mechanically, our widening bias on zero-duration spreads implies wider spreads in all sectors of the curve, all else equal. But of course, these impacts must be netted against the impact of rising term funding premium, which pressures swap spreads narrower and spread curves flatter. The combined effect is likely to be a pronounced widening bias at the front end, more modest widening pressure further out the curve, and a spread curve flattening bias overall.

Before synthesizing all of this into our swap spread forecasts for YE24, it must be said that there are of course additional sector specific drivers of swap spreads in each sector, as we also noted in our recent note on this topic. While those drivers can all be tactically important, they are unlikely to drift significantly over the medium term and are thus unlikely to materially alter our spread outlook. Indeed, swap spread deviations from the term structure baseline in the 2- to 10-year maturity sectors have been steadily converging towards zero, and have mostly been noisy in recent months (Figure 7). Even in the 20-year sector, they have been mostly stable and mean-reverting around levels cheaper than the baseline would imply, but they experienced a brief episode of richening in late May which has since retraced (Figure 8).

Figure 7: Swap spread deviations from the term structure baseline in the 2- to 10-year maturity sectors have been mostly converging towards zero

2Y, 3Y, 5Y, 7Y, and 10Y swap spread deviations* relative to the term structure of swap spreads, 3/20/2024 - 6/26/2024; bp

Source: J.P. Morgan.
* For each tenor, a baseline value is calculated as TFP * -1 * modified duration, plus zero-duration swap spread, where TFP refers to term funding premium as of that date. Deviation is calculated as actual maturity matched swap spread minus the baseline value for that tenor. For more details, see Term Funding Premium and the Term Structure of SOFR Swap Spreads

Figure 8: In the 20-year sector, deviations have been mostly stable and mean-reverting around levels cheaper than the baseline would imply, aside from a brief episode of richening in late May which has since retraced

20Y swap spread deviation* relative to the term structure of swap spreads, 3/20/2024 - 6/26/2024; bp

Source: J.P. Morgan.
* For each tenor, a baseline value is calculated as TFP * -1 * modified duration, plus zero-duration swap spread, where TFP refers to term funding premium as of that date. Deviation is calculated as actual maturity matched swap spread minus the baseline value for that tenor. For more details, see Term Funding Premium and the Term Structure of SOFR Swap Spreads

Thus, we really do not see deviations from the term structure becoming significant drivers of spreads in coming months, and instead we simply assume that recent average deviations will continue to persist in coming months. The one exception is the 30-year sector, where swap spread deviations from the term structure baseline have exhibited a systematic correlation to the Fed's balance sheet size - this is likely because the 30-year sector is even more sensitive to the Fed's balance sheet size than is captured through term funding premium. Here, we expect the Fed's balance sheet reduction over 2H24 to contribute modest additional narrowing pressure on 30-year swap spreads.   Putting all these together, our swap spread forecasts for year-end are shown in Figure 9. As we noted, we expect the widening pressure to be felt mostly in the front end and in the belly of the curve, with 2-year swap spreads ending the year nearly 10bp wider than current levels. We expect the 20-year and 30-year sector to remain mostly flat through year-end, and the spread curve to flatten.

Figure 9: Our forecast for maturity matched swap spreads at various tenors

Current (6/26/2024) maturity matched swap spread and YE24 projection* of maturity matched swap spreads in various sectors; bp

Source: J.P. Morgan.
* Fair value model forecast methodology is detailed in Term Funding Premium and the Term Structure of SOFR Swap Spreads . Projections are calculated using the projections for Term Funding Premium and Zero Duration Spreads as detailed in the above figures and assuming deviation from term structure in each sector is the near-term average

Figure 10: Long term considerations - monthly UST duration supply could increase from its current level of ~230bn 10s to over 300bn 10s in 2028

Monthly UST duration supply, current 2024 levels and quarterly projections through 2028, $bn of 10Y equivalents

Source: J.P. Morgan.

 Lastly, even though the natural horizon for this mid-year outlook piece is the second half of the year, it is worth contemplating the factors that may be at play over an even longer horizon that includes 2025 and beyond, for instance. Over such a long term horizon, one key factor that will be at play is of course Treasury supply. As our UST strategists have noted recently (see Wait till next year, J. Barry, 5/24/2024), current auction sizes are likely insufficient to meeting Treasury's financing needs in the future, and coupon auction sizes will likely need to rise beginning in 2H25 and beyond. Should those future auction size projections come to fruition, monthly UST duration supply will likely increase from its current level of ~$230bn 10s to over $300bn 10s in 2028 (Figure 10). All else equal, this would translate into a term funding premium increase of about 0.65bp/year. In turn, this could narrow 10- and 30-year spreads by 5bp and 11bp, respectively, all else equal. As with the upcoming elections, we see this more as a source of long term risk to spreads rather than as a factor that might play a significant role in 2H24.

Swap Yield Curve

As with yield levels, swap yield curves have traded in fairly wide ranges so far this year despite a Fed that has been on hold. For instance, the 2s/30s and 5s/30s swap curves have traded in 40-50bp ranges this year, while even the 1s/2s curve has traded in a nearly 40bp range (Figure 11). These changes of course reflect the policy uncertainty that has characterized the first half - as we noted earlier, markets began the year pricing in expectations of considerable rate cuts this year, but erased nearly all of that until recently, before softer data more recently has helped policy expectations converge around expectations for moderate cuts (3-4 cuts by year end). 

Figure 11: Swap yield curves have traded in fairly wide ranges in 1H24 despite a Fed that has been on hold all year

Selected statistics for benchmark swap yield curves, 1/2/2024 - 6/26/2024; bp

Source: J.P. Morgan.

Reassuringly, through it all, our empirical fair value model for various spot and forward swap curves has held up nicely in terms of accounting for moves. As readers may recall, we use a common set of five factors to explain spot and forward swap curves. These are the level of the short rates (3Mx3M forward OIS), Fed expectations (the 3Mx3M / 15Mx3M OIS curve), the size of the Fed's balance sheet (to account for QE / QT), short term inflation (2Y inflation swap yield) and long term inflation expectations (5Yx5Y inflation swap yield). Because we estimate the model over relatively long history (10-years), we include a control variable to account for the Fed's policy rate guidance, which it has used as a tool in the past. But as we look ahead over the remainder of the year, we make one minor tweak to this fair value model, to account for the observation that yield curves appear to have steeper equilibrium levels when the Fed is on hold, after accounting for all these factors. In other words, the experience of the past nine months since the last hike suggests that the intercept has likely been shifted up in each sector. To incorporate this, we re-estimate our model over ten years of history, but allowing for an intercept shift after November 2023. Details of this model are shown in Figure 12.

Figure 12: Our empirical fair value model uses a handful of market factors to explain a variety of different spot and forward yield curves …

Statistics from regressing* various spot and forward curves against 7 drivers**, model T-stats, R-squared, Standard error, current yield curve level, and current fair value/residual; current value as of 6/26/2024

Source: J.P. Morgan., Federal Reserve H.4.
* Regression from June 2014 to June 2024
** Underlying drivers are: Forward Guidance (number of years of forward policy commitment by the Fed), 3Mx3M OIS rate (%), Fed expectations curve (15Mx3M - 3Mx3M OIS curve as a proxy), size of the Fed balance sheet ($tn), 5Yx5Y inflation swap yield (%), 2Y inflation swap yield (%), intercept shift (0s before 11/1/2023, 1s after)

Before looking ahead to 2H24, we first use our fair value model to understand the relative impacts of the various factors so far this year. When one examines the moves in each driver together with the coefficients shown above, it becomes clear that the bulk of the move in swap yield curves is really due to the move in short rates and Fed expectations, and the impact of each of the other factors has been fairly minor in comparison. Moreover, the actual change in curves has been pretty much in line with expected moves that one might calculate given the move in the underlying drivers, as seen in Figure 13.

Figure 13: … and such a model has worked well in 1H24, as indicated by the fact that the actual change in numerous curves has been mostly in line with model expectations

Actual change in swap yield curves (Y-axis) versus expected change in swap yield curves curves due to underlying drivers* (X-axis) over the period 1/12/2024 - 6/18/2024; %

Source: J.P. Morgan.
* Fair value model drivers are shown in the above Figure, and expected change is calculated as the change in the underlying drivers times the coefficients from our fair value framework

As we look ahead to the remainder of the year, we begin by projecting the likely YE24 levels of the drivers in our empirical yield curve model. Given our economists’ Fed funds forecasts, we estimate that 3Mx3M OIS rates would likely trade near ~4.75% by year end, or about 40bp lower from current levels. This in turn is consistent with a minus 100bp value for the 3Mx3M / 15Mx3M curve, based on recent empirical relationships between the two. Using these, as well as our inflation strategists' forecasts for short term and long term inflation swap yields, we arrive at our projections for the various spot and forward yield curves, at a YE24 horizon. This is shown in Figure 14, along with the values for these curves as priced into current forwards. From this, we can make several observations. First, all these curves are likely to steepen by year end; the 2s/10s and 5s/30s curves, for instance, are likely to end the year 15-20bp steeper than current levels. But of course, this alone is not enough to argue for steepeners, and we must examine how much of this is already priced into current forwards. In this regard, as Figure 14 also shows, for yield curves anchored in the front end much of the anticipated steepening is already priced into current forwards. As can be seen, over 100% of the anticipated steepening in the 2s/5s and 2s/10s swap yield curves is already priced into spot and near-spot yield curve forwards. This means a flattening bias remains attractive in these sectors for now. This is less true for 2Y forward curves and/or yield curves anchored in the back end. For instance, spot 5s/30s and 10s/30s yield curve steepeners, as well as yield curve steepeners constructed using 2Y forward starting swaps are all likely to be profitable, over the remainder of the year. But given the relatively moderate nature of anticipated easing, and given how much is priced into forwards, the best yield curve strategies in 2H24 will likely involve less outright directional trades and more valuation-dependent and nuanced trade constructions.

Beyond outright macro curve trades, we expect positive carry trades on the yield curve to be profitable going forward. With the Fed likely to remain on hold for several more months and begin easing only at a moderate pace late this year, the environment should remain favorable for yield curve carry trades. Indeed, such positive carry yield curve trades have indeed become profitable since the Fed went on hold last year. One way to see this is to examine a measure we refer to as the "carry beta". This is calculated by assuming the construction of a cross section of swap curve butterfly trades each day that will be unwound in a month, and calculating the beta between the realized P/L on those trades versus the ex-ante carry on those trades. We associate this carry beta with the date on which the hypothetical trades are initiated. As seen in Figure 15, carry trades were already profitable during the later stages of the hiking cycle, but have become more so since the last hike in July of last year. We expect this to remain the case in coming months. Thus, we expect that our yield curve trading strategy in coming months will continue to revolve around positive carry. But, given the earlier discussion about curve trades needing to be less directional and more nuanced, we continue to advocate for the construction of well hedged carry trades that mitigate directional risk while enhancing carry-to-risk ratios. While this sounds repetitive, given the many times in recent months that we have advocated such a theme, it also remains the case that such trades do work, and we view this as the appropriate way to approach yield curve trading in the current regime.

Figure 14: Swap yield curves are likely to steepen in 2H24, but this is fully priced in except for spot and forward curves anchored in the belly and longer tenors

Current and YE24 projected curve levels, projected change in curves*, YE24 forward curve levels, projected YE24 curve level minus YE24 forward curve levels, and % steepening priced into current forwards**, current as of 6/26/2024; %

Source: J.P. Morgan.
* Projected change is calculated as the sum of the projected change in each variable times the partial beta of each variable from our Fair Value model
** Calculated as the difference between the forward and the current swap yield curve divided by the projected change

Figure 15: Carry trades were already profitable during the later stages of the hiking cycle, but have become more so since the last hike in September of last year

Average carry beta* over each of the specified periods in the current hiking cycle, 3/1/2022 - 5/20/2024; unit-less

Source: J.P. Morgan.
* Carry beta calculated for each date, between 01/2003 – 05/2024. Beta defined as the regression beta between P/L on 11 different butterflies initiated on that date, and the carry on the trades as of that date. Trades are assumed to be held for 1M. Butterflies included are 2/3/5, 3/5/7, 5/7/10, 10/20/30, 2/5/10, 3/7/10, 5/10/20, 5/10/30, 7/10/20, 7/20/30, and 10/20/30.

Options

As with swap yield curves, implied volatility itself has traded within a wide range year-to-date, ultimately declining across much of the surface. Short expiry and short tail structures have declined the most, with 6Mx2Y implied volatility lower by ~1.2 bp/day while 2Yx30Y implied volatility is higher by ~0.2bp/day (Figure 16).

Figure 16: Implied volatility has traded within a wide range year-to-date, ultimately declining across much of the surface

Selected statistics for various SOFR swaption structures, 1/2/2024 - 6/26/2024; bp/day

Source: J.P. Morgan.

Looking ahead, there are reasons to be bearish on gamma in the near term. After many months of significant policy uncertainty where the future cone of possible policy paths was wide enough to span deep cuts as well as further hikes, softer data is finally helping market expectations to coalesce around moderate cuts by year end (see Figure 2). This is also broadly in line with our own expectations - we look for the Fed to be patient in the near term, waiting for several months of corroboration in the data before eventually easing. The tightening of implied distributions and the thinning of the tails will likely be helpful in moderating jump risk and delivered volatility in the markets. In other words, 2H24 is shaping up to be favorable for short gamma carry strategies.

Of course, entry levels matter and we need a suitable model to decide whether entry levels are favorable or not. In addition, we also need a way to assess value across the entire vol surface including longer expiries. We have utilized an empirical fair value framework for implied volatility for this purpose. But we now make minor tweaks to our framework to adapt it slightly for use in a Fed-on-hold environment.  

As readers may recall, our fair value framework is a two-step process, where we first start by modeling the percent-yield-vol versus rate relationship assuming  the form yvol = a*exp(-by) + c, where y denotes the ATMF yield level and yvol denotes implied percent yield vol. We then estimate the parameters of this model using a cross-sectional fit, using 10-years of historical data on percent-yield vols on ATMF swaptions with several different tails and expiries. We translate our yield vol backbone to a bp vol backbone, which forms the foundation of our empirical fair value model.

We then model the deviations from this fitted curve, since implied volatility is not determined purely by yield levels. To account for the various other influences on volatility, we use a common set of variables (yield levels, market depth, real rates and monetary policy variables) to model the deviations for each structure. Details of our original model can be found in our Interest Rate Derivatives 2024 Outlook. But we make a few minor changes here, chiefly so that we can better capture the structural down-shift in equilibrium implied vol levels that can come about when the Fed is on hold. The tweaks to our fair value model are as follows. One, we remove the magnitude of Fed expectations as a variable, as it has become less significant. Two, we expand the historical window of our regression fit to go back to June 2018, to capture the behavior in the late stages of that hiking cycle. Three, as noted above, we recognize that a Fed-on-hold regime can bring different vol equilibrium levels and therefore allow for a shift in the intercept after Nov 2023 (which marked the second meeting where the Fed kept rates unchanged). This is very similar to the tweak we described earlier to our empirical model for the yield curve. Our updated empirical model for implied vol differentials from the backbone based on these factors is shown in Figure 17.

Figure 17: Our empirical model for modeling deviations of normal bp volatility from the fitted backbone

Statistics from regressing* deviations of normal bp volatility from reference** volatility against yields (%), natural log of duration weighted market depth*** ($mn), 2Y real rates (2Y swap rates minus 2Y inflation swap) (%), Fed balance sheet size ($tn), and forward guidance (in # of months), current residual (bp/day) as of 6/26/2024

Source: J.P. Morgan., BrokerTec, Federal Reserve H.4.
* Regression from June 2018 - June 2024, excluding March - June 2020
** Reference volatility calculated as yield vol times yields divided by square root of 251. Yield vol is modeled as 24.6 plus 124.9 times exponential of minus 1.1 times yields. The fit is done over 10 years using 3M, 6M, 9M, 1Y, 2Y, 3Y, 5Y, 10Y expiry and 2Y, 3Y, 5Y, 7Y, 10Y and 30Y tails.
*** Market depth is the size of the top 3 bids and offers by queue position, averaged between 8:30 - 10:30am daily. Duration weighted market depth refers to the weighted sum of market depth in 2s, 5s, 10s, and 30s using weights of 0.25, 0.5, 1 and 2, respectively.

Armed with this adapted empirical fair value framework, we can turn to projecting our outlook for implieds across the vol surface. But first we need to project the drivers themselves. We project the ATMF yields in each sector by leveraging empirical relationships between those ATMF yield levels and Treasury rates, and applying that to our UST strategists' rate forecasts. In a similar vein, we additionally use our inflation strategists' forecasts to project real yields in coming months. These forecasts call for roughly unchanged nominal yield levels and slightly higher real yields in coming months. In addition, we assume that the Fed balance sheet will continue to decline throughout 2024, as we noted earlier.

Finally, we also assume residuals will converge by year end, although the fact that current residuals are small means this impact is relatively minor. All in all, we expect implieds to decline rather modestly from current levels by YE24 (Exhibit 18). For instance, we look for 6Mx10Y swaption implied volatility to end the year near 6.4 bp/day which is 0.2 bp/day lower, and 2Yx10Y swaption implied volatility to end the year near 6.4 bp/day, which is modestly lower from current levels. We also look for the expiry curve (defined as 1Yx1Y minus 10Yx1Y implied vol differential) and the tail curve (defined as the 1Yx1Y minus 1Yx10Y implied vol differential) to go towards 1.5bp/day and 0.9bp/day respectively (versus current levels of 1.6bp/day and 1bp/day). These are all modest changes, meaning that much like with the swap yield curve, simple directional long or short vol trades are unlikely to be the best risk-reward trades in the options market in 2H24. Rather, 2-sided trading strategies that actively rely on notions of fair value are likely to prove to be important in the months ahead.

Figure 18: The outlook for implied volatility in 2H24 is one of stability

Current swaption implied volatility and YE24 projected fair value* of implied volatilities for various swaption structures, current as of 6/26; bp/day

Source: J.P. Morgan.
* Projected fair value is calculated using the fair value framework from the previous exhibit and by projecting ATMF yields using 2Y, 5Y, 10Y, and 30Y USTs and our UST Strategists projections, assuming market depth improves modestly, 2Y real rates increase in line with our inflation strategists projections, and that the Fed balance sheet declines to $6.9tn. 100% residual convergence is assumed in the projection for YE24.

Trading Principal Component Volatility

As we noted earlier, we think the backdrop is broadly supportive for short gamma carry strategies, but the fact that implieds are both close-to-fair and likely to be stable means that more nuanced volatility trading strategies are called for. Simply maintaining an outright short gamma bias is likely to produce positive returns on average in our view, but may not be attractive relative to risk, given the lack of a compelling downward bias in implieds.

One way to enhance performance in such an environment is to fine-tune fair value assessments and actively manage volatility positions from the long as well as short side, based on deviations from fair value. Other approaches involve trading volatility of one tail versus another (e.g., 6Mx2Y volatility versus 6Mx10Y) or even broader implied volatility differentials by taking fair-value views on those spreads. Indeed, in our research publications, we routinely strive to find trades of all of these types, as and when opportunities present themselves.

But it is worth asking - is there yet another way that hasn’t yet been fully explored?  In this section, we open the door to an entirely new class of relative value trades, that is based on a decomposition of volatility in any given forward rate into portions that stem from the volatility of each principal component driving rate moves.

Principal component analysis is widely understood and used in financial markets. For instance, it is common to express changes in yield levels across different tenors as the sum of (i) changes stemming from moves in a “level” factor or the first principal component, (ii) additional changes stemming from a “curve” factor or the second principal component, and (iii) so on. This is illustrated in Figure 19. The benefit of such a decomposition is that moves in each factor are designed to be uncorrelated to moves in other factors. PCA can be very useful in understanding the relative movements of highly correlated market variables, in constructing hedge ratios, and more. It is of course widely used for all of these purposes. Also, while PCA is typically computed using historical data, which makes it backward looking, we have described a way to calculate forward-looking or implied principal components by using data from the options market and the yield curve spread options market (see i-PCA: Implied Principal Component Analysis).

Figure 19: Using principal components to decompose one day’s change in yields across the curve - an illustration

An illustration of the decomposition of actual yield changes in benchmark swap yields in the 2-, 5-, 10- and 30-year sectors, as a weighted combination of a level factor (PC1), a curve factor (PC2), and higher order factors (PC3 and PC4).

Source: J.P. Morgan.
* For each principal component, we define its loadings on the 2Y, 5Y, 10Y and 30Y sector to be the impact of a 1-sigma move in that component on 2Y, 5Y, 10Y and 30Y rates respectively. For a more detailed explanation, see i-PCA: Implied Principal Component Analysis (12/6/2022). The weights may be thought of as random variables, whose realized values were 3.4, 1.6, 0.1 and 0.7 on the particular day shown above. In the scaling of principal components shown above, each weight is designed to have an implied standard deviation of 1. The implied PC1 volatility, in this example, would be square root of (6.9 ² + 7.2 ² + 6.4 ² + 5.1 ²). Ex post realized PC1 volatility may be calculated similarly, by using an ex post historical estimation of principal components

PCA has even been used to characterize moves in ATMF implied volatility at various points on the vol surface, and in constructing relative value switch trades between different points on the vol surface. But as far as we are aware, the trading of the volatility of principal components remains an unexplored area. Thus, what we discuss here is different from all of these applications of PCA. Specifically, we address two questions:

• How can an investor trade the volatility of (say) just the first principal component, as opposed to (say) trading the total volatility of 10Y swap yields?
• Why might an options market participant consider such a strategy?

To elaborate upon this a little further here, Figure 19 illustrates the idea that moves in 5Y yields, or moves in 10Y yields, can be expressed as a weighted sum of moves in PC1, PC2, and so on, each of which is uncorrelated to the other. In other words, instead of thinking of yield changes in 5s or 10s as being random variables with their own volatility, we may instead think of the weights w₁, w₂, w₃ and w₄ as random variables with some volatility. (As an aside, empirically only the first and second principal components are usually significant in US Rates markets, and we ignore the third component and beyond for ease of exposition). Thus, as a direct consequence of the decomposition illustrated in Figure 19, it follows that the squared volatility of 10Y rates (or 5Y rates) can be expressed as a weighted sum of the squared volatility of each principal component. In other words, the total volatility of 10Y yields (say) can be unbundled into a portion that comes from the first principal component, and another portion that comes from the second principal component. (Technically, it is the square of the volatility that can be undbundled, but we take liberties with the language here to enhance readability).

The same would of course be true for a different point on the curve - the volatility of 5Y rates also can be similarly unbundled into portions that come from PC1 and PC2, but the relative proportions will likely be different. For instance, 5Y rates appear relatively unimpacted by swings in PC2, given the small loading for PC2 in the 5Y sector. Thus, the volatility of 5Y tails likely stems mostly from the volatility of PC1. This is not true for other points on the curve, such as 2s or 30s, for instance.

Logically, then, we ought to be able to combine a (say) 6Mx10Y straddle (which gives exposure to the volatility of PC1 as well as PC2), versus selling a carefully chosen amount of 6Mx5Y straddles (which mostly carries exposure to the volatility of PC1 given the small loading for PC2 in the 5Y tenor), to isolate exposure to the volatility of the first principal component. More generally, by combining volatility positions on different tails using carefully constructed hedge ratios, we can isolate long or short exposure to the volatility of any desired principal component.

There is one more technical detail to mention here. As one can guess from the above discussion, the calculation of the appropriate hedge ratio to isolate exposure to the volatility of principal components will depend on the estimation of the principal components and their loadings on each tenor. But how should one estimate the principal components? Should one use implied PCA or historical PCA estimates, in which case there is the follow-on question of how much history to use. In our analysis, we use implied principal components. By choosing this, we are effectively constructing trades at entry levels that may be thought of as representing implied principal component volatility, with the P/L on such trades ultimately depending on what the realized principal component volatility is in the future.

Does this work? In other words, are the returns from such a strategy in fact correlated to ex-post realized factor volatility minus the ex-ante implied factor volatility? Figure 20 demonstrates that this is indeed the case. Rolling 3M returns on such a strategy that used 6Mx5Y and 6Mx30Y swaption volatility to isolate long exposure to volatility of the first principal component has indeed closely tracked the differential between realized PC1 volatility and implied PC1 volatility at inception.

But why might an investor want to do this? Ultimately, the hope in trading unbundled components of volatility is that there is an edge to be had in buying / selling principal component volatility vis a vis outright straddles. There is some evidence that this is in fact the case. Figure 21 shows statistics regarding rolling 3M returns from 5 different delta hedged short volatility strategies over the past three years - (i) selling PC1 volatility using 6Mx5Y and 6Mx30Y as the two legs, (ii) selling PC2 volatility using the same two legs, (iii) selling 6Mx5Y straddles, (iv) selling 6Mx10Y swaption straddles, and (iv) selling 6Mx30Y swaption straddles. All returns have been scaled by their respective standard deviations, so as to make them directly comparable across strategies. As can be seen, selling component vol can be a worthwhile endeavor. For instance, a strategy of selling PC2 volatility over the past three years has produced rolling quarterly returns that are (a) better on average, (b) have a smaller 10th percentile loss, and (c) have a better 90th percentile profit. This suggests that there may be a systematic edge to be had in focusing on the volatility of principal components, as one more avenue to generate outperformance in the options markets.

Figure 20: Returns from a strategy designed to isolate long exposure to PC1 volatility are indeed explained by the differential between ex-post PC1 realized volatility and the ex ante implied PC1 volatility

Rolling 3M returns from a long PC1 volatility position constructed using 6Mx5Y and 6Mx30Y swaption straddles* (left, bp/notional), versus the difference between historical and ex-ante implied PC1 volatility (right, bp/day)

Source: J.P. Morgan.
* Weights on 6Mx5Y and 6Mx30Y straddles are chosen to isolate exposure to PC1 while hedging exposure to PC2.

Figure 21: Selling principal component volatility can be a worthwhile endeavor, as return characteristics suggest

Selected statistics* regarding rolling 3M returns from 5 different delta hedged short volatility strategies**, 6/15/2021 - 6/14/2024; bp/notional

Source: J.P. Morgan.
* Values shown have been normalized by dividing by each strategy’s rolling returns by its standard deviation over the past three years, so as to facilitate comparison across strategies. The standard deviation is shown separately for completeness** Strategies designed to isolate exposure to PC1 and PC2 volatility use 6Mx5Y and 6Mx30Y swaption straddles as the two legs. All strategies assume daily delta hedging and monthly rolling of the options.

Trading Recommendations

• Stay short 6Mx10Y swaption straddles on a delta hedged basis, paired with a short position in Greens
  P/L on this trade is currently 1.4abp. For original trade write up, see Fixed Income Markets Weekly 2024-06-14.
• Continue to receive in 3Mx3Y and 3Mx5Y swaps versus paying in 3Yx1Y and 12Mx3M swaps
  P/L on this trade is currently -0.8bp. For original trade write up, see Fixed Income Markets Weekly 2024-06-14.
• Continue to pay in Feb 2037 maturity matched swap spreads versus receiving in USU4 invoice spreads
  P/L on this trade is currently -0.2bp. For original trade write up, see Fixed Income Markets Weekly 2024-06-14.
• Stay long Feb 37s versus selling USU4 Futures
  P/L on this trade is currently 2.9bp. For original trade write up, see Fixed Income Markets Weekly 2024-06-14.
• Maintain 3M forward 10s/30s steepeners (1:1.5 risk weighted) paired with M5/Z5 3M SOFR futures flatteners
  P/L on this trade is currently -5.8bp. For original trade write up, see Fixed Income Markets Weekly 2024-06-07.
• Maintain TU/TY invoice spread curve flatteners (1:0.35 weighted)
  P/L on this trade is currently -4.4bp. For original trade write up, see Fixed Income Markets Weekly 2024-06-07.
• Continue to Pay-fixed in 4.625% Feb ‘26 maturity matched swap spreads
  P/L on this trade is currently -3.2bp. For original trade write up, see Fixed Income Markets Weekly 2024-05-31.
• Maintain 1:0.75 risk weighted 7s/10s maturity matched swap spread curve steepeners
  P/L on this trade is currently -0.8bp. For original trade write up, see Fixed Income Markets Weekly 2024-05-31.
• Continue to Pay-fixed in 4.375% Aug ‘28 maturity matched swap spreads
  P/L on this trade is currently -1.5bp. For original trade write up, see Fixed Income Markets Weekly 2024-05-31.
• Continue to pay-fixed in 4% Feb 2034 maturity matched swap spreads
  P/L on this trade is currently -3.2bp. For original trade write up, see Fixed Income Markets Weekly 2024-05-17.
• Maintain 15Mx3M / 1YX1Y forward swap curve flatteners, paired with 20% of the risk in a long in 18Mx3M and a 24% risk weighted short in 3Mx5Y forward swaps
   P/L on this trade is currently -2.1bp. For original trade write up, see Fixed Income Markets Weekly 2024-05-03.
• Maintain 5s/30s spread curve flatteners
   P/L on this trade is currently -0.2bp. For original trade write up, see Fixed Income Markets Weekly 2024-05-03.
• Continue to overweight 1Yx10Y straddles versus a gamma-neutral amount of 1Yx15Y straddles
   P/L on this trade is currently -1.1abp. For original trade write up, see Fixed Income Markets Weekly 2024-05-03.
• Maintain 3M forward 10s/15s swap curve steepeners paired with 25% risk in 3M forward 3s/7s flatteners
   P/L on this trade is currently 1.3bp. For original trade write up, see Fixed Income Markets Weekly 2024-04-26.
• Continue to pay in 1.875% Feb 2027 maturity matched swap spreads
  P/L on this trade is currently -4.7bp. For original trade write up, see Fixed Income Markets Weekly 2024-04-26.
• Stay long A+100 1Yx5Y payer swaptions versus selling A-100 1Yx5Y receiver swaptions, delta-hedged daily, to position for a correction in skew
  P/L on this trade is currently -3.7abp. For original trade write up, see Fixed Income Markets Weekly 2024-04-19.
   
• Maintain long 65% risk weighted 1Yx10Y swaption volatility versus selling 1Y forward 2Yx10Y swaption volatility, synthetically constructed via suitably weighted 1Yx10Y and 3Yx10Y swaptions
  P/L on this trade is currently -3abp. For original trade write up, see Fixed Income Markets Weekly 2024-04-12.
• 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 -3.9abp. For original trade write up, see Fixed Income Markets Weekly 2024-04-05.
• Maintain 7s/10s swap spread curve steepeners paired with 25% risk in a 7s/10s UST curve steepener
  P/L on this trade is currently -0.2bp. For original trade write up, see Fixed Income Markets Weekly 2024-03-22.
• Maintain conditional exposure to a steeper 10s/20s swap yield curve in a selloff using 9M expiry payer swaptions
  P/L on this trade is currently 1.1bp. For original trade write up, see Fixed Income Markets Weekly 2024-03-15.
• Stay long 1Yx30Y volatility versus 1Y forward 1Yx30Y volatility, synthetically constructed via suitably weighted 2Yx30Y and 1Yx30Y swaptions
  P/L on this trade is currently -2.8abp. For original trade write up, see Fixed Income Markets Weekly 2024-03-15.
• Maintain 5s/10s off-the-run swap spread curve steepeners (100:60 weighted)
  P/L on this trade is currently -1.8bp. For original trade write up, see Fixed Income Markets Weekly 2024-03-08.
• Maintain shorts in 3Yx30Y straddles with less frequent delta hedging
  P/L on this trade is currently -4.5abp. For original trade write up, see Fixed Income Markets Weekly 2024-03-08.
   
• Maintain Z5/U6 SOFR futures flatteners paired with H6/Z6 SOFR futures steepeners (0.85:1 risk weighted)
  P/L on this trade is currently 0bp. For original trade write up, see Fixed Income Markets Weekly 2024-03-01.

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 Wednesday COB levels

• In 2H24, funding conditions will continue to be relatively benign, though we acknowledge there are risks that liquidity investors will have to grapple with in the coming months, such as continuation of QT, Treasury’s funding needs, and MMF reform
• Still, liquidity should remain abundant, particularly as a moderate path of Fed easing is expected, thus continuing to attract cash into MMFs and the front-end. This should limit any impact on EFFR/SOFR, T-bills/OIS, and CP/OIS spreads
• MMF reform is looming, though we continue to believe this should not have any outsized impact on credit spreads in the money markets
• Treasury repo clearing remains a work in progress, though concerns about readiness are emerging. MMFs are further along, while hedge funds face particular challenges given uncertainty as to how the structure of the repo markets will evolve to comply with clearing
• Total money market supply (ex-Fed) should increase by $546bn in 2H24, driven largely by Treasury bills and dealer repo. We see credit supply increasing by $190bn, mainly from credit bonds rolling into the money markets and slightly higher CP/CD net issuance
• Near-term catalysts: May JOLTS (7/2), June FOMC minutes (7/3), June employment (7/5), June CPI (7/11)

Smooth sailing with caution

A key call in our money markets outlook this year is that MMFs will remain an attractive asset class, with overall AUMs staying flat year-over-year. In the event, flows into MMFs have proven even stronger than anticipated, up $182bn YTD to $6.35tn, thanks to stickier inflation and stronger-than-expected economic growth that have led to a significant delay in Fed easing expectations ( Figure 60).

These inflows underscore the abundance of liquidity still in the financial system, most of which seemed to be sitting in the front-end. To that end, concerns that the banking system would be rapidly approaching reserve scarcity in 1H24 ultimately proved unfounded. EFFR was steadfastly stable at 5.33%, while SOFR traded in a 1bp range around 5.31%, outside of quarter-ends and tax-related payment dates. In fact, even when bank reserve balances dropped by over $300bn to $3.27tn in the second half of April, SOFR barely made a blip. At the same time, bank CP/CD spreads narrowed to nearly a post-QT low at the start of the year, only cheapening slightly as liquidity investors turned more defensive heading into March quarter-end and April tax day, before spreads began to tighten again. Furthermore, drainage at the Fed’s ON RRP slowed materially compared to last year, with balances rangebound between $300bn and $500bn, in response to a more moderate pace of net T-bill issuance so far this year ( Figure 61).

Figure 60: Flows into MMFs have proven even stronger than anticipated, up $182bn YTD to $6.35tn, thanks to stickier inflation and stronger-than-expected economic growth that have led to a significant delay in Fed easing expectations

Cumulative YTD change in taxable MMF AUMs ($bn)

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

Figure 61: Drainage at the Fed’s ON RRP slowed materially compared to last year, in response to a more moderate pace of net T-bill issuance this year

T-bill outstandings (lhs, $bn) vs. Fed ON RRP balances (inverted rhs, $bn)

Source: J.P. Morgan

In 2H24, we believe funding conditions will continue to be relatively benign, though acknowledge there are a variety of risks that liquidity investors will have to grapple with in the coming months. Even with QT taper, the Fed’s balance sheet is still shrinking, albeit at a slower pace, which means reserves are still coming closer to approaching scarcity. As of the time of writing, reserve balances registered $3.4tn. Market expectations have largely coalesced around $3.0-3.1tn as the lowest comfortable level of reserves (LCLoR), and while our Fed balance sheet expectations call for reserve balances to remain comfortably above $3tn during 2H24, signs of reserve scarcity could start to emerge towards year-end as we come closer to approaching LCLoR ( Figure 62). The good news is that we also project ON RRP balances to remain near $400bn for at least a few months and over/near $300bn for the remainder of the year, which will help to offset any funding pressures arising from balance sheet normalization.

Treasury’s funding needs also remain a concern given expectations of wider deficits over the coming years. In light of CBO’s latest 10-year projections, our Treasury strategists now expect that any additional financing needs over the balance of 2024 will be met primarily with T-bills. Additionally, there is a risk that Treasury could begin another round of coupon auction size increases sometime in mid-2025, slightly earlier than anticipated (see CBO’s 10-year budget projections: less than meets the eye?, J. Barry, 6/18/24). Upcoming US elections could further widen the deficit if the 2017 Tax Cut and Jobs Act were to be extended in some form (see US: Economic policy implications of the election, M. Feroli, 6/28/24).

Against this backdrop, with Treasury debt outstanding and primary dealer inventories already at record highs, we have already seen markets taking a bit more time to digest Treasury supply, particularly when the supply settles on statement dates. That said, given the abundance of liquidity still in the front-end (indeed, MMFs’ AUMs, dealer repo, and T-bill balances have grown by $1.3tn, $1.0tn, and $1.45tn respectively since YE22, Figure 63) and the potential growth in demand for Treasuries from other structural buyers, this should help alleviate any funding pressures over the near-to-medium term (see Yes, T-bill demand should remain robust, T. Ho, 5/30/24).

Figure 62: Our Fed balance sheet expectations for reserve balances to remain comfortably above $3tn during 2H24

Current* and projected total Fed balance sheet assets, RRP, TGA, Reserves, and Commercial bank deposits** through 2024 ($bn)

Source: J.P. Morgan., FRED, Federal Reserve H.4.1, Federal Reserve H.8
* Current as of 6/27/2024 Fed H.4. release
** Deposits as of 6/21/2024 Fed H.8. release

Figure 63: MMFs’ AUMs, dealer repo, and T-bill balances have grown by $1.3tn, $1tn, and $1.45tn respectively since YE22

MMFs’ holdings of dealer repo and T-bills (lhs, $bn) vs. MMFs’ AUMs (rhs, $bn)

Source: Crane Data, J.P. Morgan
 

With respect to the credit side of the money markets, MMF reform is top of mind, though as we discuss in more detail below, the impact on credit spreads should be minimal, unlike what we saw in 2016. What happens to credit spreads in July will be telling to the extent remaining institutional prime funds (i.e., those that have not announced conversions or liquidations) turn even more defensive to shore up liquidity heading into October. Even if spreads do cheapen however, we suspect other liquidity buyers will jump in to take advantage of the spread backup, thus limiting its widening.

Net, despite some emerging risks in 2H, we see liquidity remaining abundant in the front-end. Based on SOFR future options, market expectations have coalesced around a more moderate path of easing, to the tune of somewhere between zero and three cuts by YE24 (our economists are calling for one 25bp cut in November), which should continue to attract cash into MMFs and the front-end more broadly, all else equal. Thus, this should limit any impact on EFFR/SOFR, T-bills/OIS, and CP/OIS spreads over the coming months.

Regulation: what’s clear, and what’s unclear(ed)

Two regulations loom large among liquidity participants in 2H. First, as part of MMF reform, institutional prime MMFs will have to implement mandatory liquidity fees come this October (see A closer look at MMF reforms: Are liquidity fees swing pricing in disguise?, T. Ho, 8/7/23). To date, and largely as expected, two large internal funds and several smaller fund providers have announced their decision to exit the institutional prime market, with those funds totaling around $240bn in AUMs and comprising ~38% of the institutional prime market or ~17% of the total prime space ( Figure 64). Some are liquidating the institutional prime funds entirely, others are converting the funds to government funds, and one provider is converting to a retail prime fund.

As we have noted previously, we do not think upcoming reforms will have any outsized impact on credit spreads in the money markets (see MMF reform spotlight: Not the same as 2016, T. Ho, 5/3/24). Indeed, among the set of funds that announced either liquidations or conversions, exposure to credit was only about $7bn as of May month-end ( Figure 65). Meanwhile, of the remaining ~$390bn of institutional prime funds, CP/CD exposure totaled ~$200bn, which we estimate comprises only about 10% of the CP/CD market. While a shift towards a more rates-concentrated portfolio to shore up liquidity in the run up to October across the remaining funds could prompt spreads to widen, we note that the buyer base for money market credit is much more diversified now, with MMFs comprising less than 25% of the CP market now versus over 40% at YE15. Along with the vast and growing retail prime MMFs (up $114bn to $775bn in AUMs YTD), other liquidity investors such as LGIPs, regional asset managers, and corporates could easily step in to fill the gap in credit demand resulting from reforms (see One step at a time, T. Ho, 6/14/24). Not to mention, with spread levels remaining tight, any backup will likely be quickly absorbed, thus limiting any sort of spread widening in the near-term.

Figure 64: Two large internal funds and several smaller fund providers have announced their decision to exit the institutional prime market, totaling around $240bn in AUMs

List institutional prime MMFs that will covert to a government fund or liquidate ($bn)

Source: Bloomberg Finance L.P., J.P. Morgan
Note: AUMs are as of May-end

Figure 65: The 40% of institutional prime MMFs that are converting or liquidating held only ~$7bn in credit as of May month-end

Asset allocation breakdown of retail prime MMFs, institutional prime MMFs that are converting or liquidating, and remaining institutional prime MMFs, as of 5/31/24 ($bn)

Source: Crane Data, J.P. Morgan

Meanwhile, the implementation date of the SEC’s Treasury clearing mandate for repos (finalized in December 2023) sits a bit farther out on June 30, 2026, but concerns about readiness are already surfacing.  Recall this mandate will require covered clearing agencies to have direct participants submit all eligible repo and cash trades for central clearing.  With only about 28% of the Treasury repo market currently cleared, this mandate is a big deal, as most of the remaining ~$3.4tn would need to be migrated over to central clearing ( Figure 66).

In many respects, market participants are still trying to figure out the available models through which they will get access to central clearing in time to comply. MMFs are further along given that many already have access to sponsored repo, though there are some key details (e.g., counterparty risk exposure) that still need to be hammered out. Meanwhile, hedge funds face particular challenges in their adoption given the uncertainty as to whether they can engage in portfolio margining or whether they can engage in repo in a “done-away” model, both of which are available in the derivatives market but not in the repo market. If not, this could meaningfully raise the costs of transacting in repo. Considering all that still needs to be addressed operationally in the central clearing market, particularly with regards to access and overall plumbing, two years could be a tight timeline to address all the outstanding issues and uncertainties in the new regulatory environment. In fact, over the last month, both ISDA Chief Executive Scott O’Malia and SIFMA managing director William Thum stressed the brevity of the 2-year timeline. All in all, as the market continues to figure out different access models and the economics of engaging in those various access models, depending on how this transition unfolds, SOFR could potentially be biased upwards down the line (see Get in the clear: FAQs on Treasury repo clearing, T. Ho, 4/25/24).

Figure 66: Based on Fed data, only about 28% of the Treasury repo market is currently cleared

Gross uncleared and cleared Treasury (incl. TIPS) repo volumes ($bn)

Source: Federal Reserve Bank of New York, J.P. Morgan

Money market supply outlook

Total money market supply (ex-Fed) increased by approximately $557bn in 1H24, driven primarily by dealer repo and T-bills ( Figure 67). Bonds less than 1y also witnessed a meaningful increase of $123bn as corporate bonds rolled down the curve, and credit in the form of CP and CDs increased by $62bn YTD. Overall agency supply in the money markets was flat, with increases in discount notes and FRNs offset by decreases in coupon supply thus far this year.

In 2H24, we expect money market Treasuries (ex-SOMA) supply to increase by $385bn, mostly coming from T-bills ($274bn). Currently, the T-bill share of total outstanding Treasury debt sits above the top end of the 15-20% range, and could remain above 20% until late 2026. With Treasury likely keeping nominal coupon auction sizes unchanged over the next several quarters, our Treasury strategists expect that additional financing needs over the balance of the year will be met primarily with T-bill issuance (see Treasuries).

Furthermore, we expect dealer repo balances to increase by $101bn in 2H24 as net supply of Treasuries continues to move higher throughout the remainder of this year, at a time when primary dealer Treasury inventories have grown meaningfully. Indeed, in 1H24, primary dealers’ exposure to Treasuries surpassed their highest levels in terms of total positions ( Figure 68).

With respect to Agencies, net issuance (discos, coupons, and FRNs) in 2H24 should decline by around $130bn. While total year-end supply is unchanged from the previous forecast, the composition of supply will look different in the second half. Notably, FRNs should decrease by ~$120bn and coupons should come up by ~$50bn given the hawkish evolution of monetary policy expectations and its consequences to front-end rates (see Agencies and $-SSAs).

In short-term credit, we forecast year-end total financial CP/CD outstandings to be $1349bn, a slight increase of $37bn in 2H24. Further reserves drainage will likely persist in the second half due to QT, albeit at a slower pace, which could increase bank funding needs. We forecast non-financial CP outstandings to be $320bn, same as our previous year-end target. YTD, net issuance is up by $31bn, mainly from Tier 1 issuers and concentrated in the Health Care sector (see A deep dive into non-financial CP, T. Ho, 6/5/24). We expect total outstandings to be close to unchanged from current levels in 2H24. For the ABCP market, our year-end target is $345bn, mostly flat from YE23 levels.

Figure 68: In 1H24, primary dealers’ exposure to Treasuries surpassed their highest levels in terms of total positions

Primary dealers’ net U.S. Treasury positions ($bn)

Source: Federal Reserve Bank of New York, J.P. Morgan

Excerpted from Short-Term Market Outlook and Strategy, Teresa Ho, June 28, 2024

2024 Mid-Year Outlook

• The first half of 2024 has been relatively uneventful for US Agencies. Both DN spreads to T-bills and FRN DMs have remained within tight ranges and benchmark issuance has stayed very light
• $-SSA spreads remain are near the middle their recent ranges but remain slightly wide to their fundamental drivers. Tighter Bobl/UST spreads are supportive of narrower spreads in 2H24. We maintain our overweight stance on $-SSAs
• Net issuance from the GSEs is slightly up YTD, driven by the FHLBs. However, FHLB net issuance in the first five months of the year was less than half the increase in Fed funds volumes. This is consistent with negative net demand for advances
• We continue to expect negative Agency issuance in 2024, driven by the FHLBs. However, we make changes to our issuance forecasts at the product level: we now look for more floater and less callable supply, as a delayed and more gradual easing cycle impact demand dynamics
• We now look for $93bn of bellwether issuance across the six SSAs we cover. This implies issuance should moderate in the second half of the year

Market views

Figure 69: DN spreads to T-bills remain range-bound…

FHLB discount note/ T-bill spread; bp

Source: J.P. Morgan, FHLB

Figure 70: ...and FRN spreads have also remained within tight ranges

Spreads to ON SOFR on newly-issued FHLB FRNs by maturity; bp

Source: FHLB

The first half of 2024 was decidedly less eventful for US Agency supply and spreads than the same period in 2023. Starting at the front end, DN spreads to T-bills have remained very range-bound despite a pickup in issuance; as we’ve argued this borrowing has been opportunistic given that T-bill issuance has been negative for most of the last 3 months (see Figure 69 and US Agencies and $-SSAs, 5/28/24). Turning to FRNs, a product that has attracted more demand in recent months, valuations have also remained relatively range-bound, with spreads on newly-issued 5-to 7-month FRNs trading between 0.5-2.5bp ( Figure 70). Lastly, benchmark bullet supply has remained very light, with only two FHLB issues YTD, both 3-year maturities, with $1bn in size.

Turning to $-denominated SSAs, spreads are near the middle of their recent ranges ( Figure 71). Further, spreads continue to appear slightly wide to their fundamental drivers. Looking ahead, while we now look for more range-bound Treasury yields and HG spreads, we continue to see tighter Bobl/UST spreads through year-end, which should support tighter $-SSA spreads as well, all else being equal. In aggregate, these factors suggest somewhat narrower $-SSA spreads over the second half of the year and therefore we hold our overweight stance ( Figure 72).

Figure 71: $-SSA spreads are near the middle of recent averages...

5-year average EIB/KfW and IADB/IBRD versus 5-year Treasuries matched-maturity SOFR spread of spreads; bp

Source: J.P. Morgan

Figure 72: …but remain wide relative to their historical drivers

J.P. Morgan 5-year $-SSA spreads to Treasuries fair value model*; bp

	Current value	Coeff.	T-stat.
Intercept		-5.12	-9.9
JULI spread; bp	109	0.12	50.4
5-year UST yield; %	4.30	1.85	31.0
5y Bobl/ UST spread; bp	-186	0.04	17.6
Current residual	3
R-squared; %	73.2%
Standard Error; bp	3

Source: J.P. Morgan*Regression of average of 5-year EIB, KfW, IBRD, and IADB spreads to Treasuries over the last 5 years; bp

Agency supply

Net issuance from the three GSEs we cover has totaled $6bn in the first five months of the year, on the back of $18bn in net issuance from the FHLBs and $9bn and $3bn in paydowns from Fannie Mae and Freddie Mac, respectively. More interestingly, the composition of FHLB debt outstanding has shifted away from callables and into floaters, with also a small uptick in discos outstanding ( Figure 73). Importantly, while the increase in debt outstanding YTD has been somewhat surprising, we’d note that Fed fund volumes rose $40bn in the first 5 months of the year, more than double the increase in debt outstanding during the period. As a reminder, the FHLBs are the primary lender in this market; hence, this is consistent with slightly negative net advance demand YTD, which is more in line with our expectations ( Figure 74).

Figure 73: FHLBs net issuance of FRNs has been positive YTD, at the expense of callables...

YTD* FHLB net issuance by product; $bn

Source: FHLB
*through May 31, 2024

Figure 74: …But Fed funds volumes have risen by more than twice as much in the same period

Quarterly FHLB net issuance versus changes in Fed funds volumes; $bn

Source: FHLB, Federal Reserve
*2Q24 is through May 31, 2024
 

Looking ahead, we continue to think our overall Agency supply forecasts are appropriate. In particular, we expect -$100bn in FHLB net supply this year as banks reduce their reliance on advances as part of their liquidity plans, deposit growth remains weak, and home price growth remains modestly positive (2024 RMBS Credit Midyear Outlook, 6/28/24, John Sim). However, we now expect a different composition of supply, given the hawkish evolution of monetary policy expectations and its consequences for demand by product type.

Starting with bullets, net issuance has been very positive YTD, primarily in the form of floaters. Indeed, as Figure 75 shows, after dropping at the end of 2023 when markets priced in an aggressively dovish path for the Fed this year, the FRN share of total FHLB debt rose to 31.3% at the end of May the highest since the end of 2020. This evolution shouldn’t be surprising: with the timing of the first Fed ease getting delayed following upside inflation surprises earlier this year, it makes sense for front-end investors to continue to favor floating-rate instruments. Indeed, as our colleagues in Short Term Fixed Income research discussed earlier this month, MMFs holdings of Agency floaters have risen $103bn YTD to $477bn at the end of May (see Short-Term Fixed Income, U.S. Fixed Income Markets Weekly, 6/14/24). As we look ahead, we continue to think this community will gradually extend out of floaters as we approach the first Fed ease later this year, which should drive more negative FRN issuance through the remainder of the year.

Figure 75: Demand for floaters has recovered and they now make up a larger share of FHLB debt outstanding than callables

FRN and callable shares of total FHLB debt outstanding; %

Source: FHLB

Figure 76: Our revised issuance forecasts now look for less callable and more bullet supply over 2024

J.P. Morgan projected 2024 net debt issuance by issuer and product type; $bn

	Discos	Bullets	Callables	Long-term	Total
FNMA	0	-15	5	-10	-10
FHLMC	0	-10	0	-10	-10
FHLB	-25	25	-100	-75	-100
Total	-25	0	-95	-95	-120

Source: J.P. Morgan

Turning to callables, issuance has been negative YTD and we only see room for modest recovery from here. As we’ve discussed in the past, a more inverted 2s/5s curve reduces the attractiveness of callables, all else being equal. For reference, we came into the year expecting the Fed would start easing policy this summer, with a pace of 25bp per meeting until Fed funds dropped to 3.5%, allowing the 2s/5s curve to steepen to 5bp by year-end (see Treasuries 2024 Outlook, 11/21/23). However, after hawkish revisions to our Fed call, with the first cut in November and a gradual pace expected thereafter, we now expect this curve to stay inverted, at -35bp by YE24 (see Treasuries, U.S. Fixed Income Markets Weekly, 6/7/24). As a reminder, demand for callables traditionally rises when the yield curve is steep, as high call probabilities give investors comfort selling volatility to enhance returns. Further, implied volatility has come down, alongside greater policy clarity and a reduction in tail risks, making callables somewhat less attractive. Putting these pieces together, Figure 76 shows our updated forecasts for net debt issuance in 2024 by issuer and product, showing less callable and more bullet supply relative to our prior forecasts.

$-SSA supply

Last month, we adjusted our $-SSA supply forecasts, by raising EIB issuance by $4bn to $22bn. Following this revision, we now look for $93bn of bellwether issuance across the six agencies we cover. These forecasts suggest supply should moderate in the second half of the year ( Figure 77). Importantly, we believe our full-year issuance forecasts remain appropriate, given both the guidance from the individual agencies, as well as the traditional seasonality of supply over the course of the year ( Figure 78).

Figure 77: We look for $93bn in gross bellwether $-SSA supply this year…

KfW, EIB, IBRD, IFC, IADB, and ADB gross USD-denominated bellwether issuance as of 6/28/2024; $bn

Year	KfW	EIB	IBRD	IFC	IADB	ADB	Total
2015	30	26	21	2	10	11	101
2016	36	25	35	3	10	11	119
2017	29	18	15	1	12	19	94
2018	24	22	9	4	11	10	79
2019	23	14	15	2	8	12	74
2020	17	22	42	4	17	24	126
2021	24	19	24	2	15	20	103
2022	17	15	11	2	9	20	73
2023	26	18	20	4	11	17	95
2024 YTD	16	18	11	2	4	12	63
2024 E	23	22	17	3	11	17	93
% of 2024E	70%	82%	65%	67%	36%	71%	68%

Source: KfW, EIB, IBRD, IFC, IADB, ADB, J.P. Morgan

Figure 78: …With the pace of issuance likely to moderate in the second half of the year

Gross $-denominated SSA bellwether issuance by month, 5-year average vs YTD*; $bn

Source: KfW, EIB, ADB, IBRD, IFC, IADB, J.P. Morgan*2024 is through June 28, 2024

Slow and Steady

• We head into the second half of the year with mortgage OASs looking a bit snug, ZVs still historically wide, and the supply/demand picture looking manageable, though perhaps not overly supportive of the basis
• Year-to-date, higher coupons have generally outperformed Treasuries, thanks to a combination of carry and spread tightening; discounts have lagged on the back of technical pressures from bank portfolio reallocations
• If our house rate forecast of a single cut this year is realized, higher coupons should again outperform in the second half of the year
• Our supply forecast is unchanged for Fed runoff and net issuance, and on the demand side we’d slightly upgrade our expectation for bank demand while acknowledging that foreign buying has underwhelmed; money managers will still be the linchpin, and a combination of inflows to active managers and a continued overweight to MBS should be supportive
• Locally, mortgage and corporate spreads have exhibited strong (and opposite) correlation with rate moves, but it’s not clear those relationships would persist if we were to break out of the range
• We explore recession scenario analysis and the break-even spread widening that would make IG price changes equivalent to those on similar duration (but negatively convex) MBS into sharp rallies
• We enumerate many of the reasons proposed to explain flat conventional s-curves (points paid, low media spend, fewer appraisal waivers, originator layoffs, FICO inflation, etc.) but caution that most would be less binding into a larger rally
• Discount prints have now been very well predicted by seasonal patterns and daycount for the better part of a year, giving us (and the low coupon IO investor) more confidence in the accuracy of model valuations
• Ginnie peaks are the sole point of true excitement in prepay space; interestingly, after the “mesa” profile from January speeds reverted to the typical peak/post-peak relationship, post-peaks have somewhat underwhelmed (relative to peaks) vs. historicals, perhaps indicating that even Ginnie borrowers are harder to motivate in the current environment
• We review relative value in the FN 30 spec space; lower coupon loan balance and seasoned pools stand out with sizable OAS and ZV pickups from the improved carry profile. Lower payup stories like TX and FL provide some OAS pickup at minimal cost to ZV across the stack, while low FICO pools in the belly and high LTV pools in higher coupons provide a similar effect

Views

• Continue to trade the range with the Fed on hold and the technical backdrop mostly unchanged; mortgages are a bit snug, but not terribly so
• In specs, prefer seasoning in lower coupons and lower payup stories up the stack; it’s debatable whether the duration of high payup premiums will be as stable as just moving down to the belly

We head into the second half of the year with mortgage OASs looking a bit snug, ZVs still historically wide, and the supply/demand picture looking manageable, though perhaps not overly supportive of the basis. Year-to-date, higher coupons have generally outperformed Treasuries, thanks to a combination of carry and spread tightening; discounts have lagged on the back of technical pressures from bank portfolio reallocations ( Figure 79). If our house rate forecast of a single cut this year is realized, higher coupons should again outperform in the second half of the year. Thinking more probabilistically, lower coupons remain a viable alternative to Treasuries (though bank portfolio sales are a risk), while higher coupons can provide an alternative to corporates (more on that later). But with OASs in the 20-30bp range and the marginal buyers fairly spread sensitive, we think excess returns will need to be driven by earned spread rather than tightening.

We’ll dive into the supply/demand dynamics for the second half of the year in more detail, but at a high level, the picture ( Figure 80) looks fairly similar to what we’d come up with when we were writing our year-ahead outlook last fall. While we still expect money managers to be the most important source of MBS demand this year, we have slightly upgraded our expectation for bank buying (from -25bn to +25bn) given the stronger than expected buying in the first half. Foreign investment, on the other hand, has been underwhelming (per TIC data), and so we’re reducing our forecast (from +125bn to +100bn). Dealers have continued to grow their pool and CMO holdings, though given that they net those against TBA hedges, ‘demand’ from dealers is likely passed along to money managers as well. Our expectations for Fed runoff and net issuance remain unchanged.

Figure 80: Supply looks largely the same, and demand has only evolved slightly from our year-ahead outlook projection

Sources of agency MBS net supply and demand, $bn

Source: J.P. Morgan

One interesting trend that has emerged is that demand for Ginnies and conventionals can almost be thought of as operating on parallel tracks. Banks have a strong preference for Ginnies, and money managers have maintained a heavy overweight to higher coupon conventionals. With net issuance coming solely in those products, that’s been enough to make the picture balance; 15yrs and other refi products are steadily shrinking as a share of outstandings, given the curve inversion and lack of refi activity.

In the rest of this piece, we’ll look at mortgage spreads relative to corporates in larger rate shocks, estimating the break-even spread widening required on the IG end to net out against the negative convexity of MBS. We also consolidate the myriad reasons for why conventional s-curves remain quite flat; most would recede in importance into a sharp rally. We also look at how seasonals and daycount are largely modulating deep discount prints, implying that we aren’t seeing any significant trend drift that would lead to model recalibrations. Ginnie peak/post peak relationships also look better behaved than in prior years, even if they are well above conventional speeds. Finally, we review the spec landscape, looking at tradeoffs vs. TBA on the OAS and ZV front.

Demand

U.S. commercial banks: Bank buying in AFS/HTM has been tricky to estimate this year. Though the H.8 data showed surprisingly robust numbers to start the first quarter (+$44bn), it turned out that the first week’s reading of +$12.5bn most likely settled in late 2023, and that around $26bn came from commercial bank trading accounts (which we think represent less durable buying associated with rolls). The quarterly call reports indicated that bank MBS holdings were +$8.5bn in the first quarter, which roughly tied out with the adjusted H.8 tally of +$5.6bn.

Bank portfolio repositioning in the second quarter knocked the year-to-date total down by $26bn in one week, though it has subsequently rebounded. Currently, net buying in Q2 comes to +$5.2bn, and netting all of these factors together, YTD buying looks like +10.8bn. We’re going to put banks in for $25bn of buying for the full year, with some potential upside.

Deposit growth has been relatively muted (see Figure 81), though the same can be said for loan growth ( Figure 82); banks have been generally letting their lower yielding fixed rates burn off and replacing them with Ginnie floaters (CMOs or asset swapped fixed rates). Wherever they can find ways to raise capital, they seem to be taking the opportunity to speed the runoff of discounts with outright sales. It is very hard to project the timing and size of this activity, but we think that it will continue to be a theme for the MBS market, partially offsetting otherwise positive demand for mortgages.

Figure 81: Deposit growth has been relatively muted...

Change in Domestic Commercial bank Borrowings and Deposits, $bn

Source: J.P. Morgan, Federal Reserve

Figure 82: … as have been the changes on the asset side

Change in Domestic commercial Bank Holdings of Loans, Securities, and Cash, $bn

Source: J.P. Morgan, Federal Reserve

One wildcard for banks is what happens with Basel 3 Endgame. A recent news story suggested that the proposal had been scaled back such that the total required capital build might only be 5% (vs. the originally required 16% increase). This would likely mean that banks would be able to build to their required risk capital levels faster, and thus be less constrained by RWA over the medium term.

Most of the capital increase in the original B3e proposal comes from operational capital, rather than changes to asset risk weights, but the knock-on effects would still be felt in securities investment. When risk capital is the binding constraint, on the margin banks want to buy low RWA assets to improve their capital/RWA ratio. That means more 0% risk weight Treasuries and Ginnies (vs. higher RWA loans, or even 20% risk weight conventionals). In the very near term, banks are likely still risk capital constrained (regardless of what happens with B3e), and so they’ll continue to prefer Ginnies. But over the medium term, that regulatory-driven preference may fade sooner than previously anticipated, if these changes are finalized.

It also remains to be seen whether regulators choose to reduce mortgage loan risk weights (and put them more in line with the European B3e proposal). Doing so might could nudge the securitization rate down, on the margin, as loans could look more attractive to banks.

Money managers: We think money managers have been responsible for most of the net buying to start the year, but have done so without the active managers significantly increasing their overweights ( Figure 83). Instead, robust inflows on both the active and passive side have helped them to take down many forms of spread product ( Figure 84). The heavy overweight in higher coupon conventional 30yrs by some of the managers who have seen the largest inflows has been particularly supportive, since it has helped the market digest net supply in GSE 30yrs. With corporate spreads just modestly off their recent tights, and YTD excess returns on higher coupons giving a little more credence to the mortgage vs. corporate trade, we think that money managers can maintain their current allocations given the yields available in mortgage space.

Figure 83: Money managers have been buying MBS this year without increasing their overweights …

MBS and Corp weightings of large money managers versus the MBS and Corp weightings of the Bloomberg Agg, %

Source: J.P. Morgan, MorningstarDirect

Figure 84: … as both active and passive inflows have helped them take down many forms of spread product.

Monthly fixed income bond inflows to passive and active funds, $bn

Source: J.P. Morgan, EPFR

Inflows will remain the key consideration—they’ve been strong enough to help mortgages weather the seasonal upswing in forward origination sales, but the outlook is somewhat weaker as trailing returns on the Bloomberg Agg have softened a bit ( Figure 85). That may be a headwind in the near term, though as the year progresses and the trailing returns window shifts, 6m returns on the Agg should start to look better; that should again be supportive of inflows.

Figure 85: Trailing returns on the Bloomberg Agg have softened slightly

Trailing 6 month U.S. Bloomberg Agg returns, with projections based on a 5.07% YTW assumption, %

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

We lump insurance into the same bucket as money managers (our ‘plug’ in the supply/demand equation). It's harder to get timely stats on their buying, but anecdotally they’ve been adding, particularly in some of the longer CMO cashflows.

Foreign demand: Through April, TIC data indicates that foreign demand has been muted ( Figure 86). The Fed’s agency custody holdings series has also trended downward since November of last year ( Figure 87). The Japanese MoF data on flows into US long term debt shows that the strong start to the year may have faded more recently ( Figure 88). With the Fed on hold, currency hedged yields are still unappealing vs. local currency bonds.

Figure 86: Year-to-date foreign demand has been muted so far this year …

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 87: … and foreign official U.S. agency custody holdings

U.S. Federal Agency Securities held in foreign official accounts, $bn

Source: J.P. Morgan, Federal Reserve

Figure 88: Japanese US long term debt flows have faded recently

Japanese ministry of finance net flows of U.S. long-term debt securities, adjusted for monthly average USDJPY exchange rates, $bn

Source: J.P. Morgan, Ministry of Finance

There’s been some concern that Japanese banks are joining their American counterparts in looking for ways to unload low coupon MBS. Disclosures for the large Japanese banks lack the same detail that’s available for US institutions, so it’s very hard to quantify how much banks hold of the $242bn that TIC links to Japanese holders. That being said, the large banks do report their mark-to-market gains/losses on foreign bonds, and in aggregate the largest Japanese banks had significant losses at the end of the 2023 fiscal year. It remains to be seen whether there will be any significant net selling.

So why do we have $100bn still penciled in for the second half? We do think that as we get closer to a Fed cut, some of the traditional foreign players may be more tempted to come back into the space. Additionally, last year, the non-traditional buying localities (think Ireland, Canada, the UK) seemed to have accelerated their buying just as spreads enticed US money managers to become more active. If that trend persists this year, we’d expect to see more buying. Admittedly, our confidence in this estimate is not particularly high, however, and there’s a chance that foreign buying continues to underwhelm. That would put more burden on banks and money managers, potentially biasing spreads wider.

Supply

Our view on organic net supply remains unchanged since we wrote our 2024 Outlook late last year. While rates have provided plenty of shocks, they have done so at high enough absolute levels that it hasn’t mattered for additional cash-out activity and the purchase market remains constrained by inventory. On that front, our initial projections called for no change in the number of sales/HPI, but as the year has developed it looks more likely that sales will be depressed relative to last year by 5-10% while our non-agency team has increased their call for HPA to +5.5%. Putting these thoughts together with the first five months of issuance, we feel $175bn remains reasonable for organic net in 2024. Figure 89 shows our rough month-by-month view and how net should be backloaded in this mostly purchase-driven environment. If our forecast has a tilt to it, it would be just slightly to the higher side, depending on housing activity through the rest of the summer.

Figure 89: We’re sticking with our $175bn net call for 2024

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

Looking by product, Figure 90 shows the net issuance split for conventional 30yr/conventional “other” (refi products with shorter terms, mostly, like 15yrs) / Ginnies. Issuance outside of 30yrs remains extremely limited and runoff continues at a pace similar to where it moved to in mid-2022. The split of gross issuance between Ginnies and conventionals continues to slowly shift towards Ginnies, now at 40% versus last year’s 37.5%. In Figure 91 we use the current split along with our expectations for the full year’s net to project 2024 issuance by product. This year, its possible that 93% of net will come in Ginnies thanks to their greater share of purchase origination and the continued drag from conventional refi products.

Figure 90: Net has started to pick up thanks to seasonal trends, with growth in 30yr purchase products…

Net issuance by agency / term / product, by issuance month, $bn

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

Figure 91: …while it looks like Ginnie net issuance could overtake conventional 30yrs for the first time in over a decade

Breaking out Conv 30/Conv “other”/G2 30 gross and net issuance projections for 2024 based on an assumption around the Ginnie share

	Gross			Net
	Conv 30	Conv Other	Ginnie	Conv 30	Conv Other	Ginnie
2024 Through May	216	12	152	46	-53	59
2024 Full Year	555	31	389	140	-128	163
Ginnie Share			40%			93%

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

Relative value—vs. Treasuries and IG

As has been the case for a few years now, the question “are mortgages cheap” is still very much a question about which mortgages you’re talking about, and what you’re comparing them to. As Figure 92 shows, the coupon difference between the index and the par bond remains at historically elevated levels, with today’s 240bp spread in the 5^th percentile (by absolute difference) since 2000. In Table NaN Figure 93, we plot out a histogram of these differences over the same period just to highlight how abnormal this period is—and that’s even after the index coupon has drifted up ~60bp from its early 2022 low.

Figure 92: Even as the index coupon has drifted up 60bp from the early 2022 low…

Current coupon and index coupon (%, left) and CC-index coupon (bp, right)

Source: J.P. Morgan

Figure 93: …the current coupon / index coupon difference remains a historical outlier

Histogram of daily difference between FN 30yr current coupon and index coupon, bp

Source: J.P. Morgan

That being said, the floats in higher coupons are thickening as production remains focused on 5s-7s, and bank portfolio rebalancing activity continues to unlock specs in deep discounts. This has left the conventional 30yr stack with 10-12 tradeable coupons, depending on your definition of liquidity. The wings of the stack still have markedly different characteristics, with deep discounts offering 40-50bp ZV picks against Treasuries (and positive convexity), and production coupons offering 125bp ZV, at the expense of much less cashflow certainty. The belly offers a gradient somewhere in between.

Figure 94: OASs across the stack have settled into a tighter range

Current, 1m, and 6m treasury OAS ranges across the TBA stack in the 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 6/26/2024)

Source: J.P. Morgan

Figure 95: Current coupon OAS sits around its 2015-2018 range...

Current coupon historical treasury OAS and latest level (bp)

Source: J.P. Morgan

Figure 96: … but ZV spreads are still historically wide, thanks to the inverted curve and higher implied vol

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

Source: J.P. Morgan

None of this, of course, is new—but OASs across the stack have settled into a tighter range (see Figure 94). That makes the question of how one is using mortgages perhaps more important when determining which mortgages to buy. Broadly speaking, lower coupons remain compelling 5-7yr Treasury proxies, while higher coupons are a way to pick carry, fade vol, and arguably shed some exposure to historically tight IG corporate spreads.

On that front, there’s still no commonly agreed on way to compare mortgages and corporates. Current coupon mortgages pick ZV spread vs. similar duration corporates, but OAS doesn’t look particularly cheap vs. the bond/CDS basis (see Figure 97 and Figure 98). It may be that boiling down either of the key risks to each of the two products—that is, default risk for corporates and rate risk for mortgages—may be impossible without a ‘theory of everything’ OAS approach that simulates economic conditions and rate paths jointly. When we solve that, we’ll probably be able to retire shortly thereafter.

Figure 97: Current coupon still looks cheap to similar duration corporates on a ZV basis ...

JULI 3-5y Treasury spread, FN 30yr cc ZV spread, and difference (bp)

Source: J.P. Morgan

Figure 98: … but OAS doesn’t look particularly cheap vs. the bond/CDS basis

JULI Index Treasury spread – CDS basis and MBS Index Treasury OAS (bp)

Source: J.P. Morgan

Until then, we’re still looking for other ways to compare the two, and one possibility is running a break-even analysis in rate shocks. Locally, mortgages and corporates have exhibited some rate directionality (though with opposite signs). Corporate buyers frequently have yield bogies, and thus IG spreads have compressed as all-in yields have gone higher—at least to a certain point (~4.50% on 5s). Mortgages, in contrast, have tended to widen into rate selloffs, as technical factors (money manager inflows and bank deposit growth) are perceived to weaken into selloffs.

Figure 99: Locally, corps and MBS have experienced rate/spread directionality of the opposite sign

FN 2 OAS (left, bp) vs. 5yr UST yield (%) and JULI Index Treasury Spread (right, bp) vs. 5yr UST yield (%), observed daily in 2024

Source: J.P. Morgan

That’s locally, however—these relationships probably wouldn’t hold so tightly in larger rate moves. For example, many investors are overweight MBS (and underweight IG) because they’re worried about a recession. In that world, mortgages could stay on spread (OAS), while IG risk premium could increase. But in that world, rates are also likely to sharply rally, bringing the negative convexity of MBS into play.

MBS investors have an agreed upon framework for simulating prices in rate shock—even OAS pricing—and arguably mortgages won’t go too much tighter in OAS terms, unless the Fed’s actively involved in the market again. For IGs, it’s simple enough to compute the price impact purely attributable to rate shifts and discounting, in a vacuum. But any rates strategist will tell you that a 300bp rally doesn’t happen in a vacuum. Most investors would agree that IG corporate spreads would likely go wider into a sharp rally, because that rally would be precipitated by a shift to a weaker economy.

If we take the price changes in rate shocks in Figure 100 (which assume no move in spreads), we can then estimate how much less MBS go up in price than comparable duration corporates into large rallies. Then, using the duration of the corporates, we can estimate how much IG spreads would have to widen to have the corporates only go up in price as much as the MBS ( Figure 101). For example, we compare duration matched corporates and mortgages (using our JULI 3-5yr bucket and a blend of FN 5s/5.5s). Into a 300bp rally, the MBS gain 6pts in price, while the corporates increase 11.2pts. For the corporates to only have gone up 6pts, they would need to have widened by 139bp.

Figure 100: Into large rate moves, mortgages underperform if spreads don’t move…

Price moves of JULI 3-5yr IGs and duration matched MBS (FN 5s/5.5s), assuming parallel rate shocks and no spread changes, points per $100 of notional

Source: J.P. Morgan

Figure 101: … from which we can compute ‘break-even’ IG widening amounts to make corporates experience the same price move as MBS

Amount of spread widening on 3-5yr IGs to make their price move equivalent to an even duration blend of FN 5s/5.5s in different rate rallies, bp

Source: J.P. Morgan

There’s no doubt that this framework is overly simplistic— what happens if the rate shift isn’t parallel? What if vol moves? As mentioned earlier, the ability to simulate economic scenarios across rate paths would be more complete, but lacking that, this approach does benefit from its simplicity. In a sense, it boils down to this question: if shorter duration mortgages offer competitive yields vs. corporates today (see Figure 102), is the break-even spread widening on IGs in a recession scenario priced well? For longer duration, lower yielding mortgages like FN 2s, a similar analysis can be performed—but the positive convexity of those cashflows means that they can hold their own against corporates in rallies even without spread widening. Thinking of low coupon mortgages more like Treasuries with some additional spread pickup is probably the more appropriate comparison.

Figure 102: FN 4.5s and up outyield similar duration corporates, while lower coupons remain tighter

Yields on FN 30yrs (by coupon) and IG corporates (by JULI duration bucket) vs. duration, %

Source: J.P. Morgan

Prepayments

A few themes have stood out in 2024 prepayments looking back from the midpoint of the year. The first is the split reaction to the over 1% drop in mortgage rates early on that saw well-behaved prints in conventionals versus a more aggressive refi reaction in Ginnies. Below, we’ll go through a list of reasons why we think conventionals prepayments have been muted versus prior periods. Second, on the discount side, there has certainly been noise and excitement month-to-month (like February acceleration), but from a wider lens, 2s and 2.5s have paid close to what we would expect from normal seasonal patterns over the past year. Third, we’ll consider recent developments in Ginnies since the dramatic prepayments in Jan/Feb. While refi rates have been more aggressive in Ginnies, the behavior of post-peak pools in recent months hints towards the s-curve also being flatter on the government guaranteed side. These themes remain similar to what we saw heading into the year with the early test of speeds confirming the flatter GSE s-curve into at least “shallow” rallies.

To set the stage, Figure 103 gives perspective on what type of rates it would take to get more conventional borrowers into the refi window. Only 4% have rates of 7% or above, though a sizeable group is starting to build up just below that level. For those loans that have experienced any incentive, the conventional s-curve looks almost as flat as it did this year versus last. Figure 104 and Figure 105 show how the flatness of 2024 stacks up against refi waves years as well as other years (like 2018) where speeds were also muted. Figure 103

Figure 103: Only a sliver of conventional loans are near the refi window

Outstanding conventional 30-year loans by WAC and vintage, $bn

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

Figure 104: The 2024 YTD s-curve looks as flat as last year’s when compared to “refi waves” years...

1m CPR by incentive across various observation periods, conventional 30yrs ≥ 6 WALA, versus rate incentive (x-axis)

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

Figure 105: … and even falls near the bottom of pile when compared to “refi duds”

1m CPR by incentive across various observation periods, conventional 30yrs ≥ 6 WALA, versus rate incentive (x-axis)

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

The flat conventional s-curve

How do we explain the calm of 2024? Here is a list of key factors that we think have each contributed to flat conventional s-curves. The list is by no means exhaustive, and we’re constantly adding to it. One thing that stands out to us is that many of these factors would cease to be relevant into a large enough rally—but it’s hard to precisely say how big that rally would need to be.

Points paid: We wrote a few months ago about Freddie’s data release showing significantly more borrowers paying points, and paying more of them on average, since 2022 ( Figure 106 & Figure 107). This can impact prepayments in a few ways, one of which is not having cash reserves on hand to pay for another refinance. Additionally, it could impose a psychological hurdle to a near-term refi for only a marginal savings, as it would seem to “waste” the money spent on getting a better rate. Time since origination should erode both of these impacts and a larger rally would make no-cost refis viable.

Figure 106: The share of borrowers paying points surged after 2021...

Share of borrowers who paid points by product type, Freddie 30yrs

Source: J.P. Morgan, Freddie Mac

Figure 107: ...and the amount of points they are paying is also higher

Average points paid by product type, Freddie 30yrs

Source: J.P. Morgan, Freddie Mac

Originator Layoffs: One of the biggest mysteries of 2023 was the seeming glut of mortgage industry employees relative to volumes, which was cleared up by a massive revision in the BLS data. The revised dataset showed employment closer to the “right” level given issuance ( Figure 108). However, a 30%+ drop in employment can’t happen without friction, and it is plausible that many borrowers no longer have contact with their initial loan officer, making it more difficult to pitch refis through established personal connections.

Figure 108: The fast pace of originator layoffs means that many borrowers have lost contact with their original loan officer

Mortgage employment (k) vs. MBS issuance ($bn)

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

FICO inflation: One surprising result of the pandemic and subsequent government programs was an upward drift in credit scores as delinquencies shifted lower. This Fed note goes over the details and how it has impacted delinquency rates of late on lower credit score borrowers. For mortgages, FICO tends to impact the steepness of s-curves, in addition to the elbow shift that comes from differences in LLPAs. So a model trained to earlier periods with “lower” FICOs might overestimate what speeds look like on today’s “inflated” FICOs. There is some evidence of this drift on FICOs more applicable to the major pools (720+) from MPL ABS delinquencies. Our ABS team has seen delinquencies for 720+ borrowers trend higher at a much faster rate on the unsecured loans originated through the Fintech ecosystem ( Figure 109). According to a recent FICO blog, the degree of drift since pre-pandemic is in the range of 15-20bp (https://www.fico.com/blogs/average-us-fico-score-718), so the 740 borrower of today might prepay more like the 725 of yesteryear.

Figure 109: FICO drift can be seen in the faster pace of 720+ loans going delinquent in MPL ABS

30+ delinquencies by origination quarter for loans >=720 FICO across MPL ABS pools

Source: J.P. Morgan, dv01

SATO Borrowers and the case of the missing re-lock: The greater dispersion in rates that borrowers are getting has been a fact of life since the major sell-off in 2022. As a result, more borrowers than ever our getting rates well above the average for their time of origination ( Figure 110). Some part of that dispersion can be chalked up to SATO, but another factor is the volatility of rates within a given month. That factor also says something about borrower mentality. Why didn’t these borrowers relock their loan or go to another lender when rates moved significantly lower? It is hard to measure the impact here but its another bit of supporting evidence that the borrowers’ with the highest rates are those that are least reactive to lower driving rates.

Figure 110: During these turbulent times in rates, many loans are getting originated at rates well above the average from their origination month

Share of loans originated at different removes from the average for a month, based on origination date, for conventional 30yr loans in worst-to-deliver pools with FICO >=740 and LTV ≤ 80

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

Media Spending Down: Using the filings of public nonbanks as a guide, it is no surprise that the amount spent on advertising has gone down significantly since early 2022 ( Figure 111), making mortgage rates less front of mind for homeowners.

Figure 111: Spending on ads is well below the 2021 highs

Non-bank quarterly marketing / advertising expenses, $mm

Source: J.P. Morgan, company filings

Lack of Media Effect/Targeting: On the conventional side, it remains the case that rallies like the one in early 2024 are not driving any excitement in refi activity. In comparison, Ginnie originators have had to be more aggressive in their offerings to retain borrowers (and their MSR). One way to see the divide is in comparable refinance rates. Since October 2022, VA rates have been up to as much as a percentage point lower, while FHA rates have been 50bp lower of late ( Figure 112)

Figure 112: Rate/term refi WACs have been particularly low relative to GSE purchase rates for VA and FHA loans

Orig. WACs by issue month, split into GSE, FHA, and VA purchase and rate/term, %s

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

Fewer appraisal waivers: There are a few factors that influence the granting of appraisal waivers, chief among them likely being the GSEs confidence in the accuracy of their automated valuation models. Looking just at the issuance data, the granting of waivers has declined substantially since 2019 for the few rate/term refis getting created, another potential friction relative to the latest refi wave ( Figure 113)

Figure 113: Appraisal waivers on refis are down relative to 2019, not just the pandemic era

Share of loans in worst-to-deliver pools, <=80LTV, that were given an appraisal waiver, by purpose

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

Shorter terms offer little rate solace: The inverted curve makes it that much more difficult to give better rates on common refi products like 15yrs and 20yrs. Using the issuance data, its easy to see the connection between curve slope and term of refis ( Figure 114). As a result, if there are any borrowers out there who refuse to extend their term with refis (and require a certain amount of rate savings), they are likely sitting on the sidelines during bouts of moderately lower rates. This would contribute to a flatter GSE s-curve today versus periods like 2020-2021.

Figure 114: Borrowers who do refi are almost exclusively going into a 30yr term

Comparing the share of 30yr rate/term refis against all product types (left, %) and 2yr Treasury minus 10yr Treasury rate (right, %)

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

This list of factors hopefully conveys our view that conventional s-curves aren’t flatter today for one single reason but a host of factors that can (and likely have) ebb and wane in influence over the rest of the year. Some of them are tied to the buyers who are seeing incentive (like SATO/relocking) and others to the origination environment (layoffs/ad spending/ curve shape). Together, they can continue to keep speeds contained into modest rallies though some of these hurdles would disappear into a larger drop in rates.

Discounts

For discount speeds, changes in trend tend to happen over years rather than months, moving in tune with demographics and the macro economy rather than near-term moves in rates. Still, monthly surprises tend to raise eyebrows. As we discussed at length a few weeks ago, the changes from month-to-month should largely be impacted by daycount and seasonal trends in EHS. Since those seasonals aren’t set in stone by month, and are influential, any rearrangement in them can cause misses in discount forecasts that are rectified over the next month or two.

We can look at this phenomenon more directly using recent results as a guide. Consider 2s and 2.5s from 2021 to avoid most aging ramp interference. We can see the “implied” seasonal change for each month by looking at the daycount adjusted move in SMM (100% = no change in SMM). This can be compared to our view of “normal” seasonal factors informed by past years. Figure 115 shows the comparison. In Feb. 24, for example, the empirically measured change was 123%, while normally we would only expect an increase of 106%. The next month, however, the implied change of 117% was well below our expected 131% increase, supporting an argument for rearranged speeds rather than a level shift higher.

Approaching this experiment from the other angle, Figure 116 shows the actual accumulation of daycount adjusted SMM starting from June 2023 against what our seasonal factors would have implied. Over the course of the year, the two measures diverged by just 0.1% SMM (a less than 2% difference) as actual speeds started slightly slower before moving slightly faster during 2024.

Figure 115: Seasonal factors are a rough guide and can shift by a month or two in any given year...

The change in SMM m/m post-daycount (x-axis) against our assumed seaonal factors (y-axis) over the past year

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

Figure 116: ...which should even out over a long time period if there isn’t a fundamental change in level

Cumulative change in SMM based on actual speeds (left, blue) versus what seasonal factors alone would predict (gray, left) and the difference (right)

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

In derivs, valuations on 2s IO have moved in accordance with a cashflow that is becoming more predictable. Since the wides in October 2023, our index of 2020 2s IO bonds has tightened nearly 200bp in our model ( Figure 117). Beyond more certainty around the cashflow, a lack of new supply and updates to the YieldBook model likely contributed to the tighter spread level. It’s hard to argue for additional tightening from here on out, but these IOs still yield 6.4% with minimal duration concerns.

Figure 117: 2s IO has tightened in recent months, pointing to the predictably of the cashflow among other factors

2020 2s IO index Tsy OAS, bp

Source: J.P. Morgan

The most recent development impacting discount speeds is the surprise ruling from FHFA to allow Freddie to purchase closed-end second liens in only a limited trial with a cap of $2.5bn over eighteen months . When the proposal was first announced, there were concerns that cashouts would evaporate completely from lower coupons. The limited scope should alleviate those fears on the pool side. Of late, GSE total cashout volume has been running at nearly $5bn a month. Assuming 1/3 of that is new equity, it would take just three months to use up the capacity in GSE second lien, ignoring the additional limitation of the loan cap at under $80k ( Figure 118).

Figure 118: It would only take three months or so to use up the $2.5bn per program capacity of second liens at today’s rate of origination

Issuance of cashout loans by month, $bn

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

Ginnies

 The rate path of 2023 put Ginnies in an interesting spot where the typical “peak” pool behavior was interrupted by prohibitively high rates, depending on the coupon. This was notable in 6.5s, where a decent amount were made towards the end of 2022 at near 7% WAC, but rates were near or above that level and trending in the wrong direction when they were first eligible to refinance. The rally heading into 2024 gave lenders a chance to refi borrowers in many of the pools that missed their peak month in January and February, leading to a plateau rather than a mountain in speeds. Figure 119 shows that plateau in table form, and also shows how those in-the-money pools have paid in the three months afterwards, typically averaging in the high 20s.

How do those speeds compare to other periods? One way to see it is to take a coupon with similar moneyness (or peak speeds) and see its post-peak prepays. 2.5s in 2021 are a decent comparison, printing in the mid-50s after the early 2021 backup in rates ( Figure 120). During the second half of 2021, 2.5 multi-issuer pools paid in the high 30s to low 40s. As a result, its fair to say that the Ginnie s-curve also seems less aggressive than it was during the most recent refi wave, even if that behavior is sometimes obscured by the initially high peaks.

Relative value in FN 30 specs

There are opportunities for sizable OAS pickup across the spec space, although many options come with a heavy cost to ZV/carry. In Figure 121 and Figure 122, we show the OAS and ZV pickups across all coupons/specs, respectively, and circle the ones that provide a pickup in both (or a negligible ZV hit). The OAS gain/ZV loss tradeoff is most acute on higher coupons loan bals, where OAS pickups reach about 20bp on LLBs but come with a 40bp hit to ZV on 5.5s and 6s. Given that FN 5.5 and 6 TBA have ZV in the 120-130bp range, a 40bp hit leaves them with about as much as a belly FN 4.5/5, though with higher OAS as a baseline in addition to the significant pickup from loan balance convexity.

However, there is uncertainty around how durable the loan size prepayment protection will prove into rate rally scenarios. Generally, originators target bigger loans into refi opportunities, which helps protect smaller loan size collateral. This queuing effect may be less impactful now though because originators have excess capacity and balances on the higher coupons are not large enough to overwhelm this capacity. For investors that want to fade model risk, lower payup options like Texas and high LTV pools make sense. While they provide a smaller OAS pickup (~10bp), they do not come with any hit to ZV.

The lower coupon loan bals offer greater OAS pickup than the belly and well as positive ZV pickup, on account of faster built-in speeds making up for the payup. These seem like better overall value propositions than the belly loan bals. Our model sees wider OASs on lower coupons vs. the belly as a baseline, and the ZV cost on belly loan bals means that their ZV advantage is minimal (ZVs ~ 60bp across FN 2-3.5 loan bals and ~65 on FN 4-4.5 loan bals). Texas, Florida, and low FICO pools seem like better options in the belly, still providing an OAS pickup without the cost to ZV.

On the seasoned pool front, some OAS and ZV pickup can be gained on the lower coupons by buying 2020 and 2021 vintage ( Figure 123 and Figure 124). Then across the belly there are opportunities for sizable OAS pickup by buying 2019 and earlier vintage collateral, at the price of  some ZV spread. These pools make sense for investors that want long-term exposure to the belly of the stack. The OAS pickups are quite large (up to ~35bp on certain vintages and coupons), and should manifest in outperformance over TBA into most rate scenarios.

Recommendations

1.       Seasoned and loan bal in lower coupons (sizable pickup)

2.       FL/TX across the stack (mild payup/mild pickup)

3.       Low FICO pools in the belly, high LTV in higher coupons (mild payup, mild pickup)

Week in Review

• MBA Weekly Survey: For the week ending June 21, the purchase application index rose 0.2% w/w and was 13.1% lower than year ago levels, while the refinance index fell 0.1% w/w and was 11.9% higher than the 3-month trailing level (daycount-adjusted, not seasonally-adjusted) ( Figure 125 & Figure 126)
• Freddie Enhanced Primary Survey: For the week prior to June 27, 2024, 30-year conventional conforming fixed-rate mortgages averaged 6.86%, down 1bp from the previous week ( Figure 127).
• Primary dealer specified pool positions rose to $414.7bn (-$4.9bn w/w) as-of close trading June 19. Including TBA positions of -$355.3bn, dealers were long $59.4bn (+$0.5bn w/w) pass-throughs. Other agency MBS holdings rose $1.3bn to $29.2bn.
• Fixed-rate agency gross and net issuances were $88.9bn and $14.6bn, respectively, in May. June gross supply currently stands at $94.1bn ( Figure 128).
• Our OAS optimized portfolio favors FNCL 1.5s-2s, FNCI 1.5s, and G2 2s, while reducing exposure to FNCL 2.5s-3s & 4s-6.5s and G2 2.5s & 4.5s-5.5s. Our optimized portfolio achieves an OAS of 40 with an OAD of 5.58 versus the index’s OAS of 37 with OAD of 5.37. Cohort weights are limited to 40% redistributions in our optimized portfolio, while duration cannot change by more than 1 year and convexity can be no worse than the index ( Figure 129 & Figure 130).

Figure 125: MBA Purchase Index, calendar year overlay with daycount adjustments

Source: J.P. Morgan, MBA

Figure 126: MBA Refi Indices, seasonally adjusted

Source: J.P. Morgan, MBA

Figure 127: Primary mortgage rates, %

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

Figure 128: Gross and net fixed-rate MBS monthly issuance, $bn

Source: J.P. Morgan

Figure 129: Weights of the actual index and an OAS optimized portfolio

Weight distribution by coupon and product as for an OAS optimized portfolio as of 6/26/2024. Weights of each cohort are allowed to be shifted by at most 40%.

Source: J.P. Morgan

Figure 130: Weight redistribution for the OAS optimized portfolio

Difference between optimal and actual weights by coupon and product for an OAS optimized portfolio as of 6/26/2024. Weights of each cohort are allowed to be shifted by at most 40%.

Source: J.P. Morgan

Unwinding the first half?

• Mortgage credit spreads are already retracing wider from their tightest levels this year. We expect further widening into year-end as we enter a slower GDP-growth and election period, although a possible Fed rate cut could help
• 2.0 supply picked up as execution improved and banks became less competitive on loans. Expect supply to remain elevated and PT price spreads to widen to 1-12bk by year-end
• Non-QM has been shaped by insurance demand for loans and securities. Agency MBS 7% coupon nominal spreads hover around 120-125bp, which is where AAA non-QM has priced. Expect to see AAAs widen out to 150-175bp
• Strong housing fundamentals look fully priced into CRT spreads as well. Expect spreads to be tied to labor market strength, but limited supply could keep spreads firmer than at the top of the capital stack
• ‘21/‘22 vintage CRT B2 spreads are cheap to new issue collateral, while spreads are more in line with duration for B1s
• Delinquencies have been increasing in recent vintage non-QM and CRT, although they have moderated slightly over the last few months. We do not anticipate material losses from the recently revised Flex Mod in CRT subs
• We revise our home price forecast higher to 5.5%. The narrative remains low supply, although new home inventory is piling up as affordability remains a challenge
• We look for $110bn of gross supply for 2024, up from $97bn previously

Figure 131: RMBS credit issuance to date...

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

Figure 132: ...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 - Credit curve flattening to stall in second half

While 2023 was a year of significant spread curve flattening, we continued to see tighter spreads in 1H24. Spreads across securitized products shifted from cheap/fair to rich. Not much sits in the positive z-score space anymore ( Figure 133). It shouldn’t be a surprise then that spreads are already retracing wider from their tightest levels this year. Could this just be a quarter-end head fake or a real shift wider? We think interest levels at the tights of this year were waning. Jumbo 2.0 price spreads went inside of a point back but have now jumped wider to $1-04 back. Recent deal pricing showed less interest in LCF paper, where spreads widened to 190bp (Tsy@15c). This, in turn, likely moved the pass-through wider. We expect 2.0 pass-through price spreads to soften to $1-12 back, which is more aligned with a 200bp LCF (based on a two year regression of price spreads with LCF spreads). This would put the jumbo pass-through TOAS back into the 70s, a level where investors had a much easier time deciding whether to add (picking up 35-40bp over agency MBS).

There are some major factors to consider with our spread targets. Our targets are wider for year-end. This is, in part, linked to the belief that spreads are just too tight and hard to support. Moreover, we are heading into a period where GDP growth is expected to slow. This is typically a spread widening event. We also have an election (more volatility) and perhaps Fed rate cuts, which have opposing effects. We feel much better setting a price target for October and leaving November/December as a wildcard. The good news is that the consumer remains strong and home prices should remain resilient, despite lofty levels.

Figure 133: Securitized products shifted from cheap/fair to rich in 1H24

Z-score - number of standard deviations rich/cheap to 5yr spread average (3yr average for CRT)

Source: J.P. Morgan

Non-QM spreads are also tight, led, in part, by consistent insurance buying in loans and securities. Anecdotally, demand has slowed and this could pressure spreads a bit. Interestingly, agency MBS 7% coupon nominal spreads hover around 120-125bp. At current rates, model long-term speeds for 7s are roughly 25-30c. These are a reasonable comparison for non-QM paper, albeit absent an issuer call option and lower coupon/carry. New issue non-QM gross WACs haven been roughly 8.5% with 6.5% coupons. Spreads tightened from 150-160 (25c/call) in January to 120-125bp now, depending on the deal. Regardless, this is right on top of TBA 7s ( Figure 134). Running non-QM to full maturity gets you another 40bp, and the spread pick-up is more compelling. However, we need to look at this from an OAS perspective. Our recently developed non-QM prepayment model gives a solid advantage to non-QM. For example, we see OAS as 107bp (to modelled call) with 7s OAS as 23bp. However, this is a lot for investors to get comfortable with when very few prepay models for non-QM exist. Investors buy in to slower empirical s-curves and no call scenarios ( Figure 135). Perhaps investors won’t believe all of that OAS, and in fact, most are likely not using an OAS framework at all, yet. Moreover, the models haven’t been tested into a rate rally and when investor loans exit the prepayment penalty window. Consequently, we expect to see non-QM spreads widen back out to 150-175bp area. This would put model OAS closer to 120bp and leave plenty of room for model error.

Figure 134: New issue non-QM spreads to maturity pick under 60bp nominally to TBA 7s, but only 20bp to call

Non-QM 2023 vintage AAA A1 spreads to treasury 25c/mty and 25c/call vs. FN 7s spread to treasury (bp, to model speeds of ~30c)

Source: J.P. Morgan

Figure 135: Agency mortgages are broadly more callable than non-QM collateral

S-curves for agency MBS and non-QM by documentation type, for 6-36 WALA, 2021-present period (non-QM still in prepayment penalty window)

Source: J.P. Morgan, Fannie Mae, Freddie Mac, Black Knight, CoreLogic

Figure 136: Non-QM picks over 80bp OAS to agency MBS and 60bp OAS to 2.0. However, models are new and investors are not likely to buy-in fully to the results yet. More comfort will form as we see non-QM exit prepayment penalty windows and into a rate rally

OAS analysis comparing agency MBS, jumbo and non-QM

Source: J.P. Morgan

At this point, strong housing fundamentals look fully priced in. We expect mortgage credit spreads to be more rangebound and tied to labor market strength and the potential for Fed rate cuts later this year. While the top of the capital structure appears to have room to widen, technicals in the form of a limited supply of mortgage credit (e.g. PLS subordinates and CRT) could keep spreads a bit firmer. Calls and tenders in CRT have most investors looking at the bonds as having a 5yr (or shorter) maturity. We continue to see ‘21/‘22 vintage CRT B2 spreads cheap to new issue collateral ( Figure 138). Investors seem to prefer the increased credit enhancement of recent issue deals (recall B2 enhancement was increased to 1% from 30-70bp in earlier vintages). We think the extra home price appreciation and call/tender potential of ‘21/‘22 vintages makes them cheap to ‘24. However, for B1s, spreads appear to be a bit more in line with seasoning as credit enhancement is less in focus ( Figure 137).

Figure 137: Spreads are more in line with duration for CRT B1s, with new issue looking cheap-to-fair...

CRT B1 current DM@10c to call vs. WAL by vintage (bp)

Source: J.P. Morgan

Figure 138: ...whereas new issue B2s looks a bit more expensive relative to 21/22 vintages. Investors are paying for higher CE, despite the shorter profile and accumulated equity of more seasoned paper

CRT B2 current DM@10c to call vs. WAL by vintage (bp)

Source: J.P. Morgan

We revise our home price forecast higher to 5.5% for 2024. YTD home prices, through May, are 3.3% ( Figure 139). With seven months remaining and the stronger summer months still to come, our forecast may still be too low. The narrative remains low supply. However, new home inventory is piling up as affordability remains a challenge. Currently, about 100k new homes are being delivered/month ( Figure 140), with about 52k being sold (New Home Sales is running at 619k/year or 52k/month). If current delivery remains the same and demand holds, excess supply would grow at about 580k homes/year. Builders are already pulling back on forward builds (housing starts are declining) to mitigate but if rates remain at elevated levels well into 2025, home prices and builders are likely to suffer. Pace will be key, given that we continue to argue that the U.S. has underbuilt by about 3.5 million units since 2015 (household formation has been outpacing net delivered homes).

Figure 139: Expect home prices to be up 5.5% in 2024

Historical and projected home price growth for the Case-Shiller HPI Index

Source: J.P. Morgan

Figure 140: Months of supply of new homes is running high. Completions are adding about 100k per month. This can be a source of supply pressure, but builders are already pulling back and homes under construction/not started can be moderated

New Home Inventory (Completed, Under Construction and Not Started, in 000s, monthly SAAR, lhs), months of supply of new homes (rhs)

Source: J.P. Morgan, U.S. Census

Jumbo: More supply on the way

Jumbo 2.0 supply picked up in the first half of this year, with YTD prime jumbo issuance totaling $10.3bn, just over last year’s total of $10.2bn. Agency investor issuance stands at $2.7bn, surpassing last year’s total of just under $1bn. Rates below 2023’s peak and tighter spreads have made for more favorable execution levels in 2.0. Furthermore, banks have been less competitive on loans, adding to the securitization supply. The increase in bank holdings has tapered; banks added $53bn in residential loans YoY, far lower than the $204bn added the previous year ( Figure 141). As such, we are raising our 2.0 issuance forecast slightly to $24bn for 2024.

Figure 141: The increase in bank holdings has tapered; banks added fewer residential loans YoY than the previous year

Commercial bank holdings of residential real estate loans, not seasonally adjusted ($bn)

Source: J.P. Morgan, FRED

As background, banks increasing retail loan holdings have historically driven the retail rate lower than the correspondent and broker rate. This is clear in Black Knight’s non-conforming jumbo rate, which largely captures retail loans ( Figure 142 and Figure 143). Meanwhile, Optimal Blue’s jumbo and conforming rates largely reflect correspondent and broker channels. While the retail rate continues to be lower than the broker/correspondent rate, the spread between the two has decreased. The spread now stands at 80bp versus the last peak of 155bp in Oct ‘22 and 112bp in Nov ‘23, pointing to lower bank demand as of late. As banks remain on the sidelines for now, we expect 2.0 securitization volumes to remain steadily higher than last year. That said, the strong insurance bid on whole loans is still likely to limit supply, especially for agency investor.

Figure 142: Black Knight’s non-conforming jumbo rate largely captures retail loans

Optimal Blue jumbo and conforming rates vs. non-conforming jumbo rates from Black Knight’s McDash database (%)

Source: J.P. Morgan, Optimal Blue, Black Knight

Figure 143: The spread between retail and correspondent/broker has decreased, pointing to lower bank demand

Non-conforming jumbo rates by origination channel (%)

Source: J.P. Morgan, Black Knight

Better execution for PLS 2.0 versus the GSEs has led to a higher percentage of conforming jumbo collateral in recently issued deals ( Figure 144). With the increase in loan limits in 2024 and the median home price continuing to reach new highs (Apr ‘24 saw an all-time high of $389k), we have also seen an increase in the average original loan size of pools ( Figure 145). This could imply more convexity, and potentially faster speeds if we see a rate rally; non-conforming jumbo convexity has been marginally worse than conforming jumbo over the last year.

Figure 144: We have seen a higher percentage of conforming jumbo collateral in recently issued deals

% conforming jumbo by deal issue date. Each dot represents a prime jumbo deal

Source: J.P. Morgan, CoreLogic

Figure 145: We have also seen an increase in the average original loan size of pools

Average original loan size ($) in prime jumbo collateral by origination date

Source: J.P. Morgan, CoreLogic

For now, higher coupon jumbos are less reactive to incentives in rates, with lower negative convexity in the 2023-24 performance period than in prior years ( Figure 146). At the same incentive levels, deeply out-of-the-money jumbo pays marginally faster than in the 2021-22 period, while 50bp in-the-money jumbo pays roughly 25c versus 40c in prior years.

Figure 146: For now, higher coupon jumbos are less reactive to incentives in rates

S-curves for securitized prime jumbo, split by performance year

Source: J.P. Morgan, CoreLogic, Optimal Blue

Meanwhile, jumbo spreads have widened out relative to May’s tights, with higher coupon PTs at roughly a point back of TBA. PTs continue to offer 25-40bp of TOAS pick-up relative to same coupon agency MBS ( Figure 147). Still, this spread is at the lower end of historicals. Our on-the-run jumbo PT index shows that the spread between jumbo PT and same coupon TBA TOAS last peaked at 52bp in Jan ‘24, wider than today’s 24bp for 6cpn PTs. As mentioned, we expect some softening in the space as investor interest declines and see the PT widening to 1-12bk. This should push the TOAS closer to 70, a more attractive level relative to agencies.

Figure 147: Jumbo PTs continue to offer 20-40bps of OAS pick-up relative to same coupon agency MBS, although this spread is at the lower end of historicals

Jumbo vs. agency model runs by coupon

Source: J.P. Morgan

Non-QM: The models are coming

Primary non-QM rapidly rebounded this year with $20bn issued YTD. Over the last two years, the sector has been shaped by insurance demand for loans. The securitization share dropped from 67% in 2022 to 54% last year. Insurance demand has been strong, however, the collateral in securitizations has not materially changed from 2023. Bank statement and DSCR investor loans still make up roughly a third of monthly originations, with ‘other’ doc programs contributing another third ( Figure 148). The share of purchase loans, which have, on average, 10%-pts in higher LTVs than cash-out refis, has increased. Meanwhile, the share of low FICO (<700) collateral has dropped. Borrower FICO scores have, in fact, increased to 740. Additionally, DSCR ratios have not seen further decreases this year and the share of no ratio loans have held steady at 10-15%. Non-QM loans have significant reserve requirements. Reserves have been decreasing since 2021, but the average reserve account for purchase loans is still relatively high at $300k.

Figure 148: Bank statement and DSCR investor loans still make up roughly a third of monthly originations, with ‘other’ doc programs contributing another third

Origination ($bn) by doc type

Source: J.P. Morgan, CoreLogic

Non-QM delinquencies have been increasing since mid-2022. Recent vintages have been underperforming versus seasoned vintages at the same loan age. Generally, this has been evident across all FICO bands, but low credit borrowers have seen the highest rise in DQ rates. ‘Other’ doc types also tend to have the highest DQ rates, followed by bank statement and DSCR investor loans. Some have attributed rising DQs to insurance portfolios absorbing significant a supply of whole loans. Insurance portfolios tend to have a cleaner credit box than the broader securitization market, but it is tough to prove that adverse selection is happening in deals as we do not have access to insurance loan portfolios and the credit profile in non-QM deals has not materially worsened. Along with rising DQs, foreclosures have also been increasing. Loans are typically reported as 90+ DQ for 2-3 months before being referred to foreclosure, and roughly 1% of ‘21 and ‘22 vintage loans were reported to be in foreclosure as of the June remit ( Figure 149). The average FICO of borrowers in foreclosure is 700, and loans stay in foreclosure for about 2-5 months.

Figure 149: Foreclosures have also seen an increase, with 1% of ‘21 and ‘22 vintage loans reported in foreclosure as of the June remit

Delinquency status of outstanding 2021 to 2023 vintage non-QM loans

Source: J.P. Morgan, CoreLogic

While delinquencies have been increasing, investors have not been particularly concerned about losses. Figure 150 shows the share of ever 60+ DQ loans paid off by loan origination year. A vast majority of ever 60+ DQ loans paid off without a severity. We have identified only ~250 loans that liquidated with a loss in non-QM since 2015. Average severities have been 10-15%. We do not anticipate current DQs to translate into losses, but we look out for any pro-rata deals potentially failing the DQ trigger, which would result in classes below AAA turning sequential. Deals fail the DQ trigger when 6m average of 60+ DQs and past 12 months of mods, combined, exceed 10-20%. This threshold increases over time. At this point, no deal is close to the DQ trigger failure threshold.

Figure 150: A vast majority of ever 60+ DQ loans paid off without a severity. We have identified only ~250 loans that liquidated with a loss in non-QM since 2015

% of loans which paid off w/ and w/o severity

Source: J.P. Morgan, CoreLogic

Prepayment s-curves have significantly flattened in non-QM. Borrowers with high coupons have not been as reactive to incentive as historical performance might have suggested. Higher coupons have been printing less than the pricing assumption of 25 CPR while lower coupons have been slower than what investors have priced in. Discount ‘22 vintages have been paying under 8 CPR. We notice less of a difference in prepays among doc type in ‘22 vintage loans ( Figure 151). Meanwhile, DSCR investor loans are paying a few CPR slower in the recent ‘23 vintage ( Figure 152).

Figure 151: We notice less of a difference in prepays among doc type in ‘22 vintage loans

2022 vintage S-curves in the 2023-2024 period with incentive calculated over the conforming rate

Source: J.P. Morgan, CoreLogic

Figure 152: Meanwhile, DSCR investor loans are paying a few CPR slower in the recent ‘23 vintage

2023 vintage s-curves in the 2023-2024 period with incentive calculated over the conforming rate

Source: J.P. Morgan, CoreLogic

Prepayment and call assumptions have significantly varied across the investor base as models are not as robust here as in agency MBS or jumbo 2.0 given the short history of the sector. We have been working on a non-QM prepayment model (with our Quantitative Research team) over the past several months. Figure 153 shows preliminary results of our model across the coupon stack. The model structure is similar to our other prepayment models where rate incentives are combined with pool level credit characteristics to generate prepayment, default, and severity vectors. The model is calibrated on non-QM performance since 2015.

In higher coupons, the model is projecting slower prepays than the pricing assumption. AAAs in 9% WAC deals show ZVs at 150bp area vs. 100bp of OAS to the modeled call scenario. Lower coupons, meanwhile, show slower prepays than the market assumption of 8 CPR with 30-60bp of OAS, but these deals do not carry much call option cost. Our model prefers mid-stack 5-6% WAC deals that offer low to mid-100bp of OAS to the modeled call scenario. Note that these are still preliminary model results, subject to minor adjustments.

Figure 153: Preliminary results of our non-QM model show mid-stack 5-6% WAC deals offering low to mid-100bp of OAS to the modeled call scenario

Using J.P. Morgan non-QM prepayment model

Source: J.P. Morgan

CRT: Tenders and calls limit supply

CRT issuance has reached $5bn YTD. We anticipate the GSEs selling another $4bn by year-end, bring total supply in line with 2023. The GSEs have continued to tender bonds that do not provide any capital benefit - Freddie conducted two tender offers this year, while Fannie has done one. We anticipate a couple more rounds of tenders by year-end, which should result in marginally positive net supply for the sector.

CRT collateral in the second half of the year should look largely the same as earlier this year. Purchase loans, which have higher LTV and DTI ratios, still account for the majority of agency originations ( Figure 154). LTV and DTI ratios have seen only a marginal decline ( Figure 155), while the share of >40 DTI loans is still near an all-time high. Interestingly, new origination credit scores, which had gradually declined from pandemic highs, continued to increase and are now close to 760. Additionally, WACs will continue to increase in deals. While collateral remains largely unchanged from earlier in the year, the GSEs brought a structural change to both programs with the introduction of the bullet-like A1 note. This change better aligns the CRT structure with the GSE capital rule.

Figure 154: Purchase loans, which have higher LTV...

Weighted avg. FICO and LTV for CRT eligible collateral in agency MBS

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

Figure 155: ...and DTI ratios, still account for the majority of agency originations, which will be referenced in CRT in 2H24

Weighted avg. DTI and % affordable programs for CRT eligible collateral in agency MBS

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

As with non-QM, delinquencies have been increasing in CRT. 2022 vintage loans have underperformed, driven by weakness in low FICO and cash-out borrowers, but we have seen DQs slowly decrease or moderate slightly over the last few months ( Figure 156). Our view on CRT credit remains unchanged. The GSEs have robust delinquency workout options, which should support borrowers in distress; the recently revised Flex Modification waterfall will offer a more uniform 20% reduction in payment. We do not anticipate material losses from changes in the modification waterfall since term extension - the first step in the waterfall and not a loss to CRT - achieves the targeted reduction in payment for the majority of currently delinquent borrowers ( Figure 157).

Figure 156: 2022 vintage loans have underperformed, driven by weakness in low FICO and cash-out borrowers, but we have seen DQs slowly decrease or moderate slightly over the last few months

STACR 30+ DQ (%) by loan vintage

Source: J.P. Morgan, Freddie Mac

Figure 157: We do not anticipate material losses from changes in the modification waterfall

Annualized modification losses (%) assuming 100% of currently 30+ delinquent loans are modified

Source: J.P. Morgan, Freddie Mac

FHFA approves limited closed-end second lien pilot program

The FHFA has approved the Freddie proposal for closed-end second liens, but the program will have a significantly smaller footprint than initially anticipated. Freddie can buy $2.5bn of closed-end second liens over the next 18 months. Fannie has not made any announcements yet, but is likely to follow Freddie’s lead. The FHFA received significant pushback on the proposal from industry participants who cited concerns for expanded landing, inflationary pressures and the GSEs pushing out private market with the introduction of a second lien program. This makes sense, as a limited GSE footprint in closed-end seconds is supportive for PLS. As it is, issuance of HELOC and CES deals has already exceeded last year’s total at $6.4bn YTD ( Figure 158). New issuers like Pimco’s BRAVO, Vista Point and Goldman, have entered the space this year, and we anticipate more to come in 2H24. More closed-end second lien deals than HELOCs have been securitized thus far versus last year ( Figure 159). We maintain our forecast of $10bn.

Figure 158: New issuers like Pimco’s BRAVO, Vista Point and Goldman, have entered the space this year, and we anticipate more to come in 2H24

Original balance ($bn) of HELOC and CES securitization by shelf and year

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

Figure 159: More closed-end second lien deals have been securitized thus far versus last year

Original balance ($bn) of HELOC and CES deals by collateral type and year

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

Issuance: We look for $110bn of gross supply for 2024

We have seen $61bn of gross issuance YTD, with the bulk of supply coming from RMBS 2.0 and non-QM. We raise our issuance forecast to $110bn for 2024, slightly higher versus the $97bn we had previously ( Figure 160). This means an additional $48bn of supply through the rest of the year.

Figure 160: We raise our issuance forecast to $110bn for 2024

Issuance with FY 2024 projections across mortgage credit ($bn)

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

RMBS 2.0 issuance came in higher than expected in 1H24, particularly in agency investor. Better execution versus GSE delivery, coupled with a pull back from banks, helped add to the securitization supply. That said, supply will continue to be limited by low origination volumes and the strong insurance bid for whole loans. We raise our 2.0 forecast $3bn to $24bn. Non-QM has driven the bulk of supply YTD. Spreads have tightened meaningfully over the last few months, and we expect the pace of issuance to continue. We raise our non-QM forecast $5bn to $35bn. In CRT, Freddie and Fannie have each issued 3-4 deals YTD, in line with their issuance calendar, and both have modified their deal structures to include the bullet-like A1 class to better align with the ERCF. There have been no MI CRT deals issued so far. We keep our issuance forecast the same and expect another $4bn through year-end. In SFR, we have seen several refi deals issued YTD, driving supply up. Issuers also continue to work through their pipeline of properties acquired in ‘22, pointing to sufficient supply remaining for a steady pace of securitizations. We raise our forecast to $7bn. Home equity continues to be an area of focus. We have seen $6bn issued YTD, already exceeding the ‘23 total of $5bn, and the addition of new shelves to the sector. We keep our forecast at $10bn. Lastly, we keep our RPL forecast the same at $12bn and raise our NPL forecast slightly to $3bn.

At the doorstep of the acceptance phase

• CRE transaction cap rate expansion has been muted year-to-date through May, remaining about flat versus 2023 year-end. Levered return on equity (ignoring NOI growth and exit cap rates) has remained low as borrowing rates have remained high while LTV expansion has been limited, suggesting financing remains less accretive to returns relative to the past and this, in turn, has kept transaction activity low. We are still looking for marginal cap rate expansion to 7% by 2024 year-end, driven by more seller capitulation
• The muted transaction activity and lack of a more meaningful cap rate expansion is due to borrowers being able to ‘kick the can’ on maturing debt. Looking at private label CMBS, we see that payoff rates by original maturity dates have sunk to historical lows. The majority of loans that were originally scheduled to mature in 2023 and 2024 (through May) remain outstanding either through servicer mods/extensions or optional extensions that were exercised by borrowers. As such, most of the outstanding loans remain performing with delinquencies still muted though rising
• We see little reason for a material change in the modification/extension dynamic for the remainder of the year. However, as we look past 2024, extended loans re-defaulting remains a risk, particularly for office, where buying time may not have worked to heal wounds
• What about CMBS valuations? Broadly speaking, CMBS is no longer a ‘cheap’ asset class and spreads across the stack appear full, particularly at the top of the stack. Seasoned deals, however, do offer value for investors willing to do the credit legwork including post-mod deals (even office) where a basis reset through paydowns/cash-ins may have delevered the structures
• As for our benchmark on-the-run spread forecasts, at J+99, 10yr conduit CMBS LCF AAAs are unlikely to hit our mid-year target of J+90. Our year-end target was J+85. We revise our year-end 10yr conduit CMBS LCF AAA spread target to J+95. For 10yr Freddie K A2, we keep our year-end spread target of J+48 (J+49 currently). For conduit BBB-s, we also keep our year-end spread target of J+550-600
• The refi success rate for conduit CMBS loans scheduled to mature this year was just 63%. which is about 11pts lower than the refi success rate for 2023 maturities. However, 31% of these loans were refinanced prior to 2024. In fact, most of these prepaid loans were refinanced in 2022 and 2023 as interest rates were increasing
• The vast majority of floating-rate SASB loans that were scheduled to mature this year either exercised embedded extension options or were extended through servicer modifications. We continue to think the extension success rate for floating-rate SASB loans will remain high if sponsor continue to show commitment to the underlying properties.
• During the May remit period, the serious delinquency rate for private label CMBS reached 4.3%, representing a nearly 40bp increase since the end of last year. Office loans have continued to be the marginal driver of the rise in serious delinquencies for private label CMBS loans. Office loan delinquency rates increased by 100bp since the end of the last year to reach 6.7%
• The loan modification rate for private label CMBS has increased significantly since the start of the year to 10.5% due largely to multifamily and office loans
• Almost two years ago, we developed a set of office-related stresses for our CMBS credit model in which we ran discrete but gradual occupancy shocks to estimate loss impact on CMBS office loans. Going into 2023, in our year-ahead Outlook, we turned our 10pt occupancy shock scenario into our Base Case. Since we first crafted these scenarios, we’ve accumulated nearly two years of data on the office market and while the sector is not out of the woods yet, we think we can start to refine our stresses to cater to more local metro area stresses that are playing out. We focus on these differences for refining our office shocks for our revised loss projections
• Here is what we like for the remainder of the year: Front pay conduit CMBS A1s versus short corporates and ABS, FRESB A10Fs versus 10yr Freddie K A2s, Freddie MLs versus comparable muni Housing Finance Agency bonds, CRE CLO AAAs and floating SASB AAAs versus BSL CLO AAAs

Overview and trading themes

CRE transaction cap rate expansion has been muted year-to-date through May, remaining about flat versus 2023 year-end with blended cap rates across the major property types sitting at roughly 6.5%. Levered return on equity (ignoring NOI growth and exit cap rates) has remained low as borrowing rates have remained high while LTV expansion has been limited, suggesting financing remains less accretive to returns relative to the past and this, in turn, has kept transaction activity low ( Figure 163). There is little reason to believe this dynamic changes for the remainder of this year with less Fed cuts expected relative to the start of this year. We are still looking for marginal cap rate expansion to 7% by 2024 year-end, driven by more seller capitulation. As high for longer persists, seller capitulation on fundamentally impaired assets is an inevitability as risks and opportunity costs mount.

Figure 163: Transaction cap rates have not moved much this year and transaction volumes remain muted as high for longer persists. When will more sellers capitulate?

US commercial real estate and multifamily blended cap rates versus levered return on equity (%) and transaction volumes ($bn)

Source: J.P. Morgan, MSCI Real AssetsNote: Levered ROE = 1/(1-LTV) * (Cap Rate – (LTV*Mortgage Rate)). Does not consider future NOI growth or exit cap rates.

The muted transaction activity and lack of a more meaningful cap rate expansion is due to borrowers being able to ‘kick the can’ on maturing debt. Looking at private label CMBS (conduit, SASB, CRE CLOs), we see that payoff rates by original maturity dates have sunk to historical lows ( Figure 164). The majority of loans that were originally scheduled to mature in 2023 and 2024 (through May) remain outstanding either through servicer mods/extensions or optional extensions that were exercised by borrowers. As such, most of the outstanding loans remain performing with delinquencies still muted though rising (see CMBS credit fundamentals outlook). Even lower are actual liquidations of defaulted loans. We see little reason for a material change in the modification/extension dynamic for the remainder of the year. However, as we look past 2024, extended loans re-defaulting remains a risk, particularly for office, where buying time may not have worked to heal wounds. As Figure 165 shows, we saw a substantial portion of CMBS office loans that received modifications post-GFC re-default and ultimately liquidate for losses. A similar dynamic can play out as we step into 2025 and beyond.

Figure 164: The CMBS market is in the middle of an extension wave...

Percentage of private label CMBS loans paid off, outstanding, and liquidated by original maturity year

Source: J.P. Morgan, Trepp

Figure 165: …if the post-GFC modification wave tells us anything, there is substantial re-default risk in office loans that have been extended. This risk emerges as we step into 2025 and beyond

Conduit CMBS office loan cumulative modification rates and percentage of modified loans that are outstanding, paid off, or liquidated post-mod by deal vintage (%)

Source: J.P. Morgan, Trepp

What about CMBS valuations? Broadly speaking, CMBS is no longer a ‘cheap’ asset class and spreads across the stack appear full, particularly at the top of the stack ( Figure 166). The top of the conduit CMBS stack should largely avoid meaningful extensions in the medium term as refinance success rates remain high enough but spreads no longer look compelling to single-A corporates. The middle part of the stack (AAs and single-As) is more exposed to extension risks and has closed in on similar duration REIT bonds in recent months ( Figure 167). Conduit BBB-s remain optically wide to high yield CRE REIT bonds but with good reason. These bonds have thin enhancements and real money does not seem convinced of their relative value, meaning hedge funds and private credit remain the anchor investors for this portion of the capital stack ( Figure 168). We think new participants are needed for these cohorts for lower IG mezz spreads to move materially tighter from this point forward. Seasoned deals, however, do offer value for investors willing to do the credit legwork including post-mod deals (even office) where a basis reset through paydowns/cash-ins may have delevered the structures.

Figure 166: Valuations of high quality benchmark CMBS bonds look full

Spreads to Treasuries (bp)

Source: J.P. Morgan

Figure 167: Conduit AS to single-As have compressed materially as well. The relative value to IG REIT bonds is less interesting

Spreads to Treasuries (bp)

Source: J.P. Morgan

Figure 168: While conduit CMBS BBB-s have arguably been the best performer this year in the CMBS complex, they remain wide to high yield CRE REITs but with good reason

Spreads to Treasuries (bp)

Source: J.P. Morgan

As for our benchmark on-the-run spread forecasts, at J+99, 10yr conduit CMBS LCF AAAs are unlikely to hit our mid-year target of J+90. Our year-end target was J+85. We revise our year-end 10yr conduit CMBS LCF AAA spread target to J+95. For 10yr Freddie K A2, we keep our year-end spread target of J+48 (J+49 currently). For conduit BBB-s, we also keep our year-end spread target of J+550-600.

Here is what we like for the remainder of the year:

• Front pay conduit CMBS A1s versus short corporates and ABS. At J+92, the conduit CMBS A1s are compelling relative to 1-3yr corporates (I+71) or 3yr consumer ABS AAAs (I+48-66). These bonds have no appreciable risk of being extended, let alone taking writedowns. Tranches are small, however, so can be difficult to source.
• FRESB A10Fs versus 10yr Freddie K A2s. Voluntary prepayment speeds have slowed a lot for the FRESB market and older vintages are exposed to NYC rent regulated stories that are troubled. However, on-the-runs at J+81 offer 32bp spread pickup to its bigger brother 10yr Freddie K A2s currently at J+49. Like Freddie K A2s, FRESB A10Fs have 10yr legal hard finals so cannot extend. The program is smaller so lacks the kind of secondary liquidity that Freddie Ks enjoy but the spread pickup is hard to ignore.
• Freddie MLs versus comparable muni Housing Finance Agency bonds. Freddie Mac’s multifamily Tax Exempt program is a growing one that offers tax exemption benefits to more traditional municipal bond investors as well as those that can benefit from tax exemption. While new issue pricing has moved substantially tighter in the last few deals, they can still offer some yield pickup to comparable municipal Housing Finance Agency bonds. See here for our municipal strategists discussion and here for our primer.
• CRE CLO AAAs and floating SASB AAAs versus BSL CLO AAAs. At 1m term SOFR + 175bp+ on tier 1 CRE CLO AAAs and new issue SASB AAAs, spreads are attractive to BSL CLO AAAs pricing at 3m term SOFR + 142 at new issue. CRE CLO credit issues are real but AAAs should not experience losses although servicer extensions are a real possibility. The risk is bad headlines can cause spread widening. We would view those as buying opportunities. Well sponsored floating SASB has weathered rate volatility better than most expected with high optional extension success rates. New issue offers higher coupons and fundamentally sound collateral for the most part with a lot of new issue supply to pick from. We think this is a good time to add.

CRE sector outlooks

We recently hosted our periodic J.P. Morgan Commercial Real Estate Update call (replay here and accompanying slides here) in which our J.P. Morgan Asset Management colleague, Luigi Cerreta, did a great job of providing views on real estate fundamentals across the major property sectors. We largely agree with his views, so we recommend our readers to listen to the replay of our call.

CMBS credit fundamentals outlook

Refi success rates

Through the first five months of the year, the refi success rate for conduit CMBS loans scheduled to mature this year (excluding defeased loans) was 63%, which is about 11pts lower than the refi success rate for 2023 maturities ( Figure 169). By far the worst performing sector in the conduit CMBS space has been office, as just 36% of office loans scheduled to mature this year have successfully refinanced ( Figure 170). The refinance success rate for fixed rate SASB loans scheduled to mature this year is currently 37%, which is on par with the refinance success rate for 2023 maturities but much lower than the historical average. However, it is important to note that more than half of the fixed rate SASB loans that were originally scheduled to mature this year are performing loans that have been modified with maturity date extensions. As far as floating rate SASB and CRE CLO loans, the refinance success rate remains low, but these loans have embedded extension options, which have been heavily utilized. For a detailed breakout of loan resolutions for the private label CMBS market, please refer to Figure 189in our appendix. 

Figure 169: The year-to-date refi success rate for conduit CMBS loans is 11pts lower than the 2023 full-year rate

Refinance success rates by CMBS product type, as of May 2024

Source: J.P. Morgan, Trepp

Figure 170: Unsurprisingly, the office loan refi success rate continued to underperform that of all other property types

Refinance success rates for conduit CMBS by property type, as of May 2024

Source: J.P. Morgan, Trepp

While the overall refi success rate for conduit CMBS loans maturing in 2024 is currently at a historically low level, it is important that we understand the timing of the refinancings that have already occurred. As Figure 171 highlights, 31%, or $4.7bn, of conduit CMBS loans that were scheduled to mature through the first five months of the year were paid off prior to 2024. When we dig deeper into this specific cohort of loans, we notice that nearly 68% of them were refinanced as interest rates were increasing in 2022 and 2023 ( Figure 172). Borrowers who refinanced during this period likely experienced enough property price appreciation on their investments in the years prior that they were still incentivized to refinance even as rates were increasing. However, when looking by property type, there has been a relatively limited amount of early office refinancings during this period, highlighting how limited the refinancing prospects have been for office loans given the sector’s ongoing fundamental challenges.

Figure 172: About 68%of loans that were scheduled to mature this year and were prepaid were paid off in 2022 and 2023 when interest rates were increasing

Percentage of prepaid loans that were scheduled to mature in 2022, 2023, and 2024 by payment year

Source: J.P. Morgan, Trepp

There is also a clear delineation in debt yields based on when the loans are refinanced. Private label CMBS loans that were refinanced prior and during the year that they were scheduled to mature, generally have a higher debt yield than those that were refinanced in the years following their original maturity year ( Figure 173). This has been a consistent trend going back to 2013. During this period, debt yields on loans that were prepaid have been 423bp greater, on average, than the debt yield on loans that were paid off after their original maturity year. The gap between prepaid loans and loans that were liquidated is even greater having averaged a 700bp difference during this same period. For a detailed breakout of debt yields of loans that have been refinanced and liquidated by property type going back to 2021, please see Figure 190 in our appendix.

Figure 173: Generally, conduit CMBS loans with higher debt yields receive better refinancing outcomes

Average NCF debt yield of private label CMBS loans (excluding defeased loans) that were either paid off early, on-time, late or liquidated by maturity year, as of May 2024

Source: J.P. Morgan, Trepp

Throughout last year and this year, we repeatedly noted that conduit LCF AAA extension risk for shorter vintages are very limited considering how well refi success rates are holding up. And, while there has been debate about whether including early payoffs in the overall refi success number is fair, as we demonstrate above, a lot of what paid off early for 2023 and 2024 scheduled maturities did so in 2022 and 2023 when financial conditions increasingly become tighter. That said, the outlook for mid-to-longer duration vintages (2017+) can become more challenging as office exposures increase appreciably in these vintages. Barring a substantial rate rally and/or a significant recovery in office fundamentals, LCF AAA extension risk for these vintages are more material than earlier vintages. It seems that at least some of this is priced into the market as conduit LCF AAAs with very high concentrations to office are still trading 40bp+ wider to those with low concentrations ( Figure 174).

Figure 174: Conduit LCF AAA extension risk if more substantial for mid-to-longer duration vintages with more substantial office exposures barring a substantial rate rally and/or a significant recovery in office fundamentals as those maturities approach

Spreads to Treasuries (bp), 2018-2023 vintage conduit CMBS LCF AAAs, outliers determined using inter-quartile ranges are excluded each day

Source: J.P. Morgan, Pricing Direct

Floating-rate SASB extensions

Through May 2024, about $67bn of floating rate SASB loans reached their maturity dates, most backed by mixed-use properties. Like the 2023 maturities, the vast majority of these loans were extended. Overall, about 68% of these loans exercised embedded extension options and 6% received servicer modifications that included maturity date extensions ( Figure 175). Many of these loan modifications required the property’s sponsor to inject capital into the deal. For example, UBSCM 2018-NYCH, a deal backed by a portfolio of upper midscale hotels, received a 9-month extension after agreeing to a $10mn principal paydown. In the case of CSMC 2021-GATE, the sponsor is required to deposit nearly $45mn into a leasing reserve in exchange for a 1yr maturity date extension with a 1yr extension option. However, some loans that were expected to mature during the first five months were not extended. 12% of these loans are currently delinquent and 3% are performing as they work with special servicers on resolutions. Another 10% have been paid off and 1% have been liquidated.

Figure 175: Most floating-rate SASB loans that reached their scheduled maturity dates this year have been extended

Balance and percentage split of floating-rate SASB loans that have either been liquidated, paid off, became delinquent, optionally extended, have been modified, or are performing by property type, as of May 2024

Source: J.P. Morgan, Trepp

When we look at the data by property type, unsurprisingly, office is the clear underperformer. Office loans have one of the lowest extension success rates at just 48%, which is only behind the lodging sector and the “other” sector. Please note that the “other” sector includes relatively small asset classes like manufactured housing and healthcare. However, the “other” and lodging sectors have the highest payoff rates by far, at 35% and 25%, respectively. Some of the hotel deals that were paid off this year like BX 2018-BILT and GWT 2019-WOLF have been part of the Blackstone’s refinancing wave. In the case of office, there have been limited payoffs and high delinquencies. It is likely that the extension and payoff success rates for office will remain low relative to other property types.

Currently, there is about $161bn of outstanding floating SASB deals. Using either their 2023YE NOI DSCR or their underwritten DSCR, we estimate that about 60% of these deals have NOI DSCR less than 1.25x and 29% have a NOI DSCR less than 1.0x ( Figure 176). While DSCR levels are important in determining whether a loan can be refinanced or extended, sponsor commitment is just as important, if not more important. If sponsors are committed to these properties and are willing to inject capital into these deals, then the extension success rate for these deals will likely remain high.

Delinquencies, special servicing, and modifications

The serious delinquency rate (60d+ including non-performing matured and FC/REO) for private label CMBS continued to rise this year due, in large part, to maturity defaults. During the May remit period, the serious delinquency rate for private label CMBS reached 4.3%, representing a nearly 40bp increase since the end of last year ( Figure 177). The special servicing rate for private label CMBS loans increased 128bp the through the year so far to reach 7.4%. Also, it is important to note is the rise in loan modifications. The loan modification rate for private label CMBS is now at 10.5%, the highest level on record. When we exclude forbearance, the modification rate is 7.74%, a 111bp increase since the beginning of the year.

Figure 177: The serious delinquency, special servicing, and loan modification rate for private label CMBS loans have all climbed this year

60d+ delinquency rate (including NP matured and FC/REO), modification rate, modification rate excluding temporary forbearance, special servicing, and foreclosure rate for private label CMBS, as of May 2024

Source: J.P. Morgan, Trepp

Office loans have continued to be the marginal driver of the rise in serious delinquencies for private label CMBS loans ( Figure 178). Office loan delinquency rates increased by 100bp since the end of the last year to reach 6.7%. The serious delinquency rate for all other property types have increased slightly except for retail loans, which experienced a 50bp decrease. Given the strength of the consumer, it is not surprising to see the continued decline of delinquencies in the retail sector since the end of the pandemic.

Figure 178: Office loans have been the marginal driver for the increase in the private label serious delinquency rate

Private label CMBS serious delinquency rates (60-day+ including foreclosed/REO and non-performing matured loans) by property type, as of May 2024

Source: J.P. Morgan, Trepp

Serious delinquencies increased across all the different CMBS deal types during the first five months of the year. The conduit CMBS serious delinquency rate is up 30bp this year to 4.8% while the SASB serious delinquency rate increased just 10bp to 3.7% ( Figure 179). Please note that the SASB delinquency rate experienced a 70bp month-over-month decrease during the May remit period due to a few loan resolutions that we previously wrote about here. However,the serious delinquency rate for CRE CLOs experienced a significant increase since the start of the year as it is now at 4% after beginning the year at 2.7%. As an almost entirely floating-rate market, CRE CLO borrowers have struggled with rising debt costs as a result of the increase interest rates over the past couple of years. This, along with asset value deterioration and limited revenue growth, has resulted in increased stress in the CRE CLO market. Meanwhile, Freddie K delinquencies remain de-minimis.

Figure 179: The serious delinquency rate for all CMBS product types has increased this year

60d+ delinquency rate (including NP matured and FC/REO) by CMBS product type, as of May 2024

Source: J.P. Morgan, Trepp

As we mentioned earlier, the loan modification rate for private label CMBS has increased significantly since the start of the year. The marginal drivers of this increase have been office and multifamily loans. Due to the increased stress in the CRE CLO market, we have seen a dramatic rise in the modification rate for multifamily loans ( Figure 180). Since the beginning of the year, the multifamily loan modification rate has increased by 278bp to 8.73%. Of the multifamily loans to be modified this year, the vast majority were securitized in CRE CLO deals. As we have highlighted previously, CRE CLO managers have increasingly used loan modifications to preserve their existing CRE CLO vehicles. Unsurprisingly, we also saw a large uptick in office loan modifications as well. The office loan modification rate is now at 7.5% after beginning the year at 5.7%.

Figure 180: Modification rates have increased significantly for both multifamily and office loans this year

Private label CMBS modification rates (60-day+ including foreclosed/REO and non-performing matured loans) by property type, as of May 2024

Source: J.P. Morgan, Trepp

We expect delinquencies to continue to rise but due to increased loan modifications, we revise down our serious delinquency rate forecast for private label CMBS to 5.5% from 5.7% ( Figure 181). By deal type, we revise down our conduit CMBS serious delinquency forecast from 6.8% to 5.5%. We also must revise up our CRE CLO serious delinquency rate to 5.2% from 3.5% given the increase stress in this market.