Cross Sector P. White, L. Wash
Treasury yields are lower and spreads stable as the economic data support a gradual economic cooling and as Fed communications confirm expectations of a near-term rate easing cycle. We continue to look for a 50bp cut at the next FOMC meeting and the August employment report will be key for determining the pace of easing. We modestly revise higher our HG gross issuance forecast to $1.5tn for 2024, leaving $400bn of issuance left this year.

Governments J. Barry, P. White, A. Borges, L. Wash
Chair Powell’s dovish tone supports our forecast of consecutive 50bp cuts at each of the next two meetings, but the August employment report is still two weeks away, valuations are no longer cheap, and dealer positioning is elevated, so we stay neutral duration for now. Powell’s comments support further long end steepening, and we recommend holding 5s/30s steepeners to express our core view. Most front-end curves appear too steep relative to their drivers, but 3s/7s offers value for those who want to position for a quicker and more aggressive easing cycle. We discuss the outlook for T-bill supply and valuations. Hold tactical 5y5y inflation swap longs.

Interest Rate Derivatives S. Ramaswamy, I. Ozil, P. Michaelides, A. Parikh
The first cut is all but certain at the Sep meeting, but pace and extent remain uncertain. Policy uncertainty amplifies delivered volatility - stay bullish on vol. Tbill net issuance should turn negative in coming weeks, and term funding premium appears too low - maintain front end spread wideners and position for a flatter 5s/30s spread curve. Initiate 7s/20s weighted spread curve steepeners to benefit from relative value. Forward curve steepeners are preferable to longs in the Reds.

Short-Term Fixed Income T. Ho, P. Vohra
RRP balances have stabilized around $300bn, with limited room to decrease further. Both domestic bank borrowing in the O/N unsecured market and US banks’ participation in the repo market remain minimal, suggesting reserves have not transitioned from abundant to ample yet.

MBS and CMBS J. Sim
Mortgages had a relatively unexciting week, with lower coupons tightening modestly and the rest of the stack close to unchanged. We review 2023 CMBS property financial data, which indicates that net cashflow growth has slowed materially compared to prior years. Multifamily, in particular, has seen operating expenses increase significantly due largely to labor-related costs.

ABS and CLOs A. Sze, R. Ahluwalia
ABS spreads held firm on the week, but current strong technicals are scheduled to be tested in September due to a likely packed new issue ABS calendar, along with macroeconomic data and FOMC schedules.

Investment-Grade Corporates E. Beinstein, S. Doctor, N. Rosenbaum, S. Mantri
HG bond spreads were stable last week despite declining yields. August on track to be the 4th consecutive month of positive returns, supporting demand. We revise up our FY24 supply forecast to $1.5tr leaving $400bn to go in 2024.

High Yield N. Jantzen, T. Linares
High-yield bond yields are at a low since August 2022 and spreads are 60bp inside early August’s high. High-yield bond yields and spreads decreased 12bp and 5bp over the past week to 7.53% and 364bp. And the HY index is up +1.21% in August which boosts 2024’s gain to +6.11%.

Municipals P. DeGroot, Y. Tian, R.Gargan
Cheaper municipal valuations and the expectation for lower absolute rates as the Fed’s easing cycle proceeds, could mean that the next couple of months offer the best opportunity to buy bonds of the year and possibly the rate cycle.

Emerging Markets L. Oganes
In EM fixed income, we are MW GBI-EM local rates, CEMBI and EMBIGD. EM bond flows were -$304mn (-0.08% of weekly AUM, up from -$1.4bn).

Source: J.P. Morgan

See you in September

• Economics: The preliminary benchmark payroll revisions lowered total payroll growth for the year through March 2024 by 818k. Initial claims remained relatively unchanged at 232k for the August employment report reference week
• Treasuries: Chair Powell’s dovish tone supports our forecast of consecutive 50bp at each of the next two meetings, but the August employment report is still two weeks away, valuations are no longer cheap, and dealer positioning is elevated, so we stay neutral duration for now. Powell’s comments support further long end steepening, and we recommend holding 5s/30s steepeners to express our core view. Hold tactical 5y5y inflation swap longs
• Interest Rate Derivatives: Policy uncertainty remains elevated with OIS markets placing significant weight on both deep cutting and shallow cutting scenarios. Yields should remain sensitive to new economic and policy information which supports jump risk and elevated realized volatility. We remain bullish on volatility in short expiries. We remain positioned for swap spread curve flattening and we recommend initiating exposure to a flatter 5s/30s swap spread curve
• Short Duration: RRP balances have stabilized around $300bn, with limited room to decrease further. Both domestic bank borrowing in the O/N unsecured market and US banks’ participation in the repo market remain minimal, suggesting reserves have not transitioned from abundant to ample yet
• Securitized Products: Mortgages had a relatively unexciting week, with lower coupons tightening modestly and the rest of the stack close to unchanged. We think mortgage credit spreads can remain rangebound for the rest of the year. We see value in upper IG mezz CMBS and 5yr Agency CMBS bonds
• Corporates: HG bond spreads were stable last week despite declining yields. August on track to be the 4th consecutive month of positive returns, supporting demand. We revise up our FY24 supply forecast to $1.5tn leaving $400bn to go in 2024
• Near-term catalysts: July personal income (8/30), July JOLTS (9/4), Aug employment (9/6), Aug CPI (9/11), Aug PPI (9/12)

Over the past week Treasury yields are lower and spreads roughly stable across both corporate and mortgage markets as the economic data continue to reinforce only gradual cooling in the economy, the issuance calendar remains light, and Fed communications confirm expectations of a rate cutting cycle beginning in September. Indeed there was little fanfare from what sparse data was released this week. The preliminary benchmark revisions lowered total payroll employment growth for the year through March 2024 by 818k (0.5%), decreasing job growth by about 70k/month to a ~175k monthly run rate over the period. It is difficult to say exactly how the Fed will interpret this data, given that the revision cuts off in March 2024 and the implications for more recent months are not fully clear (see US: Benchmark brings big downward revision to (lagged) jobs, Abiel Reinhart, 8/21/24). At the same time initial jobless claims for the August employment reference week remained roughly unchanged over the week at 232k, but are 13k lower than the July reference week, and claims continue to follow a similar summer seasonal path as the last few years (see US: Good news from mostly stable jobless claims, Abiel Reinhart). The data support our view that the economy is slowing with low risks of a recession.

Instead, market participants were focused on Fed communications, which laid the ground work for an upcoming easing cycle. Even before the release of the July payrolls report, the July FOMC minutes revealed that “the vast majority” of the FOMC were ready to cut by the September meeting. Notably “many” participants noted the risks around cutting too little or too late while only “several” thought reducing rates too soon or too much could risk a resurgence in inflation (see July FOMC minutes highlighted labor market risks, Abiel Reinhart). Meanwhile this week Fed speak offered support for a “gradual” or “methodical” pace of cuts. However markets were most keen for Chair Powell’s Friday Jackson Hole speech where his comments skewed dovish. The Chair stated “the time has come” to adjust interest rates and that the Committee does “not seek or welcome further cooling in labor market conditions.” Unsurprisingly, our monetary policy NLP read Powell’s speech as more dovish than his press conference at the July FOMC meeting, and his most dovish speech since the fall of 2021 ( Figure 1, also see Fed, Powell: Review and Outlook, Joseph Lupton, 8/23/24). Whether the FOMC cuts 25bps or 50bps at the September meeting still hinges on the next monthly employment report, though Powell’s comments arguably lower the bar for the Fed to deliver a 50bp cut next month (see Cuts are coming, Michael Feroli, 8/23/24).

Figure 1: Chair Powell’s 2024 Jackson Hole speech scored as his most dovish in nearly 3 years

Jerome Powell Hawk-Dove history, using speeches, testimonies, and press conferences

Source: J.P. Morgan* For methodology, see Listen up: Upgrading J.P.Morgan’s central bank NLP machine, Joseph Lupton and Dan Weitzenfeld, 7/2/24

Figure 2: As one might expect, yields have tended to be much more sensitive to Fed-speak in periods of significant policy uncertainty

Size of the Nth percentile absolute change in the Fed sentiment index* on days with one or more Fed-speak events, grouped by policy clarity or uncertainty regime**, for various selected values of N from 10 to 90. Past 6 months; bp

Source: J.P. Morgan, Bloomberg Finance, L.P.
* The Fed sentiment index is computed as the cumulative sum of yield changes in 5-year note Treasury futures in the 15-minute period following the first Fedspeak headline on Bloomberg. Fedspeak is defined as any speech, FOMC meeting or FOMC minutes. ** Days on which the absolute difference between the total weight on deep cut scenarios and shallow cut scenarios was between 0.25 and 0.75 are classified as days characterized by Policy Uncertainty, while all other days are deemed to have Policy Clarity. For more details see Interest Rate Derivatives

As we look ahead, we continue to forecast 50bp cuts at the September and November meetings followed by 25bp per meeting thereafter. Our forecast is rooted in both the resumption of the disinflationary process which remains broad based across components, together with ongoing loosening in labor markets, which has reduced the risk prices remain sticky. With the beginning of the easing cycle less than a month away and OIS forwards pricing in 103bp of easing through YE24 compared to our own projection of 125bp, this would indicate there is value to being long duration. However we remain patient as growth data has turned more favorable in recent weeks and we do not think the debate surrounding the pace of easing will be settled before the August employment report. Moreover valuations look close to our fair value estimate, and in an environment in which term premium is biased higher, we think it’s unlikely that valuations overshoot to the rich side of fair value. Thus, we remain neutral on duration but continue to favor 5s/30s steepeners as a medium-term expression of our bullish view (see Treasuries).

Turning to inflation markets, intermediate breakevens remain cheap to our fair value framework, though the model residual narrowed on Friday, given a dovish shift from Fed Chair Powell. In recent months, the Fed’s wavering confidence in disinflation and thus a heightened focus on labor market data created a less supportive backdrop for TIPS, with real yields demonstrating lower sensitivity to nominal yields in rallies than in sell-offs. However, we argue that the backdrop is becoming more supportive for TIPS once again. Along the curve, the intermediate sector continues to offer value, and forward-starting swaps avoid the negative carry associated with spot breakeven positions. With the fundamental backdrop turning more supportive and valuations still cheap, albeit somewhat less cheap than a week ago, we maintain tactical long exposure in 5Yx5Y inflation swaps (see TIPS).

Despite the near certainty of a cut in September, the pace and extent of easing remain uncertain with OIS markets placing significant weight on both deep cut and shallow cut scenarios. In this period of elevated policy uncertainty, yields should remain highly sensitive to both new economic and policy information; Figure 2 shows the percentile moves in yields around Fedspeak events and indicates that markets are prone to greater sensitivity during periods of elevated policy uncertainty. Such a backdrop is conducive to jump risk and supportive of elevated realized volatility. Thus with significant jump risk going into August payrolls and beyond, we remain bullish on volatility in short expiries. We also remain positioned for a flattening in the swap spread curve. Front-end spreads should be biased wider as SOFR and GC rates normalize and net T-bill supply turns negative in September whereas long-end spreads should narrow against the backdrop of rising term premium. Therefore, we now recommend initiating exposure to a flatter 5s/30s swap spread curve (see Interest rate derivatives).

Turning to Fed’s balance sheet, the July FOMC minutes did not reveal much new information. The SOMA manager noted that repo rates had risen due to increased demand for financing Treasury securities and balance sheet normalization. Meanwhile, the manager concluded that reserves remained abundant but emphasized that the staff would continue to closely monitor developments in money markets. We agree, as we note domestic banks remain largely absent from overnight funding markets, whether in fed funds or Eurodollar deposits, while bank call reports show little increased net reverse repo activity YTD, as the GC/IORB spread remains subdued ( Figure 3). If we were to see increasing participation of banks within overnight markets or an increasing percentage of overnight Treasury repo transacted at or above IORB, this would signal reserve scarcity is approaching. To that end, bank reserves have been steady at ~$3.4tn while ON RRP balances have fallen by ~$700bn YTD as MMFs have shifted towards T-bills and non-Fed repo. We think there is limited room for ON RRP to move much below the current low $300 level, given operational considerations for MMFs to meet unexpected liquidity needs, and we think QT can continue through year-end (see Short term fixed income).

Figure 3: Domestic banks have remained largely absent from the overnight market, both in fed funds and Eurodollar deposits

Fed funds and Eurodollar deposits (OBFR-FF) volumes (LHS, $bn) versus Foreign OBFR volumes as a % of total (RHS,%)

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

Mortgage rates are roughly unchanged across the stack, with lower coupons modestly tighter. Along with most spread products, valuations have recovered from the local wides seen early in the month; still, with mortgages continuing to offer historically compelling yields vs. corporates, we think the technical support for the product from money managers can be maintained. We continue to prefer UIC conventionals and are underweight 4.5s and 5s. Primary rates have fallen to roughly 6.44% on GSE loans and 5.47% for VA IRRRLs, but we have not seen a surge in refi applications, likely as originators picked up some low hanging fruit earlier this month ( Figure 4, see Mortgages). Mortgage credit spreads similarly retraced much of the widening seen earlier in the month with AAA non-QM back to 130bp in new issue. We expect spreads to remain rangebound over the rest of the year as the macro outlook remains challenging while the election season quickly approaches. Primary activity has been very active this summer, but stepped down over the month from $12bn to $7bn new issue RMBS in August, and we think primary markets will be light in the next week (see RMBS).

ABS spreads largely held firm over the month with relatively muted supply and as markets focused on opportunities in higher-yielding segments. Deals were marketed at wider initial price targets given the volatility this month, but final pricing was able to track within these margins given heightened oversubscriptions. Light primary dealer issue activity lent support to the secondary market this week, allowing dealers to off-load some inventory. Private credit student loan ABS tightened by 10-15bp across the capital stack. Meanwhile performance on our bankcard ABS index remains solid. Charge-offs have increased 6bp this year to 2.2% and these low losses coupled with high yields resulted in record high quarterly average excess spreads of 19.9%. Strong credit performance, robust demand, and the back-drop of a no-recession base case scenario should sustain spreads, though we note the potential for a likely heavy new issue ABS calendar alongside increased macro volatility come September (see ABS).

Issuance was similarly relatively light within CMBS markets over the past week. Benchmark CMBS spreads were slightly tighter, and valuations look fair relative to their corporate and mortgage comparables. Conduit IG mezz tightened more, with AS to single-As tighter 8-16bp over the week. We continue to see value in upper IG mezz bonds (AS and AA) which should be more insulated from near-term risks to the data but still look cheap relative to corporate comps and to the spread curve ( Figure 5). We also still see 5yr Agency CMBS as leaning cheap to current coupon Treasury OAS and ZVs. Our analysts review 2023 conduit CMBS property financial data, which indicates that net cashflow growth has slowed materially compared to recent years, with multifamily revenues in particular suffering from rising labor-related expenses and significantly rising insurance premiums (see CMBS).

Figure 4: Primary rates have edged lower, but lock volume is off the early August highs

GSE rate/term refi lock rates (%) and number of locks (#);

Source: Optimal Blue

Figure 5: Conduit CMBS upper IG mezz are cheap on the spread curve

Residuals from linear regressions* of 10yr conduit CMBS AS and AA Treasury spreads to those of adjacent tranches; bp

Source: J.P. Morgan
*Modeled spreads linearly regresses 10yr conduit CMBS AS spreads against 10yr conduit CMBS LCF AAA and 10yr conduit CMBS AA spreads over the last 10 years (R2 = 96%,SE = 8.8bp) and 10yr conduit CMBS AA spreads against 10yr conduit CMBS LCF AS and 10yr conduit CMBS A spreads over the last 10 years (R2 = 97%,SE = 12bp)

Turning to corporate credit, JULI spreads remained rangebound over the past week within a tight 109-112bp range and the stability of spreads in recent weeks comes despite declining yields and heavy corporate issuance. Supply has been active with $103bn priced in August, surpassing the historical average of $85bn. Instead we think spreads have benefited from the reversal of the equity sell-off following more encouraging economic data, seasonally unusually strong liquidity, and retail fund inflows. Indeed trading volumes averaging $35bn/day are 67% higher than the historical August pace, bucking the typical seasonal trend of late summer depressed liquidity, while fund flows have been robust averaging $5.3bn/week this month, likely driven by positive returns. Looking ahead, we recently revised higher our gross issuance forecast modestly to $1.5tn. Specifically, we expect $400bn in supply over the remainder of the year, a step down from the current pace, and this would leave roughly $150bn of supply come September according to historical seasonals ( Figure 6). In an environment where economic growth is slow but positive and the Fed lowers rates, we would expect spreads to remain near current levels. Meanwhile if the Fed does deliver on our projected easing forecast, we expect the percentage gap in yield between HG credit vs Treasuries to be near its widest YTD and this could spur the marginal duration buyer towards credit over rates (see Corporates).

Down the capital stack, high yield spreads are 5bp tighter over the week while yields are at their lowest levels since August 2022 amid a Goldilocks narrative on growth without inflation. Month-to-date, BB spreads are tighter by 3bp while B and CCC spreads are now wider by 9bp and 33bp in August, respectively. Meanwhile the percentage of leveraged loans trading above par recovered to 42% amid improving growth sentiment, moderating retail outflows, robust CLO origination, and light capital activity. We note that the ratings profile for HY bonds is near its most benign on record, as the universe of single Bs contracted 440bp this year versus an expansion of 420bp for BBs and only a 20bp increase in CCCs. Conversely, the rating mix for loans is considerably weaker than bonds given the aggressive capital market activity earlier this year and a historic downgrade wave (see High Yield).

After a firm start to the year, US CLO supply has subsided in recent months with $12bn of new supply in August MTD compared to $24bn in May. We anticipate there will be a strain on new issuance over the coming months given challenging loan sourcing, rising recession risks, and Fed rate cuts. Spreads could soften as investors proceed with caution around market volatility. Meanwhile refi/reset activity has strong momentum with July the third heaviest month of all time at $27bn. Indeed there are 102 US managers that have refi’d/reset a deal this year compared to 42 last year as volumes have picked up (see CLOs).

Figure 6: We revise higher our gross issuance forecast to $1.5tn and expect $400bn for the remainder of the year. HG supply has exceeded the historical pattern in almost every month this year, and we expect this to continue through YE

4yr average gross issuance versus 2024 gross issuance*; bn

Source: J.P. Morgan
*The grey dashed bars are estimates based on historical pace and not our forecast

Figure 7: Cross sector monitor

Current levels, change since 8/16/24, 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
* 8/22/24 levels for iBoxx Euro HG, AA taxable munis, gold, and brent oil; 8/23/24 levels for all others

Figure 8: YTD returns on various fixed income indices; %

Source: J.P. Morgan

• Fed’s Powell: “The time has come for policy to adjust”
• Limited guidance leaves the debate over a 25bp or 50bp cut wide open; we continue to expect the latter...
• … but that will depend on the August jobs report after mixed labor market news this week
• Expected soft core PCE inflation in focus next

The path to rate cuts is now wide open. Fed communications this week—both the minutes to the late July FOMC meeting and Powell’s speech at the Jackson Hole conclave—were quite direct in signaling that policy rates will be reduced at the September meeting and presumably further in the fall and winter. Neither the minutes nor Powell shed light on the market’s current debate over a 25bp or 50bp cut next month, which makes sense given there’s almost a full monthly slate of data between now and the next meeting.

We still look for a 50bp cut in September. With risks to the inflation mandate ebbing and risks to the employment mandate growing, we think it makes sense for policy to be closer to neutral, which is likely at least 150bp below where rates are now. That said, the inconclusive nature of the Fed’s recent guidance with respect to pace suggests that the next employment report will dictate whether our forecast is realized or whether the Fed takes a more gradualist approach to easing.

Staying on the topic of the labor market, the latest news has been mixed. Once again, weekly jobless claims remained tame, consistent with a healthy jobs market (Figure 1). However, the benchmark level of employment in March 2024 was revised down by a historically sizable 818,000 jobs. Data from the Business Employment Dynamics survey through December 2023 indicate that the BLS’s preliminary estimates of job growth attributable to the net of business births and deaths may have been overstated, and so it is possible that overstatement could extend beyond March 2024.

Next week promises to be relatively quiet before the fireworks of the data deluge after Labor Day. We expect next Friday’s core PCE figure will be up 0.12%m/m, which would leave the year-ago number unchanged at 2.6%.

Cuts are coming

Chair Powell’s remarks at the Jackson Hole forum left little room for doubt that the Fed will soon be cutting rates, as he frankly put it: “The time has come for policy to adjust.” He didn’t tip his hand on the expected size of that adjustment. Instead, he said “the timing and pace of rate cuts will depend on incoming data, the evolving outlook, and the balance of risks.” While noncommittal on size, this was more open-ended guidance than what we’ve heard from some other Fed officials, who have recently spoken about “methodical” rate cuts. Overall, Powell’s remarks leaned dovish, as he stressed the FOMC’s attentiveness to labor market conditions. Powell did not express worry about the current state of the labor market, but also offered that the Fed does “not seek or welcome further cooling in the labor market.”

We continue to look for the Fed to start off the easing cycle with a 50bp cut next month, though that will depend in part on the August jobs report. As Powell noted, “The current level of our policy rate gives us ample room to respond to any risks we may face, including the risk of unwelcome further weakening in labor market conditions.” If the August jobs report validates the July weakness, then we believe they should quickly make use of that ample room.

While the labor market headlines stuck out, Powell also expressed growing confidence in inflation returning to 2 percent. In fact, much of the body of the speech was a retrospective on inflation developments over the past four years.  In that, he attributed “much of the increase in inflation to an extraordinary collision between overheated and temporarily distorted demand and constrained supply.”

Earlier in the week, the minutes to the July 30-31 FOMC meeting (which occurred just before the release of the weak July labor report) similarly skewed dovish and showed a committee more focused on labor market risks than the possibility of a renewed jump in inflation—even if they still judged current inflation as somewhat elevated. While “several” had contemplated cutting 25bp at the last meeting, a “vast majority” were prepared to cut at the September meeting, and “many” saw policy as currently restrictive.

Like many others, participants were trying to understand how much of the unemployment rise reported at the time of the meeting was due to rising labor supply and how that rise should be weighed against less concerning news from new jobless claims and layoff rates. On balance, however, they did not appear to be discounting the increase. Moreover, the Committee was anticipating the possibility of downward payroll revisions, and “several assessed that payroll gains may be lower than those needed to keep the unemployment rate constant with a flat labor force participation rate.”

The labor market giveth, and taketh away

In an otherwise fairly quiet data week, jobless claims were comforting in their recent stability but the unexpectedly large preliminary benchmark revision to payrolls cast a broader pall over markets. Data on initial claims were collected during the reference week for the payroll survey, and do not point to another step down in job growth for August. It’s unclear how much signal for upcoming job growth one should take from the 818K preliminary downward revision to March 2024 payrolls—the final revision will take place early next year—but it helped narrow some of the unusually wide gap between the establishment and household measures of employment. The revisions suggest CES job growth likely averaged around 175K for the 12 months through March, rather than the 242K per month on average as reported, although history suggests a slightly smaller final revision is common. That said, the possibility that payrolls are still being overstated since March 2024 remains: the birth-death model looks to still be adding more jobs than recent (albeit lagged) data from the Business Employment Dynamics suggest is likely (Figure 2).

Mixed signals from home sales

After a pop near the start of the year, most housing market data have trended back down over the past several months. For July, new home sales surprised with a strong 10.6% pop, whereas existing home sales modestly disappointed with a more paltry 1.3% rise. Most importantly, the July single-family new home sales release came with substantial upward revisions to prior months that suggests a better trajectory for this year (Figure 3). New home sales tend to be volatile but they also tend to lead existing home sales (as the former reflects signed contracts while the latter are not recorded until a contract is closed). However, other leading indicators of existing home sales, such as mortgage purchases applications or pending home sales, continue to point to existing home sales remaining stuck at low levels. Inventories of existing homes remain low, although as interest rates come down that could stimulate sales later this year. By contrast, inventories of new homes remain quite elevated, which is restraining price growth and likely weighing on housing starts.

Divergent flash PMI details

In a sign of stability, the all-industry flash PMI for August (54.1) was nearly unchanged from July (54.3). But beneath the headline, the details offered divergent messages. First, looking across sectors, the manufacturing index fell to 48.0. This marks the second consecutive monthly decline, reversing much of its gradual improvement since early 2023 and taking the index to its lowest level since December. The weakness in the manufacturing PMI appears broad-based across survey components. By contrast, the business activity headline for services firmed marginally to 55.2 in August, with new business activity also well supported. While there has been some speculation that services activity would cool as the post-pandemic recovery normalized and the prospect of lower interest rates gave a lift to manufacturing, in reality the services sector remains the main impetus for the continued economic expansion.

Second, the employment flash PMI sent a concerning signal ahead of the August jobs report, with the all-industry measure sliding to its lowest reading in this cycle (outside of April) of 48.9. Weakness was evident in both the manufacturing and services PMIs, although the latter has been quite choppy of late. That said, we do not tend to take much of a signal from this report for the upcoming payroll print as monthly moves in the employment PMI are not particularly informative for the jobs report.

Next week we will get inflation readings in the monthly PCE report but not much labor market news other than the weekly claims. The most important piece of information will arrive after Labor Day with the August jobs report, which might incorporate a very modest post-Beryl addition to payrolls.

Excerpted from, United States Data Watch , Michael Feroli, August 23, 2024

Your time has come

• The Treasury curve bullishly steepened this week supported by the July FOMC meeting minutes and Chair Powell’s speech at Jackson Hole, as both indicated a step in the dovish direction
• Given Powell’s concerns over further cooling in labor markets and his acknowledgement that there is significant room to lower policy rates, we think this supports our forecast of 50bp cuts in September and November, before downshifting to 25bp thereafter
• At face value, with money markets pricing a slower pace of cuts than our forecast, this suggests we should be overweight duration. However, we think it will require a weak August employment report to price a more dovish path for the Fed. With valuations no longer cheap and dealer balance sheets somewhat bloated, we prefer to be patient before adding duration
• In the meantime, this also supports a steeper curve at the long end, as the first cut approaches and the Fed begins to ease consistently. At the long end, steepeners with shorts anchored in 30s offer more value than those anchored in 20s, and we continue to favor 5s/30s to express our core view
• Front-end steepeners should outperform if the Fed eases aggressively, but 2s/5s appears too steep in our framework and incurs significantly negative carry. On margin, 3s/7s looks more attractive for those who want to position for the risk of a quicker and more aggressive easing cycle
• The period of positive net T-bill issuance is set to come to a seasonal close in early-September as we approach the corporate tax date. Longer-dated T-bills are trading at the cheap end of their recent range, but with issuance set to turn negative, dealer positions at average levels and our derivative colleagues biased toward wider swap spreads, we see scope for richening over the near term

Market views

The Treasury curve bullishly steepened this week, supported by the July FOMC minutes and Chair Powell’s speech at Jackson Hole, both of which read dovishly. The July minutes indicated the Committee’s confidence in inflation returning to 2% had increased, while in June, participants had emphasized the need to see more data. There was also heightened focus on the risk that continued easing in labor market conditions could transition to a more serious deterioration, and it’s no surprise in this context that participants saw the upside risks to the inflation outlook as moderating, while the downside risks to labor market weakening were increasing. Finally, “the vast majority observed that, if the data continued to come in about as expected, it would likely be appropriate to ease policy at the next meeting” (see July FOMC minutes highlighted labor market risks, Abiel Reinhart, 8/21/24).

Chair Powell advanced the discussion further at Jackson Hole on Friday, saying “The time has come for policy to adjust. The direction of travel is clear, and the timing and pace of rate cuts will depend on incoming data, the evolving outlook, and the balance of risks.” This was a bit of a departure from speakers earlier this week who framed the expected pace of cuts as either “gradual” or “methodical.” What’s more, the balance of risks has shifted to the labor markets, as Powell offered, “We do not seek or welcome further cooling in labor market conditions.” Finally, the Chair stated, “The current level of our policy rate gives us ample room to respond to any risks we may face.” As we discussed last week, the real Fed funds rate is approximately 2.75%, a full 200bp above the Fed’s real longer-run dot. Using either the Fed’s longer-run dot (with data back to 2012) or the Holston-Laubach-Williams r* estimate, policy rates were negligibly restrictive in 2019, and only about 100bp restrictive in mid-1995 before the first cut, suggesting policy rates are highly restrictive ( Figure 9). Unsurprisingly, our monetary policy NLP read Powell’s speech as more dovish than his press conference at the July FOMC meeting, and his most dovish speech since the fall of 2021 ( Figure 10, also see Fed, Powell: Review and Outlook, Joseph Lupton, 8/23/24).

Figure 9: Policy is highly restrictive right now, underscoring Chair Powell’s point that there is “ample room to respond”...

Real Fed funds rate*, real longer-run median Fed dot, and Holston-Labach-Williams natural rate of interest rate; %

Source: J.P. Morgan, Federal Reserve bank of New York, Bloomberg Finance L.P
* Fed funds target less core PCE oya

Figure 10: ...and today scores as his most dovish speech in nearly 3 years

Jerome Powell Hawk-Dove history, using speeches, testimonies, and press conferences

Source: J.P. Morgan* For methodology, see Listen up: Upgrading J.P.Morgan’s central bank NLP machine, Joseph Lupton and Dan Weitzenfeld, 7/2/24

On the week, 2-, 5-, 10-, and 30-year yields declined 15bp, 12bp, 9bp, and 5bp, respectively. Considering the recent sensitivities of most points along the curve, the steepening is more pronounced than would be expected by the decline in front-end yields. Figure 11 shows that adjusting for the 3-month beta to changes in 10-year yields, the short- to intermediate sector strongly outperformed this week. However, we think some mean reversion is at work here. The primary driver of most curve slopes is the market’s medium-term Fed policy expectations, and late last week, the curve had flattened more than would be implied by the move in 2-year yields. With these moves, the curve appears more appropriately priced after adjusting for the level of 2-year yields ( Figure 12).

Figure 11: The front-end outperformed on a beta-adjusted basis this week...

Actual change in yields by sector over the past week minus the change implied by the 3-month beta with 10-year yields; bp

Source: J.P. Morgan

Figure 12: ...but we think this reflects mean reversion from flat levels

Residual of 5s/30s Treasury curve regressed on 2-year Treasury yields*; bp

Source: J.P. Morgan
*6-month regression. R-squared = 87.2%, SE = 4.2bp

Looking ahead, we continue to forecast 50bp cuts in September and November and think Chair Powell’s comments today leave the door open to this outcome. Our forecast is driven by two different factors. First, the disinflationary process has resumed in a more broad-based fashion than observed in 2H23, considering there has been disinflation in core goods, services ex-shelter, and shelter. Second, the ongoing loosening of the labor markets has reduced the risk that services prices remain sticky to the upside. OIS forwards are pricing in approximately 103bp of easing over the balance of this year, versus our 125bp forecast, and this would indicate there is value to being long duration at current levels. However, the August employment report is still two weeks away, and validation of the weak July data will be needed to price in more aggressive easing this year. This is particularly important, as final demand remains strong, and our US FRI has moved higher, as we’ve recently revised up growth forecasts for the current quarter ( Figure 13). Thus, we do not think the debate about pace will be settled until the release of the August employment report on September 6.

Figure 13: Labor markets will be key in determining the pace of cuts, as our growth forecasts have been revised higher in recent weeks

J.P. Morgan US growth Forecast Revision Index;

Source: J.P. Morgan

Figure 14: 10-year Treasuries are no longer cheap after adjusting for their fundamental drivers, and we think an overshoot to the downside is unlikely in an environment in which term premium is biased higher...

Residual of J.P. Morgan 10-year Treasury fair value model*; 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 = 97.8%, SE = 20.1bp

Valuations are relatively neutral as well, as yields in the intermediate sector no longer appear cheap to their fundamental drivers: Figure 14 shows 10-year Treasuries are fairly valued when controlling for their market’s medium-term Fed policy, inflation, and growth expectations. Given that the first expected cut is now less than a month away, one could argue there is room for yields to overshoot to the downside, but as the figure shows, Treasuries have not traded significantly rich relative to their drivers since “immaculate disinflation” was the market’s narrative in late 2023. Moreover, the current environment, in which the Treasury market continues to grow more rapidly than its traditional constituents of demand, term premium is biased higher, which should make it hard for valuations to overshoot the rich side of fair value (see In the eye of the beholder, 9/12/23). We can observe this dynamic to a certain extent in the evolution of primary dealer positions, which have increased nearly 30% in coupons this year, as more price sensitive demand has yet to appear ( Figure 15). Given this backdrop, we continue to recommend patience before adding duration, but given the sensitivity of the Fed’s reaction function to the labor markets, we will reevaluate as we get closer to the release of the August employment data.

Figure 15: ...which we think is reflected in the considerable rise in dealer positions over the last year

Primary dealer positions in Treasuries (excluding T-bills); $bn

Source: Federal Reserve Bank of New York

Powell’s comments and his implicit acknowledgment that policy is significantly restrictive validate the view that there is significant room to normalize policy rates from current levels, which should support further broad steepening. The periods just prior to and around the beginning of an easing cycle are supportive of long-end led steepening, and we believe this leaves further room for the long end to steepen from current levels. Moreover, our medium-term valuation framework also suggests there is value in steepeners at current levels as well: Figure 16 shows that the market’s medium-term Fed policy and inflation expectations, as well as the Fed’s ownership share of the Treasury market have been highly significant in explaining the variation in most Treasury curve pairs over the last three years. 5s/30s is flagging as fairly valued after controlling for these drivers, while curve pairs with the long leg anchored in 20s appearing too steep, reflecting the recent cheapening of this sector. Thus, we continue to recommend 5s/30s steepeners as the medium-term expression of our bullish view as we move into an easing cycle.

Finally, most front-end pairs appear too steep and incur significantly negative carry as well. Thus, though steepeners like 2s/5s do outperform in more aggressive easing cycles, both valuations and the carry implications make these trades less attractive locally. However, among front-end pairs, 3s/7s looks too flat in this framework, and offers value for those who want to position for a quick and aggressive easing cycle.

T-bill update

Turning to the front end, T-bill net issuance has been positive for the last 7 weeks, and the stock of outstanding T-bills has risen by $290bn over the period. However, this period is coming to an end, and we forecast net T-bill issuance will turn negative in two weeks, in line with traditional seasonals ahead of the corporate and estimated tax deadline on September 16^th. Figure 17 shows our weekly T-bill net issuance and TGA projections for the coming months, assuming Treasury hits the $5bn cap at each of the four cash management buyback operations scheduled for the September 5-25 period. Looking at the details, we expect one more week of outsized net issuance, as the $35bn CMB maturing September 12 settles next week, and look for 5 weeks of negative T-bill issuance into early October. Following this period, we project more than $125bn in net issuance between mid-October and mid-November.

Figure 17: T-bill issuance should turn negative in early-September, ahead of the corporate tax date

Net T-bill issuance* (lhs; $bn) versus TGA balance (rhs; $bn)

Source: US Treasury, J.P. Morgan
*Grey bars are J.P. Morgan forecasts

Figure 18: Longer-dated T-bills are near the cheaper end of their recent ranges…

T-bill/SOFR matched maturity spreads, current levels with 3-month statistics; bp unless otherwise indicated

Maturity	Current	3m chg	3m high	3m low	3m avg	Percentile
4-week	-7.1	-3.6	-3.2	-13.3	-6.6	37%
8-week	-4.5	-2.4	2.0	-8.8	-4.1	52%
13-week	-1.7	0.6	2.3	-6.5	-2.8	75%
17-week	1.0	2.1	6.8	-2.9	-0.2	84%
26-week	5.2	6.5	8.0	-7.1	-0.1	87%
52-week	7.6	10.4	11.2	-3.3	3.0	84%

Source: J.P. Morgan

This issuance pattern should be marginally supportive for valuations in the weeks ahead. However, we would note the upcoming reductions should be concentrated in shorter-maturity T-bills. This is important because while 3- to 12-month T-bills are now trading near the cheaper end of their 3-month ranges, this is not the case for 1- and 2-month bills, which remain closer to the middle of the ranges that have persisted over the last three months ( Figure 18). Taking a step back, the cheapening in longer-dated bills should not be surprising given the recent widening in GC/OIS spreads, which we tend to view as a proxy for dealer financing costs ( Figure 19). As discussed earlier, primary dealer positions in Treasuries have risen sharply YTD, which helps explain the cheapening of Treasuries on spread to SOFR, but T-bill inventories have remained relatively stable, and most of this accumulation has occurred in coupon securities ( Figure 20). Moreover, our colleagues in derivatives strategy continue to recommend swap spread wideners at the front end. Given these dynamics, we see some room for longer-dated bills to richen from current levels over the coming weeks.

Figure 19: ...and have cheapened alongside the widening of GC/OIS spreads

12-month T-bill/ SOFR matched-maturity spread (lhs) versus 3-month GC repo/ OIS spread; bp

Source: J.P. Morgan

Figure 20: Primary dealer Treasury holdings have risen over the last year, but not in T-bills

Primary dealer positions in Treasuries*; $bn

Source: Federal Reserve Board of New York *Latest data as of as of 8/14/2024
Dashed lines represent 1-year averages

Trade recommendations

• 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: 17.4bp)

• 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: -6.7bp)

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

-   Stay short 23% risk, or $28mn notional of T 4.375% Jul-26s
- Stay long 100% risk, or $40.2mn notional of T 4.125% Jul-31s
- Stay short 84% risk, or $23mn notional of T 3.875% Aug-34s
- (US Treasury Market Daily, 8/21/24: P/L since inception: -0.4bp)

• The spread between market-implied terminal easing rate and market-implied neutral has traded in a range over the past 24-months, where the market has been unwilling to price the policy rate more than 30-50bp through implied neutral. In the past three cycles, the front-end pricing breakouts coincided with trends to higher jobless claims...
• ...Regardless of the macro-fundamental driver, the bullish consolidation richer than the 4.115-4.20% initial support zone continues to favor a positive bias and technical view that looks for an eventual breakout in late-2024 or early-2025….
• Along with the shape of the OIS curve and 2-year note pattern neckline at 3.555-3.65%, we are also focused on the -20bp through -15bp area on the 2s/5s curve chart. That has marked a barrier for multiple quarters to what we believe to be a cycle bottom pattern. We see a asymmetric risk-reward setup heading into the latter months of the year and early-2025 in favor of a breakout and transition to an aggressive steepening trend.
• The 5-year note continues to bullishly consolidate after the mid-summer acceleration to lower yields. We expect the 3.885-4.02% support zone to cap yields through late summer. A break through 4.15-4.20% is needed to derail the bullish pattern structure at this point. The rally stalled near the 3.50% Oct-Apr equal swings objective and the associated channel trend line, now at 3.475. We suspect that area will act as resistance for a few weeks but give way later in the year or into 2025.
• The 10-year note yield bullishly consolidates behind the 3.625% Apr 2023 78.6% retrace and in front of key support at 4.02-4.14%. We expect those two zones to define the trading range into the early fall. The next zone of chart resistance sits at 3.22-3.248%, targets for later this year or early 2025.
• Look for continued 5s/30s curve range trading below the 59bp Mar 2021 50% retrace and above the 28-32bp support zone in the weeks ahead. We think that range will consolidate the breakout from the 2022-2024 base pattern and give way to a more aggressive steepening trend in the months ahead.
• 10-year TIPS breakevens widen back into the 212-220bp resistance zone following the dovish Fed messaging. We are looking for breakevens to falter in that area again and for that reversal to set the market up for renewed tightening pressure into the Sep-Oct period of seasonal weakness for risky markets. Look for a retest of the 200bp area over that period.

US

Treasuries continue to trade in what look like bullish consolidation patterns across the entire duration spectrum as the Fed validates the recent dovish repricing. The Aug flight-to-quality rotation helped propel the 2-year note back toward the low yields realized early last year in the aftermath of two bank failures. Likewise, that as the money market curve adjust to nearly match the dovish extremes achieved over the past two years, with more than 230bp of total eases priced for the cycle, and the terminal easing rate priced more than 30bp through market-implied neutral. Across the board, front-end pricing tagged chart levels that now define the necklines of multi-quarter reversal patterns. Beyond the importance those key thresholds represent on the charts from a pattern perspective, the threshold on money market curve charts also represented key bifurcation zones during the cycles since the mid-1990s and when we have OIS markets to analyze. The spread between 5y5y OIS rate levels (market-implied neutral) and the forwards priced terminal easing rate has a loose inverse correlation with initial jobless claims historically ( Figure 22). While claims have been well-anchored below 260K during most of the post-COVID recovery and expansion, the spread between market implied neutral and the expected terminal easing rate has also been range bound. In each of the three slowdowns over the observed period, the money market curve not only tracked claims higher, but the market transitioned to an aggressive trending regime once it broke out.

Figure 22: The spread between market-implied terminal easing rate and market-implied neutral has traded in a range over the past 24-months, where the market has been unwilling to price the policy rate more than 30-50bp through implied neutral. In the past three cycles, the front-end pricing breakouts coincided with trends to higher jobless claims...

Initial jobless claims (sa, 1000s, left scale), spread between 5y5y OIS rate and OIS implied terminal easing rate (bp, inverted, right scale)

Source: J.P. Morgan

Regardless of the driver, the bullish consolidation on duration and yield curve charts favor an eventual release to lower yields and bull steepening breakouts in the months ahead. For the 2-year note, that bullish consolidation has unfolded richer than the 4.115% Jan yield low and 4.18-4.20% Fibonacci retracement confluence ( Figure 23). Another group of levels rests at 4.30-4.35%. The more mixed data in recent weeks leaves a backup to that second zone as a possibility, especially if the u-rate retraces some of the recent rise with the early-Sep data. The market would need to back up through the 4.595-4.66% breakout area to completely nullify the bullish trend dynamics, levels we do not see getting tested going forward. To lower yields, tactical bulls would regain traction with a break through 3.88% for another run at the 3.65% Aug extreme.

Figure 23: Regardless of the macro-fundamental driver, the bullish consolidation richer than the 4.115-4.20% initial support zone continues to favor a positive bias and technical view that looks for an eventual breakout in late-2024 or early-2025...

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

Source: J.P. Morgan, CQG

Bigger picture, the weekly and monthly time-frame momentum divergence buy signals that triggered in 2022-2023, the yield curve breakouts, late-cycle signals across other markets like USD/JPY, and recent bullish impulses to lower yield levels leave us viewing the entire 2023-2024 price action as large bullish reversal patterns ( Figure 24). We expect the front-end to pressure those levels into the fall and eventual break richer into early-2025. Initial breakout resistance rests at the 3.285% Jul 2020 38.2% retrace and 2.98% 2018 cycle cheap.

Figure 24: … The lower-frequency bullish momentum divergence signals that triggered through late-2022 and 2023, the recent impulsive bullish price action, and steepening breakouts on yield curve charts all suggest the entire late-2022 through 2024 2-year note pattern represents a large bullish trend reversal.

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

Source: J.P. Morgan, CQG

Similar to the key thresholds noted above, the 2s/5s curve has kept the area between -25bp and -12bp under steepening pressure this summer. That not only marks the upper-end of the 2023-2024 base pattern, but was also the 1989, 2000, 2006, and 2019 cycle troughs ( Figure 25). Given the dynamics identified in the money market curve pricing in prior cycles, we suspect a 2s/5s curve breakout above that area will mark the transition to an aggressive steepening trend, so we’ve added a stop-in strategy to our list of long duration proxies in the trade strategy section. On the downside, the recent corrective flattening held above support at the -33.5bp Jun 50% retrace, -34bp 50-day moving average, -36.5bp Jun 61.8% retrace, and -37.5bp 200-day moving average. We see the -41.5bp Jul 2023 38.2% retrace as a likely floor going forward.

Figure 25: Along with the shape of the OIS curve and 2-year note pattern neckline at 3.555-3.65%, we are also focused on the -20bp through -15bp area on the 2s/5s curve chart. That has marked a barrier for multiple quarters to a what we believe to be a cycle bottom pattern. We see a asymmetric risk-reward setup heading into the latter months of the year and early-2025 in favor of a breakout and transition to an aggressive steepening trend.

2s/5s curve, daily closes; bp

Source: J.P. Morgan, CQG

Figure 26: The 5-year note continues to bullishly consolidate after the mid-summer acceleration to lower yields. We expect the 3.885-4.02% support zone to cap yields through late summer. A break through 4.15-4.20% is needed to derail the bullish pattern structure at this point. The rally stalled near the 3.50% Oct-Apr equal swings objective and the associated channel trend line, now at 3.475. We suspect that area will act as resistance for a few weeks but give way later in the year or into 2025.

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

Source: J.P. Morgan, CQG

Figure 27: The 10-year note yield bullishly consolidates behind the 3.625% Apr 2023 78.6% retrace and in front of key support at 4.02-4.14%. We expect those two zones to define the trading range into the early fall. The next zone of chart resistance sits at 3.22-3.248%, targets for later this year or early 2025.

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

Source: J.P. Morgan, CQG, CME

Figure 28: The 30-year bond bullishly consolidates richer than the 4.33-4.345% Jun-Jul pattern riches and in front of the 4.40-4.50% confluence of moving averages. Look for further coiling around the 4.00% area in the weeks ahead.

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

Source: J.P. Morgan, CQG

Figure 29: Look for continued 5s/30s curve range trading below the 59bp Mar 2021 50% retrace and above the 28-32bp support zone in the weeks ahead. We think that range will consolidate the breakout from the 2022-2024 base pattern and give way to a more aggressive steepening trend in the months ahead.

5s/30s curve, daily closes; bp

Source: J.P. Morgan, CQG

Figure 30: 10-year TIPS breakevens widen back into the 212-220bp resistance zone following the dovish Fed messaging. We are looking for breakevens to falter in that area again and for that reversal to set the market up for renewed tightening pressure into the Sep-Oct period of seasonal weakness for risky markets. Look for a retest of the 200bp area over that period.

10-year TIPS breakevens, daily closes; bp

Source: J.P. Morgan, CQG

This report was excerpted from Global Fixed Income Technical, Jason Hunter, August 23, 2024

Fed shift sets more supportive foundation for TIPS

• Intermediate breakevens remain cheap to our fair value framework, though the model residual narrowed on Friday, given a dovish shift from Fed Chair Powell
• In recent months, the Fed’s wavering confidence in disinflation and thus a heightened focus on labor market data created a less supportive backdrop for TIPS, with real yields demonstrating lower sensitivity to nominal yields in rallies than in sell-offs. However, we argue that the backdrop is becoming more supportive for TIPS once again...
• ...Chair Powell stated that his “confidence has grown” on the inflation front, while the Committee is now committed to dialing back policy restriction to support strong labor markets
• Along the curve, the intermediate sector continues to offer value, and forward-starting swaps avoid the negative carry associated with spot breakeven positions. IOTAs broadly remain tight across the curve, although we acknowledge this is particularly so at the front end, and we think there’s room for the IOTA curve to flatten
• With the fundamental backdrop turning more supportive and valuations still cheap, albeit somewhat less cheap than a week ago, we maintain tactical long exposure in 5Yx5Y inflation swaps
• Inflation swap daily average trading volumes have risen 25% to $4.6bn YTD, now comprising 23.5% of TIPS trading volumes. Much of this rise has occurred in recent months and coincides with a significant increase in the asset swap share of overall inflation swap transaction volumes, to 9% in the current quarter

Market views

Over the past week, TIPS initially underperformed, but more than recovered over the second half of the week, leaving breakevens 3bp wider across the curve, as jobless claims were stable, Thursday’s 30-year TIPS reopening auction was well-received, and Chair Powell delivered dovish remarks at Jackson Hole. Specifically, the supply cleared just 0.8bp cheap to pre-auction levels, and end-user demand rose to 93.1%, the second highest takedown on record ( Figure 31). With little in the way of inflation news this week, focus was squarely on the labor market data and commentary from Fed officials, including the minutes from the July FOMC meeting. Though the preliminary estimate of the 2024 benchmark for the establishment survey showed total payroll employment was 818K lower than previously thought, a larger revision than we estimated, it is difficult to say exactly how the Fed will interpret it, given that the revision cuts off in March 2024 and the implications for more recent months are not fully clear (see US: Benchmark brings big downward revision to (lagged) jobs, Abiel Reinhart, 8/21/24). Meanwhile, jobless claims rose only slightly in the week ending Aug 17, which is the payroll survey week. Relative to the July survey week, initial claims are down 13k and the four-week average is essentially unchanged, so claims at least aren’t suggesting another step down in job growth for August (see US: Good news from mostly stable jobless claims, Abiel Reinhart, 8/23/24). Overall, the economic data leave us comfortable with the view that data is softening only slowly, with recession risks still relatively low.

At the same time, Fedspeak this week confirmed the Committee is ready to begin cutting at the next meeting, even if the pace of cuts is likely to depend on the incoming data. The July FOMC meeting minutes skewed dovish and showed a Committee more focused on labor market risks than the possibility of a renewed jump in inflation. Importantly, “several” participants had contemplated cutting 25bps at the last meeting, and it was already the case that a “vast majority” were ready to cut at the September meeting, even before the July employment report. Over much of the week, more timely commentary from Fed speakers were perceived as somewhat more hawkish, offering support for a “gradual” or “methodical” pace of cuts. However, Powell's remarks at the Jackson Hole Symposium were the main event of the week, as the Chair stated “the time has come” to adjust interest rates and that the Committee does “not seek or welcome further cooling in labor market conditions.” Whether the FOMC cuts 25bps or 50bps at the September meeting still hinges on the next monthly employment report, though Powell’s comments arguably lower the bar for the Fed to deliver a 50bp cut next month (see Cuts are coming, Michael Feroli, 8/23/24).

Figure 31: End-user demand at Thursday’s TIPS auction was the second highest on record

End-user demand* at 30-year TIPS auctions (%, LHS) versus tail (bp, RHS)

Source: J.P. Morgan
*Reflects direct and indirect bidder participation

Figure 32: Breakevens remain cheap to our fair value estimate, though the residual narrowed late in the week

Residual on J.P. Morgan 10-year breakeven fair value model*; bp

Source: J.P. Morgan
*1m-forward, seasonally-adjusted breakevens are regressed on the J.P. Morgan Commodity Curve Index (JPMCCI) as well as its square and the 3mx3m/15mx3m OIS curve; regression over the last 7years; R2=93%, SE=9bp

Against this backdrop, breakevens outperformed versus our fair value framework over the week: Figure 32 shows 10-year breakevens now appear roughly 7bp too narrow on this basis, with the residual nearly cut in half over recent sessions. As we look ahead, we think there is room for inflation to outperform further versus other macro markets, particularly given the shift in Fed reaction function articulated by Chair Powell on Friday. As we argued last week, the performance of TIPS as we approach the upcoming easing cycle is likely to depend crucially on the context in which the Fed is lowering policy rates (see TIPS Strategy, 8/16/24).

Taking a step back, recall that the Fed reaction function earlier this year made TIPS an attractive way to position long duration, given asymmetric risk/reward. Indeed, for much of the past year the primary risk to Fed expectations was further unanticipated stickiness in inflation, while on the flip side, “immaculate disinflation” suggested the Fed could ease more quickly, even without significant weakening in economic growth. In this environment, real yields exhibited a higher beta to nominal yields in rallies than in sell-offs ( Figure 33). However, this relationship shifted into the summer, as the Fed’s confidence in disinflation waned, which put greater emphasis on evidence of cooling labor markets ( see TIPS Strategy, 6/14/24). Thus, unsurprisingly, Figure 34 shows that the asymmetry of nominal yield betas shifted in the other direction in recent months, punctuated by the sharp underperformance of TIPS as yields fell in the aftermath of the July employment report. Similarly, with uncertainty on labor market and growth back in the driver’s seat, this shift also corresponded with a negative stock-bond correlations reasserting itself. At this juncture, however, with Chair Powell stating that his “confidence has grown that inflation is on a sustainable path back to 2%” and the Committee intending to dial back policy restraint “to support a strong labor market,” TIPS are likely once again to be a more attractive instrument for investors looking to add duration.

Figure 33: In contrast to the dynamic observed for much of the last year, when real yields were more sensitive to nominal yields in rallies...

Rolling 3-month betas of changes in 5-year real yields with respect to changes in nominal yields in nominal rallies versus sell-offs;

Source: J.P. Morgan

Figure 34: ....real yields have exhibited reduced sensitivity to nominal yields in rallies in recent months

Statistics for daily changes in breakevens and real yields regressed on daily changes in nominal yields, run separately for rallies and sell-offs in nominal Treasuries, for the last 3 months

		Sell-offs		Rallies
	Tenor	Beta	R-sqr	Beta	R-sqr
Breakevens	5	0.26	25%	0.52	54%
	10	0.31	40%	0.52	51%
	30	0.33	47%	0.46	46%
Real yields	5	0.74	73%	0.48	49%
	10	0.69	76%	0.48	48%
	30	0.67	78%	0.54	54%

Source: J.P. Morgan

With the fundamental backdrop turning more supportive and valuations still cheap, albeit somewhat less cheap than a week ago, we maintain tactical long exposure in 5Yx5Y inflation swaps. Along the curve, the intermediate sector continues to offer value versus the wings, and forward-starting swaps avoid the negative carry associated with spot breakeven positions in the current environment. Meanwhile, IOTAs broadly remain tight across the curve, although we acknowledge this is particularly so at the front end, and we think there’s room for the IOTA curve to flatten from current levels.

Figure 35 shows that though IOTAs widened in the aftermath of the July payrolls release, they have partially retraced narrower over the last couple of weeks, with the exception of the long end. While the front end through intermediate sectors have likely been supported by ongoing TIPS asset swap demand (discussed further below), the imminent end of the Fed-on-hold period suggests an environment that will be less ripe for carry trading. Additionally, Figure 36 shows that IOTAs have demonstrated an increased sensitivity to SOFR swap spreads over the past year, with swap spreads explaining more than 80% of the variability in 10-year IOTAs over the period. Our interest rate derivative strategists continue to argue for a flatter SOFR swap spread curve, given that ongoing QT and heavy net Treasury supply suggest higher term funding premium (TFP), and front-end spreads continue to appear substantially tight versus fair value (see Interest Rate Derivatives, Srini Ramaswamy, 8/16/24). While these considerations suggest there is more room for IOTA widening at the front end of the curve, we prefer to hold long exposure in inflation swaps relative to cash breakevens. Overall, we maintain longs in 5Yx5Y inflation swaps.

Figure 35: Following a brief widening earlier this month, IOTAs are back to pre-payrolls levels

IOTAs (derived from z-spread differentials) for hot run TIPS; bp

Source: J.P. Morgan

Figure 36: Over the past year, IOTAs have demonstrated a strong relationship with nominal swap spreads

10-year IOTAs regressed on 10-year SOFR swap spreads (bp); regression over the past year; bp

Source: J.P. Morgan

Inflation swap trading activity update

Turning to inflation swap volumes, we can see that trading activity continues to grow at a robust pace thus far this year. Using data from the DTCC swap data repository (SDR) through the middle of August, inflation swap daily trading volumes increased 25% from an average $3.6bn last year to $4.6bn YTD ( Figure 37). Moreover, the growth in the swap market continues to outstrip that of the cash market, with inflation swap daily trading volumes comprising 23.5% of TIPS volumes over the year. Turning to a more granular frequency, Figure 38 illustrates that much of this uptick has occurred over the last few months; while monthly data are quite volatile, it is notable that average daily trading volumes have increased 60% over the past three months since May to $6.4bn MTD.

Figure 37: Inflation swap volumes continue to rise on both outright and relative to the TIPS market YTD

Annual average daily trading volumes of US CPI inflation swaps (LHS; $bn) versus inflation swaps as a percentage of TIPS volumes (RHS; %)

Source: DTCC SDR, J.P. Morgan

Figure 38: More granularly, we observed a notable uptick in volumes in recent months

Monthly average daily trading volumes of US CPI inflation swaps (LHS; $bn) versus inflation swaps as a percentage of TIPS volumes (RHS; %)

Source: DTCC SDR, J.P. Morgan

Quarter-to-date, we estimate that TIPS asset swaps continue to grow as a share of overall transaction volumes, assuming that transactions with maturity dates that match outstanding TIPS reflect asset swap transactions. Figure 39 shows that, while ZC structures continue to comprise the vast majority of overall inflation swap transactions, asset swaps currently constitute 9% of overall transactions, up from 2% during 3Q23. Meanwhile asset swap transaction sizes have also expanded recently, rising 58% over the past two years to $73mn, while ZC inflation swap sizes have only increased 32% to $48mn over the same period. At the same time the maturity structure of the inflation swap market remains roughly consistent over the period, as over half of all transactions have been concentrated in the 0-5 year bucket ( Figure 40).

Figure 39: TIPS asset swaps continue to grow as a share of overall inflation swap transactions

Quarterly inflation swap trade count split by asset swaps and other (LHS, count) and asset swaps share of overall trade count (RHS, %)

Source: DTCC SDR, J.P. Morgan

Figure 40: More than half of all inflation swap transactions are concentrated in the 0-5 year sector

Distribution of US CPI inflation swap trade count based on original maturity; %

Source: DTCC SDR, J.P. Morgan

Trade recommendations

• Maintain 5yx5y inflation swap longs

-    Maintain long 100% risk, or $50mn notional of 5yx5y inflation swap (swap start: 8/20/29, swap end: 8/20/34).
(TIPS Strategy, 8/16/24). P/L since inception: +2.5bp.

False Fall

• The commencing of rate cuts next month is all but certain, but there is considerable uncertainty around pace. Uncertainty surrounding pace is apparent in option-implied probability distributions, and supported by the tone of recent Fed-speak. Such uncertainty is likely to amplify the sensitivity of yield levels to Fed-speak and/or economic data, making it a key support for jump risk and elevated realized volatility, and is significant enough to warrant a bullish stance on short-expiry swaption volatility despite seemingly rich valuations, in our view
• SOFR and GC rates are normalizing, but front end spreads remain far too narrow. Beyond valuations, one near term catalyst for a correction is likely to come in early September as T-bill net issuance turns negative
• Term funding premium remains narrow relative to fair value, and thus appears biased higher in the near term (due to a correction) as well as the longer term (due to a falling share of price-insensitive UST buyers and eventually higher duration supply). Together with the cheapness of zero-duration spreads, this argues for a flatter spread curve - position for a flatter 5s/30s maturity matched swap spread curve
• Adjusted for the broader term structure, swap spreads appear too wide in the 7-year sector and too narrow in the 20-year sector - initiate 7s/20s swap spread curve steepening exposure, weighted to hedge exposure to rising term funding premium
• We examine the near term and medium Fed expectations that are currently being priced into the swap yield curve. Currently, yield curves anchored in the front end, such as the 2s/5s curve, are fully priced to our Funds rate forecast, while forward OIS rates themselves appear high. This suggests that on a relative basis, outright longs at the front end are preferable to front end curve steepeners
• In contrast, the Fronts/Reds curve does indeed appear fairly priced to our Fed funds forecast, but forward swap curves and back-end spot curves are too flat and therefore underpricing medium term fed expectations. This makes steepeners in these sectors preferable to outright longs in the Reds and/or Fronts/Reds flatteners on a comparative basis

False Fall

New Yorkers are accustomed to many additional seasons on the calendar, one of which is colloquially referred to as "False Fall". Not quite Fall, and not quite Summer, it is a cuspy time when the temperature seems to fluctuate between the two. The US Rates markets appear to be experiencing their own False Fall, with front end yields falling initially, bouncing higher on Thursday and then back lower on Friday. All in all, forward OIS as of mid-2025 traded this week in a ~20bp range and 2- to 5-year yields traded in a ~15bp range, before finishing near the low end of the range (Figure 1).

Figure 1: Swap yields were lower this week on the back of weaker data and a dovish tilt in the FOMC minutes

Selected statistics for SOFR swap yields, YE24 OIS, and 1H25 OIS, 8/16/2024 - 8/23/2024; %

Source: J.P. Morgan.

The catalysts for the decline in yields were many - weak Canadian CPI helped spark the initial move lower, but there were other factors as well. Benchmark revisions to the establishment survey were weaker (-818K) than our expectations (-360K), the FOMC minutes were more dovish than expected especially considering that the meeting took place before the last employment report, and Fed Chair Powell's comments were quite dovish on Friday. But this week was also a reminder of the significant policy uncertainty that continues to permeate the market. As of this writing, markets are pricing in about 100bp of easing by year end. But although there is consensus among Fed speakers that policy is restrictive and easing is now appropriate, there is not yet an obvious consensus on pace. As seen in Figure 2, several speakers appear to be leaning against expectations for 50bp cuts. In other words, the pace of cuts remains undecided, and options markets  reflect that. Our favorite metric for policy uncertainty is based on inferring an implied probability distribution from the prices of calls and puts on Z4 (and/or M5) SOFR futures, and then de-constructing those implied distributions into "weights" associated with each policy outcome. Since the weights must sum to one by design, this provides an intuitive way to "read" the information in the tails of the distribution. The point of this exercise is to see if the forwards (which represent an average across outcomes) are averaging across a relatively narrow (indicating policy clarity) or a relatively wide (indicating policy uncertainty) range of policy scenarios. As we can see, the weights on deep cut scenarios are reasonably comparable to the weights on shallow cut scenarios, which suggests that dispersion remains wide and there is more uncertainty around outcomes even at a horizon that is only four months into the future (Figure 3).

Figure 2: Several Fed speakers appear to be leaning against expecting more aggressive rate cuts

Selected Fed-speak excerpts, 8/22/2024 - 8/23/2024

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

Figure 3: Policy uncertainty remains elevated, even at a horizon that is only four months into the future

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**; Feb 2024 – 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?

Therefore, it appears likely that policy uncertainty will continue to haunt markets until at least next Payrolls, which will be critical in anchoring expectations. Despite the near certainty of a cut in September, the pace and extent of easing remain uncertain, which likely means that yields will be highly sensitive to new economic or policy information that arises in coming weeks. Indeed, the evidence supports the notion that yields are much more sensitive to Fed-speak in periods of significant policy uncertainty. To see this, we turn to our long-standing "Fed-speak Index", which is constructed by (i) manually identifying key Fed-speak events, and (ii) recording the 15-minute change in 5Y UST yields following the start of the speech, and then (iii) accumulating this change in a cumulative index. This index is designed to convey a sense of direction - i.e., is Fed-speak leaning dovish or hawkish. But we can examine the magnitude of changes in this cumulative index on days where there was a Fed-speak event, grouped by periods of policy clarity versus policy uncertainty over the past six months. We do this in Figure 4, where we show percentile moves in each type of period. For instance, 90% of Fed-speak events produced a change of ~2.8bp or less in periods of policy clarity, but during periods of policy uncertainty the moves were larger and the 90th percentile move is over 5bp. In other words, the current backdrop of policy uncertainty will likely serve to amplify the effect of new information, whether from Fed-speak or from economic data.

We see such a backdrop as conducive to jump risk and supportive of elevated realized volatility in rates. This is quite evident when we look at a decomposition of month-to-date delta hedged returns from long straddle positions in various sectors. As seen in Figure 5, long volatility positions were not just profitable overall, but the portion attributable to delivered volatility (i.e., gamma P/L net of theta) was a significant fraction of the overall return. Thus, policy uncertainty has very tangible consequences and will likely continue to be supportive of a bullish view on volatility. Of course, one must always consider whether current valuations are rich enough to offset the above mentioned supportive factors.  In this regard, little has changed from last week. Specifically, in our publication last week (see Hopscotch, 8/16/2024), we examined whether policy uncertainty was significant enough to support a bullish view on volatility despite the fact that our implied volatility fair value models (which do not explicitly incorporate policy uncertainty) pointed to valuations being rich. We addressed this question by building an empirical model for delta hedged returns on volatility positions that explicitly incorporated a measure of policy uncertainty as well as the residual from our implied volatility fair value model. The conclusion then was that policy uncertainty's impact would likely be sufficient to significantly offset the effect of rich valuations, resulting in relatively low carry costs on delta hedged long vol positions. This remains true currently as well. Given this, and given the potential for significant jump risk going into Payrolls and beyond, we remain bullish on volatility in short expiries.

Figure 4: As one might expect, yields have tended to be much more sensitive to Fed-speak in periods of significant policy uncertainty

Size of the Nth percentile absolute change in the Fed sentiment index* on days with one or more Fed-speak events, grouped by policy clarity or uncertainty regime**, for various selected values of N from 10 to 90. Past 6 months; bp

Source:  J.P. Morgan, Bloomberg Finance, L.P.* The Fed sentiment index is computed as the cumulative sum of yield changes in 5-year note Treasury futures in the 15-minute period following the first Fedspeak headline on Bloomberg. Fedspeak is defined as any speech, FOMC meeting or FOMC minutes.** Days on which the absolute difference between the total weight on deep cut scenarios and shallow cut scenarios was between 0.25 and 0.75 are classified as days characterized by Policy Uncertainty, while all other days are deemed to have Policy Clarity. For more details see Figure 3

Figure 5: Long gamma positions have been profitable so far this month, as policy uncertainty helped keep delivered volatility elevated

P/L from daily delta-hedged long straddles † between 8/1 and 8/22 , estimated vega P/L*, gamma P/L** and theta P/L*** for the period; bp/notional

Source: J.P. Morgan.
† Assumed to be delta hedged daily, options are rolled on a monthly basis, assumes no transaction costs
* Vega P/L is estimated as inception vega times change in implied vol (starting implied vol minus aged implied vol on end date)
** Gamma P/L is estimated average gamma x 0.5 x monthly realized volatility x 17 (to represent the time period)
*** Theta P/L is estimated as spot premium minus aged premium at inception

Swap spreads

Swap spreads were once again narrower this week in the front end and the belly of the curve, while remaining flat to slightly wider in longer maturities (Figure 6). These moves came despite a ~5bp decline in O/N GC repo rates and a modest decline in SOFR, suggesting that stressed financing conditions were likely not the main driver behind the narrowing in front end spreads.

Looking ahead, we expect SOFR and GC repo rates to remain in check in the near term, which in turn should help support a normalization in front end spreads. But the catalyst that will likely help spur such a normalization is likely to be a turn in T-bill issuance. As seen in Figure 7, the cheapening in zero-duration spreads has coincided with the surge in T-bill supply, but this should reverse as T-bill net issuance turns negative in September. (A quick note - on any given day, we can regress maturity matched swap spreads in all the benchmark tenors versus modified duration of the corresponding bonds. We define zero-duration-spreads as the intercept from this regression, and we define term funding premium as the negative of the slope. Viewing swap spreads through the lens of these two summary metrics of the term structure offers many advantages. For more details, see Term Funding Premium and the Term Structure of SOFR Swap Spreads).

Figure 6: Swap spreads were narrower this week in the front end and the belly, and modestly wider in longer maturities

Selected statistics for maturity matched SOFR swap spreads, 8/16 - 8/23; bp

Source: J.P. Morgan.

Figure 7: The cheapening in zero-duration spreads has coincided with the surge in T-bill supply

6 week change in T-bill supply (%, left) and 6 week moving average of the residual** from our fair value model for zero-duration swap spreads* (right, inv., bp)

Source: J.P. Morgan.* For details on our methodology for calculating Zero-duration swap spreads, see Term Funding Premium and the Term Structure of SOFR Swap Spreads
** For details on our fair value framework for zero-duration swap spreads, see Interest Rate Derivatives 2024 Mid-Year Outlook

Term funding premium also appears too low relative to fair value, to the tune of ~0.3bp/year, as seen in Figure 8. Moreover, not only is term funding premium likely biased higher in the near term because of this mispricing, it is also likely biased higher in the longer term because of structural reasons (the falling share of price-insensitive investors in the UST market, and eventually higher duration supply). Taken together, this expectation of higher zero-duration swap spreads as well as higher term funding premium points to wider swap spreads at the front end as well as a flatter spread curve. Therefore, we now recommend initiating exposure to a flatter 5s/30s swap spread curve (see Trade recommendations).

Finally, it is worth noting that when we look at swap spreads in each sector in relation to the term structure (i.e., if we assume no changes to term funding premium or to zero-duration swap spreads), we still find that (i) front end swap spreads in the 2- to 5-year sector are currently trading too narrow relative to the term structure, while (ii) longer maturity swap spreads in the 7- to 30-year sector are trading too wide, but (iii) with the exception of 20-year swap spreads which appear too narrow relative to the term structure (Figure 9). Therefore, as a relative value trade, we recommend a weighted 7s/20s maturity matched swap spread curve steepener where the weighting is designed to hedge against a broader rise in term funding premium (which would flatten the overall spread curve - see Trade recommendations).

Figure 8: The residual from our Term Funding Premium model has been fairly mean-reverting and currently appears too narrow

Residual from a regression of term funding premium* versus its drivers**, past 1 year; bp/year

Source: J.P. Morgan.
* For details on our methodology for calculating Term funding premium, see Term Funding Premium and the Term Structure of SOFR Swap Spreads
** For details on our fair value framework for term funding premium, see Interest Rate Derivatives 2024 Mid-Year Outlook

Figure 9: Front end swap spreads are currently trading too narrow relative to the term structure, while longer maturity swap spreads (with the exception of the 20-year sector) are trading too wide

Residual from our fair value model* for maturity matched swap spreads, 8/22/2024; bp

Source: J.P. Morgan.
* Fair values are computed by adding our Term structure baseline swap spread values and our swap spread deviation relative to the term structure of swap spreads across each sector (see Term Funding Premium and the Term Structure of SOFR Swap Spreads for details)

Swap Yield Curve

The swap yield curve steepened this week as yields headed lower, with almost every sector of the curve now steeper relative to week-ago levels. The exception was the very front end, where the 1s/2s curve flattened modestly to the tune of 1bp (Figure 10).

Figure 10: Swap yield curves steepened in most sectors this week

Selected statistics for benchmark swap yield curves, 8/16/2024 - 8/23/2024; bp

Source: J.P. Morgan.

Our approach to explaining swap yield curve movements and taking curve views has been centered around our swap curve fair value model, which uses 5 different factors (plus a control variable to account for historical periods when the Fed has employed policy guidance as a monetary policy tool) to explain a variety of spot and forward swap curves. These factors include the short rate (3Mx3M forward OIS), medium term policy expectations (3Mx3M / 15Mx3M OIS curve), the size of the Fed balance sheet, and short term as well as long term inflation expectations. As we have noted (see our 2024 Mid-Year Outlook), such a model has been fairly effective in explaining moves in various spot as well as forward swap curves.

In the current context, the most important factors are the first two, which capture the near term as well as the medium term outlook for policy rates, but we include a full set of coefficients in Figure 11. As our economists have noted, most recently on Friday (see Cuts are coming, M. Feroli, 8/23/2024), we expect the Fed to cut rates by 50bp at the next two meetings, followed by 25bp cuts thereafter until we get to a 3% level in the second half of next year. If we assume this forecast will be realized, we arrive at a fair value of around 4.1% for the 3Mx3M OIS rate, and a fair value of -125bp for the 3Mx3M / 15Mx3M forward OIS curve. This compares with the market's pricing of 4.3% and -120bp respectively for these two factors. In other words, the 3Mx3M / 15Mx3M forward OIS curve is fairly priced relative to our forecast for policy rates, but there is room for OIS rates at the very front end to decline from current levels if our funds rate projections are realized.

Figure 11: Our empirical fair value model uses a handful of factors to explain a variety of different spot and forward yield curves, but fed expectations - both near term and medium term - are currently the most important ones

Statistics from regressing* various spot and forward curves against 7 drivers**, R-squared, Standard error, current yield curve level, and current fair value; current value as of 8/22/2024

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)

But we can use this empirical model to also ask what values for these important drivers are being priced into the swap yield curve. Said differently, if we take our empirical swap yield curve fair value model coefficients as a given, we can ask - what values for the 3Mx3M OIS rate and the 3Mx3M / 15Mx3M forward OIS curve would cause current swap yield curves to look fair. In order to avoid assuming a beta between the 3Mx3M rate and 3Mx3M/15Mx3M curve, we answer this question in two stages. First we choose swap curves that have significant coefficients with respect to the 3Mx3M rate, but do not have numerically significant exposure to the forward OIS curve. Two such examples are the 2s/5s and 2s/7s spot swap curves. From each of these, we can calculate the level of 3Mx3M OIS that would cause these swap curves to appear fair at the current levels. We walk through this process in Figure 12 - as can be seen, it would appear that swap yield curves are currently priced to a forward OIS of ~4.15%. This is quite close to levels that would be consistent with our own Fed forecast, but a good ~15bp lower than the current 3Mx3M forward OIS yield. This leads us to our first observation - on a relative basis, outright longs at the very front end are likely more attractive than front end swap curve steepeners, as these swap curves are already priced to more aggressive easing expectations than forward OIS rates themselves.

Next, armed with an estimate for the "implied" 3Mx3M OIS rate that swap curves are pricing in, we turn our attention to medium term Fed expectations (i.e., the 3Mx3M / 15Mx3M curve). To calculate the implied value for medium term Fed expectations that swap yield curves are currently pricing in, we turn to a different set of curves that have numerically significant exposure to medium term Fed expectations but not to 3Mx3M OIS rates. As it turns out, 2-year forward swap curves anchored in the belly are ideally suited to this purpose. Using three such selected curves, and assuming that the 3Mx3M OIS will reach a value of 4.15% (i.e., the estimate from above), we can solve for the value of medium term Fed expectations that would make these curves appear fair. The results of this exercise are shown in Figure 13 - as can be seen, swap curves are pricing in a 3Mx3M / 15Mx3M OIS curve of around -100bp, which is a good bit steeper than our forecast of -125bp as well as the current market pricing (the 3Mx3M / 15Mx3M OIS curve currently stands near -123bp). This leads us our second observation - forward swap curve steepeners are likely more attractive than Fronts/Reds flatteners, as the former is underpricing Fed easing expectations relative to the latter.

Therefore, we recommend that clients favor outright front end longs in place of front end steepeners on a relative basis, and favor forward curve and back end swap curve steepeners over outright duration longs in Reds, when positioning for a more dovish Fed.

Figure 12: Swap curves are currently priced to a forward OIS that is~15bp lower than the current 3Mx3M forward OIS

Current level (%), Mispricing (%)*, Beta to 3Mx3M OIS**, and Implied 3Mx3M OIS (%)*** for the 2s/5s and 2s/7s spot swap yield curves as well as the average, 8/22

Source: J.P. Morgan.
* Mispricing is calculated as Cur. Curve - Fair Value as shown in Figure 11
** Beta with respect to 3Mx3M OIS comes from the regression described in Figure 11
*** Implied 3Mx3M OIS is calculated as Mispricing / Beta wrt 3Mx3M OIS, plus the current level of 3Mx3M OIS.

Figure 13: Swap curves are pricing in a 3Mx3M / 15Mx3M OIS curve of around -100bp, which is steeper than our forecast as well as the current market pricing

Implied* 3Mx3M / 15Mx3M OIS curve for various fwd swap yield curves, 8/22; %

Source: J.P. Morgan.
* Implied 3Mx3M / 15Mx3M OIS curve is calculated as the current mispricing from our fair value model (Figure 11), assuming an implied value for 3Mx3M OIS of 4.15% (Figure 12), divided by Beta wrt 3Mx3M / 15Mx3M OIS curve, plus the current level of the 3Mx3M / 15Mx3M OIS curve.

Trading Recommendations

• Initiate 5s/30s swap spread curve flatteners
   The cheapening in zero-duration spreads has coincided with the surge in T-bill supply, but this should reverse as T-bill net issuance turns negative in September. Moreover, term funding premium likely biased higher in the near term due to mispricing, falling share of price-insensitive investors in the UST market and eventual higher duration supply. Taken together, this points to wider swap spreads at the front end and a flatter spread curve and we prefer expressing this through 5s/30s swap spread curve flatteners.
  -Pay fixed in 4.25% Jun 30 2029 maturity matched SOFR swap spreads. Buy $100mn notional of the 4.25% Jun 30 2029 (yield: 3.657%, PVBP: $447.1/bp per mn notional), and pay fixed in $97.9mn notional of a maturity matched SOFR swap (coupon: 3.354%, PVBP: $456.7/bp per mn notional) at a swap spread of -30.3bp.
  -Receive fixed in 4.625% May 15 2054 maturity matched SOFR swap spreads. Sell $24.7mn notional of the 4.625% May 15 2054 (yield: 4.107%, PVBP: $1813.5/bp per mn notional), and receive fixed in $23.9mn notional of a maturity matched SOFR swap (coupon: 3.305%, PVBP: $1869.0/bp per mn notional) at a swap spread of -80.2bp.
• Initiate 7s/20s weighted swap spread curve steepeners
   Maturity matched swap spread in the 7Y sector are wide relative to the term structure, while 20Y spreads are narrow relative to the term structure. Therefore, as a relative value trade, we recommend a weighted 7s/20s spread curve steepener where the weighting is designed to hedge against a broader rise in term funding premium (which would flatten the overall spread curve).
  -Receive fixed in 4.25% Jun 30 2031 maturity matched SOFR swap spreads. Sell $100mn notional of the 4.25% Jun 30 2031 (yield: 3.702%, PVBP: $611.3/bp per mn notional), and receive fixed in $100.5mn notional of a maturity matched SOFR swap (coupon: 3.33%, PVBP: $608.1/bp per mn notional) at a swap spread of -37.2bp.
  -Pay fixed in 4.125% Aug 15 2044 maturity matched SOFR swap spreads. Buy $20.0mn notional of the 4.125% Aug 15 2044 (yield: 4.189%, PVBP: $1336.5/bp per mn notional), and pay fixed in $18.5mn notional of a maturity matched SOFR swap (coupon: 3.441%, PVBP: $1443.2/bp per mn notional) at a swap spread of -74.8bp.
• Unwind calendar spread wideners in US Futures
  As the roll cycle is over 30% complete, we recommend unwinding this calendar spread at a loss of 0.8 ticks. For original trade write up, see Fixed Income Markets Weekly 2024-08-16.
• Unwind calendar spread narrowers in FV Futures
  For the same reason as above, we recommend unwinding at a loss of 0.1 ticks. For original trade write up, see Fixed Income Markets Weekly 2024-08-16.
• Unwind TU/TY invoice spread curve flatteners (1:0.35 weighted)
  This trade has underperformed our expectations and we recommend unwinding at a loss of 6.3bp. For original trade write up, see Fixed Income Markets Weekly 2024-06-07.
• Unwind long A+100 1Yx5Y payer swaptions versus selling A-100 1Yx5Y receiver swaptions, delta-hedged daily, to position for a correction in skew
  This trade has underperformed our expectations and we recommend unwinding at a loss of 8.5abp. For original trade write up, see Fixed Income Markets Weekly 2024-04-19.
• Maintain 10Y swap spread narrowers
  P/L on this trade is currently -0.3bp. For original trade write up, see Fixed Income Markets Weekly 2024-08-16.
• Maintain 3s/7s swap spread curve flatteners
  P/L on this trade is currently -0.5bp. For original trade write up, see Fixed Income Markets Weekly 2024-08-16.
• Maintain 0.875% June 2026 / 0.875% September 2026 swap spread curve flatteners
  P/L on this trade is currently 0.1bp. For original trade write up, see Fixed Income Markets Weekly 2024-08-16.
• Maintain longs in 6Mx5Y swaption implied volatility on an outright basis, delta hedged daily
   P/L on this trade is currently -0.4abp. For original trade write up, see Fixed Income Markets Weekly 2024-08-02.
• Maintain conditional exposure to a flatter 1s/7s swap yield curve in a selloff using 6M expiry payer swaptions
  P/L on this trade is currently -4.4bp. For original trade write up, see Fixed Income Markets Weekly 2024-07-12.
• Continue to overweight 6Mx5Y swaption straddles versus 150% of the vega risk in 6Mx30Y straddles
   P/L on this trade is currently -1abp. For original trade write up, see Fixed Income Markets Weekly 2024-07-12.
• Continue to Pay-fixed in 4.625% Feb ‘26 maturity matched swap spreads
  P/L on this trade is currently -1.7bp. 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 -1.3bp. 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 overweight 1Yx10Y straddles versus a gamma-neutral amount of 1Yx15Y straddles
   P/L on this trade is currently -2.2abp. For original trade write up, see Fixed Income Markets Weekly 2024-05-03.
   
• Continue to overweight 6Mx5Y and 6Mx30Y swaption volatility (vega weights of 0.32 and 0.76, respectively) versus selling 6Mx10Y swaption volatility
  P/L on this trade is currently 0abp. For original trade write up, see Fixed Income Markets Weekly 2024-04-05.

Closed trades over the past 12 months

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

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

• Chair Powell’s remarks at the Jackson Hole forum left little doubt that the Fed will soon cut rates. However, as expected, he did not reveal the size of the adjustment
• Our economists still expect a 50bp reduction in both September and November, followed by 25bp cuts thereafter, until reaching 3%
• ON RRP balances have stabilized around the mid to low $300bn level. We suspect there will be limited room to move further lower, in the absence of meaningfully more T-bill supply and given some MMFs’ structural need for liquidity
• Last week’s Liberty Street blog posts conclude that reserves remain abundant as of 1Q24, based on a variety of indicators
• We take a closer look at some of the discussed indicators, namely domestic borrowing in the fed funds market and upward pressure in repo rates
• Domestic bank borrowing in the overnight unsecured market remains minimal as of 2Q24, regardless of whether they are a US G-SIB or a smaller bank
• US banks’ participation in the repo market as a marginal buyer also remains low, relative to prior peaks, which suggests the percentage of overnight Treasury repo transacted at or above IORB is not signaling reserve scarcity
• Near-term catalysts: July personal income (8/30), July JOLTS (9/4), Aug employment (9/6), Aug CPI (9/11), Aug PPI (9/12)

Market commentary

In his clearest signal to date, Chair Powell’s remarks at the Jackson Hole forum left little doubt that the Fed will soon cut rates. In addition to expressing growing confidence in inflation returning to 2%, he emphasized the FOMC’s attentiveness to labor market conditions, adding that the Fed does “not seek or welcome further cooling in the labor market.” However, as expected, he did not reveal the size of the adjustment, only noting “the timing and pace of rate cuts will depend on incoming data, the evolving outlook, and the balance of risks” (see Cuts are coming. M. Feroli, 8/23/24). Meanwhile, comments from Fed officials this week were not as open-ended as Powell’s, but they also suggest that it may be appropriate to begin lowering rates soon ( Figure 42).

Figure 42: Comments from Fed officials this week were not as open-ended as Powell’s, but they also suggest that it may be appropriate to begin lowering rates soon

FOMC commentary over the past week

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

The minutes from the July FOMC meeting, held before the weak July employment report, echoed a similar dovish sentiment. Several members expressed increased confidence that inflation was moving toward the 2% target, but some participates noted growing risks from continued easing in labor market conditions could lead to more serious deterioration. While only “several” had considered a reduction of 25bp at the last meeting, a “vast majority” were ready to cut at the September meeting, with “many” viewing current policy as restrictive (see July FOMC minutes highlighted labor market risks, A. Reinhart, 8/21/24).

All of this heightens the importance of upcoming labor market data, particularly the August payrolls report, in determining whether the next move will be a 25bp-or-50bp cut. This week’s employment-related data provided evidence that the labor market might not be as strong as what we had thought, but also not severely deteriorating. BLS’ preliminary benchmark revision for the establishment survey revealed that total payroll employment in the year leading up to March 2024 was 818K lower than previously thought, a relatively large revision that was entirely concentrated in private employment. This suggests that average monthly job growth from March 2023 to March 2024 would decrease by about 70k. However, since the revision cuts off in March 2024 and the implications for more recent months are not fully clear, it is difficult to predict how the Fed will interpret this data (see US: Benchmark brings big downward revision to (lagged) jobs, A. Reinhart, 8/21/24). In more recent, higher-frequency data, weekly jobless claims rose slightly by 4k to 232k for the week ending August 10, while the four-week average has remained stable since late June. The growth of continuing claims, which surged between late May and late June, has slowed and might be leveling off (see US: Good news from mostly stable jobless claims, A. Reinhart, 8/22/24). Meanwhile, our economists found little evidence that Hurricane Beryl meaningfully impacted July’s payrolls report and believe that only a small 10-20k would be added to the August payrolls report as jobs were restored after the hurricane (See Focus: Beryl and July jobs, A. Reinart, 8/23/24).

All told, Treasury yields whipsawed once again this week. In the front-end, 2-year yields moved in a wide range, peak to trough, from 3.91% to 4.07%, ultimately closing the week 15bp lower at 3.91%. Our economists still expect a 50bp reduction in both September and November, followed by 25bp cuts thereafter, until reaching 3%. By week’s end, OIS markets are pricing in 104bp of easing by year-end, compared to 95bp last Friday, and about 36% chance of a 50bp cut at the September meeting ( Figure 43).

Figure 43: By week’s end, OIS markets are pricing in 104bp of easing by year-end and about a 36% chance of a 50bp at the September meeting

OIS-implied change in fed funds effective rate by FOMC meeting, as of 8/23/24, and 8/16/24 (bp)

Source: J.P. Morgan

Turning to the Fed’s balance sheet, the meeting minutes didn’t reveal much new information. The SOMA manager noted that repo rates had risen due to increased demand for financing Treasury securities and balance sheet normalization. Meanwhile, the staff projected that ON RRP balances should decline as T-bill issuance increases, though specific idiosyncratic factors for some ON RRP participants might sustain balances. The manager concluded that reserves remained abundant but emphasized that the staff would continue to closely monitor developments in money markets.

To that end, bank reserves have been steady at ~$3.3-3.4tn, mostly flat YTD. Meanwhile, ON RRP balances have fallen by ~$700bn so far this year, as MMFs have shifted towards T-bills and non-Fed repo holdings. Recently, RRP balances have stabilized around the mid to low $300bn level and we suspect there will be limited room to move further lower, in the absence of meaningfully more T-bill supply and given some MMFs’ structural need for liquidity. Consequently, we see QT as being in its final stages, with perhaps a few more months remaining, at which point the Fed’s balance sheet should be around $7tn by year-end, reserves at ~$3.3tn and ON RRP balances slightly below $300bn.

Tracking reserve abundance: insights from the fed funds and repo markets

Last week’s Liberty Street blog posts (see here and here) highlighted several indicators to watch for when the amount of reserves might be transitioning from abundant to ample. Among these indicators are late interbank payments, banks’ intraday overdrafts, domestic borrowing in the fed funds market, and upward pressure in repo rates, as expressed as a share of overnight Treasury repo transacted at or above IORB, assessed in conjunction with the Fed’s daily estimate of fed funds elasticity. Unfortunately, data on these indicators are either not publicly available or are published on a lagged basis, typically around quarter-ends. Based on the above indicators, the staff at the NY Fed concludes that reserves remain abundant as of 1Q24.

Even so, we thought it would be interesting to do a deeper dive into the fed funds market and look at the composition of bank borrowing. As Figure 44 shows, domestic banks have remained largely absent from the overnight market, whether in fed funds or Eurodollar deposits. This is not surprising: in a world where reserves are abundant, domestic banks do not have to borrow in the overnight market to remain funded. However, when access to funding becomes more challenging, as was the case in 2018-2019 and in early 2023, borrowing from the overnight market increases. Meanwhile, FBOs (foreign banking organizations) remain the primary borrowers of fed funds and Eurodollar deposits, though their participation is mostly driven by arbitrage behavior (i.e., borrowing in fed funds/Eurodollars and depositing the proceeds at the Fed to earn IORB). The above is consistent with call reports of banks’ net fed funds borrowing. Indeed, as of 2Q24, domestic banks have not found a need to borrow fed funds on a net basis, regardless of whether they are a US G-SIB or a smaller bank, while FBOs have maintained a steady presence ( Figure 45).

Figure 44: Domestic banks have remained largely absent from the overnight market, both in fed funds and Eurodollar deposits

Fed funds and Eurodollar deposits (OBFR-FF) volumes (LHS, $bn) versus Foreign OBFR volumes as a % of total (RHS,%)

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

We also thought it might be worthwhile to look at how banks’ involvement in repo has evolved, as their participation as marginal buyers of repo not only can help police repo rates on the upside but also signal a decline in liquidity to a point such that overnight Treasury repos trade at or above IORB. This is because banks tend to engage in repo when GC > IORB. To that end, bank call reports show little increased net reverse repo activity on the part of US G-SIBs as of 2Q24, as the GC/IORB spread remained subdued ( Figure 46). This might change in the future as QT continues to drain liquidity out of the system against a backdrop of increased Treasury issuance and elevated dealer inventories. But for now, neither the GC/IORB spread nor US G-SIBs’ net reverse repo activity suggest there is an alarming percentage of overnight Treasury repo transacted at or above IORB that would signal reserve scarcity is approaching.

Figure 45: Domestic banks, whether they are US G-SIBs or smaller institutions, have not needed to borrow fed funds on a net basis, while FBOs have maintained a steady presence

Net fed funds purchased ($bn)

Source: S&P Capital IQ, Bank Call Reports, J.P. Morgan

Figure 46: As of the second quarter, US G-SIBs have shown little increase in net reverse repo activity, as the GC/IORB spread remained subdued

Net reverse repo activity (LHS, $bn) versus GC/IORB spread (RHS, bp)

Source: S&P Capital IQ, Bank Call Reports, J.P. Morgan

Excerpted from Short-Term Market Outlook and Strategy, Teresa Ho, August 23, 2024

Dua LLPA

• Mortgages had a relatively unexciting week, with lower coupons tightening modestly and the rest of the stack close to unchanged
• Along with most spread products, valuations have recovered from the local wides seen early in the month; still, with mortgages continuing to offer historically compelling yields vs. corporates, we think the technical support for the product from money managers can be maintained
• After the early August spike in refi applications, volumes have come down but remain relatively elevated; of note, the MBA conventional refi loan size spike overstates the change in application loan size for GSE eligible loans, as measured by RALI and Optimal Blue
• We take a close look at the call protection offered by stories with higher LLPA charges, namely investor, 2^nd home, high LTV, and low FICO
• The higher rates they get come from translating those upfront fees into the rate rather than pure SATO and can be durable upon refinance dependent upon changes in the borrower’s FICO and LTV
• Looking at investor loans (with the longest history of consistent LLPAs to draw from), we show how their rates and prepayments have empirically reflected the behavior implied by the upfront fees
• 2^nd home loans now have LLPAs that exactly mirror investor loans, but 100% 2^nd pools trade well back of 100% investor
• The historical prepay record, when adjusted for the new fees, doesn’t immediately imply that 2nd homes should be more callable than investor at the same LLPAs, but there are some “known unknowns” as far as differences in underwriting frictions and borrower behavior

Views

• Continue to prefer UIC conventionals, u/w 4.5s&5s
• Prefer seasoning in discounts
• Consider 2^nd home pools which trade at payups below comparable investor pools
• Within loan bal, stick to relatively higher cuts to avoid low LTV cashouts

Mortgages had a relatively unexciting week, with lower coupons tightening modestly and the rest of the stack close to unchanged. MBS continue to offer attractive ZVs, somewhat tight OASs (in our model - others see more fair levels), and with IG corporate spreads having fully recovered from their early August widening, support for the sector from money managers is likely to remain robust. Short of a significant shift in the economic data, the market seems to have retracted to the tighter end of the stable range it has traded around this year. Spec payups were perhaps a hair firmer on 5.5s, but generally the moves were not significant.

Figure 47: Higher coupons are starting to recover from their early August widening, and lower coupons are a touch wider

Current, 1m, and 6m Treasury OAS ranges across the TBA stack in our research beta model. The black dots represent the current OAS, the blue boxes represent the 1m range, and the black lines represent the 6m range (as of 4:30PM, 8/22/2024)

Source: J.P. Morgan

Somewhat dovish Fed minutes and the BLS’s 818k downward revision to the level of non-farm payrolls saw a modest bull steepener, with one result being that we now sit at primary rates of ~6.44% on GSE loans (vs. the Aug 5 low of 6.39%) and ~5.47% for VA IRRRLs (actually below the Aug 5 low of 5.61%), according to Optimal Blue data. That hasn’t resulted in another surge in applications, however (see Figure 48 and Figure 49). Evidently, originators picked some of the low hanging fruit earlier in the month, though to be fair the running pace of rate locks is still quite a bit higher than it was two months ago.

Figure 48: Primary rates have edged lower...

GSE rate/term refi lock rates and number of locks

Source: Optimal Blue, https://www2.optimalblue.com/

Figure 49: ...but lock volume is off the early August high

VA rate/term refi lock rates and number of locks

Source: Optimal Blue, https://www2.optimalblue.com/

Most of what we see in the Optimal Blue lock data is consistent with the patterns observed in the RALI and MBA refi indices. The first full week of August (the 5^th - 9th) registered the largest increases in application volume across all of these series. One notable discrepancy, however, is in the changes in loan sizes for conventional refi apps. The MBA’s conventional refi application series includes non-conforming loans, which caused their average loan size to surge by $150k between the weeks ending 7/26 and 8/9 ( Figure 50). The series has subsequently dropped to $301k (from a high of $403k) for the week ending 8/16. The information imparted by this swing isn’t particularly useful to MBS investors, however. Since we’re focused on loans in GSE pools, Fannie’s RALI index should tell us much more about how much the average loan size on GSE eligible loans moved; by indexing the ratio of the RALI $ and RALI count, it appears that loan sizes were only up by about 14% in the same period the MBA series implies they were up nearly 60% ( Figure 51). That’s also much more in line with the loan sizes that we see in the Optimal Blue set ( Figure 52).

Figure 50: The MBA conventional refi loan size has been extremely volatile...

MBA conventional refi loan size, $k

Source: MBA

Figure 51: ...but RALI indicates a much more muted move for conforming loans

Indexed change in MBA conventional refi loan size and RALI implied loan size (ratio of dollar volume and count of applications)

Source: J.P. Morgan, MBA, Fannie Mae

Figure 52: Optimal Blue also shows that the growth in GSE lock size has been on the order of 10-15%

1wk rolling average of VA and GSE refi loan sizes, $k

Source: J.P. Morgan,https://www2.optimalblue.com/

Taken all together, we think that this points to the bulk of the expected increase in GSE refi volume being explained by the surge in applications - to estimate the UPB change in refi volume, the additional multiplier provided by the loan size should be more on the order of 10-15%. That still implies a meaningful difference in the rate/term refi volume expected to close in September vs. August (up 80%, but not 150%).

This week, we take a look at the call protection afforded by LLPAs. Generally, the stories with an LLPA angle have a first order elbow shift that can be readily observed in the empirical data for types like investor. However, there are complicating factors, since LLPAs also correlate with borrower characteristics that might imply less propensity to refi or more arduous underwriting.

A sharp rally and a surging refi index is a good reminder that call risk exists in the mortgage market and that there is a wide menu of spec pool choices that can help mitigate it. In this section, we want to think through the fundamentals of the protection that comes from the Loan-Level Price Adjustments (LLPAs) that are charged based on an individual loan’s characteristic. The most common pool types with protection that comes at least in part from the LLPA Matrix are investors and 2nds (discussed elsewhere in this piece), as well as high LTV and low FICO. Figure 53 gives the key details for each type along with stats based on their actual issuance in pool form this year.

How should we think about LLPA protection? We’ll focus on investor loans given their sizable LLPA and long history. At the time of origination, imagine two identical loans at 75LTV, with the only difference being that one is an investment property and the other is owner-occupied. Let’s say the owner loan gets a rate of X%. Then the investment property should get a rate of (X + L)%, where L = the rate equivalent of the LLPA charge for investment properties, which is currently 2.125% upfront. At a 4x mult, not unreasonable lately, L =0.53%. This is where measuring incentive/call protection gets interesting. If X was also the prevailing interest rate at the time of origination, it looks like the owner-occupied loan has 0 incentive while the investor loan has 53bp. Of course, since the same LLPAs apply on rate/term refis, into a 50bp rally we would expect these loans to pay the same (as long as other factors aren’t in play, which we will get to later), even as one shows 50bp of incentive and the other shows 103bp.

We can see this empirically in the origination data ( Figure 54). Narrowing down other characteristics like LTV and FICO so that we get a roughly similar LLPA band between owner and investor, we can look at the rate spread between the two. Over the past 12 years, it has bounced between 30bp and 70bp, with that movement based on both changes in the LLPA charge as well as the multiples being used to price the fee in rate terms.

Figure 54: Seeing the difference in origination rate caused by the LLPA charge for owner versus investor loans

Originated mortgage rate on conventional 30yr investor versus owner loans, controlling for other factors such as LTV (60-70%), FICO (>720), loan size ($250k-$450k), geography (no NY), purpose (purch + rate/term) and units (1 unit) to limit possible differences in LLPA

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

The current level of 43bp, and the 1.625% charge for 60-70LTV, implies a reasonable 3.8x mult ( Figure 55). Thinking about how that mult (and implied protection) can change into a rally, the rate spread widened out significantly from 2018 to 2020 as coupon swaps compressed and meant that 2018 originated investor loans actually gained another 20bp of elbow shift protection by mid 2019.

Figure 55: The multiple implied by the investor LLPA has typically tracked well with coupon swaps, though we haven’t seen any LLPA mults compress as much as the TBA stack

Translating the spread difference between investor/owner loans into an IO multiple based on the LLPA fee (1.75% before Sep. 2015, 2.125% until 5/1/2023, and then 1.625% for this collateral), along with the current coupon +50 to current coupon +100 multiple

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

Interpreting the prepayment data can be trickier, especially during the latest refi wave. One way we can approach it is using a similar cohort originated during the same month like our theoretical example above, in this case, April 2018. Figure 56shows how quickly the owner-occupied speeds ramped up relative to investor during the initial rate drop in 2020. We see this in many spec types, and it seems to boil down to just how capacity-constrained originators were at the start of the rally. That meant prioritizing the biggest loans, and more applicable to investor loans, those with the easiest underwriting process. Beyond their LLPAs, there is a bit more hassle to underwriting investment properties (can’t get an appraisal waiver if rental income is needed to qualify, for example), and that was enough to keep them from matching the speediest loans at the start of the rally. However, as originators bolstered their ranks, the investor cohort eventually paid the same as owners by late 2020. Looking at a 2021 s-curve, this fact is reflected in the neat 50bp elbow curve between the two implied by the LLPAs at the time and a 4x mult ( Figure 57).

Figure 56: The most clean owner loans were far faster than any elbow shift would imply during the initial 2020 rally…

Speeds on April 2018 originated loans (owner vs investor) for UMBS 30yr, >$300k loan size, <75LTV, >720 FICO, 1 unit

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

Figure 57: …but the elbow shift eventually re-asserted itself in 2021 as capacity mattered less

UMBS 30yr s-curve observed in 2021 for >$300k, <75LTV, >720 FICO, owner versus investor

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

To wrap-up, in theory the most important aspect of LLPA call protection comes from the implied elbow shift. Now, each category (occupancy, high LTV, low FICO) has its own quirks that goes along with it. In addition, when looking at the refi rate (no cashout), the LLPAs can change as the borrower’s credit profile evolves over the course of the loan. This is particularly true for the LTV and FICO stories, where the protection is often more transient, especially during times of rapid home price appreciation.

Over the past few years, the GSEs have adjusted LLPAs several times, changing levels on the overall FICO/LTV grid and for specific loan characteristics (see MBS Market Commentary: Three Shall Be the Number). After the smoke cleared on these adjustments, 2^nd home and investor loans delivered on or after May 2023 emerged with identical LLPAs. Despite this, as well as the fact that 2^nd home and investor loan characteristics now broadly mimic one another, investor pools trade at higher payups than 2^nd homes ( Figure 58 & Figure 59). Some of this could be perceived differences in underwriting/behavior between the borrower sets or a greater comfort in the long history of investor pools, but the LLPAs and loan stats alone would argue for an increase in payups on 2^nd home loans.

In this piece, we’ll explore what is available in the prepay data for comparing the two pool types. To start, we look back at 2021 speeds (when loans had significant incentive, but before the LLPA changes) and consider how they might shift and look with today’s fees. Observing that fully ramped investor and 2^nd home s-curves should be similar in today’s environment, we then compare curtailments on ramping loans. Though we find that both 2^nd home and investor collateral provide protection against non-spec curtailment spikes, 2^nd home curtailment spikes are slightly higher than investor curtailment spikes, so this story is still slightly worse than investor, but only for a tick or two due to these ramp speeds.

Figure 58: New prod 2^nd home pool payups are trading 7-14tks back of investor pool payups …

Conventional investor vs. 2^nd home average payups for 6s-7s as of 8/22/2024, 1/32s

Source: J.P. Morgan

Figure 59: … and this gap cannot be explained by repline differences in loan characteristics, since recent 2^nd home and investor collateral eligible for separate pooling looks similar for recent production

UMBS 30 2^nd home vs. investor loan stats, for collateral with >50bp incentive, 250k-800k orig. lnsz, FICO > 720, excluding NY, low FICO, and high LTV pools, grouped by issuance years

		Orig. Bal ($bn)	WAC	LTV	Loan size	% 1 unit	SATO
2020 - 2021	2nd home	2.7	3.8	77.1	363.6	99.9	76.0
	Investor	40.1	3.8	65.7	374.2	72.3	85.0
2023 - 2024	2nd home	11.1	7.2	70.5	404.6	100.0	39.0
	Investor	23.5	7.2	69.1	383.9	83.8	37.0

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

The big remaining question mark when considering the call protection of investor versus 2^nd home loans is possible underwriting and behavioral differences. For underwriting, it is common for borrowers on investment loans to use their rental income to qualify, adding to the documentation burden and making the process more manual. This friction doesn’t exist on 2^nd homes. The behavioral part is harder to ascertain; the change in LLPAs means the historical data isn’t a great guide. On the surface, both sets of borrowers would seem to have ample reason to keep track of refinance opportunities, given their exposure to real estate.

Returning to what we can quantity, looking at the degree of 2^ndhome call protection with the new LLPAs is difficult since most loans have been deeply out-of-the-money during the period of 2023. Instead of looking at recent s-curves, we compare ramped 2^ndhome and investor s-curves observed in 2021, and then considered how those curves might shift based on today’s LLPAs. At that time, when investor loans had higher LLPAs than 2^ndhomes and 2^ndhome LLPAs were negligible, there was a clear elbow shift between investors and 2^ndhomes ( Figure 60).

Figure 60: When investor loans had higher LLPAs than 2^nd homes, there was a clear elbow shift between investors and 2^nd homes …

UMBS 30 2^nd home and investor 8-30 WALA s-curves observed in 2021, for collateral with 200k-800k orig. lnsz, FICO > 720, excluding NY, low FICO, and high LTV pools, with minimum balances of $500mm and s-curves

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

Figure 61: … but adjusted for LLPAs and LTV, investor and 2^nd home eligible collateral eligible for separate pooling would nearly have paid the same

Left-hand chart, without the original investor s-curve, with new 2^nd home and investor s-curves adjusted for new LLPAs using investor / 2^nd home FICO / LTV / unit distributions, assuming a 5x mult for LLPA changes and 5 pt. lower orig. LTVs for 2^nd homes

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

To reflect potential refi behavior more accurately in the present environment, we first take 2021 2^nd home and investor loans and estimate how their LLPAs would change (based on individual stats like LTV/FICO) with the new grid. These changes in fee can be translated into elbow shifts (and plotted into an s-curve) by putting them in rate terms. We choose a 5x mult for this calculation (1% fee = 20bp move in rate incentive) as a reasonable level given what we saw in investor versus owner issued rates during 2021.

Beyond mechanically adjusting these s-curves, we also recognize that the presence of 2^nd home LLPAs now affects borrower downpayment behavior in and of itself. For instance, before 2022, a 2^nd home borrower could lever up from 75 to 80 LTV without incurring any additional LLPA charge, but now, doing so would mean increasing their rate due to 1.25% more in upfront fees. Borrowers taking out 2^nd home loans are motivated to pay more principal than they were before 2022; stepping down an LTV bucket is equivalent to paying points while getting to apply the amount paid directly to principal. Moreover, borrowers taking out loans for 2^nd homes likely have primary residences with built up home equity, and they could consider tapping that equity to increase the downpayment on their 2^nd home.

With these considerations in mind, we adjust LTVs observed in the 2021 s-curve down by 5 points in addition to adjusting 2021 s-curves based on changes in LLPAs. After this reconstruction, we see that ramped investor and 2^ndhome s-curves are roughly in line with one another ( Figure 61). This gives some empirical comfort to the fundamental truth that the LLPAs are the same and should provide equivalent protection as discussed in the earlier section of this piece.

time borrowers of owner-occupied homes because they can finance their interest payments with rental income, and thus are under less pressure to pay down their loan at inception.
 

Figure 62: 2^nd home and investor collateral protect against the worst of the 1-4 WALA curtailment spikes, though 2^nd home curtailments are slightly faster than investor curtailments

UMBS 30 owner occupied (not first time borrower), 2^nd home, investor, and owner occupied (first time borrower) 1 mo. curtailment CPR aging ramps observed Apr.-Jun. 2024, for purchase-only collateral, >250k orig. lnsz, FICO > 720,

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

Our model is currently tuned to slightly faster 2^nd home loans speeds roughly equivalent to a 25bp elbow/+0.1 refi multiplier relative to investor loans, using the characteristics of the FNCL 6.5 TBA ( Figure 63 and Figure 64). On this collateral, our model thinks payups should be 23 ticks for investor pools (2 ticks below market) and 14 ticks for 2^nd home (3 ticks above). Even with this more aggressive calibration on speeds than LLPAs alone would suggest, the 9 tick model payup gap is 5 ticks larger than the market is currently pricing 6.5s. Pools of course will vary relative to the TBA in other stats (loan sizes are lower) so its worth looking at the exact collateral.
 

Figure 63: Our model is currently calibrated to 2^nd home loans prepaying faster than investor…

FN 6.5 TBA CPR projections in our beta research model, vs. investor and 2^nd home (spec types are generated by overriding FN 6.5 occupancy type)

Source: J.P. Morgan

Figure 64: … reflected in a slight elbow shift and higher refi multiplier

FN 6.5 TBA fwd LT CPRs under a range of parallel rate shocks scenarios (flipped to mimic incentive, 0 set at a rates up 100bp scenario), in our beta research model, vs. investor and 2^nd (spec types are generated by overriding FN 6.5 TBA occupancy type)

Source: J.P. Morgan

Week in Review

• MBA Weekly Survey: For the week ending August 16, the purchase application index fell 5.2% w/w and was 8% lower than year ago levels, while the refinance index fell 15.2% w/w and was 27.7% higher than the 3-month trailing level (daycount-adjusted, not seasonally-adjusted) ( Figure 62 & Figure 63).
• Freddie Enhanced Primary Survey: For the week prior to August 22, 2024, 30-year conventional conforming fixed-rate mortgages averaged 6.46%, down 3bp from the previous week ( Figure 64).
• Primary dealer specified pool positions rose to $456.5bn (+$4.8bn w/w) as-of close trading August 7. Including TBA positions of -$392.8bn, dealers were long $63.7bn $59.8bn (+$3.9bn w/w) pass-throughs. Other agency MBS holdings rose$0.4bn to $30.8bn.
• Fixed-rate agency gross and net issuances were $95.9bn and $19.0bn, respectively, in July. August gross supply currently stands at $85.9bn ( Figure 65).
• ICI Total Bond Long-term Mutual Fund and ETF Weekly Flows: Inflows were +$6.3bn for the week of August 14, and +$26.2bn for the four weeks leading up to August 14 ( Figure 66).

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

Source: J.P. Morgan, MBA

Figure 66: MBA Refi Indices, seasonally adjusted

Source: J.P. Morgan, MBA

Figure 67: Primary mortgage rates, %

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

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

Source: J.P. Morgan

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

Source: J.P. Morgan, ICI

Non-QM to agency

• AAA non-QM is back to 130bp in the new issue and 6 coupon jumbo PTs are now at 1-04bk, tighter than pre-August rate volatility. We expect spreads to remain rangebound through the last four months of the year as the macro outlook remains challenging and election season quickly approaches
• After we introduced our framework to value the option cost in non-QM, investors were quick to point out the value of the 100bp step-up in the AAA to A coupon
• In other words, there are likely some rate scenarios where the issuer is indifferent between calling the transaction when the collateral price is at a modest discount and paying the 100bp coupon step-up post the four year call
• We implemented a new functionality in MSC that allows us to call the transaction at different collateral prices
• The middle of the coupon stack sees the most impact when the deal is called below par. For example, 6% coupon AAAs structured off 6-7.5 WAC collateral lose 15-30bp of OAS for every 2-point decrease in the collateral price threshold
• Non-QM borrowers are ineligible for agency loans due to a lack of documentation, prior credit event and/or specific loan features. However, can these borrowers eventually qualify and refinance into an agency loan?
• We quantify how non-QM borrowers typically transition to other loan types, providing clarity on borrower refinancing behavior
• Our new non-QM prepayment model will be released in MSC today, August 23rd. We detail how to run the model in MSC

Figure 70: RMBS credit issuance to date...

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

Figure 71: ...and spreads

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

Market Commentary

Mortgage credit spreads retraced much of the widening seen in the first week of August. AAA non-QM is back to 130bp in new issue and 6 coupon jumbo PTs are now at 1-04bk, tighter than pre-August rate volatility. We expect spreads to remain rangebound through the last four months of the year as the macro outlook remains challenging and election season quickly approaches.

New issue has continued to be very active throughout the summer. July saw $12bn of new issue RMBS price. August has seen a moderate slowdown with $7bn issued thus far, and the primary market is expected to be light in the last week of the summer.

Non-QM collateral call price vs. option cost

In July, we introduced a framework to value the call option cost in non-QM. In this framework, we run non-QM deals through OAS rate paths, calculate the collateral price at a defined spread in each month post the three-year call date and terminate deals in scenarios where the collateral price is above par. After we introduced this framework, investors were quick to point out the value of the 100bp step-up in the AAA to A coupon. In other words, there are likely some rate scenarios where the issuer is indifferent between calling the transaction when the collateral price is at a modest discount and paying the 100bp coupon step-up post the four-year call. This means that we would be undervaluing the call option cost if we allow deals to be called only when the collateral price is above par. We are currently working on a framework to correctly value the step-up in our non-QM call framework. In the meantime, we have added a new functionality in MSC to allow us to call a given transaction at a different (specified) collateral price.

Figure 72 shows model results when we vary collateral call thresholds, keeping the rest of the framework the same. We terminate deals when the collateral price exceeds $90 to $100, in 2-point increments. The middle of the coupon stack sees the most impact when the deal is called below par. For example, 6% coupon AAAs structured off 6-7.5 WAC collateral lose 15-30bp of OAS for every 2-point decrease in collateral price threshold. Higher coupons, on the other hand, see less of an impact, with AAAs structured off 9 WAC pools losing only 10bp OAS when calling the deal at $98 instead of $100. This is not surprising as the call is in the money across most of the scenarios for higher WAC paper, while cuspy coupons end up losing significant value even with small changes in call assumptions.

Figure 72: The middle of the coupon stack sees the most impact when the deal is called below par

J.P. Morgan non-QM prepayment model run

Source: J.P. Morgan
 

Non-QM to agency curing

Non-QM borrowers are not eligible for agency loans due to a lack of documentation, prior credit events and/or specific loan features. However, can these borrowers eventually qualify and refinance into an agency loan? Documentation type is a big differentiator of prepayment speeds in non-QM, and its composition has changed over the years. When the non-QM sector first emerged, full documentation borrowers accounted for approximately 30% of originations. Today, this cohort represents only 5% of non-QM loans ( Figure 73). It is likely that a significant portion of full documentation borrowers have refinanced into agency loans, especially following the QM rule change in 2021. In this section, we quantify how non-QM borrowers typically transition to other loan types, providing clarity on borrower refinancing behavior.

Figure 73: Documentation type is a big differentiator of prepayment speeds in non-QM, and its composition has changed over the years

Securitized non-QM originations by loan documentation type

Source: J.P. Morgan, CoreLogic

For this analysis, we use Black Knight’s McDash database to identify non-QM loans and observe what loans borrowers refinance into.¹ McDash does not explicitly identify non-QM loans, so we use whole loans with a spread over the jumbo rate >75bps (for jumbo loans) or spread over the conforming rate >75bps (for conforming loans) as a proxy for the sector. Figure 74 shows loan origination balances for the segment we identify as non-QM in McDash versus IMF (‘expanded credit’) and all securitized non-QM. The balance of loans we identify as non-QM in McDash is small, and the proportion where we can identify the loan the borrower transitioned to next is even smaller (see black bars in the figure) – around $600mn to $4bn in each vintage bucket. In some vintages like ‘21 and ‘22, McDash non-QM balances are close to, or higher, than the IMF origination volumes. Nevertheless, as we show below, the loan strats are comparable, and in the absence of more precise ways to identify non-QM, we work with this dataset below, with these caveats.

Figure 74: The balance of loans we identify as non-QM in McDash is small, but in the absence of more precise ways to identify non-QM, we work with this data below

Origination balance by loan vintage for McDash non-QM WL with an identified next loan after being paid in full, all McDash non-QM WL, IMF expanded credit and all securitized non-QM

Source: J.P. Morgan, Black Knight, IMF

Figure 75 shows collateral strats comparing the identified McDash non-QM loans to CoreLogic securitized non-QM loans. Average spread by vintage ranges from 180-242bp over the conforming rate, which is in a similar, if not slightly higher range than securitized non-QM. The vintage-level FICO, LTV and WACs are comparable. Our McDash AOLS are slightly smaller than securitized non-QM, which could mean that we are capturing some additional agency loans.

Figure 75: Loan strats are fairly comparable between the identified McDash non-QM and CoreLogic’s securitized non-QM

Collateral strats for identified McDash non-QM vs. CoreLogic securitized non-QM

Source: J.P. Morgan, Black Knight, CoreLogic

With this dataset, we look at what kinds of loans non-QM borrowers transitioned to after paying the initial non-QM whole loan in full. We first look at the existing mortgage universe. Purchase loans, which refer to borrowers who sold their prior home, or who are new home owners, have historically accounted for the majority of loan purpose types at roughly 65% of transactions ( Figure 76). In ‘20 and ‘21, however, in the midst of historically low rates, purchase loans accounted for only ~40% of transactions, while rate/term refis increased to 35-45%, exceeding cashout refis.

McDash is only able to identify the agency type of the next loan in refi transactions and not purchase transactions, so we limit our analysis to the former. In refi transactions, earlier vintages saw more borrowers transition to GSE loans, with the share peaking in the 2018 vintage ( Figure 77). Later vintages have seen more of a shift towards WLs, but the volume of borrowers prepaying for refi has also drastically decreased. Across the vintages, borrowers that transition to Ginnie loans from the initial non-QM WL see the largest %-pt decrease in WAC, followed by GSE and then WL.

Figure 76: Purchase loans have historically accounted for the majority of loan purpose types, while refi transactions increased in ‘20 and ‘21

Refi/cashout/turnover split for all mcdash by loan vintage and loan purpose, for all McDash loans

Source: J.P. Morgan, Black Knight

Figure 77: Earlier vintage non-QM WLs saw more borrowers transition to GSE loans after payoff, with that share peaking in 2018 vintage loans

Proportion of loans by loan vintage and next loan type. Only includes borrowers that did refi transactions

Source: J.P. Morgan, Black Knight

We break this down further. Non-QM borrowers who transition to Ginnies largely do cashout refi transactions, and this share has remained roughly the same over time ( Figure 78). These borrowers could potentially be less financially savvy in terms of refi options, or, more likely, need the cash on hand for other purposes. On the other hand, borrowers who transition to GSE loans largely do a mix of cashout and rate refis ( Figure 79). With very little balance in later vintages, we take these shares with a grain of salt.

Figure 78: Borrowers who transition to Ginnies largely do cashout refi transactions…

Refi transaction type by loan vintage for non-QM WLs that transitioned to a Ginnie loan, share by loan count. Only includes borrowers that did refi transactions

Source: J.P. Morgan, Black Knight

Figure 79: …while borrowers who transition to GSE loans do a mix of rate and cashout refis

Refi transaction type by loan vintage for non-QM WLs that transitioned to a GSE loan, share by loan count. Only includes borrowers that did refi transactions

Source: J.P. Morgan, Black Knight

Limitations in the dataset prevent us from further breaking out the data by initial documentation type. Still, we can make some inferences based on what we know from securitization originations. For instance, more recent vintages reflect more investor and bank statement loan performance and fewer full doc loans, and vice versa for earlier vintages, although it is hard to attribute changes in performance to doc type alone given the vastly different economic environments. While the dataset we have remains a fraction of the overall market, this gives us a better understanding of borrower behavior in refi transactions in non-QM.

Non-QM prepayment model will be released today

Our new non-QM prepayment model will be released in MSC at close today, August 23rd. Figure 80 shows the temporary MSC settings required to correctly run the model in MSC. These settings will become the default in the upcoming October MSC release. At that point, users will only need to select Research Settings in the calculator.

For now, please use the settings in the table below. Users will need to set three Environment Variables under Preferences (gear icon, top right of MSC) and add some additional columns and column inputs. Once this is done, users can run both the loan-level and pool-level models. When running OAS, we recommend using the pool-level model, which can be set by using Run Pool Level Model = Yes.

Users will be able to run our non-QM call framework using the following specifications:

• To run our call value framework, (1) select Pricing Call under Additional Outputs and (2) add OAS Call Spread column from Column Configuration.
• To run the call with different collateral price, use the PRICING_CALL_THRESHOLD environment variable.
• Users can also run deals to an immediate three year call by setting CMO Run to Call = Yes.

Figure 80: Our new non-QM prepayment model will be released in MSC as of close today, August 23rd

Source: J.P. Morgan

2023 CMBS financials update - operating expense growth moderates but still running hot

• It was a relatively quiet week for CMBS this past week with only two new issue deals pricing. Benchmark CMBS spreads were slightly tighter with 10yr conduit CMBS LCF AAAs 1bp tighter week-over-week at J+102 and 10yr Freddie K A2s 1bp tighter at J+52.
• Full year 2023 conduit CMBS property financial data indicates that net cashflow (NCF) growth has slowed materially compared to the previous two years. This slowdown was due to decelerating top-line revenue growth and operating expense growth, which, while decelerating, remained elevated
• By core property type, we see that lodging and office saw NCF shrink in 2023. We are beginning to see the sticky effects of inflation on an asset class like lodging that is highly operationally levered with an expense ratio of about 70% historically
• We have seen operating expenses grow materially for multifamily when examining Freddie K data but top-line revenues for Freddie K (and conduit CMBS for that matter) loans grew by 6.8%, leading to a relatively healthy bottom-line NCF growth of 6.1% in 2023. This supports the notion that while rent growth on new leases may have decelerated materially, rent growth for lease renewals likely held up much better
• We find that labor-related expenses broadly categorized as payrolls and expenses (property management employee costs) and repairs and maintenance in the OSAR data contributed most to elevated operating expense growth in the multifamily sector over the last few years
• The growth in insurance premiums have been eye-popping and may continue to grow at elevated rates as it appears climate risk is underpriced in the real estate market. Unsurprisingly, metro areas in climate/natural disaster affected areas (broadly Florida and Texas MSAs) have seen the most acute growth in insurance premiums

Weekly market snapshot

Market commentary - upper IG mezz look cheap

It was a relatively quiet week for CMBS this past week with only two new issue deals pricing compared to a more frenetic pace in recent weeks. Benchmark CMBS spreads were slightly tighter with 10yr conduit CMBS LCF AAAs 1bp tighter week-over-week at J+102 and 10yr Freddie K A2s 1bp tighter at J+52. Benchmark CMBS valuations are fair relative to their corporate and mortgage comps as they have been for some time now. Conduit IG mezz tightened more, with AS to single-As tightening by 8-16bp week-over-week versus their corporate comps that were roughly flat. We had noted better relative value in IG mezz in recent weeks (see here and here).

Spread volatility around major economic data releases (particularly employment reports) remain a concern but we view loss risk to IG mezz as very limited under our Base Case loss projections ( Figure 83). Assuming our house view of a slowing economy avoiding a recession materializes, we should continue to see the spread curve bullishly flatten. Given near-term risks to data, we favor the upper IG mezz bonds (AS and AA bonds) that still look cheap to corporate comps but also to the spread curve ( Figure 84).

Figure 83: We continue to see limited writedown risks for conduit IG mezz

J.P. Morgan Conduit CMBS loss model cumulative loss projections

Vintage / Property Type		Base	Bearish
All 2.0	2010-24	5.0%	6.0%
Vintage	2010	2.6%	2.6%
	2011	2.7%	2.7%
	2012	4.1%	4.1%
	2013	3.8%	3.8%
	2014	4.9%	4.9%
	2015	4.4%	4.5%
	2016	4.9%	5.2%
	2017	5.5%	6.3%
	2018	5.2%	6.2%
	2019	6.9%	8.4%
	2020	6.5%	8.1%
	2021	5.8%	7.4%
	2022	7.6%	9.9%
	2023	4.7%	6.1%
	2024	2.2%	3.1%
Property Type	Office	8.8%	11.8%
	Retail	4.3%	4.3%
	Lodging	3.4%	3.4%
	Multifamily	1.7%	1.7%
	Industrial	1.4%	1.4%
	Other	1.8%	1.8%
Average Loss	AAA	0.0%	0.0%
By Original	AA	0.0%	0.2%
Rating	A	0.7%	3.2%
	BBB+	7.7%	13.1%
	BBB	6.3%	26.6%
	BBB-	9.8%	16.4%
	BB	33.9%	46.3%
	B	55.9%	62.4%
	NR	81.6%	83.4%

Source: J.P. Morgan

Figure 84: Conduit CMBS upper IG mezz are cheap on the spread curve

Residuals from linear regressions of 10yr conduit CMBS AS and AA Treasury spreads to those of adjacent tranches (bp)

Source: J.P. MorganNote: Modeled spreads linearly regresses 10yr conduit CMBS AS spreads against 10yr conduit CMBS LCF AAA and 10yr conduit CMBS AA spreads over the last 10 years (R2 = 96%,SE = 8.8bp) and 10yr conduit CMBS AA spreads against 10yr conduit CMBS LCF AS and 10yr conduit CMBS A spreads over the last 10 years (R2 = 97%,SE = 12bp).

While Agency CMBS has tightened week-over-week, we still see 5yr Agency CMBS as leaning cheap relative to current coupon Treasury OAS and ZVs ( Figure 85).

Figure 85: 5yr Agency CMBS still leaning cheap relative to current coupon mortgage OAS and ZV

5yr Freddie K A2 (actual) versus modeled spread and difference between actual and modeled spreads (residual), bp

Source: J.P. MorganNote: Modeled spreads linearly regresses 5yr Freddie K A2 spreads against current coupon FNMA 30yr Treasury OAS and current coupon FNMA 30yr Treasury ZV spreads over the last 5 years. R2 = 72.4%,SE = 8.7bp.

Ratings Tracker

Figure 87: CMBS ratings downgrades to upgrades ratio since March 2021

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

Primary Markets

2023 CMBS financials update - operating expense growth moderates but still running hot

Full year 2023 conduit CMBS property financial data indicates that net cashflow (NCF) growth has slowed materially compared to the previous two years. This slowdown was due to decelerating top-line revenue growth and operating expense growth, which, while decelerating, remained elevated ( Figure 89). In our view, as top-line revenue growth continues to normalize, expense management will become a greater focus going forward.

Figure 89: Conduit CMBS net cashflow growth has moderated substantially on slower revenue growth and elevated operating expense growth

2014-2019 vintage conduit CMBS same-store annual changes in revenues, operating expenses, and net cashflows

Source: J.P. Morgan, Trepp

By core property type, we see that multifamily, industrial, and retail conduit CMBS loans were able to grow NCF in 2023 while lodging and office saw NCF shrink ( Figure 90). Looking back to longer same-store data from 2019, we see that conduit office loans have seen sequential declines in NCF with 2023 full-year NCF now sitting 6% below underwritten levels ( Figure 91).

Figure 90: Multifamily, industrial, and retail properties backing conduit CMBS grew net cashflows in 2023 while lodging and office saw net cashflows shrink...

2014-2019 vintage conduit CMBS same-store year-over-year change in revenues, operating expenses, and net cashflows by property type, as of 2023

Source: J.P. Morgan, Trepp

Figure 91: ...leading to aggregate office and lodging net cashflows that are now below underwritten levels

2014-2019 vintage conduit CMBS same-store net cashflow changes, cumulatively indexed to underwritten levels = 100

Source: J.P. Morgan, Trepp

The structural ails of the office market are understood but more interestingly, conduit lodging loan NCF is, not only lower in 2023 year-over-year, but is about a point lower relative to underwritten levels as slowing top line revenue growth is met with growing operating expenses ( Figure 92). We had long thought that the inflation pressures posed risks to NCF growth in the lodging market. We are beginning to see the sticky effects of inflation (labor, cost of goods sold) on CRE, especially in an asset class like lodging that is highly operationally levered with an expense ratio of about 70% historically ( Figure 93).

Figure 92: Conduit CMBS loans backed by lodging are seeing net cashflows drag...

2014-2019 vintage conduit CMBS lodging same-store revenue, operating expense, and net cashflow changes, cumulatively indexed to underwritten levels = 100

Source: J.P. Morgan, Trepp

Figure 93: ...given high operating leverage as revenue growth slows and expense growth remains high

2014-2019 vintage conduit CMBS same-store operating expense ratio to revenues by property type

Source: J.P. Morgan, Trepp

Figure 94 shows the sensitivity of NCF growth to revenue growth for the core property types according to their operating leverage (operating expense to revenue ratios) under a hypothetical fixed 6% expense growth scenario. At this expense growth assumption, the breakeven revenue growth for lodging is 4-5%. The other property types, which are less operationally leveraged, have breakeven revenue growth rates of either 1-2% (retail and industrial) or 2-3% (multifamily and office).

Figure 94: Under a 6% expense growth assumption, breakeven revenue growth is in the +1-5% vicinity

Hypothetical NCF growth under different revenue growth assumptions at each property type’s average operating expense ratio and a +6% operating expense growth assumption

Source: J.P. Morgan

Figure 95 shows vintage and property type cuts of conduit CMBS same-store financials performance.

Multifamily requires a special mention as the sector has faced some headwinds due to elevated construction deliveries. As we wrote in an earlier weekly, we believe multifamily fundamentals are bottoming out as supply deliveries should wane materially next year and demand should remain durable. However, one often-discussed risk to multifamily, is the growth in operating expenses.

Indeed, we have seen operating expenses grow materially for multifamily when examining Freddie K data ( Figure 96). Operating expenses ballooned by 8.8% in 2022 but moderated to 7.6% in 2023. Interestingly, despite the rapid deceleration of multifamily effective rent growth on new leases in 2023, we found that top-line revenues for Freddie K (and conduit CMBS for that matter) loans grew by 6.8%, leading to a relatively healthy bottom-line NCF growth of 6.1% in 2023. This datapoint further supports the notion that while rent growth on new leases may have decelerated materially into heavy supply deliveries in recent quarters, rent growth for lease renewals likely held up much better, allowing landlords to pass on a good chunk of expense growth to tenants. Still, operating expense ratios to revenues have ticked up over the years as expense growth has exceeded revenue growth in the years following the pandemic.

Figure 96: Freddie K multifamily operating expense growth moderated in 2023 versus 2022 but remain elevated. That said, rent growth kept up in 2023

Freddie K backed multifamily same-store annual changes in revenues, operating expenses, net cashflows, and operating expense to revenue ratio

Source: J.P. Morgan, Trepp

What’s driven expense growth in recent years? Using a subset of the Freddie K 2019-2023 same-store financials data that has clean, detailed expense breakouts, we find that labor-related expenses broadly categorized as payrolls and expenses (property management employee costs) and repairs and maintenance in the OSAR data contributed most to elevated operating expense growth over the last few years given the categories' large share (~40%) of total operating expenses ( Figure 97). Combined, the labor-related expenses contributed to about 40% of the total operating expense growth in 2023 of 7.6%.

Figure 97: Labor-related operating expenses and rising insurance premiums have been the largest drivers of Freddie K multifamily operating expense growth

Freddie K backed multifamily same-store annual changes in the largest operating expense categories, their contribution to total operating expenses, and their contribution to the year-over-year 2023 total operating expense growth

Source: J.P. Morgan, Trepp

The growth in insurance premiums have been eye-popping (a phenomenon that continues to impact real estate broadly given increasing climate/natural disaster risks) but comparatively, insurance premiums still only account for ~7% of total operating expenses (up from ~4% in 2019). However, given how much this category jumped in 2023, it was the single largest contributor to overall operating expense growth at 28%.

Real estate taxes are another large expense item, but the growth rates here have been tamer relative to labor-related costs and insurance premiums. The labor-related and real estate tax expense categories can moderate going forward. Labor costs should track broader employment trends, which is cooling and given the ~20% decline in multifamily property prices since QT began, real estate tax assessment values, at least in theory, should moderate, which may lead to lower real estate tax growth rates.

Insurance premiums may continue to grow at elevated rates as it appears climate risk is underpriced in the real estate market. Unsurprisingly, metro areas in climate/natural disaster affected areas (broadly Florida and Texas MSAs) have seen the most acute growth in insurance premiums ( Figure 98). Here, we use a larger sample of Freddie K loans that captures loans with valid same-store OSAR data over 2022 and 2023 only to increase information value when cutting the data to MSA level. These are also some of the areas where total operating expense growth is being significantly driven by the outsized growth in insurance premiums.

Figure 99 shows the expense line items that feed the top 4 OSAR expense categories that contributed the most to total operating expense growth in 2023.

Weekly Tracker

Figure 100: Delinquency rate

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

Source: J.P. Morgan, Trepp

Figure 101: Delinquency cure rates

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

Source: J.P. Morgan, Trepp

Figure 102: Specially serviced rate

Conduit CMBS percentage of loans in special servicing (%)

Source: J.P. Morgan, Trepp

Figure 103: Office RTTO indexed to pre-pandemic levels

Kastle System Back to Work Barometer, weekly

Source: Kastle Systems, Bloomberg Finance L.P.

Cross Sector Spreads

Cross Sector Spreads (continued)