Opening
The packet is dominated by one story: a global bond sell-off that nothing yet seems to stop, colliding with a Fed that is still flirting with a September hike, a yen whose strength is now policy-engineered, and equity desks split between staying pro-risk and going tactically defensive. Today's NFP preview and next week's ECB/CPI prints are the near-term catalysts.
Evidence IDs · P002 · P032 · P021 · P030
Top read
Natixis: 'Who will stop the bond sell off, please?' — French 30-year above 5% for the first time since 2008, QT plus a Q4 US Treasury absorption wall of $628bn, and no buyer of last resort in sight.
Evidence IDs · P032
Market posture
Posture is cautiously pro-risk but hedged: GS GOAL stays OW equities for 12m while tactically neutral for 3m, JPM flags quantity-based financial conditions still supportive even as price-based metrics tighten, and BofA argues midterms amplify dispersion more than vol — favour cheap, targeted hedges over index protection. Bond carry math (DB) means the worst of negative total returns is likely behind us even if yields grind higher.
Evidence IDs · P021 · P027 · P001 · P006
Themes
Bond sell-off: normalization, not (yet) fiscal crisis — and carry now cushions the ride
French 30-year above 5% for the first time since 2008, but DB says Treasuries can rally in total-return terms even to 5.5%.
Verdict
Higher global yields are the base case, but the mechanism is normalization out of financial repression plus supply indigestion, not a fiscal-credibility break; long-duration total returns are far more resilient than in the early 2020s.
Analysis
Natixis frames the sell-off as momentum with no visible stopping force: eurozone/UK QT is still withdrawing support just as a wall of issuance looms, and in the US a further $628bn net supply must be absorbed in Q4 — with competition from JGBs and hyperscaler-friendly credit. France is the pressure point (30-year OAT above 5%, 10-year OAT-Bund spread at 88bp, highest since 2012). DB pushes back on the fiscal-panic narrative: the latest rise is 'a continuation of the normalisation after the financial repression of the 2010s', a second-order fiscal effect at most. Crucially, the carry math has changed: the Bloomberg US Treasury Total Return Index is up ~1% over the past year despite 10-year yields rising ~60bps, and 10-year yields could reach roughly 5.5% in a year (6.4% over two) before total returns turn negative. The counterargument — that this is the start of de-dollarisation/Treasury privilege loss — Natixis says is only truly testable if the AI-productivity thesis disappoints. Portfolio consequence: don't extrapolate early-2020s bond losses; shorting duration now fights positive carry, while European sovereign spreads (France, BTP-Bund ~84bp) remain the cleaner stress expression.
Tension
Normalization versus indigestion: DB sees cyclical normalization; Natixis emphasizes QT and supply absorption risk.
What changed
French 30-year crossed 5% (first since 2008); OAT-Bund spread at post-2012 wides; US Q4 net supply of $628bn now quantified.
Desk views
Natixis sees nothing halting the rise in yields: QT in the eurozone and UK leaves investors facing the issuance wall, with US Q4 absorption of $628bn the key indigestion risk.
Evidence IDs · P032
DB argues the rise is normalization, not fiscal panic, and that carry means 10-year yields could hit ~5.5% over a year before total returns go negative — making shorts expensive and the worst bond losses likely behind us.
Evidence IDs · P006
Investor implication
Stay short-biased on European sovereigns/spreads rather than Treasuries; avoid paying up for duration shorts given carry; watch Q4 supply and hyperscaler credit competition as the swing factors.
Trade expression
Long OAT-Bund spread widening / BTP-Bund stress; neutral-to-carry-positive Treasuries rather than outright shorts.
Invalidation
A recession shock, AI-led equity bust, or credible fiscal consolidation that breaks the yield momentum.
Transition
The same yield pressure is what forced the US Treasury into interventions — and pushed Japan into the largest FX operation on record.
Evidence IDs · P032 · P006 · P007
Yen: intervention, quid pro quo, and stretched BoJ pricing
Verdict
Yen strength is policy-backed and likely to persist near-term, but markets have overpriced BoJ aggression: MUFG sees a September hike as likely yet a >25bp move as very unlikely.
Analysis
Japan's intervention since end-July reached JPY15.4tn (~USD 96.5bn) — nearly three times the July 2024 operation — plus at least USD 10-15bn sold directly by the US Treasury, which UniCredit reads as Washington protecting Japan's USD 1.11tn Treasury holdings ahead of its 9 September buyback program. Yet the joint operation has been less successful than 2024's solo action: USD/JPY fell less than 10 big figures versus 20+ then, and the pair remains not much below 160 (low of 155.21 on 3 August). MUFG notes Takata's speech hinting the BoJ could hike by more than 25bps has driven 30bps of tightening priced for this month and ~100bps over a year — pricing MUFG calls stretched: a hike this month 'will only just confirm a pick-up in the pace from every six months to every three months'. The quid pro quo narrative (Bessent urging Japan to 'stop the reflation', GPIF's unusual August meeting) supports a structural yen-revaluation story, but fiscal concerns and Japan's trade deficit cap the upside. Portfolio consequence: yen strength is credible but the path is intervention-managed; front-end JGB/BoJ pricing is the stretched leg, not spot.
Tension
Markets price ~100bps of BoJ tightening over 12m; MUFG doubts any single hike exceeds 25bps.
What changed
Record JPY15.4tn intervention with US Treasury participation; Takata's >25bp hint; September hike now largely priced.
Desk views
MUFG: BoJ pricing looks stretched — 30bps priced for this month and ~100bps over a year, but MUFG sees >25bps as unlikely and even back-to-back moves doubtful.
Evidence IDs · P030
UniCredit: this intervention was ~3x July 2024's and involved the US Treasury (likely to stop Japan selling Treasuries), yet was less successful — USD/JPY fell under 10 big figures vs 20+ in 2024.
Evidence IDs · P034
MUFG JPY Monthly: a September hike is largely priced as part of the quid pro quo for US support, but rapid hikes toward neutral are uncertain and fiscal concerns weigh on the yen.
Evidence IDs · P031
Investor implication
Fade outsized BoJ hike pricing rather than spot yen strength; watch 18 September BoJ and any renewed joint-intervention signals around 160.
Trade expression
Long JPY vs high-carry G10 funded legs; receivers in Japanese front-end vs market-implied ~100bps of tightening.
Invalidation
BoJ delivering a >25bp hike, or US pressure on Japan receding (removing the quid pro quo bid).
Transition
Yen strength was one knock-on of the dollar/rates complex — where the Fed's own path is the bigger open question.
Evidence IDs · P030 · P031 · P034
Fed: September hike odds hinge on CPI, not payrolls
Verdict
A September hike remains the base case for several desks, but the August CPI print on 11 September — not Friday's NFP — is the decisive report.
Analysis
CACIB expects a 'mixed bag' August jobs report: NFP rebounding to +65k after July's -23k decline, unemployment edging up to 4.2% on participation payback, and AHE +0.3% MoM taking the YoY rate down to 3.1%, the lowest since 2021 — reinforcing the Fed's view that labour is not an inflation source. Most FOMC members see the labour market as stable, so an in-line report changes little; CACIB explicitly flags the August CPI (11 September) as the more consequential report before the September FOMC. ING agrees the bar to derail a September hike is high for second-tier data, retains a USD-upside preference on front-end rates and energy, but flags the risk that higher back-end yields prompt more Treasury interventionism and a revamp of the debasement trade. GS's desk adds nuance: Fed speakers say employment data is 'just not key for them at the moment', yet payrolls still move the USD and front end, with slight negative skew risk into Waller and NFP. Market pricing sits near 64% odds of a hike (JPM). Portfolio consequence: long USD/front-end payer positioning is consensus but crowded; the asymmetric risk is a soft CPI that unwinds September-hike pricing quickly.
Tension
Labour-market weakness (July NFP -23k) versus a Fed focused almost entirely on inflation.
What changed
July's outright NFP decline and downward revisions; September-hike pricing dipped from 18bp to 15bp as the oil rally stalled.
Desk views
CACIB: expect NFP +65k, unemployment 4.2%, AHE YoY 3.1% — but CPI on 11 September is the more consequential report for the September FOMC.
Evidence IDs · P002
ING: bar to drive the Fed away from a September hike is high; retain USD upside, but watch Treasury interventionism and the debasement trade as the main risk.
Evidence IDs · P026
GS trading desk: Fed speakers say employment data is not key, yet payrolls still move USD/front-end; slight negative USD skew into Waller and NFP.
Evidence IDs · P022
Investor implication
Position for CPI as the volatility event; USD longs are consensus and vulnerable to a dovish inflation surprise.
Trade expression
Own CPI-event vol rather than pre-positioning heavily in USD; payers in the front end only with cheap convexity.
Invalidation
A major NFP downside surprise trimming September odds, or CPI confirming labour-led disinflation.
Transition
If the Fed does hike into resilient growth, equities face a rates speed limit — which is exactly how the strategy desks are positioning.
Evidence IDs · P002 · P026 · P022 · P028
Equities: stay invested, but the rotation is real and hedges are cheap
Verdict
Remain OW equities for 12m but tactically neutral for 3m; buy dips, diversify, and use selective overlays — with midterms a dispersion story, not a vol story.
Analysis
GS GOAL keeps a modestly pro-risk 12m stance (OW equities, N bonds, commodities, cash; UW credit) on slowing-but-sturdy growth and healthy earnings, but warns near-term volatility could rise amid rates vol, seasonal weakness, midterms and geopolitics — and that 'bond markets have become a speed limit for equities'. JPM's US Market Intel moves to a Tactically Cautious/Neutral stance, a genuine disagreement with GS's tactical neutrality-toward-risk. BofA's vol work resolves the election question: S&P options price just a 0.82% post-election move with little VIX event premium, and with easing equity leverage and low correlations the midterms should be rotation- rather than vol-inducing — so avoid overpaying for index protection and prefer cheap hedges (S&P PDOs ~50% cheaper than put spreads) plus targeted IWM Nov puts where regional bank/small-cap event vol looks underpriced. JPM Flows & Liquidity adds the quantity-based counterpoint: credit and money creation show financial conditions remain supportive even as price-based metrics tighten, and investors are long equities/bonds only modestly. Portfolio consequence: the risk is not a broad drawdown but dispersion — AI capex beneficiaries have sold off while breadth broadened — so express views at the sector/single-stock level, not the index.
Tension
GS stays pro-risk for 12m; JPM goes tactically cautious/neutral. Midterms: BofA sees dispersion, JPM's election series implies broader event prep.
What changed
AI capex beneficiaries selling off with significant broadening after 1H's narrow rally; US Treasury interventions in long-dated Treasuries weighing on the Dollar and boosting Gold.
Desk views
GS GOAL: remain OW equities for 12m, tactically neutral 3m; equities outperformed bonds as longer-dated yields neared post-GFC highs, and bonds are now a speed limit for equities.
Evidence IDs · P021
JPM US Market Intel: moving to a Tactically Cautious/Neutral stance with an updated Monetization Menu and a Q4 roadmap of bull/bear event outcomes.
Evidence IDs · P028
BofA: midterms amplify dispersion more than vol — S&P options price only a 0.82% post-election move; favour cheap hedges (S&P PDOs, VIX calendar collars) and IWM Nov puts for targeted protection.
Evidence IDs · P001
JPM F&L: quantity-based metrics (credit, money creation) show financial conditions still supportive despite price-based tightening; market liquidity deteriorated most in cash USTs.
Evidence IDs · P027
Investor implication
Stay long equities at the 12m horizon but fund it with cheap, targeted hedges; rotate toward broadening winners rather than crowded AI capex names.
Trade expression
OW equities with S&P PDO overlays and IWM Nov puts; express dispersion via sector pairs rather than index vol.
Invalidation
A bond yield shock that breaks the 'speed limit', or earnings revisions rolling over.
Transition
Within equities, Europe offers the valuation-and-cash-flow diversification the US no longer does.
Evidence IDs · P021 · P028 · P001 · P027
Europe: strong micro meets challenging macro — and the ECB may be less hawkish than priced
Verdict
European equities remain attractive on earnings, cash flow and diversification, but gas-driven yield pressure and an overly hawkish ECB curve are the key risks.
Analysis
The macro headwind is real, though: DB's Early Morning Reid records fresh record yields across Europe on relentless gas gains (front-month TTF at €73.60/MWh, highest since January 2023), with the 10yr Bund at a post-2011 high of 3.37% and France's OAT at a post-2008 high of 4.25%. ING argues markets look too hawkish on the ECB — three hikes by April 2027 embedded in the swap curve 'seems overly hawkish considering core inflation has remained so well behaved' — and that policymakers may care more about widening bond spreads than second-round inflation, arguing for a less hawkish message than priced. CACIB expects the ECB's 10 September projections to revise 2027 HICP up (+30bp to 2.6%) on gas/electricity second-round effects while trimming 2026. Portfolio consequence: long European equity beta with a hedge against the gas/yield channel; the ECB meeting is a two-sided event — a less hawkish ECB helps bonds and equities, but energy-driven HICP revisions argue against dovish repricing.
Tension
Equity-friendly earnings momentum versus gas-price-driven bond and spread pressure; hawkish ECB pricing versus ING's spread-first read.
What changed
Bund at post-2011 high (3.37%), OAT at post-2008 high (4.25%), TTF at highest since January 2023; ECB projections due 10 September.
Desk views
GS: 2026 EPS growth forecast raised to 15%; Europe trades at 14.5x forward P/E with a 5.3% FCF yield vs <2% for the 10 largest US stocks; favour Renewables and HALO, UW Autos/Chemicals on Chinese competition.
Evidence IDs · P024
ING: ECB swap curve embedding three hikes by April 2027 looks overly hawkish given well-behaved core inflation; policymakers may prioritise widening spreads over second-round risk.
Evidence IDs · P026
CACIB: expect ECB to revise 2027 headline HICP up 30bp to 2.6% on gas/electricity second-round effects, 2026 down 10bp to 2.9% on food weakness.
Evidence IDs · P003
Investor implication
Own European equity diversifiers but monitor TTF and OAT-Bund spread as the risk trigger; ECB on 10 September is the near-term pivot.
Trade expression
Long European equity vs US mega-cap tech as a cash-flow/diversification pair; watch OAT-Bund spread for the macro hedge leg.
Invalidation
A sustained gas spike that forces genuinely hawkish ECB action and crushes bond-proxy sectors.
Transition
Europe's gas problem is itself a commodity story — and commodities extend well beyond energy.
Evidence IDs · P024 · P007 · P026 · P003
Gold: corrective after the 2% recommitment, but the medium-term bid is intact
Verdict
Near-term correction risk after the Jackson Hole-driven sell-off, but the long-end/rotation narrative supports renewed upside once the front end stabilizes.
Analysis
GS's metals desk explains the mechanics: the rally from 4050 to 4550-4600 was independent of the short end, but the sell-off happened because the 2-year moved up after Jackson Hole's recommitment to the 2% inflation target. The desk sees support at 4250-4300 (4250 near the 50d MA), notes two-way flows with clients buying calls into weakness, and argues topside convexity is mispriced at ~22 vol — 'if you were of the opinion that there is a chance that the Fed does not hike, you should be buying 5500 gold calls'. The structural case sits behind this: GS GOAL notes US Treasury interventions in long-dated Treasuries weighed on the Dollar and boosted Gold, and UBS's 4Q26 outlook emphasizes that strong central bank activity has been a key support for the metal, with investor demand the swing variable. The disagreement in the packet is between the corrective read (front-end-driven) and the structural bullish read (debasement/central-bank demand). Portfolio consequence: gold is the natural hedge to both a Fed pause and a fiscal/debasement scenario — express via cheap calls rather than delta, given the vol reset.
Tension
Corrective near-term (2y-driven sell-off) versus structurally bullish (central bank demand, debasement trade).
What changed
Jackson Hole recommitment to the 2% target drove the sell-off via the 2-year; spot rebounded ~1.4% as Treasury yields retreated.
Desk views
GS metals trading: sell-off was 2y-driven after the 2% recommitment; support 4250-4300; topside convexity mispriced at ~22v — buy 5500 calls if you think the Fed might not hike.
Evidence IDs · P022
UBS 4Q26 gold outlook: strong central bank activity has been a key support; investor demand is the key swing factor for the yellow metal.
Evidence IDs · P033
Investor implication
Use the correction to own cheap upside convexity into CPI/FOMC; gold doubles as the debasement-trade hedge ING flags.
Trade expression
Buy 5500 gold calls (mispriced convexity at ~22v) rather than outright delta into the correction.
Invalidation
A hawkish CPI that pushes the 2-year decisively higher and forces further leveraged unwinds below 4250 support.
Transition
Away from the headline metals complex, one agricultural market is quietly tightening under a weather risk few are watching.
Evidence IDs · P022 · P033 · P021 · P029
High-conviction calls
Higher global yields remain the base case, but Treasuries now offer enough carry that total returns stay positive until 10-year yields reach roughly 5.5% — shorting duration fights the coupon.
Evidence IDs · P006
US midterms will amplify dispersion more than volatility: with only a 0.82% post-election move priced and low correlations, avoid index-level election protection and hedge cheaply and selectively.
Evidence IDs · P001
BoJ pricing is stretched: a September hike is likely, but MUFG sees a move larger than 25bps as very unlikely despite ~100bps priced over 12 months.
Evidence IDs · P030
August CPI on 11 September, not Friday's NFP, is the consequential report for the September FOMC.
Evidence IDs · P002
CEMEA rates: markets have overestimated inflation/rate risks — nearly half of covered CEMEA central banks have cut since the war began; stay relatively dovish, especially CEE-3 and South Africa, and bullish Frontier FX carry.
Evidence IDs · P019
Must reads
The clearest articulation of the bond sell-off's drivers — QT, the $628bn Q4 absorption wall, and France above 5% — with the de-dollarisation test condition.
Evidence IDs · P032
The carry math that changes how the yield sell-off should be traded: 5.5% (1y) and 6.4% (2y) breakevens for negative Treasury total returns.
Evidence IDs · P006
The anchor asset-allocation stance: OW equities 12m, tactically neutral 3m, with bonds as the equity speed limit.
Evidence IDs · P021
The definitive read on stretched BoJ pricing and the US-Japan quid pro quo behind yen strength.
Evidence IDs · P030
The RBA's worsening trade-off: a firmer July CPI against a deepening housing downturn, with a November hike to 4.60% as base case.
Evidence IDs · P018
Synthesis
One thread runs through the packet: the end of the financial-repression regime. Yields normalizing toward post-GFC highs (P006, P032) forced US Treasury interventions that boosted gold and weighed on the dollar (P021), pushed Japan into record intervention and a policy-driven yen revaluation (P031, P034), and turned bonds into a speed limit for equities (P021) — while quantity-based financial conditions remain supportive enough (P027) that the pro-risk stance survives, expressed through dispersion and cheap hedges rather than outright index risk (P001).
Evidence IDs · P006 · P032 · P021 · P031 · P034 · P027 · P001
The near-term catalyst stack is dense: NFP Friday, CPI 11 September, ECB projections 10 September, BoJ 18 September, and the potential Trump-Xi summit on 24 September for the US-China truce — each capable of repricing the September-hike debate that sits behind the dollar, gold and rates complex.
Evidence IDs · P002 · P003 · P030 · P004
Under the radar
GS flags that ~70% of global sugar exports come from Brazil, India and Thailand — regions particularly exposed to El Niño drought/flood risk through the June 2026-May 2027 event. Raw sugar has already risen 18% since 6 August on Brazilian sugar content below the 10th percentile, but the near-term downside risk is a recovery in sugar content confirmed by delayed UNICA reports, which could trigger a speculative-positioning pullback. The larger medium-term risk is precautionary export restrictions. Note CANE sits among the top RS leaders (+21.3% 1M), confirming the move is already being expressed.
Evidence IDs · P016
GS argues China's policymakers may see gradual RMB appreciation against the USD as serving multiple goals simultaneously: reducing tariff pushback, supporting RMB internationalization by offsetting the negative rate differential, and gaining global manufacturing share. With CNY (and KRW, +15% in eight weeks on a once-in-a-generation terms-of-trade shock) already on an appreciating trend, JPY is now Asia's biggest FX outlier — and MUFG notes potential G20 coordination over currency policy toward China as a forward watch item.
Evidence IDs · P017 · P008 · P031
Closing
The regime is shifting, not breaking: yields normalize with carry cushioning the pain, interventions (US Treasury in Treasuries, MOF/US Treasury in USD/JPY) are now routine policy tools, and equities stay owned but expressed through dispersion and cheap convexity. CPI on 11 September and the BoJ on 18 September decide whether the September-hike consensus survives the week.
Evidence IDs · P006 · P021 · P034 · P002 · P030
Watch next
- ISM services today and NFP Friday: CACIB projects NFP +65k with unemployment back to 4.2%; a major downside surprise could trim September-hike odds, but CPI on 11 September is the decisive report.
- ECB projections on 10 September: watch for the expected 2027 HICP upgrade (+30bp to 2.6%) and whether the ECB's message is less hawkish than the three-hikes-by-April-2027 curve pricing.
- BoJ meeting 18 September: 30bps priced for this month against MUFG's <25bp call — a >25bp hike would invalidate the stretched-pricing fade.
- Potential Trump-Xi summit on 24 September: critical for extension of the US-China Busan Truce, including continued suspension of China's critical mineral export controls.
- Q4 US Treasury supply: $628bn net to be absorbed; any sign of indigestion or competition from JGBs/hyperscaler credit is the bond sell-off's escalation trigger.
- RBA: September hike finely balanced (45% probability per GS) with swap markets pricing just under 70%; a November hike to 4.60% is the base case against a deepening 9% peak-to-trough housing decline.