Research Digest — last 24h

Verdict

The rate repricing is strengthening, not fading, and is transmitting into equities, FX and credit expectations; whether it is near completion is genuinely contested.

Analysis

Natixis states the higher-for-longer scenario 'is not fading but strengthening,' with Treasury yields up 10bp in the 30-year to 17bp in the 10-year, intraday 10yr and 30yr at 19-year highs of 5.13% and 5.41%, and the 10yr TIPS yield at 2.77%, a level not seen since 2008 (P031). It attributes the move to three factors: Brent back above $100 at $103, a US flash composite PMI at its highest since July 2021, and a weak 5-year Treasury auction that confirmed duration buyers remain on the sidelines (P031). DB corroborates the scale, calling it the biggest 10yr jump since April 2025 and noting the 5yr rose above 5% for the first time since 2007, with futures pricing a 71% chance of an October Fed hike (P016). GS supplies the fundamental input: September flash services and manufacturing PMIs rose against consensus for a slight decline to their highest since March and May 2022, and 'both PMIs also pointed to intensifying price pressures' (P022). The causal chain is data strength plus oil feeding hike expectations, which lifts real yields, which compresses equity multiples. The counterargument comes from two directions. Citi argues rising real yields have compressed valuations as 'a rational de-rating rather than fear' and that credit markets 'continue to send a positive signal despite higher rates,' with banks more willing to lend and hyperscaler leverage low (P014). JPM goes further, saying it would 'buy this dip,' that yields are moving toward fair value near 5.5%, and that similar episodes have historically been buying opportunities (P026). MUFG sits between, calling the move one with 'the hallmarks of a pain trade and forced selling' that 'could have further to run' (P029). The investor consequence is that duration-sensitive equity leadership and rate-differential FX are the transmission points, and the next observable test is whether the October Fed hike gets fully priced and whether the 5-year auction channel improves.

Tension

Strong data justifies further hikes (P016, P022, P031) versus the move looking stretched and forced (P023, P029) versus yields approaching fair value and equities a buy (P026).

What changed

Futures moved to price a 71% chance of an October Fed hike, and the 5yr yield crossed 5% for the first time since 2007 (P016).

Desk views

Natixis reads the session as a regime signal, with real yields at 2008 highs and a weak 5-year auction confirming duration buyers are absent.

Evidence IDs · P031

DB quantifies the move and the driver set: PMIs plus oil, with the 5yr above 5% for the first time since 2007 and a 71% implied October hike probability.

Evidence IDs · P016

GS supplies the activity and inflation input, noting PMIs beat a consensus for slight decline and that both surveys pointed to intensifying price pressures.

Evidence IDs · P022

Citi dissents on the equity read-through, treating the de-rating as rational and credit as still supportive.

Evidence IDs · P014

JPM is the outlier bull, framing yields as approaching fair value and the episode as a buying opportunity.

Evidence IDs · P026

Investor implication

If the October hike gets fully priced, the discount-rate channel keeps pressure on long-duration equity leadership; if the auction and oil channels ease, the JPM/Citi buy-the-dip case gains traction. This is conditional interpretation, not a forecast.

Trade expression

Not established in cited evidence.

Invalidation

A reversal in oil below $100 and a PMI rollover would undercut the inflation-impulse leg (P031, P022).

Verdict

The dollar advanced with the US yield selloff, but the cited evidence splits on whether this is durable carry or an overextended move vulnerable to correction.

Analysis

MUFG states plainly that 'higher US yields across the curve helps the dollar,' with the selloff in US fixed income leading global bond markets, and notes the 2-year yield jump this month is around 60bps, close to surpassing the March move (P029). DB records the Dollar Index at 101.13, +0.56% (P016). ING dissents on durability: 'the dollar rally has accelerated, and the move is starting to look stretched,' with DXY breaking above 101.0 on strong PMIs, higher oil and soft risk sentiment, and it expects a correction to the 100-100.5 area if an October hike is not fully priced (P023). The mechanism is the rate-differential and carry channel: higher US front-end yields raise the dollar's relative appeal, and JPM's FX desk reports the franchise was 'a buyer of USD across all sectors yesterday as Equities fell and global yields ripped,' with corporate USD demand the largest seen this month, especially against AUD (P025). GS adds a structural flow angle: net US outflows have accelerated in recent months, and while the hurdle to impact the broad dollar is higher given US market size, directional flows matter more when they shift from historical averages, as is now the case (P005). The counterargument is that the move is flow- and sentiment-driven rather than fundamentally anchored, which ING makes explicit (P023) and MUFG echoes by calling it a pain trade with forced selling (P029). The investor consequence is that dollar longs carry both carry support and correction risk, and the next test is upcoming US data and whether the October hike gets fully priced (P023).

Tension

Rate-differential and flow support (P029, P025, P005) versus a stretched move vulnerable to correction (P023).

What changed

DXY broke above 101.0 on strong PMIs, higher oil and soft risk sentiment (P023).

Desk views

MUFG ties dollar strength directly to the US fixed-income selloff leading global bond markets, while flagging pain-trade dynamics.

Evidence IDs · P029

ING warns the rally is stretched relative to fundamentals and sketches a correction path if the October hike is not fully priced.

Evidence IDs · P023

JPM's FX desk describes the flow: broad USD buying as equities fell and yields ripped, with the largest corporate USD demand this month against AUD.

Evidence IDs · P025

GS frames US net outflows as a shift from historical averages, which raises their FX relevance despite the size of the US market.

Evidence IDs · P005

Investor implication

Dollar longs are supported by carry but exposed to a correction if the October hike is not fully priced; the AUD leg is where flow evidence is most specific (P025).

Trade expression

Not established in cited evidence.

Invalidation

A DXY correction toward 100-100.5 would validate the stretched-move view (P023).

Verdict

The cited evidence supports a de-rating from higher real yields while earnings hold up, which favors dispersion and quality over broad beta; the buy-the-dip case is contested.

Analysis

Citi's core claim is that 'higher rates have yet to dent corporate fundamentals, and earnings growth is broad across sectors,' with rising real yields compressing valuations as 'a rational de-rating rather than fear' (P014). BofA's volatility work supplies the market-structure consequence: dispersion 'continues to be the defining theme in U.S. equities,' with single-stock vol facing upside risks from boom-bust dynamics and stock correlations remaining low as markets re-assess the future, and it argues the recent semis-driven pullback in single-stock vol offers an opportunity to reload dispersion in a defensive vega-neutral manner (P009). The mechanism is that a higher discount rate hits long-duration, high-multiple equities hardest while idiosyncratic earnings drivers dominate index-level moves, so cross-sectional dispersion rises even as the index de-rates. JPM's morning briefing supports the earnings leg, noting strong cyclical momentum with September flash PMIs beating expectations and a sharp rise in employment, and its capex survey confirming strength across key end markets (P026). The counterargument is directional: JPM would buy the dip through a broadening trade, citing equity valuations as a cushion and geopolitical de-escalation as a possible off-ramp (P026), while Citi keeps a constructive and unchanged equity stance (P014). BofA's dispersion view is not a directional call and explicitly notes the entry level is elevated versus history, though it argues the dotcom analogue suggests further room (P009). The investor consequence is that index-level beta is a weaker expression than dispersion or quality tilts, and the next test is whether earnings breadth holds as real yields stay elevated (P014, P026).

Tension

De-rating with resilient earnings and rising dispersion (P014, P009, P026) versus the buy-the-dip broadening case (P026).

What changed

Dispersion has continued its multi-year trend higher in 2026, and the recent semis-driven pullback in single-stock vol reopened the entry (P009).

Desk views

Citi separates the rate-driven de-rating from fundamentals, keeping a constructive equity stance and favoring quality and short duration in fixed income.

Evidence IDs · P014

BofA argues dispersion is the defining theme and that the semis-driven single-stock vol pullback is an entry, while conceding the entry level is elevated versus history.

Evidence IDs · P009

JPM would buy the dip through a broadening trade, citing valuations as a cushion and geopolitical de-escalation as an off-ramp.

Evidence IDs · P026

Investor implication

If real yields stay elevated, dispersion and quality expressions are better supported than index beta; the buy-the-dip case depends on yields approaching fair value, which is a JPM opinion, not an established fact (P026).

Trade expression

Not established in cited evidence.

Invalidation

A broad earnings deterioration would break the de-rating-not-fear framing (P014).

Verdict

Trend-following positioning is eroding across equity indices and rotating toward USD, creating a mechanical amplification channel for the rate-driven selloff.

Analysis

BofA reports that CA equity positioning deteriorated as short-term trend signals weakened across every index tracked, with stop-outs concentrated among medium- and slower-moving trend followers carrying the largest long exposure, while faster models had already trimmed longs (P010). The specific triggers matter: the S&P 500 saw its first stop-outs of the episode with the next nearest sell trigger about 1.3% away, and the Euro Stoxx 50 closer still at roughly 0.5% (P010). In the Russell 2000, Nasdaq-100 and Nikkei, short-term signals have turned negative, suggesting the fastest CAs are moving short, even though Russell 2000 medium- and long-term trends remain bullish (P010). The mechanism is mechanical rather than fundamental: as trend signals break, systematic de-risking forces selling independent of valuation, which can amplify a rate-driven drawdown. BofA also notes CTAs are buying USD and remain short USTs after the Fed hike, with trend followers maintaining stretched shorts across the curve (P010). JPM's flows work adds a related positioning fact: momentum signals for bond futures moved back to more elevated shorts for 10y USTs and Bunds after the sell-off, and remain in more extreme short territory for OATs versus Bunds and extreme long territory for energy product futures (P024). The counterargument is that these are positioning and flow observations, not directional forecasts, and the same BofA note records that Russell 2000 medium- and long-term trends remain bullish, implying the de-risking is concentrated in faster cohorts (P010). The investor consequence is that the next 1-2% of index downside could trigger further mechanical selling, and the next test is whether the S&P 500 and Euro Stoxx 50 sell triggers are breached (P010).

Tension

Mechanical de-risking pressure (P010) versus still-bullish medium- and long-term trends in small caps (P010).

What changed

The S&P 500 saw its first stop-outs of this episode, and short-term signals turned negative in Russell 2000, Nasdaq-100 and Nikkei (P010).

Desk views

BofA details the cohort-level stop-out pattern and the proximity of index sell triggers, plus the CTA rotation into USD and short USTs.

Evidence IDs · P010

JPM documents the positioning backdrop in bond futures, with elevated shorts in 10y USTs and Bunds and extreme shorts in OATs versus Bunds.

Evidence IDs · P024

Investor implication

If the S&P 500 and Euro Stoxx 50 sell triggers are breached, mechanical selling could amplify the move; this is a flow-amplification channel, not a directional forecast (P010).

Trade expression

Not established in cited evidence.

Invalidation

A trend reversal that lifts short-term signals back positive would remove the mechanical selling pressure (P010).

Verdict

Clients bought US equities led by institutions and hedge funds, but rotated decisively down-cap and away from large caps, with buybacks slowing.

Analysis

BofA reports clients were net buyers of US equities after big de-risking-like outflows the prior week, buying both single stocks ($1.3bn) and equity ETFs ($1.7bn), led by institutional clients who have been net buyers in 3 of the past 4 weeks, with hedge funds also net buyers while retail clients were net sellers for the eighth straight week (P008). The rotation is the story: clients bought mid and small/micro cap equities versus selling large caps, with record mid-cap inflows, the fifth largest when normalized by Russell Midcap market cap, and BofA favors mid caps for 2H given less risk from rate hikes but a similar earnings recovery versus small caps (P008). The mechanism is that higher rates disproportionately threaten the most rate-sensitive, longest-duration large-cap growth cohort, pushing allocators toward mid caps with comparable earnings recovery but less rate-hike risk. The countervailing flow fact is that corporate client buybacks slowed for a second week, with annualized cumulative buybacks year-to-date tracking below 2024-25 levels though above 2010-23 annual levels, and rolling 52-week buybacks as a percentage of S&P 500 market cap at the lowest since August 2021 (P008). Sector flows were also lopsided: clients sold stocks in 7 of 11 sectors, led by Industrials, which saw its largest weekly outflows in the data, while Health Care saw record inflows (P008). The counterargument is that this is a single week of flow data and BofA's own framing ties the mid-cap preference to a rate-hike-risk assessment that could shift with the rate path (P008). The investor consequence is that the marginal buyer is rotating down-cap and away from large caps, which is consistent with the de-rating and dispersion themes, and the next test is whether the mid-cap inflow pace and the buyback slowdown persist (P008).

Tension

Record mid-cap inflows and institutional buying (P008) versus slowing buybacks and large-cap selling (P008).

What changed

Clients turned net buyers of US equities after big de-risking-like outflows the prior week, with record mid-cap inflows (P008).

Desk views

BofA documents the flow rotation: institutional and hedge fund buying, retail selling, record mid-cap inflows, large-cap selling, and slowing buybacks.

Evidence IDs · P008

Investor implication

The marginal buyer is rotating down-cap and away from large caps, consistent with the de-rating and dispersion themes; the buyback slowdown removes a large-cap support (P008).

Trade expression

Not established in cited evidence.

Invalidation

A reversal to large-cap inflows and a buyback reacceleration would undercut the rotation thesis (P008).

Verdict

The cited evidence supports a broadening AI-driven productivity and capex story with competitive pressure intensifying, while JPM flags that AI capex is crowding out other capex.

Analysis

GS reports Anthropic launched Claude Opus 5.5, scoring 58 versus 53 for GPT 6 Astra on an independent aggregate, particularly strong in coding and agentic knowledge work, using less compute and more than 30% faster than Opus 5, at $4 input / $20 output per 1M tokens versus Astra at $10/$50, so only 40% of Astra's token price (P007). GS calls this significant, saying Anthropic appears to have overtaken Astra on the leading independent aggregate with a significantly lower price point, strengthening competitive pressure on OpenAI (P007). GS's other note frames the demand side: the launch and usage of Muse has been the main driver of equities in recent days, first with CPU stocks rallying on the view that agents would increase CPU usage, and it argues the proliferation should be enormous as large corporates underwrite safety, infrastructure and distribution, removing friction and rent-seeking from the economy as a productivity boost (P004). JPM's flows work supplies the corporate-finance mechanism: with cash flows rising faster than capex, corporate financing surpluses have been rising steeply in recent quarters, approaching historically high levels outside crisis periods, and one explanation is that AI-related capex has been crowding out other capex, containing overall capex growth at the aggregate level (P024). JPM adds that these surpluses reduce the need for corporate-sector funding at an aggregate level and allow corporates to sustain buybacks, particularly outside the tech sector (P024). The counterargument is embedded in the same evidence: if AI capex is crowding out other capex, the aggregate capex impulse is weaker than the headline AI narrative suggests, and the productivity gains GS describes are prospective, not realized (P024, P004). The investor consequence is that the AI trade is broadening from hardware into productivity and corporate-finance effects, and the next test is whether the competitive price pressure GS documents translates into broader adoption and whether the capex crowding-out persists (P007, P024).

Tension

Broadening AI productivity and capex (P004, P007) versus AI capex crowding out other capex and containing aggregate capex growth (P024).

What changed

Anthropic's Opus 5.5 overtook Astra on the leading independent aggregate at 40% of Astra's token price (P007).

Desk views

GS documents the competitive shift: Opus 5.5 leads on the independent aggregate at a much lower price point, strengthening pressure on OpenAI.

Evidence IDs · P007

GS frames the demand side, arguing agentic proliferation removes friction and rent-seeking as a productivity boost.

Evidence IDs · P004

JPM supplies the corporate-finance mechanism, noting AI capex is crowding out other capex and financing surpluses support buybacks outside tech.

Evidence IDs · P024

Investor implication

The AI trade is broadening from hardware into productivity and corporate-finance effects, but the crowding-out evidence implies the aggregate capex impulse may be weaker than the AI headline suggests (P024).

Trade expression

Not established in cited evidence.

Invalidation

Evidence that AI capex is additive rather than crowding out would undercut the aggregate-capex caveat (P024).

Verdict

The cited evidence supports a base case of a tactically stable but deliverable-light summit that extends top-level stability into H1 2027 while the underlying economic squeeze continues.

Analysis

DB's base case for the 24 September Trump-Xi summit is a meeting heavy on positive optics but light on deliverables, and it assesses that a successful tactical summit amid strategic competition is the realistic best-case outcome, extending top-level stability into H1 2027 and minimizing near-term risks of decoupling, a security crisis, or a disruptive trade war (P017). DB notes the 3 November US midterms could incentivize the Trump administration to seek deals as voter cost-of-living concerns are highest since April, and possible deliverables include progress on a Board of Trade for non-sensitive goods, agricultural purchases, targeted investment deals, and an extension of the 2025 Busan agreement, which postponed a range of mutual economic restrictions until November 2026 (P017). CACIB's EM weekly shares the cautious-optimism framing, expecting both sides to aim to avoid escalation, with the extension of the October 2025 Busan Truce, tariff exemptions on $30bn of goods, and AI guardrails as the key topics to watch (P001). The mechanism is that a truce extension removes a near-term tail risk of tariff escalation, supporting EM and China-exposed risk assets, while the underlying strategic competition keeps the medium-term squeeze in place. The counterargument is that DB explicitly says a breakthrough is unlikely and the underlying China-US economic squeeze is set to continue, so the summit is a risk-reduction event rather than a resolution (P017). CACIB adds a liquidity caveat: month-end and quarter-end flows plus thin liquidity in some EM economies on holidays including Korea (24-25 September), China (25 September and 1-7 October) and Taiwan (25-28 September) could distort price action (P001). The investor consequence is that the summit is a volatility-suppressing catalyst with limited fundamental change, and the next test is whether the Busan extension and tariff exemptions are actually secured (P017, P001).

Tension

Tactical stability and truce extension (P017, P001) versus an unlikely breakthrough and a continuing underlying economic squeeze (P017).

What changed

The summit is the near-term catalyst, with the Busan agreement extension as the key deliverable to watch (P017, P001).

Desk views

DB frames the summit as optics-heavy and deliverable-light, with a truce extension extending stability into H1 2027 while the economic squeeze continues.

Evidence IDs · P017

CACIB expects cautious optimism with both sides avoiding escalation, and flags the Busan Truce extension, tariff exemptions and AI guardrails as key topics.

Evidence IDs · P001

Investor implication

A truce extension would remove a near-term tariff-escalation tail risk for EM and China-exposed assets, but the medium-term squeeze remains; this is a risk-reduction event, not a resolution (P017).

Trade expression

Not established in cited evidence.

Invalidation

A failure to extend the truce or a re-escalation would undercut the stability base case (P017).

Verdict

Oil above $100 is reinforcing inflation concerns and hike expectations, but GS forecasts China's crude imports to stay low, which caps one upside channel while leaving Middle East escalation as the main risk.

Analysis

Natixis identifies Brent moving back above $100, up 4% at $103, as the first of three factors driving the bond rout, reigniting inflation concerns (P031). DB records Brent at $102.25, +3.38% (P016). MUFG notes Brent is up close to 6% from the lows and that the speech by President Pezeshkian of Iran at the UN did not suggest prospects for imminent peace were good (P029). Westpac adds the supply-side detail: oil rebounded sharply as Pezeshkian's comments tempered optimism around reopening the Strait, while US officials indicated the administration was working with refiners to voluntarily curb diesel exports (P033). The countervailing evidence is GS's oil analyst: a sharp decline in China's crude imports has helped stabilize the oil market over the last six months, and using two approaches GS forecasts China's crude imports to stay low in the coming months if prices remain elevated, so it continues to view a possible escalation of strikes on Middle East crude production and export infrastructure, not higher China imports, as the main upside risk to its crude price forecast (P021). GS also notes Persian Gulf oil exports remain about 6mb/d below pre-war levels despite a September rebound, and that high-frequency seaborne Chinese crude net imports picked up 6% (0.5mb/d) from August in September so far but remain nearly 3mb/d below seasonal norms (P021). The mechanism is that oil above $100 feeds headline inflation and hike expectations, which lifts real yields, but the China-import offset means the demand-side upside is capped, leaving supply-side escalation as the dominant risk. The investor consequence is that oil is currently a rates and inflation input more than a demand story, and the next test is whether Middle East escalation materializes or whether China imports stay low as GS forecasts (P021).

Tension

Oil above $100 reinforcing inflation and hike expectations (P031, P016, P029) versus GS's forecast that China's crude imports stay low, capping demand-side upside (P021).

What changed

Brent moved back above $100, up 4% at $103 (P031).

Desk views

Natixis names Brent above $100 as the first driver of the bond rout, reigniting inflation concerns.

Evidence IDs · P031

GS forecasts China's crude imports to stay low if prices remain elevated, and views Middle East escalation as the main upside risk.

Evidence IDs · P021

Westpac ties the oil rebound to Pezeshkian's comments on the Strait and US efforts to curb diesel exports.

Evidence IDs · P033

Investor implication

Oil is currently a rates and inflation input more than a demand story; the main upside risk is Middle East supply escalation, not China demand (P021).

Trade expression

Not established in cited evidence.

Invalidation

A sharp recovery in China's crude imports would validate the demand-side upside channel (P021).