The Curve Bites Back: Soft CPI, Higher Long Yields, and a Duration Miss

Scorecard: Last Week's Calls

Do not chase the equity record, AI and semis neutral (Medium): Correct. The S&P 500 set a fresh record close of 7,785.76 on 14 August, then slid to 7,745.06 on 17 August as a tech-led selloff met rising yields (FRED, 17 August). Sitting on our hands paid.
The implicit July CPI framing, consensus near 3.4% headline (gradeable consensus call): Correct. Headline CPI landed exactly at 3.4% y/y, up 0.1% m/m, core 2.5%, all in line (BLS, 12 August). Energy fell 1.5% on the month even as it ran up 14.7% y/y.
Long gold, target $4,500 over three months, add dips toward $4,200 (Medium): Correct and tracking. Spot held roughly $4,385 on 17 August, up modestly on the week and above our zone (Fortune, 17 August). It never revisited $4,200, so the add did not trigger.
Own the ECB September hike via front-end rates, neutral EUR, overweight European financials, fade EUR/USD above 1.16 (Medium): Tracking correct. The euro-area Q2 GDP second estimate confirmed 0.4% q/q and 1.0% y/y on 14 August, keeping the hike debate live (Eurostat, 14 August). EUR/USD touched a two-month high of 1.1614 before easing to about 1.158, so the fade above 1.16 was live (Regal, 17 August).
Small constructive GBP tilt above 1.34, add dips toward 1.33, cautious gilts (Low): Correct. UK Q2 GDP rose 0.4% with a 0.3% June monthly beat (ONS, 13 August); cable held about 1.355 (Yahoo Finance, 17 August).
Neutral-to-small-long AUD, buy dips toward 0.685 (Medium): Correct and tracking. The RBA held at 4.35% and its 11 August Statement kept upside inflation risks front and centre; AUD/USD held about 0.707, never testing our 0.685 dip-buy (RBA, 11 August; Yahoo Finance, 17 August).
Overweight IG, underweight HY, scale into HY only past 3.25-3.5% OAS (Medium): Correct and untriggered. HY OAS sat at 2.71% on 13 August, still far from our scale-in (FRED/ICE BofA, 13 August).
Broad dollar neutral-to-short (implicit, from the prior book): Correct and tracking. DXY eased to about 99.5, near a two-month low, on softer US data (Trading Economics, 17 August).
Structurally long large-cap Indian IT exporters plus financials, trim if USD/INR breaks 96.5 (Medium): Mixed, leaning negative this week. The Nifty 50 fell for a fifth straight session to 24,287.65 on 17 August, with IT the biggest drag as TCS and Infosys dropped (Reuters, 17 August). The rupee weakened only to about 95.4, so the 96.5 trim stayed untriggered.
Long silver (Medium): Roughly flat. Silver held about $65.5 on 17 August, then gave back ground to about $63 on 18 August in a broad metals pullback (Forbes, 17 August; Kitco, 18 August).
Add duration on backup toward 4.70%, target 4.40% on the 10-year, hold through CPI (High): Wrong this week, and we own it. The backup to 4.70% arrived and then some: the 10-year rose to 4.72% and touched a 19-month high of 4.75%, while the 30-year topped 5.31%, its highest in 19 years, despite the soft CPI that was supposed to help (CNBC, 17 August). Our entry triggered and went straight offside; the 4.40% target is not just unmet, it is moving away.
Sellers of Brent strength, target high $70s within a month, contrarian (Medium): Wrong this week. Brent settled at $90.87 on 17 August as the Hormuz deal collapsed (CNBC, 17 August). Our own rule said pause the fade at $90; the fade got run over first.
Underweight energy equity (implicit): Wrong this week. Oil rallied and energy outperformed a soft tape, with Chevron up 1.41% on the 17 August down day (Regal, 17 August).
Short USD/JPY toward 150, contrarian, entered offside (Medium): Wrong and getting worse. USD/JPY climbed to about 159.1 by 17 August as the late-July intervention faded and no follow-up landed (Vantage, 17 August). The short is further offside than last week.

Hit rate: eight clean or tracking wins, four clear misses, two mixed. The wins were the disinflation read, gold, European and UK rates, credit discipline, the soft dollar, and above all not chasing the equity record. The misses all rhyme: we underestimated how much the long end and oil would trade on supply and risk premium rather than the Fed path.

Key Levels Dashboard

Brent crude, ~$91/bbl, up ~8% on the week. Settled $90.87 on 17 August as Iran ruled out an interim-deal extension (CNBC, 17 August).
WTI crude, ~$84.50/bbl, up ~7% on the week. Closed $84.50 on 17 August on Hormuz supply fears (CNBC, 17 August).
Gold spot, ~$4,385/oz, up ~1% on the week. Held its ground through higher yields, then dipped toward $4,333 on 18 August (Fortune, 17 August; Kitco, 18 August).
Silver spot, ~$65/oz, roughly flat on the week. Gave back gains to about $63 on 18 August in a broad metals pullback (Forbes, 17 August).
Federal Reserve funds target, 3.50-3.75%, unchanged. Held 9-3 on 29 July; next decision 16-17 September (Federal Reserve).
ECB deposit rate, 2.25%, unchanged. Next meeting 10 September with fresh projections (ECB).
BoE Bank Rate, 3.75%, unchanged. Next decision 17 September (Bank of England).
RBA cash rate, 4.35%, unchanged. Held 11 August with a hawkish Statement (RBA, 11 August).
RBI repo rate, 5.25%, unchanged. Held unanimously 5 August, neutral stance (RBI).
Latest US jobs print, July -23k, unemployment 4.1%, unchanged. No new payrolls this week; next report 5 September (BLS, 7 August).
S&P 500, ~7,745, down ~0.2% on the week. Record close 7,785.76 on 14 August, then slipped on rising yields (FRED, 17 August).
US 10-year Treasury, ~4.72%, up ~6bp on the week. Touched a 19-month high of 4.75% mid-week; 30-year hit 5.31% (CNBC, 17 August).
DXY, ~99.5, down ~0.2 on the week. Near a two-month low on softer US data (Trading Economics, 17 August).
EUR/USD, ~1.158, up ~0.3% on the week. Hit a two-month high 1.1614 before easing (Regal, 17 August).
GBP/USD, ~1.355, up ~0.4% on the week. Held above 1.34 after the UK GDP beat (Yahoo Finance, 17 August).
AUD/USD, ~0.707, up ~0.1% on the week. Steady after the RBA's hawkish hold (Yahoo Finance, 17 August).
USD/INR, ~95.4, rupee down ~0.2% on the week. Pressured by oil and a soft Nifty (Investing.com, 17 August).
USD/JPY, ~159.1, yen weaker by ~1.3 on the week. Intervention gains faded further (Vantage, 17 August).
Nifty 50, 24,287.65, down ~1.2% on the week. Fifth straight decline on Mideast and IT weakness (Reuters, 17 August).
US HY OAS, ~2.71%, tighter ~1bp on the week. Still deep inside our scale-in trigger (FRED/ICE BofA, 13 August).

Executive Summary

The week delivered exactly the CPI we expected and exactly the bond market we did not. July headline inflation came in at 3.4% y/y, 0.1% m/m, core 2.5%, all in line, and cooler than feared once energy pass-through faded (BLS, 12 August). Producer prices and retail sales that followed were soft, University of Michigan sentiment fell to 51.0 in early August from 55.2 in July, below the 54.5 Reuters consensus (University of Michigan Surveys of Consumers, 15 August), and by 17 August a 90% majority of economists, 94 of 104 in the August 12-17 Reuters poll, expected the Fed to leave rates unchanged at its September meeting, with CME FedWatch pricing a hold near 70% (Reuters, 17 August). That is our disinflation-and-hold thesis, confirmed.

Yet the 10-year rose to 4.72% and touched a 19-month high of 4.75%, and the 30-year topped 5.31%, a 19-year high (CNBC, 17 August). The front end fell as the hold got priced; the long end sold off on term premium, on record corporate bond supply, on fiscal worry, and on a suspicion that a Warsh-led Fed is content to let inflation run above target. Investment-grade companies have sold nearly $1.5 trillion of bonds year to date, a 36% jump on last year, with AI-related debt about 15% of the total, per Barclays; as BMO's Ian Lyngen put it, "a record pace of corporate bond issuance has added substantial duration supply, with consequences for the outright level of yields as well as the shape of the yield curve and term premium" (Bloomberg, 17 August). That bear-steepening is what hurt our duration add, which is the honest headline of the week. Add a Middle East that went the wrong way, the June memorandum expired, Iran ruled out extending it and threatened to escalate, and Brent settled at $90.87, and two of our positions took the punishment together (CNBC, 17 August).

The forward path is now about who wins the tug of war in the long end. If the hold sticks and term premium stabilises, the front end anchors and the curve does the work; if oil stays bid and supply keeps pressing, the long end can grind higher regardless of the Fed. We position for the first with front-end duration and a steepener, and we stop trying to catch the long bond.

Signal-Filter Takeaways for Clients

Being right on the Fed was not enough. The trade that mattered this week was the shape of the curve, not the level of the funds rate, and we were positioned for the level. Own the front end and the steepener, not the outright long bond.
Oil is no longer a clean fade. With the Hormuz deal dead and transits near a standstill, the risk premium is real. We respect the $90 line we drew ourselves and re-engage as sellers only on evidence flows are normalising or above roughly $93.
Gold is earning its keep as the hedge that works when both stocks and bonds wobble. It held through a hawkish long end. Keep it core.

01 Middle East: The Deal Dies, the Premium Returns

What happened. The interim framework did the opposite of what markets hoped since our last edition. The June memorandum of understanding expired on Monday 10-11 August, Iran ruled out extending it, and by 17 August a senior Iranian official told Reuters that Tehran was prepared to escalate in and around Hormuz (CNBC, 17 August). Iran and Oman have agreed the coordinates of new shipping routes, but Tehran insists a bilateral deal does not reopen the strait, and demands the US end the war, lift the blockade and pay compensation first (Al Jazeera, 8 August). The UAE reported Iranian attacks on two ADNOC-linked tankers on 14 August; the US threatened the "economic isolation" of Iran and said its blockade could run "indefinitely" (CNBC, 14 August). Transits stayed depressed: roughly 8 to 14 vessels a day against about 130 before the war (CNBC, 11 August). Energy Secretary Chris Wright said on 11 August that "the seven-day average for oil leaving the Strait of Hormuz is currently up to almost 9 million barrels per day, total oil flows are currently averaging approximately 15 million barrels per day" (X, 11 August), though the figure is contested: Kpler put crude Hormuz exports at just 1.74 million bpd for the week of 3 August, and Kpler's Matt Smith told CNN "it is not possible to reconcile the disparity between what we see and what he is quoting."

Why it matters. The clean "fade every spike" setup we ran for weeks is gone for now. The June-to-August pattern was buy-the-dip-in-oil on escalation, sell-the-rip on deal hopes; the deal is now off the table and the premium is stickier. That said, the striking fact is how much crude still moves despite the standoff, which is why Brent sits near $90 and not near $110.

Scenarios

Base case, 55%: no deal, sporadic attacks, transits stay low, Brent mid-to-high $80s to low $90s.
Downside, 30%: a strike lands or the blockade tightens, Brent back toward $100.
Upside, 15%: a genuine de-escalation or a workable Iran-Oman arrangement, Brent slides to the low $80s.

Market implications

Asset class: Long-dated crude call optionality still earns its premium; outright oil longs are chasing.
Currency and flows: Bad for the rupee and the yen as energy importers, marginally supportive of the Canadian and Norwegian complexes.
Sector rotation: Energy equity is a hedge, not a conviction long, and it worked this week.
Entry and exit: Re-enter the Brent fade above roughly $93 or on confirmed normalisation; stand aside between $85 and $92.

What to watch. Whether transits recover off the 8-to-14 vessel range, any confirmed strike, and whether Iran-Oman produces a signed text.

House view. We hold a small long-dated call tail against a strike and we do not chase spot; the fade is paused, not abandoned. Medium conviction.

02 Energy Markets: OPEC+ Adds Into a Closed Strait

What happened. The paradox sharpened. Seven core OPEC+ members, Saudi Arabia, Russia, Iraq, Kuwait, Kazakhstan, Algeria and Oman, approved a 188,000 bpd September increase at a virtual meeting on 2 August, completing the phased rollback of the 1.65 million bpd 2023 voluntary cut, with the next meeting set for 6 September; a Q4 pause is expected but was not formally confirmed in the communique (Reuters, 2 August). The EIA cut its Q3 Brent average to about $85 and assumed Hormuz constraints persist through August, with disruptions near 0.6 million bpd lingering into end-2027 (EIA, 11 August). Brent still rose to $90.87 and WTI to $84.50 on 17 August because the deal collapsed, not because the barrels vanished (CNBC, 17 August).

Why it matters. This is a market with more supply coming and a chokepoint that will not clear. That combination caps the upside and floors the downside, which is why we think the mid-$80s is the gravity point once the acute risk premium bleeds off. Our fade got run over at $90, so we respect our own stop and re-engage higher.

Scenarios

Base case, 55%: Brent oscillates $85 to $92 as supply and risk premium offset.
Downside for prices, 25%: covert flows normalise, OPEC+ barrels bite, Brent to low $80s.
Upside for prices, 20%: escalation, Brent through $95 toward $100.

Market implications

Asset class: Prefer optionality to linear exposure; sell strength rather than buy breakouts.
Currency and flows: Energy-importer currencies stay pressured while crude holds $90.
Sector rotation: Keep energy equity as a hedge only.
Entry and exit: Sell Brent strength above ~$93, stand aside below, cover if a strike lands.

What to watch. The 6 September OPEC+ meeting for any change to the Q4 pause, and Kpler transit counts.

House view. Contrarian and disciplined: we fade the next spike back toward the mid-$80s, small size, given OPEC+ supply and still-substantial Gulf flow. Low conviction after this week's stop-out.

03 United States: Right on Inflation, Run Over on the Long End

What happened. July CPI printed 3.4% headline, 0.1% m/m, core 2.5%, in line and benign, with energy down 1.5% on the month (BLS, 12 August). Producer prices and retail sales softened, Michigan sentiment fell to 51.0 on 15 August, and hike odds collapsed: CME FedWatch cut the September hike to 42% on CPI day and toward roughly 30% after soft PPI, with a hold near 70% by 17 August; the August 12-17 Reuters poll had 94 of 104 economists expecting a hold and 80 seeing no change through year-end (Reuters, 17 August). And yet the 10-year rose to 4.72%, touched 4.75%, and the 30-year hit a 19-year high of 5.31%, driven by term premium, record corporate bond supply, fiscal worry, and fear of Fed complacency on inflation (CNBC, 17 August). Fed Chair Warsh made no public remarks in the window and delivers his first Jackson Hole keynote on 28 August.

Why it matters. This is the flip. Last week we raised the duration add to High conviction and said we would hold through CPI; the CPI was fine and the trade still lost because the market repriced the term premium, not the Fed path. We were solving for the wrong variable. The front end behaved exactly as our thesis predicted, so the fix is to move there and to own the steepener rather than fight the long bond.

Scenarios

Base case, 55%: Fed holds in September, front end anchors near 4.15% on the 2-year, long end stays heavy, curve steepens.
Downside, 25%: oil and supply push the 10-year above 4.85%, equities wobble further.
Upside, 20%: term premium stabilises and the 10-year drifts back toward 4.55%.

Market implications

Asset class: Front-end duration over long duration; 2s10s or 5s30s steepener.
Currency and flows: A heavy long end plus a soft front end is dollar-neutral, which fits our soft-dollar bias.
Sector rotation: Long-duration equities are exposed to the long-end move; that is why the S&P slipped from its record.
Entry and exit: Add front-end duration now; add long-end only above 4.85%; trim the steepener if 5s30s runs past prior wides.

What to watch. Warsh at Jackson Hole on 28 August, the 19 August FOMC minutes, and the pace of AI-related bond issuance.

House view. Flip flagged: we cut the outright 10-year add, own the front end, and run a curve steepener. Medium conviction, down from High, because the driver is supply and term premium, not the policy path we called correctly.

04 Europe: Growth Confirmed, the Hike Debate Lives

What happened. The euro-area Q2 GDP second estimate confirmed 0.4% q/q and 1.0% y/y on 14 August, its strongest quarterly pace since Q1 2025, helped by AI-related investment and one-off factors (Eurostat, 14 August). Final July HICP is due 19 August, the day after this edition, with the flash at 2.9%. The ECB meets 10 September with fresh projections; Lagarde has flagged upside inflation risks and some governors questioning whether a hike is warranted.

Why it matters. Growth that holds above expectations keeps the September hike genuinely on the table and validates owning it via front-end rates. Nothing this week changed the thesis; the GDP confirmation strengthened it.

Market implications

Asset class: Own the ECB hike via front-end euro rates.
Currency and flows: Neutral EUR; the pair testing 1.16 is a fade level, not a breakout.
Sector rotation: Overweight European financials into a steeper curve and a live hike.
Entry and exit: Fade EUR/USD above 1.16, do not press shorts below 1.14.

What to watch. Final July HICP on 19 August and ECB commentary into 10 September.

House view. Long the ECB hike via front-end rates, neutral EUR, overweight European financials. Medium conviction, unchanged.

05 United Kingdom: A June Rebound Keeps the Pound Bid

What happened. UK Q2 GDP rose 0.4% q/q, matching consensus and slowing from Q1's 0.6%, but June monthly GDP jumped 0.3% against expectations of zero, helped by services, a heatwave and World Cup spending (ONS, 13 August). Cable held about 1.355. The BoE, which cut to 3.75% on a hawkish 6-3 vote on 30 July, next decides on 17 September.

Why it matters. The June beat takes the edge off recession worry and supports our small constructive sterling tilt. It does not change the BoE's cautious path, but it does argue against pressing shorts.

Market implications

Asset class: Cautious on gilts given the global long-end selloff.
Currency and flows: Constructive GBP above 1.34.
Sector rotation: No fresh call.
Entry and exit: Add cable dips toward 1.33, fade toward 1.36.

What to watch. The 20 August Blue Book revisions and the 17 September BoE decision.

House view. Small constructive GBP tilt above 1.34, add dips toward 1.33, cautious gilts. Low conviction, unchanged.

06 Australia: A Hawkish Hold, Confirmed by the Statement

What happened. The RBA held at 4.35% on 11 August and its quarterly Statement kept the tightening bias explicit: inflation "still too high," risks "skewed to the upside," and Bullock talking down the housing downturn as a reason to ease (RBA, 11 August; ABC, 11 August). Market pricing put roughly a 50% chance of a hike by November. National dwelling values fell 0.7% in July. AUD/USD held about 0.707.

Why it matters. A central bank still leaning hawkish with a firm labour market underpins the currency and validates buying dips rather than chasing. The Statement was the event, and it delivered.

Market implications

Asset class: No change; rates biased to stay restrictive.
Currency and flows: Neutral-to-small-long AUD.
Sector rotation: No fresh call.
Entry and exit: Buy AUD dips toward 0.685, trim toward 0.72.

What to watch. The RBA's next decision at end-September and any shift in the November hike pricing.

House view. Neutral-to-small-long AUD, buy dips toward 0.685. Medium conviction, unchanged.

07 India: The Index Keeps Slipping, IT Leads It Down

What happened. The Nifty 50 fell for a fifth straight session to 24,287.65 on 17 August, down about 1.2% from our baseline, with IT the biggest drag as TCS, Infosys and HCL all fell more than 1% (Reuters, 17 August). Oil near $90 and stalled US-Iran talks kept sentiment cautious; the rupee weakened to about 95.4, still short of our 96.5 trim trigger (Investing.com, 17 August). The RBI, which held at 5.25% on 5 August, next meets in late September or early October.

Why it matters. This is the position that is testing our patience. The structural case for large-cap Indian IT and financials is intact, but the tape is weak and oil is a live headwind for an importer. We stagger entries rather than add aggressively, and the 96.5 rupee line remains the discipline.

Scenarios

Base case, 60%: the Nifty bases in the 24,200 support zone as earnings season ends.
Downside, 25%: oil above $95 pushes the rupee past 96.5 and forces a trim.
Upside, 15%: oil rolls over, IT rebounds, the index reclaims 24,700.

Market implications

Asset class: Structurally long large-cap IT exporters and financials.
Currency and flows: Watch the rupee against 96.5; oil is the swing factor.
Sector rotation: Add financials on weakness; IT is a hold, not an add, until the tape steadies.
Entry and exit: Stagger entries into weakness, trim only if USD/INR breaks 96.5.

What to watch. USD/INR versus 96.5, Brent, and the RBI's late-September meeting.

House view. Structurally long large-cap Indian IT exporters plus financials, trim only if USD/INR breaks 96.5. Medium conviction, unchanged, but on a shorter leash after five down sessions.

08 Precious Metals and Safe-Haven Assets

What happened. Gold held roughly $4,385 on 17 August, up modestly on the week, then eased toward $4,333 on 18 August in a broad metals pullback as yields and oil pressured rate-sensitive assets (Fortune, 17 August; Kitco, 18 August). It never revisited our $4,200 add zone. Silver held about $65.5 before slipping toward $63 on 18 August. The World Gold Council's 17 August note reframed last week's calmer data as an inflection, and official demand remains a support: the People's Bank of China added about 20 tonnes in July, its 21st consecutive monthly purchase, and central banks bought a record-for-any-second-quarter 289 tonnes in Q2, up 62% year on year (World Gold Council, 30 July; Bloomberg, 7 August).

Why it matters. Gold did what bonds could not: it hedged. It held its ground through a hawkish long end and an equity wobble, which is precisely the diversification we own it for. The 18 August dip is noise, not a thesis break.

Market implications

Asset class: Long gold, long silver.
Currency and flows: A soft dollar and central-bank demand underpin bullion.
Sector rotation: No fresh call.
Entry and exit: Add gold toward $4,200, silver on pullbacks; trim gold above $4,500.

What to watch. Real yields, the dollar, and Warsh at Jackson Hole.

House view. Long gold, target $4,500 over three months, add toward $4,200; long silver. Medium conviction, unchanged.

09 Sovereign and Credit Conditions

What happened. HY OAS sat at 2.71% on 13 August, essentially unchanged and still historically tight, while the Treasury long end sold off hard, the 30-year at a 19-year high of 5.31% (FRED/ICE BofA, 13 August; CNBC, 17 August). The divergence is the story: credit is priced for a soft landing even as the risk-free long end reprices term premium.

Why it matters. Tight spreads plus a rising risk-free rate is an uncomfortable combination, because the all-in yield on HY is rising for the wrong reason. We stay up in quality. The scale-in trigger is nowhere close.

Market implications

Asset class: Overweight IG, underweight HY.
Currency and flows: No fresh call.
Sector rotation: Favour higher-quality, shorter-spread-duration credit.
Entry and exit: Begin HY scale-in only past 3.25-3.5% OAS.

What to watch. Whether spreads finally widen as the long end grinds higher.

House view. Overweight IG, underweight HY, no scale-in yet. Medium conviction, unchanged.

10 Other Flashpoints

Japan and the yen: USD/JPY rose to about 159.1 by 17 August, further from our 150 short target, as the late-July coordinated intervention faded with no follow-up; the BoJ's July Summary of Opinions flagged faster hikes, and a September or October move is live (Vantage, 17 August; CNBC, 12 August). We stay short, sized small, and we own that it is our worst-performing call. Medium conviction, offside.
AI concentration and semis: Nvidia reports Q2 FY27 after the close on 26 August, the next big test; the S&P's slip from its 14 August record was led by tech as yields rose (Reuters). Neutral stance intact.
US trade policy: The 10-12.5% forced-labour tariffs from 24 July remain a slow inflation drip; no new material development this week.
Ukraine and Russia, G7 cohesion: No material new development; G7 secondary-sanctions pressure stays background.
US midterms: 3 November, a rising catalyst; no new material development this week.
Cyber: No genuinely new flashpoint this week worth adding.

Upcoming Events and Catalyst Calendar

19 August: Euro-area final July HICP. Confirms the 2.9% flash and the energy pass-through the ECB is watching into 10 September.
19 August: July FOMC minutes. Colour on the 9-3 hold and how close a September hike really is.
20 August: UK ONS Blue Book revisions. Could reshape the recent GDP path.
26 August: Nvidia Q2 FY27 earnings. The single biggest test of the AI-capex trade and megacap breadth.
27-29 August: Jackson Hole symposium, Warsh keynote on 28 August. His first as Chair; the framework signal 19 days before the FOMC.
5 September: US August payrolls. The next read on a labour market that printed -23k in July.
6 September: OPEC+ meeting. Watch for any change to the flagged Q4 pause.
10 September: ECB decision and projections. Our owned September-hike call is graded here.
11 September: US August CPI. The inflation read the Fed sees before it decides.
16-17 September: FOMC. Hold near 70% priced; the long-end reaction is the risk.
17 September: BoE decision. Our sterling and gilt views are tested.
Mid-to-late September: BoJ meeting. A hike would validate the short-yen thesis.
Late September or early October: RBA and RBI decisions. Australia's hawkish bias and India's neutral stance in focus.
29 October: ECB decision. Follow-through on the September projection path.
3 November: US midterms. A rising macro and fiscal catalyst.

Recommendations (Staged)

Immediate, rates: Cut the outright 10-year duration add and shift to front-end duration (2-5 year) where the September hold is pricing; run a 2s10s or 5s30s steepener. Add long-end exposure only above a 4.85% 10-year. This is the flip from last week.
Immediate, gold: Stay long, keep the $4,500 three-month target, add toward $4,200. Treat the 18 August dip as an entry, not an exit.
Immediate, oil: Hold the small long-dated call tail against a strike; keep the fade paused between $85 and $92; re-enter as a seller above ~$93 or on confirmed transit normalisation.
Near-term, Europe: Hold the ECB-hike-via-front-end position and European financials into 10 September; fade EUR/USD above 1.16, do not press below 1.14.
Near-term, UK: Keep the small constructive GBP tilt above 1.34, add dips toward 1.33, stay cautious on gilts.
Near-term, Australia: Keep neutral-to-small-long AUD, buy dips toward 0.685, trim toward 0.72.
Structural, India: Stay long large-cap IT exporters and financials, stagger entries into weakness, trim only if USD/INR breaks 96.5.
Structural, credit: Overweight IG over HY; begin HY scale-in only past 3.25-3.5% OAS.
Structural, Japan: Hold the short USD/JPY toward 150 as a contrarian position, sized small while offside; cut it if a September BoJ hike fails to materialise and the pair breaks above 161.

Thresholds That Change the Calls

10-year above 4.85% on sticky oil or supply: add long-end duration to the front-end book.
10-year back below 4.55% with term premium easing: reduce the steepener, the long-end thesis is normalising.
A confirmed Hormuz strike or blockade tightening: re-arm the oil long, drop the fade.
A durable Hormuz reopening or confirmed flow normalisation: press energy underweight, cut the call tail.
Brent above ~$93: re-enter the sell-strength fade.
HY OAS past 3.25-3.5%: begin HY scale-in.
Gold above $4,500: trim; toward $4,200, add.
DXY reclaiming 102: revisit long-dollar; below 98, press the short.
USD/INR breaks 96.5: trim the India tilt.
USD/JPY above 161 with no BoJ move: cut the short.

Markers of Concern

Levels are a dated snapshot, mostly 17 August intraday or close, with some 18 August updates noted; markets move.
Scenarios are probabilities, not forecasts.
The long-end selloff is being driven by term premium and supply as much as by the Fed, so a correct Fed call does not guarantee a correct duration call, as this week proved.
A negative July payroll with triple-digit downward revisions remains a growth warning that a benign CPI does not erase.
Our duration and oil misses were correlated: both traded on supply and risk premium rather than the macro path, and that correlation could persist.