Fed Holds, Korea Falls: A Hawkish Stance + Concentration Risk

THE GLOBAL MACRO PICTURE

Oil reversed sharply at the start of August. Brent fell around 7% to below $84 after Trump announced fresh Iran talks and confirmed he had called off a planned military strike, then dropped a further 6% by August 5, breaking well below $80. That unwound a large part of July's 25% surge, which was itself driven by renewed US-Iran hostility and supply disruption stretching from the Strait of Hormuz to the Red Sea. The rest of the de-escalation trade moved in lockstep: silver climbed above $58/oz, U.S. and European equities rose the same morning, and European natural gas fell alongside crude. Hedges unwound as fast as they were built. After a July peak of $94.26 on the 23rd, the pattern is now legible: energy risk is being priced as a repeating, reversible event rather than a durable repricing.

The Fed held its policy rate at 3.50-3.75% for a fifth consecutive meeting, but the 9-3 vote, with three dissents favouring a hike, carries more information than the hold itself. The statement continues to frame inflation as above the 2% goal and attributes it to supply shocks in specific sectors, primarily energy, rather than broad demand. June's data fit that frame, with headline CPI cooling to 3.5% and core to 2.6% as energy costs fell during the brief lull in the conflict. The lull ended in July when Brent rose 25% again, and the July CPI print due August 12 is the first read on whether that reversal feeds through. On labour, the Committee's language points to stability rather than softening, with job creation keeping pace with the workforce and unemployment barely moved. Markets have drawn the obvious conclusion, with CME FedWatch pricing a 56.9% probability of a September hike.

Policy is splitting across regions. The ECB moved first, raising by 25bps in June on Middle East inflation pressure before holding in July on continued uncertainty over the shock's size and duration. The PBOC has moved the other way, though less than headlines suggest; its early-2026 pledge was for broadly "moderately loose" policy, but stronger growth and imported energy inflation have pushed it toward a more cautious, data-dependent stance, with structural tools favoured over blanket cuts. Fiscal policy shows a sharper divide. The U.S. is running a deficit near 5.8% of GDP in fiscal 2026 and emerging markets excluding China widened theirs to roughly 4.7% in 2025, both expansionary, while Europe has far less room with euro area growth forecast at just 1.1% for 2026.

REGIONAL ANALYSIS

U.S. equity markets have broken into record territory, with the S&P 500 clearing its June 2 all-time high on August 3, supported by strong AI-linked earnings and the oil-price relief that followed Trump's statement. That strength sits awkwardly against sharply higher rate-hike probabilities since earlier in July. The 2-year Treasury yield, the market's clearest gauge of near-term Fed expectations, has risen to approximately 4.19%, consistent with the near-57% priced for September. The "higher for longer" stance set out in our previous report remains intact, and the margin for a policy misstep narrows as valuations stretch further from the rate backdrop.

European equities benefited from the retreat in oil and the broader risk-on tone, with the Euro Stoxx 50 up in the first week of August. Growth remains the region's binding constraint, with the Eurozone projected to expand just 1.1% in 2026 on IMF and KPMG estimates, a downgrade driven largely by the Middle East energy shock. The ECB held its deposit rate at 2.25% on July 23 while flagging a likely September hike on energy-driven inflation risk, which would compound growth headwinds if realised. Combined with tighter fiscal space than the U.S., this leaves Europe the most exposed developed market should global rates remain elevated for an extended period.

Outside Korea, Asian markets have been mixed. Lower energy costs benefit import-dependent economies, while a firmer dollar and tighter global financial conditions have pressured others ahead of central-bank meetings. On August 3, most regional indices excluding Hong Kong's Hang Seng closed lower, tracking weakness in technology shares tied to the unwind in the broader chip sector.

ASSET CLASS PERFORMANCE

Equities

The S&P 500 sits at an all-time high on AI-linked earnings and oil relief, and the Euro Stoxx 50 has followed the same risk-on impulse. Korea is the outlier. KOSPI closed at roughly 6,257 on August 3, down 5.12%, after an approximately 18% single-day surge in the prior session, before recovering to around 6,600 by August 5. Two markets, one macro backdrop, and a divergence of roughly 20 percentage points over a single month.

Bonds

The front end is doing the work of pricing the hike. The 2-year yield peaked at 4.37% during the Hormuz escalation cycle before easing to 4.18% as Trump pivoted to Iran diplomacy, while the 1-year moved from 3.66% to 3.92% before settling near 3.82%. Yields at these levels raise the cost of maintaining leverage and keep short-duration government paper competitive against risk assets, which is the mechanism by which a hawkish hold transmits into positioning rather than into headlines.

Commodities

Brent broke below $80 on August 5, unwinding July's 25% surge on Trump's Iran talks announcement and an OPEC+ output increase. Precious metals held their bid regardless, with gold up to $390 on the GLD proxy and silver above $58/oz, reflecting hedging against both yield and geopolitical uncertainty rather than a directional call on either.

CONCENTRATION RISK, NOT AN AI UNWIND

South Korean equities suffered the worst monthly drawdown of any major global equity market in July 2026. KOSPI fell as much as 33% intramonth before a violent partial rebound, with an 18% single-day surge on July 31 followed by a 5.1% pullback on August 3. For context, the 1997 Asian financial crisis took the index down 27% and the 2008 global financial crisis 23%.

Most of the damage traces to a shake-out in Samsung Electronics and SK Hynix, which together account for more than half of KOSPI's market weighting and roughly two-thirds of this year's index gains. The trigger was twofold: China's CXMT completed a blockbuster chip IPO signalling domestic memory self-sufficiency, and SK Hynix posted record Q2 revenue of 79.3 trillion won that still missed consensus estimates of 84 trillion, puncturing AI-capex optimism. What followed was a record 29.2 trillion won, approximately $19.7 billion, of leveraged positions unwinding simultaneously. The speed and magnitude of the move point to forced selling via margin calls and algorithmic risk-control triggers rather than a genuine collapse in underlying fundamentals, which explains why foreign investors were selectively re-buying the largest names even as the index crashed.

The clearest illustration was the collapse of Situational Awareness, the AI-focused hedge fund run by former OpenAI researcher Leopold Aschenbrenner. Fund assets fell from a peak of $45 billion in early July to roughly $10 billion by July 30, after reported leverage of up to 400% triggered a wave of margin calls from its prime broker. Unable to meet those calls, the fund was forced into a fire sale of its leveraged public holdings, including SK Hynix and CoreWeave, to Ken Griffin's Citadel at a distressed discount. It is a textbook case of leverage amplifying losses in a concentrated trade, and the unwind fed forced selling into names already under stress from the broader KOSPI shake-out.

The critical point is what did not happen. U.S. equities were left almost entirely unscathed, with the S&P 500 holding within a four-point range over the same window and closing it around 2.5% higher. Korea functioned as a high-beta proxy for global AI-sentiment risk, an early and liquid instrument for cutting AI-linked exposure before rotating out of higher-cap U.S. names, rather than as a leading indicator dragging U.S. indices with it. This was concentration risk expressing itself, not a broad AI-sentiment collapse.

THE POLICY BIND

The U.S. inflation profile arguably calls for a further rate hike. U.S. equity markets, still leaning on AI-linked large-cap earnings for momentum, would perform better with a cut. Trump meets that tension from a position of unusual weakness, with his approval rating falling to 35% in the Reuters/Ipsos poll released on August 3, and voters trusting Democrats over Republicans on economic stewardship for the first time in nearly a decade. A hike that further destabilises AI-sensitive equities, layered on a fresh market memory of the KOSPI shock, would compound a run of losses for an administration with little political capital left to spend.

This is where the least-discussed tail risk sits. The temptation to lean on the Fed for a cut, even where the inflation data argues otherwise, would undermine the central bank's credibility on its 2% target, a scenario Chair Warsh has explicitly tried to preempt. A policy error driven by political pressure does not require anyone to be wrong about the economy; it only requires the timeline to be wrong.

LINKED RISKS, NOT SEPARATE STORIES

The September calendar concentrates the problem. The FOMC meets on the 16th and 17th, the BOJ on the 17th and 18th, back to back, with roughly 57% priced for a U.S. hike and the BOJ still normalising policy from ultra-low levels. The U.S.-Japan rate differential, and with it the funding cost of the yen carry trade, will be repriced on consecutive days into positioning that has proven repeatedly capable of moving faster than the policy that triggers it. Should the KOSPI-driven deleveraging in AI and chip exposure accelerate, the same forced-selling mechanism seen in Situational Awareness's collapse could spill into yen-funded positions, since a stronger yen and falling carry-trade assets tend to reinforce each other in a self-feeding unwind.

Three implications follow for positioning into Q4. First, a hawkish hold argues for quality, cash-generative U.S. equities over rate-sensitive and highly leveraged segments; the 2-year is already pricing this, and any hawkish surprise from the August 12 CPI print would hit leveraged positions hardest. Second, energy hedges are best sized for short-duration volatility rather than as a structural bullish call on crude, given that each Hormuz escalation has proven reversible within days once alternate supply routes and diplomatic off-ramps materialise. Third, monitoring leverage build-up in AI-adjacent funds, and stress-testing portfolios against a KOSPI-style forced unwind, is now a materially more urgent exercise than it was a month ago, because the same concentration dynamic that amplified a sector correction into KOSPI's worst month on record is present in other major benchmarks.

The dates to watch are August 12 for the U.S. July CPI print and the OPEC monthly report, September 16 and 17 for the FOMC, and September 17 and 18 for the BOJ. The scenario that matters is not any one of these risks in isolation. It is the tail where a Fed hike, a carry-trade unwind, and a fresh AI-leverage unwind arrive together. Positioning that treats these as independent stories understates the case.