The Weekly Alpha Digest

Macro Update

Start with the mechanics, because the narrative built on top of them has almost nothing to do with what happened. The fund opened in late 2024 with $225m, compounded to somewhere between $20bn and $24bn by the middle of this year, and was up around 439% net at the end of June. It is now close to $10bn, most of that locked in private positions it cannot exit. CNBC reported the listed portfolio changing hands in a single block ahead of Thursday’s open, and Bloomberg named Citadel as the buyer. The holdings overlapped heavily with our own thematic universe, and not one of them saw its operating outlook deteriorate over the weeks it was being dumped.

A liquidation of that size still cannot bend a global factor unassisted, and it had company. Korean retail had accumulated leveraged single-stock positions worth almost 3% of free float, an exposure ratio somewhere between four and seven times what any comparable market carries, spread across an estimated 300,000 to 400,000 margin accounts that were closed out inside a fortnight. Goldman’s prime data shows tech gross being cut harder than at any point since the series began. Set a large forced seller alongside a leveraged retail base and the ordinary rebalancing that follows a record quarter, and the month requires no fundamental explanation whatsoever. Contract pricing in DRAM, high-bandwidth memory bookings and fuel cell backlogs all moved the other way while it was happening.

What lit the fuse was policy. Two Warsh press conferences have now passed with almost nothing said, and the silence is deliberate: he wants the Fed to occupy less space in market pricing. Score the language rather than the substance, though, and each appearance comes out tighter than the one before, at the same time as our own inflation tracking points down and keeps pointing down. Silence read as hawkishness is not costless. A leveraged, long-duration equity position is rarely undone by time; it is undone by a jump in real yields, and long-end reals were climbing precisely as the most crowded trade in global markets began to slip.

Both how far under consensus the coming prints sit, and the full series behind that call are reserved for Pro, but we can share our conclusion. A central bank that publishes a 2% target, refuses to forecast, and then refuses to tighten with inflation running above target has already told you what it expects, whatever it declines to say out loud. A hike this year looks unlikely to us, and we are content to sit opposite the real-rate panic while that resolves.

The weekend’s more consequential development had nothing to do with the Middle East. Washington and Tokyo bought yen together on Friday, the first jointly executed operation of its kind in fifteen years, and by Sunday both had confirmed it and signaled a readiness to do it again. The trigger was almost comic. A press photographer at the Camp David cabinet session got a clean shot of the Treasury Secretary’s notes, where the instruction “Buy Japanese Yen (JPY) $5-10 bil” sat waiting to be carried out. It was duly carried out. USD/JPY has since retreated from four-decade highs around 164 into the high 150s, and the United States achieved that without parting with a single dollar. The inversion is the whole point. In 2011 the problem was a yen that had grown too strong. This time the object is a ceiling on the dollar, installed without any official change in dollar policy.

How it was financed is the more interesting question, because the answer identifies which currency pair absorbs the strain if a one-off hardens into a standing policy, and that is the input that a position size actually turns on. We work through it, with the positioning data, in Pro. What is worth stating in the open is the alignment. A Treasury talking its own currency down against its principal funding pair, and endorsing tighter policy abroad while doing so, arrives in the same week our inflation work has the United States and the euro area diverging by the widest margin in years, in the same direction. Add growth odds stabilizing, our rising-inflation probability well off its highs, liquidity still running better than the consensus allows, and a refunding calendar with no supply shock buried in it. That combination has historically supported risk and weighed on the dollar. It is also the configuration in which a month like July reads as noise.

Geopolitical Update

The conflict has run in a repeating cycle all year, and it has just turned over at the top again. Threats escalated to language about force not deployed since the Second World War, and then, with the market braced for the follow-through, came a stand-down justified by a deal whose “perimeters” had supposedly been settled, a reopened Hormuz included. We noted a fortnight ago that the threats against shipping matched the register used immediately before the early-April peak almost word for word. Neither the cast nor the script has changed since.

None of this reflects an inability to keep fighting. March consumed interceptors and munitions at a rate nobody intended to sustain, but the tempo since operations resumed has been deliberately low, which is the slide into a grinding low-intensity conflict we expected rather than a supply problem. Washington can carry on. It has decided it would rather not. Track the objectives US officials have named in public and the contraction is stark: a long opening list narrowed to enrichment plus the strait through April, and now enrichment itself pushed down the queue while everything bends toward getting traffic moving again, up to and including an arrangement in which Tehran collects transit fees, possibly shared with Muscat. Having recognized the position he had put himself in, Trump chose to stop making it worse, and he has swallowed a good deal of public humiliation rather than widen the war or commit troops. We expect this to settle into a second memorandum over the next few weeks, then a quiet stretch, then a smaller third episode at some point beyond the midterms.

The market’s instinct here is the wrong one. A second deal has far less stored relief to release, because no fear was priced in on the way up. In the spring, dated crude traded through $120, the Houston differential blew out to $7 a barrel as buyers competed for American cargoes, and any hint of a settlement was worth tens of dollars. That differential is now closer to a dollar and crude never revisited its highs. Physical participants priced this round as a risk they already understood, which leaves very little to hand back when the handshake finally arrives. There is no relief rally available in an episode that never frightened anybody, and buying one is a trade on the previous conflict rather than this one.

Refined products are the exception. Cracks remain above their 2022 highs, and it has been a strong year for refiners, so if Gulf shipping, freight and insurance normalize together, diesel and gasoline margins have room to compress. That is a margin trade rather than a directional crude view, and it is certainly not a reason to buy everything. Yemen is the loose end. The Houthis entered late, announced a blockade of Saudi shipping on their own initiative, and are not reliably steered from Tehran, so an agreement covering the Gulf leaves them exactly where they are. Two chokepoints need watching now rather than one.

Russia earns a paragraph, because a funding problem is being widely misread as a hard stop. Debt auctions were shelved indefinitely on July 20, and the quarter has raised 8.8bn rubles against a 1.5 trillion plan, which is a rounding error on the target. The bid exists; the price does not. Moscow will not pay what the market is asking, which puts its tolerance somewhere below the 16.5% the ten-year currently offers. Nor was the sovereign fund ever what paid for this war. Debt issuance did the heavy lifting, and with the auction channel shut, state banks are warehousing the paper against central bank liquidity, which is money printing routed through an intermediary. The binding constraint is the cost of financing rather than its availability, and it will surface as inflation rather than as a policy announcement.

Portfolio Update

We are leaving the ETF allocation untouched, and that is a decision rather than a default. Open gains across the book stand near 17% on current marks, with the memory and bottleneck sleeve, a fifth of the portfolio, still the dominant contributor, which is precisely the exposure July was supposedly writing off. Coming through the factor’s worst month on record in the black, while holding several of the securities being liquidated into it, is an outcome we are content with. Sleeve-level marks and the full position detail sit on the Pro portfolio page.

On fundamentals, we would own most of what that fund was forced to sell. Its listed book mapped closely onto our own bottleneck and electricity exposures; the operating picture strengthened while the selling was underway, the seller no longer exists, and the largest block of technical supply in this cycle was absorbed in a single print. Standing alongside the buyer in that transaction appeals rather more than standing alongside the running commentary.

Timing is a separate question, and we are not pretending otherwise. July did lasting damage to how the theme is perceived as much as to its chart. Technical levels have gone, momentum signals are fully short, the Korean retail base has been wiped out, and every institution that bought the story near the peak now has months of supply to work off into any rally. None of that mends quickly, whatever the operating data says, so we build at our own pace and reserve capital for the retest. Standing still is also a decision.

Beyond the ETF book, nothing has changed, and conviction is, if anything, higher. A short dollar position remains the core holding. We stay with European gas into a heating season carrying the smallest storage buffer in fifteen years, with the power build-out, and with gold as protection against currency debasement. We hold no directional view on crude while two chokepoints drive the headlines, and we continue to lend against the real-rate panic at the front of the US curve.

None of it has to be done this morning. Having survived last month, we can afford to be selective in this one.