Positioning and Outlook - 2026-08-18

Published 2026-08-18

Stocks recovered from an Iran-driven selloff as AI earnings and rate-hold expectations lifted technology shares, but widening high-yield spreads and a deteriorating corporate credit-survival screen point to mounting stress beneath the surface.

Recent Events

The past few weeks were dominated by the Iran conflict and its spillovers: stocks fell as Middle East escalation landed just ahead of the Federal Reserve's rate decision, OPEC cut its global oil demand growth forecast amid disruptions around the Strait of Hormuz, and the dollar stayed feeble as rate-hike bets dwindled. Inflation data later in the period supported expectations that the Fed would hold, and Wall Street gained as AI earnings lifted technology shares, while China's effort to ease the Iran-war aluminium shock came, by the reporting's own account, at a cost. Credit told a less comfortable story, high-yield spreads widened before flattening out, and the corporate credit-survival screen kept roughly half its coverage in the danger zone, with the average score slipping to 22 out of 100. Single-name items were secondary but worth noting: On Holding trimmed its sales outlook after missing quarterly revenue targets, a central bank speech from Cook addressed the outlook for the U.S.

Upcoming Events

The forward calendar runs quiet before it turns heavy: the nearer term offers the PBOC's loan prime rate fixing on August 20, while the real macro spine arrives in September, with the ECB's Governing Council on the 10th, the FOMC decision on the 16th and the Bank of England's MPC barely a day later, a cluster that compresses the window in which the tape has to absorb three policy signals at once. Around that, the scheduled fare is moderate-impact filler: a run of short-dated bill auctions through mid-August and a handful of phase 2 readouts, including ensifentrine and raludotatug deruxtecan from MRK and rilpivirine from JNJ. Earnings supply the background hum, with 285 reports slated across the window, though the nearest names on the slate are small-caps rather than tape-movers.

Macro Projections

The forward calendar is dense, GDP and the PCE inflation gauge on August 26, payrolls on September 4, the ECB on September 10, but the cluster that sets the outlook is the August CPI print on September 11 into the FOMC on September 16, with the Bank of England a day behind. Kalshi prices a 63% chance that CPI runs above 3.3% year over year (88% above 3.2%), and with a cut before 2027 at just 16.3% and the two-year yield pinned at 4.17%, the near-term mechanism is direct: a hot print keeps the Fed locked on hold (hold propensity sits at 0.85) and transmits straight into more pressure on the long end, while a soft print is the only path that revives cut pricing. The fork is explicit: at 3.5% or higher the long-end selloff extends toward a 30-year yield above 5.5% and the hold regime hardens into hike-risk repricing; at 3.1% or lower the steepening stalls and cut pricing revives into year-end. The months-to-quarters regime is split, credit is calm and still tightening, with high-yield spreads (the extra yield junk bonds pay over Treasuries) at 2.67% as of August 14, down from 2.85% at end-July, the CCC-versus-BB quality reading steady near 2.0 and false-bottom risk low, while the stress sits in duration and currencies: the 30-year Treasury at 5.32% is a 95.4th-percentile reading last exceeded in June 2007 and still accelerating, against a 10-year at 4.74% (58.8th percentile) that is rising but not extreme, the signature of a long-end term-premium move confirmed by the curve steepening to +0.53. G7 currency volatility, how sharply the major exchange rates are swinging, is elevated and accelerating with the dollar through 160 yen, and the carry model holds disorderly-unwind odds at 38.9%, up from 34% but plateaued rather than building; equity volatility is calm, the VIX at 15.19, even as hedging builds underneath, with the tail-risk gauges SKEW at 138.4 and VVIX at 93.9 both accelerating. The dominant tension is that the two loudest news themes, geopolitics and rates, point to stress that tier-one market data refuses to confirm: the Iran–Hormuz conflict is live, Iran says the strait stays shut until US conditions are met and Houthi forces attacked shipping within the past two days, with a geopolitical-risk index at 144.5 off a recent 225 peak, yet Kalshi prices Brent above $90 at only 18%, down 15 points in 24 hours, against a 17.4-million-barrel crude build, so the model reads a genuine unpriced tail of roughly 30–35% for Brent above $90 within 30 days, about 15 points of gap, with the inventory cushion keeping the base case range-bound. On the multi-year view the chain lengthens and confidence thins: US equities are priced for perfection at a cyclically adjusted price-earnings multiple of 42.4 and a 3.98% earnings yield against 5.38% in developed markets ex-US and 5.98% in emerging markets, and the long-end term-premium build reads as the early edge of a structural repricing rather than a cyclical wiggle, weighable, base-rate-grounded channels, not guarantees, and held at lower confidence than the nearer mechanisms. The calm-credit regime itself carries roughly a 30% chance of breaking by Q4, flagged in advance by a weekly high-yield close above 3.0% with a rising VIX, and the framing is falsified outright if spreads gap above 3.0% on any weekly close, if the 30-year falls back below 5.0%, or if Brent settles above $95, any one would break the calm-credit, long-end-premium, unpriced-oil-tail trinity the outlook is built on; the petroleum inputs, dated August 7, are the oldest in the stack and are discounted accordingly.

Positioning

With high-yield credit spreads near 2.67% pointing to a calm credit backdrop, the model portfolio sizes each position in proportion to its screening-kernel composite score, a modelled blend of expected upside drift against drawdown probability, so the highest-conviction names carry the largest equity weights while weaker scores stay small. Around that core, overall risk appetite is calibrated by a set of structural cause-and-effect channels, including merger-arb completion, oil-supply shocks feeding headline inflation, and commodity production losses, each carried with its full historical firing record rather than treated as a forecast. Those mechanisms have fired with hit rates clustered in the 80%-plus range across a combined five dozen or so instances, and each is discounted for the share the market already prices, which keeps them as weighable inputs to sizing rather than directional bets. The resulting weights, scores, and the dossier behind each holding are set out in the tables that follow.

Model Portfolio

The model portfolio is, first and foremost, a book held in waiting: money-market exposure and the unallocated sleeve together make up the overwhelming share of the allocation, kept as dry powder for staged deployment as the credit and volatility picture resolves. The only committed position beyond cash is a modest weight in the Invesco DB Commodity Index Tracking Fund, a diversified, futures-based commodity sleeve that lends the book some sensitivity to raw-material prices while the rest sits parked. In shape, then, this is a defensive, liquidity-heavy portfolio, capital preserved rather than deployed, with a single real-asset outrider alongside.

Track Record

The model's track record is young, about two months of evaluated ideas since mid-June, and the early scoreline is more sobering than flattering. Just over half of recommendations beat their benchmark a day out, but that edge erodes with time: fewer than half are ahead at one week, and only about a third at four weeks, where average alpha runs around negative 1.6 percent. The three-month horizon has yet to produce its first scored outcomes, so the sample remains thin and no durable verdict should be drawn from it. The point of the exercise is the discipline rather than the grade: every call is written down before the outcome is known and tallied when the horizon arrives, the misses alongside the winners.

Macro forecasts: 75 graded, 393 open — each call is scored against what actually happened when its horizon arrived.