Positioning and Outlook - 2026-08-18

Published 2026-08-18

Equities steadied as strong AI earnings and rate-hold expectations offset the Iran conflict's spillover into oil and commodity markets, but credit stayed fragile, with the high-yield model keeping its defensive cash posture and the credit survival screen edging weaker.

Recent Events

The past few weeks were dominated by the Iran conflict's spillover into markets: stocks fell on the Mideast escalation ahead of the Federal Reserve's rate decision in late July, OPEC cut its global oil demand growth forecast as disruptions hit the Strait of Hormuz, and China moved to ease the war-driven aluminium shock, though at a cost, while the dollar softened as rate-hike bets dwindled. Equities nonetheless found a footing as AI earnings lifted tech and the inflation data reinforced rate-hold wagers, even as On Holding tempered its sales outlook after missing quarterly revenue targets. Beneath the index level, credit stayed fragile, the high-yield spread gauge went from widening to flat and kept its defensive cash posture throughout, and the credit survival screen slipped from an average of 23 out of 100 with 50 names in danger to 22 out of 100 with 51. The period also brought scheduled remarks from Cook on the outlook for the US and Alaskan economies, Class I device recalls touching Boston Scientific's neuromodulation unit and AVID Medical, and an 8-K flagging a material agreement alongside director and officer changes, a full slate of threads for the days ahead.

Upcoming Events

The forward calendar is distinctly back-loaded. The next several sessions offer mostly routine bill supply, a few mid-stage trial readouts from Merck and Johnson & Johnson, and a busy but diffuse earnings slate, 285 reports in the window, adding breadth rather than one dominant print. The first macro marker is the People's Bank of China's loan prime rate fixing on 20 August, after which the schedule compresses sharply: the ECB's policy meeting on 10 September, the FOMC decision on 16 September and the Bank of England's rate decision a day later place three high-impact central-bank events in the space of a week. That cluster forms the spine of the period, concentrating most of the scheduled event risk into a handful of sessions and leaving the tape little room to reset between them.

Macro Projections

The run into mid-September is calendar-dense, GDP and the PCE price index on 8/26, payrolls 9/4, the ECB 9/10, CPI 9/11, the FOMC 9/16 and the Bank of England 9/17, but the hinge is the inflation print landing five days before the Fed. Prediction markets price August CPI above 3.3% year-on-year at 63% (and above 3.2% at 88%), so the near-term mechanism is close to binary: a hot print locks the hold, a cut before 2027 is priced at just 16.3%, and transmits straight into more long-end pressure, while a soft print is the only channel back to revived easing. The model carries that as an explicit fork: at 3.5% or higher the long-end selloff extends, the 30-year goes through 5.5%, and hold pricing hardens into hike-risk repricing; at 3.1% or lower the steepening stalls and cut pricing rebuilds into year-end. Underneath, the months-to-quarters regime is calm credit against stressed duration: high-yield spreads (the option-adjusted premium junk bonds pay over Treasuries) have tightened to 2.67% from 2.85% at end-July with the CCC-versus-BB quality mix stable near 2.0 and false-bottom risk low, while the 30-year at 5.32% sits at its 95.4th percentile, last higher in June 2007, and the 10-year at 4.74% is only mid-range, which, with the curve steepening to +0.53 and the 2-year pinned at 4.17%, marks this as a long-end term-premium move consistent with a Fed going nowhere (hold propensity 0.85). The second cyclical pressure point is currency: G7 FX volatility, how sharply the major exchange rates are swinging, is elevated and accelerating with the dollar through 160 yen, and the carry model holds the odds of a disorderly unwind at 38.9%, a plateau stepped up from 34% rather than a risk still rising; equity volatility itself stays calm at a 15.19 VIX, but hedging is quietly building, with SKEW (the price of crash protection) at 138.4 and VVIX (expected swings in the VIX itself) at 93.9, both accelerating. The dominant tension is that both lead news themes, geopolitics and rates, point to stress the market data refuse to confirm: Iran says the Strait of Hormuz stays shut until US conditions are met, Houthi attacks on shipping are two days old, and the geopolitical-risk index sits 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; the model reads a genuinely unpriced tail at roughly 30–35% odds of a $90 print within 30 days, though the inventory cushion (petroleum data as of 8/7, the oldest input and discounted accordingly) argues the base case stays range-bound. Confidence thins as the horizon lengthens: over the multi-year view the anchor is valuation, with US equities priced for perfection at a 42.4 cyclically adjusted price-earnings multiple and an earnings yield of 3.98% against 5.38% in developed markets ex-US and 5.98% in emerging, a 140-to-200-basis-point gap, and the model assigns roughly 30% to the calm-credit regime breaking by Q4, with a weekly high-yield close above 3.0% alongside a rising VIX as the lead indicator. The framing stands or falls on three tripwires, high-yield spreads gapping above 3.0% on any weekly close, the 30-year back below 5.0%, or Brent settling above $95, because these are weighable cause-and-effect channels rather than predictions, and any one breach would signal that the calm-credit, long-end-premium, unpriced-oil-tail trinity has broken.

Positioning

Position sizes in the model portfolio are a direct readout of conviction: each name's weight scales to its screening-kernel composite score, a modelled blend of expected upside drift weighed against drawdown probability, so the highest-scoring names carry the largest allocations. That conviction is being expressed against a calm credit backdrop, with high-yield spreads near 2.67%, which leaves the book comfortable leaning into its stronger signals rather than defensively trimming them. Calibrating the overall risk appetite are a handful of structural mechanisms, merger-arb completion dynamics, an oil supply shock channel into headline inflation, and commodity production losses, each carried as a weighable input with its full firing history, misses included, and discounted for the share of the move the market already prices. The result is a book whose shape is systematic rather than discretionary, with the holdings that follow reflecting where the composite scores, and the mechanism evidence behind them, actually point.

Model Portfolio

The model portfolio is sitting almost entirely on the sidelines: half the book is parked in money-market exposure, explicitly designated as dry powder, with the balance left unallocated. Nothing is currently committed to risk assets, so the posture is one of patience rather than conviction, capital is being held back for staged deployment as the credit and volatility picture resolves. In effect, the book's full weight remains available to rotate once conditions clarify, making this a waiting position rather than an expressed market view.

Track Record

Every idea the model puts forward is written down before the outcome is known and scored when its horizon arrives, and the tally below counts all 757 evaluated calls rather than a flattering subset. The record so far is candidly mixed: just over half of one-day calls have beaten their benchmark, yet average alpha at that horizon is fractionally negative, and both hit rate and alpha fade as the window stretches, by four weeks only about a third of ideas have finished ahead. Two cautions cut against reading too much into any of it: the entire sample spans just 65 days of one market stretch, so the hundreds of observations are far from independent, and the three-month horizon has yet to mature, with its first outcomes due around mid-September.

Macro forecasts: 75 graded, 396 open — each call is scored against what actually happened when its horizon arrived.