TL;DR
For August 2026, the defensive tilt deepens. Healthcare (XLV) extends to #1 at 41%, Semiconductors (SOXX) trims to 30%, Defense (ITA) holds at 29%. Kevin Warsh's first FOMC delivered a hawkish hold — rates unchanged but the dot plot flipped toward hikes — which fired our pre-armed trim trigger (#6) and cut SOXX by 5 points.
5-Round Consensus Summary
| Round | Topic | Outcome |
|---|---|---|
| R1 | June review | +8.42% hit · alpha +9.98%p vs SPY (best in 5 months) |
| R2 | Macro pivots | Warsh hawkish hold · CPI +4.17% reaccel · UMich 44.8 · curve -18bp |
| R3 | Composite score | 52.4 (v2-rc) / 53.3 (v1) — neutral-bearish, lower band |
| R4 | Dissent check | "Cut SOXX -7pp" rejected — the fired trigger sanctions exactly -5pp |
| R5 | Final allocation | XLV 41% · SOXX 30% · ITA 29% |
Why the shift
Warsh's debut meeting held rates at 3.50–3.75% but revamped the statement (341 to 130 words), stripped the easing bias, and lifted the 2026 dot-plot median from 3.4% to 3.8% — six of eighteen officials now pencil in two hikes. That is precisely the "hawkish hold" our June trigger was armed for, so SOXX was trimmed 5 points into XLV mechanically, not by discretion. Beneath the surface the cracks are widening: CPI re-accelerated to +4.17% YoY, consumer sentiment collapsed to 44.8 (a third straight monthly drop), and the yield curve flattened 18bp as the dollar broke 101.
The contrarian point — narrowing breadth, not decoupling
Semiconductors rose +12% for a third straight month while the S&P fell -1.55% and the Nasdaq -3.22%. The popular read is that AI has "decoupled" and SOXX is now a safe haven. We disagree: a single sector melting up while the index falls is the classic late-cycle top signature (1999, 1972), not safety. With the Fed removing the duration premium, the most-extended asset has the least cushion — SOXX's peak was +14.5%, about 97% of our auto-trim threshold. The answer is not to sell, but to trim to the trigger and finally quantify the stretch (a SOX P/E percentile build is now our top priority).
Not investment advice. Past performance does not guarantee future results.