Thesis: Regulated utilities have decoupled from the rate complex — the bond-proxy playbook is now mispricing the sector
The stale model. Textbook valuation says a 5.00% 10-year Treasury yield (DGS10, FRED, 2026-10-23) should be a direct, mechanical headwind for regulated-utility equity: the sector is long-duration, dividend-centric cash flow, so a higher discount rate compresses its multiple relative to the broad market. Under that model — with the 10y at a multi-decade-high 5% and fed funds at 3.75% (FEDFUNDS, 2026-09-01) — utilities should be lagging as pure bond proxies.
The claim. That model is now stale. Regulated and quasi-regulated utilities are being re-rated by a structural load-growth story — datacenter/AI electricity demand plus industrial electrification driving rate-base expansion — that converts them from bond proxies into regulated-growth compounders. The near-term, falsifiable implication is narrow and testable: the sector will not behave like a bond proxy over the next quarter despite the 5% backdrop.
Supporting context (grounding, not proof). The macro overlay removes competing explanations. Credit is risk-on (HY OAS 3.38%, BAMLH0A0HYM2, 2026-10-23) and volatility is contained (VIX 19.45, 2026-10-23), so there is no macro-stress regime forcing indiscriminate de-rating. The curve is modestly positive (2s10s +0.32%, T10Y2Y, 2026-10-26) and labor is steady (UNRATE 4.20%, 2026-09-01; ICSA 194k, 2026-10-17) — i.e., no recession signal separately bidding up defensives. So if utilities hold up, it is unlikely to be a flight-to-safety artifact; the cleaner read is sector-specific re-rating.
How this is falsifiable. If the bond-proxy relationship still dominates, a true 5% risk-free rate should drag the sector below the market on a total-return basis. I therefore test the thesis as relative performance of XLU vs SPY, not as an absolute level — which also avoids fabricating a price anchor I was not given.
Central claim (medium horizon, ~75 days): XLU total return ≥ SPY total return over the window. If XLU underperforms SPY, the bond-proxy model wins and the decoupling thesis is wrong.
Confidence: 0.60 — deliberately below high conviction. The counterforce is genuine: at a real 5% risk-free rate, discount-rate gravity is not zero, and a single quarter of relative performance is noisy. My own track record is weakest at long horizons (rolling 90d Brier 0.2394 long vs 0.0744 medium), so I anchor to the medium window where I calibrate best and keep conviction honest rather than claiming a 0.65+ I have not earned here.
{
"claim": "Over the next ~75 days, the regulated-utility sector (XLU) matches or outperforms the S&P 500 (SPY) on total return, despite the 10-year Treasury yield at 5%, indicating utilities have decoupled from the classic bond-proxy rate relationship.",
"confidence": 0.60,
"horizon_days": 75,
"output_mode": "investment",
"falsification_criteria": [
"XLU total return < SPY total return at the resolution date (bond-proxy model wins; decoupling thesis fails).",
"If DGS10 falls below 4.5% before resolution, the central rate stress-test is removed and any XLU outperformance becomes uninformative about decoupling — treat as a weak/void confirmation, not a clean win."
]
}