Gold & SPX Ratio-Momentum Rotation (QuantifiedStrategies)
A fully-disclosed two-asset rotation from quantified-strategies: hold whichever of S&P 500 / gold is favored by the price-ratio-vs-moving-average signal, decided at each month end. Unlike the gold-spx-switching-strategy (Returnstream) variant there is no cash/T-bill leg and no absolute-momentum filter — you are always in one of the two assets.
Rules (fully disclosed in the free article)
- Monthly closing prices; S&P 500 price index (Yahoo
^GSPC) and monthly spot gold. - Compute the S&P 500 / gold price ratio and its 20-month SMA.
- Ratio above SMA next-month signal → long S&P 500 for the coming month.
- Ratio below SMA → long gold for the coming month.
- Rebalance at month close, repeat. Always fully invested.
Claimed results (author backtest, 1985/86 – Jul 2020, no costs stated)
| Metric | Rotation | S&P 500 B&H | Gold B&H |
|---|---|---|---|
| CAGR | 9.9% | n/a (not stated) | n/a |
| Max DD | 22% | 55% | n/a |
Equity chart (100k, Jul 1986 → Jul 2020): rotation ≈ 2.35M, SPX ≈ 1.4M, gold ≈ 0.55M. Per-asset accumulated gains “pretty equal” — SPX months and gold months each contributed roughly half the strategy’s return.
Librarian assessment
- Signal is fully specified → trivially replicable; unlike returnstream the paywall hides nothing essential.
- Dividend exclusion is the big one:
^GSPCis price-only, and the strategy sits in SPX most months. SPX total return ≈ price return + ~2%/yr over this window, so dividend-adjusted SPX buy-and-hold likely matches or beats the 9.9% figure; the drawdown edge (22% vs 55%) is the more durable part of the claim. - Edge is window-dependent: the equity curves only separate around/after the GFC; before 2008 the rotation trailed SPX. Backtest stops Jul 2020, missing gold’s 2024–25 run.
- 20-month SMA is a high-smoothing choice; only that length was tested — parameter sensitivity unexamined by the author.
Open questions
- Does the claim survive dividend-adjusted SPX and realistic costs?
- Sensitivity to SMA length (10/15/20/30 months)?
- Would adding the Returnstream-style cash leg improve the DD profile further, or just cut returns?
- Intra-ratio direction confusion: article body uses SPX/gold ratio, summary image says gold/SPX with the same buy rule — replication must pin the direction down.
See claims qs-rotation-beats-both-buyhold, qs-rotation-gfc-divergence, qs-rotation-asset-contributions-equal and test plan gold-spx-ratio-rotation-replication.