Asset Class Momentum Rotational System

Der Drei-Gate-Härtetest · echt nur, wenn es alle drei nimmt und adversarische Widerlegung übersteht

Gate 1

Survivorship-frei

immune

sauberes Panel

Hier eliminiert

Gate 2

Placebo ≥ P95

P58

~116 von 200 Körben übertroffen

Gate 3

Kostenbewusstes Netto

RF +5.87

positiv, nicht zertifiziert

Gescheitert

Schlechtester 12-Monats-Abschnitt (RF)-0.55

−1.00 Untergrenze0

Jede Strategie hier — Gewinner eingeschlossen — verliert in ihren schlechtesten 12 Monaten. Tiefe ist ehrlicher Kontext, nicht das Urteil.

Am Glücks-Gate abgelehnt — sein Netto-Rang war nicht besser als Zufallskörbe (unter der P95 Skill-Linie).

Each month, this system ranks seven asset-class ETFs (SPY, EFA, EEM, TLT, IEF, GLD, VNQ) by their trailing 12-month return and holds the top three, per Faber's relative-momentum approach. The idea is that recent relative strength carries forward across asset classes.

What we found

The momentum ranking added nothing detectable. Its placebo percentile was 58, meaning random selections of the same ETFs did about as well — there is no measurable rank-skill in the momentum signal. An equal-weight buy-and-hold of all seven ETFs outperformed the rotation over the test window, so the results reflect passive asset-class exposure (beta) rather than any timing or selection edge. We reject it as a standalone effect.

How we tested it

2005–2026 test windowmodelled liquidity-aware costssurvivorship immune

Tested on a set of liquid asset-class ETFs (SPY, EFA, EEM, TLT, IEF, GLD, VNQ), monthly rebalancing. Realistic modelled costs.
Placebo / robustness test: real result vs random baskets or shuffled signals (real vs the 95th percentile of random)

Source: Faber (2010) relative-momentum across asset-class ETFs; QuantPedia
Read the paper ↗

← The Academic Audit — all 54 studies

Research, not investment advice. “Validated” factor-legs are market-neutral diversifying building blocks with a losing worst year — none is a standalone tradeable strategy. Metrics are cost-aware and modelled (not live fills); the 2005–2026 test window is out-of-sample versus the source paper. Dollar figures are not returns and are omitted by design.