Sector rotation

Definition

Sector rotation overweights and underweights industry sectors based on where the economy sits in its cycle (the classical version) or on relative price strength (the momentum version), usually implemented with sector ETFs. The premise: sector leadership is not random — it follows cycle position, rate environment, and persistence effects that can be measured.

How it works / structure

• Cycle version: map macro state (growth/inflation/rates direction — pillar-8 entries) to historically favored sectors (early-cycle cyclicals, late-cycle energy/staples etc.); the schedule is folklore-adjacent and treated as hypothesis, not fact (see evidence).
• Momentum version: rank sectors by trailing relative strength, hold leaders, rebalance on schedule — industry momentum with documented evidence (Moskowitz-Grinblatt 1999).
• Parameters (engine-executable): sector universe (e.g. 11 GICS ETFs), signal (macro-state map vs N-month relative strength), holding count, rebalance frequency, benchmark, and the exposure accounting below.
• Exposure honesty: sector tilts decompose heavily into factor/beta exposures (strategy-factor-investing); attribution against factor benchmarks is required or rotation skill is claimed for beta.

When it applies

ETF-implementable universes, monthly-scale horizons, and macro theses that want equity expression without single-name risk (lens-market, lens-macro). Cycle versions require an honest answer to “where are we in the cycle?” — a question NBER answers only in hindsight.

Risk profile & failure modes

• Cycle-dating lag: regimes are identified retrospectively; live cycle calls are frequently wrong, and the rotation schedule’s damage concentrates at turning points.
• Whipsaw at rebalances: sector leadership mean-reverts at short horizons; over-frequent rotation pays friction to chase noise.
• Concentration correlation: sector portfolios are lumpy — a 3-sector tilt can be one macro factor in disguise (risk-correlation-exposure).
• Backtest-schedule folklore: the classic early/mid/late- cycle sector map is derived from few cycles and rarely snooping-adjusted — labeled folklore unless a cited test accompanies it.

Evidence & limits

Moskowitz-Grinblatt (1999) documented strong industry-level momentum (industry winners persisted) — the momentum version’s evidence base, with the usual decay caveats. Cycle-schedule rotation lacks comparable peer-reviewed support: cycle definitions are retrospective (NBER) and samples contain roughly a dozen complete cycles — too few for reliable schedules. The platform treats rotation as industry momentum (evidenced) or as a macro thesis (graded per thesis), never as a calendar.

Falsifiable-thesis examples

Illustrations only, not signals:

• “The top-3 sectors by 6-month relative strength will outperform the equal-weight sector average over the next quarter” — falsified by the realized spread.
• “Energy will outperform the broad index in the next two quarters given [stated macro condition]” — falsified by the pair’s returns.

Cross-references

• Evidence engine: strategy-momentum (industry variant)
• Exposure decomposition: strategy-factor-investing, port-exposure-netting
• Macro state inputs: regime-rate-environments, macro-inflation-linkages, lens-macro
• Implementation: instrument-etf, port-allocation-frameworks