Skip to main content

Sector Relative Strength

Updated Jun 23, 2026 at 8:47pm

Research Draft Medium 1,231 words

Sector relative strength is the practice of ranking and rotating across the broad equity market's component sectors — operationally the eleven GICS sectors and their proxy ETFs (e.g. XLK technology, XLF financials, XLV healthcare, XLE energy, XLU utilities, XLP staples, XLY discretionary, XLI industrials, XLB materials, XLRE real estate, XLC communication services) — by measuring each sector's performance relative to a common benchmark (usually the S&P 500). The goal is top-down: tilt capital toward leadership and away from laggards rather than picking single names blind. The core tension is that two very different ideas travel under this one banner — a measured, persistent tendency for strong sectors/industries to keep outperforming over the next 1–12 months (industry momentum, which has real academic support), versus a discretionary economic-cycle story ("which sector leads which phase of the business cycle") that is far softer than its popularity implies.

How it's measured

The atom is the relative-strength ratio: divide a sector's price (or total-return index) by the benchmark, Sector / SPX. A rising ratio line means the sector is outperforming regardless of whether both are going up or down; a falling line means underperformance. Practitioners then rank all eleven sectors by ratio trend (and by raw trailing return) to produce a leadership table.

Relative Rotation Graphs (RRG), developed by Julius de Kempenaer (launched on Bloomberg terminals in 2011, on StockCharts in 2014), formalize this into two normalized indicators plotted as a scatter:

  • JdK RS-Ratio (x-axis): a normalized measure of the level of relative strength vs the benchmark.
  • JdK RS-Momentum (y-axis): the rate of change of RS-Ratio — momentum of the relative strength.

Both are centered on 100. This carves the plane into four quadrants that securities tend to traverse clockwise: Improving (weak but accelerating, upper-left) → Leading (strong and accelerating, upper-right) → Weakening (strong but decelerating, lower-right) → Lagging (weak and decelerating, lower-left). A "tail" shows the recent path; tail length and direction are read as much as the current quadrant.

How it's used in practice

  • Leadership tilt (the robust core). Overweight sectors in the Leading quadrant or with the strongest rising RS-ratios; underweight or avoid those in Lagging. This is a top-down momentum filter: pick names from leading groups, fade weak ones.
  • Rotation anticipation. Improving→Leading transitions are treated as the highest-conviction additions; Leading→Weakening as a profit-taking / trim cue; the clockwise path is used to anticipate the next group to rotate up.
  • Group context for stock selection. A stock in a leading sector has a tailwind; the same setup in a lagging sector is a tougher trade. This dovetails with O'Neil-style "buy leaders in leading groups" thinking.
  • The economic-cycle map (the soft layer). The Sam Stovall / S&P sector-rotation model maps sectors to four business-cycle stages: early-cycle favors economically sensitive groups (industrials, technology, financials, discretionary); mid-cycle is associated with technology and communication strength; late-cycle rotates toward defensives (healthcare, staples, energy); recession favors classic defensives (utilities, staples, healthcare) before the recovery bid returns in late recession. This is a useful mental framework for where we are, but it is prescriptive folklore, not a timing edge (see below).

Standing & evidence

This is the section where the two ideas must be cleanly separated.

Industry/sector momentum — documented and meaningful. Moskowitz & Grinblatt ("Do Industries Explain Momentum?", Journal of Finance 1999) found a strong, prevalent momentum effect in industry components of returns — buying past-winning industries and shorting past-losing ones was significantly profitable even after controlling for size, book-to-market, individual-stock momentum, and microstructure, and they argued industry momentum accounts for much of the better-known individual-stock momentum anomaly. Later work (e.g. Grobys & Kolari, Journal of Financial Research 2020) extended this, documenting a short-formation (one-month) risk-managed industry-momentum spread that they report as significantly priced in the cross-section. So the relative-strength persistence underpinning a leadership tilt has real empirical support — though, like all momentum, it is conditional, subject to violent "momentum crashes," and partly eroded by trading costs.

Business-cycle sector rotation — much weaker than presented. The discretionary "which sector leads which phase" mapping does not hold up well under testing. Molchanov & Stangl ("The Myth of Business Cycle Sector Rotation," International Journal of Finance & Economics 2024) found no systematic sector outperformance where the popular model says it should occur; at best, conventional sector rotation produced only modest outperformance that diminished once realistic transaction costs and the practical difficulty of timing business-cycle turning points were accounted for. When they let any industry's excess return predict other industries' future returns, predictability was not significantly different from random chance. Cycles also do not repeat cleanly — phase boundaries are only knowable in hindsight, sector composition drifts (technology's weight today bears little resemblance to prior cycles), and policy/rate regimes differ each time. Treat the cycle map as a low-confidence narrative overlay, not a rule.

Strengths & limitations

  • When it works: trending, sector-dispersed markets where leadership is durable — the leadership-tilt / RS-momentum approach captures real persistence. Sector dispersion gives the ranking something to bite on.
  • When it fails: (1) Whipsaw / mean-reversion regimes and sharp risk-on/risk-off reversals, where last month's leaders become next month's laggards (momentum crashes). (2) Low-dispersion melt-ups, where almost everything moves together and the ranking is noise. (3) Cycle-timing misuse — acting on the Stovall map as if phase→sector were deterministic; the turning points are unknowable in real time and the historical "edge" largely vanishes after costs and mistiming.
  • #1 misuse: lending the credibility of measured industry-momentum to the discretionary cycle-rotation story. They are not the same claim. The persistence of leadership is evidenced; "buy industrials because we're in early-cycle" is not.
  • Regime/timeframe dependence: RS-Ratio/RS-Momentum readings are highly sensitive to lookback length and benchmark choice; a sector can be "Leading" on a weekly RRG and "Lagging" on a daily one. RRG quadrants are descriptive, not predictive — they label the present, they do not guarantee the clockwise path continues.

System relevance

This node is the sector-level member of the relative-strength & rotation branch — it pairs with the broad relative-strength definition (stock-vs-benchmark RS) and defers all swing-specific operational mechanics (exact entry/stop/target, hold period, position sizing on a sector-leader trade) to the Swing Trading branch; don't duplicate those here. For the Augustus trade-setup agent, the decision-useful output is a directional bias (favor setups in leading sectors/groups, demand more confirmation against laggards), supplied as one weak-to-moderate contextual input — not a standalone signal. Hard caveat for the agent: weight the leadership-momentum signal above the economic-cycle map, and never treat "we are in X phase, therefore buy sector Y" as a verified edge — Molchanov–Stangl shows it largely isn't one after costs and mistiming.

Sources