Direct Answer
Industry rotation applies relative-strength and momentum analysis one tier below the GICS sector level, to individual GICS industries or sub-industries, using the same ratio-chart methodology described in Sector Relative Strength and Momentum. Because a GICS sector's return is a market-cap-weighted blend of every industry inside it, a sector can post a flat or modestly positive aggregate return while its constituent industries are rotating sharply — for example, semiconductors rallying hard while software lags within Information Technology, with the two moves largely canceling out at the sector level.
A sector-level relative strength view alone misses this. It answers "is Technology as a whole attracting capital?" but not "which part of Technology is actually attracting it?" Industry-level rotation analysis fills that gap, and matters most when a sector's aggregate RS is near flat or ambiguous — that is exactly the condition under which real leadership changes are most likely to be hidden inside the average.
Key Takeaways
- A sector return is a weighted average, and averages hide dispersion: Two industries of similar weight moving in opposite directions can produce a sector-level return close to the benchmark even though real rotation happened underneath.
- Industry rotation uses the same methodology as sector RS, one level down: The ratio-chart approach and lookback conventions from sector relative strength apply directly to industries — there is no separate formula to learn.
- The GICS hierarchy is the map you need first: Knowing which industries and sub-industries sit inside which sector, and at roughly what weight, is a prerequisite covered in GICS Sector Taxonomy and How to Use It.
- Divergence is most informative when the sector level looks unremarkable: A sector already showing a large, unambiguous move doesn't need an industry-level check as urgently as one sitting near flat, where offsetting moves are the most plausible explanation.
- Clean industry-level proxies are scarcer than sector ETFs: Many GICS industries lack a dedicated ETF, so industry-level analysis often requires a custom-weighted basket or an index provider's sub-index rather than a single ticker.
- Industry rotation narrows the search; it doesn't replace sector context: The sector still sets the benchmark, factor tilt, and macro backdrop the industry sits inside — see the Sector & Industry Analysis hub for the full research process.
What Is Industry Rotation?
The GICS hierarchy organizes every public company into a sector, an industry group, an industry, and a sub-industry — four tiers of increasing granularity, explained in full in GICS Sector Taxonomy and How to Use It. Sector-level analysis, the subject of Sector Relative Strength and Momentum, treats each of the 11 GICS sectors as a single unit and measures its price performance against a benchmark like the S&P 500.
Industry rotation takes that exact same relative-strength methodology and applies it one tier lower, to the industries (or sub-industries) inside a single sector. Instead of asking "is Technology outperforming the S&P 500?", industry rotation asks "which industries inside Technology are outperforming Technology itself, or the S&P 500 directly?" The math is identical — a ratio of the industry proxy's price to the benchmark's price, tracked over a 6-to-12-month window — only the unit being measured changes.
Why Can a Sector Look Flat While Its Industries Diverge?
A GICS sector's return is a market-cap-weighted average of every industry that sits inside it. Averages, by construction, can obscure the individual components that produced them. If one industry inside a sector rallies sharply while another industry of comparable weight declines by a similar magnitude, the two moves substantially offset each other in the sector-level number — the sector can post a return close to its benchmark even though a real, tradeable rotation occurred among its parts.
This condition is common inside broad, economically diverse sectors. Information Technology, for instance, contains businesses with very different demand drivers: semiconductor companies are tied to capex cycles and AI infrastructure buildout, while enterprise software companies are tied to corporate IT budgets and subscription renewal cycles. These sub-groups do not always move together, and a period where one is in favor while the other is out of favor is exactly the kind of setup a sector-level-only view will misread as "nothing happening in Technology."
The practical implication is a filter, not a replacement: when a sector's relative strength reading is near flat or ambiguous, that is the signal to check the industry level before concluding there is no rotation worth acting on.
How Do You Apply Relative-Strength Methodology at the Industry Level?
The mechanics are the same as sector-level RS, described step by step in Sector Relative Strength and Momentum: divide an industry proxy's price by the benchmark's price for each day in a lookback window, and track whether the resulting ratio is rising or falling. A rising ratio means the industry is gaining ground on the benchmark; a falling ratio means it is losing ground — independent of the industry's absolute price direction.
The one practical difference is proxy availability. Every one of the 11 GICS sectors has a liquid, single-ticker SPDR ETF (XLK, XLF, XLE, and so on), but far fewer of the underlying industries do. Some industries have a reasonably clean single-ticker proxy — SMH or SOXX for semiconductors, IGV for software, XOP for oil and gas exploration and production — while others require building a custom market-cap-weighted basket from a handful of representative constituents, or relying on an index provider's published industry sub-index. The lookback conventions carry over unchanged: 6-to-12 months is standard, often with the most recent month excluded to reduce short-term reversal noise.
Worked Hypothetical Example: Technology Looks Flat, Two Industries Diverge Sharply
The following figures are a clearly labeled hypothetical, constructed to illustrate the mechanism — not a report of actual market returns for any specific period.
| Component | Approx. Weight in Sector | 3-Month Return | Weighted Contribution |
|---|---|---|---|
| Semiconductors industry | 30% | +9.0% | +2.70 pp |
| Software industry | 35% | −5.0% | −1.75 pp |
| Rest of Technology sector | 35% | +1.0% | +0.35 pp |
| Technology sector (weighted total) | 100% | — | +1.30% |
Weighted sector return = (0.30 × 9.0%) + (0.35 × −5.0%) + (0.35 × 1.0%) = 2.70 − 1.75 + 0.35 = +1.30%. Against a hypothetical S&P 500 return of +1.0% over the same three months, Technology's sector-level relative strength is a barely positive +0.3 percentage points — the kind of reading a sector-level-only process would likely wave off as "no meaningful rotation in Technology this quarter."
The industry-level view tells a different story. The semiconductor industry's relative strength versus the same S&P 500 benchmark is +9.0% − 1.0% = +8.0 percentage points of outperformance — a strong RS reading by any standard. The software industry's relative strength is −5.0% − 1.0% = −6.0 percentage points of underperformance, an equally strong reading in the opposite direction. Both of these signals are fully invisible in the sector-level +0.3pp number; they only appear once the sector is decomposed into its constituent industries.
What Are the Limits of Industry Rotation Analysis?
Industry-level analysis trades breadth for depth, and that tradeoff has real costs. Proxy quality is uneven — a single-ticker ETF for an industry is convenient but may itself be concentrated in a handful of names, meaning the "industry" return can be dominated by one or two large constituents rather than reflecting the industry broadly. Custom baskets avoid that problem but require more work to build and maintain, and introduce judgment calls about which constituents to include and how to weight them.
Industry-level samples are also smaller and noisier than sector-level samples by construction — fewer constituents means single-company news (an earnings surprise, an M&A announcement, a regulatory action) can move the whole industry reading in a way that would barely register at the sector level. This makes industry-level RS signals more prone to short-term noise, and reinforces the case for the same reversal-dampening lookback conventions (6-12 months, excluding the most recent month) used at the sector level.
Finally, industry rotation is a complement to sector analysis, not a substitute for it. The sector still determines the benchmark, factor exposure, and macro backdrop a given industry operates inside — see Economic Cycle and Sector Rotation and the Sector & Industry Analysis hub for that broader context.
FAQ
What is industry rotation?
Industry rotation is the application of relative-strength and momentum analysis one tier below the GICS sector level, to individual GICS industries or sub-industries. It uses the same methodology as sector-level relative strength — comparing an industry's price performance to a benchmark over a trailing window — but applies it to a narrower, more homogeneous group of businesses inside a single sector.
Why can a sector look flat while its industries are diverging sharply?
A GICS sector's aggregate return is a market-cap-weighted blend of every industry inside it. If one industry rallies while another of similar weight declines, the two moves can largely offset in the sector-level average, producing a return close to the benchmark even though real leadership rotation happened underneath. The sector-level number is a summary statistic, and summary statistics can hide dispersion.
How do you apply relative-strength methodology at the industry level?
The same ratio-chart approach used for sectors works for industries: divide an industry proxy's price (an industry ETF, a custom-weighted basket, or an index provider's industry sub-index) by the benchmark's price over a trailing window, then track whether the ratio is rising or falling. The lookback conventions are identical to sector-level RS — 6 to 12 months is standard, often with the most recent month excluded to reduce short-term reversal noise.
What are the limitations of industry-level rotation analysis?
Clean, liquid industry-level proxies are less available than sector ETFs — many GICS industries have no dedicated ETF, forcing analysts to build a custom-weighted basket from index constituents or rely on a data vendor's industry sub-index. Industry-level samples are also smaller, so a single large constituent can distort the industry's apparent return. Industry rotation should supplement sector-level analysis, not replace it, since the sector view still determines the overall benchmark and factor context an industry sits inside.
What should industry rotation be used alongside?
Industry rotation works best paired with the GICS taxonomy (to know which industries sit inside which sector and at what approximate weight) and with sector-level relative strength (to know whether the broader sector itself is attracting or losing capital). Industry rotation narrows the search within a sector; it does not replace the sector-level and macro context that determines whether the sector is a reasonable place to be looking in the first place.
Related Guides
- GICS Sector Taxonomy and How to Use It The four-tier GICS hierarchy — sector, industry group, industry, sub-industry — and how stocks are classified into it.
- Sector Relative Strength and Momentum The ratio-chart relative-strength methodology this page applies one tier down, at the industry level.
- Leading, Weakening, Lagging, Improving A four-quadrant classification framework for describing how a sector's or industry's relative strength and momentum are evolving.
Sources
Disclaimer
This article is for educational and informational purposes only and does not constitute personalized investment, financial, or legal advice. The worked example on this page uses hypothetical, illustrative figures and does not represent actual returns for any sector, industry, or ETF. Past performance and relative-strength patterns do not guarantee future results. Trading involves risk, including the possible loss of principal.