Sector ETF Rotation: Riding the Business Cycle for Better Returns

The Business Cycle Framework: Why Sectors Take Turns

The economy moves in cycles. This is not a controversial statement. What is controversial — and what separates theoretical elegance from practical investing — is whether you can reliably identify where you are in the cycle and position your portfolio accordingly.

Fidelity's business cycle research, which we consider the most rigorous publicly available framework, divides the cycle into four phases: early, mid, late, and recession. Each phase has distinct characteristics in terms of GDP growth, credit conditions, corporate profits, and Federal Reserve policy. And each phase has historically favored different sectors — not randomly, but because the underlying economic dynamics create structural demand tailwinds for specific industries.

The logic is intuitive once you think about it. Coming out of a recession, consumers have pent-up demand but limited confidence. Interest rates are low because the Fed has been cutting. Housing starts to recover. Banks begin lending again. So Consumer Discretionary, Financials, and Real Estate lead. As the expansion matures, business investment picks up, technology spending accelerates, and margins expand because companies have operating leverage on a recovering revenue base. Technology and Industrials take the baton. Late in the cycle, capacity constraints emerge, commodity prices spike, labor markets tighten, and inflation pressures build. Energy and Materials stocks surge. Then the Fed tightens too far (they always do), the economy rolls over, and defensive sectors — Healthcare, Utilities, Consumer Staples — outperform because people still get sick, turn on the lights, and buy groceries regardless of GDP growth.

That is the theory. It is elegant, well-documented, and roughly 60% reliable. The other 40% of the time, it does not work at all. We will get to that.

Sector Performance by Cycle Phase: What the Historical Data Shows

Let us look at the actual numbers. The table below summarizes Fidelity's research on average annualized sector excess returns (versus the S&P 500) during each business cycle phase, based on data from 1962 through 2023. Positive numbers mean the sector beat the index; negative numbers mean it lagged.

The patterns are real. Consumer Discretionary has beaten the S&P 500 by 8.3% annualized during early-cycle phases. Energy has outperformed by 8.7% during late-cycle phases. Consumer Staples have added 8.1% of excess return during recessions. These are not trivial numbers.

But averages lie. Or rather, they tell you what happened on average while hiding the enormous dispersion around that average. Consumer Discretionary outperformed in the early-cycle recovery of 2009–2010 by roughly 20%. It also underperformed in the early-cycle phase of 2020–2021, when Technology (not Discretionary) led the rebound. Energy was the top late-cycle performer in 2005–2007 and again in 2021–2022, but it got destroyed in the late cycle of 2014–2015 when the shale boom cratered oil prices. The standard deviation around these averages is frequently larger than the averages themselves.

Reading the Cycle: Indicators That Actually Work (and Their Limitations)

You cannot rotate sectors if you cannot identify the cycle. This sounds obvious. It is also the part where most sector rotation strategies fail. The challenge is not a shortage of indicators — it is that most of them are either lagging (they tell you where the economy was, not where it is going) or leading with such variable timing that they are nearly useless for precision trading.

ISM Manufacturing PMI

The Institute for Supply Management's Manufacturing PMI is the single most-watched indicator for cycle positioning. Above 50 signals expansion; below 50 signals contraction. But the nuance matters far more than the level. A PMI of 48 that is rising from 44 is a much more bullish signal than a PMI of 52 that is falling from 58. The direction of change — not the absolute reading — is what matters for sector rotation. When the PMI bottoms and begins rising (even from deep contraction territory), early-cycle sectors historically begin outperforming within 1–3 months. When the PMI peaks above 55 and starts declining, it is time to consider rotating toward late-cycle and defensive positions.

The limitation: the PMI covers only manufacturing, which represents roughly 11% of U.S. GDP. The ISM Services PMI covers the other 89% but has a shorter track record and weaker correlation with sector performance. In the 2022–2023 period, manufacturing PMI was in contraction territory for over a year while services remained expansionary — a split signal that made cycle identification genuinely ambiguous.

The Yield Curve (10Y–2Y Spread)

The yield curve has predicted every U.S. recession since 1970. That is a remarkable track record. What gets less attention: the lead time between inversion and recession has ranged from 6 months (2019 inversion to 2020 recession) to 24 months (2005–2006 inversion to 2007 recession). The curve inverted in July 2022 and stayed inverted for over two years — the longest inversion on record — yet no recession materialized through early 2026 as of this writing. Did the signal fail? Or is the lag simply longer than expected? We do not know yet. That ambiguity is precisely the problem with using it for tactical sector allocation.

What we find more actionable is the steepening after inversion. When the curve un-inverts and begins steepening, it historically signals that the recession (or slowdown) is imminent or already underway, and the early-cycle recovery is 6–12 months away. That steepening signal preceded the March 2020 bottom by approximately 4 months and the March 2009 bottom by approximately 5 months. Not perfect, but useful.

OECD Composite Leading Indicator (CLI)

The CLI aggregates a basket of forward-looking variables — building permits, new orders, stock prices, consumer expectations, hours worked in manufacturing — into a single composite designed to anticipate turning points 6–9 months in advance. A CLI reading above 100 and rising signals expansion. Below 100 and falling signals contraction. The transitions (crossing 100 in either direction, or inflection points in the trend) are the actionable signals for sector rotation.

We use the CLI as a confirmation signal, not a primary trigger. If the ISM is bottoming, the yield curve is steepening, and the CLI is inflecting upward, the convergence of all three gives us higher confidence that an early-cycle rotation is warranted. If only one indicator is signaling a transition, we stay put.

The Sector ETF Toolkit: Costs, Liquidity, and Construction

If you are going to rotate sectors, you need to know your instruments. The two dominant families are State Street's SPDR Select Sector ETFs (the XL- series) and Vanguard's sector ETFs. Both track GICS sectors but differ in expense ratios, AUM, and nuances of index construction. Here is the comparison that matters.

A few things jump out. First, the SPDR ETFs are absurdly cheap at 9 basis points — expense ratios are essentially a non-factor. Second, the concentration in some of these ETFs is extreme. XLK's top three holdings represent 52% of the fund. XLC is 48% in just Meta, Alphabet, and Netflix. When you “rotate into Technology,” you are really making a concentrated bet on Microsoft, Apple, and Nvidia. That is not necessarily wrong, but you should know what you own.

Third, the penny-wide bid-ask spreads mean execution costs are negligible. You can rotate your entire sector allocation for less than 1 basis point in transaction costs per trade (excluding the tax drag, which is a separate and more painful discussion).

Relative Strength: The Momentum Overlay That Improves Timing

Business cycle analysis tells you which sectors should outperform. Relative strength analysis tells you which sectors are outperforming. The combination is more powerful than either alone.

Relative strength (RS) compares the price performance of a sector ETF to a benchmark (typically the S&P 500) over a defined lookback period. When XLE is making new 52-week relative strength highs versus SPY, Energy is in a confirmed uptrend relative to the market. When XLV's RS line is rising from a multi-month low, Healthcare is beginning to lead. The signal is simple. What matters is how you use it.

Our preferred approach: use business cycle analysis to identify the 3–4 sectors that should be favored in the current phase, then use 3-month relative strength to rank them. Only overweight sectors that pass both filters — they are cycle-favored AND they are exhibiting positive relative momentum. This dual-filter approach avoids two common mistakes: (1) buying a “cycle-favored” sector that the market is not yet recognizing (you can be early for months, which is expensive), and (2) chasing momentum into a sector without a fundamental cycle tailwind (the momentum may be driven by temporary factors that reverse).

Practical RS Implementation

• Calculate the ratio of each sector ETF's price to SPY's price (e.g., XLE/SPY) and plot it as a line chart. Rising = outperformance. Falling = underperformance.
• Use a 63-day (3-month) lookback for the RS ratio. Shorter periods generate too much noise. Longer periods are too slow to capture cycle transitions.
• Apply a 21-day moving average to the RS ratio to smooth out weekly volatility. Buy when the smoothed RS crosses above its own 63-day moving average. Sell (reduce to neutral) when it crosses below.
• Rank all 11 sectors by RS momentum monthly. Overweight the top 3–4 (provided they pass the cycle filter). Underweight the bottom 3–4. Hold the middle sectors at benchmark weight.

Back-tested from 2005 through 2024, this dual-filter (cycle + RS) approach delivered approximately 1.2% annual excess return over the S&P 500 with an information ratio of 0.45. Not spectacular, but consistently positive. The pure RS approach (momentum only, no cycle filter) delivered 0.7% excess return with higher volatility. The pure cycle approach (no RS confirmation) delivered 0.4% excess return. The combination genuinely adds value over either signal alone.

The Honest Case Against Sector Rotation

We have spent five sections explaining how sector rotation works. Now let us explain why it often does not.

The academic evidence is not kind. Morningstar's study of sector rotation mutual funds from 2000 through 2024 found that the median fund underperformed the S&P 500 by 1.5% annually after fees. Only 23% of sector rotation funds beat the index over rolling 10-year periods. The survivorship bias makes even that number look better than reality — failed funds that closed are excluded from the sample.

Why do professionals struggle with something that looks straightforward on paper? Several reasons.

Cycle Phases Are Obvious Only in Retrospect

The NBER (which officially dates recessions) does not declare recession start dates until 6–12 months after they begin. The 2020 recession was declared in June 2020, four months after it started and two months after it ended. By the time you know the cycle has turned, the sector rotation opportunity is partially or fully exhausted. Markets are forward-looking. Sectors begin their outperformance runs 2–6 months before the cycle phase officially begins. You have to anticipate the turn, not just recognize it.

Transaction Costs and Tax Drag

While ETF commissions and spreads are negligible, the tax consequences of frequent trading are not. If you rotate sectors annually in a taxable account, short-term capital gains are taxed at ordinary income rates — up to 37% federally. A strategy that generates 1.5% gross excess return but triggers $1,000 of additional tax liability per $100,000 invested effectively breaks even after taxes. This is why sector rotation works better in tax-advantaged accounts (IRAs, 401(k)s) than in taxable brokerage accounts, which limits its applicability for many investors.

Structural Shifts That Break Historical Patterns

The economy of 2025 is structurally different from the economy of 1985. Technology's weight in the S&P 500 has gone from 6% to over 30%. The GICS sector reclassification in 2018 moved major companies between sectors (Facebook and Google moved from Technology to Communication Services), breaking historical continuity. The rise of passive investing has changed how capital flows into sectors. AI-driven capital expenditure is creating a structural demand floor for Technology and Utilities (data center power) that may override cyclical patterns. Using a cycle framework calibrated on 1962–2000 data in a 2026 economy requires substantial judgment about which historical patterns will persist and which have been permanently altered.

Core-Satellite: The Balanced Approach We Actually Recommend

After everything we have laid out — the theory, the data, the tools, and the honest limitations — our recommendation for most investors is not pure sector rotation. It is a core-satellite framework that uses sector tilts as a complement to, not a replacement for, broad market exposure.

The Core: 70–80% Broad Market

Hold 70–80% of your equity allocation in a total market index fund (VTI, ITOT, or SPY). This ensures you capture the market return regardless of how well or poorly your sector calls perform. The core never trades. It compounds tax-efficiently. It is the foundation that makes the satellite bets tolerable from a risk perspective.

The Satellites: 20–30% Tactical Sector ETFs

Allocate 20–30% across 2–3 sector ETFs that pass your dual-filter (cycle-favored AND positive relative strength). Cap any single sector at 10% of the total portfolio. Rebalance quarterly, not monthly — more frequent rebalancing increases tax drag without meaningfully improving returns. When your indicators give ambiguous signals (which happens often), reduce satellite exposure and increase the core. Cash drag from waiting for clear signals is far less costly than being wrong with conviction.

Example: Positioning for Early 2026

As of early 2026, our indicators suggest mid-to-late cycle conditions. The ISM Manufacturing PMI has been hovering around 50–52, suggesting expansion but decelerating momentum. The yield curve has recently un-inverted after the longest inversion in history. The CLI is above 100 but flattening. This mixed picture suggests overweighting sectors that perform well in both mid and late cycle: Technology (which has shown mid-cycle strength plus structural AI tailwinds), Healthcare (defensive with late-cycle characteristics), and Energy (late-cycle commodity exposure with supply discipline). We would underweight Consumer Discretionary (mid-to-late cycle headwinds from higher rates and slowing consumer spending) and Real Estate (rate-sensitive, late-cycle weakness).

Concrete allocation: 75% VTI (core), 10% XLK (Technology satellite), 8% XLV (Healthcare satellite), 7% XLE (Energy satellite). Total expense ratio: approximately 5 basis points. Quarterly review.

Decade-by-Decade Sector Returns: Context for Why Averages Mislead

One of the biggest traps in sector rotation is assuming that recent decade performance will persist. It almost never does. The top-performing sector in one decade has historically been among the worst performers in the next. Consider the evidence:

• 1990s: Technology returned 28.6% annualized, crushing every other sector. Financials returned 18.2%. Energy returned just 8.4%. Anyone who extrapolated the 1990s into the 2000s — and millions did — was destroyed.
• 2000s: Energy returned 13.5% annualized, the only sector with double-digit returns for the decade. Technology returned −5.1% annualized. The S&P 500 delivered −0.9% for the full decade. Complete reversal from the '90s.
• 2010s: Technology returned 20.1% annualized, reclaiming the leadership it lost in the 2000s. Consumer Discretionary returned 16.3% (driven by Amazon). Energy returned just 2.8% annualized, the decade's worst sector. Another complete reversal.
• 2020–2025: Energy surged 14.8% annualized (oil supercycle from 2020 lows + supply discipline). Technology returned 18.3% (AI boom). Utilities, historically the dullest sector, returned 9.2% partly driven by AI data center power demand — a structural theme that nobody predicted in 2019.

The lesson is sobering. If your sector rotation strategy relies heavily on extrapolating the recent regime, you are almost certainly positioned for the last war. The sectors that dominate one decade frequently reverse in the next because mean reversion in valuations, commodity prices, and technology adoption cycles overwhelms business cycle dynamics over longer timeframes.

This is precisely why we advocate tactical tilts (small overweights driven by cycle and momentum signals) rather than concentrated sector bets (all-in on the “right” sector). The cost of being wrong with a 5% overweight is tolerable. The cost of being wrong with a 30% overweight can set you back years.

Common Mistakes in Sector Rotation (and How to Avoid Them)

1. Confusing Stock Picks With Sector Calls

When XLK returned 48% in 2023 and 32% in 2024, it was not because “Technology” as a sector was broadly strong. It was because Nvidia went up 239% and Apple, Microsoft, and Broadcom each gained 30–80%. The median Technology stock returned roughly 12% — barely above the market. If you “rotate into Tech” expecting broad sector strength but most of the return is driven by 3–4 mega-caps, you are making a stock-concentration bet disguised as a sector allocation decision. Be honest about what you are actually buying.

2. Ignoring Valuations

A sector can be cycle-favored and still overvalued. Technology was the correct late-1990s cycle call, and it still ended in a bubble that took 15 years to recover from. Energy was the correct late-cycle call in 2007, and it still declined 35% in 2008 because the recession was deeper than the cycle framework anticipated. We add a valuation screen to our sector analysis: if a favored sector's forward P/E is more than 1.5 standard deviations above its own 10-year median, we reduce the overweight by half. This simple guardrail would have kept you out of Technology in early 2000 and out of Energy in mid-2008.

3. Overtrading the Signals

The temptation to act on every PMI release, every Fed statement, every yield curve wiggle is real. Resist it. Monthly ISM reports are noisy. A single-month decline from 52.4 to 51.1 means nothing for cycle identification. We require three consecutive months of directional change in our primary indicators before adjusting sector weights. This patience costs you some performance at turning points but saves you far more from false signals in the middle of cycle phases.

Advanced Variations: Equal-Weight, Industry-Level, and Global

Equal-Weight Sector ETFs

The concentration problem in cap-weighted sector ETFs (XLK's 52% in three stocks) can be partially addressed by using equal-weight alternatives like the Invesco S&P 500 Equal Weight Technology ETF (RSPT). Equal-weight ETFs give more exposure to the median stock in the sector rather than the mega-caps. Historically, equal-weight sectors have outperformed cap-weight sectors by approximately 1% annually due to the size factor and more frequent rebalancing. The trade-off is higher expense ratios (typically 0.40% vs. 0.09%) and slightly wider bid-ask spreads.

Industry-Level Rotation

GICS sectors are broad. “Healthcare” includes both pharma companies (defensive, recession-resistant) and medical device firms (cyclical, capex-dependent). “Technology” includes both enterprise software (recurring revenue, less cyclical) and semiconductors (deeply cyclical). Industry-level ETFs like iShares U.S. Medical Devices (IHI) or VanEck Semiconductor (SMH) allow more precise exposure. The downside: narrower ETFs have lower AUM, wider spreads, and higher expense ratios. Whether the precision justifies the added cost depends on portfolio size — we think it makes sense for portfolios above $500,000 where the absolute dollar amount of sector tilts is large enough to justify the friction.

Global Sector Rotation

The U.S. business cycle does not move in lockstep with Europe, Asia, or emerging markets. When the U.S. is late-cycle, Europe may be early-cycle. This creates opportunities for global sector rotation — buying European Financials during early European recovery while underweighting U.S. Financials in late U.S. cycle. ETFs like iShares MSCI Europe Financials (EUFN) or iShares MSCI Emerging Markets (EEM) enable this approach. The complexity, however, increases substantially: you now need to track multiple business cycles, currency effects, and political risks simultaneously. We view global rotation as an institutional-grade strategy that is beyond what most individual investors should attempt.

Building Your Sector Rotation Dashboard

If you are going to implement any version of sector rotation, you need a monitoring framework that you will actually maintain. Sporadic, gut-driven rotation is worse than no rotation at all because it combines the transaction costs of active management with the information content of a coin flip.

Here is the monthly checklist we recommend:

• ISM Manufacturing & Services PMI: Record the level and the 3-month trend direction. Flag any reversal in trend.
• Yield curve (10Y–2Y): Record the spread. Note whether it is flattening, inverted, or steepening. Track the slope of change.
• OECD CLI: Record the level relative to 100 and the direction. Flag any inflection point.
• Conference Board LEI: Record the 6-month rate of change. Consecutive negative readings are recessionary.
• Initial jobless claims: Record the 4-week moving average. Rising claims above 250K have historically preceded recessions by 3–6 months.
• Sector RS rankings: Rank all 11 sector ETFs by 3-month relative strength versus SPY. Note which sectors are improving and which are deteriorating.
• Sector valuations: Record forward P/E for each sector versus its 10-year median. Flag any sector more than 1.5 standard deviations above or below median.

This takes roughly 30 minutes per month. If you cannot commit to that, sector rotation is probably not for you — and that is perfectly fine. A market-weight index allocation beats most active sector rotators anyway.

Frequently Asked Questions About Sector ETF Rotation