Week 16: Sector Rotation — Which Slice of the Market Wins in Each Phase

---

1. Why This Is Important

The S&P 500 is not one thing. It is eleven things stitched together under a single ticker. In any given calendar year, the spread between the best-performing GICS sector and the worst can run 60-90 percentage points. In 2022, energy returned +64% while communication services returned -38% — a 100-point spread inside a single index. The "market" returned -18%. None of the eleven sectors actually did that.

If you understand which slice of the market wins in each phase of the business cycle, you have the rough outline of one of the durable alpha sources: sector rotation. Capital cycles between sectors on predictable macro and rate shifts. Ride the rotation and the cycle pays you. Fight it and you spend a decade underperforming a passive S&P investor who never opened the app.

Four reasons this matters even if you decide not to rotate.

It is a real alpha source. Unlike stock picking — where the

academic evidence is ugly — sector tilts have empirical backing. Defensive sectors outperform in late-cycle and recession regimes; cyclicals outperform in early- and mid-cycle. The pattern is messy, the timing is hard, but the structural story has played out repeatedly since the data started in the 1960s.

It tells you what the market is thinking. When utilities

and staples lead for three months while financials and discretionary lag, the market is voting on a slowdown — and doing it before any economist has confirmed one. Reading sector leadership is reading the macro bid in real time.

It contextualises your equity drawdown. When your tech

stocks fall 30% and energy is up 50%, the issue is not "tech is broken." It is that the discount rate for long-duration cash flows just doubled. Knowing the cycle prevents you from selling the wrong sector at the wrong time.

Most retail rotation loses. Alpha is the rare gap, not the

default — and the data on retail sector rotators is consistently worse than buy-and-hold the index. Knowing the framework is necessary to even attempt it, and even then most attempts fail. Decide honestly whether you are running rotation as a hobby or as a real edge.

This lesson covers the eleven GICS sectors and their SPDR ETFs, the canonical cycle map, why the map breaks (2020 flipped the script in four weeks), the momentum-vs-mean-reversion question at the sector level, and the brutal honest reality of retail rotation P&L.

---

2. What You Need to Know

2.1 The Eleven GICS Sectors and Their SPDR ETFs

The Global Industry Classification Standard (GICS), maintained jointly by S&P and MSCI, divides every listed company into one of eleven sectors. State Street's SPDR family runs an ETF for each, all with expense ratios under 0.10% and combined assets above $200B. They are the canonical retail vehicles for sector exposure.

Sector	ETF	What's in it
Technology	XLK	Microsoft, Apple, NVIDIA, Broadcom, semiconductors, software
Financials	XLF	JPMorgan, Berkshire, banks, insurers, exchanges
Health Care	XLV	UnitedHealth, Eli Lilly, J&J, drugmakers, devices, hospitals
Consumer Discretionary	XLY	Amazon, Tesla, Home Depot, autos, hotels, restaurants, retail
Communication Services	XLC	Meta, Alphabet, Netflix, telecoms, media (split out from tech in 2018)
Industrials	XLI	GE, Caterpillar, Boeing, defense, rails, airlines, machinery
Consumer Staples	XLP	P&G, Costco, Walmart, Coke, Pepsi, household goods, tobacco
Energy	XLE	ExxonMobil, Chevron, oil and gas E&P, services, refiners
Utilities	XLU	NextEra, Duke, electric and water utilities, regulated monopolies
Real Estate	XLRE	American Tower, Prologis, REITs (split out from financials in 2016)
Materials	XLB	Linde, Sherwin-Williams, chemicals, metals, paper, packaging

Two notes on the structure.

First, XLRE and XLC are recent. Real estate was carved out of financials in September 2016; communication services was carved out of technology and discretionary in September 2018. Most of the long historical data we have is therefore on nine sectors, not eleven. For the 2010-2024 backtests we run later, treat XLRE and XLC as "data starts mid-decade" and don't over-fit to their short history.

Second, GICS classifications are not always intuitive. Amazon is discretionary, not tech. Tesla is discretionary, not industrials. Visa and Mastercard are tech, not financials. Walmart is staples, Costco is staples, but Target is discretionary. The mental model "tech = software + chips" is tighter than the actual XLK basket; the mental model "financials = banks + insurers + exchanges" is roughly right after XLRE was carved out.

2.2 The Canonical Cycle Map

The folk wisdom — repeated in every CFA textbook and most strategist notes — is that sectors lead and lag in a predictable order through the four phases of the business cycle. The map looks like this:

The story behind the map.

• Early recovery — rates have been cut hard, the yield curve is

steep, credit is loosening, consumer pent-up demand starts spending. Financials earn on the steep curve and on falling loan losses. Consumer discretionary rides the wage and confidence rebound. Industrials ride the inventory restock.

• Mid-expansion — growth is broad, capex turns on, business

software and infrastructure spending accelerates. Technology and communication services lead as long-duration cash flows re-rate higher under still-low rates. Industrials keep working.

• Late expansion — capacity is tight, inflation lifts, the Fed

is hiking. Energy and materials lead because their input is the inflation. Health care starts working because investors begin reaching for quality and stable demand.

• Recession — earnings are falling, rates are being cut, fear is

the dominant emotion. Consumer staples, utilities, and health care lead — the products people need rather than want. Long Treasuries, gold, and cash are the cross-asset winners.

The map is real. It is also not nearly as clean as the textbook draws it, which is the next subsection.

2.3 The Yearly Heatmap — The Map Versus Reality

The chart below is the actual 2010-2024 record: each year's annual total return for each SPDR sector ETF, coloured by decile within that year. A green cell means the sector was in that year's top decile; a red cell means it was in the bottom decile.

Read the chart with three questions.

First, does the canonical map hold? Pick any clean cycle phase on the chart and check. 2010-2011 was textbook early recovery, and financials underperformed badly (XLF -17% in 2011) because the European debt crisis hit US banks hard. That is not in the cycle map. 2018-2019 was textbook late-cycle, and energy was the worst-performing sector both years. Also not in the map.

Second, what years does the map fit? 2017 — mid-expansion — tech leads, materials and industrials work, staples lag. Clean. 2022 — inflation shock that rhymes with late-cycle — energy crushes everything, materials and staples hold up, tech and discretionary collapse. Also clean.

Third, what years break it? 2020 broke it in four weeks. The COVID recession started in February 2020 with utilities and staples leading, exactly as the map says. By April, with the Fed at zero and Congress having passed $5T of fiscal stimulus, technology had flipped to the year's leadership and energy was finishing -33%. The recession was officially over in April. The "recession sector" playbook had three months of validity in a downturn that re-rated in weeks rather than years.

The honest read: **the cycle map is a statistical tendency, not a calendar.** It works in the average year of the average cycle. Individual years are dominated by individual catalysts — the 2008 financial-system crisis, the 2014-2016 oil collapse, the 2020 COVID shock, the 2022 rate shock, the 2023 AI mania — and these catalysts override the cycle whenever they show up.

2.4 Momentum vs Mean Reversion at the Sector Level

This is the momentum-vs-mean-reversion question, scaled down from the whole market to the eleven slices. Two playbooks both have empirical support:

• Momentum. Buy the sector that has been working. Hold for 3-12

months. Sell when it stops working. The trend-following literature shows positive risk-adjusted returns at the sector level over decades — the same way it does at the asset-class level.

• Mean reversion. Buy the sector everyone hated last year, sell

the one everyone loved. At horizons longer than a year, sector rankings revert. Energy in 2020 (worst sector, then best in 2021). Tech in 2022 (worst, then best in 2023). The cycle of leadership.

Both are true. The hard part — exactly as in #8 — is recognising which regime is active. Some markets reward momentum; others reward mean reversion. Most of the academic evidence points to **3-12 month momentum and 3-5 year mean reversion** as the rough horizons that have separated. Inside that band, the strategies cancel.

Practically, the retail investor who buys "the sector that just had a great year" is usually expressing mean-reversion-as-mistake — they are buying at the top of a 12-month momentum run that is about to reverse. The investor who buys "the sector that just had its worst year in a decade" is more often right than wrong, if they can hold through the additional drawdown that frequently comes before the revival. Either playbook needs the discipline to size for survival and a clear stop on what proves the thesis wrong.

2.5 The Brutal P&L of Retail Rotation

Here is the part the textbooks leave out. The interactive lab below this lesson lets you pick up to four sectors and plot their 2010-2024 cumulative return against the S&P 500 with a single toggle. Run a few combinations:

• All-tech rotation (XLK only). Beats SPY handily over the full

window. End wealth roughly 2.0x the index. Looks like genius. Until you check the year-by-year, where 2022 cost you -28% while the index lost -18%, and most retail attention-driven rotation trades happen after the +50% year, not before it.

• Buy the previous year's winner, every year. Trail the

S&P 500 by 4-6 percentage points annually over 2010-2024. The textbook momentum trade, executed naively, loses to passive.

• Buy the previous year's loser, every year. Roughly matches

the S&P, with much higher volatility. The textbook mean-reversion trade, executed naively, ties.

• Equal-weight all eleven. Slightly lags SPY (which is

cap-weighted) over the window. Works in years where the mega-caps are not running away (2014-2016, 2022); lags in years when they are (2020-2021, 2023-2024).

The decade's retail rotators have lost not because rotation is wrong, but because the version they ran was: chase last quarter's winner, sell on the first 10% drawdown, repeat. That is not rotation. That is performance-chasing in fancy clothes.

The professional version of rotation — overlay on top of a passive core, tilt by a few percentage points based on the cycle and positioning data, never bet the portfolio on a single sector — is where the alpha lives. The retail version, run as a substitute for the passive core, almost always underperforms.

The honest framing: **most retail readers are better off with the S&P 500 and a 10-15% sector tilt at most.** That tilt can give you the win-some flavour of the cycle without the structural risk of betting the portfolio on next year's leadership map.

2.6 The Practical Toolkit

Three ways retail investors actually use sector ETFs:

Tilt, don't bet. Overweight one or two sectors by 5-10% of

the portfolio relative to S&P weights. If the tilt is wrong by 5 percentage points of return, the portfolio impact is 25-50bps — survivable. If the tilt is right, you get a modest premium over the cycle. This is the only rotation discipline that survives most retail temperaments.

Defensive flip on hard signals. The yield curve invert, the

PMI drop below 50, jobless claims trend up — these are late-cycle signals. Some investors run the rotation only at these triggers: tilt to staples, utilities, health care, and reduce equity beta. Hold the defensive tilt until the curve un-inverts and the Fed cuts. This is Horace's preferred shape: asymmetric risk-on / risk-off rather than continuous rotation.

Alpha tranches in conviction sectors. Within a

sector you understand, you can own the seniors (the cap-weighted ETF), the juniors (mid-cap names), and the explorers (small-cap options or single names). When the sector cycles into favour, the tranches re-rate in order, and the explorers carry the biggest payoff. Energy 2020-2022 was the textbook example — XLE up 1.6x, mid-cap E&Ps up 3-5x, distressed shale survivors up 10-30x. This is the playbook for sector conviction, not a recipe for blanket rotation.

The interactive lab below lets you mix the sectors and see how each combination behaves against SPY. Run a few. The discipline the chart will teach you, faster than any textbook, is that beating SPY by tilt is harder than the cycle map makes it look.

---

3. Common Misconceptions

"The cycle map is a calendar." It is a statistical tendency.

In any individual year, sector returns are dominated by the year's specific catalyst (oil shock, banking crisis, COVID, AI mania). The cycle is one variable; it is rarely the dominant one.

"Tech is always the winner." Tech led 2013-2021 and 2023-2024

and was the worst sector in 2022, the worst sector in 2008, and one of the worst over 2000-2002. The "always" is a recency bias from a decade of falling rates.

"Defensives never grow." Healthcare returned 200%+ over

2010-2019 with materially lower volatility than the S&P. "Defensive" means lower beta, not zero growth. The compounding is real, just smoother.

"Energy is dead." Energy was the worst sector for most of

2014-2020 and the best sector in 2021 and 2022. ESG mandates that dumped it on the way down had to buy it back at higher prices on the way up. Buying what passive flows have abandoned was the trade.

"I'll rotate based on the headlines." Headlines lag price by

3-6 months. By the time CNBC is running "the recession is here" on the chyron, defensives have already led for two quarters. Rotation works only if you front-run the consensus, which is most retail traders' weakness, not strength.

"All sectors revert to the mean." Energy didn't from 2014

to 2020 — it kept being the worst sector for six straight years. Mean reversion is a long-horizon tendency that can take a decade to play out, and "right but early" is operationally indistinguishable from "wrong" — the market can stay irrational longer than you can stay solvent.

"Sector ETFs are diversified." XLK is 45% Microsoft + Apple +

NVIDIA. XLC is 40% Meta + Alphabet. XLY is 25% Amazon + Tesla. The "sector" is often a leveraged bet on three names. The S&P 500 is more diversified than most of its sector slices.

"Sector rotation is the same as factor investing." It isn't.

Factors (value, momentum, quality, low-vol) cut across sectors. You can be long the value factor and own all eleven sectors. Sector rotation is one specific tilt; factor investing is a different framework that often produces opposite sector bets from the cycle map.

---

4. Q&A Section

Q: What's the simplest sector tilt I can run on top of an S&P 500 core? A: Hold 80-90% in SPY (or VOO), and use the remaining 10-20% to overweight one or two sectors by conviction. If you think we're late-cycle, that 10-20% goes to XLV and XLP. If early-cycle, XLF and XLY. Rebalance annually. The portfolio risk impact stays modest, and you participate in the rotation at the margin.

Q: How do I tell which phase of the cycle we're in? A: Yield curve slope, PMI level, jobless claims trend, and Fed policy direction together give you a 70%+ confident read. We cover the indicators in detail in Week 10. Roughly: yield curve inverted + PMI below 50 + claims rising + Fed cutting = late-cycle into recession. Curve steep + PMI rising + claims falling + Fed cutting hard = early recovery. Most signals never give you the textbook configuration; you go with the majority of indicators.

Q: Can I just buy whatever sector worked last year? A: That is the naive momentum trade and over 2010-2024 it has slightly underperformed the S&P. The reason is selection bias — the previous year's winner is often the one most likely to mean-revert in the next year. The 12-month momentum signal that academics document works at 1-3 month rebalancing horizons, not annual.

Q: What about leveraged sector ETFs? A: Avoid. Daily-rebalanced 2x and 3x ETFs decay over time due to volatility drag, often losing money even when the underlying sector is roughly flat. They are intraday speculation tools, not investments.

Q: Is energy a buy in 2026? A: This course doesn't pick names. The framework: energy has underperformed since the 2022 peak, oil is range-bound, ESG selling has eased, and the S&P 500 weight is below 4%. By the principle of buying what passive flows have abandoned, the structural setup is more attractive than the ten years from 2010 to 2020. Whether that translates to next year's return depends on demand, geopolitics, and the dollar.

Q: How does sector rotation interact with the four-tranche framework? A: Once you have a sector conviction, the tranches tell you how to express it. ETF for the senior exposure. Mid-cap names you've researched for the junior leg. Long-dated calls on small-cap names or pre-revenue specs for the explorer leg, sized so the loss is survivable. The cycle gives you the sector; the tranches give you the expression.

**Q: What's wrong with rotating into staples and utilities for the recession?** A: Two things. One, by the time the recession is the consensus narrative, those sectors have already had their move; you're late. Two, in a deflationary recession (1929-1932, 2008-2009) even staples and utilities sell off; they only outperform on a relative basis. The defensives playbook works best in shallow recessions with quick Fed responses (2001, 2020). It can hurt in long bear markets where everything goes down together.

Q: Why did the 2020 cycle break the map? A: COVID was a four-week recession from a markets perspective. The Fed cut to zero in 9 days, Congress passed $5T of stimulus in 6 weeks, and the discount rate on long-duration cash flows collapsed. Tech (long-duration) re-rated immediately; energy (short-duration commodity demand) collapsed because nobody was driving. The map assumes a multi-quarter transition. 2020 didn't give one. Future shocks may also not give one.

Q: How do US sector classifications differ from international? A: GICS is global. The same eleven sectors exist in Europe and Asia. Sector weights differ — the US is roughly 30% tech, Europe is 5%; Europe is heavier in financials and industrials. But the framework transports. We restrict to US-only across this course; ex-US sector tilts are out of scope here.

**Q: Should I use sector ETFs in a tax-advantaged account or taxable?** A: Either works. Sector ETFs are tax-efficient — most are organised as RICs with low turnover. The tradeable advantage is that rotating between sector ETFs in a taxable account triggers capital-gains events; in a 401(k) or IRA, it doesn't. If you plan to actually rotate, do it in the tax-deferred account.

**Q: What's the single highest-conviction sector tilt for someone who follows the structural-flow argument?** A: Horace's bias is toward what passive flows have abandoned. In April 2026 that is a coin-flip among energy (post-2022 peak, ESG-divested), small-cap value (a decade out of favour), and the financial sector (post-2023 regional bank scare). None of these are textbook recommendations; they are positions taken on top of a broad-index core, sized so being wrong for two years doesn't threaten the rest of the portfolio. The core is still the S&P 500.