Sector Rotation: Following the Business Cycle

Four Phases, Reasonably Stable Across Textbooks

The business cycle, for the purposes of rotation, reduces usefully to four phases. The labels vary across textbooks; the substance does not.

1. Early Recovery. The economy emerges from recession. The Fed has been cutting and policy rates are low. Consumer confidence rebuilds from a depressed base. Corporate earnings lift off floors that looked permanent six months earlier. Equity valuations, on most measures, are cheap.
2. Mid-Expansion. Growth accelerates and broadens. Employment rises. Consumer spending compounds. Earnings grow strongly. This phase is ordinarily the longest, and ordinarily the most remunerative for equity holders.
3. Late Expansion. Growth peaks. Inflation rises. The Fed begins raising rates and the yield curve flattens. Labour markets tighten; commodity prices firm. Valuations are high. Warning signs accumulate while optimism remains intact — a characteristic combination, and one that has preceded most of the cyclical reversals worth studying.
4. Contraction. Growth turns negative. Earnings fall. Layoffs follow. The Fed pivots back to cutting. Equities decline, sometimes disorderly. This is the painful phase, and also the phase that lays the ground for the next recovery.

The full cycle, peak-to-peak, has historically run five to ten years, with considerable variation around that range. The 2001 recession was shallow and brief; the 2008-2009 contraction was deep and ran eighteen months. The next one will differ in its own way again, which is rather the point of phase identification being a discipline rather than an algorithm.

The Cyclical / Defensive Sort

Before getting to the sector-by-phase table, the prior taxonomy deserves to be drawn cleanly. The eleven S&P 500 sectors sort, with one or two awkward cases, into two groups. Cyclical sectors have earnings that track the business cycle with some leverage. Defensive sectors have earnings that hold up tolerably regardless of the macro state. Everything else in the rotation framework is built on this distinction.

Cyclical Sectors

• Technology (XLK). Enterprise and consumer technology budgets expand in good times and contract in bad. Beta is high by construction: the sector outperforms in bull markets and underperforms in bear markets, and the magnitude of that asymmetry is substantial.
• Consumer Discretionary (XLY). Non-essential spending — restaurants, retail, travel, entertainment. Booms when households feel wealthy and falls back when they pull in. A reasonable leading indicator of household balance-sheet stress.
• Financials (XLF). Banks, insurers, capital markets. Beneficiaries of rising rates through net interest margin and of economic growth through loan demand; acutely exposed to financial crises, as 2008 established with some thoroughness.
• Industrials (XLI). Manufacturing, transportation, construction. A direct read on business investment and infrastructure spending.
• Materials (XLB). Mining, chemicals, metals. Demand tied to industrial production; pricing tied to commodity curves.
• Energy (XLE). Oil, gas, refining. Driven by commodity prices and global demand, with a geopolitical overlay that does not always coincide with the business cycle. The 1973 and 1979 oil shocks were external to the US cycle and dominated it anyway, which is worth keeping in mind whenever a clean phase model starts to look like the only model required.

Defensive Sectors

• Healthcare (XLV). Pharmaceuticals, medical devices, hospitals. Demand is biological rather than economic. Revenue stability is the distinguishing feature.
• Consumer Staples (XLP). Food, beverages, household products. Procter & Gamble, Coca-Cola, Walmart. Unexciting by design; resilient by mechanism.
• Utilities (XLU). Electricity, gas, water. Regulated monopolies, stable cash flows, high dividends. Bond-proxy behaviour, which means sensitivity to long rates as well as to the cycle — a feature easily overlooked in environments where the curve is moving.
• Real Estate (XLRE). REITs. Defensive by virtue of rental cash flows, but rate-sensitive on the financing side. The 2022 drawdown made the latter point with some force.
• Communication Services (XLC). A hybrid: defensive telecoms alongside cyclical media and entertainment. The sector behaves ambiguously, and one is rather better off thinking about its constituents individually than treating the wrapper as a coherent allocation.

Which Sectors Earn in Which Phase

The Dollar Sits Behind the Table

Rotation textbooks tend to present the phase table as if it were complete. It is not. The exchange rate sits behind the cyclical/defensive sort, and on occasion overrides it. Two episodes are worth keeping in front of the reader. The Plaza Accord of September 1985, by which the G5 deliberately weakened the dollar against the yen and the deutsche mark, produced a multi-year tailwind for US multinationals and for materials in particular — sectors that in a closed-economy framework would have looked unpromising at that point in the US cycle, and which proceeded to outperform anyway because the currency dimension dominated the domestic one for the better part of three years. The Asian Crisis of 1997, conversely, sent capital into the dollar, hammered emerging-market-exposed names, and rewarded domestically-oriented defensives well before the US cycle itself rolled. In both cases, a sector model that ignored the currency dimension would have allocated to the wrong sleeve. The lesson is dull but important: rotation in an open economy is a two-axis problem, and the second axis is the dollar. One can solve the problem by tilt rather than by overhaul, but pretending the second axis does not exist is a cheaper fix than it appears, with a longer bill at the back end.

Implementation: A Working Framework

A serviceable sector-rotation framework has five components.

1. Identify the current phase. The relevant inputs are GDP growth, employment data, inflation, the shape of the yield curve, and leading indicators. The Conference Board’s Leading Economic Index is a competent composite. None of these is sufficient on its own.
2. Overweight the leading sectors. If the assessment is mid-expansion, tilt toward technology and industrials. If contraction appears to be beginning, shift toward healthcare and consumer staples. The correct tilt is ten to twenty per cent of portfolio weight, not a wholesale reallocation.
3. Underweight the lagging sectors. Reduce exposure to sectors that historically underperform the current phase. Shorting them is a separate decision and requires its own evidence.
4. Use sector ETFs. The Select Sector SPDR ETFs (XLK, XLF, XLE, XLV, and the rest) collapse the rotation into a single trade per sector. Stock selection inside a sector is a separate exercise, and a more speculative one.
5. Review monthly. Macro conditions evolve on a timescale of months, not days. A monthly review of the data and of sector relative strength is sufficient. Over-trading is the ordinary failure mode, and rather harder to avoid than the textbooks acknowledge.

Relative Strength as Confirmation

Economic assessment supplies the thesis; relative strength supplies the confirmation. The arithmetic is straightforward: divide a sector ETF’s price by the S&P 500 (SPY) price and observe the ratio. A rising ratio means the sector is outperforming the market; a falling ratio means it is not.

When the macro reading says “overweight financials” and the XLF/SPY ratio is rising, the trade has two independent pieces of evidence behind it. When the macro and the tape disagree, the correct response is usually to wait. Markets are sometimes ahead of the data, and sometimes wrong; the disagreement itself is informative, and reading it is most of the skill in discretionary rotation.

Pacing the Stage

Sector rotation has something of the rhythm of a Grand Tour. The general classification is not won in the first week, and the rider who attacks every climb on the opening day rarely makes it to the third week with legs intact. The peloton sets the pace; a domestique sits on a wheel; the leader saves himself for the mountain stages and the time trial. A sector book run by attacking every move on the tape is the cycling equivalent of pulling on the front of the bunch from kilometre zero — admirable energy, predictable result. The rotational discipline is closer to the team that conserves through the flat days, identifies the two or three stages that will decide the race, and commits then. Cycles, like Grand Tours, are won by reading which stage one is on, and by accepting that most of the days do not require an attack.

What Late-Cycle Looked Like in 2007

An archived editorial portfolio from this period ran through a textbook late-expansion-to-contraction cycle. Through 2005 and 2006 — mid-to-late expansion — the best-performing positions were concentrated in technology, energy, and small-cap growth: the classic late-cycle winners. As 2007 turned into early 2008, the macro backdrop shifted decisively toward contraction. The largest losses clustered in cyclical names caught on the wrong side of the turn — mining, financial-adjacent companies, and speculative small-caps reliant on easy credit.

A rotation into healthcare, staples, and cash as the cycle turned would have meaningfully compressed the 2008 drawdown. This is not hindsight dressed up as analysis; it is the working lesson of the discipline. The cycle turns, and a portfolio that does not turn with it accepts the full amplitude of the turn.

A Rules-Based Momentum Overlay

For traders who would rather not rely on subjective phase identification — a defensible preference, given how often phases are disputed in real time — a simple momentum overlay does much of the work.

1. Monthly screen. At month-end, rank all eleven sector ETFs by trailing three-month return.
2. Buy the top three. Allocate equally among the three sectors with the strongest three-month momentum.
3. Rotate on exit. If a sector drops out of the top three on the following month’s screen, it is replaced by whichever sector has taken its place.
4. Cash filter. If the S&P 500 is trading below its 200-day moving average, move the sleeve to cash or short-duration Treasuries. This keeps the strategy out of the market during the major drawdowns, which is where most of the long-run damage occurs.

The virtue of this approach is that it does not require a phase call. It follows what is in fact working, and rotates into new leadership with a modest lag that the trend-following discipline compensates for. It allocates the phase identification one was attempting to dispense with to the price series, which is a more candid arrangement than it first appears. Less elegant than a discretionary cycle view, and probably more robust. One can hold both views without contradiction; the better operators tend to.

Common Errors

• Rotating too late. Markets are forward-looking. Rotation needs to happen three to six months before the economic data confirms the shift; by the time a recession is officially dated, defensive sectors have ordinarily outperformed cyclicals by ten to fifteen per cent already.
• Ignoring the cycle entirely. A portfolio permanently concentrated in a single sector will be brilliantly right for part of the cycle and quite badly wrong for the rest. The compounded outcome is worse than most investors assume.
• Over-rotating. A jump from one hundred per cent technology to one hundred per cent utilities is not rotation; it is speculation under a different name. Rotation means tilting. Broad diversification is preserved; adjustments happen at first remove.
• Fighting the central bank. When the Fed is cutting, cyclical and growth sectors lead. When it is raising, defensives close the gap. Aligning with monetary policy is not a prediction — it is a recognition of the mechanism.
• Chasing last quarter’s winner. A sector that has outperformed for several quarters may be closer to the end of its run than to the beginning. Momentum and macro evidence together are more reliable than either alone. One hopes the lesson does not have to be relearned each cycle, though on the record it generally does.

Base case: rotation is a tilt discipline, executed monthly, anchored to a phase reading and confirmed by relative strength. Tail risk: the currency axis moves, or a commodity shock arrives off-schedule, and the phase model allocates the portfolio to a sector that the dollar has just demoted. Both cases occur. Neither retires the framework.

See also: The Federal Reserve and Interest Rates · What Is Inflation? · What Is a Bull Market? · What Is a Bear Market? · ETFs