Pillar 06 · Equities · 11 GICS Sectors
Once rotation has named the sectors to own — this is the evidence underneath that call: how each of the eleven actually behaves against the market.
Eight-panel decomposition of how the eleven S&P 500 sectors behave relative to the market — industry beta and rolling sensitivity, Jensen's alpha, relative strength and drawdowns, correlation regimes, a multi-factor model that fuses sector returns with the Treasury curve, rotation cycles, cross-sectional dispersion, and each sector's contribution to the index.
Beta answers one question: when the whole stock market moves 1%, how much does this industry move? It is the single most fundamental measurement in finance, calculated as β = cov(industry, S&P 500) ÷ var(S&P 500). A beta of 1.0 means the industry moves in lockstep with the market. The rolling beta below recomputes this over a moving 12-month window so you can watch each industry's sensitivity drift over time.
| Sector | N | Full-Period β | Latest 12M β | Δ vs First Window |
|---|
CAPM says an industry's return should be explained entirely by its beta and the market. Alpha is whatever's left over: α = R_industry − [rf + β × (R_market − rf)], where rf is the risk-free rate (here the 3-month T-bill). Positive alpha means the industry earned more than its market risk justified — it was either genuinely mispriced or carries a reward CAPM can't see. The rolling version reveals when each industry was in or out of favour.
| Sector | β | Annualised α | Latest 12M α |
|---|
Relative strength divides a sector's cumulative growth by the S&P 500's: cum_industry ÷ cum_S&P. A rising line means the sector is outperforming the market; a falling line means it's lagging. Plotting all eleven sectors together gives an instant picture of the rotation cycle. Drawdown is a separate risk lens — the worst peak-to-trough fall — because a sector can have low beta yet still suffer deep, sudden drops in specific episodes.
| Sector | Total Return | vs S&P | Max Drawdown |
|---|
Correlation measures how tightly a sector's moves track the market, on a scale of −1 to +1. The catch: it isn't constant. In a crisis, everything falls together and correlations spike toward 1.0 — exactly when the diversification you were counting on evaporates. This tab tracks each sector's rolling 12-month correlation with the S&P 500 so you can see those regime shifts directly.
Beta only explains the market piece. This tab runs a richer regression that fuses the sector data with the Treasury rates data: R_industry = α + β_mkt·R_S&P + β_rates·Δ10Y + β_credit·ΔHY-spread + ε. It separates how much of each sector's return comes from the broad market versus moves in interest rates versus credit stress — turning two datasets into one unified model.
β_rates means the sector gets hurt when long yields rise — a real risk in a hiking cycle. A large positive β_credit means it suffers as credit spreads widen (risk-off).⚠ The high-yield credit-spread factor (BAMLH0A0HYM2) only begins 2023-05-30, so β_credit is estimated on the overlapping months and is less stable than the market and rate loadings.
| Sector | α (ann.) | β Market | β Rates (Δ10Y) | β Credit (ΔHY) | R² |
|---|
Textbook theory says different sectors lead at different points in the economic cycle: Financials and Consumer Discretionary in early expansion; Technology and Industrials mid-cycle; Energy and Materials late-cycle; Utilities, Health Care and Staples in contractions. Rather than trust the textbook, this tab tests it empirically. It uses the yield-curve slope (10Y−3M, from the rates data) as a proxy for the cycle phase, sorts every month into a regime, and shows which sectors actually outperformed in each.
| Regime | Slope Range | Cycle Proxy | Months | Top Sector |
|---|
Dispersion is the cross-sectional spread of sector returns in a given month: dispersion_t = std(all sector returns on month t). When dispersion is high, sectors are going their own separate ways — which sector you pick matters a lot. When dispersion is low, everything moves together and macro forces dominate, so sector selection barely helps. Plotted against credit spreads, it becomes a regime gauge built entirely from the data on this site.
A broad index's return is the sum of its parts: contribution_i = weight_i × return_i. Decomposing it shows whether the market's gains are broad-based (many sectors pulling together) or concentrated in a handful — as the 2023–2024 rally was, where a few mega-cap tech names drove most of the gains. This tab breaks the equal-weighted sector index into exactly how much each sector contributed over time.
⚠ Market-cap weights are not available in the source data, so contribution uses an equal-weighted proxy (each sector weighted by its share of the stock count). This measures breadth of participation, not exact cap-weighted S&P index points.