How US sectors performed when real yields were zero bound
Since 2 Jan 1985, the record reads this regime on 12.9% of trading days, across 11 episodes with a typical run of 126 sessions. Across those sessions since 22 Dec 1998, Information Technology did best at +15.8% a year and Energy did worst at -5.1%, against +8.4% for the S&P 500.
Figures as of 6 Nov 2026 · recomputed every trading day
The regime in the record
Sector by sector
Annualised mean return on the session after each regime day. Sorted by the regime figure.
| Asset | In this regime | All days | vs S&P 500 | Up sessions | Volatility | Sessions |
|---|---|---|---|---|---|---|
| Information Technology XLK | +15.8% | +11.8% | +7.3% | 55% | 20.7% | 1,356 |
| Consumer Discretionary XLY | +11.3% | +10.6% | +2.8% | 54% | 19.7% | 1,356 |
| Health Care XLV | +11.2% | +8.8% | +2.8% | 54% | 17.6% | 1,356 |
| Financials XLF | +8.9% | +7.9% | +0.4% | 52% | 23.4% | 1,356 |
| Industrials XLI | +8.6% | +9.2% | +0.2% | 53% | 20.2% | 1,356 |
| Consumer Staples XLP | +5.7% | +5.6% | -2.8% | 53% | 14.7% | 1,356 |
| Utilities XLU | +3.6% | +5.7% | -4.9% | 53% | 19.4% | 1,356 |
| Materials XLB | +2.1% | +8.6% | -6.3% | 53% | 21.5% | 1,356 |
| Energy XLE | -5.1% | +10.1% | -13.6% | 49% | 27% | 1,356 |
| Real Estate XLREsince 2015 | -5.9% | +5.5% | -16.1% | 52% | 21.9% | 736 |
| Communication Services XLCsince 2018 | -9% | +11.3% | -4.7% | 56% | 28.3% | 275 |
Context
| Asset | In this regime | All days | vs S&P 500 | Up sessions | Volatility | Sessions |
|---|---|---|---|---|---|---|
| S&P 500 SPY | +8.4% | +10.1% | — | 54% | 17.8% | 1,356 |
| Nasdaq 100 QQQsince 1999 | +16.9% | +12.9% | +8.4% | 56% | 19.9% | 1,356 |
| Russell 2000 IWMsince 2000 | +7.7% | +9.6% | -0.7% | 53% | 22.4% | 1,356 |
| Long Treasuries TLTsince 2002 | +2.7% | +1% | -5.8% | 53% | 16.3% | 1,356 |
| Gold GLDsince 2004 | -1.1% | +11.4% | -9.5% | 51% | 16.4% | 1,356 |
How this regime is defined
Real interest rates reads Zero bound in the daily regime record.
Method and limits
Descriptive statistics over the published regime record — not a backtest, a forecast or advice. Each regime day is credited with the NEXT session's return, so a regime never earns the move that revealed it. Returns are ETF price returns (dividends excluded), annualised as 252 × the mean session return; volatility likewise. Sector funds start in December 1998 (Real Estate in 2015, Communication Services in 2018); the regime record starts in 1980. Labels are the engine's as currently published and are recomputed when its method improves. A figure with fewer than 60 sessions is not shown.
This page's data as JSONQuestions
- Which sectors did best when real yields were zero bound?
- Information Technology +15.8%, Consumer Discretionary +11.3%, Health Care +11.2%. Annualised mean next-session returns since 22 Dec 1998; the S&P 500 returned +8.4%.
- How common is this regime, and how long does it last?
- It appears on 12.9% of trading days since 1985, in 11 separate episodes. The typical run is 126 sessions; the longest lasted 332 sessions.
- Is this regime in force today?
- No. It was last seen 28 Jul 2022 – 1 Sep 2022. The daily nine-axis reading is on the Regime Radar.
- Is this a backtest?
- No. It describes how assets behaved on the session after each day the regime was read, with no portfolio, costs or selection. A backtest would test a rule; this records a history.