XLE vs Brent one-session lead-lag: does XLE lead crude more than crude leads XLE?
Energy equities are supposed to price what crude hasn't figured out yet. On the evidence, XLE does get a head start — just a small one. Across 718 overlapping sessions, an XLE move today correlated 0.147 with Brent's move tomorrow, a result that's hard to write off as noise. Run the relationship the other way and it evaporates: 0.027, statistically a coin flip.
Scale is the caveat. Same-day correlation between the two is 0.463, so they mostly move together in real time and the one-session lead is a thin residual on top of that. The gap between the two directional leads is +0.120, with 99% of bootstrap resamples keeping XLE ahead — though the reported test on the gap itself doesn't cleanly pin the size down, only the direction.
That makes it a genuine lean rather than a tradable edge. The cross-correlation table, regression slopes, and bootstrap detail are below.
Over the past ~3 years, does XLE's close-to-close return lead Brent crude's next-day return by more than Brent crude leads XLE's next-day return? I expect XLE to lead Brent by one session because energy equities price supply-demand expectations ahead of spot crude.
How this was measured
Daily XLE close-to-close returns from minute bars were joined to the daily Brent value series on common business dates. Two one-session lead coefficients were estimated: corr(XLE_t, Brent_t+1) and corr(Brent_t, XLE_t+1), with OLS slopes and R² as auxiliary measures. A bootstrap over session pairs quantifies whether the correlation gap differs from zero. A cross-correlation table from lag -5 through +5 places the one-session lead in context against same-day and multi-day dynamics.
The key numbers
Reading the numbers
Over 718 shared trading sessions, XLE and Brent moved together on the same day far more than either led the other: same-day correlation was 0.463, versus 0.147 for XLE today to Brent tomorrow and 0.027 for Brent today to XLE tomorrow. The XLE-leads gap of +0.120 points the right way, but it isn't statistically distinguishable from zero.
The charts
Each dot is one session pair: XLE's close-to-close return on the horizontal axis and Brent's return the next day on the vertical. The cloud is wide in both directions, spanning roughly minus 9% to plus 9% for XLE and minus 15% to plus 13% for Brent, with both averages sitting near zero, so there is no obvious diagonal line running through it. That visual looseness matches the weak 0.147 correlation: knowing XLE's move today tells you only a little about Brent's move tomorrow.
This bar chart compares the two competing one-session leads: XLE today to Brent tomorrow at 0.147, versus Brent today to XLE tomorrow at 0.027. The XLE bar is about five times taller, which is the +0.120 gap the report highlights and is the direction the hypothesis expected. But the key caveat is scale: both bars are small, and the same-day XLE-Brent correlation of 0.463 is far larger than either one, meaning the two markets mostly reprice together rather than one reliably going first.
The rolling 60-session lines show how unstable the lead relationship is through time: the XLE-to-Brent line averages 0.153 but swings from minus 0.188 to 0.480, while the Brent-to-XLE line averages 0.028 and ranges from minus 0.359 to 0.403. The XLE line is generally higher, but the two cross repeatedly and both dip negative, so the lead is a tendency rather than a steady rule. Ending readings of 0.182 versus 0.082 look supportive, but the history says the gap comes and goes, consistent with a lead that is directional yet not statistically clear.
Cross-correlation by lead-lag k
| k_sessions | direction | n | pearson_r | p_value |
|---|---|---|---|---|
| -5 | Brent leads XLE by 5 | 713 | -0.0308 | 0.4119 |
| -4 | Brent leads XLE by 4 | 714 | 0.0092 | 0.8054 |
| -3 | Brent leads XLE by 3 | 715 | -0.0392 | 0.2955 |
| -2 | Brent leads XLE by 2 | 716 | -0.0001 | 0.9981 |
| -1 | Brent leads XLE by 1 | 717 | 0.0269 | 0.4716 |
| 0 | Same session | 718 | 0.4627 | 0 |
| 1 | XLE leads Brent by 1 | 717 | 0.1468 | 0.0001 |
| 2 | XLE leads Brent by 2 | 716 | -0.0233 | 0.5329 |
| 3 | XLE leads Brent by 3 | 715 | -0.0248 | 0.508 |
| 4 | XLE leads Brent by 4 | 714 | 0.0182 | 0.6264 |
| 5 | XLE leads Brent by 5 | 713 | 0.0403 | 0.283 |
One-session lead summary
| direction | n | pearson_r | ols_slope | ols_p | r2 |
|---|---|---|---|---|---|
| XLE -> Brent next-day | 717 | 0.1468 | 0.2778 | 0.0001 | 0.0215 |
| Brent -> XLE next-day | 717 | 0.0269 | 0.0142 | 0.4716 | 0.0007 |
The takeaway
Yes — XLE does lead Brent more than Brent leads XLE, but the edge is real only in one direction and modest in size. Over 718 overlapping sessions from late 2023 through August 2026, an XLE move today showed a 0.147 correlation with Brent's move tomorrow (p ≈ 0.0001, roughly a 1-in-12,000 fluke), while Brent today versus XLE tomorrow came in at 0.027, which is statistically a coin flip (p = 0.47). The gap between those two leads is +0.120, and 99% of bootstrap resamples put XLE ahead — though the reported two-sided test on the gap itself came back at 0.50, so the size of the edge isn't cleanly pinned down, only its direction. Perspective matters here: same-day XLE-Brent correlation is 0.463, far larger than either next-day figure, meaning the two assets mostly move together in real time and the one-day head start is a thin residual, not a separate signal. The XLE-to-Brent slope of 0.278 versus 0.014 the other way tells the same story — a useful tilt, and the regression explains only about 2% of next-day variance in crude. Bottom line: the direction of your thesis is supported, the magnitude is small, so treat it as a genuine lean rather than a tradable gap you can lean on hard.
The fine print
- The bootstrap tells a slightly mixed story: the 95% interval on the lead gap [0.029, 0.216] excludes zero and 99% of resamples favor XLE, yet the reported two-sided p on that gap is 0.50 — so the direction is convincing but the size is not.
- Brent daily values follow a T+1 publication convention, so a Brent return may only be confirmable a day later — this is a lead-lag observation, not a same-clock executable signal.
- XLE holds integrated oils and refiners, not pure upstream producers, which dilutes the clean supply-demand-expectations channel.
- 717 sessions is about three years and only one energy regime set; a 2022-style supply shock could produce a different lead pattern.