Does XLE's first 30 minutes after a Brent -1% day predict its open-to-close return?
After Brent crude slides more than 1%, the next morning's first half-hour in XLE tends to separate accumulation from de-risking. Across 222 oil-shock days, an early positive start led to a higher close 68% of the time, while a negative start closed higher just 30.8% of the time. That spread, backed by a rank correlation near 0.52, suggests the open is not noise — it carries real information about where the session ends.
This study tests that relationship directly: how well XLE's first-30-minute return predicts its same-day open-to-close return after a Brent selloff. The full evidence — correlations, regression slope, the split by early-session sign — is in the analysis below.
Over the past ~3 years, when Brent crude falls more than 1% on day t, does XLE's first-30-minute return on day t+1 predict its same-day open-to-close return? I expect a positive first 30 minutes after oil-shock selloffs to signal institutional accumulation and a rebound into the close, while a negative first 30 minutes marks continued de-risking and a weak close.
How this was measured
Brent daily returns below -1% define oil-shock days. XLE regular-session minute bars are collapsed into trading sessions. The first-30-minute return is measured from the 09:30 open through the 09:59 close, and the same-day open-to-close return is measured from the daily open to the last regular-session close. Each Brent trigger day is matched to the next available XLE session. The relationship is estimated with Pearson and Spearman correlation, OLS slope, and a Welch two-sample split by first-30-minute sign.
The key numbers
Reading the numbers
Across 222 Brent -1% shock days, XLE's first 30 minutes averaged -0.03% and the close averaged -0.07%, but the two moved together (r=0.49). Positive early-tape days closed +0.49% on average versus -0.55% for negative early-tape days, a 1.04 point gap that is statistically clear.
The charts
Each dot is one of the 222 shock days, and the cloud tilts upward from the lower-left to the upper-right: days with a stronger first 30 minutes tend to have a stronger open-to-close return, matching the idea that early buying after an oil selloff carries into the close. The tilt is statistically clear, with a Pearson correlation of 0.49 and p below 0.001, and the OLS slope of about 0.93 means a 1% move in the first half hour is associated with nearly the same move in the close. But the dots are widely scattered, and the R-squared of 23.8% is a reminder that the first 30 minutes only partially explains how the day ends.
The two boxes show the key sign split: on days when the first 30 minutes were positive, the average close was +0.49%, while on days when the first 30 minutes were negative, the average close was -0.55%. That roughly 1.04 percentage point gap is not a fluke (Welch p=0.000), so the sign of the early tape after an oil shock is a useful directional read. The boxes overlap on the tails, though, so a positive early tape does not guarantee a green close and a negative early tape does not guarantee a red one.
Open-to-close return by first-30-min sign
| First30 sign | N | Mean OC | Median OC | P(OC>0) |
|---|---|---|---|---|
| All | 222 | -0.0007 | -0.0004 | 0.482 |
| Positive | 100 | 0.0049 | 0.0046 | 0.68 |
| Negative | 120 | -0.0055 | -0.0041 | 0.3083 |
The takeaway
Yes — after a Brent crude drop of more than 1%, XLE's first half-hour is a real same-day tell, not noise. On the 100 days the first 30 minutes were positive, XLE closed higher 68% of the time with an average gain of +0.49%; on the 120 negative first-half-hour days, it closed higher just 30.8% of the time and averaged -0.55%. The rank correlation is 0.52 and the Pearson correlation is 0.49, with p-values so small that this being luck is essentially impossible. That said, the first 30 minutes explains only about 24% of the day's variance, so it is a meaningful edge rather than a crystal ball. The practical takeaway: after an oil-shock selloff, early strength in XLE leans toward a rebound into the close, while early weakness leans toward another weak close — a useful tilt, not a guaranteed outcome.
The fine print
- Oil-shock days cluster during crude downtrends; 222 events may overstate the independent sample size behind the correlations and t-test.
- Brent dates use the index date as the shock day and match to the next XLE session; if the index reflects release date rather than value date, alignment could shift by one session.
- The first-30-min window stops at 09:59 ET; including the 10:00 bar can change the result, and no controls for SPY or macro news are applied.
- 'Institutional accumulation' is an interpretation; the data show return continuation, not order flow or who was buying.