CVX 5-day realized beta to Brent around quarterly earnings
It sounds intuitive: when Chevron reports earnings, company-specific news should swamp the oil tape, at least for a few sessions. To test that, we measured CVX's five-day realized beta to Brent around its last 11 quarterly releases and compared it with the twenty sessions before each report.
The average event beta came in lower — 0.187 versus 0.248 — but the median moved the other way, and only 4 of 11 releases actually showed a lower beta. A few large moves dragged the mean down, while the paired tests came nowhere near significance.
So the simple decoupling story does not hold up. The full event-by-event breakdown, including the releases that drive the result, is below.
Over the past ~3 years, does CVX's 5-day realized beta to Brent crude drop significantly around its quarterly earnings releases relative to the prior 20 sessions? I expect company-specific earnings news to dominate crude-beta pass-through, so Chevron temporarily decouples from oil until the macro tape reasserts.
How this was measured
Build daily CVX and Brent returns aligned on common trading sessions. For each CVX quarterly earnings release with a known reported_date, anchor t0 as the first trading session on or after the release. Estimate an OLS beta of CVX returns on Brent returns over the five sessions t0..t+4 and over the twenty sessions t-20..t-1. Compare the paired event-minus-prior betas with a one-sided paired t-test and a Wilcoxon signed-rank test, and report the fraction of releases where the event beta was lower than the prior beta.
The key numbers
The charts
Event-level beta comparison
| reported_date | event_5d_beta | prior_20d_beta | diff |
|---|---|---|---|
| 2024-02-02 | -0.3553 | -0.0302 | -0.3251 |
| 2024-04-26 | 0.1522 | -0.0881 | 0.2403 |
| 2024-08-02 | 0.55 | 0.0836 | 0.4665 |
| 2024-11-01 | -0.1124 | 0.2111 | -0.3235 |
| 2025-01-31 | -0.0626 | 0.397 | -0.4596 |
| 2025-05-02 | 0.1676 | 1.1162 | -0.9486 |
| 2025-08-01 | 0.2941 | 0.1674 | 0.1267 |
| 2025-10-31 | 0.1919 | 0.1861 | 0.0058 |
| 2026-01-30 | 0.4944 | 0.3764 | 0.118 |
| 2026-05-01 | 0.3107 | 0.1502 | 0.1605 |
| 2026-07-31 | 0.4316 | 0.1553 | 0.2762 |
The takeaway
No — over the last three years Chevron did not reliably decouple from crude around earnings; if anything, the typical earnings week looked slightly more oil-sensitive, not less. The average 5-day event beta was 0.187 versus 0.248 over the prior 20 sessions, a paired difference of -0.060, which looks directionally like the drop you expected. But that average is doing all the work for the hypothesis while the underlying pattern says the opposite: the median difference was +0.118, and only about 36% of releases (4 of 11) actually saw a lower beta around earnings. The negative mean comes from a handful of sharp moves — February 2024 at -0.33, November 2024 at -0.32, January 2025 at -0.46 — offset by a +0.47 swing in August 2024, so a few events are carrying the whole result. Statistically this is a coin flip: a one-sided paired t-test gives p = 0.32 and the Wilcoxon signed-rank test gives p = 0.45, neither close to meaningful, and with just 11 releases the test has very little power to detect a real effect even if one exists. Practically, the takeaway is that you should not assume Chevron's earnings temporarily sever the crude link — there's no tradable decoupling signal here, and any strategy that assumes one is leaning on noise from a dozen observations.
The fine print
- Only 11 earnings events; a 5-day OLS beta is extremely noisy, so this test can't rule out a small real effect.
- The mean and median disagree in sign — a few large outlier events drive the headline average.
- Single-factor Brent beta only: no control for the broad equity market, so a lower event beta could reflect general market noise rather than oil-specific decoupling.
- Event window anchored to the first session on or after the reported date; after-hours release timing and intraday price alignment with Brent settlement aren't modeled.