Does XOM 20-day realized beta to Brent fall after positive EPS surprises?
Earnings beats are supposed to make XOM trade more like a company and less like a crude proxy. Over the past three years, the opposite showed up. Across the 8 quarterly reports with a positive EPS surprise, XOM's 20-day realized beta to Brent averaged 0.24 before the print and 0.34 after — higher, not lower.
The direction was consistent, too: beta fell in only 2 of the 8 events, and the largest swings were increases. Correlation tells the same story. With 8 observations and noisy 20-day beta estimates, this is inconclusive rather than proof of a reverse effect — but there is no support here for the decoupling thesis.
The full event-by-event breakdown, methodology and significance tests are in the analysis below.
Over the past ~3 years, does XOM's 20-day realized beta to Brent crude fall after a positive quarterly EPS surprise relative to the 20 days before the report? I expect a supermajor earnings beat to shift XOM's returns from oil-beta pass-through to company-specific capital-return and cost-control news, so realized correlation with Brent should decline.
How this was measured
Built daily XOM returns from minute bars and aligned Brent daily settlement prices onto the XOM trading calendar. Screened XOM earnings for reports with reported surprise percentage above zero. For each qualifying report, located the first XOM trading day on or after the report date and measured realized beta as the OLS slope of XOM daily returns on Brent daily returns over the 20 trading days before the report (t-20..t-1) and the 20 trading days after the report (t+1..t+20). The primary comparison is the paired change in pre- vs post-event beta; correlation is reported as a supplement.
The key numbers
Reading the numbers
Across 8 positive EPS surprises, XOM's average 20-day beta to Brent actually rose from 0.24 to 0.34, not fell, and it dropped in only 25% of cases. The paired tests (p=0.76 and p=0.90, well above the usual 0.05 threshold) say this sample gives no reliable evidence of a decline either way.
The charts
These two side-by-side boxes show the 20-day oil beta for each XOM report window, measured before the earnings release and again after it. The 'after' box sits higher than the 'before' box — its mean moves from roughly 0.24 to 0.34 and its top whisker reaches 0.73, versus a high of 0.95 before — so the typical event pushed XOM's returns more in line with Brent, the opposite of the hypothesis. What deserves your eye is the upward shift in the middle of the boxes rather than any single outlier: only one quarter of the 8 events saw beta fall, so the decline you were looking for is the minority case here, not the pattern.
This scatter plots each report as one dot, with its pre-surprise beta on the horizontal axis and its post-surprise beta on the vertical axis; points above the diagonal mean beta rose after the beat. Most dots sit above that invisible 45-degree line, and the pre-beta range runs from about -0.02 (essentially no oil linkage) up to 0.95, while post-beta stays in a tighter 0.07 to 0.73 range. The takeaway for your question is that the movement is not a clean rotation into company-specific news — the low-beta events tend to drift up and the high-beta ones drift down, which is the noise the report flags rather than a systematic post-earnings decoupling from Brent.
Per-event XOM-Brent 20-day beta around positive EPS surprises
| report_date | surprise_pct | pre_beta | post_beta | delta_beta | pre_corr | post_corr |
|---|---|---|---|---|---|---|
| 2024-02-02 | 13.24 | 0.178 | 0.067 | -0.111 | 0.262 | 0.051 |
| 2024-04-26 | 0.98 | -0.019 | 0.261 | 0.28 | -0.037 | 0.417 |
| 2024-08-02 | 6.47 | 0.112 | 0.54 | 0.428 | 0.096 | 0.811 |
| 2024-11-01 | 2.13 | 0.269 | 0.383 | 0.114 | 0.75 | 0.454 |
| 2025-01-31 | 10.97 | 0.22 | 0.726 | 0.506 | 0.283 | 0.769 |
| 2025-05-02 | 1.15 | 0.953 | 0.228 | -0.725 | 0.77 | 0.444 |
| 2025-08-01 | 5.13 | 0.055 | 0.187 | 0.133 | 0.074 | 0.252 |
| 2026-07-31 | 11.68 | 0.161 | 0.353 | 0.192 | 0.471 | 0.728 |
The takeaway
The answer is no — XOM's 20-day beta to Brent did not fall after positive EPS surprises; it went the other way. Averaged across the 8 qualifying reports, realized beta was 0.24 in the 20 days before the report and 0.34 in the 20 days after, a mean increase of about +0.10 (median +0.16). Beta actually declined in only 2 of the 8 events, so 75% of the time it rose, and the biggest moves were upward — +0.43 after the August 2024 report and +0.51 after January 2025. The one-sided test for a decline came in at p = 0.76, and the rank-based version at p = 0.90, both nowhere near the 0.05 bar, which is what you'd expect if there were no effect at all. With just 8 events and noisy 20-day beta estimates, this is inconclusive rather than proof of a reverse effect — but there is zero support here for the idea that a beat decouples XOM from crude. Practically: don't build a trade or a risk model around the notion that earnings beats mute XOM's oil beta over the following month.
The fine print
- Only 8 earnings events with complete windows — thin evidence for any paired test, so treat the direction as a hint rather than a verdict.
- Small and large beats are pooled; if only big surprises matter, averaging them with marginal ones would wash the effect out.
- Each 20-day beta comes from roughly 20 daily returns and blends the earnings reaction with oil moves, macro news, and other company headlines.
- Brent settlement prices were forward-filled onto XOM trading days and the event anchor can shift for after-hours releases, so daily timing is approximate.