BKR daily-return beta to Brent crude – high vs normal oil-vol regimes
Elevated oil volatility was supposed to make BKR trade as a hotter crude proxy: tighter hedging demand and a busier drilling outlook should, in theory, push the stock’s sensitivity to Brent higher. Over the past three years, it didn’t. Across 741 overlapping BKR-Brent trading days, the mean 60-day beta to crude was 0.263 when Brent’s 20-day realized volatility sat above its historical median, versus 0.255 in normal-vol regimes. A 0.009 gap, with a p-value of 0.646, is indistinguishable from noise.
The high-vol bucket was messier—beta dispersion roughly doubled to 0.311 from 0.160—but the center of the distribution barely moved. The hedging-demand story fails to show up in realized betas, and the data give no reason to expect BKR to turn into a more leveraged oil play when turbulence spikes.
The full breakdown below details the regime classification, the rolling OLS methodology, and the tests behind the null.
For BKR over the past ~3 years, does its daily-return beta to Brent crude increase in months when Brent's 20-day realized volatility is above its historical median? Elevated oil volatility signals more hedging demand and drilling activity, making BKR's stock trade as a more levered bet on crude during turbulent regimes.
How this was measured
Resampled BKR minute bars and brent_daily_df to daily close prices, computed daily percentage returns. Brent 20-day realized volatility (rolling std of daily returns, window=20) served as the oil-vol signal; days where this 20d RV exceeded the full-sample median were classified as the 'elevated oil vol' regime, others as 'normal'. BKR's market beta to Brent was estimated with a rolling 60-day OLS regression (covariance/var approach). The mean of these 60-day betas was compared across the two regimes using Welch's t-test (two-tailed, unequal variance). Higher mean beta during high vol would indicate that BKR trades as a more levered bet on crude when oil turbulence is elevated.
The key numbers
Reading the numbers
Across 682 usable days, BKR's average Brent beta was 0.263 in high-volatility months versus 0.255 in normal months — a tiny gap of 0.009. The p-value of 0.646 means this difference is easily within the range of random chance, so the data do not show that oil turbulence makes BKR a more levered crude play.
The charts
This line tracks BKR's rolling 60-day sensitivity to Brent from late 2023 through mid-2026. The line swings a lot, dipping below zero near -0.21 at its lowest and spiking above 1.0 at its highest, with the average sitting near 0.26. Notice the most recent reading is only about 0.12, so the current beta is well below the period average. The wide swings show that beta is not constant, but the key question is whether those swings line up with oil-volatility regimes — which the next chart checks.
The two bars compare average beta in elevated-volatility months versus normal-volatility months. The left bar, at 0.263, is barely taller than the right bar at 0.255 — a difference of less than one hundredth of a beta unit. That gap is not meaningful: the statistical test gives a p-value of 0.646, meaning a difference this size would be unsurprising if the true beta were identical in both regimes. So the evidence does not support the idea that BKR becomes a stronger or more leveraged bet on crude when Brent's realized volatility rises above its median.
Beta summary across oil-vol regimes
| Regime | N | Mean beta | Std beta |
|---|---|---|---|
| Elevated oil vol (RV20 > median) | 346 | 0.263 | 0.311 |
| Normal oil vol (RV20 ≤ median) | 336 | 0.255 | 0.16 |
The takeaway
Answer: no—BKR’s daily-return beta to Brent did not rise in months when oil volatility was elevated over the past ~3 years. Days with Brent 20-day realized vol above the median (0.0187) showed a mean 60-day beta of 0.263, versus 0.255 in normal-vol days—a +0.009 difference that is economically trivial. The t-stat of 0.46 and p-value of 0.646 put that gap squarely in coin-flip territory: the data cannot distinguish it from zero, so the hypothesis that turbulent oil makes BKR a more levered crude play is not supported. If anything, the elevated-vol bucket was just messier, with beta std of 0.311 against 0.16 in normal times, but the center of the distribution barely moved. The result is a clean null: no evidence of a regime-dependent beta. The takeaway: over this window, the hedging-demand/drilling-activity story doesn’t show up in realized betas—don’t count on BKR automatically becoming a hotter oil proxy when volatility spikes.
The fine print
- Rolling 60-day betas overlap, so high-vol and normal-vol observations aren't independent; the t-test's degrees of freedom are inflated and the result is a bit too optimistic.
- The vol cutoff is the in-sample median (0.0187); out-of-sample, the regime split could shift.
- High-vol beta estimates were far more dispersed (std 0.311 vs 0.16), so that regime's average is less stable.
- Only one horizon (60-day beta) and one oil-vol window (20-day) were tested; other combinations might behave differently.