AI Research BPmacro:brent_daily

BP beta to Brent crude: bottom vs top quintile of Brent 20-day realized volatility

0.423
BP beta to Brent — bottom quintile
The research question

For BP over the past ~3 years, is the daily-return beta of BP to Brent crude higher when Brent's 20-day realized volatility is in the top quintile than when it is in the bottom quintile? I expect BP to trade as a higher-beta crude proxy during high-vol regimes because traders reprice the supermajor more aggressively to oil swings when the commodity signal is loudest.

How this was measured

BP and Brent daily returns were computed from daily closes over the available overlapping ~3-year window. For each trading day, Brent 20-day realized volatility was computed on trailing Brent returns and shifted one day so the regime label at day t uses only data through t-1; days were then split into bottom and top quintile regimes using full-sample 20th/80th percentiles. Within each regime, BP's beta to same-day Brent returns was estimated with ordinary least squares; an interaction model tested whether the slope difference between the top and bottom quintiles is statistically distinguishable from zero.

The key numbers

BP beta to Brent — bottom quintile
0.423
N=137 days; Brent 20d RV lower 20%
BP beta to Brent — top quintile
0.235
N=137 days; Brent 20d RV upper 20%
Beta difference (top minus bottom)
-0.188
Positive = BP is higher-beta during high crude-vol regime
Interaction p-value
0.0276
Interaction p=0.0276 < 0.05 -> slope difference is statistically clear
Avg Brent 20d realized vol — bottom quintile
19.6788%
Annualized, trailing 20d, shifted one day
Avg Brent 20d realized vol — top quintile
65.7867%
Annualized, trailing 20d, shifted one day

Reading the numbers

BP's beta to Brent was actually lower in the high-vol regime (0.235 vs 0.423), the opposite of the expectation. The gap is statistically clear (p=0.0276), so the pattern is unlikely to be random noise.

The charts

BP beta to Brent crude by Brent realized-volatility regime
What this chart says

The bar on the left, for bottom-quintile Brent volatility, is nearly twice as tall as the bar on the right for top-quintile volatility — 0.423 vs 0.235. That is backwards from the hypothesis: BP acted less like a crude proxy when oil was most volatile, even though average realized vol jumped from 19.68% to 65.79%. The statistically clear difference (p=0.0276) means this isn't just noise; in the data, calmer oil markets are when BP's daily returns track Brent more closely.

Regime-level OLS estimates

Vol regimeN daysBP betaBeta std errR squaredAvg Brent 20d RVAvg BP daily retAvg Brent daily ret
Bottom quintile1370.42270.0750.19060.1968-0.0008-0.0002
Top quintile1370.23460.03030.30690.65790.00080.0009

The takeaway

The answer is no — and in fact the result runs the other way. BP's daily beta to Brent was lower, not higher, during Brent's high-volatility regime: 0.235 in the top quintile versus 0.423 in the bottom quintile, a difference of -0.188. That gap is not noise: the interaction p-value is 0.0276, meaning there is roughly a 3-in-100 chance this difference would appear by luck, so the contrast is statistically clear at conventional confidence levels. Both regimes have 137 trading days, and the quintiles are genuinely distinct — average annualized Brent 20-day realized vol is 19.7% in the low bucket versus 65.8% in the high bucket. So the practical takeaway is that during the calmest crude-oil stretches BP has tended to track same-day Brent moves more closely, while in chaotic oil markets that same-day relationship loosens. For anyone using BP as a crude proxy, the intuitive "louder oil signal, higher beta" story is not supported by this three-year window; the relationship is real but opposite to the hypothesis.

The fine print