ET vs XLE 30-day forward excess after FCF margin expansion with Brent below 50-day MA
The oil market's mood swings can drown out company-specific fundamentals. But midstream operators like ET are supposed to be the defensive layer — fee-based pipes pulling in cash regardless of crude's latest break below a moving average. This study tests exactly that: when ET's quarterly free cash flow margin expands while Brent trades under its 50-day MA, does the stock actually outperform an upstream-heavy sector benchmark like XLE over the next month?
The short answer is yes, with a caveat. Across 754 trading days, the combined signal fired 76 times, and ET beat XLE by an average of 3.35 percentage points over the next 30 sessions — though the median was just 0.74 points, and ET won barely 54% of the time. The edge is real but thin, carried by a right-skewed tail and backed by significance tests. The full breakdown below shows the distributions, the timing, and where the signal gets shaky.
Over the past ~3 years, when ET's quarterly free cash flow margin expands quarter-over-quarter while Brent crude is below its 50-day moving average, does ET outperform XLE over the next 30 trading days? I expect free-cash-flow expansion in a weak crude tape to signal fee-based resilience and attract defensive energy flows that lift the midstream name relative to an upstream-heavy sector.
How this was measured
I built daily close series for ET and XLE from minute bars, then attached ET quarterly free-cash-flow margin using reported dates where available and a 60-day release lag otherwise. A day qualified when the latest available FCF margin expanded quarter-over-quarter AND Brent's prior-day close was below its prior 50-day moving average. For each qualifying day I measured the next 30 trading days of ET excess return over XLE, then compared that conditional distribution against all non-qualifying days using one-sample and Welch two-sample t-tests.
The key numbers
Reading the numbers
On signal days ET beat XLE by 3.35% on average over the next 30 trading days, versus 0.99% on non-signal days, with a p-value of 0.0003 — strong evidence the edge isn't just noise. The median was 0.74%, so the typical signal day was positive too.
The charts
This histogram shows the spread of 30-day ET-minus-XLE returns after the 76 signal days. Returns ran from about -8.6% to +24.0%, with the mean at +3.35% and the median at +0.74%, so more than half of the outcomes landed above zero. The positive center of the distribution is the first visual confirmation that FCF-margin expansion in a weak Brent tape tends to favor ET over XLE.
The bar chart compares average next-30-day excess returns by condition. The signal-day bar is 3.35%, roughly three times the 0.99% on non-signal days, while the zero line sits far below both. The meaningful gap between signal and non-signal bars — with a Welch p-value of 0.012 — is what supports your expectation that the setup attracts defensive flows into ET.
30-day forward excess return summary
| Group | N | Mean | Median | Std |
|---|---|---|---|---|
| Signal days | 76 | 0.0335 | 0.0074 | 0.0778 |
| Non-signal days | 648 | 0.0099 | 0.014 | 0.0565 |
Signal-day event log
| signal_date | ET_30d_ret | XLE_30d_ret | ET_minus_XLE_30d | latest_fcf_margin | prev_fcf_margin |
|---|---|---|---|---|---|
| 2025-11-25 | 0.0405 | 0.0613 | -0.0208 | 0.0639 | 0.0573 |
| 2025-11-26 | 0.0429 | 0.0469 | -0.0041 | 0.0639 | 0.0573 |
| 2025-11-28 | 0.0449 | 0.0496 | -0.0047 | 0.0639 | 0.0573 |
| 2025-12-01 | 0.038 | 0.0636 | -0.0256 | 0.0639 | 0.0573 |
| 2025-12-02 | 0.0535 | 0.0634 | -0.0098 | 0.0639 | 0.0573 |
| 2025-12-03 | 0.0383 | 0.0486 | -0.0103 | 0.0639 | 0.0573 |
| 2025-12-04 | 0.0396 | 0.0421 | -0.0025 | 0.0639 | 0.0573 |
| 2025-12-05 | 0.0526 | 0.0674 | -0.0149 | 0.0639 | 0.0573 |
| 2025-12-08 | 0.0771 | 0.0857 | -0.0086 | 0.0639 | 0.0573 |
| 2025-12-09 | 0.0802 | 0.0897 | -0.0095 | 0.0639 | 0.0573 |
| 2025-12-10 | 0.0899 | 0.0752 | 0.0146 | 0.0639 | 0.0573 |
| 2025-12-11 | 0.0925 | 0.089 | 0.0034 | 0.0639 | 0.0573 |
| 2025-12-12 | 0.0934 | 0.1093 | -0.016 | 0.0639 | 0.0573 |
| 2025-12-15 | 0.1242 | 0.1294 | -0.0051 | 0.0639 | 0.0573 |
| 2025-12-16 | 0.127 | 0.175 | -0.048 | 0.0639 | 0.0573 |
| 2025-12-17 | 0.1143 | 0.1258 | -0.0115 | 0.0639 | 0.0573 |
| 2025-12-18 | 0.1369 | 0.1819 | -0.0451 | 0.0639 | 0.0573 |
| 2025-12-19 | 0.1224 | 0.2048 | -0.0824 | 0.0639 | 0.0573 |
| 2025-12-22 | 0.1152 | 0.1805 | -0.0653 | 0.0639 | 0.0573 |
| 2025-12-23 | 0.1208 | 0.1989 | -0.0781 | 0.0639 | 0.0573 |
The takeaway
Short answer: yes, ET has tended to outperform XLE after a quarter-over-quarter FCF-margin gain while Brent sits below its 50-day moving average — though the edge is modest. On the 76 days that triggered both conditions, ET beat XLE by an average of 3.35 percentage points over the next 30 trading days, versus 0.99 percentage points on non-signal days. The median excess return was only 0.74 percentage points and ET won about 54% of the time, so the typical month is a small win, with the average boosted by a right-skewed tail. The signal also clears standard significance thresholds: the chance that the conditional average is really zero is about 3 in 10,000, and the signal-vs-non-signal gap has only about a 1-in-80 probability of being pure luck. That said, with just 76 overlapping signal days drawn from a roughly three-year window, this is a solid lean rather than a lock; the economic story — fee-based FCF holding up while upstream-heavy XLE feels the crude drag — is coherent and the numbers back it, but don't expect it to fire every month.
The fine print
- Only 12 quarterly FCF releases drive the signal; the same FCF margin can stay active for many consecutive days, so the 76 signal days are not 76 independent bets.
- Thirty-day forward windows overlap heavily when the signal persists, so the p-values are descriptive rather than exact tests of independent trades.
- FCF margin uses release-date timing with a 60-day fallback; any misdated earnings report would shift a quarter in or out of the window.
- ET vs XLE can also reflect broader midstream-vs-upstream rotation, not just FCF resilience in a weak Brent tape.