Why Energy Stocks Are at a Record While Oil Is Down 31% From Its High — and What Happened Last Time
Research note. Computed from daily closes in our maintained TradeStation price database (XLE from 1999, seventeen energy constituents from 2010 or their listing), Brent crude spot from FRED, and our own 3-2-1 crack spread built from the NY Harbor gasoline and diesel series. The historical studies are frozen through August 20, 2026; the readings, scoreboard and charts continue to update. See How we checked it.
Energy stocks are at a 28-year record because the refining margin is at a 20-year record, and the oil price is close to a bystander. Since Brent peaked on April 7, 2026 the three large refiners in the sector ETF have gained 36–46%; the majors and the producers that make up most of the fund are flat. The one prior time the sector topped out looking like this (November 2022, crude 29% off its high, the crack spread at $60) the margin halved within a year and XLE lost 12%. My read: this is a margin trade wearing an energy-sector label, and the crack spread, not crude, is the number that decides how it ends.
- XLE closed at 63.75 on August 20 — its 23rd record close of 2026, up +42.3% on the year — with Brent at $95, -31% from its April high.
- The equities never priced the spike: over the six months to August 20, Brent gained +37% and XLE +18%. They lagged on the way up and had less to give back.
- The 3-2-1 crack spread is $71 a barrel, the 99.9th percentile since 2006; its 200-day average is the highest on record.
- Across 13 completed record-with-crude-down episodes since 1999, XLE’s next year was a median +11.9% — near its baseline — and the outcome tracked what crude did next, not where it started.
A record that crude did not deliver
The Energy Select Sector SPDR closed at 63.75 on August 20, its third record close in four sessions and its 23rd of the year. It first cleared the June 2014 peak on January 28 and is up +42.3% for 2026, against +12.3% for SPY. Brent is up too, +55% on the year from $61, but the crude story this year is a spike and a retreat: $138.21 on April 7, $95.29 at the latest print. A third of the barrel’s peak value is gone. The equities barely noticed.
Measured from the April 7 crude peak, XLE is +5.8% and Brent -31.1%. Measured over the six months to August 20, the order reverses: Brent +37%, XLE +18%. That second pair is the one I keep coming back to. Energy equities did not track the barrel up to $138 in the spring. They priced something closer to the 200-day average, which sits near $85, so there was no $138 in the stock prices to unwind when crude gave it back. Over the quarter to August 20 the two went opposite ways, XLE +5.2% against Brent -18.4%, and the 63-session return gap between them stood at +23.5 pts, the 96th percentile of its history since 1999.
Energy equities and the barrel, three years
Who did the work: refiners, and almost no one else
Split the sector by what each company actually sells and the record stops being a puzzle. Refiners buy crude and sell gasoline and diesel; a cheaper barrel with firm product prices is a wider margin for them. Since the April crude peak, Valero, Marathon Petroleum and Phillips 66 have gained between 36% and 46%. Exxon and Chevron, which between them are the largest part of the fund, are up 1% and 2%. The producers — ConocoPhillips, EOG, Occidental, Diamondback, Devon — sit between −2% and +8%. Oilfield services are mixed. The only non-refiner doing refiner-like work is Targa, a midstream name with its own volume story.
The arithmetic follows. A fund that is +5.8% since April 7 while its heaviest constituents are flat has had that gain delivered by a minority of its weight. Year to date the refiners are up 86% to 120%. Strip them out and the energy sector in 2026 looks like a sector that tracks a $95 barrel up from $61 — respectable, and well short of a record.
Energy sub-groups since the crude peak (Apr 7, 2026)
Live — price return from the April 7 close to the latest close. Year-to-date in gray.
The number underneath: a 99.9th-percentile refining margin
The 3-2-1 crack spread, two barrels of gasoline plus one of diesel less three barrels of crude, was $71.33 a barrel at the latest print. In the twenty years of daily data we hold, only the weeks around May 2022 were higher; the record is $75.89 on May 13, 2022. What is new is persistence. The spread’s 200-day average is $42.62, the highest in the record, above the $41.32 it reached in October 2022. A spike is a weather event. A record 200-day average is a regime, and it is the regime the refiners are priced for.
This is also where the honest version of the story has to admit what it does not know. The diesel leg is doing most of the lifting (the diesel crack is above $100), and a margin that wide is a statement about product supply, refinery outages and export demand as much as about crude. None of that is visible in an oil price. Which is exactly the point: if you were watching Brent to judge the energy trade this year, you were watching the wrong line.
The line that explains the record
What happened the last time: November 2022
XLE’s 2022 high came on November 15, 2022 at 47.04. Brent that day was $94.30, -29% from its March 8 peak of $133.18, and the crack spread was $60.06. Energy equities at a high, crude well off its peak, a refining margin near the top of its range. It is the same picture as today, drawn in the same colors.
Twelve months later the crack spread was $28.74, less than half. Brent had drifted to $77.73, a modest move for crude. And XLE was -11.8%, having been -18.6% at the May 31, 2023 low: -6.9% after a quarter, -16.0% after two. The equities did not follow the barrel, which barely moved. They followed the margin, which collapsed. That is the base case I carry into this one, and it is also why I would not describe what comes next as an oil call.
The obvious objection is 2014, when XLE also set a high in June and then lost a quarter of its value inside a year. But that high came with Brent at $114, near its own peak, and the year that followed was a crude crash from $114 to $62 with the crack spread flat at $17–22. 2014 was an oil story; the equities fell because the barrel did. 2022 was a margin story; the equities fell while the barrel held. Today rhymes with 2022.
Does a record with oil down predict anything? Fifteen episodes say: only through crude
I ran the mechanical version to make sure the 2022 analogue is not a story I picked because it fit. Since 1999 there have been 202 trading days on which XLE closed within 3% of its 52-week high while Brent sat 15% or more below its own — 3.0% of all days. Cluster qualifying days within 63 sessions of each other and the record holds 15 episodes, 13 of them old enough to grade a full year. The two most recent, December 2025 and May 2026, are not.
Over the following year XLE returned a median +11.9% with 8 of 13 positive, against a baseline of +8.6% and 61% positive for any day since 2000. The setup, by itself, is close to noise. What it is not is a top signal. The spread of outcomes is decided by crude’s next move: in the 7 episodes where Brent was higher a year later, XLE’s median was +27.5%; in the 6 where Brent was lower, -3.8%. The three most recent completed cases (March 2022, October 2022, November 2024) all fell in the second group.
What would change my mind
My position is that energy equities at a record are a bet on a 99.9th-percentile refining margin persisting, and that the 2022 path — margin reverts, equities follow it, crude is a sideshow — is the base case. Three markers settle it, each on a page that updates daily. First, the crack spread’s 200-day average: if it turns down and the front spread trades below $50, the refiners lose the thing they are priced on, and I would expect XLE to give back its post-April gain regardless of what Brent does; watch it on the crack spread page. Second, Brent itself: if it reclaims its 200-day average near $85 and then $110, the producers and majors re-engage and this becomes the 7-of-13 branch of the study, where the median next year was +27.5%. That is the outcome that proves my margin framing too narrow. Third, the test that would prove me wrong outright: XLE still above 63.75 in February 2027 with the crack below $45. That would mean the sector is being priced on something this page has not measured, such as capital discipline, buybacks or a reserve story, and the margin framing would be wrong.
How we checked it
XLE’s record is taken from its split-adjusted daily closes since 1999; the fund had a 2-for-1 split, so the 2014 peak of 49.75 on this page is half the nominal price of the time and the comparison is like-for-like. Brent is the FRED spot series (DCOILBRENTEU), which publishes two business days behind the equity close, so every XLE-versus-Brent figure uses the latest Brent print on or before the equity date. The crack spread is our own computation, (2 × NY Harbor gasoline × 42 + NY Harbor diesel × 42 − 3 × WTI) ÷ 3, with the 200-day average from the same file. Percentiles are simply where a reading ranks in its own history.
The record-state study counts a day when XLE closes within 3% of its trailing 252-session high and Brent closes 15% or more below its own; qualifying days within 63 sessions of each other merge into one episode anchored on the first, and forward returns are graded only where the full window exists. The baseline samples every fifth trading day from 2000 onward. The sub-group scoreboard uses seventeen large constituents grouped by business and shows price returns without dividends or index weights. It explains who moved; how much of the fund each move represents is a separate question. Both frozen studies end August 20, 2026; only the exhibits labeled live update.
Frequently asked questions
Why are energy stocks going up when oil is going down?
Because the part of the sector that earns more when crude falls is doing the work. Since Brent peaked on April 7, 2026, the three large refiners in the energy ETF have gained 36–46% while the integrated majors and the producers are within a few percent of flat. Refiners sell gasoline and diesel and buy crude, so a cheaper barrel with firm product prices widens their margin. The 3-2-1 crack spread, the standard measure of that margin, sits at the 99.9th percentile of its 20-year history, and its 200-day average is the highest on record.
Do energy stocks follow the oil price?
Over long stretches, yes: the daily returns of XLE and Brent have a correlation of about 0.41 since 1999, and the sector rarely sustains a record while crude collapses. Over weeks and months the link is loose. Energy equities price a slower-moving number than spot crude, closer to the 200-day average and the futures strip than the front month, so they lagged the spring spike on the way up and had less to give back on the way down.
What happened the last time energy stocks hit a high with oil well below its peak?
November 15, 2022. XLE closed at its 2022 high with Brent at $94, 29% below its March peak, and the 3-2-1 crack spread at $60. Over the next twelve months the crack fell to $29, Brent drifted to $78, and XLE lost 11.8%, with an 18.6% drawdown at the May 2023 low. The equities followed the margin down while the barrel barely moved.
Is a record high in XLE a sell signal?
Not by itself. Across the 13 completed episodes since 1999 in which XLE closed within 3% of a 52-week high while Brent sat 15% or more below its own, the sector’s median return over the following year was +11.9% with 8 of 13 positive, against a +8.6% baseline. What decided the outcome was crude’s next move: a median gain near +27% when Brent rose over the following year and a median loss when it fell. The setup describes the market; it does not time it.
What is the 3-2-1 crack spread and why does it matter here?
It is the gross margin of a simple refinery: the value of two barrels of gasoline plus one of diesel minus the cost of three barrels of crude, quoted per barrel. At publication it was $71.33 against a 20-year record of $75.89 set in May 2022. Because refiners are the one energy sub-group whose profits rise when crude falls, the crack spread is the number the current rally depends on, well ahead of the oil price.