Oil Prices in Real Terms, 1950 to 2026

Crude oil, nominal and real — five-year steps to 2005, every year after

Domestic first purchase price. The real column is the Institute's computation; the nominal and CPI columns are both primary and are printed so that any reader can redo the arithmetic.

Nominal source: EIA Monthly Energy Review, August 2026, Table 9.1 (Crude Oil Price Summary), page 171 — Crude Oil Domestic First Purchase Price. Deflator source: BLS historical CPI-U table, U.S. city average, all items, annual averages (historical-cpi-u-202402.pdf). Base month July 2026 at a CPI-U of 333.918, from BLS Consumer Price Index for All Urban Consumers (CPI-U), U.S. city average, all items, 1982–84=100, not seasonally adjusted; Table 1 of the July 2026 CPI news release, last modified August 12, 2026. The deflator column is printed so the real column is not something a reader has to take on trust: multiply the nominal figure by it and the real figure follows.

A note on which nominal series this is, because it is not the one on a screen. The domestic first purchase price is what producers received at the wellhead for domestic crude, averaged across a year. It is not WTI, not Brent, and not a spot quote, and it runs below the benchmark headline in most years. It is used here because it is the longest continuous United States series a public agency publishes, which is the only way to reach 1950 at all. Do not set a figure from this table against a spot benchmark taken from another one.

Where the real peak actually sits, and it is not where the nominal peak sits

The highest nominal annual figure above is $95.99 in 2013. The highest real figure is $142.11, in 2011 — a year whose nominal price of $95.73 looks unremarkable beside its neighbours. The lowest real figure in the series is $27.37 in 1970, against a nominal price of $3.18. The real range across seventy-three years is therefore about 5.2 to one, while the nominal range is about 30.2 to one. Inflation supplies the difference between those two ratios, and it is not a small difference.

The 1980 row is the one most often mis-stated. Its nominal $21.59 is remembered as the great oil-shock price, and in July 2026 dollars it is $87.49 — below several years of the 2010s. The 1979–81 shock was enormous as a change and as an economic event; as a price level in today's money it was not the peak. Anyone reasoning about the present from 1980 as the historical high-water mark is reasoning from a number that has been beaten more than once since.

Set the current episode against that. The highest monthly first-purchase price in the 2026 series is $104.88 in May 2026, which is 73.8% of the 2011 real peak. That is a serious price. It is not an unprecedented one, and a page that told a reader otherwise would be wrong on its own arithmetic.

2024, 2025 and 2026 — nominal only, and the reason is stated

These rows carry no real column. That is a limitation of the deflator, not an oversight.

The real column stops at 2023 and the 2024, 2025 and 2026 rows carry nominal figures only. The reason is stated rather than papered over: the BLS historical annual-average table the Institute holds ends at 2023, and an annual average CPI for a year is not available until that year is complete. The Institute does not interpolate a deflator. For the 2026 rows the distinction is close to immaterial — the base month is July 2026, June 2026 CPI-U was 333.952 against July's 333.918, a difference of one hundredth of one per cent — so a 2026 nominal figure is within a rounding error of its own real value on this base. For 2024 and 2025 it is not immaterial, and those two rows should be read as nominal.

Both tables run most recent first, so the monthly rows are a sequence read upward from the bottom rather than a set of levels to be scanned. January 2026 at $58.44 was the low of the year and the month before the disruption; May 2026 at $104.88 is the high. That is a rise of $46.44 a barrel, or 79.5%, in three months, on a series measuring what refiners actually paid rather than what a screen showed. The June 2026 row is blank because the source edition does not carry it, and a blank is what an unavailable month looks like here.

The 2008 records, deflated — and three different kinds of measurement

These three figures are quoted interchangeably in commentary. They are not interchangeable.

Deflated on the July 2008 CPI-U of 219.964 against a base of 333.918 for July 2026, a factor of 1.518058. All three records were set in the same month, so one factor applies to all three.

The settlement record — $145.29 on July 3, 2008 — becomes $220.56 in July 2026 dollars. That is the figure a reader needs in order to place the present episode, and it is roughly 2.10 times the highest monthly first-purchase price of 2026. Anyone describing the current market as comparable to 2008 in price terms is out by a factor of more than two.

The distinction between the first two rows is not a technicality and it recurs constantly. A settlement is the official daily close of the front-month contract and is the figure a clearing house uses. An intraday high is a price at which some quantity traded during a session, which may have been one lot for one minute. Both are real; they measure different things; and a comparison taking an intraday figure from one episode and a settlement from another has manufactured a difference out of the convention rather than out of the market.

The pump records — and one of them survives deflation

This is the section where the arithmetic changes the story being told this week.

The prior nominal high on the weekly all-types diesel series (EMD_EPD2D_PTE_NUS_DPG, which begins March 1994) was $5.810 in the week ended June 20, 2022. EIA changed the methodology of the diesel series on June 13, 2022 — before the comparator week — so the two readings sit on the same side of the break and the comparison is clean. This is a nominal record only; see the real-versus-nominal ledger, where it is not one.

What the deflator's own energy components say about this episode

Published July 2026. Read the last two rows against the middle four.

Source: BLS CPI news release for July 2026, released August 12, 2026; CPI-U, U.S. city average, not seasonally adjusted. The all-items row carries a twelve-month change and no index level, because the level of the all-items index is the base of every real figure on this page and is printed in the deflator note above rather than twice.

Energy rose 14.7% over twelve months against an all-items rate of 3.4% — a factor of 4.3. Inside energy the split is sharper than the aggregate: fuel oil at 39.1% and motor fuel in the twenties, against electricity at 4.2%. Electricity and piped gas have barely moved.

That contrast is a mechanism, not a curiosity. A liquid fuel is priced off a world market that clears daily and passes a supply shock through to a pump within weeks. Electricity is a regulated network good priced off a fuel mix in which oil is nearly absent, through rate cases that take months or years. So an oil disruption appears in the energy-commodities line almost immediately and in the electricity line hardly at all, and a reader tracking “energy inflation” as one number will mistake a liquid-fuels event for a general one. It is the same distinction the barrel-to-gallon ledger draws between the cost of the barrel and the price at the pump.

Where the real column stops, and what the Institute will not do about it

The real column stops at 2023 and the 2024, 2025 and 2026 rows carry nominal figures only. The reason is stated rather than papered over: the BLS historical annual-average table the Institute holds ends at 2023, and an annual average CPI for a year is not available until that year is complete. The Institute does not interpolate a deflator. For the 2026 rows the distinction is close to immaterial — the base month is July 2026, June 2026 CPI-U was 333.952 against July's 333.918, a difference of one hundredth of one per cent — so a 2026 nominal figure is within a rounding error of its own real value on this base. For 2024 and 2025 it is not immaterial, and those two rows should be read as nominal.

The rest of the reference