In Part 1 of this series, I laid out some of the surprising facts in the EIA data following the US/Israeli strikes on Iran. In particular, I noted that US crude production didn't increase (it actually edged slightly lower), and that the surge and retreat in crude exports during the week ending 3/13 suggested something more complicated than a simple “high global prices pulled US barrels overseas.” In this follow-up, I'll offer what I think is the fuller story, aided by an expanded version of the table from Part 1 that now covers five weeks of data and includes crude oil prices and gasoline trade flows.
The Updated Table
I concluded Part 1 with a comprehensive table showing that the actual flows of crude oil and petroleum products was more nuanced than “armchair economist” logic would suggest. Yet in reproducing the table below, I’ve made some additions: First, I included another week of data (now ending March 27). But I also included the spot prices of West Texas Intermediate (WTI) and Brent crude (measured at mid-week, to better represent the price signals that market participants were actually responding to during each week).
I also included a new section on motor gasoline production, imports, exports, and net exports, to get a firm handle on the interaction between global markets and US gasoline availability.
A few things pop out immediately from the expanded table. First, WTI crude rose 47% in five weeks, from $65/bbl to a peak of $96. Second—and we will return to this issue later in the post—Brent crude rose even faster, hitting $118/bbl by the week of March 20, producing a Brent-WTI spread of $22. (Under normal conditions that spread is $3–6.) Third, and directly relevant to my original, glib explanation for rising US prices at the pump, gasoline exports out of the US did not increase after the attacks; they actually fell, from 1,067 kb/d pre-crisis to as low as 829 kb/d by the week of March 27.
Why Didn’t US Production Rise?
As a good armchair economist, I would have expected that a 50% price spike would cause US oil producers to ramp up their pumping rates. So why does the table above show domestic crude production as essentially flat (even slightly falling), week after week?
The answer is one of timing. I still believe that “supply curves slope upward,” meaning that a higher price would elicit a larger quantity of supply brought to market. But in the oil business, the relevant time horizon to see such a response is in months, not weeks, because it’s harder to turn the dial up than down. Specifically, once a specific oil well is up and running, it produces at rates determined by geology and engineering, not short-run market prices. A producing well is essentially running at its technically optimal rate, dictated by reservoir pressure, wellbore design, and surface equipment.
You don't “turn up” an oil well the way you can push a refinery from 89% to 93% utilization. Notice in the table that refineries did exactly that—crude inputs to refineries rose from 15,841 kb/d to a peak of 16,598 kb/d, an increase of about 5% in three weeks, because ramping up refinery throughput is an operational decision that can be made quickly.
Increasing crude production, by contrast, is a capital investment decision with a long lag. In US shale, that cycle takes roughly three to six months at minimum. And that’s assuming producers are convinced the price increase is durable. A conflict where a ceasefire is actively being negotiated is a scenario where a rational producer hesitates. It is risky to commit capital to a new well that takes six months to start producing, in a scenario where prices might crash back to $65 if the Strait of Hormuz reopens by the summer.
(On the other hand, if the global price of oil crashes, then the owners of existing wells—particularly ones that were just barely breaking even—can certainly dial down the rate of extraction. This effectively holds barrels off the market, and carries them forward to future years when the global price might have recovered. There is thus a fundamental asymmetry, where the rate of crude extraction can be slowed down very quickly, while speeding it up is technically difficult.)
Why Didn't the Brent-WTI Spread Get Arbitraged Away?
This is the more interesting puzzle, and it gets to the heart of why my original podcast explanation was incomplete. With Brent at $118 and WTI at $96—a $22 spread versus the normal $5—you would expect US exporters to be shipping every available barrel overseas and pocketing the difference. Instead, as the table showed, crude exports actually fell from their pre-crisis level.
Several real-world frictions explain this. First, the same conflict that closed the Strait of Hormuz also disrupted global tanker markets more broadly. War-risk insurance premiums spiked across the region, and shipping companies rerouted vessels away from the Gulf. Getting a barrel from the US Gulf Coast to a Brent-priced destination in Europe or Asia became significantly more expensive, directly eating into the ostensible arbitrage profit.
Second, most crude export volumes are sold on term contracts, not spot. A refinery in Rotterdam that contracted for US crude three months ago is still receiving those barrels on schedule. The $22 spread is real but only accessible on the marginal barrel—the incremental volume above existing contracts—which is a much smaller share of total trade than the headline volumes suggest.
Third, and perhaps most important for understanding the table: US refineries were running hard and competing aggressively for domestic crude. As noted above, refinery inputs rose by 757 kb/d over the period. That domestic demand for crude partially offset the incentive to export it. This is likely part of why WTI rose substantially even while Brent outpaced it: the domestic refinery system was bidding up the price of the crude that would otherwise have been available for export.
So the original logic I laid out in the podcast episode was correct, insofar as it goes: A higher global price of oil would tend to attract more US exports of crude, leaving less available for the domestic population. But what was also happening as soon as war broke out, was a surge in demand for crude (and downstream products) in the US as well. That’s why the WTI benchmark exploded too; the high US price was necessary to keep the crude oil at home, in spite of the spiking global price amidst the Iranian conflict.
So What Did Actually Happen to Gasoline Prices?
Let me return to the original question that started this whole investigation: why did US gasoline prices rise, given that America gets very little crude directly from the Persian Gulf?
My original answer on the podcast, namely that high global prices pulled US crude barrels overseas (and thereby leaving less for American refiners), was not what happened, as I’ve explained above. Rather than having to make do with less oil (as I had thought), instead US refinery runs increased. US gasoline production increased after war broke out, hitting 9,888 kb/d in the week of March 6. (Gasoline imports ticked up modestly as well.) And yet even though American refiners were cranking out above-average runs of gasoline, and even though our net exports of gasoline also fell (compared to the pre-crisis level), even so, the table shows that “gasoline stocks” fell steadily from 253 MMbbl to 241 MMbbl over the five weeks, a drawdown of 12.2 million barrels. How can this be? If production of gasoline is up, exports of it are down, and the inventories of gasoline are also falling…where is all the extra gas going?
The explanation is that the government “gasoline stocks” measurement refers to formal stockpiles in the “upstream” portion of the industry, such as refiners, terminals, and pipelines. But the official inventory numbers (at this aggregate level) would miss a wave of precautionary inventory building downstream the supply chain: gas stations topping up their underground storage tanks, trucking fleets filling their depot tanks, and motorists rushing to fill the family SUV even though it had been sitting at a half tank when news of war broke out. (Although the data don’t explicitly show it, I’m guessing that the surge in demand for gasoline was not because Americans decided to drive more, once war broke out.)
The Bottom Line
The updated data in the table tell a story along these lines: When war broke out and interrupted the flow of crude oil out of the Persian Gulf, the price of Brent soared. But at the same time, downstream US demand for gasoline (presumably to bulk up inventories) soared as well. US refiners thus ran hot, rushing to fill this precautionary demand. This pushed up the WTI price of oil, keeping it at home. (Other problems with international shipping allowed crude to stay in the US without rising fully to the Brent price.)
If my updated explanation is accurate, and if the Strait remains constrained, we should expect to see increased US crude production by the fall. Sooner than that, we should expect to see either a surge (compared to pre-crisis levels) in gasoline exports, as the surplus in US production (relative to motorist consumption) could turn its focus away from US inventories and instead flow to the gasoline-starved world or we should expect US refineries to return to pre-crisis production rates, settling in to the new normal of less crude to go around.
If my updated framework is essentially correct, what can we say about the social function of the precautionary inventory additions by US companies? Ironically, the “price gouging” story could fit the facts in the table loosely enough. After all, one might interpret the numbers as showing greedy middlemen “holding gas off the market” and thereby driving up the price, all under the cover of war panic.
However, even if such a cynical view accurately described the motivations of the businesspeople in question, it wouldn’t negate the validity of my original discussion: In a situation such as ours, where crude oil and its products might be far more scarce in 6 months than the present, we want social mechanisms to encourage motorists to cut down on their driving, and for the major players to “hoard” gasoline.
Only when it is clear that the danger has passed (perhaps with a solid ceasefire agreement), would it make sense for various firms to allow their stockpiles to shrink to more normal levels. And notice, that is exactly the response that our profit-and-loss capitalist system rewards.
Dr. Robert P. Murphy is the Chief Economist at infineo, bridging together the dependability of Whole Life insurance policies with the benefits of blockchain-based finance.
Twitter: @infineogroup, @BobMurphyEcon
Linkedin: infineo group, Robert Murphy
Youtube: infineo group
To learn more about infineo, please visit the infineo website.