The price elasticity of US shale oil supply: Insights from a natural experiment

The 2026 Iran war provides a natural experiment that allows model-free estimation of the price elasticity of US shale oil supply. This column argues that the monthly and quarterly elasticities are effectively zero. This finding is supported by data on well completion times, survey data, and economic theory. The analysis contradicts higher oil supply elasticity estimates from recently proposed non-standard panel regression specifications, which attribute the supply response to shale oil producers strategically reducing inventories of drilled but uncompleted wells. The findings provide indirect support for the view that oil prices are mainly driven by demand, rather than supply, shocks.

When the Iran war broke out in 2026, some market observers and economists expected US shale oil production to accelerate substantially in response to higher global oil prices. A common view was that if any oil producers would be able to respond quickly to this price signal, it would be US shale oil producers, who operate with much shorter production cycles than other producers (e.g. Bornstein et al. 2021). Such a production response would have provided much-needed relief at a time when crude oil became increasingly scarce in the world. 

In this column, we analyse how much US production of shale oil responded to higher oil prices in the first month of the war and in the first quarter including the months from March 2026 to May 2026. A common measure of the responsiveness of oil production to price signals is the price elasticity of oil supply, defined as the percent change in US oil production relative to an exogenous percent change in the price of oil over the same time period induced either by a shock to the global demand for oil or a shock to the supply of oil elsewhere in the world raising demand for US oil (see Kilian (2022) for an in-depth review of the oil supply elasticity definition and earlier estimates). The geopolitical oil supply disruption associated with the 2026 Iran War provides a clean example of such an event. It not only caused an exogenous and large surge in global oil prices, but it demonstrably raised the demand for US oil from Asia. We focus on how much US shale oil production responded to this price signal.

The left column of Table 1 relates the percent change in US shale oil production in March to the observed percent increase in the average West Texas Intermediate (WTI) spot price in March. The implied ratio is effectively zero, as is the estimate of the corresponding quarterly elasticity.

Table 1 Monthly and quarterly price elasticity of supply during the 2026 Iran war

Why this elasticity estimate makes economic sense

Our elasticity estimates are not only consistent with theoretical arguments in Anderson et al. (2018), who show that the short-run oil supply elasticity is zero if adjusting oil production is costly, as is inevitably the case in practice, but also with evidence from industry sources. 

One piece of evidence is from the Dallas Fed Energy Survey, which, in the second quarter of 2022, asked oil company executives directly how many months it would take them to raise production in response to an unexpected increase in the price of oil. The survey responses are fully consistent with the one-month price elasticity of oil supply from new wells being zero. In fact, the modal response is four to six months and more than half of firms require more than half a year to complete a new well.

A second piece of evidence comes from data on US well completion times provided by the data analytics firm Kayrros. Our sample includes 21,299 wells from all US oil basins and their characteristics in 2023-24, which are the most recent data currently available. The completion time of a new well is the time it takes from the rig arriving at the well site to the oil well starting to produce. The median duration between the rig arrival and the well starting to produce is 5.7 months. Only 0.8% of wells in this sample were completed in less than one month. This evidence as well immediately calls into question the view that the one-month price elasticity of oil supply from new wells is non-negligible, as there typically is an additional one- or two-month lead time to prepare the drilling site that is not included in the completion time. 

Even if we focus on the subset of wells that are already drilled, but uncompleted (known as DUCs), which tend to have shorter completion times, there is little support for a much larger supply elasticity. The Kayrros data show that the median duration between the arrival of a completion crew at the DUC site and the completion of the well is 1.4 months with a maximum of eight months. Only 10% of DUC completions occur within three weeks, and there are indications that this percentage is substantially inflated by measurement error. Considering the additional time it takes to arrange for a completion crew, these data cast doubt on the view that DUCs could explain a one-month price elasticity of shale oil production much larger than zero, all the more so as there were only 2,357 DUCs nationwide in February 2026, most of which were part of the working inventory of producers and not available for immediate completion. 

Finally, there is evidence that one should not have expected the price elasticity of oil supply to be any larger even at the annual horizon. Given the time it takes to complete a well, as of the end of March 2026, shale oil producers would have looked at oil futures prices for delivery in 2027 when deciding whether to drill and complete a new well. Figure 1 illustrates that at that horizon WTI futures prices were close to the $67 per barrel price prevailing right before the war broke out, making it impossible to lock in higher oil prices and thus providing little incentive for production increases. Thus, one would expect a near-zero elasticity not only at the one-month and one-quarter horizons, but throughout the year of 2026 and into 2027.

Figure 1 WTI futures curve, 31 March 2026

Source: Bloomberg.

What about alternative much larger shale oil supply estimates recently reported in the literature?

While our finding of a zero quarterly price elasticity of shale oil supply is supported by the panel regression estimate in Newell and Prest (2019), it is at odds with some more recent findings in the literature based on non-standard panel regressions that allow for an additional effect on oil production from the three-month oil futures spread. For example, Bjørnland et al. (2021) reported a one-month price elasticity of shale oil supply from North Dakota as high as 0.9, while Aastveit et al. (2026) reported an estimate of 0.62 based on a more comprehensive data set. Their higher elasticity estimates are entirely driven by the coefficient on the oil futures spread in the regression. Their interpretation of this coefficient is that shale oil producers reduce DUC inventories strategically, as the futures price exceeds the spot price, creating a supply response missed by standard regressions.

It is readily apparent that this interpretation cannot be valid. There simply are not enough DUCs available for completion to generate a price elasticity of oil supply of 0.62. Under plausible assumptions, this would have taken the completion of more than 12,000 DUCs in record time in March 2026. There were only 2,357 DUCs nationwide in February 2026, according to Kayrros, so this is plainly impossible. While one could debate the details of this computation, the fact is that the number of DUCs slightly increased from 2,357 in February 2026 to 2,381 in March, rather than falling sharply, as predicted by Aastveit et al. (2026). Nor does it seem plausible that a large number of DUCs could be completed merely in weeks, given the evidence on completion times we discussed earlier.

A partial explanation of this puzzle is that the 0.62 elasticity estimate reported in Aastveit et al. (2026) refers not to a point estimate of the supply elasticity, but to an upper bound. The underlying regression implies an elasticity value anywhere between 0 and 0.62, depending on the value of the three-month futures spread, rather than a point estimate of the average elasticity or, for that matter, the elasticity in March 2026. Approximating the elasticity for March 2026 using the derivations in Aastveit et al. (2026) suggests a one-month shale oil supply elasticity of only 0.12 rather than 0.62. Even this much smaller elasticity is still six times larger than our estimate, however, and still implausible. Given an elasticity of 0.12, it would have taken the completion of about 2,400 DUCs within two weeks, which does not seem credible given our earlier evidence.

Concluding remarks

A natural question is why researchers care so much about the one-month price elasticity of shale oil supply beyond the immediate interest in the oil supply response to the 2026 geopolitical oil supply shortfall. The reason is that the US price elasticity of oil supply effectively determines the global price elasticity, which can be expressed as a weighted average of the supply elasticities for US shale oil and conventional crude oil anywhere in the world. 

The global oil supply elasticity is an important parameter in structural vector autoregression (VAR) models of the global oil market. It determines whether the price of oil is primarily driven by shocks to global oil demand or global oil supply. For example, it has been shown in the literature that – if this elasticity is close to zero – the price of oil is primarily driven by oil demand shocks. Our analysis provides indirect support for this view.

It should be noted that even if the shale oil supply elasticity were much larger, this would not have invalidated earlier findings in the literature based on structural models that impose supply elasticities close to zero. The reason is that these studies were based on data that pre-date the US shale oil era. Our evidence that the global oil supply elasticity has remained close to zero even after the surge in US shale oil production suggests that models embodying tight bounds on the global oil supply elasticity remain relevant even today and that models that do not impose such bounds should be viewed with caution.

Source : VOXeu