Supply Flexibility in the Shale Patch: Evidence from North Dakota

Abstract

We analyse if supply flexibility in oil production depends on the extraction technology. In particular, we ask to what extent shale oil producers respond to price incentives by changing completion of new wells as well as oil production from completed wells. Using a novel well-level monthly production data set covering more than 15,000 crude oil wells in North Dakota, we find large differences in response between conventional and unconventional (shale) extraction technology: While shale oil wells respond significantly to spot future spreads by changing both well completion and crude oil production, conventional wells do not. Our results suggest that firms using shale oil technology are more flexible in allocating output intertemporally. We interpret such output pattern of shale oil wells to be consistent with the Hotelling theory of optimal extraction. JEL-codes: C33, L71, Q31, Q40

The effect of oil supply shocks on US economic activity: What have we learned?

“…Nevertheless, recent work by Baumeister and Hamilton () has prompted academics and policymakers to revisit the way prior information is used to conduct inference in structural vector autoregressive models of the oil market. Indeed, they obtain estimates of the effect of oil supply shocks that are considerably larger than suggested by earlier studies (e.g., Kilian, ; Kilian & Murphy, , ) and larger than any microeconometric estimates of that elasticity (e.g., Anderson, Kellogg, & Salant, ; Bjørnland, Nordvik, & Rohrer, ; Newell & Prest, ). Hence our interest in analyzing how alternative modeling assumptions map into different conclusions regarding the importance of oil supply shocks.…”

“…Both the economic theory in Anderson et al () and extraneous microeconometric estimates argue against priors with large probability mass on higher values of the price elasticity of oil supply. Estimates based only on data from North Dakota by Bjørnland et al () imply a slightly higher global price elasticity of 0.04, whereas those of Newell and Prest () imply an elasticity of 0.01. The highest plausible bound is distinctly smaller than the posterior elasticity estimate of 0.15 reported by Baumeister and Hamilton ().…”

“…To further investigate the role of the prior on the short‐run price elasticity of oil supply, α qp , we reestimated BH19 after truncating the prior on α qp in a manner that shifted probability mass closer to zero. Specifically, we first truncated the prior at the 0.0258 bound imposed in K12 and KM14; we then allowed for larger bounds motivated by recent microeconomic studies (e.g., Bjørnland et al, ). We found that using a tighter prior on α qp , which rules out the possibility that the elasticity could be very large, considerably reduces the differences between the estimated short‐run price elasticity of supply obtained in BH19 and KM14, yet it considerably increases the estimate of the elasticity of demand in production for BH19.…”

“…Anderson et al () find that oil production from existing wells in Texas does not respond to changes in oil production. Using data for North Dakota, Bjørnland et al () estimate an elasticity of 0.035 for conventional wells and 0.041 for shale oil wells.…”

“…Motivated by the estimates in Bjørnland et al (), we evaluate the sensitivity of the results to a prior for α qp , where we consider an upper bound of 0.04 on the supply elasticity ( α qp ∈(0,0.04]). This bound is justified as follows.…”