Fracking and Structural Shifts in Oil Supply.

AuthorWalls, W.D.
  1. INTRODUCTION

    Oil producers in North Dakota, Texas, and a few other regions have helped to catapult America to its position of prominence as the world's leading producer of crude oil today. It is apparent that the level of production in America has risen rapidly since the adoption of horizontal drilling and hydraulic fracturing (fracking) production techniques (see Figure 1). Clearly the oil supply curve has shifted outward. What is not readily apparent from descriptive statistics on U.S. oil production is how the adoption of these production methods has changed the structure of oil supply in the U.S. "The peculiar features of the shale revolution have altered the nature of supply in critical ways. Not only are entirely new sources technologically viable, they can be brought online in tiny increments--the cost of a productive shale well is three orders of magnitude smaller than Arctic or deep-sea projects" (Dimitropoulos and Yatchew, 2018, pp. 683-684). (1) Recent work by Anderson et al. (2018) and Newell and Prest (2019) demonstrates--using theoretical and empirical evidence-that the primary margin of adjustment of oil supply to price changes is through increased drilling activity and production from new wells and not by additional production from existing wells. The analysis in this paper drills down a little deeper into this way of thinking by examining how oil supply may differ across regions using conventional technologies and regions using hydraulic fracturing production methods.

    There is much anecdotal evidence about how technological innovation may have changed the cost structure of producing crude oil, particularly in tight oil formations where conventional production techniques were not economically feasible. (2) But a large part of the reason that hydraulic fracturing has shifted oil supply is due to the reduction in uncertainty surrounding crude oil exploration and production using this technology. Consider the discoveries of new additions to proved reserves by three channels: (i) extension of existing fields through enlargement of the production reservoir area; (ii) new reservoir discoveries in old fields; (iii) new field discoveries. The actual expansion of reserves for each channel is plotted in Figure 2. Among these three categories, new reservoir discoveries in old fields is the most uncertain outcome because it requires costly deep drilling in the search for new reservoirs. Extension is the least risky, particularly with innovations in horizontal drilling and hydraulic fracturing. From 1990 to 2008, the composition of reserves expansion by extension, new field discoveries, and new reservoir discoveries in old fields was 61%, 23% and 16%, respectively. However, since 2008 the composition of new reserves additions has changed to 87%, 7%, and 6%, respectively. The increasing contribution of extensions to new proved reserves has been driven primarily by horizontal drilling and the adoption of hydraulic fracturing. As a result, the search for new proved reserves via extension has become far less risky. (3) Newell et al. (2019) convincingly reason that conventional oil and gas investments resemble "trophy hunting," with high risks compensated by high rewards; however, modern unconventional extraction resembles a "manufacturing process" in which operators are much more certain of their production prospects.

    In this paper, we provide econometric estimates of the supply relation for groupings of U.S. oil producers that differ in their use of hydraulic fracturing. We find that tracking is associated with i) supply responses that are asymmetric with respect to price increases and decreases; ii) a much larger supply response with respect to price rises than is the case for non-fracking producing regions; and iii) a faster speed of adjustment to price changes. Because shale oils account for about one-half of U.S. oil output, these attributes of supply can have important implications for the world oil market, particularly if they are the marginal producers. First, price increases cause a rapid increase in tight oil output, making tracking oil producers a primary beneficiary of price increases. Second, price decreases cause a much smaller decrease in tight oil output; even in a falling price environment, the decrease of tight oil output is limited. (4) The features of post-shale-boom U.S. oil supply are helpful in explaining the impotence of recent OPEC policy actions intended to elevate and stabilize the world price of crude oil after causing shale operators to exit the industry.

    In the following section we set out a straightforward oil supply model that incorporates i) the possibility of asymmetries with respect to price; ii) partial adjustment of supplier behavior over time; and iii) endogenous structural change in the supply relation. In Section 3 we estimate the oil supply model as a fixed-effects panel with a possible (endogenously determined) structural break for each region where changes in production have been driven mainly by hydraulic fracturing--North Dakota, Texas, Colorado, and New Mexico--and in all other regions as a whole; short-run production elasticities, long-run elasticities, and speeds-of-adjustment are calculated from the estimated supply relations. We then report a number of refinements and extensions to the baseline oil supply model in Section 4. We show that firms' financial management in practice is consistent with our empirical results and discuss how the asymmetry of marginal oil supplies may have severely limited OPEC's ability to manage the world oil market in Section 5. Final conclusions are summarized in Section 6.

  2. OIL SUPPLY WITH STRUCTURAL CHANGE, ASYMMETRIES, AND PARTIAL ADJUSTMENT

    Table I reports and summarizes estimates of crude oil supply elasticities reported in the scholarly literature. Of the numerous studies listed, only three yield non-negative estimates for short-run supply elasticity: 1) the model used by Hogan (1989) that was estimated on aggregate data, 2) the model of Rao (2018) estimated on California well-level data, and 3) the model of Newell and Prest (2019) for unconventional production estimated on well-level data from five U.S. States. Estimates of the elasticity of crude oil reserves with respect to price are uniformly positive.

    The estimates vary starkly among the different studies of oil supply. A particularly interesting finding in the Newell and Prest (2019) model is that well-level unconventional oil supply has a positive supply elasticity while the elasticity for conventional oil wells is close to zero. This finding is consistent with the observation of Anderson et al. (2018) that the main margin of adjustment for oil supply is drilling activity and not additional production from existing wells. (5) For the technological reasons discussed above in Section 1, we expect the unconventional oil producing regions to have a supply structure that differs from the conventional oil producing regions.

    In the following sections, we propose an oil supply modelling framework that 1) captures differences in the oil supply relation between fracking and non-fracking producing regions; 2) allows for the possibility of a structural break in the oil supply relation (motivated by the boom in shale oil production that began in 2008 as shown in Figure 1 above); and 3) allows for asymmetries in supply responses to increases and decreases in the price of oil.

    2.1 Decomposing Price Changes

    Gately and Huntington (2002) suggest a decomposition of price changes to estimate the asymmetric effects of price increases and decreases.'' Following the approach of Gately and Huntington (2002), we decompose crude oil first-purchase price {p,) in each period using the equation (1) below,

    [Please download the PDF to view the mathematical expression] (1)

    where ln(p,) is the logarithm of the real first purchase price of crude oil at time t; [p.sub.0] is the initial log price of crude oil; [prise.sub.1] denotes the cumulative increase (rises) of ln(p); and pet, represents the cumulative decreases (cuts) of \r\(p). The sample used for estimation spans January 1986-February 2019. (7) As an example of the price decomposition, the first purchase price for Texas is displayed in Figure 3.

    2.2 Oil Supply Model for Shale Regions

    With the price variable decomposed as set out above, we propose to estimate the price elasticity of supply for four shale-rich regions in the U.S.: North Dakota, Texas, Colorado, and New Mexico. (8) We explicitly account for structural change in supply behavior that may be associated with the adoption of hydraulic fracturing production methods. We propose a structural change model where the breakpoints are unknown and jointly estimated with the coefficients of the supply equation. (9) Specifically, consider a model with one unknown breakpoint as set out below:

    [Please download the PDF to view the mathematical expression](2)

    where ln(Q,) is the log of monthly production (barrels) at time t and In ([Q.sub.t-1]) is lagged production; t is a time trend and T is the unknown date when the structural break occurs. [I.sub.2] (t

    [Please download the PDF to view the mathematical expression] (3)

    [Please download the PDF to view the mathematical expression] (4)

    In model (2) there are two regimes. The estimated coefficients a, (3 and y differ between two regimes: [T.sub.0 ][less than or equal to]t

    2.3 Oil Supply Model: Non-Shale Producing Regions

    It is reasonable to consider whether the presence of a structural break was present in the non-shale producing regions of the U.S.; in these regions the reserves of shale oil are either absent or not fully developed. If the changes in supply behavior are driven mainly by fracking, we would not expect to find a similar structural change in the non-shale producing regions. To examine this possibility, we estimate a fixed-effect panel threshold model for all onshore production regions other than North Dakota, Texas...

To continue reading

Request your trial

VLEX uses login cookies to provide you with a better browsing experience. If you click on 'Accept' or continue browsing this site we consider that you accept our cookie policy. ACCEPT