Shale Oil Is Not A Ponzi Scheme: Evidence From Decline Curves

|
Includes: AMR, AREX, CDEV, CXO, EGRGF, ESTE, GSTC, HK, JAG, PXD, REN, ROSE, VNOM, WLL, XEC
by: Laurentian Research

Summary

More than two years into the oil upswing, shale oil developers are thriving, with many of them either already generating or on the way to generate positive free cash flow.

However, some investors still cling to the notion that shale developers run on an ever-faster treadmill in capital needs and thus operate a Ponzi scheme.

Here, I intend to use empirical hard data to demonstrate shale oil as a Ponzi scheme is a pure fallacy which can potentially blind investors from seeing opportunities in shale.

Shale oil is not a Ponzi scheme, as shown by the decline curves

The shale revolution was launched in 1997 and became operative in full throttle since the 2000s. It has greatly improved the energy independence of the U.S. (see here); because of record-setting shale oil production, the U.S. net oil imports, a measure of the domestic oil sufficiency, declined from over 12.5 MMbo/d in 2005 to 3.7 MMbo/d as of 2017 and are projected to drop further to 2.4 MMbo/d this year and to 1.6 MMbo/d next year (Fig. 1)(see here). The rising production of cheap shale gas and the resultant coal-to-gas conversion did the heavy lifting for greenhouse gas emission reduction in the U.S. (Fig. 2)(see here).

Fig. 1. U.S. net imports of crude oil and petroleum products (Mbbl/d). Data source.

Fig. 2. U.S. natural gas marketed production in Bcf. Data source.

In spite of the valuable societal benefits that are derived from the shale revolution, the business of shale exploitation got little respect. As the oil crash brought the industry to its knees, various financial media and environmentalist organizations talked down the shale developers in an air of palpable schadenfreude.

  • One of the more popular rhetorics presents a shale E&P project as a perverse black hole, gorging an ever larger amount of capital while producing little financial return. Operationally, the shale developer has to drill an increasing number of wells to stay ahead of the accelerating natural decline of production, as though he were jogging on a treadmill that goes ever faster (Fig. 3). Don't feel sorry for the shale oil industry, opined the conspiracy theorists, because it is running the biggest Ponzi schemes known to mankind, pumping in cheap money in billions to inflate the acreage valuation so as to create a mirage of financial viability for a time, only to enrich the select few (see, e.g., here).

Fig. 3. Shale developers were said to be running on an accelerating treadmill. Cartoon credit.

I can understand the doubt about shale's economic viability when the oil price was down - which is part of human nature - but it surprises me the treadmill metaphor appears to have taken a life of its own even as the shale developers regain vigor as oil price rises. Investors who still believe the negative talks about shale oil may have missed out opportunities offered by a great number of shale developers.

In this article, I intend to introduce some salient facts which, as you will see, contradict with the false accusations against shale oil. Specifically, I will derive an average decline curve from the production data for wells drilled since 2009 in the Permian Basin, the most productive shale provinces in the world; I will then demonstrate, in the life of a shale development project, an ever decreasing number of new wells are actually needed to offset the natural decline thus to maintain a flat production profile; my argument is further strengthened by two additional factors, i.e., the impact of the oil price crash on the accounting profitability of E&P companies and the effect of capital efficiency gains and operating cost-cutting at shale operators.

The decline curve

Once the initial production ramp-up is done, production decline starts thanks to the changes in reservoir physiochemical properties and reservoir drive mechanisms induced by continuous extraction of fluids (oil, natural gas, and water) from the reservoir.

  • The decline curve for a particular reservoir can be derived from empirical data to show the pace at which production is expected to decline over the lifetime of that reservoir.
  • An aggregate decline curve can even be derived for multiple reservoirs. For an unconventional (shale) play, the aggregate decline curve reflects the declining production property for wells drilled in various areas of the play and across the various benches. The derivation of the aggregate decline curve is an especially pertinent procedure and economical practice for our particular purpose - to understand the production profile and economics of an entire shale play.

Knowing the decline curve can help us estimate the production profile, reserves, and present value of an asset, a reservoir, or a play.

The dataset
We built a dataset of aggregate average daily production rates for the Permian Basin from 2009 through May 2018, with data sourced here. The data separate oil (bo/d) and natural gas (Mcf/d) production, although we consider the total production (boe/d). Under each particular year of first oil, the data points are listed by the number of months since the well started flowing.

Table 1. The number of wells put on stream in each time period and the number of months in production for the dataset.

The number of wells had increased steadily from 50 in 2009 to 2,684 in 2014 before the oil price crash brought it down to 2,573 in 2015 and 2,215 in 2016. It started to rise again in 2017, totaling 3,364 in the year and projected to reach a record of around 3,600-4,000 (Table 1).

It is worth noting that data are not so "smooth" for the earliest years, e.g., 2009 and 2010, and particularly so for the earliest months of those years.

The decline type

There are three types of production decline, i.e.,

  • the hyperbolic decline, and
  • its two degenerated special cases, i.e., the exponential decline (aka, constant fractional decline) and the harmonic decline (Fig. 4)(see here).

Fig. 4. Three types of decline model. The vertical axis is the decline rate, while the horizontal axis the production, where q denotes production and t, time. Source.

The Permian production declines along a hyperbola (Fig. 5), which is confirmed by the monthly decline rate (i.e., -dq/(q dt)) plotted both against t (the month since in production flow) as depicted in Fig. 6 and against q (production rate) as illustrated in Fig. 7. The Permian wells clearly do not follow the exponential decline model, where the decline rates are constant, or the harmonic decline model, where the decline rates fall on a sloped straight line when plotted against the production rate.

Fig. 5. The monthly q vs. t plots for the Permian Basin in 2009-2017. q, production rate; t, months in production flow. The author's calculation and illustration based on data sourced here.

Fig. 6. The monthly dq/(qdt) vs. t plots for the Permian Basin in 2009-2017. q, production rate; t, months in flow. The author's calculation and illustration based on data sourced here.

Fig. 7. The monthly dq/(q dt) vs. q plots for the Permian Basin in 2009-2017. q, production rate; t, months in flow. The author's calculation and illustration based on data sourced here.

Parameters for the decline curve

The production rate q at time t in a hyperbolic decline is described by the following equation (see here),

where q0 is the production at the initial time, while a and b are shale play-specific constants.

  • q0 appears to be approaching 1,000 boe/d after having climbed an S-shaped trajectory (Fig. 8).

Fig. 8. The initial production rates, historical and projected. The author's calculation and illustration based on data sourced here.

  • Constants b/a and a are estimated to be 0.681 +/- 0.344 and 0.581 +/- 0.101, respectively. The monthly decline rates derived from these constants are depicted in Fig. 9.

Fig. 9. The monthly decline rates determined from the raw data using the hyperbolic decline equation. The author's calculation and illustration based on data sourced here.

Another approach, perhaps a simpler one, is to calculate the average yearly decline rate from the raw data. The error in the estimates increases for wells older than 5 years due to the limited number of wells drilled back then (Table 1; Fig. 10).

Fig. 10. The decline curve for the Permian Basin as determined from yearly average decline rates, shown with standard deviation. The author's calculation and illustration based on data sourced here.

Testing of the decline curves

From the same q0, the average initial production rate determined from the dataset, the production profile is calculated using the monthly decline rates derived above (Fig. 11).

Fig. 11. The backtesting of the decline curves derived from the hyperbolic equation (left) and the yearly average decline rates (right). The author's illustration.

The production profile merely presents the same raw data in a new format. It should not have any indicative value as to the accuracy of the decline rates.

The production profile derived from the estimated monthly decline rates retrodict the dataset very well in the first few years but as for the earlier years, the prediction tends to be higher than the raw data. Such a deviation is expected because the older wells are biased toward less productivity (Fig. 8). As more productive wells drilled in recent years age, this apparent deviation is anticipated to disappear. Overall, I am satisfied with the quality of the monthly decline rates derived above.

To maintain a flat production profile ...

Using the decline curve derived above, we can construct a production profile for, e.g., a shale project and a shale producer. Fig. 12 shows the production profile under the assumption that 100 wells are drilled and completed in each of the ten years. In this case, the production is found to grow at a CAGR of 8.1%.

Fig. 12. The production profile if 100 wells are to be completed in the first and each subsequent year, assuming that the initial production rate is 1,000 boe/d. The author's calculation.

Suppose again 100 wells are drilled and completed in the first year and further assume that, in each of the subsequent nine years, a fixed number (but fewer than 100) of wells are to be drilled and completed. We find that the production growth rate decreases, that a flat production profile can be achieved if 44 wells are drilled and completed in each of the nine years and that even if no wells are drilled in these nine subsequent years, production will only decrease by 22.1% per year (Table 2).

Table 2. The number of wells drilled in the 2nd-9th year and the corresponding CAGR of production growth from the high of the first year to the end of the ten-year period. Please note 100 wells are to be drilled in the first year.

The proverbial treadmill metaphor states that an increasing amount of capital needs to be plowed into a shale project just to keep the production flat. However, based on empirical data, our model clearly demonstrates the treadmill metaphor does not hold up:

  • Production will grow even if the level of investment in drilling and completion decrease moderately.
  • To maintain a flat production, the operator only needs to drill and complete 44% of the wells as in the first year, i.e., it can spend 56% less in each subsequent year than in the first.
  • Even if drilling is halted altogether, in a ten-year time span, the production only drops by 22% on average (although production declines faster in the first few years and slower later on).

Discussions

What about the hundreds of billion dollars lost during the oil downturn?

To answer that question properly, we have to review the SEC accounting rules governing how the E&P companies report their reserves. Simply put,

  • The SEC requires E&P companies to report the value of reserves as PV-10, i.e., the present value of the estimated future oil and gas revenues, reduced by direct expenses and discounted at an annual rate of 10%.
  • The SEC Final Rule of December 31, 2008, requires companies to estimate proved reserves using oil and natural gas prices based on the unweighted arithmetic average of the first-day-of-the-month price for each month within the 12-month period prior to the end of the reporting period (unless prices are defined by contractual arrangements, excluding escalations based upon future conditions).
  • The SEC Final Rule defines the term proved oil and gas reserves as “those quantities of oil and gas, which, by analysis of geoscience and engineering data, can be estimated with reasonable certainty to be economically producible, from a given date forward, from known reservoirs, and under existing economic conditions, operating methods, and government regulations, prior to the time at which contracts providing the right to operate expire, unless evidence indicates that renewal is reasonably certain regardless of whether deterministic or probabilistic methods are used for the estimation.

The shale producers valued the PV-10 of their reserves using SEC pricing of $96.04/bo, $95.05/bo, $96.78/bo, and $94.99/bo in 2011, 2012, 2013, 2014, respectively. But what follows is quite dramatic come-down: $50.28/bo, $42.75/bo, and $48.72/bo in 2015, 2016, and 2017, respectively (see here) [Edited].

Rattled by such a precipitous drop in oil prices, the shale producers scrambled to write off the reserve impairments on their income statements in 2015 and 2016 [Edited]. However humongous the accounting losses may be on paper, the physical proved reserves, if properly booked, did not evaporate as the oil price tanked; they have slept soundly in the subsurface reservoirs. Therefore, with the effect of this artificial effect of non-cash write-off removed, the shale oil producers' cash profitability situation is not nearly as bad as meets the eye.

Capital and operational efficiency

In our model (Fig. 12; Table 2), we assume everything else including the drilling and completion costs stays the same. That assumption turns out to be quite conservative in reality.

Drilling and completion costs have consistently declined over the years, thanks to technological innovations, operational efficiency gains, use of lower-cost materials (e.g., local sands and recycled water), and learning from the offset operators. Pad drilling drives down drilling and completion costs on a per-well basis.

  • For example, Whiting Petroleum (WLL) has cut drilling time and improved completion technology (Fig. 13).

Fig. 13. Whiting Petroleum, as an example, achieved capital efficiency through drilling efficiency gains and improved completion technology. Source.

Capital efficiency has also been achieved in other ways:

  • For example, Concho Resources (CXO) and Jagged Peak Energy (JAG), e.g., use 3D seismic-enabled technology to precisely target the reservoir (Fig. 14).

Fig. 14. Concho Resources (right) and Jagged Peak Energy (left) uses 3D seismic-enabled precision targeting technologies to get the maximum exposure to the reservoir. Source and source.

  • Improvement of completion technique enabled the operators to achieve higher and higher average 30-day production rate for wells completed in 2015-2018 in major shale plays including Bakken, Permian, Eagle Ford, DJ Basin, and Mid-Continent (Fig. 15), a trend we have encountered in Fig. 8.

Fig. 15. The 30-day initial production rate for horizontal wells. Source.

On the other hand, the operating costs have consistently decreased across all shale plays. The increased scale of operation helps dilute the fixed operating costs over a larger total production, resulting in a drop in lease operating expense (aka, LOE) and general and administrative expense (aka, G&A) on a per-boe basis. The building out of the infrastructure leads to lower transportation and marketing costs, in spite of local differential blow-outs due to exporting pipeline constraints.

All of these capital and operational efficiency gains make it possible for shale developers to drill and complete more wells with the same amount of capital. This further weakens the treadmill argument.

How about the shrinking pool of tier-1 shale?

Some shale developers assess the quality of shale in their leases using an informal tier system, i.e., tier-1, tier-2, and tier-3, in terms of either the well-level estimated ultimate recovery of oil and gas (aka, EUR) or the internal rate of return (aka, IRR). Locally-proper type wells as defined by appraisal drilling reflect the quality of the area, which is then projected unto all identified drilling locations in the area.

Depending on subsurface geology and the prevailing well economics, a shale play may include certain domains in it with a concentration of tier-1 identified drilling location. These domains are sometimes called the sweet spots or the core of the play.

The drilling up of tier-1 locations, as worried by some, may force the operators to branch out of the sweet spots and settle on drilling less economical tier-2 and tier-3 locations, which will, of course, drain capital and generate unattractive returns. Mark Papa of Centennial Resource Development (CDEV) holds this view (see here).

However, the recent history of shale development in the U.S. has proven the concern over exhausting tier-1 locations is overblown:

  • The improvement in completion technology and the ensuing increase in EUR have turned some former tier-2 or tier-3 assets into tier-1 performers. For example, Whiting Petroleum (WLL) has significantly expanded the areal extent of the core of the Bakken play (Fig. 16).

Fig. 16. The expansion of the sweet spot in the Bakken play by Whiting Petroleum. Source.

  • Reduction of drilling and completion costs further enhances well economics and expand the inventory of identified drilling locations, as seen, e.g., in Approach Resources (AREX)'s Permian properties (Fig. 17).

Fig. 17. The reduction of well costs and improvement in EUR achieved by Approach Resources. Source.

The subsurface geology cannot be changed, but shale developers can come up with better technologies to tap into that geology. As shale developers climb the learning curve, as technology advances, the above trend will further extend the runway of shale development. In addition, there are still so many benches yet to be developed, especially in the multi-stack plays such as Permian.

Even if the vast majority of the high-quality shale has been developed in the primary recovery, who can rule out the possibility that the advancing technology makes it economical to recover more oil in the secondary and tertiary recovery which comes after the primary phase of development? I do not have a definite answer to that question, but I know it is proven a losing proposition to bet against the ingenuity of the enterprising American oilmen, who happen to be some of the earliest users of supercomputing, have pioneered the application of virtual reality decades before Silicon Valley (e.g., here), and supply to NASA for its outer space operations (see here).

  • Currently, the industry's average oil recovery factor in the shale plays is typically well below 10%, as opposed to the 30% on average for conventional reservoirs (see here). Even a few percentages of increase in the recovery factor would result in the additional production of billions of barrel of shale oil.

Investor takeaways

Sometimes, wrong theories can outlive its usefulness, e.g., that of flat Earth. Investors tend to harbor some irrational theories when it comes to the oil industry as well. The accusation of shale oil as a Ponzi scheme, in particular, has kept many investors from investing in the U.S. shale developers, thus missing out some once-in-a-generation opportunities. I believe it is high time that we slay the dragon and let go of the fear.

Fears have different origins; our fear, e.g., of snakes is innate, while others result from ignorance. The fear of shale oil as a Ponzi scheme is built upon the incorrect notion that production from shale declines at an ever faster pace, thus gorging an increasing amount of capital without generating an adequate return, which is an ever faster treadmill so to speak.

To shed a light on the validity of such allegation, I set out to derive the aggregate decline curve for shale oil and gas production in the Permian Basin, the top unconventional play in the world. Shale production in the basin is found to be of the hyperbolic form, characterized by a faster initial decline in the beginning and slower decline subsequently.

Based on such a decline curve, I further reckon that - everything else held the same - production will keep growing as long as the annual investment is maintained at over 44% of the level as in the first year. This refutes the notion that shale oil is a black hole for capital.

Furthermore, advancement in drilling and completion technologies in combination with efficiency gain due to the economies of scale have made it possible for shale developers to use less investment to produce more oil. This further underlines the fallacy of the belief that shale oil is a Ponzi scheme. Therefore, it is suggested that you have a few shale oil names on your watch list.

Disclosure: I am/we are long ESTE, HK, VNOM, WLL, XEC.

I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it (other than from Seeking Alpha). I have no business relationship with any company whose stock is mentioned in this article.

Editor's Note: This article covers one or more stocks trading at less than $1 per share and/or with less than a $100 million market cap. Please be aware of the risks associated with these stocks.