Verifying oil and gas estimates

Performing due diligence on the accuracy of oil & gas reserve estimates requires two elements:

  1. An understanding of the art and science of reserve estimation.  This blog post will provide a brief overview of reserve estimation techniques.
  2. Technical information on the methodology used and its key assumption.  Places to look for technical disclosures would be investor presentations, CORRESP filings, and (occasionally) the 10-Ks / annual reports.  However, this process can be tedious whenever investor presentations are difficult to locate.

Different viewpoints on accurate reserve estimation

The decline curve of a well will depend on many factors such as:

  1. The permeability and porosity of the reservoir.  The website Geomore.com has a good summary on the concepts behind permeability and porosity.
  2. Other geological characteristics of the reservoir such as pressure, naturally occurring fractures, etc. etc.
  3. Whether the well was drilled horizontally, directionally, or vertically.
  4. Well spacing.  Closely spaced wells may (or may not) cannibalize the production of other nearby wells.
  5. The completion techniques used: the number of stages of hydraulic fracturing, the type of proppant (e.g. sand, ceramics) used, composition of ‘secret sauce’ fracking fluids, etc. etc.
  6. The type of enhanced recovery techniques used:  artificial lift, compression, water injection, CO2 flooding, etc. etc.

Because of these variables, different wells will have different decline curves.  In the case of ‘unconventional’ wells using the latest technology (e.g. horizontal drilling, multiple stages of hydraulic fracturing, new formulations of fracking fluids, etc.), estimation may be difficult because there is no historical data for the latest technologies.

Ryder Scott has an interesting and insightful overview of various decline curve analysis (DCA) techniques in its newsletter article “Promising DCA techniques emerge as widespread public debate focuses on shale gas forecasts” (alternate link).  The article provides an excellent overview of different estimation techniques and different viewpoints on the appropriateness of the assumptions behind DCA estimates.  For DCA using Arps equations, the key assumptions are the “b-factor” and “Dmin” (minimum decline) values.  Higher b-factor values and lower Dmin values will lead to more aggressive reserve estimates.  Conversely, lower b-factor values and higher Dmin values will lead to more conservative estimates.

One of the controversies noted in the Ryder Scott article is that the use of Arps equations can lead to abuses, especially when a ‘bolt-on’ Dmin value is not used to improve estimates.

“There are companies that have publicly stated that they use Arps with no Dmin,” said Lee. “I believe that Arps, with or without Dmin, is subject to abuse and is, in fact, abused regularly.”

While Sentieo (as a company) does not necessarily agree or disagree with Mr. Lee’s views, the Sentieo platform can be used to find companies that disclose the use of Arps equations without the use of a Dmin value.  The platform can be used to find companies that disclose the b-factor but remain silent on the Dmin value; such companies may not be using a Dmin value in their reserve estimates.

Further reading

For more information on reserve estimation, you can look into resources such as:

  1. White papers from engineering firms.
  2. Academic literature.  You can search through such articles via Google Scholar.
  3. Articles from shale skeptics such as Arthur Berman (e.g. here and here).
  4. SEC correspondence filings that highlight the SEC’s concerns.

Disclaimer: While there is a wide range of viewpoints on estimating oil and gas reserves, please note that Sentieo does not express an opinion as to the superiority of any one method over another.  Please do your own research and form your own opinions.

CORRESP filings

Sometimes the SEC will ask a company about its reserve estimates.  CORRESP filings can be useful for a few reasons:

  1. Some companies will disclose their estimates’ underlying assumptions to the SEC confidentially rather than disclosing their assumptions publicly.  (*To be fair, there may be commercial reasons for keeping information confidential so that competitors do not benefit from it.)  The decision to keep reserve estimation details private will be reflected in the public record.
  2. If the company does disclose publicly, the CORRESP filing will contain the key technical assumptions.
  3. Based on the questions that the SEC asks, investors could infer what the SEC’s concerns are.

In the Sentieo database, five companies have six CORRESP filings that mention b-factors, as shown in the image below (click the image to enlarge it):

b-factor-disclosures-in-CORRESP-filings

These documents may be very relevant to oil and gas investors.  Chesapeake’s entire response to the SEC from October 2011 (EDGAR, snt.io) is one of the most interesting CORRESP filings.

Finding technical data with Sentieo

With Sentieo, the process is fairly straight-forward.  A single search for “b-factor” will simultaneously search through EDGAR filings and investor presentations.  The punctuation of b-factor does not matter.  Sentieo will find instances of both b-factor and b factor.

b-factor-search-with-markup

Because it is highly unlikely that a Dmin figure will be presented without the associated b-factor, it is not necessary to search for Dmin.

Finding technical data without Sentieo

Without Sentieo, the process is trickier.  Search tools such as Crtl+F (Command+F) are usually sensitive to the hyphen in b-factor.  One method is to perform a search for b-factor and a second search for b factor.  A different method is to search for factor.  However, a search for factor may find instances of “risk factors” and other language that is not relevant to reserve estimates.

A subtle but important issue is that some investor presentations cannot easily be found on EDGAR or the company’s website.  For example, ECR has investor presentations that mention its b-factor assumptions.  In the Sentieo database, there are only 4 instances of ECR mentioning b-factor.  None of these instances occur in EDGAR filings as all the mentions occur in investor presentations, as shown below.

ECR-b-factor-mentions-in-Sentieo

Checking EDGAR manually, I could not find an 8-K filing corresponding to the June Investor presentation.  For items filed during June and July of 2015, only 3 of the 8-Ks were large enough to correspond to an investor presentation.  All three of those 8-Ks correspond to a different topic ($550M in 8.875% Senior Notes due 2023).

ECR-on-EDGAR
http://www.sec.gov/cgi-bin/browse-edgar?CIK=ecr&owner=exclude&action=getcompany

The company’s investor relations website at ir.eclipseresources.com (archive.org) contains a link to the November presentation at:

http://ir.eclipseresources.com/sites/eclipseresources.investorhq.businesswire.com/files/doc_library/file/ECR_Corporate_Presentation_-_November_2015_Final.pdf

Changing the URL from November to June does not lead to the June presentation PDF.  The events-calendar page (archive.org) isn’t much help either as it only contains links to the September and August presentations.

Google

Using Google to search the EclipseResources.com website for PDFs containing the word June (site:eclipseresources.com filetype:PDF June) does not work.

The best way to find the June presentation is to search Google for ECR June Presentation.

ECR-june-presentation-on-Google

The second search result contains a link to the June presentation.

https://s3.amazonaws.com/images.hartenergy.com/Interactive+Article/Utica/docs/Eclipse-June_2015+Utica.pdf

Without Sentieo, you would need:

  1. To know that the June presentation should exist.
  2. To tediously track down a copy of the presentation even though Google cannot locate it on EclipseResources.com.

As you can see, it is very easy to miss ECR’s June investor presentation! Sentieo maintains its own archive of investor presentations to make it much easier to find difficult-to-locate presentations.

*Disclaimer: I express no opinion on ECR’s compliance with Regulation FD.

10-Ks

In the Sentieo database, only 3 E&Ps mention the phrase b-factor in their 10-K or exhibits to their 10-K.  Evolution Petroleum’s reserve engineers state the following in Exhibit 99.6 attached to EPM’s FY2013 10-K (EDGAR, snt.io):

Production is expected follow a hyperbolic decline with initial decline rates of 99.99417% for oil and 91.669% for gas (as defined in the PhdWin economics program) and a hyperbolic exponent (“b” factor) of 1.50, which accounts for the production transitioning from early transient flow to (pseudo) steady-state flow through the multiple porosity, permeability, and fracture systems within the reservoir.

For the two 10-Ks that followed, it seems that Evolution Petroleum no longer used Pinnacle Energy Services for its reserve estimates.  This likely explains why the subsequent 10-K exhibits no longer mention the word b-factor.

CPE and AREX are the two other stocks that mention “b-factor” in their 10-K.  AREX’s latest 10-K does not state the actual number(s) for its b-factor assumption.  Instead, AREX provides some detail on the methodologies used for its reserve estimates (EDGAR, snt.io).  As for CPE, the company’s reserve engineers mention the b-factor but do not seem to disclose the actual number used (EDGAR, snt.io).

In general, 10-Ks contain technical information less frequently than CORRESP filings and investor presentations.  Investors who find technical disclosures useful should be searching through investor presentations and CORRESP filings as well as 10-Ks.

Peer comparisons

There are some situations where two or more publicly-traded companies will report data for the same play.  For example, both MPO and SD report technical data for their Mississippian plays.  In such instances, investors could compare the different assumptions used.  Such an analysis could reveal the relative aggressiveness/moderation in reserve estimation between the two companies.  However, differences in estimates could be explained by:

  • Differences in completion techniques.
  • Geological differences between wells.
  • Differences in well spacing.
  • Differences in enhanced recovery techniques.
  • Any number of other variables.

Closing thoughts

The Geomore.com website linked to earlier provides an easy-to-understand introduction to petroleum geology.  The website highlights various technical details that investors could look into such as permeability, porosity, seismic data, etc.  In general, there is a wealth of information that oil and gas investors may want to look at to perform due diligence.  Sentieo can be used to quickly find disclosures of:

  • Permeability.
  • Porosity.
  • Seismic data.
  • Assumptions and cash flow forecasts used to derive NPV estimates.
  • Pressure-based and volumetric-based estimates of reserves.  Some engineers will supplement DCA with other methods to see if the DCA estimate is reasonable.
  • Water saturation.  (Too much water can cause a well to become uneconomic.)
  • Studies on well spacing.  Oil and gas operators may perform experiments to understand the amount of cannibalization that occurs with more closely spaced wells.
  • Key details of lease contracts.
  • Contracted rates for midstream infrastructure such as gathering pipelines, compressors, etc.
  • Dmin values if Arps equations were used.  For shale plays, the lack of a Dmin value could be a red flag.

Omissions of such data could be just as important as what a company chooses to disclose.  (*In some cases, omissions in 10-K reports may exist due to SEC regulations that prevent companies from making certain technical disclosures.)  It can be interesting to compare the data that a private buyer would demand versus the information provided in public disclosures to institutional (and retail) investors.  In general, oil and gas companies disclose few or none of the items on the list above.  While you may be surprised at what Sentieo does find, you may be even more surprised at what Sentieo doesn’t find.

Links

Modified Hyperbolic Decline Curve (Updated) – This spreadsheet may be helpful in understanding the effect of changes in the assumptions behind Arps equations.

New call-to-action