In a 2006 interview with Charlie Rose, Warren Buffett said he couldn't "understand" the company General Motors (NYSE:GM)--that it was too complicated for him to figure out. But when Charlie Rose pushed him further, he finally admitted what he really meant: "I don't know whether you can exist with the kind of obligations they've got."
We find ourselves in the same position when it comes to Bellatrix (NYSE:BXE). While we can't say with certainty that Bellatrix will file for bankruptcy between now and 2020 (when the first of the term debt comes due), we simply cannot see how-by what means-they can avoid bankruptcy or some other form of substantial shareholder value destruction between now and that time.
Today in Part 2 (go here for Part 1) we are going to focus very narrowly on the issue of drilling inventory. Specifically, we do not think Bellatrix has the quantity and quality of future drilling locations that it suggests through its presentations. BXE's inventory claims (number of "locations") rest on two assumptions: 1) that the Spirit River is homogeneous (consistent) enough across their acreage to duplicate their past results in the undeveloped sections of their acreage, and 2) that the vertical layers within the Spirit River are separated enough to be targeted by different wells (thereby allowing BXE to count more "locations" stacked on top of each other). You can think of the first assumption as a matter of areal (lateral or horizontal) extension (i.e. Does the quality sustain as we move north, south, east, and west?) and the second assumption as a matter of vertical extension (i.e. Does the quality sustain as we drill laterals above and below the first Spirit River wells?).
Assumption #1: Areal Consistency
The success or failure of wells in the area seem to depend on whether or not an individual well intersects a system of pre-existing natural fractures. We believe BXE is hinting at this point on slide 14 of their January presentation where they say "Open and closed fracture systems evident in rock core and to a lesser degree in rock cuttings." We see the same issue in the 12/4/2014 press release in reference to the Second White Specks formation: "Regional vertical wells that encounter the interval in naturally fractured areas have demonstrated encouraging hydrocarbon recovery..." Lastly, in the Q3 2016 earnings call, CFO Ed Brown makes the clarifying point that "The Cardium is very heterogeneous in nature." This was in response to a question asked as to what price would make the Cardium economic. Ed was making the point that it varies even within the small amount of acreage BXE has held on to.
The reason natural fracture systems matter is because they provide what's called "secondary porosity", meaning an additional amount of storage space on top of the "primary porosity" (which is the storage space from the pores within the rock layer itself). Consequently, this is not something we can make up for with hydraulically fracturing. You can't create this kind of "secondary porosity" after the fact and then wait millions of years for the hydrocarbons to get around to filling up the fractures. It's either there or it isn't. Horizontal drilling and fracking help in this kind of play by increasing the odds of intersecting one of these pre-existing fracture systems, but they don't solve the basic problem of whether or not you have enough secondary porosity to begin with (i.e. enough volume to make the play economic). Our suspicion that the Spirit River is this kind of "secondary porosity" play is further corroborated by the wide dispersion of well results shown in slide 17 of the January presentation (i.e. some of the wells do very well because they intersect multiple natural fracture systems while others do poorly because they either intersected few or none at all).
[Side Note: The role of natural fracture systems in a true shale play (such as the Marcellus, Barnett, etc.) is the exact opposite of the role they play in conventional and "tight" quasi-conventional plays like the Spirit River, Mississippi Lime, etc. In the latter case the natural fractures provide the secondary porosity necessary to make the play economic, but in the case of shales, this secondary porosity is not needed. Instead, the natural fractures actually serve to undermine the effectiveness of fracking. In a shale, the goal is to effectively turn the rock layer into rubble ("rubblization"). Pre-existing fracture networks interfere with this process by providing a path of least resistance that absorbs all the energy from the frack. This was one of George Mitchell's major breakthroughs when he figured out how to unlock the Barnett Shale (the first shale to be unlocked).]
Porosity (primary and secondary) is not the only variable that determines whether or not a well will be successful. There's also the thickness, permeability, water saturation, and a few other factors. These all go into the PIIP (Petroleum Initially In Place) and related calculations that E&Ps turn into contour maps. In Slide 15 of BXE's January presentation we can see one such map. While the map does not explicitly say that it is this kind of map, we think that is the most likely guess. Alternatively, it could be some kind of thickness or net pay map, but either way it reveals the same basic point that the Spirit River is not uniform in quality across BXE's acreage. Notice first of all how the wells that have already been drilled (both on and off BXE's acreage) are concentrated on the thickest contours. And notice how the areas with the least drilling are in areas where the contours get thinner. Indeed, where the contours fade to nothing (represented by a pure white background) there are no wells at all.
Unsurprisingly, if we go back to slide 15 from the December 2016 presentation we see the same pattern manifested in the presence (or absence) of old historical vertical wells (indicated by hollow circles and circles with 'x's through them):
It's also worth pointing out that BXE and its peers have done relatively little development in the areas that correspond to these "dead zones." Taking all of this together-the heterogeneity, natural fracture dependency of the Second White Specks, Cardium, and Spirit River; the BXE contour map; and the pattern of old historical wells-leads us to believe that the quality of wells drilled in these dead spots (if they are ever drilled) will be substantially below the quality of BXE's historical wells (those drilled in the sweet spots).
The last point we'll make on this point (areal extension) is that comments made in the recent divestiture press releases on the effects of reserves aggregation suggest that Sproule has been using what's call the "probabilistic method" (as opposed to the "deterministic method") for reserves estimation. This is both more appropriate for wells in a "secondary porosity" play and provides a greater uplift for the reported reserves in such a play, which makes us feel just slightly more confident that this is the nature of the play. If the Spirit River were more like a homogeneous shale play, the deterministic method would be more likely used and there would be no aggregation effects to worry about (and disclose in the divestiture press releases).
Assumption #2: Vertical Stacked Potential
Just to remind you, we're dealing here with the question of whether or not BXE has a meaningful inventory of future drilling locations. We just addressed the issue of areal consistency and now we are going to address the issue of vertical consistency (i.e. can they count four wells for every one well on the map based on the idea that there are four different vertically stacked layers that can be separately targeted).
We doubt the claim that there is stacked potential within the Spirit River formation itself for the following reasons:
The entire Upper Mannville (Spirit River) is roughly 150 meters (500 feet) thick from top to bottom. Within that, the Fahler A, B, and Notikewin members are roughly 90 meters (300 feet) thick from top to bottom. BXE's type curve is based exclusively on wells in the Notikewin and Fahler B members, and if we measure from the bottom of the Notikewin member to the top of the Fahler B member, they are only 30 meters (100 feet) apart.
No matter how you measure it, though, it is a small amount of separation between these different layers. It's the rule and not the exception for fracks to spread vertically several hundred feet. Here's a graph produced by Halliburton back in 2012 showing the vertical extension of hydraulic fractures in the Barnett and Marcellus shales.
If we just eyeball it, we can see that it's typical for fractures to spread 500 feet in one direction or the other. This means that when BXE fracks the Notikewin, Fahler A, Fahler B, and even the Wilrich, they are most likely producing from all four members simultaneously. Just because they drill a horizontal well in the Fahler B member, call it a "Fahler B well", and frack it, that doesn't mean it's really just producing from the Fahler B. If they are producing from these other members (which we think is likely), then when they go back in to drill these other "locations", they won't be getting anywhere near the same results, since they will be drilling into a formation already partially depleted.
There's also the issue of the natural fracture systems, which increase the likelihood of vertical connections between the different layers-plus the fact that fracks have generally been getting larger since 2012 (when the Halliburton data was compiled).
In this kind of situation-with the layers so close together-you need what's called a "frack barrier" to make the case that these layers are truly separate and not just one single producing layer. A frack barrier is a stronger rock layer separating the others that is harder to frack than the surrounding rock, thus preventing a frack in the Fahler B from getting into the Notikewin, and so on. We haven't heard any mention of a frack barrier from the company's management, and we see nothing in the information provided to make us assume that such a frack barrier exists. What's more, this still would not resolve the issue of pre-existing natural fractures creating a connection between the layers.
We suspect instead that BXE is merely hoping that--just maybe--something like this could work. Chesapeake and Epsilon Energy tried to make a similar claim back in 2014 that the Upper and Lower Marcellus could be treated as separate and distinct targets. They were never really able to prove this concept, though, with a truly stacked test. Instead they did a pseudo-stacked test were there was actually a lot of lateral (horizontal) separation. Most likely, it was the lateral separation and not the vertical separation that allowed them to produce semi-decent results.
With the rug pulled out from under assumptions #1 and #2, it seems BXE might not have much of a drilling runway going forward. This matters a lot because even if we take for granted that BXE's wells are economic, they would need to be able to develop enough wells to scale those economics and overcome the current debt load. It's easy to see, for instance, how if BXE had only one more well that it could drill, that well wouldn't make much of a difference for the company as a whole, even if it was going to be twice as economic as the current type curve.
We are still waiting for well production data from the Alberta Energy Regulator, which we suspect will corroborate what we've said so far. And if it actually weighs against what we've said so far, we'll be sure to let you know.
We will be posting Part 3 in a few days which will include our model runs and outputs as well as some supporting evidence revealed by the differences between the Canada vs. U.S. compliant reserve reports.
Disclosure: I/we have no positions in any stocks mentioned, and no plans to initiate any positions within the next 72 hours.
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.