Description

Industry Overview

It is time to take a fresh look at the frac sand space on the long side.  US Silica (SLCA) has been written up several times on the VIC, both as a long and a short.  Please see those posts for additional background on the space.  Roc924’s most recent short write-up was quite timely and has largely played out as expected.  This is a cyclical business where additional capacity can be added relatively quickly, and assets should not be valued at many multiples of replacement cost.  Significant structural changes and recent capacity addition announcements in the sand market have caused great concern among investors.  The space now trades below replacement cost and at low single digit free cash flow multiples, with both the previously bullish buy side and sell side seeming to throw in the towel after a new greenfield mine is announced every week and recent comments by Haliburton, the largest buyer of frac sand, that sand intensity, or the amount of sand pumped per well, may have peaked.

Oh yeah, oil below $50 hasn’t helped either.  Our less dire view:

• The sand market is still very tight.  Long-term contracts favoring sand producers currently being signed don’t fit with the bearish view on supply.  Sand customers feel the urgency to secure access to sand supply.   

• Earnings are accelerating upward from negative during the oil downturn that began in 2014.  Stocks in the space are cheap on 2018 numbers (3-4x EBITDA, 3-5x cash earnings).  SLCA and HCLP have clean balance sheets and quality assets (scale and logistical superiority really matter).  Earnings should accrete to equity.  HCLP will in the near future initiate a high double digit yielding distribution.

• Sand intensity will still grow, though perhaps not at the ~20%+ CAGR previously seen.

• New capacity is needed.  Current rig count understates sand demand.  Permian sand will be half of demand and one fourth of supply by 2019.  Northern White will set the price as marginal cost provider.

• Supply and demand will eventually meet, probably in 2019-2020.  Sand demand is estimated at around 75mm tons this year, growing to over 100mm tons next year on a flattish rig count from here.  Based on company comments and our research into the mines announced, we expect Permian capacity to grow to 25-30mm tons by the end of next year.

• Hi-Crush will have earned more than half of its market cap by the time sand pricing has a chance to rollover.  Depending on commodity prices, earnings are unlikely to crater in 2019 like many fear.  At this point, Hi-Crush gets little credit (if any) for its Northern White production, which represents over 75% of capacity.  The market is also assigning no value to the option on higher oil prices and the resulting increase in sand demand.    

We admittedly have some doubt as to what this industry looks like after 2019 as there are many moving parts.  However, the stocks are valued as though the industry will be terminally impaired after 2019.  We chose this idea partially for those reasons and hope to hear other members’ views.  

Company Background

Hi-Crush is a frac sand provider with 13.4 million tons of annual production capacity from four low cost Northern White mines (10.4mm tons) and one new Permian mine (3mm tons).  The company also has the largest network of rail terminal and sand distribution assets across the country.  Hi-Crush, an Avista-sponsored entity, was taken public at a share price of $17.00 in 2012.  Avista is a leading private equity firm with significant investing and operating expertise in the energy industry, having sponsored other successful public companies such as Carrizo (CRZO).  Much of the company’s growth has been though asset drop downs from Avista, growing from one mine in 2011 with one million tons of capacity to five mines today with over 13 million tons of capacity.    

What is frac sand?

Frac sand is the high-quality quartz sand used as proppant in the completion of oil and gas wells.  Proppant is any material, including sand, resins and ceramics, used to prop open fractures in a well.  Proppant is pumped into the reservoir in a fluid, typically water, under high pressure, thereby creating a network of new fractures and lodging the proppant therein.  The American Petroleum Institute (API) sets standards for characteristics such as geology (>99% silica), mesh size (size of the grain), sphericity/roundness (the rounder the better), solubility (will it dissolve in the frac fluid), crush resistance (how well it holds up under pressure), and turbidity (measure of purity).  Frac sand is further categorized as either Tier 1 Northern White or Tier 2 Brown sand.  Northern White is generally considered to be of higher quality for its higher silica content and crush resistance as well as lower solubility and turbidity.  Thus, it is typically the most expensive frac sand.  The best Northern White sand is found in the Wisconsin dome and is railed to oil and gas basins.  Most of the brown sand comes from central Texas due to its proximity to drilling activity, but can be found in a variety of places.  New Permian sand coming online will be categorized as Tier 2 Brown.  Most consider brown sand inappropriate for wells deeper than 8,000 feet where pressure is greater than 5kpsi.  This makes brown sand a non-starter for areas like the Bakken and most gas plays.  Buyers of sand can choose which sand suits their needs based on cost, efficacy, and interaction with other elements of the reservoir/frac job.

Will the frac sand market be oversupplied in 2018?

We think no.  Here are SLCA’s comments on supply/demand from their Q1 call:

We believe our industry will remain tight in the near future due to three main factors.  First, our industry must add capacity to meet customers' needs. Our internal estimates and current sale-side reports estimate industry sand proppant demand to be about 75 million tons here in 2017, growing to over 100 million tons in 2018, with some estimates as high as 147 million tons. Our industry will be short capacity and we cannot let sand become the bottleneck for the completions industry.  Second, all sands is not fungible within that 100 million-plus tons of projected 2018 demand. Unlike many industrial products, there's a lot of friction in the sand market for a variety of reasons, including logistics, quality differences and mesh sizes. Therefore, we on average should see 20% to 25% more total supply than demand before our markets come into balance. So, for example, if 2018 demand is 110 million tons, that implies its supply and demand balance around 135 million tons of effective capacity. Today, even after estimated reactivations of idle capacity, our industry would only have approximately 90 million tons of effective capacity, thus leaving a 45 million tons shortfall versus projected 2018 needs. And third, even all the likely capacity additions that are being talked about are not enough. We think there could be an additional 10 million to 15 million tons of brownfield capacity added in the next 12 to 18 months, including our own expansions; and perhaps, as much as 20 million to 25 million tons of greenfield capacity being added locally in the Permian, all of which will be needed if current demand estimates prove accurate. Even if our estimated 35 million tons of potential brownfield and greenfield additions come on line, the market will still be short. There's several implications of this on a national level. First, our markets are expected to stay very tight. Most of the major sand suppliers, including U.S. Silica, are running flat out today. Demand is expected to continue growing faster than supply for the foreseeable future, and as such, we expect to continue pricing recovery and improving margins in our sand sales. Second, customers are coming to us to lock in sand supply for the next three to five years. We're using these discussions to form deeper relationships with the companies that we expect to be the long-term winners. We're also working on the next generation of agreements that can better weather the cycle.

Summarizing the comments above, the industry has 90 million tons effective capacity today, with another 40 million tons coming online in the next 18 months (high side estimate), totaling 130 million tons in capacity going into 2019.  Demand will grow from 75 million tons in 2017 to over 100 million tons in 2018.  With the sources of friction mentioned above, 100 million tons of demand needs at least 125 million tons of capacity to not be in a shortage position.  Consensus demand numbers are in the same ballpark:

You’ll see that sand demand declined with the rig count in 2015 and 2016.  But, 2017 demand is roughly 50% higher than 2014 on less than half the rig count as a result of the secular trends of higher sand intensity, more horizontal rigs, longer laterals and more wells per year due to higher rig efficiency.  IHS predicts frac sand demand to grow at a 22% CAGR through 2022 primarily as a continuation of these trends.  Our own back of the envelope build up to demand gets to a similar, though higher number than any street estimate.  Spears & Associates predicts sand intensity will increase to 9 thousand tons per well in 2018.  Using that number and the roughly 20,000 wells that will be drilled annually at the current rig count, we get to annual frac sand demand of over 160 million tons.  Here is a basic table with the key variables:  

While the output for sand demand may seem intuitively too high, we think the downturn has masked the step changes in sand intensity, and, as a result, the risks to the demand estimates are to the upside with most, if not all of these variables.  Market confidence in oil prices above $50 will result in a rig count above 1,000 rigs.  We don’t expect much more out of rig efficiency, but we should see further gains from more pad drilling.  As previously mentioned, sand intensity may grow at a slower pace, but it has not peaked.  The above table also does not account for any of the vertical rigs running.  We think the market does not properly appreciate how significant the multiplier effects of these variables can be.  For example, going from 900 to 1,000 rigs and assuming 15 days per well instead of 18 increases sand demand by 55 million tons (the high side of Permian capacity additions).  Hi-Crush uses a more conservative estimate of 100-130 million tons of annual demand at current run rate activity levels.

We’ll also note that many estimates for demand this year do not properly consider the build in drilled uncompleted wells (DUCs).  Much of the rig activity this year has not translated into sand consumption as the DUC count has grown from 5,586 wells at year end to over 7,000 wells at the end of July.  That continued DUC build indicates that 2017 sand demand could be understated by 15 million tons.  DUCs should naturally grow with the rig count but have been higher than some expected due to shortages of sand and pressure pumping crews, among other reasons.  

We admit that where the supply side shakes out is much more difficult to predict than demand.  With the industry seeking to grow capacity by 50% in a year, we think the number is more likely to come in below reported estimates for a variety of reasons we’ll discuss in the next section.  It is also important to make the distinction that the US sand market is made up of a handful of unique regions – Permian, Marcellus/Utica, Eagle Ford, Bakken, Midcontinent.  Much focus has been on the Permian as it represents the largest sand demand and the largest amount of forthcoming regional supply.  The Permian was not only blessed with some of the largest oil resources in the world, it also has a substantial amount of sand that is suitable for hydraulic fracturing.  Many of the other regions are not so lucky and will rely almost exclusively on Northern White for its frac sand supply.  We also need to consider grain size supply and demand as the natural breakdown of grain sizes produced may not match what the market demands.  Recent trends show a continued shift to finer grains.  These finer grains (40/70 and above) represent approximately 80% of demand, but only about 60% of supply.  As discussed in the SLCA comments above, this is one of the main reasons for production capacity on practical basis needing to be at least 25% higher than demand.

Putting it all together, we see both capacity and demand growing to around 125 million tons or higher in 2018.  As such, we do not see a significant risk of oversupply until at least 2019, and we don’t think we are paying for much beyond that time in the sand stocks.  As long as demand stays on trend, supply needs to continue to grow in tandem or the market will experience shortages and price spikes.  

What are the cost structure differences between Northern White and regional (Permian) sand?

Here is an illustrative table comparing per ton metrics of Northern White vs. Permian (these generally represent HCLP’s cost structure) ($ per ton):

With a glance at the table above, one can see why investors are concerned.  The key lines to highlight here are rail and trucking costs.  Since Northern White sand is railed from Wisconsin to the Permian or other areas, it has substantially higher transportation cost as compared to regional sand, which is only trucked from the local mine to the wellsite.  Trucking costs for Northern White are typically less than in-basin regional sand because the rail terminals are generally closer to drilling activity, but this cost savings is not enough to make up for the rail cost.  Hi-Crush’s Kermit mine was the first Permian sand mine to come online (July 31).  While Permian-produced sand is in its infancy, it is already causing a major disruption in the frac sand industry because of the cost advantage it is expected to enjoy vs. Northern White.  Costs for Permian mines are projected to be less than half of those of Northern White delivered in basin.  This arbitrage has encouraged the development of significant Permian supply, which will be coming online over the next two years.        

One important caveat to discuss on the table above is that it does not account for any adjustment as Permian capacity comes online.  For example, many sand suppliers have discussed trying to push back on rail costs to get their cost structure more inline.  Also, many expect trucking costs to explode if even a portion of the local Permian sand is developed.  Since trucking distances are longer for local sand, trucking cost inflation disproportionately hurts Permian mines.  Richard Schearer, Emerge Energy Service’s CEO, commented on this dynamic on their last call:  

“We have seen some operators convert from regional brown sands to Northern White, as surging trucking rates are making rail transport at Northern White, more cost competitive. Our terminal network is critical in helping our customers optimize their supply chains. So, we see the physical properties and the logistical benefits of Northern White contributing to a continued bright future…But certainly in the Texas market we're seeing not only a surge in pricing, that again varies even by trucking company. But importantly, we're seeing a shortage of trucks now. So, as a result of that, that gap is narrowing to the point that the better quality material needed from Wisconsin becomes more competitive as we look to improve our logistics position here, both with competitive rail rates that we're fighting for everyday and making progress on, and then the transload sites that we have in place. So, as we see more and more frustration with some of the trucking logistics challenges, we've actually had more customers begin to look at bringing northern white sand down as a result.”

Any potential cost structure adjustments still likely don’t close the gap between Northern White and Permian.  As long as a significant arbitrage exists, market forces will encourage further Permian sand mine development.

How much Permian sand is coming online?

No one really knows.  Permits filed would indicate it could be over 50 million tons of annual capacity, equating to 50% of total current US sand supply.  The public sand companies handicap anywhere from 15-35 million tons of greenfield Permian projects that will come online by the end of 2018.  We tend to agree with the high-end of those numbers.  Gating factors include finding appropriate sand/geology, securing access to water/electricity,/highways, determining what to do with protected lizard populations, ordering specialized equipment, and hiring labor with necessary skills.  A significant amount of the projects in the often quoted 50+ million ton number are speculative projects with no financing or customer contracts in place.  Permits only cost a few hundred dollars to file.  However, we do expect the majority of these projects to move forward.  The cost arbitrage is just too high, and the sand market is too tight.  We think project financing is not a permanent roadblock for good projects.  In addition, while many of the public companies talk about the constraints around Permian sand, they all want to be there.  Some of this is a defensive move in case they are wrong about how significant Permian supply could be.  Also, the public companies are best positioned to move these projects forward with the experience of constructing greenfield mines and relationships with customers and suppliers.  Below is a table that Jefferies published in March listing greenfield frac sand projects (regional and Northern White).  It is important to note that the projects in red were not specifically sanctioned at the time.   I’m sure there is an updated list, but we have not seen it.   

*Source: JEFCO OFS Report (3/22/17)

We have created one below for the Permian projects we are aware of (it is admittedly not comprehensive).  Subsequent to the JEFCO table being published, Silica announced a 4 million ton Permian mine that is included in their numbers above (i.e., it is not incremental).  Not on this list are Encap-backed Black Mountain’s four million ton Permian mine (expandable to 8mm) and Unimin’s five million ton Permian project.  Preferred Sand is a KKR-backed company that recently filed its S-1.  (Good luck getting the IPO done in this market.)  Preferred has two Permian mines that are projected to be online in early 2018 with total annual capacity of 3.6 million tons each.  We understand that the Wilks Brothers’ mine in our table below is the mine on which Fairmount recently announced an agreement.  One can get as granular (pun intended) as one wishes here.  The takeaway here is that significant sand production is coming online in the Permian, though perhaps not as much as many fear.   

(millions of tons capacity)

Is Permian sand a substitute for Northern White?  Will it meaningfully displace Northern White sand?

Somewhat, and eventually, respectively.  We see Permian sand as the biggest risk to the stock, but think it is more than priced in.  Plus, it can’t really hurt Hi-Crush for the next several years.  We don’t know yet what the ultimate appetite for Permian sand will be; initial data points indicate that it will be quite strong.  The fact that HCLP contracted 90% of its Permian capacity for multiple years with significant customer prepayment before the mine opened is very telling.  We also think this is an indication of just how tight the sand market is currently.  We recently spoke to a pressure pumping customer of Hi-Crush who indicated that spot prices of Northern White sand in the Permian have grown to well over $100 per ton.  This would include the cost of rail, trucking and potentially a containerized solution.  

Of course, if sand companies had an endless supply of Northern White sand at low prices, we would not see any Permian mines coming to market.  In this low oil price environment, E&P companies are trying to push on any part of their cost structure that they can.  The potential cost savings of regional sands are too large to ignore.  In addition to cost consideration, sand is critical to the supply chain.  As more and more E&P companies enter “manufacturing mode” with their operations, they need to secure long-term sources of sand supply.  While some E&P companies say they will not consider using Permian sand, many companies have already contracted with Permian sand providers as their primary source of sand.  In addition to the unnamed HCLP customers, Encana (ECA) plans to source local sand for its entire Permian operation.  Other companies are planning to use both Northern White and regional sand in the Permian and view local sand as a low-cost way to increase sand intensity.  

Permian sand capacity also will skew heavily (80%) toward 100 mesh sand (HCLP’s Kermit mine is 100% 100 mesh).  Yet, Permian demand is more evenly split between the fine grades – 100 mesh and 40/70 sand.  Of the Permian 40/70 that will be produced, we’ve heard of issues with crush resistance, so broad adoption looks unlikely at this point.  Regardless, 40/70 sand used in the Permian will still need to be primarily sourced from Northern White mines.  This is something like 20 million annual tons in a market of 45 million total tons in demand.  Hi-Crush commented on the Q2 call that they “have been selective in our approach to contracting the valuable 40/70 that our Northern White mines can supply cost effectively to the Permian through our owned and operated in-basin terminals”.  Thus, pricing for 40/70 sand is likely to stay quite high in the Permian even as local sand supply comes online.  Some speculate that 100 mesh sand could become oversupplied in the Permian, which could negatively impact Hi-Crush’s Kermit mine.  Given the many moving parts, no one knows for sure how supply/demand will shake out for one particular grade of sand in one particular market.  However, Hi-Crush can’t be financially impacted by a 100 mesh oversupply for the next several years due to its fixed price contracts at Kermit.  We can worry about that in 2024 when those contracts come up for renewal.  It is also important to note that Permian sand will likely displace other regional Texas sands like Voca, which are trucked from central Texas and are priced off of Northern White in this tight market.  As trucking costs increase, this supply may no longer be competitive.  At around 30% of Permian supply today, other regional sand displacement could be a significant offset for the Permian sand coming online.  

Finally, Permian sand is a substitute for Northern White only in the Permian.  Permian sand could conceivably saturate that individual market, but it will likely not be railed to other markets as it would lose its structural cost advantage once it pays to rail out of the basin.  Regardless of what happens with Permian supply and demand, half of the sand demand in the country will be in other places and will be primarily supplied by Northern White.  Areas like the STACK/SCOOP in Oklahoma, which had little sand demand at the prior peak in 2014, will likely demand and receive any displaced NW tons.  Given the shorter rail distance, these areas will also enjoy lower rail costs.  Pressure pumping customers and E&P customers with assets in different areas may prefer to contract with Northern White mines for its flexibility in grade and its ability to be shipped anywhere in the country.  As such, while we do view the coming Permian capacity as a major structural change to the energy industry’s sand flows, we think Northern White still has an important place in the supply system.

Is sand intensity peaking?

The answer is no.  The secular trend of increased sand intensity is well established.  See the following graph of Permian sand intensity to illustrate:

At the prior peak in 2014, average proppant per well was 3 thousand tons, doubling to 6 thousand tons in 2016.  Spears & Associates estimates that proppant loadings will grow to 9 thousand tons in 2018.  (We do not purchase Spears research, but we have a lot of respect for them.  Here is a link to a Spears blog post that discusses their views on sand intensity: https://spearsresearch.com/insider/?offset=1500631200183&reversePaginate=true ).

However, Haliburton, the largest buyer of frac sand, called the trend into question, indicating on their Q2 call that this was the first quarter in many years in which sand pumped per well was down.  While we don’t think they would lie about it, the statement confused us and took sand stocks to a whole new level of pain.  We can speculate why the largest buyer of sand would go out of its way to highlight the point to investors as sand prices continue to rise quarter after quarter.  Prior to the HAL comment, concerns primarily had to do with supply; now we have to deal with the prospect of demand decreasing.  The statement was confusing to us because it goes against every other data point we can find (and we looked…a lot).    Haliburton’s peers were asked on their calls if they were seeing the same thing; the answer was unanimously no.  Flip through almost any E&P presentation, and you’ll see subsequent generations of completion recipes with higher sand loads yielding higher and higher productivity.  We have seen some companies on the leading edge of sand intensity experiment with slightly lower sand loads (e.g., PE, MTDR, CLR, APC (in DJ Basin)).  We know of no company pulling back from higher sand loads as a result of this experimentation.  Clearly, there is some elasticity to sand loading per foot; if oil prices stay low and sand prices continue to rise, we’ll hit a point of diminishing returns on the incremental pound of sand added.  It does not appear that we are there yet.  On the other hand, if oil prices increase, customers will  be less price sensitive, and incremental sand will have a higher ROI.  Finally, the gap between average and leading edge sand intensity is still quite large – Hi-Crush highlights “super fracs” of 25 thousand tons in its presentation.  Clearly, the secular trend is still very much intact.

Do long-term customer contracts smooth out the cycle / give sand suppliers any visibility?

To some degree.  Many sand producers had fixed price, take-or-pay contracts prior to the downturn.  When oil prices collapsed, certain customers literally could not pay their service providers and the sand companies realized their contracts did not do a good job of protecting them from a defaulting customer.  Many sand supply contracts were either torn up or cancelled with a termination fee, leaving the sand producers holding the bag and forcing them to shutter mines and pay for railcars they weren’t using.  Newer contracts signed in the last couple of years are more complicated but have more teeth so to speak and should hold up better.  Since they were signed during a downturn when sand prices were low, the companies structured them to reprice quarterly with spot market prices for sand and/or oil prices, providing upside as drilling activity recovered.  The sand companies have chosen to not disclose the detail of these contracts for competitive reasons; however, we understand they have certain stabilizing mechanisms such as collars and floors as well as clauses regarding penalties for non-performance.  Leading edge minegate pricing has more than doubled from the lows to around $60 per ton, though contract repricing provisions and remaining fixed price contracts will result in a delay in average revenue per ton catching up with spot prices.  Typical Northern White contracts have a three to five year duration; average Northern White contract duration for Hi-Crush is currently three years.  Hi-Crush has one significant legacy fixed price contract at its 2.86 million ton capacity Augusta facility.  This contract was amended in Q3 2016, locking in very low prices, and is set to reprice later this year, which should provide for a decent uplift in average revenue per ton going into Q4.

Hi-Crush’s contracts for its new Kermit mine are long-term fixed price take-or-pay.  The anchor contract for one third of the capacity (1mm tons annually) is with a large, well-capitalized E&P company and has a duration of seven years.  Hi-Crush has stated that the anchor contract provided for a contribution margin of $35 per ton, implying a landed price of $60/ton.  We understand that the three million ton mine is now 90% contracted with multi-year contracts that provide at least $35 per ton in contribution margin.  

What is the significance of Hi-Crush’s infrastructure / logistical assets?

All of Hi-Crush’s four Northern White mines are located with direct access to Class I railroads, including Union Pacific and Canadian National, which eliminates the need to truck to the rail site like some peers.  Also, all of the mines are capable of loading unit trains, which is a train consisting of 120 cars shipping only sand directly to one destination, which makes for the most efficient and cost effective rail operations.  Hi-Crush equates unit trains to driving on the express lane of the same major highway.  Some basic math: each rail car carries 200k lbs. of sand x 120 railcars = 12 thousand tons on a full unit train.  This roughly equates to sand for only two horizontal wells, the industry needs thousands of unit train trips every year.  Hi-Crush shipped three unit trains every two days in the second quarter, equating to 68% of its total railcars shipped.  Since rail is the most expensive and time consuming piece of the Northern White supply chain, having efficient rail operations is critical to success.   

Hi-Crush owns and operates 11 terminal facilities with Class I railroad access across Texas, Colorado, Pennsylvania, Ohio and New York (six of which are unit train capable), with 74 thousand tons of rail storage capacity and 120 thousand tons of silo storage capacity.  Hi-Crush’s terminals in the northeast make it the largest sand infrastructure owner serving the Marcellus/Utica.  This infrastructure, strategically located as close as possible to activity in the various unconventional basins, is key to executing just-in-time sand deliveries and allowed the company to avoid many of the bottleneck issues that peers experienced in the second quarter.  Hi-Crush shipped 77% of its sand in the second quarter through company-owned terminals.  The company’s soon to be completed (October) Pecos terminal will be the first unit train capable terminal with silo storage in the Southern Delaware Basin.  This terminal is directly on the Union Pacific main line, which will allow for the cheapest rail cost possible.  Along with its other two terminals in Odessa and Big Spring and its storage at Kermit, Hi-Crush has an enviable logistics position to serve the Permian.  Around 80% of Permian sand consumption is within 50 miles of Hi-Crush’s Permian footprint, thus minimizing trucking distances/cost.  Moving the radius out to 75 miles covers 95% of Permian activity.  Generally-speaking, an owned terminal adds $5-10/ton of contribution margin capture, or $100mm of midpoint margin at full utilization.  As Hi-Crush sells more sand in-basin (64% in Q2 vs. 49% last year), it has more opportunity to capture margin from its sand and that of third parties.  Hi-Crush estimates that its terminals have a total cost of approximately $250 million.  

Propstream is Hi-Crush’s last mile delivery solution of purpose-built cubic sand containers and conveyors.  Proppant is loaded into Propstream containers at Hi-Crush’s in-basin terminals and then put on flatbed trucks, saving time and money vs. commonly-used pneumatic trucks and providing for less demurrage and wellsite congestion.  SLCA has a similar solution, which they estimate has 85% faster unload time and 30% less truck trips as compared to pneumatic trucks.  Propstream’s mobile enclosed conveyor system PropBeast feeds sand into the blender, reducing sand emissions by more than 90% as compared to pneumatic trucks.  Importantly, PropBeast meets the OSHA regulations going into effect in 2018 that limit respirable sand emissions.  These regulations are a significant threat to most of the industry currently relying on pneumatic trucks.  Propstream is owned 100% by HCLP and buys equipment from PropX, in which HCLP has a $17 million investment. Each Propstream system consists of two PropBeast conveyors, some 100 PropX containers and the HCLP personnel to operate them.  There are currently four PropStream crews operating in the Permian as well as the Northeast.  The company expects to have nine PropX systems deployed by yearend.  Each system generates over $1.5 million in EBITDA, meaning Propstream will be a $15 million EBITDA contribution for 2018.  

As the amount of sand moved around the country increases, efficient, integrated logistics are increasingly important.  The “last mile” is in many cases the most expensive piece of the delivered ton of sand, especially for regional sand where there is no rail component.  As such, when producing / transporting a commodity like sand, the ability to save the customer money on the delivery cost and reliably deliver on time becomes a key differentiator vs. competitors that may just sell sand.  Thus, a commodity product becomes a specialized service.  Hi-Crush often describes itself as a logistics company that happens to sell sand.  As E&P companies increasingly purchase sand from the sand companies instead of going through the pressure pumper, a turnkey solution is critical.  For example, Schlumberger manages all of its sand logistics in one central “war room”, staffed with dozens of people.  An individual E&P company is unlikely to have anywhere near that type of capability to manage logistics.  In the prior cycle, service companies would heavily mark up the sand an also make money on transportation/logistics.  In displacing the service company in managing logistics, Hi-Crush now has the opportunity to capture that additional margin without the customer having to pay more.  

What about consolidation?

Regardless of what happens with supply and demand, we expect consolidation in the space to accelerate.  We mentioned the likelihood of the public companies buying more speculative Permian mines above.  We also see a high likelihood of consolidation among the public companies in an industry that is already fairly consolidated.  SLCA has publicly stated that it would like to buy more sand assets and/or companies and sees itself as the consolidator in the industry.  Finally, we think it makes a lot of sense for large service companies like Schlumberger (SLB) and Haliburton to own sand mines.  SLB indicated on its first quarter call that it is pursuing additional ownership of sand operations to diversify its supply chain and decrease reliance on third parties.  HAL is less likely with its capital-light business strategy.  Many pressure pumpers invested in sand mines in the prior peak to secure capacity.  The move makes strategic sense because sand is such a critical piece of the pressure pumping supply chain.  We’ve heard of many instances over the last few months where frac crews are waiting on sand for days.  This is time and money lost for all.  Patterson-UTI (PTEN) estimates that sand related pass throughs are as much as one-third of its pressure pumping company’s revenue.  Bringing this in-house and controlling the supply chain provides many financial, competitive, and logistical advantages for service companies.      

Capital Structure / Cash Flow / Valuation

Hi-Crush Proppants LLC, which owns HCLP’s general partner, was formed in 2010 by Avista Capital Partners (61% ownership) and members of the management team (the remaining 39% ownership).  This entity is entitled to any incentive distribution rights (IDRs) paid by HCLP as the distribution grows.  In addition, as mentioned above, these insiders own 23.2% of the outstanding HCLP units (13.2% Avista, 10.0% management and directors), which means they receive ordinary course distributions on limited partner units as well.    

Hi-Crush had an average diluted unit count of 91.6 million in the second quarter.   The company completed several equity issuances in 2016 and 2017 to delever the balance sheet and acquire assets, primarily via dropdowns from the sponsor.  As of June 30, 2017, Hi-Crush had $194 million of debt, almost all of which was a senior secured term loan that matures on April 28, 2021.  This term loan has an accordion feature that allows the company to draw an additional $100 million with amendment.  The cost of the term loan is LIBOR + 2.75% (4.9% currently).  We expect this term loan to be refinanced prior to maturity and that a similar amount of debt becomes a permanent part of the capital structure absent some sort of M&A transaction.  The company also has an undrawn $75 million revolver ($17 million letters of credit outstanding).  Cash balances at the end of the second quarter were $27 million.  Net debt was $167 million.  

For a look at 2018 cash flow, we assume 90% utilization on Northern White capacity and 100% utilization at Kermit.  The company will be at practical full utilization as it exits Q3.  We’ll assume Northern White pricing doesn’t improve beyond some guided improvement for Q3.  Cost per ton is guided to be in the $12 per ton range.  We back into $65 per ton of Kermit revenue to arrive at the guided $35 contribution margin.  We assume Northern White rail cost per ton is $40 and trucking costs are $10 per ton.  Kermit trucking costs per ton are assumed to be $20.  Obviously, there are no rail costs for Kermit.  In our 2018 model, Northern White contribution margin is $25 per ton, which is consistent with management’s expectation of greater than $20 contribution margin for next quarter even at less than full utilization.  This basic model does not distinguish between NW sand sold FOB vs. in-basin, but rather assumes all sand is sold in basin.  Thus, our transportation costs will appear higher than what the company reports, but contribution margin should be similar.  We are using the high side of guided G&A of $8-9 million per quarter, and we assume that debt stays flat for interest expense.

The company will spend $115 million to $125 million this year on construction of the Kermit mine, the Pecos terminal facility and additional Propstream systems.  As an MLP, Hi-Crush pays no significant income tax and makes estimates for maintenance capex to calculate distributable cash flow.  The company uses $1.35 per ton produced for maintenance capex, which is non-cash but clearly a real expense due to depletion of sand reserves.  After this year, HCLP will have no significant growth projects and free cash flow should ramp meaningfully.  The company will need to make a working capital investment (primarily receivables and inventory) as revenue grows and Kermit ramps production, but we will not hit them for this in our valuation.

We think the following is a fair representation of what 2018 holds.  HCLP looks extremely cheap at 2.3x distributable cash flow, 3.2x EBITDA, less than book value and $64 per million tons of annual capacity.  As a point of reference, consensus 2018 EBITDA is $251 million.  We keep asking ourselves where we could be significantly wrong as the low single digit multiples indicate numbers are going way down soon.  We have seen some analysts start to take 2018 numbers down on fears that capacity additions will hurt pricing in the back half of 2018.  Given Kermit’s fixed price contracts, there really is not much price risk to it, up or down, for several years.  As mentioned above, capacity additions will likely put downward pressure on pricing at some point, but that probably won’t be until 2019 at the earliest.  We actually expect sand prices to increase through the balance of this year, and new Permian mines won’t have much impact on 2018 pricing.  We have a constructive view on pricing after 2018 for reasons mentioned above, but admit that visibility is diminished.  

Looking at the valuation another way, what is the scenario where the current price make sense?  After all, buying cyclical companies when they look cheap is often not a great idea.  The stock is almost back to its prior low range when oil was in the high $20s and the company had significantly less cash flow and infrastructure assets and significantly more debt.  It is hard for us to come up with a scenario where cash earnings go below $1 per share (less than an 8 P/E currently) even slashing Northern White prices.  As long as the Northern White mines can cover G&A and interest expense, Kermit earns more than $1 on its own for the next several years.  That $1 earnings scenario would involve many other higher cost NW mines shutting down first and taking significant capacity off of the market.  A significant downturn in the rig count due to lower oil prices is highest probability for such a scenario.  This is clearly a risk that a prospective HCLP owner would have to underwrite (or hedge); we would not be posting this idea if we were not bullish on oil.  We still see a tremendous margin of safety in a lower for longer environment.   

One final valuation perspective – the amount that the company paid for the Kermit mine just six months ago is now 40% of the current enterprise value and almost half of the market cap.  Presumably, it is worth much more today, now generating cash and covered by multi-year contracts for 90% of capacity.  Backing out another $100 million of value for Propstream (~5x EBITDA), which is agnostic to sand source, leaves just over $400 million for Northern White mines and all of the company’s infrastructure assets (12 terminals, 8 of which are in the Marcellus and not directly exposed to regional sand).  Those Northern White and terminal assets are generating over $200 million of run rate EBITDA, which likely goes up next year.  These assets may indeed be a cigar butt (we don’t think so), but we think we have more puffs remaining than the market price implies.

(in millions, except share price)

Distributions / IDRs

Hi-Crush is currently in the no man’s land of being an MLP that doesn’t pay a distribution.  Hi-Crush paid its first distribution of $0.475 per share in July 2013 and continued to make distributions until August 2015, at which point shareholder distributions ceased due to the oil downturn.  The company has indicated it will resume its distribution this year.  Estimates for what they will pay are all over the map.  Some analysts, who apparently don’t read company press releases, think no distribution will be paid given the uncertain environment.  Others have them paying around $2/unit next year (25%+ yield).  The company previously targeted debt/EBITDA of 2.5x and distribution coverage of 1.2x.  Lessons learned after cutting the distribution result in new targets of 1.5x for both debt to EBITDA and distribution coverage, though no new distribution policy has been formally announced.  We think the distribution will start off somewhere north of $1/unit annually and grow from there.  Excess coverage will provide substantial flexibility to pursue new growth projects or repurchase units.  At the current valuation, we hope it will be the latter.  We have discussed a buyback with management and expect some comment either way at the company’s analyst day in September.  Here are the comments on distributions from the last conference call:

Laura Fulton – CFO – “our focus is on resuming the distribution in the latter part of 2017 and starting off at an amount that is meaningful to unitholders, but also very sustainable and something that we’re allowed to grow over time.”  

Robert Rasmus – CEO – “We feel we’ve done the bulk of our capex investing now…a priority for us is to create value and total return for our unitholders and a major portion of that is also to return cash to the unitholders in the form of distributions.”

IDRs will not be paid until the company pays unitholders an annual distribution of $2.185 per share, at which point the GP receives 15% of the incremental distributions and can eventually reach 50% of incremental distributions.  Thus, Hi-Crush’s general partner is highly incented to grow the distribution beyond this level again.  A meaningful unit repurchase program at recent prices would allow the company to reach the higher IDR more quickly and make the ongoing distributions per unit more sustainable as the same amount of cash flow is distributed over fewer units.  We view distribution resumption as a major catalyst for the stock, not only because of the juicy income it will provide, but also because it will allow for MLP index inclusion and institutional ownership.

Mine Assets

Kermit Mine Facility

Hi-Crush completed construction of its 3-million-ton annual capacity Permian mine in Kermit, TX at the end of July.  The company originally acquired the 1,226 acre site and 55 million tons of frac sand reserves in February 2017 through its purchase of Permian Basin Sand for $275 million.  The company issued shares at $18 to partially fund this acquisition.  Production facilities cost an additional $50 million.  Sand produced will be 100% 100-mesh.  The Kermit facility is strategically located in Winkler County within 75 miles of 95% of sand consumption in the Delaware and Midland basins.  Operating costs per ton are expected to be $10 per ton.  As a part of its purchase of the Kermit property, it secured options to increase its mining operations.  Our sense is that Kermit capacity could easily be expanded to 4 million tons.   

Wyeville Mine Facility

Wyeville is HCLP’s lowest cost facility, with operating costs of less than $15 per ton, and is one of the lowest cost Northern White facilities around.  The vast majority of Wisconsin mines have costs above $20 per ton.  As a result, Wyeville remained highly utilized during the downturn as other higher cost mines were shuttered.  The Wyeville facility is a 971-acre facility with integrated rail infrastructure, connected to a Union Pacific Railroad mainline. Completed in 2011 (and expanded in 2012), the Wyeville facility has processing capacity of 1.85 million tons of 20/100 frac sand per year.  Assuming production at that capacity, and based on Hi-Crush’s reserve report, the Wyeville facility has an implied reserve life of 41 years.  

Blair Mine Facility

The Blair facility, HCLP’s newest Wisconsin mine, was completed in March 2016, and has operating costs that rival Wyeville.  Blair has integrated rail infrastructure as well, and connects to a Canadian National Railroad mainline.  Completed in March of 2016, the facility has a processing capacity of 2.86 million tons of 20/100 frac sand per year and an implied reserve life of 41 years.

Augusta Mine Facility

The Augusta facility was originally a greenfield project of Avista.  It is the company’s highest cost facility, and as such was idled for much of 2016.  Completed in 2013 and expanded in 2014, the Augusta facility has a processing capacity of 2.86 million tons of frac sand per year (multiple mesh sizes) and 41 million tons of reserves (reserve life of 14 years).  Augusta has eight on site rail spurs, and just like the Wyeville facility, it is connected to a Union Pacific Railroad mainline.   

Whitehall Facility

Whitehall commenced production in late 2014 and has a processing capacity of 2.86 million tons of 20/100 frac sand per year, targeting the Marcellus and Utica plays.  It also has integrated rail infrastructure, and similarly to the Blair facility, makes use of a connection with the Canadian National Railroad mainline. The Whitehall facility has 80 million tons of reserves and an implied reserve life of 28 years.  Whitehall was idled during the second quarter of 2016 and resumed production in early 2017.

Why should we own Hi-Crush vs. other names?

1. Low cost, strategically located mines.  Hi-Crush believes it has the lowest operating costs among Northern White mines.  It has first mover advantage in the Permian and has de-risked its cash flow stream with a multi-year fixed price contract.

2. Significant logistics assets and last mile solution.  See detail above.  

3. Relatively clean balance sheet.  Should be less than 1x debt/EBITDA on 2018 numbers.

4. High insider/PE ownership.  Insiders own 23% of the LP and all of the GP’s IDRs.  We think the high insider ownership combined with the carrot of hitting the IDR split next year makes for sound capital allocation going forward and a massive return of capital.    

5. Appealing valuation.  While the entire sand space is heavily discounted on 2018 numbers, HCLP stands out as an outlier at 2.3x P/DCF, 2.6x EV/EBITDA and a potential high double digit yield.

6. Forthcoming capital returns.  We don’t know what the initial distribution will be, but we see around $3 of distributable cash flow to play with.  We hope unit repurchases are a big portion of this given the valuation.  (If you are a unit holder, please call management and say you want a buyback.)        

Public Comps

Source:  The Bloomberg.  Consensus EBITDA estimates.

SLCA is considered by many to be the leader in the frac sand space and is the largest of the public companies by enterprise value.  The company has high quality mines and infrastructure along with a widely accepted last mile solution.  The stock trades at an attractive valuation and probably does well from here.  SLCA sees itself as a consolidator.  

FMSA has too much debt to provide any margin of safety.  The company only recently moved into the Permian.  It also has no last mile solution.  Despite these negatives, it trades at higher multiples than the rest of the group.  

We have not looked as much at EMES.  The debt load is a non-starter for us.

SND came public late last year at $11 per share.  The company has a significant net cash position and looks cheap.  The company still needs a Permian presence and last mile solution.

Key Risks

Lower oil prices – HCLP is highly correlated with oil prices and is quite volatile.  Even if you buy our thesis that a lower price environment is priced in, the stock will likely trade with oil.  

More mine announcements – Sand stocks go down with each new mine announcement.  We think we’ll see fewer mines announced going forward, but it will still be an overhang.  We think many of the yet to be announced transactions will be public mine companies buying speculative Permian mines (similar to the FMSA deal).  This does not add incremental capacity to the already announced pipeline but makes the likelihood of that capacity moving forward much more likely.  Some of the public miners still want a Permian presence (SND, EMES), so we think this is a high probability.   

Sand oversupply – We discussed this risk in detail above.  We are comfortable that stock prices impute a horrendous supply scenario.  However, we could obviously be wrong on the amount of supply or where the stocks will trade.

Distribution foot fault – One of our big worries is that the company pays a lower distribution because the yield would be “too high”.  We’ve seen this before with poor management teams that want to empire build with shareholder dollars.  Given the alignment here, we think this is a low risk.

Changes in grade preference / technological disruption – E&P preference for sand grade has changed dramatically over the last several years.  Companies now are gravitating toward 100-mesh, a grade that once was considered waste sand.  HCLP has some flexibility in the grades it produces, but we could imagine a scenario where they produce less of what the market demands.  Kermit can only produce 100-mesh, but its contracts protect it from potential grade preference changes for the foreseeable future.  Some unknown technology that becomes a substitute for frac sand or reduces its demand would obviously be disastrous for the frac sand group.  

Avista fund liquidation – 13% of the LP units are owned by Avista sponsored private equity funds.  These funds will eventually come up against a fund termination date and either need to sell units or distribute in kind, which could cause selling pressure on the stock.  It could also precipitate a sale to a peer or service company.

Catalyst

Increased earnings

Higher oil prices, rig count, frac crew count

Clarity on sand industry supply/demand

Resumption of distribution

Unit repurchase program

MLP index inclusion

Takeout by industry peer or customer