|Shares Out. (in M):||58||P/E||8.3||7.0|
|Market Cap (in M):||560||P/FCF||0||0|
|Net Debt (in M):||55||EBIT||0||0|
I own the common units of CONE Midstream Partners (NYSE: CNNX). This is a quality unlevered 100% fee-based MLP with a 9.5% yield that is focused on gathering natural gas in the Marcellus for production coming from the joint venture between its sponsors, CONSOL Energy (NYSE: CNX) and Noble Energy (NYSE: NBL). CNNX has a 20 year dedication on 516,000 net acres and a right of first refusal on an additional 186,000 net acres, with upside potential from the Utica horizon on that same acreage. The distribution is well covered at 1.44x in the most recent quarter, and is at the lowest level of the IDR splits. There is decent visibility towards a 15% distribution CAGR over time.
The units are trading at $9.60, down 60% YTD and down 56% from the Sep-2014 IPO price of $22. I believe they are worth $15 ½ right now (immediate 60% upside), and that they could be worth $24-26 by 2017 (150-170% upside).
The downside is minimal. How much higher does the discount rate need to be? When this security was trading with a 7% yield not long ago, 10% would have seemed like a credible downside case. Now we’re at 10%. Should it be 12%? 15%? 20%? Even if the distribution got cut 10% and had a 1.1x coverage ratio (implying a 30% reduction in distributable cash flow), valuing it with a 12% yield would make the units worth $6.80. So, you’ve got up to 14-16 points of upside and maybe 3 points of downside. The upside:downside ratio could be as high as 5:1.
Why is it so cheap?
Falling natural gas prices: The selling pressure that pushed the units down to their recent low of $8.58 was driven by a variety of issues that came together to create a perfect storm. It started with commodity prices. The MLP asset class is generally known for pipelines (i.e. toll road business models) where the key variable to how much money they make is the volume of hydrocarbons being transported through systems, not the prices of the commodities themselves. However, as oil and gas prices fell over the last year, investors became concerned that E&P companies would reduce their production growth rates, and they lowered their expectations for volumes passing through MLPs as a result. Natural gas prices are down by roughly 50% in the last 52 weeks. This certainly lowers the growth trajectory at CNNX, but a temporary low growth/no growth environment is certainly a far cry from the distribution becoming unsustainable. There will be a supply response by E&P companies, and demand continues to grow.
MLP panic selling: The search for yield as a result of the low interest rate environment drove many investors into the MLP asset class. Unsurprisingly, some clever operators took advantage of that demand for income by creating MLPs out of businesses that are fundamentally incapable of sustaining the kinds of stable distributions investors expect from the asset class. Cats and dogs shouldn’t mate, but people will try it anyway if you throw money at them. My short writeup on Seadrill Partners (NYSE: SDLP) from Feb-2014 is an example of a business that should never have been put in an MLP structure and has performed very poorly.
Other examples include Transocean Partners (NYSE: RIGP) and frac sand producers Hi-Crush Partners (NYSE: HCLP) and Emerge Energy Services (NYSE: EMES), to name just a few. These securities have done horribly, with distributions that are expected to be cut, have been cut, or have been eliminated entirely in HCLP’s case. HCLP is down 82% YTD and EMES is down 92% YTD.
One of the most prominent dividend cuts just occurred at Kinder Morgan. They took on too much leverage before the downturn and found themselves between a rock and a hard place. It is no longer an MLP, but it used to be, and it still serves as a kind of bellwether for the group. Six months ago it had a $90 billion market cap. It’s now $34 billion. A lot of people have been burned.
With ~2/3 of the MLP asset class owned by retail investors, selling pressure in certain names has turned into a stampede out of the group. Many MLPs are selling for EBITDA multiples that are as low as they got during the financial crisis, with distribution yields that are as high as they got during the financial crisis. I acknowledge that an MLP that has grown since then and is now higher in the IDR splits than it was six years ago deserves to be cheaper because a greater portion of its distributable cash flow is being siphoned off to its GP, but even accounting for that I think a return to financial crisis valuation levels says something about how cheap the sector has gotten.
Small float: CNNX’s float is somewhat small at <$200 million (~1/3 of the total units), making it difficult for institutional investors to step in and take advantage of the selloff.
Interestingly, CNNX is significantly cheaper than three comps that are also 100% fee-based Marcellus gathering MLPs with E&P sponsors. I will spend a good amount of time casting CNNX’s business into relief against those peers to see if any issues appear beyond CNX’s coal exposure that would justify its highly discounted relative valuation. As a quick preview, all four companies I will compare have their own plusses and minuses, but I will hopefully demonstrate that CNNX is overall not materially worse on the fundamentals of its business. In fact, CNNX is much better off than its peers in some ways.
Here are the Marcellus MLP comps. Incredibly, AM and EQM are currently the 6th and 7th lowest yielding MLPs in the entire Alerian MLP Index, which is yielding ~9%, while CNNX yields more than the index even though it is 100% fee-based, has almost no debt, has above average distribution coverage, and is at the bottom rung of the IDR splits.
The E&P sponsors for the comps are Antero Resources (NYSE: AR), EQT Corp. (NYSE: EQT), and Rice Energy (NYSE: RICE).
History of CONE Midstream
CONSOL Energy was founded as a coal company in 1864. They are located near Pittsburgh. At some point in the not too distant past they realized the long-term outlook for coal was in trouble. The central issue was their utility customers were moving away from coal towards natural gas because of economics and regulations. CNX somewhat redefined their business as being one where they serve the power generation market in a more general sense. They wanted to stay in Appalachia and they wanted opportunities they could invest a lot of capital in, which would be necessary in order to transform the company. On that basis it would make sense to develop a gas production business.
The first step was to produce gas from coal bed methane on their existing mining properties, which they began to ramp in the early 2000s. The second step was the acquisition of Dominion Resources’ Appalachian E&P business in 2010 for $3.5 billion, which included 491,000 undeveloped Marcellus Shale acres, mainly in southwest Pennsylvania. They subsequently leased up additional acreage. This occurred just as the Marcellus was heating up.
15 months later they sold half the Marcellus operation to NBL for $3.5 billion, effectively getting their cost basis out of the asset but still retaining 50%. The two companies agreed to jointly fund the development of 663,350 total acres, with NBL operating the wet gas acreage and CNX operating the dry gas acreage. CNX wanted to bring in NBL for two reasons: first, because the additional capital would help accelerate development and second, because NBL would bring shale development expertise that CNX lacked.
The state of Pennsylvania probably has an underappreciated role in the history of the oil and gas business. The world’s first commercial oil well was drilled there in 1859. However, by the time the Marcellus was unlocked in the last decade the state didn’t really have a modern E&P industry. That meant producers would have to invest capital to develop their own gathering infrastructure in order to connect their wells to the major transmission pipelines that could move their gas to market. NBL/CNX did this and so did other producers in the area like AR, EQT and RICE.
NBL/CNX called their midstream system CONE Gathering, and decided to IPO a partial interest in it as CONE Midstream. They staffed it with a mix of executives from NBL and CNX. The logic of taking it public made a lot of sense at the time. It would allow them to get a valuation on the midstream system’s cash flows of 12-13x EBITDA, whereas if those assets remained inside their corporate parents they would continue to be valued in the marketplace at 5-7x EBITDA multiples that are typical for E&P companies. Moreover, it would allow them to immediately free up >$400 million of capital by offering 1/3 of the entity to the public, and also raise additional low-cost capital in the future with further drop downs.
Importantly, CNX/NBL would still control CNNX through their ownership of the GP as well as benefit from their ownership of the incentive distribution rights.
CNNX had certain elements of its business formalized as part of the preparation for an IPO. For example, they were given the 20 year acreage dedication for the Marcellus horizon along with the right of first refusal on additional acreage starting Sep-30-2014. Note that the ROFR does not include the Utica horizon, which exists under much of the Marcellus horizon across the acreage. I will address that later.
They also established a formal pricing structure:
$0.40/mcf fixed price for volumes of dry gas transported
$0.55/mcfe for wet gas
$5/bbl for volumes of condenstate
2.5% annual escalator that kicks in Jan-01-2016
(Actual unit revenues have remained fairly constant at $0.50 – 0.52/mcfe for the last 8 quarters.)
CNNX was taken public at $22/unit with a 3.9% yield in late 2014. 1/3 of the L.P. units were offered to the public, with NBL and CNX each retaining 1/3. CNNX owns approximately 60% of the total gathering system. NBL and CNX jointly own the remaining 40%, which appears as a non-controlling interest on CNNX’s financial statements. That 40% can be dropped down to CNNX in pieces over time.
The distribution growth CAGR guided to on the roadshow was 20% for 5 years. Gas prices were over $4/mmbtu at the time, and the depth and quality of the resource within the acreage dedication made the guidance entirely plausible. Because of that perceived visibility the units traded up to a high of $31.62 on Nov-13-2014, for a yield of 2.7%.
How I Got Interested
NBL acquired Rosetta Resources (ROSE) in a 100% stock deal that closed in Jul-2015. (Some of you might recall my disastrous writeup of ROSE in Mar-2014.) I researched NBL before the acquisition closed in order to determine whether we should continue as shareholders. I took note of the basic facts surrounding CNNX’s operations but didn’t see it as an opportunity worth pursuing even though the units had fallen since the IPO. They were still barely yielding 5%.
Natural gas prices subsequently fell throughout the summer and fall, and NBL/CNX began laying rigs down. NBL talked with investors about allocating less and less capital to the Marcellus asset for 2016. MLPs in general began to sell off. CNNX’s units got smoked and traded down to a >10% yield. I felt like a 10% yield was enough of an inducement to figure out whether the security was mispriced.
Outlook for Natural Gas
Let’s back up and look at whether the market even wants more natural gas. The United States consumed approximately 70.5 bcf/d in 2014, up ~20 bcf/d in the last 10 years. According to Simmons, incremental gas demand over the next five years will be 15.8 bcf/d. There are five main drivers to their forward growth projections:
8.9 bcf/d LNG exports (these volumes are largely contracted on a take or pay basis)
2.3 bcf/d of incremental exports to Canada/Mexico
2.2 bcf/d of power generation (1.1 bcf/d coal to gas switching, 0.4 bcf/d to replace nuclear retirements, 0.7 bcf/d electricity load growth)
2.0 bcf/d to feed industrial production of chemical products (methanol, ethylene, and ammonia)
0.3 bcf/d new transportation opportunities
Whether incremental demand ends up being 15.8 bcf/d or some lower number like 5.0 bcf/d, the point is that gas demand is growing because it is cheap and abundant. If the market is growing, CNNX at least has a chance to grow by virtue of a rising tide lifting all boats, assuming new wells on its acreage dedication are economic, and assuming CNX and NBL have the money to drill and complete new wells. Even if domestic demand doesn’t absorb any more gas, the LNG export contracts will ensure that a decent amount of incremental gas will be consumed.
Relative Cost Position of the Marcellus
Assuming the market wants more gas, does the Marcellus deserve to supply it? You can look at the rankings of gas price breakevens by basin from almost any investment bank or consultancy, and the Marcellus will usually be the lowest cost basin on average, often followed by the Utica. Top down analysis like that is not enough to put the question to bed because there are sometimes wide variations among producer cost structures within a basin. For example, there could be a producer in a high cost basin that is competitive with the best basins because of some company-specific reasons. And, it’s not hard to find high cost operators in the best basins. However, the top down conclusions about the Marcellus reconcile with everything I have seen when looking at individual producers on a bottoms up basis. A number of Marcellus producers have leaseholds that work under $2/mcfe, and some that work under $1/mcfe. The tremendous capital efficiency of the Marcellus explains why it has been able to grow from just a few percent of national production in 2008 to over 20% today. Later on I will discuss how the economics of CNNX’s acreage dedication in particular stacks up against the comps.
Outlook for CNX’s Coal
Let’s also have a look at CNX’s coal business before we get more into the specific details of CNNX. Coal is an important element of this story because it accounts for ~60% of CNX’s LTM segment EBITDA as of Q3’15. CNX will need those cash flows to support the development of the Marcellus acreage.
The cornerstone for this segment is three longwall mines in Pennsylvania that produce high energy-content thermal coal (13,000 btu/lb) and have a combined 28.5 million tons of annual production capacity.
My framework for thinking about the outlook for coal is simple. Coal basically has one purpose, which is to generate electricity. That application has historically accounted for >90% of demand. Thermal coal has historically been a >1 billion ton market in the United States, and has historically fueled half of the nation’s electricity generation. Because of growing competition from low cost natural gas, and because of coal plant shutdowns driven in part by environmental regulations (as a rule of thumb one MWh of electricity fueled by coal creates ~2x the CO2 emissions created by one MWh of electricity from natural gas), coal has lost significant market share. 2015 YTD industry production is down >90 million tons vs. the same period a year ago, and coal’s share of electricity generation has fallen to ~34%.
Many regulated utilities have retrofitted their coal-fired plants with scrubbers over the last 10 years to remove SO2 and NOx emissions (scrubbers do not remove CO2), often at a cost of tens of millions or hundreds of millions of dollars. The Supreme Court has established that utility shareholders are entitled to a fair return on prudently invested capital for assets that are used and useful in providing service to customers. If scrubbed plants are abandoned, ratepayers will still be on the hook for those capital costs. It is unlikely that state regulators would choose to abandon those investments so soon after approving them.
Even if the market gods could wave a wand and mandate that all states eliminate coal-fired power generation as quickly as possible, it would still take many years. There are a number of states, particularly near CNX’s mines, where coal accounts for 80-85-90-95% of the generation mix.
If you assume 185 GW of the nation’s 300 GW of coal plants have been retrofitted with scrubbers, that those plants are the only remaining source of thermal coal demand and they run with a 60% capacity factor consistent with the asset class’s historical average, that the average heat rate of those plants is 11 mmbtu/MWh, and that CNX’s coal contains 13,000 btus per pound, the market would still require 413 million equivalent tons of their coal. That level of demand appears to intersect with the supply cost curve at an energy equivalent price of nearly $50/ton – plenty of room to make a profit given CNX’s weighted average cash costs of $35/ton (Q3’15 Pennsylvania volumes of 5.7 million tons at $33.50/ton, Q3’15 Virginia volumes of 0.9 million tons at $44.73/ton, assumes no volumes from “other” coal operations). By my math, at $50/ton they would generate ~$400 million of EBITDA. The Virginia mine actually produces metallurgical coal which is typically more valuable than thermal coal, but I am not assuming any premium in these realizations. Consensus EBITDA for the coal business is approximately $500 million per year for the next couple of years, based on the sell-side models that have been published within the last few months.
CNX has 59/45% of its 2016/17 Pennsylvania production already priced at $53 ½ and $54, along with 64/7% of its 2016/17 Virginia production priced at $56 ½ and $54.
Based on their forward contracts and their position on the cost curve, I think it’s somewhat safe to say CONSOL’s coal business will generate a decent amount of EBITDA to support the development of the E&P business over the next couple of years.
The wild card appears to be how messy the coal industry’s liquidation becomes at the middle to high end of the cost curve. One antidote to the downward pressure on prices is that utility customers value stability of supply. I am hearing examples of premiums of several dollars per ton being paid to the producers with low cost mines and decent balance sheets. I’ll ask a rhetorical question – if you were a utility, knowing that your customers’ highest concern is that the lights come on when they flip the switch, are you going to try to save a couple dollars per ton buying coal from a producer who is on the brink and may not be around to deliver? CNX can offer a surety of supply that is as good as any miner except maybe Alliance Natural Resource Partners (NYSE: ARLP).
With that behind us, let’s compare the gathering MLPs across a variety of standard metrics and see what stands out.
CNNX’s profitability is middle of the road, so they’re fine in that regard. Their distribution coverage is robust at the same time their maintenance capital spending is at the front of the pack. Sometimes you wonder if MLPs are skimping on maintenance in order to support an unsustainable distribution, but that doesn’t appear to be the case here.
The ratio of operating cash flow to capex, or the amount of total capital spending they have been able to fund with internally generated cash flow, has been somewhat low, but that is mainly a function of the fact that the entity is so new. As you would expect, EQM’s ratio is the highest since it is the most mature (they are in a phase where they can really leverage centralized infrastructure), and RMP’s is the lowest since it is the newest.
CNNX is at the very bottom rung of its IDR splits so they still have a lot of runway before they start giving up significant incremental cash flows to the GP. AM is already at the 15% incremental split and EQM is at 50%. Also, the ratio of CNNX’s highest distribution split level to the minimum quarterly distribution is 150% - the same as the Marcellus peers – so the terms of CNNX’s IDRs are just as fair to their common unit holders.
For a broader perspective beyond just the Marcellus MLP comps, the >20 constituents of the Alerian MLP ETF (NYSE: AMLP) gave up nearly 20% of their total distributable cash flow to their GPs in the most recent quarter vs. 2% at CNNX. 13 of the constituents are already at the 50% incremental split, including 4 of the biggest 6 by market cap.
Interestingly, CNNX has much less leverage than Marcellus peers and other MLPs, meaning that if things get really bad in terms of natural gas prices and organic volume growth, they have more room than peers to “manufacture” distribution growth through additional drop-downs. It’s worth noting the >20 constituents of AMLP are levered ~4x on average using 2016 consensus EBITDA vs. 0.6x at CNNX. Crucially, leverage is so low at CNNX that they could fund dropdowns 100% with debt for a while, negating the need to tap equity markets at a 9.5% yield.
Self-help growth valuation: CONE Gathering generated $150 million of gross annualized EBITDA in Q3’15, of which ~58% was net to CNNX, or $87 million. By my math, if CNNX acquired the $63 million of EBITDA they don’t already own at a 10x multiple funded with 100% debt at a 6% interest rate, then assuming a 1.1x generic distribution coverage ratio, they could increase the annual distribution by 47% from $0.912 to $1.34. Assuming an 8% distribution yield the units would be valued at $16.70, for 75% upside. I chose those figures simply because they sounded conservative.
This scenario would require $630 million of new debt ($63 million EBITDA * 10x multiple). Their $250 million unsecured revolver maturing Dec-30-2019 has $193.5 million available on it, so a decent amount of any new debt would need to be funded externally, but it’s not a tall ask. It would increase pro forma leverage to 4.6x, which is within the revolver’s 5.0x covenant. Management has said they would be comfortable going up to 5.0x for a transaction, although they would prefer to live in the range of 3.0 – 3.5x over the long-term.
EBITDA interest coverage would be 3.7x ($150mm total EBITDA divided by $40.4 million of interest, comprised of $38 million from the $630 million new debt at 6%, $2.1 million from the $56.5 million existing credit line balance at 3.75%, and $500k from the $193.5 million unused revolver capacity with a 25 bps commitment fee) vs. the minimum covenant level of 3.0x.
I am not predicting this 100% debt-funded drop down scenario will happen. Nor am I banking my thesis on it. I’m just trying to show what the maximum potential for self-help is at this point in time, and it is significant.
It’s not as if the group’s guidance is going in a positive direction while CNNX’s is going in a negative direction. AM and RMP have increased full year 2015 guidance for EBITDA and distributable cash flow once so far this year, while CNNX has increased it twice. EQM had a big increase in Q1’15, but it was driven by a drop down from EQT.
Distribution Growth Projections
Let’s have a look at consensus estimates for distribution growth over the next few years. Again, it’s not as if the group is going in a positive direction while CNNX is going in a negative direction. Their expected growth at this point in time is lower than the group’s, but is 47% cumulative distribution growth through 2018 so bad that investors need a 9.5% current yield to get them interested? 47% growth is no joke. Even if they only achieve a fraction of that growth it’s enough to make a 9.5% current yield attractive.
Since there is nothing obviously wrong with CNNX at the MLP level that I believe would fundamentally justify such a higher cost of capital than peers, let’s look at the sponsors’ E&P businesses to see if we can find trouble.
Note that all three of CNNX’s comps are sponsored by a single pure-play gas-focused Appalachian producer, whereas CNNX is jointly sponsored by a company in Noble Energy for whom the Marcellus is just one of 7 key producing areas. Noble can allocate capital amongst a number of basins, whereas the other sponsors currently only have one basin to absorb their development capital.
Sponsor Asset Bases
Let’s confirm that CNNX’s sponsors control enough resources to support a growing midstream business, or at least maintain its current size. There are several ways to evaluate this. I am leaving NBL out of this section because they don’t break out the Marcellus.
Note: The (G) stands for gross and the (N) stands for net.
One basis of comparison is proved reserves. CNX has less than AR and EQT, but this figure is on a net basis. Since NBL has approximately 2 Tcfe of proved reserves in Appalachia the amount available to support CNNX’s business is approximately 9 Tcfe (6.8 Tcfe + 2.0 Tcfe), which is a competitive figure vs. 10.7 Tcfe at EQT and 12.7 Tcfe at AR.
However, proved reserves are defined by the SEC and the definition has certain limitations. For instance, you can only book reserves if you intend to drill them within 5 years, even if they are highly economic and you can’t get to them all until year 6. Maybe you have so much awesome acreage that you couldn’t possibly hope to drill it all in 5 years, and therefore you can’t book all the hydrocarbons as proved reserves?
To get a fuller picture of the gas available to CNNX and its peers we have to look at 3P resources. Since this is not an SEC-defined term we have to rely on the companies to provide those figures using their own methodology. I’m sure E&P companies abuse this metric since there is a lot of discretion in how you come up with the number, but it should have at least some informational value. CNX has a staggering amount of 3P resources.
To be clear, the Utica horizon is not explicitly covered by CNNX’s 20 year acreage dedication, but it would be almost impossible for another midstream company to provide the gathering on those locations at a cheaper cost to the J.V. than CNNX because they underlie the Marcellus horizon on acreage where CNNX is has already built infrastructure.
Importantly, virtually all the Marcellus acreage in the CNX/NBL joint venture is already held by production (HBP’d). This is crucial. On the one hand, you could say this is bad for CNNX because if a smaller portion were held by production then it would force CNX/NBL to drill more wells and pull forward production volumes, which would drive greater volume growth and earnings at CNNX.
On the other hand, having such a high HBP percentage is better for the long-term financial health of the sponsors because it doesn’t put them in a situation where they might be forced to spend significant amounts of capital at a moment in time when they are strapped for cash (or when prices are low), and it allows them to develop their acreage more efficiently (and more cost effectively) by choosing when and where to drill on the basis of what makes the most economic sense rather than simply where they need to hold acreage.
I would not want to be AR or RICE with those low HBP levels and having to drill wells in a world where benchmark gas prices are below $2/mmbtu.
Sponsor Growth Estimates
This is where the rubber meets the road. Production volumes by the sponsors have more direct influence on the MLPs’ EBITDA from year to year than anything else.
I have excluded Noble’s consensus estimates from the tables below because Noble doesn’t give much granular standalone guidance for their Marcellus operations and there aren’t consensus estimates for that asset. The other producers in the sponsor group are more or less pure plays, so I am showing their consolidated production and capex estimates as a proxy for the underlying growth available to their MLPs even though some of it is coming from other horizons besides the Marcellus.
CNX’s consensus production growth estimates might be lower than Antero’s and Rice’s over the next couple of years, but 31% cumulative growth is solid. CNX has been guiding to 20% production growth in 2016. You don’t need any growth in order to make a 9.5% yield on CNNX attractive – you just need flat volumes – but a 9.5% yield with 20% underlying growth would be outstanding. For the record, if Henry Hub stays at or below $2/mmbtu and local northeast price stay even lower I would not expect the full 20% guidance figure to materialize.
CONSOL’s capex is expected to decline more than peers over the next couple of years, but it would be a mistake to interpret falling capex to mean falling volumes. CNX and NBL combined have over 100 drilled but uncompleted wells (DUCs), and they only need to complete 30-35 of them to hold volumes flat in 2016. For perspective, the joint venture partners tied 136 gross wells to sale over the last 12 months.
As a very general rule, drilling-related activities account for 1/3 of the total cost of shale wells and completion activities account for 2/3 of the cost. The nice thing about completing DUCs is that the drilling cost is already paid for, so the decision to complete them allows you to tie a new well to sales while only laying out 2/3 of the total cost. (Again, that is a very general rule).
CNX/NBL were running 10 rigs at the time of the IPO last fall. I mentioned earlier they are not running any rigs at this point. In contrast, AR is still running 10 rigs in Appalachia and RICE is running 5 (dropping to 3 in 2016). I think people look at the disparity in rig counts between the sponsors and see it as an easy to way draw a contrast between the MLPs. It’s easy to take a mental shortcut and conclude CNNX’s volumes are about to go into free fall while volumes at the other MLPs are still growing up and to the right. The shortcut ignores the nuance of the DUC inventory. It also ignores the fact that the wells on CNNX’s acreage dedication are on par with the comps, as I will discuss later. Their acreage is just as worthy of capital investment as anyone else’s. Given the relative cost parity of the acreage, I think it’s fair to question whether storming ahead with a growth agenda when gas prices are ~$2/mmbtu is driven by considerations other than pure economics, e.g. low HBP levels. I think it’s more a sign of weakness than a sign of strength. I’ll just have to agree to disagree with the market for now.
Overall, the joint venture partners can plan for a severe drop in capex in 2016 without subjecting CNNX to a decline in volumes. It’s possible they would have to complete more than 30-35 wells to hold production flat in 2017 because production has been growing so quickly until now (at Q3’15 CNX’s net LTM Marcellus Segment production was +46.3% YoY; NBL’s was + 32.4% YoY) that they have a big growth hump to offset. I don’t know what CNX’s Marcellus segment decline rate is. A number like 35% wouldn’t surprise me. If you just hold volumes flat for a couple years the corporate decline rate will come way down as the base of wells matures. In other words, it’ll require less and less capital to hold volumes flat as time goes by after they get through the hump. Shale wells decline very quickly at first and then the decline rates settle down after a couple of years.
CNX’s E&P-specific capex guidance for 2016 is $400 – 500 million, down 44% YoY. NBL will be spending some amount of money too (TBD). I believe completing the 30-35 wells required to hold production flat in 2016 would cost <$200 million.
It’s worth asking how much of the E&P comp group’s capex will be funded through internally generated operating cash flows, vs. external sources like new borrowings or asset sales. That’ll tell us something about the likelihood the sponsors will actually be able to invest that capital. Interestingly, CNX is the only E&P in the group expected to be FCF positive over the next couple of years, and NBL has the smallest outspend of the remaining companies. I just don’t see the economic logic of significantly outspending cash flow to drill and complete new wells when gas prices are ~$2/mmbtu and most of the production from a shale well comes in the first few years.
Low gas prices would probably rank high on investors’ list of concerns about the Marcellus right now. Even though CNX and NBL have great acreage and lots of it, if they were relatively unhedged for 2016/17 while the group was very well hedged I could see that as a possible justification for CNNX trading at a higher yield than peers.
The table below shows the portion of consensus consolidated production estimates for each sponsor that is already covered by fixed price hedges.
A few things stand out to me. First, while CNX and NBL are virtually wide open for 2017 and 2018, so is EQT Corp., whose MLP is yielding 600 bps less than CNNX. Second, RICE has significant open exposure given the level of outspend they are expected to generate over the next few years. Third, AR is very well hedged. I think it’s fair to say CNNX is at a disadvantage relative to AM and RMP when it comes to the sponsors’ hedge protection, but that disadvantage is tempered by the fact that AM and RMP intend to significantly outspend cash flow in a very weak commodity price environment, and by the fact that CNX/NBL have more flexibility in dialing back capex without losing their leaseholds due to their high HBP levels.
Let’s look at the sponsors’ leverage. It should give us additional insights as to their capacity to keep investing in their acreage and driving or maintaining their production volumes.
CNX is levered on 2016 EBITDA by more than a turn over the next closest peer on an “as reported” basis. I like to adjust the leverage for the after-tax values of their hedge books, as if they were to immediately close out all their profitable hedges and use the after-tax proceeds to pay down debt. Under that methodology, Antero’s pro forma leverage improves the most, falling to 1.8x because of their massive hedge book. CNX only falls to 4.3x – high but manageable. NBL falls modestly to 2.8x. (The table does not include the value of NBL’s Q4’15 hedges; currently $180 million after-tax; would drop them to 2.7x.)
However, I don’t think modifying the leverage with the just hedge gains is the whole story. Producers sometimes enter into firm transportation (FT) contracts for takeaway pipeline capacity. These obligations are take or pay. They can be like assets when local prices are weak relative to benchmarks and the firm capacity gives you access to distant markets with higher prices, but if you can’t use the capacity for some reason and you still have to pay, well, you’ll be in a world of hurt. Just ask Chesapeake.
I calculate the NPV of FT using a 7% discount rate (it’s a nice middle of the road number – nothing scientific) in order to measure the debt equivalents associated with the obligations, and I add it to the pro forma debt to arrive at a fully adjusted leverage figure. On that basis, AR’s adjusted leverage blows out, and CNX/NBL are the lowest in the group because of their modest FT obligations. To be fair, AR and RICE could probably sell some of their firm transportation to other producers if they couldn’t use it all, but would it be for a high enough price to cover their costs and make them whole? It seems like the MLP market’s attitude towards these enormous obligations is, “what could go wrong?” The credit markets seem to have noticed, however. It’s interesting that AM only yields 20 bps more than EQM even though AR’s bonds are yielding several hundred basis points more than EQT’s.
Also worth noting are the companies’ relative levels of total interest bearing debt per mcfe of PDP reserves at YE2014. CNX and NBL are very good in this regard. Also, CNX and NBL are right in line with the group in terms of their credit ratings.
CNX is in the process of trying to sell $1.0 – 1.4 billion worth of E&P assets. It’s an open question what kind of valuations they will get in this weak commodity price environment, but if they can pull off deals at half-decent valuations they should be able to make a material dent in their $3.7 billion of total debt.
In summary, CNX and NBL have enough debt that it needs to be respected, but their credit profile doesn’t suggest they are impaired in terms of their ability to develop their Marcellus assets and drive volumes through CNNX.
Sponsor Cost Structures
There is a lot of shale gas in this country. The cost curve for the industry is relatively flat compared to other commodities, although it is generally accepted that the Marcellus (and also the Utica) reside at the low end of the curve. Given the possibility that gas prices remain low for an extended period of time, I wanted to see if the acreage in the CNX/NBL joint venture is low cost enough to justify continued investment and maintain production. In other words, do they have the same “right” to produce as their Marcellus comps?
I looked at it from two perspectives. First, I looked at it on the basis of current unit costs reported in their GAAP financial statements. That will tell us where they stand currently. Second, I compared their individual well economics using data from their investor presentations. The industry is continuously improving. Individual well data should be more leading edge (i.e. taking into account increased efficiencies and dramatic service company price reductions in the last year), whereas the GAAP data is more influenced by historical costs. Unfortunately, individual well economics only reflect half cycle returns, whereas the GAAP figures give you a better picture of full cycle costs. If you compare the producers using both sets of data and CNX looks good both ways, then it’s probably a good indication that their rock is good enough to justify investment as long as commodity prices are supportive.
NBL doesn’t break out Marcellus segment unit economics, but CNX does (along with Utica segment economics), so I used CNX’s figures as proxy for the JV as a whole.
On the GAAP figures below, CNX’s Marcellus and Utica segments look very competitive.
CNX provides five different type curves for their Marcellus acreage so I have broken them out in a separate table. I would also like to note that the CNX well cost figures are from Jan-2015. The competitiveness of their wells on a money forward basis is likely understated by these figures since they probably do not fully incorporate the benefits of service cost deflation and further efficiencies that have occurred in the last year, whereas the well costs for the other three producers are very recent.
In spite of that handicap, CNX’s Marcellus wells look very competitive.
CNNX has significant upside almost no matter how you want to value the units. Here are a few ways I have looked at it.
If CNNX was simply valued on the basis of the $0.228 current quarterly distribution using RMP’s 5.9% yield (the highest of the three comps), it would be worth $15 ½ immediately (60% upside)
If the current quarterly distribution were capitalized at the Alerian MLP Index’s historical yield spread of approximately 400bps over the 10 year Treasury’s current YTM of 2.3%, you would get to a 6.4% yield and a $14 ¼ valuation (50% upside). Coincidentally, you would get a similar result if you applied the Alerian MLP Index’s historical average yield spread of 100bps over the Moody’s Baa-rated corporate bond index’s current yield of 5.5%
The Moody’s index had an average absolute yield over the last 15 years of 6.4%. Adding the 100bps historical MLP spread and capitalizing CNNX’s current distribution at 7.4% would value the units at $12.30 (30% upside)
The 10 year Treasury had an average yield over the last 15 years of 4.1%. Adding the 400bps historical spread and capitalizing CNNX’s current distribution at 8.1% would value the units at $11 ¼ (18% upside)
Applying the Marcellus MLP comp group’s average yield of 4.6% to CNNX’s current quarterly distribution of $0.228/unit would make the units worth $20 (>100% upside)
In the scenario I outlined earlier where CONE Gathering drops down the remaining ~40% of the gathering system, CNNX would be worth >$16 on a generic 8% yield (65% upside)
If CNNX simply grew the quarterly distribution at a 10% CAGR for the next two years, from $0.228/unit to $0.276/unit, and received the group’s average yield of 4.6%, the units would be worth $24 (150% upside)
A 15% CAGR under the same scenario would value the units at $26 (170% upside)
As I mentioned at the very beginning, even if the distribution got cut 10% and had a 1.1x coverage ratio (implying a 30% reduction in distributable cash flow), valuing it with a 12% yield would make the units worth $6.80 (27% downside).
Maybe that downside case is not harsh enough. I can come up with a scenario that is worse. You can always be incrementally more pessimistic than the most recent bear case until you get to zero, but our job as investors is to assess the range of outcomes along with their probabilities. When I lay out the facts about this security, it is my judgment that it has much greater upside than downside, and that the probability of an attractive return is high. I have made my bet and now it’s time to see if I’m right.
If you are unsure of my bull case for CNNX but are still struck by the relative valuation gap vs. its peers, another way to play would be to pair a long position in CNNX with a short position in AM. However, I am not doing that because I think gas prices are at the low end of their likely range. If they go higher AM will probably go higher too and I don’t want to give away some of my profits on CNNX to a short position in AM.
CNX Bankruptcy: It would be bad for CNNX if some combination of low coal and/or natural gas prices persisted for long enough that CNX was forced into bankruptcy. CNX’s bonds are yielding 16-17%. However, the existing wells would still produce for the benefit of the creditors, and those volumes would still be forced to flow through CNNX’s gathering system. If nothing else the situation would create uncertainty around CNNX among investors and depress the valuation. It would also slow down the development of the E&P assets, possibly to the point where volumes would decline.
However, CNNX had a distribution coverage ratio of 1.44x in the last quarter so there is some room for volumes to fall before having to cut the distribution, assuming a coverage floor of 1.0x. Assuming all the lost revenue of $0.51/mcfe fell straight to the bottom line, gathered volumes could fall by 128 mmcfe/d net to CNNX, or by 220 mmcfe/d grossed up for the 40% noncontrolling interest. Gross gathered volumes in Q3’15 were 1,156 mmcfe/d, so gross volumes could fall by 19% (220/1,156) before having to effect a distribution cut. In reality, the company would probably cut the distribution in advance of crossing that threshold.
Admittedly I have not studied the history of how gathering MLPs have fared during the bankruptcies of producers they serve. My understanding is that these MLPs are generally in a strong commercial and legal position because it is essential for the value of the estate that the MLP’s assets continue to move product to market and generate cash flows. Gathering companies have apparently continued to get paid in recent bankruptcies and continue to be treated as preferred vendors.
Fee schedule reductions: I am also focused on the theoretical risk of a downward repricing to CNNX’s fee schedule. The gathering fees were originally set at a level that would allow the midstream entity to earn a mid-teens ROIC. The gathering agreement with CNX/NBL calls for an annual historical and 10 year forward ROIC analysis. If the ROICs are too far out of whack with the mid-teens target, then either side has the opportunity to ask for an adjustment. CONE Gathering’s LTM ROIC as of Q3’15 was 15.6%, so it’s right in the middle of the fairway. However, most MLPs do not have boards that represent the interests of common L.P. unitholders. In this case CNNX is clearly controlled by CNX and NBL so there is risk that if push came to shove on pricing that the sponsors’ financial interests would win out at the expense of CNNX’s. And surely the fee schedule could be revisited during a CNX bankruptcy. For what it’s worth, NBL has their own midstream assets in the DJ Basin that they want to bring public (they recently pulled the offering due to market conditions), so you would think they have an incentive to be reasonable when it comes to the treatment of CNNX in order to protect their reputation as a sponsor.
The sponsors get acquired: Just as a thought exercise, let’s assume both the sponsors were acquired by a much bigger producer. A producer that was bigger than NBL would likely be operating in many basins and have many options with respect to capital allocation. They may choose to allocate to so little capital to the Marcellus that production goes into long-term decline. Since the acreage is >90% held by production they could allow the declines to begin without giving up the future optionality of the acreage. That would be a clear negative for CNNX’s cash flows.
The provision of this report does not constitute a recommendation to buy or sell the security discussed herein. The report is an example of the author’s / authors’ company write-ups / research process; its breadth and coverage may differ materially from other such reports. Certain statements reflect the opinions of the author(s) as of the date written, are forward-looking and/or based on current expectations, projections, and/or information currently available to author(s). The author(s) cannot assure future results and disclaim any obligation to update or alter any statistical data and/or references thereto, as well as any forward-looking statements, whether as a result of new information, future events, or otherwise. Such statements/information may not be accurate over the long-term.
|Subject||Re: bigger picture|
|Entry||12/30/2015 05:28 PM|
Thank you for your kind words. Coincidentally, I read the Raging Capital presentation yesterday. CEQP is a much different animal than CNNX - certainly more complicated. I don't know enough about it yet to have an opinion. There are probably many babies that have been thrown out with the bathwater in MLP land. You could burn a lot of man hours sifting through them all. CNNX happened to fall into my lap. Given my time constraints I have been buying AMLP aggressively.
|Entry||01/01/2016 06:33 PM|
Thank you for your questions.
Just to be clear, CNNX has not received any minimum volume commitments from its sponsors, so it’s not as if CNX has some sort of fixed dollar obligation it might choke on if it can’t produce enough gas. The only thing that is fixed is the price per unit in the fee schedule.
I think the benefits to CNX of a fee reduction would be more than offset by the destruction of value it would cause to its stake in CNNX. The units owned by CNX and NBL are actually subordinated units. The subordination period lasts until Sep-30-2017. The biggest fee reduction CNX/NBL could obtain from CNNX without causing the distribution to fall below the minimum quarterly level is $0.11/mcfe, assuming a 1.0x distribution coverage ratio and all else equal. Last quarter’s pricing was $0.51/mcfe, so that would be a ~20% reduction.
If they got a fee reduction of $0.12/mcfe or greater, it would save CNX at least $25 million per year. (That’s equivalent to 4.6% of CNX’s Q3’15 annualized EBITDA of $544 million.) However, it would cause the distributions on the subordinated units owned by CNX to cease, costing them $17.5 million per year in distributions.
What if they just got the fees cut by $0.11/mcfe and saved themselves a little bit less than the $25 million but without having to forgo the distributions? That would have other knock-on effects.
CNNX’s Q3’15 annualized EBITDA of $87 million would drop by $27 million to $60 million. The fee reduction would probably cause a further loss of confidence in the marketplace. CNNX is currently trading at 7.3x EBITDA, or $9.85 per unit. What if the multiple compressed to 5x? The combination of the lower multiple on reduced EBITDA would value the units at $4.20, causing the value of CNX’s stake in CNNX to drop by $110 million.
It would also compromise the ability of CNX/NBL to drop down additional interests in CONE Gathering into CNX. As of Q3’15 there was $63 million of annualized EBITDA that could still be acquired. A 10x multiple would generate cash proceeds of $630 million, or $315 million net to CNX for their half interest. The $63 million would fall by $19.5 million to $43.5 million. They wouldn’t be able to command a 10x multiple I assumed in the writeup if the common units were trading at 5x. Let’s be charitable and give them an acquisition multiple of 8x. 8x on $42 million would generate cash proceeds of $350 million, or $175 million net to CNX. $175 million instead of $315 million equates to value leakage of $140 million.
Now we’re up to $250 million in value leakage as a result of a $0.11/mcfe fee reduction.
There’s more. All else equal, if CNX/NBL dropped the remaining ~40% of CONE Gathering into CNNX on the terms I assume in the writeup, then on current volumes the GP/IDRs would be generating $9.8 million of cash flow. Capitalized at 6%, those cash flows would be worth $165 million, or $82.5 million net to CNX. With a little bit of organic volume growth the GP/IDRs could become worth substantially more because at that point the IDRs would be in the high splits. With a fee reduction the GP/IDRs would basically be worthless. Yeah, they would still be generating a little bit of cash flow, but who would honestly want to buy them?
Now we’re up to at least $332.5 million of value leakage as a result of a $0.11/mcfe fee reduction.
If CNX/NBL leave the fees alone and proceed in a manner that is supportive of the MLP, the valuation should eventually re-rate. The $1.34/unit distribution I modeled in the self-help/drop down scenario could reasonably be expected to get capitalized at RMP’s distribution yield of 5.9%, valuing the units at $22 1/2, up from $9.85 currently. That kind of uplift would be worth $243 million to CNX’s 19.1 million units.
In summary, CNX can save $25 million per year by obtaining a $0.11/mcfe fee reduction, but there is nearly $600 million of total value that would go out the window. I think it’s a no-brainer to leave the fee schedule alone.
|Subject||New CNX guidance|
|Entry||01/08/2016 05:25 PM|
CNX released 2016 capital and production guidance on Wednesday. They are guiding for 15% production growth YoY vs. 20% previously, but on just $205-325 million of E&P capex vs. $400-500 million previously. The efficiency gains continue. The 15% production growth assumes 73 gross completions, compared with the 30-35 completions I had discussed as being necessary to hold production flat.
Overall I thought the guidance was positive for CNNX.
|Subject||Re: New CNX guidance|
|Entry||01/12/2016 02:15 PM|
Thanks for the report. Do you have an opinion on CNXC? It trades like it is going to file, but I sure don't see that anytime soon. Or, a fund has redemptions and has to sell at any price.
|Subject||Re: Re: New CNX guidance|
|Entry||01/12/2016 02:25 PM|
I don't have a current opinion. I looked at it briefly on the IPO and concluded that it had more leverage than I would have preferred. As a general observation it seems like there is a lot of fear-based selling across resource industries right now.
|Subject||Distribution increased 3.6%|
|Entry||01/25/2016 05:10 PM|
They raised the distribution by 3.6%, from $0.228/quarter to $0.2362/quarter. The units are now yielding 10.3%.
|Entry||08/22/2016 05:12 PM|
The units have clearly re-rated since my initial writeup. At this point I think investors
CNNX is trading at 11.3x distributable cash flow vs. 14.4x for the group. With that