Description

2017.10.28 ARLP ($20.35) Long Thesis

I believe Alliance Resource Partners, L.P. (ticker ARLP) is an interesting long.  ARLP is a thermal coal producer operating in the Illinois Basin (“ILB”) and Northern Appalachia (“NAPP”).  Historically, Alliance Holdings GP, L.P. (ticker AHGP) owned the GP and incentive distribution rights (“IDRs”), along with ~31mm units of ARLP.

ARLP was written up by Gandalf in Feb 2015.  AHGP was written up by thistle933 in Dec 2009, Dec 2012, and April 2016.  The write-ups and message threads provide some really excellent background.

In short, I believe ARLP has over 60% upside over 2 years in a base case ($29 share price plus >$4 of distributions along the way; I’ve cut and pasted summary numbers at the end).  A reasonable bull case produces more than a double in that time frame.  A reasonable bear case implies a share price decline of nearly 40%, but distributions soften the blow to a total loss of ~25%.  Of course, when discussing coal, a draconian bear case is never out of the question.  On the other hand, a blue sky case could drive nearly a triple.

The structure of Alliance dramatically changed just 3 months ago.  ARLP issued another 56.1mm MLP units to AHGP and the IDRs were retired.  AHGP is now essentially a shell holding units of ARLP.  This was step 1 in what will ultimately be the simplification of the two publicly traded entities into just ARLP.

I believe that the opportunity today exists for several reasons.  First, outside of hedge funds trafficking in distressed coal debt and reorg equities arising from such debt, coal is not a sector that is currently “in favor” in the public markets.  Nor is coal generally a good topic of conversation in polite company, unless one bashes it.  It would be hard to think of a more hated industry amongst the population broadly.  Moreover, while met coal is volatile, met producers have been minting money of late.  (Look no further than HCC’s upcoming $11+ per share special dividend coming soon on a share price of $26.01 – after another special dividend paid earlier this year to sponsors pre-IPO).  But ARLP is a thermal coal company.  Thermal coal has been in significant decline.  So it’s bad enough that ARLP is a thermal coal company, but it is an MLP, making it difficult or impossible for a large chunk of institutional investors to own it or cover it.  The MLP sector as a whole is also out of favor, as demonstrated by the Alerian Index total return YTD of negative 10% in the face of QQQ’s total return YTD pushing +29%.  Investor flows have not been positive for the MLP sector, unsurprisingly.  And among MLP investors, ARLP is largely orphaned because it is a coal MLP, not a traditional midstream gas MLP.  Of course, ARLP now has over a $2.5B market cap, but currently only ~1/3 is free float, or less than $1 billion.  I think it is an easy argument to make that there are plausible “technical” or non-fundamental reasons for ARLP to be cheap today.  Suffice it to say that I am not pitching the momentum flavor of the month with this one.

However, that still leaves the fundamental reason for why I believe ARLP is cheap today.  After all, ARLP is trading at a 9.9% distribution yield and at over a 16% levered free cash flow yield, despite less than a single turn of debt leverage.  At the midpoint of fy17 guidance, it is trading at 5.1x EBITDA and 6.7x EBITDA less capex.  The market is implicitly betting that ARLP’s earnings power and cash flows will decline very substantially in the next several years, as price realizations fall further.  

Joe Craft is the owner operator leading ARLP and the prior write-ups do an excellent job discussing his massive ownership and historical capital allocation prowess.  Just at a high level, ARLP has CAGR’d at nearly 17% in the 18+ years since IPO until now.  The scale, earnings power, and distributions have grown exponentially over the years, yet the share count was essentially flat since 2003 (aside from the recent jump in MLP units in exchange for retiring the IDRs, which didn’t materially alter the earnings power per MLP unit, etc).  

I view Joe as proactive, conservative, and prudent.  The most recent annual letter clearly demonstrates it.  As an owner operator, Joe is focused on the long term.  For instance, after years of growing ARLP’s distribution, ARLP slashed it 35% in Q2’16 in one fell swoop.  In a year when peers were filing for or making their way through bankruptcy, ARLP paid down $269 million of debt in 2016 and brought their DCF coverage ratio to nearly 2x.  ARLP paid up a bit to term out the bulk of their debt to 2025 with basically no covenants.  ARLP also slashed production and cut costs significantly.

In connection with step 1 of the simplification announced 3 months ago, ARLP also announced a 14% increase in its distribution, and noted that it would be back to growing its distribution again going forward.  ARLP also noted that after historically feeling uncomfortable repurchasing MLP units (largely due to perceived or real conflicts between ARLP and AHGP), it would now be in a position to repurchase MLP units going forward.  (They also noted the ability to issue MLP units now in connection with potential acquisitions, again now that conflicts that would have necessarily arisen between ARLP and AHGP have now been eliminated with the retirement of the IDRs).  

The bearish interpretation of the transaction exchanging the IDRs for MLP units is that Joe is bearish on the long term prospects for coal and ARLP’s earnings power.  After all, the vast majority of his ownership has been in AHGP with IDRs in the high splits, and now he has moved to MLP units where you’d prefer to be situated if cash flows decline.  There may be truth to that, but there was no need to immediately raise the distribution by 14% and remark that going forward it would be back to growing again.  Moreover, ARLP has talked about being comfortable historically at ~1.5x DCF coverage ratio and the lowest annual coverage ratio going back to 2012 was 1.50x.  ARLP also speaks to the board looking at distributions on the basis of a 5 year plan, contemplating the ability to pay it comfortably even in a stress scenario.  While no one has a crystal ball, and while ARLP did slash the distribution once before, materially raising the distribution as they just did 3 months ago doesn’t seem like the first course of action to take if they were truly all beared up on the longer term outlook.  It seems more likely that Joe believes EBITDA and cash flows will be higher over time, not lower.

In conferences and road shows since then, management has made it clear that they have been befuddled by the market’s reaction to their news as they expected a hugely positive reaction.  They expected the distribution yield to compress, not expand.  Then, after yesterday’s close, ARLP raised the distribution further, but only by 1% sequentially this time.  It is possible that the outlook has suddenly deteriorated so markedly that ARLP has lost much of its recently newfound confidence in the future.  But I think the more likely scenario is that management has decided they are not getting credit for the distribution and that the share price is cheap.  Now that unit repurchases are a tool they can use, it seems to me the 1% sequential distribution increase was purposefully done to be able to demonstrate continued “growth” of the distribution, but at essentially the lowest possible rate.  We will find out in the next year or two as we take stock of where the excess cash gets invested over that period of time, but I suspect that this latest distribution announcement was done to maximize the amount of cash to be invested in repurchases (or opportunistic acquisitions), while still being consistent with prior comments about the distribution being back on the path of growth again.

ARLP unsecured bonds due 2025 have traded up since issuance earlier this year, now yielding 6.25% to maturity.  While they have a significant equity cushion beneath them, these bonds have no covenants (except a change of control put that is worthless today).  It seems clear that the bond market and equity market view the issuer quite differently.  Not only is the equity distribution yield much greater than the bond yield, but it is growing, while the bond coupon is fixed.  For direct US taxable owners, the MLP units are also tax advantaged while the bonds produce ordinary income.  Granted, many would argue that the bond markets today aren’t rationally priced, but I still thought the comparison between ARLP bonds and units was worth making.

Ultimately, I think ARLP comes down to the future cash flows.  The market is saying cash flows are going to collapse in the coming years.  But if the current 16%+ levered FCF yield with less than 1x of debt is sustainable, then ARLP will be an excellent investment.  With that much cash and that little debt, the potential scenarios would all be good ones.  The cash can be returned to stockholders or invested accretively.  Or perhaps ARLP gets bought out, by a strategic or financial buyer.  (ARLP essentially began as an MBO led by Joe Craft).  At a high level, ARLP cash flows come down to volumes, price, and costs.  

On the volume side, I think there is more upside than downside from here for ARLP.  ARLP sold <37 mt in 2016 after slashing production.  2017 guidance points to 39 mt at the midpoint, and ARLP has the capacity to be producing ~50 mt.  My base case assumes 40 mt of volume in fy19, compared to 44 mt in my bull case and 35 mt in my bear case.  

On the cost side, ARLP has demonstrated very strong cost control.  EBITDA expense per ton has gone from $38.15 in 2012 to $31.09 in 2016 and $27.74 in the first half of 2017.  Much of this cost improvement has come from a positive mix shift of production to lower cost mines, with a greater mix of longwall production and where ARLP still has significant unutilized capacity.  I assume costs ~$29 per ton in each of 2017-2019 across my different cases.  If ARLP can grow its volumes, its costs should benefit from a still higher mix of production at its lowest cost mines, separate from fixed cost leverage.  On the other hand, royalties and production taxes are tied to pricing.

That leaves price as the key driver and risk – both to the downside and potentially the upside.  I would note that ARLP price per ton has fallen substantially from $56.28 for fy12 to $50.76 in fy16 and $45.42 in H1’17.  My base case assumes further erosion to $43.50 in fy18 and $42.50 in fy19.  My bull case assumes $45 in fy18 and $45.50 in fy19.  My bear case assumes $42 in fy18 and $37.50 in fy19.

As such, Coal EBITDA / ton sold goes from $20.77 in fy14 and $19.67 in fy16 to fy19 levels of $13.75, $16.62, and $9.07 in my base, bull, and bear cases, respectively.

I think it’s also important to note that coal stockpiles are down materially, despite a relatively unprecedented series of mild winters and the latest mild summer.  I think the biggest risk in the shorter run is that coal stockpiles and unpriced tons are largely still high enough that utilities have bargaining leverage.  But fundamentally, the fact that stockpiles have come down materially in spite of the weather is encouraging.  In theory, weather can also surprise in the opposite direction.  An unusually cold winter (or two, please!) along with utilities shift towards more spot buying and last-minute contracting could materially shift the bargaining leverage to the coal producers.

Meanwhile, we should get some resolution in the coming months.  A number of high priced legacy contracts that have kept high cost and stressed ILB operators producing tons will be expiring at the end of December.  It remains to be seen whether those utility customers decide to move those tons from the current producers to better capitalized and “safer” suppliers like ARLP, and drive those high cost producers to shut in those tons, or if they insist on keeping those tons in place at higher prices to preserve more sources of supply.  But it seems some decisions in that regard have already been made that should be beneficial to ARLP and based on recent higher cost competitor mine closure announcements.  And if API2 stays strong to continue incentivizing significant ILB exports, the supply / demand balance for ILB could surprise positively and be tighter than expected for 2018.

I won’t spend much time thinking about a blue sky scenario, especially since I’m no natural gas expert.  But I will briefly wonder if the US could ever potentially export enough gas to drive a world gas price?  It seems implausible today, but many extremes in general commodity and specifically natural gas pricing have seemed implausible before they happened.  Moreover, what if the odds of this happening don’t rise to the point of it ever being expected per se, but merely to the point where stock prices discount the probability being something noticeably north of zero?

It seems politics may represent more free, albeit remote, call options.  First, repeal of the Clean Power Plan may not dramatically slow or prevent further coal fired power plant retirements, but it can’t hurt in that regard.  Rick Perry’s letter to FERC around reliability and resilience would seem to be a clear boon for coal producers if anything comes of it.  While the reaction from the gas and environmentalist lobby and was swift and strong, I was surprised that ARLP units did not react positively at all.  While the odds seem very low, the impact is potentially tremendous, and not priced in today.

Even tax reform could be beneficial for ARLP.  If pass through entities get a 25% tax rate for their income passed through to their owners, that would represent a materially lower rate than the 39.6% many high net worth MLP owners pay on the ordinary income portion they receive annually and on that ordinary income recapture upon liquidation or exit from the MLP (as a result of being passed through tax depreciation and depletion in prior years of owning the MLP).  And if the interest expense deduction for corporate America gets ratcheted back and drives less future supply / issuance of corporate debt, then presumably fixed income proxies like MLPs would see their yields tighten, all else equal.

Catalysts

Cash flows over time

ARLP unit repurchases

Step 2 fully collapses AHGP into ARLP sometime in 2018, driving significantly better trading liquidity as free float essentially doubles

Risks

Mild weather in the short term

Long term naturally gas prices surprise materially to the downside

ARLP is the last man standing in thermal coal, but with reduced earnings power

Historical and Case Summaries

Catalyst

Cash flows over time

ARLP unit repurchases

Step 2 fully collapses AHGP into ARLP sometime in 2018, driving significantly better trading liquidity as free float essentially doubles