Coal miners in the U.S. are unloved – Patriot is in bankruptcy, Wilbur Ross is telling investors to stay away, and the Obama administration is trying hard to shut coal plants down.  But a cost-based analysis of the market suggests that Alliance (either the LP – ticker ARLP – or the GP – ticker AHGP) is a good investment at these prices.

The Illinois Basin is one of the four main coal-producing regions in the U.S.  Coal has been produced in the ILB since the 18^th century.  The central location of the ILB mines, and the rivers running through the ILB, have meant low transportation costs to big utilities like the TVA and Southern Company.  Rivers are an exceptionally cheap form of transportation – up to 10x cheaper than by rail – and provide access both throughout the eastern U.S. and to overseas markets through terminals near New Orleans.

By the 1970s, ILB coal production was ~140 million tons/year, out of ~600 million total U.S. coal production.  But then the high sulfur content of the coal gave us acid rain, and the Clean Air Act of 1990 imposed a cap-and-trade system for sulfur emissions.  The increased cost of burning ILB coal relative to low-sulfur CAPP and Powder River Basin coal drove ILB production down below 100mm tons, even as total U.S. production grew to 1.1 billion tons.  PRB coal filled much of the gap, with most of the value accruing to the railroads (it costs about $10/ton to scoop it out of an open pit, but the pit is in the middle of Wyoming – a key reason why Buffett owns a railroad).

But sulfur emission trading led utilities to explore different means of lowering their costs, which stimulated the development of scrubber technology.  Today, 58% of U.S. electricity produced from coal is generated by a plant with a scrubber installed.  Installation of a scrubber is funded by an increase in rate base, and operating costs are minimal, especially when offset against sales of the sulfur to chemical companies.  Scrubbers have changed high-sulfur coal from a pollutant into a cost-reducer.

At the same time, CAPP miners exhausted the easy-to-get coal.  As you dig deeper into the ground, costs and risks increase.  And sometimes the six foot seam splits into two three-foot seams, crushing your production yields.  Cash costs per ton for a top-notch CAPP mine are now north of $60/ton (Alliance’s two CAPP mines, which are now just 5% of current production, have been around $60, but spiked in the third quarter to $68).  For a less-than-great mine, they exceed $70/ton.  In the ILB, costs are $30-32/ton for a room-and-pillar underground mine, or in the 20s for a long wall, or a surface mine.

Faced with adverse cost trends, many CAPP producers skimped on safety.  The Upper Big Branch disaster in April 2010 was the worst in U.S. coal mining since 1970.  It killed 29 miners, put Massey on the block, and provided even more justification for the Obama administration to crack down on CAPP companies.  While MSHA has made it more difficult in the ILB as well, their restrictions are particularly expensive for an older mine operating at deeper depths.  And it’s gotten harder to blow the tops off of mountains in West Virginia.  CAPP coal is unlikely to get cheaper to mine anytime soon. 

At a cash cost of $70/ton, you need a selling price of at least $80/ton to cover maintenance cap ex and generate a small profit.  But $80/ton for CAPP coal implies a natural gas price of $4.50-4.75.  Utilities are paying $3.80/Mcf for one-year-out natural gas, and so there has been a lot of coal-to-gas switching.  EIA data show total Appalachian production (Central and Northern) down to 336 million tons in 2011, from 488 million in 2000.  And production for 2012 is likely to be lower.

By contrast, ILB production was 116 million tons in 2011, up 10% from 2010.  Alliance and Peabody management have stated that their current ILB selling prices of $51-52/ton are competitive with natural gas at $3.25-3.50/Mcf.  Analysis of likely plant dispatch decisions by a TVA or Southern Company utility manager confirm this math.

With ILB cash costs of $32/ton, Alliance is making $20 of cash contribution margin per ton sold.  Maintenance cap ex is about $5/ton, and there are central SG&A costs of ~$2/ton leaving $13/ton of unlevered FCF.  This makes management’s long-held goal of 30% cash-on-cash returns for growth spending on new mines seem reasonable.  Cash build cost of adding a ton of production was $40 in the mid-2000s.  With cost inflation, and the more capital-intensive (but ultimately lower operating cost) long walls going into the new Tunnel Ridge and White Oak mines, cost has likely risen to north of $50/ton of capacity.  $13 of cash flow is a 26% return on $50 of cap ex.  And the one-year forward gas price of $3.80 suggests some room to push prices up on new production from $52/ton.

ARLP has been one of the best performing stocks on U.S. exchanges since it came public in 1999.  From $8/share it has grown to $55/share, while paying out $24 in dividends, and never raising equity capital.  Leverage has remained modest, currently at 1.3x EBITDA, compared with 3.1x for Peabody, and 3.8x for Arch.  But return on capital (EBIT over debt plus equity) is 28% for Alliance, versus 14% for Peabody, and 4% for Arch.  This disparity – at least 2x the ROIC of the next best player – has been true for more than a decade.

Why?  Key is management’s operating skill, which has led to minimal safety issues and a well-paid, non-union workforce.  These achievements have enabled Alliance to open a new mine only when a significant portion (more than half) of production has been contracted at attractive prices.  While pricing in contracts is only firm for the first year, and then tails off over years two and three, Alliance’s record of safety and reliability has meant no force majeure events to stress out utility managers worried about keeping the lights on.  And during periods of high spot relative to contracted prices, Alliance management has refrained from declaring “price majeure”, unlike many other players in this industry.  Reliability means a lot to utility managers thinking in decades.  We have been told by a manager at TVA that there are only three coal miners he wants to rely on – Alliance is one of them.  And this fact has allowed them to get predictable volumes and good pricing, driving the high returns on capital over time.

Underpinning all of this is CEO Joe Craft’s savvy decision to build capacity in the ILB when it was out of style.  A $30-40/ton cost advantage is a handy thing to have when your 30% return target only needs a $20/ton contribution margin.  It has paid off handsomely for him.  He owns 22 million units of AHGP, or 37% of units outstanding.  At a current dividend rate of $2.88/unit, he is receiving $63 million of annual dividends, compared to his 2011 compensation of $598K.  With these incentives in place, it’s not surprising that he acts so clearly like an owner – this is one of the most respected CEOs in the energy industry.

Big structural cost advantage, high ROIC, low debt, great management.  What could go wrong?  The big wild card is carbon emissions, and the Obama administration’s hostility to coal.  Attempts in the first term to pass cap-and-trade stalled, and seem unlikely to be revived, as they ultimately would increase energy costs to the middle class, which is not high on Obama’s list.  But he has directed the EPA to require emissions standards for new power plants that make it impossible to build a new coal plant, as well as restrictions that may force the retirement of older, less-efficient coal plants.  On their own, these executive actions might close plants equal to 10-15% of demand, or reduce U.S. coal production from about a billion tons to something more like 850-900 million tons.  

This seems like a minimal impact, though.  Environmental think tank World Resources Institute reported last month that China and India are planning to build 1,200 new coal plants, which would consume another 4 billion tons of coal per year.  Given that both countries are already coal importers (China produced 3.6 billion tons in 2011, and imported 177 million, and India produced 535 million and imported 90 million) , and the ease of export from ILB through the U.S. river system, there should be ample export demand.  The EIA estimates that U.S. exports of coal will be 125 million tons in 2012, up from an average of 56 million tons/year over the past decade.  Note that the cost of shipment from ILB to China is about $50/ton, and recent thermal coal pricing in China has been about $100/ton.

Could a CO2 tax or cap-and-trade system make a big dent in Alliance’s business?  It’s true that burning coal releases twice the CO2 as burning natural gas.  And it’s true that any serious attempt to address carbon emissions will put a price on CO2.  If CO2 were taxed (or priced through a permit trading scheme) at $20/ton (the minimum recommended starting point according to many in DC, and compared to the price in Europe of about €6/ton last week), paying $3.25 for gas at the well-head would now be equivalent to paying a coal miner $30/ton, not $52/ton.

This risk hangs over the sector, as it will be difficult to make money selling at $30 with fully-burdened cash costs of $39 in the ILB, or $79 in CAPP.  But it’s hard to see how $30/ton is going to happen.  That price would make virtually all of CAPP and ILB production uneconomic.   This poses several problems for a politician.  First, to replace 452 million tons of annual coal production would require 11 Tcf of natural gas (one ton of coal is roughly 24 Mcf of gas).  That would be a 45% increase in U.S. production of natural gas (about 24 Tcf/year), or a more-than-doubling of U.S. shale gas production (about 10Tcf/year).  It’s not clear that could happen easily, or that the well-head gas price would stay at $3.25.

Second, utility managers like having a blend of power sources.  Horizontal fracking is a young technology, and the price of natural gas has historically been highly volatile.  Try finding someone willing to sell gas forward at today’s prices in large quantities.  It would be equally hard to persuade a utility manager used to thinking in decades (which is basically all of them) to give up entirely on a reliable power source.

Third, coal-production is rather well-correlated with electoral importance.  Destroying the coal industry is not a great way to win the White House, and definitely complicates getting elected to Congress in many places.

Finally, if the objective is a reduction in coal burn, then the logical coal to stop burning is the high cost coal.  That would be CAPP coal, and there is a lot of it to stop producing before ILB coal is at the top of the cost curve.

What about valuation?  Alliance is finishing construction of three new mines – Penn Ridge, Gibson South, and White Oak.  Once complete, these are forecasted to allow total production of about 48 million tons of coal, of which about 75% will be ILB, and 25% will be Northern Appalachian mines (though lower cost due to use of long wall equipment).  Historically, Alliance retained 30-50% of distributable cash flow in order to reinvest for growth, which has been great for shareholders.  In 2011, Alliance made just under $18/ton of EBITDA, and is forecasting $16-17/ton for 2012, as costs are a bit higher due to the start-up of the new mines.  The new mines will be operating at full capacity in 2015.  If you assume $18/ton of EBITDA, and that all distributable cash flow is paid out as a dividend, then ARLP would pay an $8.80/unit dividend, and AHGP would pay a $6.75/unit dividend.  These would be yields of 16% and 15%, respectively, which seems cheap for a high ROIC business, especially with good reinvestment opportunities as CAPP continues to fade.

Given the history of reinvestment, and assuming no shift that clobbers ROICs, I suspect that about 2/3 will get paid out, and 1/3 reinvested.  Whatever decision is made, we can be confident that Joe Craft will be thinking like an owner.

Note that I have calculated FCF multiples in the write-up summary as share price to distributable free cash flow.  Also, EBIT is EBITDA less maintenance cap ex, and enterprise value for AHGP includes the value of the 21.2 million shares of ARLP that AHGP does not own.  That's a pretty theoretical number in my mind, compared with the cash flows available for distribution to GP and LP.

Should you own the LP or the GP?  The LP is safer, as it is not leveraged to the IDRs like the GP.  But Joe’s ownership is almost all at the GP level.  Historically, it has been a smart choice to have the benefit of the GP’s IDRs, and to be where the CEO’s own bet is.