Coal

The question

After a year in which the coal cohort materially underperformed the S&P 500 and most listed names failed a standard absolute-return hurdle, is there a defensible thesis for holding the seven that cleared the relaxed screen -- and does the concentration itself carry information about where the cohort's cash-flow durability actually lives?

Base rates

The reference class is commodity cohorts whose secular-decline narrative diverges from their near-term cash-flow reality: US tobacco manufacturers through the 1998-2005 window, US onshore conventional oil through 2009-2012, and integrated steel through 2015-2018. In each case, the cohort's total-return profile was driven by two features the market under-weighted: very high free-cash-flow yields at depressed multiples, and a bifurcation in end-market exposure that the headline framing missed. The base rate for holding the survivors through the tail of a secular-decline narrative is roughly the 55th-70th percentile of resource-sector sleeves over a three-year horizon, with the overwhelming majority of the return coming from cash distribution rather than multiple expansion.

The coal cohort trailed SPY's roughly 28-29% one-year return by a wide margin through the period ending March 2026. The screen had to drop the absolute-return hurdle to populate a book at all. That is the thesis, not an embarrassment -- the imputed-expectations gap is widest precisely when a cohort has just underperformed sharply and the survivors are the names whose underlying cash flows did not break.

A second base rate worth naming: metallurgical coal prices have historically tracked integrated steel spreads with a six-to-nine-month lag, while thermal coal prices track natural-gas basis and grid-reliability spreads. The two commodities are frequently conflated in the headline cohort and are now on visibly different cycles.

Why consensus is wrong

The sell-side frames coal as a single secular-decline cohort. That framing was accurate through roughly 2019, when thermal demand was falling and metallurgical demand was cyclical around a flat trend. Two things have changed since. First, thermal coal's marginal-demand story in the US has inverted: grid operators facing AI data-center load growth, coincident retirements of aging gas peakers, and transmission-queue backlogs are extending coal-plant operating lives rather than accelerating retirements. PJM, MISO, and ERCOT have each announced extensions on specific units through 2032 or beyond. Second, metallurgical coal has decoupled from the thermal decline narrative on steel-cycle fundamentals that are independent of the grid mix.

The second piece the consensus misses is capital return. The cohort's survivors run 40-60% free-cash-flow yields at current prices with balance sheets largely rebuilt from the 2015-2020 cycle. Payout ratios on variable-distribution frameworks have run at 60%+ of free cash flow. At these yields, the cash return alone clears a reasonable equity hurdle over three years regardless of multiple path.

Third, royalty and coke-adjacent exposures inside the cohort carry different cash-flow profiles than the operating mines themselves, and the market prices them together.

Position construction

This is a deliberately concentrated seven-name book. The concentration mirrors the water-infrastructure precedent where only three names cleared the hurdle and we held only those three. Here the cohort yielded seven, and we hold seven. HCC was excluded because it sits inside a separate steel-adjacent idea. The book is thin because the cohort is thin, and diluting with screen-fails would import the cash-flow deterioration the screen was designed to filter out.

The seven names cluster into three groups.

Diversified thermal anchors (~40%). BTU at 20% is the global seaborne and US domestic thermal franchise with exposure across PRB, Illinois Basin, and Australian assets. ARLP at 20% is the Illinois Basin thermal producer with the most consistent distribution history in the cohort -- the master limited partnership structure and coverage profile make it the closest analogue to a yield anchor.

Metallurgical coal (~36%). METC (20%) is the Central Appalachian metallurgical producer with the cleanest inventory and lowest all-in sustaining cost in the book. AMR (~16.4%) is the largest metallurgical name by reserve base, with export-terminal flexibility that lets the cash flow ride the global steel cycle rather than US domestic demand. Together these two positions carry the steel-cycle exposure that is structurally decoupled from the thermal narrative.

Royalty and adjacency (~23.6%). NRP (~11.1%) is the mineral-rights royalty trust -- non-operating cash flow from coal and soda ash with essentially no reinvestment requirement. HNRG (~6.9%) is the Illinois Basin thermal producer with an integrated power-generation asset that captures the grid-firming thesis directly. SXC (~5.6%) is the domestic coke producer with contracted supply to US integrated steel mills; the revenue model is fee-based take-or-pay, not commodity-priced.

Asymmetric payoff

If thermal coal retirement curves stay deferred on the announced schedule, metallurgical spreads hold near current levels, and the cohort continues returning 40-60% of free cash flow annually, the book returns roughly 14-22% annualized over three years, with the majority coming from cash distributions. If steel demand contracts and grid firming pivots faster toward batteries and gas peakers, the book returns roughly -15 to -25%. If a Chinese or Indian steel re-acceleration tightens seaborne metallurgical markets, or if AI data-center siting forces additional US thermal extensions, the right tail is 30-45%.

At a 50% base, 30% bear, and 20% bull weighting, expected value is roughly +8 to +14% annualized against an SPY base rate near +8%. The expected-value edge is narrower than in other energy books because the downside scenarios are more severe; the payoff is genuinely asymmetric, not symmetrically skewed.

Three things that would change our mind

1. Two or more announced coal-plant retirement extensions being reversed by state-level regulatory or political action, removing the grid-firming leg of the thermal thesis.
2. Seaborne metallurgical coal benchmarks falling below $150 per tonne for two consecutive quarters while global steel production declines -- signalling that the met-thermal decoupling is collapsing back into a single cycle.
3. A cohort-wide move by the surviving operators to rebuild growth capex and suspend variable distributions, which would invalidate the cash-return thesis that currently underwrites the book's expected return.

What we're explicitly NOT betting on

We are not betting on a thermal coal renaissance or a reversal of long-run secular decline. The thesis requires only that retirements slip by five-to-eight years against the market's expectation -- a much weaker claim. We are not holding HCC; it sits in a separate steel-adjacent idea. We are not holding the Chinese or Indonesian listed thermal operators; they failed the screen on different grounds than the US cohort. And we are not padding the book with names that failed the one-year absolute-return hurdle to look more diversified -- the seven-holding concentration, like the three-holding water-infrastructure precedent, is deliberate. Diluting it would reduce expected return without reducing the risk that actually matters, which is that any one of the three sub-cohorts -- thermal, metallurgical, or royalty -- is misread.