Ag

The question

Are fertilizer producers, grain merchandisers, and row-crop equipment makers priced for a prolonged destock and a multi-year equipment spending air pocket, or for a normalization of nutrient demand, a mean-reverting replacement cycle, and a structural food-security bid that has been absent from the tape for eighteen months?

Base rates

The reference class is prior ag down-cycles that followed a price spike: 1996-2001 after the 1995-96 grain run, 2013-2016 after the 2010-2012 spike, and 2023-present after the 2021-2022 spike. In each prior episode, fertilizer equity drawdowns from peak averaged 45-55% and bottomed 18-30 months into the destock; forward three-year returns from that trough, measured as a diversified nutrient basket, averaged +16-22% annualized with roughly 70% hit rate. Row-crop equipment OEMs followed a similar pattern with a two-quarter lag, because the order book drains before the used-equipment market clears.

The cohort has trailed SPY trailing-1Y by a wide margin. The index compounded roughly 28.6% while the ag complex was flat to down. That underperformance is the thesis. The imputed-expectations gap between cycle-trough multiples on DE, NTR, MOS, and CF and the normalized mid-cycle earnings those businesses produced from 2017-2019 is the setup.

The nutrient cycle has its own base rate. Nitrogen, phosphate, and potash each have distinct supply curves, but the meta-pattern is that prices below marginal cash cost do not persist more than four to six quarters because the marginal tonne exits. Potash is currently trading near the 75th percentile of global cash cost; phosphate is near the 85th. Nitrogen is the only leg that still has a demand question attached.

For equipment, the base rate is fleet age. The North American large-tractor and combine fleet is at a 22-year average age, roughly two standard deviations above the 1995-2025 mean. Fleet age this stretched has never mean-reverted without a multi-year replacement up-leg.

Why consensus is wrong

Consensus reads the last eighteen months as the start of a secular de-rating rather than the bottom of a destock. Three pieces of the story are mis-specified.

First, global grain carryout. Stock-to-use for corn and wheat is near the 20th percentile of the last twenty years. A single below-trend harvest in either hemisphere moves the nutrient pull from cautious to urgent because growers in a tight balance sheet buy every pound of nitrogen the agronomy book calls for. The market is pricing the carryout as if it were the 70th percentile.

Second, Chinese and Russian export posture. China has operated nitrogen and phosphate export windows rather than steady flow for three consecutive years; Russian and Belarusian potash has been routed around sanctions at material friction cost. The consensus supply model treats these tonnes as fungible. They are not. A tonne that has to clear two extra intermediaries carries a persistent cost basis that the spot print does not capture.

Third, the equipment destock. Dealer inventories on used combines and high-horsepower tractors are still elevated, and consensus assumes another full season of discounting before the channel clears. The offsetting variable is that new-order cadence has already bottomed at roughly 55% of the 2013-2019 run-rate. When the channel clears -- and fleet age guarantees it will -- the order book snaps rather than ramps because there is no intervening used-equipment buffer to absorb the shift.

Position construction

The book has one 20% anchor, four cluster sub-books, and a long niche tail.

Anchor. DE at 20% is the cleanest read on row-crop equipment replacement, with precision-ag recurring revenue providing a floor under the cyclical trough. The incremental unit economics on large-frame tractors at normalized volumes are the strongest single driver in the book.

Potash and phosphate cluster (~27.7%). NTR at ~15.2% is the integrated potash-plus-retail franchise with through-cycle free cash flow. MOS at ~5.1% adds concentrated phosphate exposure near the 85th percentile cash cost position. SQM at ~6.8% captures the specialty potash and iodine franchise; the lithium optionality is a secondary feature rather than the thesis. IPI is not in the book -- the small-cap potash pure-play did not clear liquidity screens.

Nitrogen (~7.4%). CF at ~7.4% is the North American nitrogen cost-curve leader with structural natural-gas feedstock advantage versus European producers. Sizing reflects the fact that nitrogen is the only nutrient where the demand question has not fully resolved.

Grain merchandisers (~27.3%). ADM at ~16.6% and BG at ~10.7% are the vertically integrated origination, storage, and processing businesses whose earnings power is less about crop price direction than about volatility and basis. Both trade at mid-cycle free cash flow yields.

Equipment, ex-anchor (~15.5%). CNH at ~7.1% and AGCO at ~4.4% round out the row-crop OEM exposure across Case, New Holland, Massey Ferguson, and Fendt. TTC at ~4.0% adds lawn-care and turf equipment with a distinct housing-linked demand curve.

Niche and protein (~2.7%). CALM at ~2.5% is the egg producer with a fleet-of-farms model whose pricing is driven by the biosecurity cycle. NC at ~0.2% is the de minimis coal-and-minerals royalty position; small enough that its idiosyncratic risk does not move the book. SMG is not in the book -- the consumer lawn-and-garden channel did not clear screens.

Asymmetric payoff

If the nutrient destock clears over the next four quarters, the equipment order book normalizes on fleet age, and grain carryout stays at or below the 30th percentile, the weighted book returns roughly 18-26% annualized over three years. If carryout builds for two consecutive seasons and the destock extends, the book returns roughly -5% to +3% with dividend support from DE, NTR, ADM, and BG providing a floor. If a meaningful global harvest shortfall develops in 2026 or 2027, the right tail is 35-45% with multiple expansion concentrated in the nutrient names.

At 55% base, 25% bear, and 20% bull, expected value is roughly +14 to +20% annualized against an SPY base rate near +8%. The payoff is asymmetric because cycle-trough multiples, cost-curve positions at the high end of the global supply stack, and a fleet-age mechanic that cannot be deferred indefinitely all compress the bear case while the food-security right tail remains open.

Three things that would change our mind

1. Global corn and wheat stock-to-use rebuilding above the 50th percentile on two consecutive harvests with above-trend yields in both hemispheres, which would remove the tightness that makes a nutrient bid self-enforcing.
2. Chinese nitrogen and phosphate export windows widening to steady-flow rather than episodic release, with policy language indicating a multi-year change rather than a tactical adjustment -- this would collapse the friction premium the book depends on.
3. Large-tractor and combine dealer new-order cadence failing to inflect above 70% of the 2013-2019 run-rate by the second half of 2026, signaling the fleet-age mean reversion has a demand rather than a supply problem.

What we're explicitly NOT betting on

We are not betting on a specific corn, wheat, or soybean price print. We are not betting on a particular lithium price recovery through the SQM position. We are not betting on any single regulatory outcome -- tariff, export-control, or biofuels mandate. We are not betting on the timing of Chinese urea export quotas. We are not betting on a weather event. The thesis requires only that nutrient demand normalizes off a destock, that an aged equipment fleet behaves like every prior aged equipment fleet, and that grain merchandiser volatility stays in its historical distribution. All three are strictly weaker claims than picking a crop-price direction, and the book is sized for them.