Mosaic Co/The MOS

Thesis

Mosaic Co (MOS) mines phosphate rock in Florida and potash in Saskatchewan and sells branded fertilizer products (MicroEssentials, K-Mag) plus generic DAP/MAP/MOP into a global market. The product is fungible, the price is set in Vancouver, Sao Paulo, and the Persian Gulf, and Mosaic's job is to be a low-cost producer with long-lived reserves (Esterhazy, Belle Plaine, Florida) and a Brazil distribution arm.

The scorecard tells the story bluntly. Composite 60. ROIC 10y average 4.32%, ROIIC 5y 2.18% — both below any reasonable cost of capital, meaning the company has destroyed economic value over a full cycle even before counting reinvestment. FCF conversion is a weak 46.6%. Net debt/EBITDA looks fine at 0.29x but interest coverage is only 2.7x, meaning EBITDA is volatile enough that even modest leverage bites at the trough. TTM P/E is 42.09 vs. a 10y average of 28.65, and EV/FCF is a staggering 166.83 — the cyclical denominator is collapsing while the price is not.

Reverse-DCF implied growth of 3.95% is what the market is paying for. Owner earnings TTM are $0.32B against an enterprise value implied by the multiples; the deterministic IV range is $12.94 (base) to $16.41 (high) versus a $23.15 price — a P/IV of 1.79x. Even the high case requires a 30% drawdown to enter with a margin of safety. There is no price math here that works for a Buffett-Munger compounder. The qualitative work below confirms the quantitative verdict: this is a commodity-cycle stock, not a compounder, and it fails the circle-of-competence test on rule four (predicting commodity prices). Recommendation: Avoid at this price; revisit only below $11.

Moat

Mosaic operates in fertilizer — a commodity where the molecules from Saskatchewan, Belarus, Russia, Morocco, Saudi Arabia, and Florida are interchangeable at the farm gate. Buffett's framework demands an enduring moat that protects 'excellent returns on invested capital' [6]. Mosaic's 10-year average ROIC of 4.32% is the empirical answer: whatever moat exists is not wide enough to generate returns above cost of capital across the cycle. Let me work through the five moat types.

1. Pricing power — NONE. Mosaic is a price-taker. Phosphate (DAP/MAP) and potash (MOP) prices are quoted weekly in benchmark publications and move on Chinese export quotas, Belarusian sanctions, Black Sea shipping, and Brazilian soybean acreage — none of which Mosaic controls. The MicroEssentials premium product line is a real but small differentiator; it represents enhanced-efficiency fertilizer with sulfur and zinc co-granulated, but agronomically substitutable. Buffett's See's Candy test [3] — can you raise prices every year regardless of input cost? — fails immediately.

2. Switching costs — NONE. A farmer in Mato Grosso choosing between Mosaic, Yara, Nutrien, and ICL pays the landed price. There is no integration, no recurring subscription, no proprietary application equipment. Damodaran's switching-cost framing [2] applies to software lock-in, not bulk commodity bags.

3. Network effects — NONE. None of the standard mechanics (more users = more value) apply to bulk fertilizer.

4. Intangibles (brand, IP, regulatory) — NARROW at best. MicroEssentials is a real brand with patent protection on the granulation process, and the Mosaic Fertilizantes Brazil distribution network has scale advantages in last-mile logistics. But these are local, not global, and the patent shield is finite. Damodaran's note [1][2] that legal moats often come with regulated price caps doesn't fully apply here, but the durability test does: a moat that must be continuously rebuilt is no moat at all [6]. Florida phosphate operations carry a perpetual EPA/state Consent Decree (2015) governing gypstack management — a regulatory cost, not a competitive shield, since competitors face equivalent regimes.

5. Cost advantages — NARROW, the only real moat. This is where Mosaic has something. Esterhazy and Belle Plaine are world-class potash deposits with multi-decade reserves and solution-mining/conventional cost positions in the second quartile of the global cost curve. Florida phosphate rock benefits from co-located mining, beneficiation, and chemical conversion. Mosaic also owns Ma'aden Wa'ad Al Shamal joint-venture exposure in Saudi Arabia — a first-quartile cost position. But the global cost curve is dominated by Russian Uralkali, Belarusian Belaruskali (when not sanctioned), Canadian Nutrien (lower cost, more potash than Mosaic), and Moroccan OCP (lowest-cost phosphate rock by a wide margin). Mosaic is mid-cost in potash and mid-to-high cost in phosphate. A $10B competitor stress test [the standard prompt] is moot — the competitors already exist (Nutrien at $30B, OCP state-owned, Yara at $10B) and have spent the capital. The moat is not a moat; it's a cost curve position.

Erosion risks. (a) Belarusian potash returning to global markets at full volume; (b) Chinese phosphate self-sufficiency reducing import demand; (c) green ammonia/precision agriculture reducing nutrient intensity per bushel over 20 years; (d) Saudi Ma'aden expansion adding low-cost phosphate supply that Mosaic itself partially funded.

Buffett's test [3]: 'a moat that must be continuously rebuilt will eventually be no moat at all.' Mosaic must continuously sustain $1.2-1.4B/yr of capex on mine development, gypstack closure, and Brazil logistics just to hold position. This is not See's Candy; it is closer to a regulated utility without the regulated returns.

Moat verdict: NARROW.

Management & Capital Allocation

The CEO is Bruce Bodine (since early 2024), a Mosaic insider who came up through operations. The CFO and management team are competent operators with deep mining backgrounds. The 10-K/A amendment in March 2026 was housekeeping (shareholder count disclosure and an Esterhazy QP consent), which is mildly positive — they amended quickly rather than burying an issue.

Let me grade the five capital allocation choices.

1. Reinvestment in the operating business. Mosaic has invested heavily in MicroEssentials capacity, the Esterhazy K3 ramp, the Ma'aden phosphate JV in Saudi Arabia, and Brazil distribution buildout. ROIIC over the trailing 5 years is 2.18% — well below cost of capital. The Saudi Ma'aden investment is illustrative: Mosaic helped finance and provided technology to a low-cost competitor that now sells into the same Asian markets. The honest assessment is that incremental capital has earned sub-cost-of-capital returns through a partial cycle, partly because the cycle was unfavorable and partly because the projects competed with their own future end-markets. Grade for reinvestment: D.

2. Acquisitions. The defining acquisition was Vale Fertilizantes in 2018, which created Mosaic Fertilizantes (the Brazil business) for ~$2.5B. It bought scale in the world's most attractive fertilizer demand market. Integration was rocky (currency, logistics, Brazilian inflation), but the strategic logic was sound. No major acquisitions since. Grade: C+.

3. Debt management. Net debt/EBITDA at 0.29x is conservative. The balance sheet score of 19/25 reflects this. Interest coverage at 2.7x is uncomfortable given the cyclicality — at the 2016/2019 trough, EBITDA can fall 60%+, which would push coverage to ~1x. Management has been disciplined about not over-levering at peaks. Grade: B.

4. Buybacks. Share count is down 1.02% over 10 years — essentially flat. Mosaic has run multiple repurchase programs (2021, 2022 authorizations). The critical Buffett test is the average P/IV at which buybacks were done. Without disclosed-by-tranche pricing, we can infer from history that 2022 buybacks were executed at ~$50-70 per share against an IV that was likely in the $20-30 range — i.e., P/IV well above 2x. This is precisely the value-destructive buyback pattern: buying back stock at the peak of the cycle when cash is flush, then issuing or pausing at the trough. The current $23 price is closer to IV but still 1.79x base IV; further buybacks here would still be marginal. Grade: D.

5. Dividends. Mosaic pays a modest dividend (recently around $0.84-0.88/sh annual), with cuts during the 2016-2017 trough and increases during 2021-2022 peak. The pattern is reactive to the cycle rather than a smoothed return-of-capital policy, which is honest but not exemplary. Grade: B-.

Communication quality. 10-K disclosures are thorough, segment reporting is clean (Phosphates / Potash / Mosaic Fertilizantes / Corporate), reserves disclosure follows SK-1300 with QP consents (the recent Esterhazy amendment). Earnings calls are factual, with appropriate emphasis on prices and volumes; management does not over-promise on cycles. Grade: B+.

Synthesis. The capital allocation pattern is the classic commodity-cycle trap: reinvest and buy back at the peak (2011-2014, 2021-2022), conserve at the trough (2016-2019, 2024-2025). The aggregate result is the ROIC and ROIIC numbers — sub-cost-of-capital. This is not bad people; it is good people in a structurally bad business. Buffett's distinction [6]: 'if a business requires a superstar to produce great results, the business itself cannot be deemed great.' Conversely, even great managers cannot make a price-taking commodity producer earn sustained excess returns.

Capital allocator: C.

Industry Structure

Porter's Five Forces on global fertilizer (phosphate + potash):

1. Threat of new entrants — LOW. Greenfield potash mines (Jansen, BHP) cost $5-10B and take a decade. Phosphate rock is geographically concentrated (Morocco/Western Sahara holds ~70% of global reserves). Capital intensity, permitting, and reserve quality are real barriers. This is the most attractive of the five forces for Mosaic — and the only one that is genuinely defensive.

2. Bargaining power of buyers — MODERATE-HIGH. Buyers are large agricultural cooperatives (CHS, Helena, Nutrien Ag Solutions in NA), Brazilian distributors (Heringer, Yara Brasil), and Indian/Chinese state buyers (IPL, Sinofert) who negotiate annual contracts. State buyers in particular extract concessions in benchmark price negotiations. Farmers themselves can defer application by 1-2 seasons without yield collapse, so buyer demand is elastic at the price-setting margin.

3. Bargaining power of suppliers — LOW-MODERATE. Sulfur, ammonia, and natural gas are the key inputs. Sulfur is a refinery byproduct with periodic gluts. Ammonia is increasingly tight as global natgas prices fluctuate. Equipment vendors (Caterpillar, Komatsu, FLSmidth) have some leverage but not pricing power over Mosaic.

4. Threat of substitutes — LOW NEAR-TERM, MODERATE LONG-TERM. There is no chemical substitute for phosphorus or potassium in plant biology — these are essential nutrients. Manure recycling, precision agriculture, and biological nitrogen-fixation reduce intensity per bushel but do not eliminate demand. Over 20+ years, agtech could compress nutrient demand growth from 2-3% to 0-1%.

5. Rivalry among existing competitors — VERY HIGH. This is the killer force. Nutrien (potash leader, $30B mkt cap, Saskatchewan), Uralkali and Eurochem (Russia), Belaruskali (Belarus, sanctioned but volume-leaking), OCP (Morocco, state-owned, lowest-cost phosphate), ICL (Israel, Dead Sea), K+S (Germany), Mosaic (US/Canada/Brazil), and SQM (Chile). Eight global players in a market with no product differentiation and government-influenced supply behavior. China has alternated between major exporter and self-sufficiency hoarder for phosphate. Belarus and Russia are wildcard exporters whose volume return after sanctions easing would crush prices. The 2022 spike (potash to $1,200/t) and 2024-2025 reversion ($300-400/t) is the rivalry signature.

Value pool location and trajectory. The value pool sits with the lowest-cost producers (OCP for phosphate, Belaruskali/Uralkali for potash) and with state-influenced exporters who can flex volume for geopolitical objectives. Mosaic captures middle-of-curve economics. The value pool has been stagnant or declining in real terms over 20 years; the 2008 and 2022 spikes were real but short-lived. Long-term grain demand grows 1.5-2%/yr but supply elasticity of fertilizer is high.

Industry Verdict: Poor.

Inversion (Bear Case)

I am now short Mosaic. Here is my pitch.

The single event that kills this. China resumes large-scale phosphate exports while Belarus simultaneously gets sanctions relief and flushes potash back into Brazil and India at distressed prices. This is not hypothetical — China was a swing exporter as recently as 2020, and Belaruskali was the world's #2 potash exporter pre-2022. Either alone is a 25% price-down event for Mosaic's primary nutrients; both together would mark a fresh trough. Earnings would collapse from a current TTM that already produces a 42x P/E down to negative free cash flow within four quarters. Stock target: $9-11.

Why the moat is narrower than bulls think. Bulls cite 'irreplaceable' Florida phosphate and Saskatchewan potash reserves. The data refutes this. ROIC has averaged 4.32% over a full decade encompassing two price spikes. If the moat were real, ROIC would be 12-15% across the cycle, like a See's Candy or a Coca-Cola. It isn't. The 'reserves' are a balance sheet asset that doesn't earn excess returns because eight other producers have equally good reserves and three of them (OCP, Nutrien, the Russians) have lower cost positions. The MicroEssentials premium is 5-8% of EBITDA — a rounding error against the cyclical denominator. Mosaic's own joint-venture investment in Saudi Ma'aden helped fund a lower-cost competitor — the moat is so weak that the company arms its rivals.

Why management is worse than it appears. They look conservative on leverage today (0.29x net debt/EBITDA), but interest coverage is only 2.7x at TTM EBITDA — meaning at a -50% trough EBITDA scenario coverage drops to ~1.4x and refinancing risk emerges. The buyback record is the bigger tell: the bulk of buybacks were executed in 2021-2022 at $50-70 per share, when intrinsic value (using normalized through-cycle FCF) was likely $20-25. That is shareholder destruction at scale, dressed up as capital return. Compare to Nutrien, which has been more disciplined about averaging in. The Mosaic CEO transition (Bodine, 2024) brought in an operations lifer at the start of a downcycle — exactly the wrong moment for an inwardly-focused operator vs. an externally-focused capital allocator.

What bulls are extrapolating that won't hold. The bull narrative leans on (a) Indian/Brazilian demand growth, (b) energy-transition tailwinds for Saskatchewan/American mining, (c) China structurally pulling back from exports, and (d) green ammonia constraining nitrogen substitution. Each has a counter. (a) India and Brazil are exactly the markets Belarusian and Russian potash will flood when sanctions ease. (b) The 'friend-shoring' premium has already been priced — Mosaic's stock is up from sub-$10 in 2020 to $23 today on this narrative. (c) Chinese export quotas are policy-driven and have flipped within months historically. (d) Nitrogen and potash/phosphate are not substitutes; this is a category error.

The most dangerous extrapolation is the EV/FCF of 166.83. To justify $23 you have to believe FCF is about to multiply 4-5x and stay there. The cycle history says it will multiply 2x, hold for 18 months, and revert. The market is paying for permanent peak earnings on a cyclical denominator.

Valuation trap (multiple compression / regime change). TTM P/E is 42 vs. a 10-year average of 28.65. The 10-year average is itself elevated because it spans two spike periods. Cycle-low P/E for fertilizer producers historically compresses to 6-9x trough earnings, which would imply $6-9 per share at the next trough. The asymmetry is brutal: at 1.79x base IV today, the upside requires a sustained price spike (low odds, hard to time) while the downside is a return to historical norms (high odds, just need to wait). The reverse-DCF implied growth of 3.95% sounds modest but is 2x the actual 5y ROIIC of 2.18% — the market is paying for growth this company has not delivered with its incremental capital.

Additional regime risks: (1) farm-bill changes reducing US fertilizer subsidies; (2) precision-ag adoption compressing nutrient intensity 1-2%/yr over a decade; (3) gypstack environmental liabilities (the 2015 EPA Consent Decree) escalating; (4) Brazilian real depreciation eating Mosaic Fertilizantes margins.

If I am right, the stock could be worth $9-11 within 2-3 years.

Lollapalooza Bias Check

Biases active in me right now as the analyst:

Anchoring. I am anchored to the $23.15 current price as a reference point. The IV range is $12.94-16.41. If I had been handed only the IV range and the business description without the price, I would not have spent a sentence considering a buy thesis. The price drags my brain into 'maybe the market knows something' territory. Discipline: ignore the price tag, value the business, then look at the price.

Authority. Buffett bought Mosaic-adjacent and ag-adjacent names (DE, Posco) in the past, and several smart fundamental funds (Greenlight historically, others) have rotated through fertilizer cyclicals. The pull is to assume someone smart sees something. The deflection: Buffett also explicitly avoids commodity producers that lack pricing power, and most fundamental investors who own MOS treat it as a trade, not a compounder.

Recency. The 2022 fertilizer spike is fresh in mind. Russian invasion, Belarusian sanctions, $1,200/t potash, $1,000/t DAP. The recency bias suggests the next spike is around the corner. The data says spikes occur roughly every 7-10 years and revert within 24 months. Recency overweights the spike and underweights the long flat.

Confirmation bias. I started this analysis with a strong prior that commodity producers fail Munger's circle-of-competence test on rule four. I have to actively look for the bull case. Honestly considered: MicroEssentials brand premium (real but small), low-cost JV in Saudi (meaningful but partly arms competitors), Brazil distribution moat (real, locally). The bull case exists but does not change the math at $23.

Deprival super-reaction. Not strongly active here — I do not own MOS and have nothing to lose by saying pass.

Incentive bias. Compounder framework rewards 'pass' more than 'buy'; the cost of a wrong-pass is opportunity cost (invisible), the cost of a wrong-buy is realized loss (visible). This biases me toward Avoid/Too Hard. The discipline: am I dismissing a real opportunity to feel safe? In this case, the math is too clear — 1.79x base IV, sub-cost-of-capital ROIC, fail on circle-of-competence — for the incentive bias to be load-bearing.

Social proof and commitment: not active. I have not previously published on MOS and have no exposure.

On balance, the active biases (anchoring + recency) push toward being more constructive than the data warrants. The conclusion (Avoid) is robust to bias correction.

10-Year Outlook

In ten years (2036), will Mosaic still mine phosphate in Florida and potash in Saskatchewan and sell branded and generic fertilizer into a global commodity market? Yes, with very high confidence. The reserves are multi-decade, the assets are not strandable, and the demand for P and K is biologically anchored.

Will the customer base be larger? Marginally. Global cropland is roughly flat; nutrient demand grows at population + per-capita-calorie growth, call it 1.5-2%/yr in volume terms. Brazil and India will be larger; North America and Europe will be flat to declining. So unit volumes ~20-25% higher.

Will profit per customer (ton) be higher? This is the wrong question for a price-taker. The honest answer is: profit per ton will average across cycles roughly where it has averaged for the past 30 years — modest, with periodic spikes. There is no structural reason to expect a permanent step-up.

Will the moat be wider? No. The competitive set is well-defined (eight global producers), the cost curves are slowly evolving but Mosaic is not gaining position, and new entrants (BHP Jansen) are slowly adding capacity at the low-cost end. If anything, the moat erodes as Saudi Ma'aden ramps and Belarusian volume normalizes.

Single biggest threat over 10 years: a combination of Belarusian/Russian potash market normalization and Chinese/Moroccan phosphate capacity additions creating a sustained 5-7 year glut. Probability: meaningful — 30-40%.

The business will exist. The compounder thesis will not. ROIC will continue to average 3-6% across the cycle, and the stock will trade in a band ($8-30) that mean-reverts around an IV that is structurally below the cyclical price.

CONFIDENCE: medium