The Trade

Long ARCO (Arcos Dorados Holdings) at $7.68. McDonald's exclusive master franchise operator across 20 Latin American countries, 2,400+ restaurants, 700 million people, 62% digital penetration, 2-3x market share vs nearest competitor in every major market.

The stock trades at 4.3x normalized EV/EBITDA. Coca-Cola FEMSA, the closest comparable LatAm franchise operator, trades at 8x. The global QSR franchisors — MCD, QSR, YUM — trade at 15-20x. ARCO is at half its nearest comp and a quarter of its parent's multiple.

The market is pricing a cyclical consumer trough as structural impairment. We think the bear case is cyclical and the bull case is structural. Three independent cost tailwinds are converging under a $159 million accounting distortion that's confusing every model on the street.

12-month probability-weighted target: $10.96 (+43%). Expected value including dividend: +46%.

The 42-Point Divergence

EWZ (iShares MSCI Brazil ETF) is up 39.1% over the past year. ARCO is down 3%. That's a 42 percentage point divergence.

Run the regression: ARCO's alpha vs a SPY + EWZ model is -20.4% annualized. The market isn't punishing Brazil. The market is punishing ARCO specifically — while Brazil rallies, while the real appreciates 11%, while every LatAm-exposed company from Coca-Cola to Herbalife to Globant guides to currency tailwinds.

This divergence has three explanations, all cyclical:

1. Brazil QSR traffic is down mid-to-high single digits. Lower-income consumers are squeezed by 15% Selic rates. Management said "the situation is going to last for a while." Real comp growth is near zero.

2. The $159M accounting mirage. A one-time Brazil tax credit dumped $106 million into 2025 EBITDA. Reported EBITDA is ≈$630M. Normalized EBITDA is ≈$520M. The $110M gap means that 2026 will look like a decline year on a reported basis even if the underlying business improves.

3. JP Morgan and Bradesco both downgraded to Neutral in late 2025. When the banks that cover you step back, the marginal seller shows up. RSI is 16. This is capitulation territory.

All three are cyclical. None are structural. McDonald's confirmed this when they offered ARCO a new 20-year master franchise agreement starting January 2025 — you don't hand 20-year exclusive rights to a franchise operator you're worried about.

What the Market Isn't Pricing

Three cost tailwinds are converging simultaneously. Each is independently verifiable. Together they set up EBITDA margin expansion even on flat real volumes.

1. The Liability Management Trade (Idiosyncratic)

In December 2025, ARCO's CFO Mariano Tannenbaum engineered something no other LatAm company has matched. He replaced $135M of 6.8% USD sustainability-linked bonds with local BRL bank debt synthetically hedged to an effective USD cost of 2.53%. In a 15% Selic environment.

That's a 430 basis point rate reduction. ≈$4.5M/year pretax interest savings plus enhanced tax deductibility from local Brazilian interest expense. It also eliminates the ESG step-up covenant risk embedded in the old bonds — if ARCO missed its sustainability targets, the coupon would have stepped up. Nobody on the call asked why they exited that structure.

We corroborated this across six LatAm companies doing liability management. KOF issued MXN bonds at 9.12%. Embraer's average debt cost went UP to 9.1%. Localiza is focused on duration, not rate reduction. YPF is getting 6.5% in Argentine pesos. Nobody is achieving 2.53% synthetic USD cost.

The mechanism: McDonald's franchise creditworthiness gives ARCO privileged access to bilateral bank debt that distressed Brazilian corporates cannot get. The derivative overlay converts BRL exposure to synthetic USD at favorable basis swap spreads. This isn't luck — it's a structural advantage of being the McDonald's franchise system in a high-rate environment where weaker credits are going bankrupt.

Bloomberg on March 14: "Brazil's Corporate Debt Drama Is Entering a New Chapter." Record bankruptcies. Two major firms restructuring $13.5B combined. ARCO is playing a different game.

2. The Cattle Cycle Divergence (Partially Systematic)

The US and Brazil are at opposite points in the cattle cycle. This is one of the more interesting cross-border disconnects in food right now.

US: Herd at 75-year low. Tyson closing beef plants on "tighter cattle supply." Texas Roadhouse: ≈7% commodity inflation, beef is "nearly all" of it. Restaurant Brands/Burger King: 20%+ beef inflation, franchisee profitability falling from $205K to $185K per unit. Darden: "near-record beef cost sustained longer than expected." No relief until 2027+ when herd rebuilding starts.

Brazil: Peak slaughter. Darling Ingredients CEO on their Q4 call: "Now go south Brazil, beef large, large now. Extremely full all plants down." Zoetis confirms favorable supply. USDA projects the cycle beginning to reverse in H2 2025 with herd retention, but current supply is massive. Datagro projects slaughter down 9% in 2026 — meaning the relief window is open NOW but closing over the next 2-3 quarters.

ARCO's Q4 2025 was the first quarter with food & paper cost improvement after beef ran up 30% over 12 months. Two consecutive quarters of sequential improvement. Management: "we feel confident in our ability to continue recovering gross margin."

Alsea (ALSSF), the most direct LatAm QSR peer, confirmed the same trend on their Q4 call: "positive news on price of beef, price of chicken, positive trend compared to the previous year." FX appreciation on dollarized raw materials is helping everyone.

The ARCO-specific layer on top of the industry trend: royalty equalization (Brazil royalty rate increase in 2025, which compressed Brazil margins by 160 basis points) is now fully lapped. That headwind is gone as of January 2026.

3. The G&A Cut ($10M+ Annualized)

Already completed in early 2026. Headcount reductions across all three divisions and corporate. Restructuring charge of $8.7M taken in Q4 2025 and excluded from adjusted EBITDA. Management gave a specific number: "our ongoing cost base has been reduced by more than $10.0 million on an annualized basis."

This is the simplest tailwind. It's done. It's quantified. It flows through starting now.

The Bear Case (Taken Seriously)

Brazil consumer weakness is the single largest driver — roughly 25% of thesis variance — and we have zero edge in predicting when it turns. Selic is at 15%. The central bank is doing "cautious easing." Lower-income consumers, who are ARCO's core demographic in Brazil, are retrenching. Industry traffic is down mid-to-high single digits.

Management guided Brazil comps "in line with inflation" — that's zero real growth. They said "second quarter and on" for improvement, but also said "the situation is going to last for a while." We're assigning 35% probability to Brazil traffic turning positive by Q3 2026, which means we're betting AGAINST management's timeline. If we're wrong and traffic recovers, great — that's the bull case. But we're not relying on it.

Argentina comp deceleration is mathematical, not operational. SLAD comps were +49.5% in Q4 2025, driven by hyperinflation passthrough. As Argentina normalizes in 2026, headline comps will decelerate sharply regardless of business health. This will look bad to anyone screening on comp growth without understanding the denominator effect.

No 2026 financial guidance. Management gave CapEx (105-115 openings, $275-325M), dividend ($0.28, +17%), and directional commentary on costs — but no EBITDA or revenue targets. With the $159M tax benefit distorting 2025, this creates an information void. The absence of guidance after a year of inflated EBITDA is not confidence-inspiring.

The cattle cycle is turning. Datagro projects Brazil slaughter down 9% in 2026 as herd retention begins. Current beef supply abundance may be 2-3 quarters from tightening. If ARCO gets 2 good quarters of food cost relief and then beef re-accelerates, the margin expansion thesis has a shorter runway than the base case assumes.

BRL forward curves project weakening. 5.45+ by September 2026 vs the current ≈5.25. If EM risk-off (tariffs, dollar strength) pushes BRL past 6.0, the FX tailwind reverses and USD EBITDA compresses even on good local results.

Factor Decomposition: Where's the Edge?

75.4% idiosyncratic variance in the SPY + EWZ model. Passes the 75% threshold, but barely.

Edge concentration in the tight thesis (margin expansion + earnings mispricing): ≈52%.

The honest truth: the largest single driver (Brazil consumer, 25%) is the one where we have no informational advantage. We can't predict when 700 million Latin Americans start spending again. What we CAN see is that three independent cost tailwinds are converging under an accounting distortion that's creating artificial noise in the reported numbers — and the market is pricing ALL of that as structural impairment.

Scenarios

Bear (30%): $5.50 (-28%)

Brazil consumer deepens. Selic stays 14-15%. Cattle cycle turns faster than expected — beef re-accelerates H2. BRL weakens past 5.5 on EM risk-off. 2026 normalized EBITDA: $480M. Multiple compresses to 3.5-4x.

Base (45%): $11.25 (+47%)

Volume flat in real terms. Cost tailwinds materialize: +$15M beef, +$10M G&A, +$20M royalty lapping, +$15M new openings, -$15M Argentina decel. 2026 normalized EBITDA: $565M (+9%). Multiple expands to 5-5.5x as street normalizes.

Bull (25%): $17.00 (+121%)

Brazil consumer recovers Q2-Q3. Selic cuts accelerate. Traffic turns positive. All base case tailwinds PLUS volume recovery and operating leverage. 2026 normalized EBITDA: $640M (+23%). Multiple expands to 6.5-7x toward KOF range.

Probability-weighted target: $10.96 (+43%)

EV = 0.30 x (-28%) + 0.45 x (+47%) + 0.25 x (+121%) = +43.0%
Dividend: +3.65%
Total: +46.7%
Stress test at 45% bear: +28.1%

Upside/downside ratio: 6.1x. Even under adversarial assumptions (45% bear probability), the EV is strongly positive.

The Options Signal

The equity market is capitulating. The options market is not.

May 15 expiry (covers May 13 earnings): 1,215 call OI at the $8 strike. P/C ratio 0.03 — that's 36.8x calls vs puts. For a $1.6B market cap stock with 2,700 total options OI, 1,215 contracts at a single strike is a concentrated bet. Someone with capital expects ARCO above $8 by May expiry.

August expiry: 363 call OI at $8, 113 at $10, 115 at $12. The $12 strike implies +56% upside by August.

IV at 51.6% on May (110th percentile vs 52-week range). The market is pricing a big move. The positioning says it's to the upside.

Equity flow (forced selling, RSI 16) is diverging from options flow (call accumulation into earnings). This is a classic contrarian pattern — the last sellers are institutional rebalancers and momentum funds being forced out, while options market participants are positioning for the turn.

Entry and Sizing

Entry: Tranche into a 2-4% position.

Why not wait for earnings? Ex-div March 30 captures 3.65% yield. If Q1 confirms the cost thesis, the stock gaps up and the 1,215 call OI at $8 tells you someone else has already figured this out. Retail edge is positioning before confirmation, not after.

Why not all-in now? No hard catalyst for 7 weeks. Brazil macro could deteriorate. BRL weakened 2.7% in March on Middle East tensions. Tranching respects the timing uncertainty.

Alpha: 3.9% (conservative) to 12.1% (probability-weighted), midpoint 8%.

What Kills the Thesis

1. Brazil traffic accelerates DOWN (not just flat — worsening). Would signal structural QSR demand destruction, not cyclical trough.
2. MFA renewal terms are punitive — higher royalties, restricted territory, or performance clawbacks.
3. BRL collapses past 6.0 — not just weakening to 5.45, but real EM crisis. Destroys the USD P&L.
4. Beef costs spike 20%+ again in H2 before margin gains are realized. Closes the cost relief window before it delivers.

None of these are our base case. But any of them would require reassessing the thesis.

What Would Make This Better

• Insider buying at these levels. Not present yet. If the CEO or CFO starts buying, ESCALATE.
• MFA renewal terms disclosed with favorable economics. Structural re-rating.
• Sell-side initiates a note on normalized EBITDA vs reported. Removes the fog.
• Brazil QSR traffic turns positive in Q1 or Q2. The 35%-probability event that would compress the path to re-rating.

Conviction

The bear case is real but cyclical. Brazil consumers are squeezed at 15% Selic. Traffic is down. Argentina comps will decelerate. The cattle cycle will eventually turn.

The bull case is structural. A 20-year McDonald's franchise across 20 countries. 62% digital penetration with no LatAm peer close. 2-3x market share in every major market. A CFO engineering 2.53% synthetic debt costs that no comparable company can match. Three independent cost tailwinds converging simultaneously.

At 4.3x normalized EV/EBITDA, the market is pricing this business like it might not survive. McDonald's just handed it a 20-year contract. The options market is loading calls into earnings. RSI is 16.

The current price requires Brazil consumer weakness AND beef cost re-acceleration AND FX reversal AND permanent street confusion — simultaneously. Joint probability is low. Expected value is +46%. Upside/downside is 6:1.

Build the position in tranches. Starter now. Confirm at earnings. Full conviction on volume recovery.

Evidence