Setup

Pamella Dana announced she wasn't seeking reelection on April 7, 2026. She bought $84,000 of AMTB on May 7 anyway — one of nine directors who acquired 3,761 shares each, same day, same price ($22.44). Total board-level capital deployed: $760K before Q2 data is public. Dana had no future option value from the board seat. She bought anyway.

This is a 57-day catalyst position into Q2 earnings July 22. At that horizon the Sharpe reaches 0.69. The 15-month hold doesn't clear 0.5 — the market isn't discounting enough to justify the drift.

AMTB is a South Florida community bank (≈$8B assets) mid-turnaround on a credit cycle. Its CEO vacancy closed May 18. Its largest funding risk — a $1.9B Venezuelan deposit base the 2025 10-K called a concentration hazard — is now being called a growth catalyst by the same management team that wrote the risk factor.

What the Filing Says

The May 26 8-K is a Reg FD investor presentation (Item 7.01, Exhibit 99.1). Lower evidentiary tier than a binding agreement, but management guidance lives here.

Carlos Iafigliola was appointed permanent President and CEO effective May 18, following a competitive search that included external candidates. He was CFO (2020-2023) and COO (2023-2025) before serving as Interim CEO for six months. Twenty-two years at the company. The transition is continuity, not disruption — he was already making the decisions. Compensation terms are pending; an 8-K/A is forthcoming. Watch CIC provisions for strategic signaling.

Management cited "preliminary reactivation of Venezuelan economy" driving deposit growth. They guided total deposits to $8B in Q2 (from $7.9B at Q1), with 8-10% cumulative growth for 2026. Average Venezuelan account balance: $46K. These are HNW diaspora deposits — not domestic wages, which remain under $1/month in Venezuela — and the mechanism runs through oil sector income flowing to private accounts, enabling diaspora relatives to accumulate and hold USD.

Q2 NIM guided 3.45-3.50%, down from 3.82% in FY2025. Management cited "developments in April and May" as the driver. The Q1 transcript provided the mechanism: broker CD costs ran ≈15bps wider than expected ($5M headwind in Q1) as AMTB used brokered funding to bridge deposit growth, while classified loan exits shifted the asset mix toward lower-yield but cleaner originations. On a $8B asset base, 35bps of NIM compression = ≈$25M annual NII headwind.

Q1 NPAs: $191.6M, up from $186.9M at year-end. Exits occurred ($38.4M in loan sales, $14.3M in paydowns) but three new commercial relationships added $10.5M to classified. Special mention loans fell $54M — a leading indicator — but net credit improvement is marginal. ACL/NPL coverage: approximately 46%, critically thin.

Buyback: $34.8M of $40M authorization deployed at $22.44/share average — the same price directors paid.

What the Market Thinks

Current price $22.46 = 1.00x TBV. Four analysts: 75% bullish, consensus target $25.50 (+13.5%). Options market is too illiquid to extract probabilities (45 OI on the July chain; ATM IV 81.3% vs 30-day HV 23.6% — 3.4× premium is noise at this volume).

Back-solving from scenario structure (bull $27.30 / base $22.83 / bear $19.25, P(base)=40%, P(bear)=20% fixed): the current price implies P(bull) = 34.7%. We assign 40%. The gap is 5.3pp — $1.44/share, +6.4% EV over 15 months. Annualized Sharpe 0.18. The 57-day window around July 22 reaches 0.69.

Analyst consensus at $25.50 implies near-zero probability of the bear case. We assign 20%. The edge is better calibration on the downside, not higher directional conviction.

Why the Gap Exists

The 10-K risk factor is four months stale. The 2025 10-K was filed February 27, 2026. Maduro was captured January 3. OFAC removed acting president Rodríguez from the SDN list April 1. The language describing Venezuelan deposits as a structural concentration hazard was written into a different geopolitical reality. The cross-sector Q1 confirmation is independent and broad: Copa Airlines returned to five Venezuelan cities, CVX and Shell confirmed JV mobilization, SLB called Venezuela "an exciting growth opportunity," Arcos Dorados cited Venezuelan local currency appreciation as a consumer activity indicator. These companies don't share a research department. The sell-side covering AMTB — four analysts, Florida community bank specialists — has not synthesized this.

The Venezuela factor doesn't appear in trailing return data. OLS regression over 90 trading days shows ILF (LATAM ETF) loading of β=0.036, t=0.46, accounting for 0.1% of AMTB's trailing variance. The deposit impact of Venezuelan economic reactivation won't appear in financial data until the Q2 10-Q (due August 2026). A factor the model can't see yet cannot be priced. The forward idio lives precisely where the trailing regression is blind. KRE (regional banking ETF) explains 73.4% of trailing variance — the standalone long is ≈70% a sector bet. The idiosyncratic thesis sits in the remaining 30%.

Nine directors bought, including the one who's leaving. Dana's purchase is the detail that doesn't have an easy institutional explanation. Directors buying simultaneously means coordination — they saw the same Q2 deposit pipeline data, the same credit trajectory, the CEO decision. They bought at $22.44, the same average as the corporate buyback. This is not in any analyst report.

The NIM Problem and the Honest Offset

"The source is the cure" overstates it. Venezuelan deposit growth helps NIM at the margin — estimate 5-10bps on $150M in incremental deposits (relationship deposits at ≈100bps lower cost than broker CDs). On an $8B balance sheet that's approximately $1.5-2.5M in annual funding cost savings, or 2-3bps of NIM. Not 35bps.

The actual NIM recovery requires: (1) classified loan exits to complete, restoring asset yield as clean lower-risk originations replace charged-off relationships; (2) broker CD dependency declining organically as deposit growth reduces reliance on brokered funding; (3) rate environment cooperation (asset-sensitive balance sheet benefits from stable or rising rates). The "developments in April and May" language almost certainly reflects continued broker CD market pressure — per Q1 management: "a little over $5 million" in elevated broker costs, 15bps wider than expected. That's a funding cost issue, not an asset quality issue, and it resolves when brokered funding rolls off. Timeline: 2-4 quarters. Venezuelan deposits accelerate that timeline modestly.

The floor management is calling (3.45-3.50%) is not conservative — it assumes broker CD costs normalize by mid-2026 and Venezuelan deposits partially replace them. Peer NIM is expanding (BKU guiding up to 3.06-3.20%; SBCF expanded 17bps QoQ to 3.83%), which confirms AMTB's compression is credit-remediation-specific, not structural.

Who's Selling

Nine directors bought simultaneously at $22.44. Someone is on the other side. The most likely counterparties: momentum and index rebalancing accounts (stock sitting below 50-day MA at 78% of 52-week range, slight one-month negative), Florida CRE credit-concern holders who bought on the improvement narrative in late 2025 and are trimming at cost, and potentially some compliance-driven institutional trimming of Venezuela-adjacent exposure post-Maduro (OFAC counterparty uncertainty has real institutional risk management implications). None of these are informed sellers with a view on Q2 deposits. That's the edge.

Risks

1. Credit builds require ACL replenishment. ACL/NPL coverage at ≈46% ($79.3M ACL / $176.1M NPLs) is critically thin. A defensible floor for a bank with this credit profile is 60-70%. Rebuilding to 65% on current NPLs requires ≈$35M in additional provision. If NPLs grow to $210M (bear case: new inflows outpace exits), 65% coverage requires ACL of $136.5M — a ≈$57M build from current. At Q1 net income of $17.6M/quarter, a $28-30M/quarter incremental provision load turns the bank to a net loss for 2-3 quarters. That is the path to $19.25 (0.86x TBV). It requires both credit deterioration AND regulators pushing for reserve rebuilding simultaneously — not negligible given where coverage sits.

2. Venezuela doesn't move deposits. Macro corroboration is strong — the recovery is happening. The unproven link is whether Venezuelan economic reactivation translates into deposit growth at AMTB at the 8-10% pace implied. Management asserts the linkage. No independent validation exists yet (no competitor bank with comparable disclosure, no Q1 deposit pipeline data external to management).

3. NIM floor doesn't hold. Broker CD costs persisting into H2 2026 or asset mix continuing to shift toward lower yields would push NIM below the 3.40% floor management described. Each additional 10bps below that floor is ≈$7-8M annual NII headwind.

4. Turnaround being run by a finance executive. Iafigliola was CFO then COO — not a credit workout specialist. The COO seat is also still interim (Adrian Rodriguez). The operational role closest to the classified loan book has no permanent owner.

5. OFAC compliance tail. Core PdVSA sanctions technically remain in place. Venezuelan customers with indirect sanctioned-entity exposure create regulatory liability. AMTB's compliance infrastructure matters here and is not externally visible.

Catalysts

CEO comp 8-K/A (imminent, reveals CIC structure) → COO appointment, P=55% (by Sept 30) → Q2 earnings July 22 (deposits vs $8B, NIM floor, NPA trajectory — the binary) → Q2 10-Q ACL update (August) → Q3 credit inflection test (November).

What Would Change Our Mind

• Q2 deposits below $7.8B — Venezuela reactivation not flowing into accounts
• NPAs above $210M in Q3 — new inflows outpacing exits, ACL replenishment forced
• ACL reduced — management concealing deterioration rather than remediating it
• CFO departure mid-turnaround — the one clean execution signal remaining
• BKU or SBCF discloses Venezuelan deposit stress — directly contradicts the moat

Evidence