Popular Inc (NASDAQ: BPOP) — The Deposit Moat Is Already Priced

Published: 2026-09-11 · Verdict: Hold · Entry price: $135 · Price target: $175 · Research confidence: High (90%)

Executive conclusion

Analyst Take

Popular is a high-quality bank at a price that now requires high-quality outcomes. At the September 11, 2026 closing price of $163.71, the shares trade at approximately 11.1 times trailing estimated earnings of $14.79, 1.86 times June tangible book value of $87.94, and 1.63 times reported common book value of $100.38. The approved $0.90 quarterly dividend produces a 2.2% indicated yield. The recommendation is HOLD, with a twelve-month base-case value of $175, including one year of dividends, and a preferred accumulation price of $135. The latter would equal about 9.1 times current trailing earnings and 1.54 times June tangible book, offering more protection against credit, margin, and multiple normalization. [S1][S12][S26]

The central thesis is that Banco Popular de Puerto Rico owns an unusually valuable deposit, payments, distribution, and relationship franchise. June 2025 office-assigned data attributed 64.6% of Puerto Rico deposits to Banco Popular, compared with 15.8% for FirstBank and 11.7% for Oriental. A Federal Reserve review separately identified Banco Popular as the territory’s leading institution in deposits, small-business lending, and home mortgages. That scale is economically visible in a 1.10% second-quarter cost for deposits excluding Puerto Rico public funds, extensive transaction-fee income, and the ability to fund a $79 billion balance sheet primarily with deposits. [S1][S13][S14]

The investment case has strengthened materially since 2024. Second-quarter 2026 net income was $278.2 million, diluted EPS was $4.35, reported net interest margin was 3.66%, and CET1 was 16.08%. Management raised 2026 net-interest-income guidance to 8%-9%, introduced a 14%-17% through-cycle ROTCE objective, authorized another $1 billion of repurchases, and subsequently obtained board approval for the 20% dividend increase. The shares outstanding have fallen about 20% since 2021. These are strong operating and capital facts. [S1][S3][S5][S8][S26]

Two corrections materially temper the draft’s conclusion. First, the reported 17.02% ROTCE is a non-GAAP measure whose denominator adds back average unrealized losses on available-for-sale securities and losses retained from securities transferred to held-to-maturity. Return on tangible common equity before that denominator adjustment was 20.12%. The adjustment makes the measure more comparable across rate-driven AOCI cycles, but it must not be combined mechanically with the $87.94 tangible-book denominator in a residual-income model without acknowledging the measurement mismatch. Second, the draft’s probability-weighted valuation arithmetic was wrong: weighting approximately $101 bear, $175 base, and $216 bull values at 25%, 50%, and 25% produces about $167, not a value in the mid-$170s. That corrected expected value is only modestly above the current price. [S1]

The strongest bull counter-case is that Popular is in the early stages of a structural earnings improvement. Reinvestment of low-yield securities, a low-beta nonpublic deposit base, moderate loan growth, transformation benefits, and a declining share count could keep normalized tangible returns near 16%-17%. First BanCorp and OFG already demonstrate that Puerto Rico banks can generate high-teens returns, while Bank of Hawaii shows that a durable island deposit franchise can command roughly two times tangible book. Under the bull assumptions, about $17 of forward EPS and a 12.5-times multiple produce value above $210 before dividends. [S18][S19][S20]

The strongest disconfirming evidence is credit formation. The $71.3 million charge-off on a Puerto Rico telecommunications borrower can reasonably be excluded when estimating recurring losses, producing a 0.33% second-quarter charge-off rate. It cannot be erased from capital economics, and the same quarter brought two unrelated commercial nonaccruals totaling approximately $129 million. That followed two large commercial nonperformers totaling about $188 million in the third quarter of 2025, one of which was the telecom credit. Management’s characterization of the problems as borrower-specific remains a hypothesis, not independent evidence. [S1][S2][S5][S22]

Governance also requires proof. Javier Ferrer publicly said he was retiring to focus on his health and family, so the reason is not wholly undisclosed as the draft suggested. Specific medical circumstances appropriately remain private. Nevertheless, Ferrer served only fourteen months as CEO, received a $1.62 million short-term award, a $2.6 million restricted-equity award, and a twelve-month consulting agreement at $100,000 per month. Jorge García’s internal promotion preserves institutional knowledge, but the second CEO handoff in fourteen months, accompanied by CFO and chief-risk-officer changes, raises the execution hurdle. [S1][S7]

Investment conviction is moderate; evidence quality is high for current financial condition and only medium for earnings durability. The next decision sequence is concrete: new commercial nonaccrual formation; criticized-asset movement; nonpublic deposit balances and cost; securities reinvestment yield; net expense savings; average repurchase price; and García’s capital target. The call would become more constructive if trailing ROTCE remained at least 15% through two quarters of higher credit cost while nonpublic deposits and tangible book per share compounded. It would become more defensive if unrelated commercial nonaccruals continued above $100 million per quarter, nonpublic funding cost rose without offsetting asset yields, or capital distributions slowed because losses or regulation consumed the apparent surplus.

Stock Price Action — Five-Year Event Map

The stock’s five-year record spans a post-pandemic liquidity boom, the 2023 regional-bank panic, a margin recovery, and a rerating toward premium island-bank status. BPOP closed at $74.53 on September 13, 2021 and $163.71 on September 11, 2026, an unadjusted price gain of 119.7% before dividends. The interval’s intraday low was approximately $49.34 on March 15, 2023; its high was $179.24 on August 17, 2026. The latest 52-week range was $108.74-$179.24, placing the current price 8.7% below the high and roughly 78% of the way from low to high. The stock is 31.2% above its September 11, 2025 close of $124.80. These prices are observations; explanations of the moves are inference unless tied to a dated disclosure. [S12]

• September 2021-January 2022: BPOP rose from the mid-$70s to an intraday $99.49 in January 2022. Popular earned $934.9 million in 2021, aided by post-pandemic liquidity and favorable reserve dynamics. Higher-rate expectations plausibly contributed to the share move, but daily prices cannot allocate causality among rates, sector positioning, earnings, and capital expectations. [S3][S12][S25]

• January-December 2022: the shares fell to $66.32 at year-end even though reported net income reached $1.10 billion and diluted EPS reached $14.63. Reported earnings included large Evertec-related gains and a deferred-tax valuation benefit, while rising rates produced securities marks and future deposit-cost uncertainty. The price decline and financial items are facts; the proposition that investors discounted the nonrecurring earnings is an inference. [S3][S12]

• March 2023 banking panic: BPOP reached its five-year low near $49.34 as failures of mainland regional banks intensified concerns about uninsured deposits, securities duration, and liquidity. Popular’s subsequent 2023 annual report discussed the wider banking-sector shock. Popular’s customer funding was less dependent on the concentrated venture deposits implicated in those failures, but its large securities portfolio and bank classification made it exposed to the same market factor. [S12][S24]

• Second half 2023 through 2024: BPOP recovered to $82.07 at the end of 2023 and $94.06 at the end of 2024. Deposits remained stable, CET1 remained strong, the securities-book valuation gap began to heal, and management resumed buybacks in the third quarter of 2024. The recovery is consistent with improving confidence in funding durability; it does not prove the 2023 price was entirely irrational because 2023 net income had fallen to $541.3 million. [S3][S23][S24]

• April-August 2025: an April low near $78 was followed by a move above $125. The company reported improving margins, loan growth, and repurchases, and completed the planned succession from Ignacio Alvarez to Javier Ferrer. The direction of fundamentals supports the rerating, although rate expectations and bank-sector rotation also influenced returns. [S3][S12]

• October 2025 credit reset: the stock moved from $128.04 on September 26 to an intraday $108.74 on October 30. On October 23, Popular disclosed that nonperforming loans had risen $190.6 million sequentially, principally because two unrelated commercial exposures of $158.3 million and $30.1 million entered nonaccrual. That filing supplies company-specific evidence for part of the selloff; it is more defensible than attributing the entire move to generalized regional-bank anxiety. [S12][S22]

• January-July 2026: BPOP rose from $124.52 at 2025 year-end to $170.88 after the July 23 results. Full-year 2025 net income had risen 36%, first-half 2026 diluted EPS reached $8.13, margin improved, and shares continued to be retired. On July 23 itself, BPOP opened at $175.12 and closed at $170.88 despite strong earnings, consistent with investors balancing higher guidance and capital return against credit formation and a fuller valuation. That interpretation is plausible, not demonstrable from the tape alone. [S1][S3][S12]

• August-September 2026: the stock reached $179.24 intraday on August 17 and then declined to $163.71. Ferrer sold 35,000 shares on August 10 at a weighted-average $175.433 and retired at month-end. The transaction and leadership change are facts; attributing the subsequent decline to either event would overstate the evidence because sector rotation, profit-taking, and rate expectations are competing explanations. [S7][S10][S12]

The five-year gain reflects real earnings, capital, and franchise progress. It also means the inherited cheapness thesis is stale. In 2023, investors could buy near or below tangible book amid system stress. The current multiple assumes a durable premium return and leaves much less room for borrower, funding, or management surprises.

Verdict: The event map supports the quality of the operating recovery, but the approximately 120% price appreciation has converted valuation from a catalyst into a load-bearing assumption. Company-specific credit formation explains more of the October 2025 drawdown than the draft acknowledged, while daily price action still cannot prove causality. [S12][S22]

Business Overview

Popular, Inc. is a Puerto Rico corporation and regulated financial holding company. Its principal banking subsidiaries are Banco Popular de Puerto Rico, or BPPR, and Popular Bank, its New York-chartered mainland subsidiary. BPPR provides retail and commercial banking, mortgage origination and servicing, cards, auto and equipment finance, leasing, trust, brokerage, investment, and insurance-related services across Puerto Rico and the U.S. Virgin Islands. Popular Bank operates primarily in New York, New Jersey, and Florida and increasingly emphasizes commercial relationships. Popular common stock trades directly on Nasdaq; it is not an ADR, MLP, partnership, or K-1 security. Tax consequences depend on the shareholder’s residence and account, but the listed security is ordinary corporate common equity. [S2][S3][S7]

The business is readily understandable as four linked engines. First, Popular gathers transaction, savings, time, commercial, municipal, and Puerto Rico government deposits. Second, it deploys the funding into loans, Treasury and agency securities, and cash. Third, it earns fees from deposit services, debit and credit cards, mortgage banking, insurance, brokerage, asset management, trust, and other operating services. Fourth, retained income and surplus regulatory capital fund technology, organic growth, dividends, and repurchases. The essential drivers are earning-asset volume and yield, deposit mix and cost, credit losses, fee intensity, operating expense, and the common equity regulators and management require.

At June 30, 2026, Popular had $79.0 billion of assets, $70.2 billion of deposits, $39.75 billion of loans held in portfolio, $31.26 billion of investment securities, $1.46 billion of borrowings, and $6.43 billion of stockholders’ equity. Deposits represented about 98% of deposits plus borrowings, and the loan-to-deposit ratio was only 56.6%. This is therefore not just a lender with excess cash. Nearly $31.3 billion of securities makes portfolio yield, duration, collateral availability, AOCI, and reinvestment strategy central to earnings. [S1][S2]

The ending deposit composition must be stated correctly. The draft mixed average deposit categories with the $70.2 billion period-end total. At quarter-end, nonpublic deposits were $47.53 billion: $15.09 billion of demand deposits, $23.38 billion of nonbrokered savings/NOW/money-market accounts, $79.5 million of brokered savings/NOW/money-market deposits, $8.11 billion of nonbrokered time deposits, and $873.1 million of brokered CDs. Puerto Rico public deposits were $22.70 billion: $11.44 billion of demand, $10.35 billion of savings/NOW/money-market balances, and $918.3 million of time deposits. Public deposits represented 32% of consolidated deposits and 39% of BPPR deposits. [S1][S2]

Those public balances are both an asset and a qualification. Their second-quarter cost was 2.61%, compared with 1.10% for core deposits excluding public funds. They are tied to short-term rates with a lag, require high-quality pledged collateral, and move with taxes, disbursements, federal transfers, and government cash management. Because they are collateralized, they do not carry the same uninsured-run profile as uncollateralized institutional deposits; because they are institutionally managed and seasonally volatile, they are not equivalent to sticky household transaction balances. Management expected them to fall from $22.7 billion to a $20-$22 billion range for the rest of 2026. [S1][S2][S5]

Revenue is recurring but not contractually fixed. In the second quarter, $693.4 million of net interest income plus $180.5 million of noninterest income produced $874.0 million of revenue before provision and expenses; net interest income supplied 79%. The fee mix was broader than the draft described: service charges on deposits were $39.0 million, debit-card fees were $31.5 million, credit-card fees were $34.8 million, other banking fees were $12.1 million, insurance fees were $12.6 million, brokerage and asset-management fees were $10.0 million, trust fees were $7.8 million, mortgage-banking income was $6.3 million, and other operating income was $26.4 million. The draft incorrectly labeled the $31.5 million debit-card figure as deposit-service fees. [S1]

Revenue stability is moderate rather than contractual: annual net interest income plus noninterest income was approximately $2.48 billion in 2021, $2.94 billion in 2022, $2.64 billion in 2023, $2.80 billion in 2024, and $3.05 billion in 2025. The range is manageable for a bank, but the composition moved with rates, securities yields, deposit repricing, and one-time gains. Net income was substantially more volatile because provision, taxes, transaction items, and securities effects amplify movements in pre-provision revenue. [S3][S4][S23][S24][S25]

The loan portfolio is diversified by product but concentrated by geography. June loans included about $20.1 billion of commercial credits, $1.7 billion of construction loans, $8.8 billion of mortgages, $2.0 billion of leases, and $7.2 billion of consumer loans. Approximately 70% of consolidated loans were associated with Puerto Rico operations, and real-estate-related categories represented 54% of loans. Non-owner-occupied commercial real estate was $5.62 billion, including roughly $674 million of office exposure, while multifamily loans were about $2.4 billion and concentrated in New York, South Florida, and Puerto Rico. The office balance was only about 1.7% of total loans, but CRE and multifamily still require market-by-market analysis. [S1][S2]

BPPR is the economic core. Its second-quarter net interest income was $589.9 million with a 3.85% margin, compared with $113.1 million and 3.17% at Popular Bank. BPPR also held about $28.0 billion of the $39.75 billion loan portfolio and nearly all public deposits. Popular Bank supplies geographic diversification and access to much larger mainland credit pools, but its deposit cost was 2.73% and it lacks BPPR’s island-wide distribution advantage. The strategic rationale is selective commercial growth, not recreating the Puerto Rico branch network. [S1][S3][S6]

Customer value comes from convenience, reliability, embedded payments, product breadth, and local underwriting knowledge. A household can combine direct deposit, cards, bill payment, savings, mortgage, insurance, and investments. A business or municipality can combine operating accounts, merchant settlement, payroll, cash management, credit, trust, and collateral services. Each product can be moved independently, but moving the whole relationship creates documentation, systems, permissions, account-history, and behavioral costs.

Unrecognized economic assets include the Popular brand, customer relationships, transaction histories, licenses, fraud and underwriting data, branch and ATM coverage, payment connectivity, and the trained workforce. These assets are not fully recorded in tangible book. Their economic value is evidenced indirectly by leading market positions, a low nonpublic deposit cost, broad fee generation, and the capacity to serve virtually every major Puerto Rico customer segment. Goodwill of roughly $790 million is recorded separately and is excluded from tangible book. Potential recovery of reserved mainland deferred-tax assets is contingent upside rather than a current franchise asset available to shareholders. [S2][S3][S14]

Bank cash-flow statements require specialized interpretation. Deposits are simultaneously funding and a customer product; loans and securities are operating assets economically even when accounting classifications place their cash flows outside operating activities. Conventional enterprise value, EBITDA, working capital, and industrial free cash flow can therefore mislead. Net income, ROA, common-equity returns, tangible-book growth plus distributions, loss formation, liquidity, and regulatory capital generation are more decision-useful.

Verdict: Popular is a comprehensible deposit-funded spread-and-fee business with an exceptional Puerto Rico relationship franchise. Its revenue base is diversified but still 79% net-interest income, and its low loan-to-deposit ratio makes securities and public-funding management just as important as loan production. [S1][S2]

Industry Dynamics

Puerto Rico banking operates inside the U.S. dollar, federal deposit-insurance, Federal Reserve, and U.S. payments systems. Banks face federal capital, liquidity, anti-money-laundering, cybersecurity, consumer-protection, fair-lending, and safety-and-soundness rules, together with Puerto Rico requirements. Entry therefore requires more than a mobile application: a challenger needs a charter or regulated partner, common capital, risk and compliance infrastructure, fraud controls, customer trust, deposit access, experienced lenders, and either physical distribution or sustained digital acquisition spending. [S2][S3][S14]

The territory’s banking structure is highly concentrated. FDIC Summary of Deposits-derived data counted only six institutions with Puerto Rico offices in June 2025 and calculated a statewide deposit HHI of 4,606. BPPR held $54.14 billion, or 64.62%, of office-assigned deposits across 158 offices; FirstBank held $13.27 billion, or 15.84%; Oriental held $9.82 billion, or 11.72%; and Citibank held $5.36 billion, or 6.40%. The statewide figures are descriptive rather than formal antitrust-market definitions, and internal office assignment can alter where corporate and public deposits appear. The gap is nevertheless too large to be dismissed as measurement error. [S13]

A Federal Reserve CRA evaluation supplies an independent operating cross-check. It described BPPR as the leading Puerto Rico institution in deposit share, small-business lending, and home-mortgage lending, with major competitors Citibank, FirstBank, and Oriental. The review also found delivery systems readily accessible throughout the assessed communities. The CRA rating is not a safety-and-soundness opinion, but the distribution and market-rank observations support the franchise analysis. [S14]

The addressable market is not appropriately estimated by applying a generic bank penetration rate to GDP. Popular competes for household and commercial deposit balances, transaction volume, cards, consumer credit, commercial and construction loans, mortgages, leasing, trust and investment assets, insurance distribution, and selected mainland commercial relationships. Puerto Rico is economically the core domestic market even though it is a U.S. territory; the mainland business is a separate U.S. growth pool. Foreign-currency translation and international revenue are not material operating drivers. [S2][S3]

Near-term macro evidence is supportive but less uniformly positive than the draft implied. Puerto Rico’s seasonally adjusted unemployment rate was 5.8% in June 2026, low by historical standards but up from 5.6% a year earlier. Nonfarm payrolls were approximately 955,000, only about 1,000 above June 2025. This describes a resilient rather than rapidly expanding labor market. The low unemployment level supports consumer credit; the year-over-year increase and flat payroll count warn against extrapolating a new structural boom. [S16]

Federal and government flows are similarly two-sided. The fiscal 2026 certified budget totaled $32.7 billion, including $14.2 billion of federal funds. Yet the same budget said federal funding had declined about $1.2 billion and held back 5% of most agency spending for eight months to protect against further reductions, revenue weakness, or higher Medicaid costs. FEMA’s cumulative recovery obligations remain a meaningful source of contractor activity and liquidity, but obligations, disbursements, loan demand, and final bank earnings occur at different times. [S15][S17]

Puerto Rico’s structural constraints include population aging and migration, energy cost and reliability, storm exposure, fiscal oversight, and reliance on federal transfers. Reconstruction can support contractors, consumption, public balances, and infrastructure lending for years. It can also produce uneven timing: public cash may accumulate in deposits before projects create private credit demand, while permitting, procurement, grid limitations, or federal changes delay deployment. Popular already owns most local deposits, so long-term value is more likely to come from wallet share, service productivity, pricing discipline, and per-share capital management than from large incremental deposit share.

Competition is stable at the center and more intense at the edge. FirstBank and Oriental are substantial, well-capitalized competitors. National banks can target corporate, wealth, card, or digital relationships without building island-wide retail networks. Online banks and fintechs can bid directly for rate-sensitive savings or provide payments, lending, brokerage, and merchant tools. Management described Puerto Rico loan pricing as rational but acknowledged aggressive competition from community banks in Florida, large institutions in New York, and online deposit pricing at Popular Bank. [S5][S6][S18][S19]

Competition differs by product. Primary transaction accounts compete on convenience, reliability, branch and ATM availability, digital quality, fraud control, and habit. CDs and online savings compete much more directly on rate. Commercial loans compete on relationship knowledge, certainty, structure, covenant quality, and spread. Cards compete on underwriting, rewards, acceptance, and servicing. Mortgages compete on price, speed, government-program expertise, and servicing. Trust and investment services depend on confidence and adviser continuity. Popular’s advantage is therefore strongest when the customer uses several linked services and weakest in a single rate-comparison product.

Current industry profitability is attractive but not unique to Popular. First BanCorp reported second-quarter ROA of 2.02%, ROE of 19.49%, a 4.87% margin, and CET1 of 16.96%. OFG reported ROA of 1.86%, ROTCE of 17.18%, a 5.40% margin, and a higher 1.22% charge-off rate. Popular produced 1.41% ROA and 17.02% adjusted ROTCE. Smaller rivals’ ability to match or exceed current returns is disconfirming evidence against treating market share as automatic monopoly profit. [S1][S18][S19]

A supply-side capital-cycle lens explains why the oligopoly can remain profitable without being invulnerable. Banking capacity cannot expand solely because current margins are attractive: loans consume capital and liquidity, compliance has high fixed costs, experienced lenders are scarce, and deposit acquisition can be expensive. Conversely, three well-capitalized incumbents can still overproduce credit if they chase a limited set of high-quality borrowers. The most useful early warnings are falling spreads, looser structures, accelerating construction commitments, or rising brokered and promotional funding—not branch counts alone.

Foreign low-cost labor or manufacturing capacity is not a direct competitive threat because an FDIC-insured relationship cannot be imported like a manufactured product. Software development, servicing, and call-center activities can be outsourced, and global cyber attackers can exploit digital channels. The economically relevant substitutes are national-bank scale, digital distribution, instant rate comparison, and fintech interfaces rather than wage arbitrage.

Verdict: Puerto Rico’s three-bank core, regulatory fixed costs, customer trust, and distribution requirements support above-average industry economics. The counterweight is that peers already earn excellent returns, marginal products are increasingly price-transparent, and federal support is declining at the same time that public deposits remain unusually large. [S13][S15][S18][S19]

Competitive Position

Popular’s clearest advantage is the combination of distribution density, relationship breadth, transaction habit, and data. A 64.6% office-assigned deposit share, 153 Puerto Rico retail branches at year-end 2025, and 582 Puerto Rico ATMs make BPPR a default financial interface for many households, employers, merchants, municipalities, and institutions. Scale spreads technology, security, marketing, and compliance costs across a larger base and supplies more opportunities to cross-sell cards, loans, insurance, investments, trust, and cash-management services. [S3][S13][S14]

A moat claim must connect to an economic outcome. Popular’s moat is supported by the 1.10% cost of deposits excluding Puerto Rico public funds, $117.5 million of quarterly banking fees, a broad noninterest-income mix, and leading small-business and home-mortgage positions. If the moat weakened, one would expect nonpublic deposit attrition, higher relative deposit cost, lower primary-account usage, falling fee retention, or deterioration in loan pricing and acquisition expense. Current evidence does not show those outcomes, but public disclosure lacks account-level retention and product-depth cohorts. [S1][S14]

Brand matters because banking requires repeated trust. Popular was founded in 1893, and its name is embedded in Puerto Rico commerce. A brand lowers perceived risk when customers entrust payroll, savings, payment credentials, borrowing relationships, and business operating cash to an institution. Brand is not an accounting asset, and heritage by itself does not establish pricing power. Its economic relevance is supported only insofar as customers maintain low-cost balances, use multiple services, and tolerate neither inferior digital service nor reliability failures.

The moat is a reinforcing system rather than a patent-like barrier. Branches acquire and service customers; digital channels lower transaction costs; cards and payments increase engagement; transaction data improve fraud detection and underwriting; commercial relationships create operating deposits; product breadth raises wallet share; and scale funds technology and compliance. Each component can be copied individually. Reproducing all of them at island-wide scale while earning an acceptable return is considerably harder.

Customer switching costs are moderate and relationship-dependent. Opening another savings account is easy, and borrowers can solicit competing loan quotes. Moving direct deposits, automated bills, merchant settlement, payroll, treasury permissions, cards, historical statements, and several family or business products is more burdensome. Municipalities and institutions also face procurement, collateral, documentation, and system-integration work. Switching costs therefore rise with relationship breadth; they are operational and behavioral, not contractual lock-in.

Deposit quality is heterogeneous. Noninterest-bearing operating accounts and ordinary savings balances are more franchise-like than brokered CDs or public funds. Brokered savings and CDs totaled approximately $953 million, modest against $70.2 billion of deposits. Public deposits were $22.7 billion, cost more than twice as much as nonpublic deposits, required pledged securities, and can move with the fiscal calendar. A sound moat analysis assigns the greatest value to nonpublic operating balances and does not apply the 64.6% market-share label indiscriminately to every deposit dollar. [S1][S2]

Local credit knowledge may reduce information asymmetry but cannot eliminate concentration risk. Popular’s long operating history should improve collateral knowledge, borrower access, and workouts. Consumer losses and delinquencies have generally remained manageable. However, local knowledge did not prevent a $155 million telecom exposure from becoming nonperforming, nor two further commercial borrowers totaling $129 million from entering nonaccrual. Relationship banking can improve selection; it can also create familiarity risk or allow single-name exposures to become economically material. [S1][S5][S22]

First BanCorp provides the most direct operating challenge. BPPR’s Puerto Rico deposits were roughly four times FirstBank’s office-assigned balance, yet First BanCorp reported higher recent ROA and ROE. At $27.69, First BanCorp traded around 2.18 times its $12.68 tangible book, above BPOP’s 1.86 times. The market therefore does not reward scale alone; it rewards high returns, credit outcomes, and distributable capital. [S12][S13][S18]

OFG offers a different comparison. It reported a 5.40% margin and 17.18% ROTCE, similar to Popular’s adjusted tangible return but with a 1.22% annualized charge-off rate. At $52.66, OFG traded near 1.69 times its reported $31.12 tangible book. Popular’s larger low-cost funding base merits some premium, but that premium is justified only if scale converts into superior through-cycle credit-adjusted returns rather than merely a lower headline margin. [S12][S19]

Bank of Hawaii is a structural analogue rather than a competitor. Both banks serve island economies with high local recognition, geographic concentration, storm exposure, and limited natural market expansion. Bank of Hawaii traded near two times its $38.01 tangible book at $75.80 while producing lower current ROA but exceptionally low losses. The comparison shows that geographic concentration need not preclude a premium multiple; it also shows why even a few large commercial nonaccruals can matter to BPOP’s valuation. [S12][S20]

Popular Bank’s mainland position is substantially less protected. It competes in New York, New Jersey, and Florida against community banks willing to price aggressively, national institutions with deep products and technology, and online banks bidding for deposits. Management has rejected an expensive de novo branch buildout and instead emphasizes selective commercial teams and relationship deposits. Mainland loan growth should be judged by relationship-funding attachment, origination spreads, risk grades, and vintage loss—not volume alone. [S3][S5][S6]

Digital execution can widen or erode the moat. If mobile service, fraud controls, and digital origination meet customer expectations, branches become a trust and advice layer rather than a high-cost transaction channel. If service lags, fintechs can capture valuable interactions while Popular retains the fixed cost of physical distribution. The 2025 rollout of a commercial cash-management platform and fully digital consumer-credit origination shows investment, but public data do not yet establish incremental customers, lower unit cost, or durable product-level returns. [S3][S9]

Verdict: Popular has a real but qualified competitive advantage built on nonpublic deposits, payments, brand, distribution, and bundled relationships. Smaller Puerto Rico banks’ excellent returns, the economics of public deposits, mainland competition, and recent commercial-credit events prevent treating dominant market share as monopoly economics. [S1][S18][S19]

Growth History and Forward Opportunities

Recent growth has come from margin recovery, moderate loan expansion, securities reinvestment, fee activity, and a shrinking share count rather than rapid balance-sheet expansion. Loans increased 4.1% year over year to $39.75 billion at June 2026. Deposits increased 4.5%, but $3.0 billion of the sequential increase came from Puerto Rico public funds, while ending deposits excluding public funds fell $414 million from March and rose $1.23 billion from June 2025. Management said average nonpublic balances performed better than the period-end comparison. [S1][S5]

Management continued to expect loan growth near the low end of its 3%-4% range. Large Puerto Rico transactions were closing more slowly, public-works financing remained further out, U.S. construction payoffs restrained balances, and New York multifamily conditions remained selective. Moderate growth can create value if spreads and losses compensate equity; faster growth would be a warning if it required weaker structures or promotional funding. [S1][S5][S6]

The largest visible earnings opportunity is securities repricing. Popular had $31.26 billion of securities at June 30. During the second quarter, it purchased about $1.1 billion of Treasuries with roughly 2.8 years of duration and a yield near 3.9%. Average Treasury balances rose $1.4 billion sequentially, helping increase the fully taxable-equivalent securities yield by 15 basis points. As legacy low-yield holdings mature or amortize, reinvestment can offset deposit and loan-yield pressure. The benefit depends on market rates, prepayment speeds, collateral needs, and management’s willingness to take duration. [S1][S5]

Net interest income remains the main growth driver. Management raised full-year 2026 guidance from 5%-7% growth to 8%-9%, citing public-deposit volume and earning-asset deployment, while expecting margin to remain broadly stable through year-end. This is management guidance rather than a fact about future earnings. It can be tested against quarterly net interest income, average public and nonpublic balances, securities yields, loan spreads, and deposit costs. [S1][S5]

Fees offer smaller but more diversified growth. Management raised its quarterly noninterest-income range to $165-$170 million after $180.5 million in the second quarter. Cards benefited from purchase activity, while the quarter also contained a $3 million unrealized gain within equity-method investment income. Sustainable growth should be measured through card purchase volume, accounts, service usage, and recurring fees rather than simply annualizing the strongest quarter. [S1]

Transformation could improve operating leverage, but disclosures require careful reconciliation. The 2026 expense outlook embeds roughly $50 million of efficiency actions, while the proxy reported that 2025 initiatives had produced about $20 million of sustainable annual savings. Those are not contradictory if the former includes later or broader actions, but gross initiatives are not the same as net expense reduction. Wage inflation, cybersecurity, compliance, cloud migration, and digital investment can absorb benefits. [S5][S6][S9]

Puerto Rico reconstruction remains an opportunity without being a perpetual growth engine. Federal obligations and a $14.2 billion federal component of the fiscal 2026 budget support contractors, infrastructure, consumption, and public balances. The same budget reported a $1.2 billion reduction in federal funding and prepared for further cuts. The investable question is therefore conversion: how much public liquidity becomes well-priced private credit and recurring fee activity after delays and competitive bidding. [S15][S17]

Popular Bank provides a larger addressable market but a weaker moat. Selective C&I, owner-occupied real estate, multifamily, and construction relationships in New York and Florida can diversify earnings. The good version produces relationship deposits and returns above the cost of common equity; the poor version adds competitive assets funded by expensive online balances. Management’s refusal to chase loan volume is sensible, but it must be verified through deposit attachment and vintage performance. [S5][S6]

Repurchases are an important per-share growth engine. A fully used $1 billion authorization at $164 would retire approximately 6.1 million shares, about 9.5% of June shares, before considering earnings, option issuance, price changes, and regulatory limits. The arithmetic illustrates capacity, not a forecast. At 1.86 times tangible book, value creation depends much more on sustainable returns than it did when shares were repurchased near $96 in 2024. [S1][S8]

Potential mainland deferred-tax-asset realization is upside rather than a product strategy. Management said net operating losses begin to expire in 2028 and rejected acquiring a business solely to consume the tax asset. The right catalyst is sustained profitable mainland growth that creates taxable income, not a transaction whose operating return is inferior to its tax benefit. [S2][S5]

Verdict: The product and service outlook is credible but primarily margin-, fee-, and per-share-driven. Securities reinvestment, moderate loans, deeper customer relationships, and repurchases can sustain growth; shrinking federal support, slow transaction timing, and tougher mainland competition constrain a high-volume thesis. [S1][S5][S15]

Financial Quality

Popular’s five-year history shows a consistently profitable franchise with substantial earnings cyclicality and important comparability traps.

Revenue is net interest income plus noninterest income, not industrial-company sales. The figures reconcile Company Financials multi-period statements to the annual and quarterly filings. [S1][S3][S4][S23][S24][S25]

The 2022 peak was not a clean recurring run rate because transactions involving Evertec and deferred-tax valuation changes materially affected income. The 2023-2024 trough was also not a neutral run rate: higher deposit costs and low-yield securities accumulated during excess-liquidity years compressed profitability. The current period benefits from the opposite process as deposit pricing eases and securities reprice upward. Earnings are therefore above a neutral cycle but not necessarily at the absolute peak because reinvestment and capital optimization remain incomplete.

Margin is load-bearing. Reported NIM improved from 3.24% in 2024 to 3.49% in 2025 and 3.66% in the second quarter of 2026. Fully taxable-equivalent margin reached 4.17%. Second-quarter net interest income of $693.4 million increased 3.5% sequentially and 9.8% year over year; one additional day contributed about $5 million. Securities income, commercial-loan growth, and higher public-deposit balances were positive, while interest expense on deposits was an offset. [S1][S3]

Deposit costs provide the strongest quantitative moat evidence. Total cost fell from 1.78% in the prior-year quarter to 1.57%, while core cost excluding Puerto Rico public deposits was 1.10%. BPPR’s total deposit cost was 1.32%, compared with 2.73% at Popular Bank. The consolidated number hides this difference. Future earnings depend not only on benchmark rates but on how much nonpublic operating money remains at BPPR versus public, time, brokered, and mainland balances. [S1]

Profitability should be assessed with bank-appropriate measures. Second-quarter ROA was 1.41%, reported return on average common equity was about 15.2%, and management’s adjusted ROTCE was 17.02%. Conventional ROIC is not decision-useful for the consolidated bank because deposits and debt are operating funding and cash and securities are operating assets. Subtracting deposits from invested capital would manufacture unstable results. ROA, common-equity returns, tangible-book growth plus distributions, credit-adjusted spread, and regulatory capital generation are the appropriate analogues. [S1][S2]

The ROTCE reconciliation is essential. Popular first calculated a 20.12% tangible common return before adjusting for unrealized AFS losses and losses remaining from AFS securities transferred to HTM. It then added approximately $825 million of average AFS losses and $184 million of average transferred-HTM losses back to average tangible common equity, producing 17.02%. The methodology prevents falling AOCI from mechanically shrinking the denominator and inflating return. It is reasonable, but it means the 17.02% measure should not be compared mechanically with a price divided by period-end tangible book that includes AOCI. [S1]

Credit is the main challenge to earnings quality. At June 30, loans held in portfolio were $39.75 billion, nonperforming loans were $413.4 million, and the NPL ratio was 1.04%. The allowance was $784.8 million, or 1.97% of loans and 190% of NPLs. Total nonperforming assets, including $83.7 million of loans held for sale and $49.6 million of OREO, were $546.7 million. Headline coverage is substantial, but reserve ratios are not substitutes for name-level concentration when a single borrower can create a $71 million charge-off. [S1][S2]

Quarterly net charge-offs were $104.1 million, or 1.05% annualized. The resolved telecom relationship produced $71.3 million of the charge-off; excluding it, management calculated 0.33%. That adjustment is useful for estimating recurring loss but incomplete as a risk conclusion. Two unrelated commercial borrowers totaling $129 million entered nonaccrual, and nonconsumer NPL inflows were $137 million. Management raised full-year charge-off guidance to 65-80 basis points. [S1][S5]

The sequence began before the latest quarter. In the third quarter of 2025, NPLs rose $190.6 million as two commercial credits of $158.3 million and $30.1 million entered nonaccrual. One became the telecom exposure resolved in 2026; the other produced an initial $13.5 million charge-off. Thus, the latest quarter did not introduce the first large-credit problem—it resolved one known name while adding two more. This weakens, but does not disprove, the claim that each problem is idiosyncratic. [S22]

Allowance varies sensibly by portfolio. At June 2026, consolidated coverage was approximately 1.56% for commercial, 0.83% for construction, 0.91% for mortgages, 0.90% for leasing, and 5.00% for consumer loans. Card coverage was about 6.85%, personal-loan coverage 5.25%, and auto coverage 4.37%. Higher consumer frequency justifies higher percentages. Large commercial exposures can nevertheless create more loss dollars than years of ordinary consumer charge-offs, even at a lower portfolio-average reserve rate. [S1]

Commercial real estate deserves proportional rather than sensational treatment. Non-owner-occupied CRE was $5.62 billion; retail represented about 32%, hotels 19%, and office 12%. The $674 million office balance was roughly 1.7% of total loans. Multifamily was approximately $2.4 billion, with more than half in New York and another meaningful balance in South Florida. The modest office percentage reduces a common regional-bank risk, while geographic and sponsor concentration still require criticized-asset disclosure. [S2]

The securities book is large but primarily duration rather than credit risk. Available-for-sale securities are marked through AOCI. Held-to-maturity securities had a carrying value around $6.20 billion and fair value near $6.18 billion at June 30, although book value retained losses from earlier AFS transfers. Most holdings are U.S. Treasury, agency, or government-sponsored instruments. Approximately $7.3 billion of HTM securities were pledged for public and trust deposits, making encumbrance and collateral release relevant to liquidity. [S2]

Accounting conservatism is mixed rather than uniformly strong. CECL recognizes expected loan losses and AFS marks flow through equity, both conservative features. HTM classification suppresses subsequent fair-value movement in reported equity; tax-equivalent and adjusted-return measures require reconciliation; and 2026 presentation reclassified certain reciprocal deposits as nonbrokered and conformed prior-period categories. None of these changes is inherently improper, but economic liquidity and funding quality cannot be read from accounting labels alone. [S1][S2][S3]

Off-balance-sheet credit exposure is material. June commitments included $7.00 billion of unused credit-card lines, $4.36 billion of commercial lines, $1.07 billion of construction lines, $291 million of other consumer commitments, $135 million of standby letters of credit, $9 million of commercial letters, and $22 million of mortgage commitments—approximately $12.88 billion in total. Utilization is uncertain and card lines may generally be reduced, but stress draws can consume liquidity and capital. [S2]

Other obligations include operating leases, representations and warranties, guarantees, litigation contingencies, frozen defined-benefit plans, and $192.7 million of trust-preferred securities guaranteed by Popular. Direct Puerto Rico government loans and investments are much smaller than public deposits and should not be conflated with them. No reviewed filing identified an off-balance-sheet item that independently implies a solvency hole, but commitments make a point-in-time funded-loan balance incomplete. [S2][S3]

Reported operating cash flow was approximately $1.01 billion in 2021, $1.02 billion in 2022, $687 million in 2023, $675 million in 2024, and $878 million in 2025. With approximately $197 million of 2025 capital expenditure, an industrial calculation would produce about $681 million of free cash flow. That is a useful reconciliation, not distributable bank cash: loan growth, securities deployment, deposit movement, collateral, and regulatory capital are economically operating items despite statement classifications. [S3][S4]

Physical capital intensity is low—premises and equipment were about $732 million in June—but economic capital intensity is high because every loan and security requires funding, liquidity, loss capacity, and regulatory capital. Growth creates value only when spread and fees exceed expected losses, operating cost, taxes, and the required return on that common equity.

Trailing diluted EPS of approximately $14.79 is an analyst estimate calculated by replacing the first half of 2025 with reported first-half 2026 results. It reconciles to the current Company Financials trailing P/E. Annualizing the $4.35 second quarter would yield $17.40, but doing so would assume one strong quarter’s provision, balance seasonality, and tax mix recur. [S1][S4]

Verdict: Financial quality is high but not uncomplicated. Capital, funding, spread recovery, and tangible-book growth are strong; adjusted ROTCE requires a denominator reconciliation, the credit sequence now spans three reporting periods, and industrial ROIC or free cash flow should not replace bank-specific return and capital measures. [S1][S2][S22]

Capital Allocation

Popular’s allocation hierarchy is organic balance-sheet support, technology and transformation, common dividends, repurchases, and selective M&A. The company has substantial capital above current regulatory minimums and is intentionally reducing excess common equity. That is rational when stressed losses are covered and shares trade below conservative intrinsic value; it becomes procyclical if management distributes capital at a premium immediately before credit or regulatory requirements rise. [S1][S5][S8]

The net share-count record is strong. Year-end common shares were approximately 79.9 million in 2021, 71.9 million in 2022, 72.2 million in 2023, 70.1 million in 2024, and 65.7 million in 2025; June 2026 shares were 63.87 million. The small 2023 increase shows why gross buybacks must be reconciled with issuance and compensation. Across the full period, the net reduction was about 20%. [S1][S3][S23][S24][S25]

In 2024, Popular spent roughly $217 million to repurchase approximately 2.26 million shares, an average near $96. In 2025, it repurchased about 4.6 million shares for approximately $502 million, an average near $109. First-half 2026 repurchases were about $280 million, including $125 million in the second quarter, and exhausted the prior $500 million authorization. The board then authorized another $1 billion with no expiration, while management indicated an expected $300-$400 million of additional 2026 purchases. [S1][S3][S8][S9]

Price discipline now matters more. Purchases near $96 occurred around 1.3 times contemporaneous tangible book and have generated clear market value. At $164 and 1.86 times June tangible book, a buyback raises EPS and adjusted ROTCE mechanically but creates economic value only if the franchise sustains returns sufficient to justify the premium. The authorization is capacity, not a commitment, and its terms permit modification or suspension.

The quarterly common dividend was formally approved at $0.90 on August 20, payable October 1 to holders of record September 10. Annualized payments of $3.60 equal about 24% of trailing estimated EPS, leaving meaningful earnings retention. Dividend coverage is strong, but subsidiary-dividend capacity, regulatory approval, stress losses, and risk-weighted-asset growth govern the practical payout ceiling. [S1][S26]

The combined payout is much larger than the dividend ratio suggests. At 63 million shares, the annualized dividend is roughly $227 million. Adding a $350 million illustrative second-half repurchase amount would place annualized common distributions above half of current earnings; using the entire $1 billion authorization would exceed one year’s normalized earnings. The strategy is capital optimization, not merely an income policy.

The recent acquisition record is sparse. Popular has not completed a transformative banking acquisition during the reviewed five years. The 2022 Evertec transactions monetized and restructured an investment and commercial relationship rather than adding a conventional bank franchise. Management remains open to mainland opportunities but explicitly rejected buying a company merely to use tax assets. With no recent major acquisition, there is no current M&A return record to reward; a future deal should be tested against repurchases, integration cost, deposit quality, and fully allocated risk-adjusted return. [S3][S5]

Reinvestment centers on customer platforms, cybersecurity, compliance, branches, cloud systems, process redesign, and employee capability. Popular rolled out commercial cash management, digital consumer-credit origination, mortgage-servicing changes, and cloud ERP work. The proxy credited 2025 initiatives with approximately $20 million of sustainable annual savings, while 2026 guidance incorporates broader efficiency actions around $50 million. Net operating leverage—not gross savings—should determine whether these investments earn acceptable returns. [S3][S5][S9]

Stock compensation is modest relative to earnings but is not free. The substantial net share decline shows repurchases more than offset employee grants and vesting. Company Financials and filing review did not identify a material open-market insider purchase during the reviewed period. Forms 3, 4, and 5 instead showed grants, vesting, withholding, routine dispositions, and Ferrer’s August sale. Absence of a purchase is not proof of insider pessimism. [S3][S9][S10][S11]

The compensation framework has meaningful alignment and discretion. In 2025, 81% of the CEO’s target pay and an average 65% for other named executives was at risk. Annual incentives use adjusted net income, adjusted ROTCE, strategic transformation, and individual goals. Half of target long-term awards are performance shares based equally on relative three-year TSR and absolute three-year average adjusted ROTCE; half are time-vested restricted stock, generally over four years. [S9]

The completed 2023-2025 award illustrates the trade-off. BPOP’s 85.3% three-year TSR ranked second among 42 comparator banks, earning 150% of target for that half. Three-year average adjusted ROTCE of 11.02% fell between a 9.5% threshold and 13% target, earning 71.76% for the other half. Combined vesting was 110.88%. Relative TSR captured a successful recovery, but adjusted returns give the committee exclusion discretion and can reward a sector rebound that begins from distress. [S9]

Governance protections include stock-ownership requirements, clawbacks, prohibitions on hedging and pledging, payout caps, and risk-adjustment authority. A qualification is related-party credit. The proxy disclosed approximately $27.0 million of loans to parties related to director Alejandro Ballester and a $29.8 million current commercial-building loan involving interests related to chairman Richard Carrión. Audit Committee approval and current payment status reduce concern, but the exposures deserve disclosure in a bank where credit judgment is central. [S9]

Ferrer’s transition economics are also a governance cost. The 8-K disclosed a $1.62 million cash award, a $2.6 million one-year restricted-stock award, up to three years of medical coverage, and $100,000 per month for twelve months of consulting. The agreement supports continuity and contains restrictive covenants; its generosity after a fourteen-month CEO tenure should nevertheless be weighed against the board’s succession planning. [S7]

Management incentives emphasize earnings, ROTCE, TSR, and capital efficiency. Those metrics align with shareholders when credit and absolute capital constraints remain binding. They can also encourage accelerated buybacks, balance-sheet optimization, and adjusted-item exclusions. The most informative behavioral evidence is therefore future repurchase price discipline and risk performance, not incentive labels alone.

Verdict: Capital allocation has created substantial per-share value, particularly through earlier repurchases at lower multiples. Dividend coverage and dilution control remain strong. The new authorization faces a materially higher valuation hurdle, and transition awards plus related-party credit modestly weaken an otherwise credible governance record. [S7][S8][S9]

Changes and Headwinds — Last Two Years

The largest operating change since 2024 is the restoration of net interest income and returns. Net income rose from $614.2 million in 2024 to $833.2 million in 2025, while adjusted ROTCE improved from 9.85% to 13.04%. First-half 2026 net income reached $523.9 million, and second-quarter adjusted ROTCE reached 17.02%. Higher securities yields, lower deposit costs, loan growth, fee activity, and fewer shares contributed. External rates created both the earlier pressure and later reinvestment opportunity; internal deposit management, underwriting, investment timing, expense actions, and repurchases determined the realized outcome. [S1][S3][S4]

Funding composition changed materially. Total deposits rose from $64.9 billion at 2024 year-end to $66.2 billion in 2025 and $70.2 billion in June 2026. Public deposits rose to $22.7 billion and supplied most of the latest sequential increase. Management’s first-quarter expectation of $18-$20 billion became stale after the June inflow; the updated expectation was $20-$22 billion for the rest of 2026. The revision is evidence of seasonality and forecasting limits, not proof of permanently higher public liquidity. [S1][S5][S6]

Credit moved from broadly benign consumer performance to repeated large commercial events. Two commercial loans totaling about $188 million entered nonaccrual in the third quarter of 2025. One was the telecom relationship, which produced a $71.3 million charge-off and sale of the remaining $83.7 million in July 2026. Two additional unrelated borrowers totaling $129 million entered nonaccrual in the second quarter. The company raised charge-off guidance to 65-80 basis points. [S1][S5][S22]

The strategic framework also changed. Management formalized three objectives—customer primacy, simplicity and efficiency, and top-tier performance—and raised the long-term return ambition from a sustainable 14% adjusted ROTCE toward a 14%-17% through-cycle range. The completed compensation cycle’s adjusted return was only 11.02%, so the new target represents a substantial step-up. It may reflect better margin, productivity, and lower intended excess capital; it has not been demonstrated through recessionary credit. [S3][S5][S9]

Leadership changed twice. Ignacio Alvarez retired as CEO in June 2025, Ferrer served from July 2025 through August 2026, and García became CEO September 1. Soriano moved from chief risk officer to CFO, while Sousa became chief risk officer. The appointments are internal and preserve institutional knowledge, but simultaneous movement across CEO, finance, and risk roles raises the need for stable controls and consistent disclosure. Ferrer stated that health and family motivated his retirement, and the 8-K said his board resignation did not result from disagreement. [S1][S7]

Capital policy became more aggressive. Popular completed a $500 million authorization, added a $1 billion authorization, and raised the quarterly dividend 20%. CET1 nonetheless increased to 16.08% in the second quarter because income exceeded capital use. Management also discussed potential risk-weighted-asset relief from evolving standardized capital rules; any benefit remains an estimate until rules are final and implemented. [S1][S5][S6][S8]

Markets and facilities evolved incrementally. Popular Bank continued its commercial-led strategy in New York and Florida, while Popular exited mainland mortgage origination and optimized Puerto Rico servicing. The company continued investing in branches, digital origination, cash management, cloud ERP, and security rather than undertaking a large branch expansion. Premises and equipment increased to about $732 million at June 2026. [S2][S3][S6]

No material accounting-policy change transformed the economics. The notable presentation change was the 2026 recharacterization of approximately $916 million of reciprocal deposits as nonbrokered, with the prior-year presentation conformed. CECL, AFS fair value through AOCI, and HTM amortized-cost accounting remained central. Investors should not misread the classification change as organic deposit improvement. [S1][S2]

The environment improved in level but not uniformly in direction. Puerto Rico unemployment remained historically low at 5.8%, but it rose year over year and payroll growth was nearly flat. Federal funds remain large, but the fiscal 2026 budget reported a $1.2 billion decline and created a spending holdback. These contradictions make it unsafe to attribute the full earnings improvement to a structurally stronger island economy. [S15][S16]

Verdict: The last two years produced materially better margins, earnings, tangible book, and capital distributions, but also repeated commercial-credit formation and unusually rapid senior-management turnover. Both external liquidity and internal execution improved results; neither should be treated as permanent. [S1][S7][S15][S22]

Risk Analysis

The most plausible ordinary downside is an earnings and multiple reset, not insolvency. If charge-offs remain around 90 basis points, margin falls 25 basis points, operating expenses rise 5%, and repurchases slow, annual EPS could fall near $11.50. Applying an 8.5-times multiple would place the equity near $98 before dividends. This scenario requires neither a bank run nor a deep recession; it requires current high returns to prove cyclical.

Commercial concentration is the leading incremental risk. The telecom exposure was large enough to add roughly 72 basis points to the quarterly annualized charge-off ratio. Two new borrowers entered nonaccrual while that credit was resolved. Investors cannot evaluate borrower collateral, industry, sponsors, or workout severity from public disclosure. Consequently, NPL formation and criticized-asset movement deserve more weight than management’s description of each name as discrete. [S1][S5]

Consumer exposure is diversified but macro-sensitive. Consumer allowance was about 5%, including higher coverage for cards, personal loans, and auto. Those reserves recognize frequency and unsecured risk. A recession, hurricane, energy outage, or government retrenchment could correlate unemployment, used-car values, unsecured credit, mortgages, and small-business performance. Historically low unemployment is a buffer, but the June rate had already risen year over year. [S1][S16]

Funding risk differs from the 2023 uninsured-deposit archetype. About one-third of deposits were collateralized public funds, and brokered deposits were comparatively small. Public balances can nevertheless move quickly for fiscal reasons, while collateral requirements encumber securities. Their departure would release collateral but might reduce earning assets or require higher-cost replacement funding. Aggregate deposit share does not eliminate timing, price, or collateral risk. [S1][S2]

Interest-rate risk is two-sided. Falling rates can lower public and market-sensitive deposit costs but reduce variable-loan yields and future securities reinvestment returns. Rising rates can raise asset yields over time while increasing deposit competition and AOCI losses. The current reinvestment tailwind reflects the maturation of low-yield assets. Extending duration to protect near-term income could recreate mark-to-market vulnerability.

Public-policy dependence is more nuanced than a simple reconstruction boom. Federal obligations support economic activity, but the certified budget reported falling federal funds and a precautionary holdback. A reduction in public deposits would not necessarily signal deposit-franchise weakness, yet simultaneous declines in public funds, contractor activity, loan demand, and fees would reveal greater macro dependence than the current valuation allows. [S15][S17]

Geographic concentration magnifies storm, grid, fiscal, political, and demographic shocks. Popular Bank only partially diversifies the group because BPPR remains the principal earnings engine. Federal aid can replenish deposits after a disaster, creating a later liquidity benefit, but that does not remove the initial operational disruption, credit loss, property damage, or customer hardship.

Operational risk is economically material. Popular’s payments, cards, digital channels, data, and branch network make it a large target for cybercrime and fraud. A major outage or conduct failure could produce restitution, regulatory restrictions, remediation expense, and customer attrition. No current filing establishes such a catastrophic event, but the combination of technology transformation and executive turnover raises execution sensitivity. [S2][S3][S7]

A catastrophic investment loss would require multiple defenses to fail together: a severe island or fiscal shock, meaningful deposit outflow, broad commercial and consumer defaults, operational disruption, insufficient insurance or federal response, and capital distributions that had already reduced the cushion. Another path would be a large cyber or conduct event that triggered restitution and regulatory constraints while damaging customer trust.

A literal total loss is remote. Common equity of $6.4 billion, a $785 million loan allowance, substantial collateral, deposit funding, and federal regulation stand ahead of failure. The plausible path would be cumulative credit and securities losses sufficient to exhaust common equity, followed by inability to issue capital or sell the franchise before receivership. A subjective five-year probability below 1% is reasonable but is an analyst estimate, not an observed frequency. Shareholders can still experience a 50%-70% drawdown long before total loss becomes probable.

Valuation magnifies manageable operating risks. At 1.86 times tangible book, evidence that normalized returns are 12%-13% rather than 15%-17% can compress the multiple even while book value remains positive. The 2023 price history shows how quickly bank risk premiums can change. High CET1 mitigates fundamental loss, not share-price volatility. [S1][S12]

Verdict: Solvency risk is low, while earnings and multiple risk are meaningful. The most actionable warning would be another quarter of more than $100 million of unrelated commercial NPL formation, especially if accompanied by rising nonpublic funding cost or weaker capital generation. [S1][S2][S22]

Valuation Discussion

At $163.71 and 63.87 million June shares, Popular’s equity value is approximately $10.46 billion. Enterprise value is not useful for a deposit-funded bank because deposits, borrowings, cash, and securities are operating inputs. Current decision-useful multiples are approximately 11.1 times trailing estimated EPS of $14.79, 1.86 times tangible book of $87.94, and 1.63 times common book of $100.38. The approved annualized dividend of $3.60 yields about 2.2%. [S1][S4][S12][S26]

The trailing EPS estimate replaces the first half of 2025 with reported first-half 2026 results. It is an analyst calculation, not company guidance. Annualizing second-quarter EPS would produce $17.40, but that embeds the quarter’s tax rate, provision, fee items, and balance seasonality. The lower trailing denominator is the more conservative starting point.

A residual-income identity helps frame expectations. With a 10.5% cost of common equity and 3% long-run tangible-book growth, a 1.86-times tangible-book multiple implies a sustainable return of roughly 17%. That is close to the high end of management’s 14%-17% objective. The comparison is directional because Popular’s reported 17.02% adjusted ROTCE adds back AOCI-related losses to its denominator, while the $87.94 tangible book used in the market multiple includes AOCI. A fully consistent model must normalize both return and book. [S1][S5]

The unadjusted tangible return of 20.12% shows why current earnings can support a premium today, but it is also flattered by a smaller AOCI-reduced tangible-equity denominator. The adjusted 17.02% removes much of that denominator effect. Neither number proves through-cycle economics. Investors must estimate future loss cost, deposit beta, expense intensity, capital needs, and the normalized securities yield.

Peer valuation is broadly consistent with the market assigning Popular a good—but not uniquely superior—franchise. At September 11 prices, First BanCorp traded around 2.18 times tangible book while reporting 19.49% ROE and 2.02% ROA. OFG traded near 1.69 times tangible book with 17.18% ROTCE and higher charge-offs. Bank of Hawaii traded near 1.99 times tangible book with lower current ROA but exceptionally low credit losses. Popular’s 1.86 times and 17.02% adjusted ROTCE sit within this island-bank range. [S12][S18][S19][S20]

The peer set is not interchangeable. First BanCorp and OFG compete directly in Puerto Rico but differ in funding mix, capital, consumer exposure, and loss rate. Bank of Hawaii shares geographic concentration and deposit-franchise characteristics but operates in a different economy. Large mainland regionals provide rate and capital comparisons but lack Popular’s local share. A mechanical median would conceal more than it reveals.

Popular’s own history is equally important. The shares traded near or below tangible book during the March 2023 panic and were repurchased around $96 in 2024. The current premium reflects real improvements in margin, adjusted returns, capital distributions, and tangible-book growth. Historical valuation percentiles must therefore be recalculated after the price and earnings gap; an inherited statement that BPOP remains cheap because it once traded near book is stale. [S3][S12][S23][S24]

Explicit scenarios clarify the distribution:

These are analyst estimates, not management forecasts. The draft’s probability arithmetic was incorrect. A 25% bear, 50% base, and 25% bull weighting yields approximately $167: $25 from the bear case, $88 from the base, and $54 from the bull. The result is only about 2% above the current price and demonstrates why modest base-case upside is insufficient by itself.

The market is right that Popular deserves more than a distressed regional-bank multiple. BPPR’s funding scale, capital, tangible-book growth, payments position, and per-share execution are real. The fragile expectation is that current adjusted returns can persist through higher credit formation, less public liquidity, and a new leadership regime.

Dividend-discount analysis alone understates value because the cash dividend is only about 24% of trailing EPS and repurchases are a major distribution channel. Treating every repurchase dollar as value creation would overstate value because buying above intrinsic value transfers wealth away from continuing holders. Residual income and tangible-book-plus-dividend growth provide the better framework.

Required return is a major sensitivity. At 15% sustainable return and 3% growth, a 9.5% cost of equity supports a multiple near 1.85 times, while an 11% cost supports only about 1.50 times. On $87.94 of tangible book, that difference is about $31 per share. The factor model’s market, small-size, financial-sector, and credit-risk exposures are consistent with a risk premium that can change abruptly, but the statistical coefficients do not establish business causality. [S21]

Verdict: Current valuation is defensible only if Popular sustains roughly mid- to high-teens tangible returns with controlled credit and disciplined capital return. The corrected probability weighting shows little expected-value discount, while inconsistent ROTCE and tangible-book denominators must not be combined without adjustment. [S1][S12]

Variant Perception

The observable debate is no longer whether Popular is a troubled Puerto Rico bank. The rerating, premium-to-book valuation, and earnings-call questions indicate that investors recognize the deposit franchise, margin recovery, and excess capital. Thoughtful questions focused on whether ROTCE was peaking, how low CET1 could move, public-deposit seasonality, margin stability, whether commercial problems were isolated, expense growth, mainland competition, M&A, deferred-tax assets, and repurchase pace. [S5][S6]

The strongest bull case is structural. Popular can reinvest low-yield securities, retain low-beta nonpublic deposits, grow loans modestly, deepen card and cash-management relationships, convert transformation efforts into net savings, and retire 5%-7% of shares annually while CET1 moves toward a more efficient target. Under this view, a 16%-17% adjusted ROTCE is the new normal rather than a cyclical peak. Puerto Rico reconstruction provides a multi-year activity runway, and Popular Bank supplies selective growth optionality.

Evidence against the strongest bull formulation is already visible. Federal funding is declining, nonpublic deposits fell sequentially at quarter-end, public deposits are costly and volatile, and commercial-credit formation has repeated. Popular’s adjusted ROTCE has not been tested through a recession or a period of materially weaker government liquidity. [S1][S15][S22]

The strongest bear case is not bank failure. It is that the market has capitalized a favorable conjunction of public liquidity, historically low unemployment, securities reinvestment, lower deposit cost, tax-exempt income, and aggressive repurchases. Commercial problems are beginning to season while a new management team changes capital policy. Under this view, normalized adjusted ROTCE is 11%-13%, and a 1.2-1.5-times tangible-book multiple is more appropriate.

Evidence against the strongest bear formulation is Popular’s actual funding and capital performance. Core deposits cost only 1.10%, CET1 was 16.08%, tangible book grew 16.6% year over year, and margin held at 3.66% despite rate volatility. The telecom asset was sold, consumer credit improved, and First BanCorp and OFG demonstrate that high-teens returns are not implausible in Puerto Rico. [S1][S18][S19]

The differentiated view is narrower: Popular’s franchise is better than a generic regional bank, but the stock already reflects much of that superiority. The market may be granting through-cycle status to a return objective announced during favorable conditions before enough credit, federal-funding, and leadership evidence exists. The variant rests on a higher proof standard, not a prediction of imminent deterioration.

Five load-bearing assumptions define the debate:

1. Nonpublic deposits remain low beta. Current evidence is the 1.10% cost and 2.7% year-over-year ending-balance growth. Falsifier: average balances fall more than 3% year over year while cost rises more than 20 basis points without comparable asset-yield improvement. [S1]

2. Commercial problems remain discrete. Current evidence is the 0.33% ex-telecom quarterly charge-off rate. Falsifier: unrelated commercial NPL formation exceeds $200 million cumulatively over two quarters or criticized assets rise materially. [S1][S5]

3. Securities reinvestment offsets rate pressure. Current evidence is the $1.1 billion of purchases near 3.9% and higher securities yield. Falsifier: reported NIM falls below 3.45% without a deliberate liquidity build or faster balance-sheet de-risking. [S1][S5]

4. Capital return remains value-creating. Current evidence is the net 20% share reduction and 16.08% CET1. Falsifier: aggressive repurchases above two times tangible book precede declining tangible book per share or a forced distribution pause. [S1][S8]

5. Leadership continuity survives the transition. Current evidence is the promotion of experienced internal executives. Falsifier: further senior-risk departures, material strategy reversal, control deficiencies, or repeated guidance errors before September 2027. [S7]

The factor model is a positioning diagnostic, not a business classification. As of September 10, it estimated market exposure of 0.98, growth exposure of -0.66, small-size exposure of 0.44, financial-sector return exposure of 0.37, and credit-risk exposure of 0.35. Residual momentum was 0.036, residual Sharpe 0.073, and residual volatility 0.199. R-squared was 0.466, so more than half of return variation remained unexplained by the included factors. Sector coefficients are statistical exposures, not legal or causal facts. [S21]

The model explains why company fundamentals do not control every price move. Bank, credit, size, value-growth, and broad-market rotations can move BPOP independently. The small negative modeled quality exposure does not contradict reported profitability; it reflects return covariance under the model’s definitions.

Relevant prior research principles were revalidated selectively. The rule that common returns deduct noncommon claims applies, but Popular’s noncommon distributions are too small to alter the thesis materially. The rule requiring valuation to be recomputed after a large price gap is directly relevant and overturns any stale near-book cheapness claim. Retrieved biotechnology, AI-adoption, marketplace, and manufacturing learnings lack causal relevance to a deposit-funded bank and were rejected rather than transferred.

Verdict: Consensus correctly recognizes franchise improvement and capital capacity but may be early in treating current returns as through-cycle. The decisive variant is the evidence threshold applied to credit, public liquidity, and the return denominator—not a disagreement over whether BPPR is the market leader. [S1][S5][S13]

Fact vs. Interpretation

The distinction matters because management commentary is not independent confirmation. Credit isolation, net transformation savings, capital-rule benefits, and through-cycle returns remain testable hypotheses. [S1][S2][S5]

Open Questions

• What were criticized and classified commercial balances at September 30, excluding the resolved telecom exposure, and how did internal risk grades migrate? [S1][S2]

• What collateral, sponsor support, industry, and expected-loss severity apply to the two new nonaccruals totaling approximately $129 million? [S1][S5]

• How much of the $22.7 billion public-deposit balance is operationally stable, what is its expected quarterly range, and how much collateral is encumbered against it? [S1][S2]

• Can management reconcile the approximately $20 million of sustainable savings reported for 2025 with the broader approximately $50 million of actions embedded in the 2026 expense outlook? [S5][S9]

• What CET1 operating target will García use after final capital rules, stressed losses, buybacks, and potential noncommon capital issuance? [S5][S7]

• What valuation discipline governs the pace of the new $1 billion authorization? [S8]

• What board succession and retention measures address the second CEO change in fourteen months and simultaneous CFO/CRO changes? [S7]

• What proportion of Popular Bank growth is funded by relationship deposits, and what are fully allocated returns by origination vintage? [S3][S6]

• How much reserved mainland deferred-tax value can be realized before relevant losses begin expiring in 2028? [S2][S5]

• Can Popular disclose account-level retention, primary-account status, product depth, and funding cost by household, business, and public customer? [S1][S14]

What Must Be True

Bull tests

• Returns: trailing-four-quarter adjusted ROTCE remains at least 15% through 2027 without excluding recurring credit or transformation costs. Monitor GAAP common income, average tangible equity, AOCI adjustments, and every non-GAAP reconciliation. The starting reported measure is 17.02%, but its denominator adjustment must remain transparent. [S1]

• Credit: charge-offs excluding the sold telecom exposure remain below 50 basis points, unrelated commercial NPL formation stays below $100 million per quarter, and criticized assets do not rise materially. Monitor the NPL roll-forward, provision, loan-level reserve changes, and recoveries. [S1][S2][S22]

• Funding: average nonpublic deposits remain stable or grow, while their cost stays below roughly 1.35% absent higher benchmark rates. Monitor demand balances, nonbrokered savings, time deposits, brokered funding, public balances, and collateral. [S1][S2]

• Margin: securities reinvestment and loan pricing keep reported NIM above 3.50% through 2027. Monitor portfolio purchase yield, duration, public and nonpublic deposit beta, loan spreads, and the difference between reported and tax-equivalent margin. [S1][S5]

• Capital allocation: net shares decline at least 4% annually without CET1 falling below a prudent stressed buffer or tangible book per share declining because repurchases were too expensive. Monitor average purchase price, CET1, stress assumptions, and tangible-book-plus-dividend growth. [S1][S8]

• Execution: revenue grows faster than normalized expense across four quarters, and management reconciles gross initiatives to net savings. Monitor personnel, technology, promotion, efficiency, and customer-service metrics. [S1][S5][S9]

Bear tests

• Credit broadening: two consecutive quarters each produce more than $100 million of unrelated commercial NPL formation, or NPLs exceed 1.5% of loans. That would materially falsify the isolated-credit narrative. [S1][S22]

• Deposit erosion: average nonpublic deposits fall more than 3% year over year while their cost rises. That would indicate that public liquidity was masking weakness in the customer franchise. [S1][S2]

• Return reversion: trailing adjusted ROTCE falls below 12.5% without a severe recession. That would indicate the 14%-17% objective relied on temporary spread, tax, or capital effects. [S1][S5]

• Capital misallocation: Popular repurchases aggressively above two times tangible book and then pauses distributions because credit or regulation consumes capital. That would demonstrate procyclical allocation. [S1][S8]

• Leadership disruption: further senior-risk departures, a material control deficiency, or a major strategic reversal occurs before September 2027. That would outweigh the continuity benefit of internal promotion. [S7]

• Macro dependence: public deposits fall below $18 billion while loan demand, fees, and contractor credit quality weaken together. That would show that government and federal flows carried more of the earnings base than assumed. The risk is credible because the fiscal 2026 budget already reported a $1.2 billion decline in federal funding and Puerto Rico unemployment rose year over year. [S15][S16]

The thesis does not require every quarterly metric to be perfect. It requires evidence that Popular’s nonpublic deposit advantage converts into durable credit-adjusted returns after public liquidity, securities repricing, capital optimization, and large commercial exposures are normalized. Conversely, the bear case does not require failure of the bank; it requires proof that investors paid a high-teens-return multiple for a franchise that earns low-teens returns through the cycle. [S1][S2][S5][S15][S16]

Public source appendix

• S1: Popular — Q2 2026 financial results and supplement — primary company financial release; published 2026-07-23; Pages 1-7 and Tables A-R: earnings, margins, deposits, loans, credit, capital, guidance, and non-GAAP reconciliations
• S2: Popular — Q2 2026 Form 10-Q — SEC filing; published 2026-08-07; Financial statements, MD&A, securities, deposits, CRE concentrations, commitments, government exposure, taxes, and credit notes
• S3: Popular — 2025 Annual Report — SEC-filed annual report; published 2026-03-11; Pages 4-12 and Items 1, 1A, 7, and 8: business, strategy, ten-year financial history, capital, credit, branches, and financial statements
• S4: Company Financials — BPOP multi-period statements, profile, ratios, and valuation — configured financial-data cross-check; publication date unavailable; Exchange-qualified NASDAQ:BPOP; annual 2021-2025 and trailing June 2026 statements, per-share data, profitability, cash flow, and valuation; reconciled to filings
• S5: Company Financials — Q2 2026 earnings-call transcript and Popular webcast — management transcript and webcast; published 2026-07-23; Prepared remarks and Q&A: ROTCE objective, credit, public deposits, NII, expenses, competition, deferred taxes, M&A, and repurchases
• S6: Company Financials — Q1 2026 earnings-call transcript and Popular webcast — management transcript and webcast; published 2026-04-23; Prepared remarks and Q&A: capital, deposits, margin assumptions, transformation, competition, mainland strategy, and M&A
• S7: Popular — leadership-transition Form 8-K — SEC filing; published 2026-07-23; Item 5.02: Ferrer retirement, compensation and consulting; García, Soriano, and Sousa appointments
• S8: Popular — capital actions Form 8-K — SEC filing; published 2026-07-23; New $1 billion repurchase authorization, completion of prior authorization, and planned dividend increase
• S9: Popular — 2026 definitive proxy statement — SEC proxy filing; published 2026-03-26; Executive and director compensation, incentive metrics, performance-share results, ownership, related-party transactions, and governance policies
• S10: Javier Ferrer — August 2026 Form 4 — SEC insider filing; published 2026-08-12; Table I: August 10 sale of 35,000 shares at a weighted-average $175.433 and post-transaction holdings
• S11: SEC — Popular filing history — primary filing index; publication date unavailable; Trailing 60 months of Forms 10-K, 10-Q, 8-K, DEF 14A, 3, 4, and 5 reviewed for material financial, governance, and insider events
• S12: Company Financials — BPOP and peer price history — market-data cross-check; published 2026-09-11; Exchange-qualified daily OHLC data through September 11, 2026; current prices cross-checked to company disclosures
• S13: Puerto Rico deposit market structure — authoritative-data derivative; published 2025-06-30; FDIC Summary of Deposits-derived June 2025 institution deposits, office-assigned shares, branch counts, and HHI
• S14: Federal Reserve — BPPR 2025 CRA performance evaluation — regulatory evaluation; published 2025-06-23; Pages 3-7: institution description, distribution, market leadership, competitors, products, and accessibility
• S15: Puerto Rico Oversight Board — fiscal 2026 certified budget — government oversight release; published 2025-06-27; Budget totals, $14.2 billion federal component, $1.2 billion federal-funding decline, uncertainty, and 5% spending holdback
• S16: BLS — June 2026 state and territory employment — government statistics; published 2026-07-21; Tables 1 and 3: Puerto Rico unemployment, labor force, and nonfarm payroll employment
• S17: FEMA — Puerto Rico from recovery to resilience — government report; published 2024-09-01; Cumulative disaster obligations, reconstruction programs, and implementation context
• S18: First BanCorp — Q2 2026 results — peer primary release; published 2026-07-22; Net income, EPS, NIM, ROA, ROE, tangible book, deposits, credit, and CET1
• S19: OFG Bancorp — Q2 2026 results — peer primary release; published 2026-07-21; Q2 tangible book, ROA, ROTCE, margin, allowance, nonperformers, and charge-offs
• S20: Bank of Hawaii — Q2 2026 results — peer primary release; published 2026-07-27; Q2 earnings, ROA, common-equity return, NIM, tangible book, deposits, loans, and asset quality
• S21: The factor model — BPOP diagnostic — internal quantitative diagnostic; published 2026-09-10; September 10, 2026 exposures, residual signals, alpha, R-squared, and adjusted R-squared
• S22: Popular — Q3 2025 financial results — SEC-filed earnings release; published 2025-10-23; Net income, margin, provision, NPL increase, two large commercial exposures, charge-offs, capital, and repurchases
• S23: Popular — 2024 Annual Report — SEC-filed annual report; published 2025-03-31; MD&A, financial statements, deposits, capital, repurchases, AOCI, credit, and risk factors
• S24: Popular — 2023 Annual Report — SEC-filed annual report; published 2024-03-28; Shareholder letter, MD&A, financial statements, 2023 banking stress, funding, securities, and price context
• S25: Popular — 2021 Form 10-K — SEC filing; published 2022-03-01; Five-year baseline for business, earnings, balance sheet, cash flow, shares, and capital
• S26: Popular — approval of $0.90 quarterly dividend — SEC-filed company release; published 2026-08-20; Board approval, record date, and October 1, 2026 payment date