OZK — Bank OZK

Bank OZK has spent four decades becoming the dominant construction lender in America, generating a net interest margin of 4.56% — nearly a full percentage point above the regional bank average — through a model built on real estate professionals, conservative underwriting, and the discipline to walk away when terms are unfavorable. A charge-off spike concentrated in a handful of identifiable office and life sciences credits compressed 2025 earnings to $6.18 per share while leaving tangible book value still growing, the core franchise structurally intact, and the stock at $47 — roughly seven and a half times earnings and one times tangible book value. At that price, the market is collecting admission to watch a cyclical setback while forgetting to charge for the franchise that produced it.

---

The commercial real estate industry has spent the past three years working through a cycle that most participants expected to be over by now. Office occupancy in gateway markets has stabilized at between 45% and 65% of pre-pandemic levels — down from expectations of a faster return to normalcy — and life sciences real estate, which surged during the COVID-era biotech boom, has absorbed the dual punishment of rising interest rates and a pullback in venture funding to early-stage companies. These two sub-sectors have inflicted real losses on the banks that financed their development. Regional banks with concentrated commercial real estate exposure have become a source of investor anxiety since the 2023 banking stress, and a vocabulary of alarm — "unrealized losses," "concentration risk," "duration mismatch" — has been applied to any bank that looks different from a conventional deposit gatherer.

Bank OZK looks different from a conventional deposit gatherer. Its construction loan portfolio is approximately 60% of total assets. It originates billions of dollars of commercial real estate loans each year. Its net interest margin exceeds almost every regional bank peer. Its efficiency ratio of 33% is barely half the industry average. When the 2025 charge-off results arrived, the stock sold off as if the difference between OZK and an undisciplined CRE lender had been erased. It has not been.

Commercial construction lending is a specialty within banking that most large institutions handle poorly, and that most have exited at various points in the past decade. The work requires genuine real estate expertise: the ability to evaluate a developer's track record, underwrite the cost and demand assumptions underlying a construction project, assess collateral value through a development cycle spanning two to five years, and manage a complex process through cost overruns, lease-up timelines, and interest rate changes that can alter the economics of the finished product. The typical credit officer at a large institution is trained to evaluate stabilized income-producing properties. Construction lending before stabilization is a different discipline — the cash flows do not yet exist, the collateral value is largely theoretical until the project completes, and the underwriter is betting on cost, timing, and demand in markets where any of those variables can shift materially.

The U.S. commercial real estate mortgage market processed approximately $499 billion in borrowing and lending activity in 2025, a 22% increase from 2024. Construction lending is a subcategory of that market, and a subcategory that most banks have repeatedly exited during periods of credit stress. The void has been filled partly by private credit firms — Blackstone, Ares, and others — that have captured an estimated 15-20% of institutional CRE lending in select markets. But private credit lacks the regulatory clarity, the deposit-funded cost of capital, and the execution velocity of a bank operating at scale. Bank OZK has operated in this niche continuously, across the 2008 financial crisis, the post-pandemic rate cycle, and the current commercial real estate stress, because it is institutionally staffed to do so.

Bank OZK is an Arkansas-headquartered bank with $40.8 billion in total assets, operating across nine states through more than 250 branches. Its revenue model is the traditional bank model — the spread between what it earns on loans and what it pays on deposits generates approximately 93% of revenues as net interest income. What makes it structurally different is the origin of that spread.

The bank's Real Estate Specialties Group — RESG — is staffed not by conventional bankers but by real estate professionals who specialize in construction financing across 60 metropolitan markets. RESG originated $13.8 billion in construction loans in 2022, making it the largest construction lender in the United States by a wide margin — roughly double the combined originations of JPMorgan Chase and Wells Fargo in the same period. The group's competitive strategy is consistent and explicit: target average loan-to-cost ratios below 50% against industry norms of 70-80%, work only with developers who have proven track records and tested experience, and withdraw from markets where pricing does not reflect genuine risk-adjusted economics. When competitors retrenched during 2008, OZK stayed in the market on its own terms. When the CRE market was most active in 2021 and 2022, OZK's originations peaked — not because the bank was chasing volume but because favorable terms were briefly available at scale.

The competitive advantage here is structural and specific. A 45% average loan-to-cost ratio means the borrower's equity absorbs the first 55% of any loss before OZK suffers a dollar of principal impairment. No other major construction lender maintains this level of conservatism consistently across the cycle. The operational advantage is equally concrete. An efficiency ratio of 33% means OZK spends 33 cents of every revenue dollar on operations; the average regional bank spends more than 55 cents. This gap is not an artifact of Arkansas real estate costs — OZK originates loans across California, New York, Texas, and Florida, the most expensive real estate markets in the country. The efficiency gap reflects a focused product set, a lean organizational structure, and decades of discipline at the direction of a CEO who has been in the seat since 1979.

The net interest margin superiority follows directly from the loan mix. Construction loans are floating-rate, priced at a premium to conventional commercial real estate loans, and the relationship-driven origination process allows OZK to set terms that reflect the genuine risk-adjusted economics of the lending. In 2024, that NIM was 4.56% — down from 5.16% in 2023 as deposit funding costs rose through the rate cycle, but well above large bank averages of 2.5-3.5% and most community bank ranges of 3.5-4.5%. The compression is a cyclical effect; the structural margin superiority is not.

Metric	Bank OZK	Regional Bank Avg	Large Bank Avg
Net Interest Margin (2024)	4.56%	~3.8%	~2.9%
Efficiency Ratio (2024)	33.3%	~57%	~62%
Return on Assets (2024)	1.92%	~1.1%	~0.9%
Net Charge-Off Rate (2024)	0.20%	~0.45%	~0.50%

OZK reported GAAP net income of $700.3 million in 2024 — a record — earning $6.14 per diluted share. Net interest income reached $1.55 billion. The efficiency ratio of 33.25% reflected the benefit of RESG loan yields repricing upward through the rate cycle on floating-rate structures. In 2025, total net interest income reached a new record of $1.59 billion, but a credit event in the second half consumed most of the earnings momentum. Full-year net charge-offs in 2025 totaled $160 million — up from $57 million in 2024. The fourth quarter alone produced $98.3 million in charge-offs, compared to $12.4 million in Q4 2024. The result was 2025 net income of $699.3 million and diluted EPS of $6.18 — essentially flat to 2024, because the credit losses absorbed what would otherwise have been a meaningful earnings increase on a record interest income base.

The charge-offs were concentrated in identifiable problem credits. A Boston office redevelopment project produced a $72.4 million charge. A Chicago life sciences development contributed another $9 million. Further losses appeared in Santa Monica creative office space, Baltimore, and Seattle. These are specific sub-sectors — urban office and early-stage life sciences real estate — where fundamental demand destruction, not ordinary credit cycle stress, has occurred. Office demand in gateway cities has not recovered toward pre-pandemic levels; life sciences real estate built on 2021 venture capital enthusiasm has encountered a more sober capital environment. The bank's allowance for credit losses stands above $600 million against a $33 billion loan portfolio — a reserve ratio of approximately 1.8%. The capital position remains strong: tangible common equity-to-tangible assets was 12.52% at year-end 2024. Tangible book value per share has grown from approximately $27 at the end of 2021 to $41.48 at year-end 2024 and approximately $45 at year-end 2025, compounding at roughly 13% annually despite the elevated charge-offs of the credit cycle.

The bank does not face the structural vulnerability that made Silicon Valley Bank fragile. OZK's loan portfolio is predominantly floating-rate — it benefits when rates are elevated rather than suffering duration losses — and its funding base is retail and institutional deposits that have repriced gradually through the cycle. The short thesis comparing OZK to SVB confuses concentration in an asset class with structural balance sheet risk. These are different problems. OZK has only the first of them, to a degree that is monitored and deliberate.

George Gleason has served as Chairman and CEO since 1979 — an unbroken 47-year tenure spanning multiple recessions, multiple credit cycles, and the transformation of the bank from a small Arkansas community lender to the dominant U.S. construction lending franchise. He owns approximately 12-16% of the company, a stake worth $600 million or more at current prices. His 2024 total compensation was $7.33 million — 80% of it in performance-based equity — slightly below market average for comparable-sized institutions. He has not sold meaningful shares in recent years. The CFO, Tim Hicks, joined in 2009 and rose through internal ranks. The President, Brannon Hamblen, came from RESG leadership. This is a management team built on the franchise, not imported to fix it.

The capital allocation record reflects long-term shareholder orientation. OZK has increased its dividend for 29 consecutive years — through multiple recessions, the financial crisis, and the pandemic. The current dividend of $0.43 per quarter represents approximately 3.6% yield and a payout ratio of roughly 28% of earnings — conservative enough to retain capital for growth while providing a tangible, growing cash return. Management authorized $250 million in share repurchases in 2024, executing buybacks at an average price of $44.45 — below tangible book value, making each repurchased share immediately accretive to the per-share asset base of remaining shareholders. A management team that buys below book while raising dividends, builds reserves while maintaining capital ratios above 12%, and does all of this without a single year of dividend reduction in three decades is not engaged in capital misallocation.

The five-year operating table shows the business across a full cycle:

Year	RESG Originations	NCO Rate	NIM	Efficiency Ratio	TBV/Share
2021	~$5B	~0.05%	—	~31%	~$27
2022	$13.8B	~0.10%	—	35.5%	~$32
2023	$7.22B	0.13%	5.16%	~38%	~$37
2024	$5.41B	0.20%	4.56%	33.25%	$41.48
2025	~$5.0B	0.49%	~4.2%	~33%	~$45

RESG originations peaked at $13.8 billion in 2022 — a year when construction financing demand was elevated and OZK could set favorable terms at volume. They fell by nearly half to $7.22 billion in 2023 and further to $5.41 billion in 2024, approaching the pre-pandemic origination level of $5.67 billion in 2019. This is not market share loss. It is discipline. When the CRE market tightened and developers could not underwrite favorable economics, OZK reduced volume. Total RESG commitments outstanding peaked at $34.5 billion in March 2024 and have been declining deliberately — management capped individual new loan sizes at $500 million to limit concentration risk and has publicly stated the goal of reducing RESG from 64% to 50% of the total portfolio.

The charge-off column is the evidence test for the underwriting discipline the business claims. Through 2021 and 2022 — years of extraordinary lending volume and a rapidly rising rate environment — net charge-offs were effectively zero. In 2023, as the CRE market began to stress, they remained a negligible 0.13% of average loans. The 2025 spike to 0.49% is real — but the Boston office credit alone accounted for nearly half of the Q4 charge: a single, identifiable urban office development in one of the most supply-damaged markets in the country. The bank's RESG portfolio contains approximately 292 construction credits across 60 metropolitan areas, the majority in multifamily housing, hospitality, and industrial properties rather than office. A portfolio where one troubled office credit in Boston represents nearly half of a full quarter's charge-off spike is, by construction, not a portfolio with systemic underwriting failure.

Tangible book value per share grew from approximately $27 to approximately $45 across the four-year period — compounding at roughly 13.5% annually — even as the charge-off cycle produced losses that would have severely impaired a more conventionally underwritten portfolio. The efficiency ratio held consistently below the industry average throughout, providing the operating leverage that makes OZK resilient to credit losses that would threaten a high-cost operator. The combination of operating efficiency and conservative loan-to-cost ratios is the structural equation: when the cycle turns and loans go bad, OZK has more collateral buffer to absorb losses and more operating margin to absorb the provisioning impact on earnings.

The next phase of growth comes from two sources. First, RESG stabilization: management expects the portfolio to stop declining as existing construction loans repay on completed projects and new originations at favorable economics replace them. The RESG pipeline remains active; the bank has simply capped loan sizes and tightened terms in the current environment. Second, the Corporate and Institutional Banking division: OZK's CIB business has grown from 18 employees to 97 over the past two years, expanding into asset-based lending, middle-market corporate lending, and specialty finance. Management expects CIB to reach a portfolio size comparable to RESG by 2027. This part of the thesis requires monitoring rather than certainty — CIB is building, not built. But the team size and capital commitment suggest the bank is serious about the pivot, not testing it tentatively. OZK has captured approximately 1% of the $499 billion annual U.S. commercial real estate lending market in origination flow — and essentially zero of the institutional lending market it is now pursuing through CIB. The division, with 97 professionals and infrastructure in place, is at approximately the stage RESG occupied two decades ago.

The stock trades at $47.48 against year-end 2025 tangible book value of approximately $45 — a price-to-tangible-book ratio of roughly 1.05. It trades at 7.7 times 2025 GAAP earnings of $6.18 per share. Regional bank peers trade at 15-16 times earnings; the U.S. banking industry average is approximately 11 times. The market is pricing OZK as a cyclically impaired, CRE-concentrated regional bank with unclear forward earnings. The numbers describe a bank with a 33% efficiency ratio, a 4.2%+ net interest margin, a 47-year owner-operator holding 12-16% of the equity, 29 consecutive years of dividend increases, and credit losses that — when examined specifically — are concentrated in a handful of identifiable sub-sector credits that are already provisioned.

Adjusting 2025 earnings for the excess charge-offs relative to a normalized credit cycle — the $160 million in 2025 charge-offs versus $57 million in 2024, with the $103 million difference representing the concentrated office and life sciences impairments — produces normalized after-tax earnings of approximately $6.90 per share and normalized pre-tax earnings of approximately $8.84 per share. At $47.48, the stock trades at approximately 5.4 times normalized pre-tax earnings. The regional bank peer average implies an earnings yield approaching 7%; OZK's normalized pre-tax earnings yield is approximately 19%. That gap represents the entirety of the 2025 discount applied permanently to a franchise that is not permanently impaired.

The bear case deserves a direct answer. The most credible version argues that the Q4 2025 charge-offs were the leading edge of a longer credit deterioration cycle — that additional problem credits exist in the RESG portfolio in urban office and life sciences markets, that 2026 will produce another round of elevated losses, and that CIB will not scale fast enough to replace RESG origination volume. In this scenario, normalized earnings stay depressed, tangible book value growth slows from 13% toward 6-8% annually, and the stock drifts toward $38-40 where it is priced at 5-6 times depressed earnings. This is the scenario to monitor: not a systemic failure of the RESG underwriting model, but a longer tail on the specific sub-sector impairments already identified. Management has characterized the remaining problem credits as identifiable and limited; the allowance of $600 million-plus provides meaningful coverage. But those representations deserve verification over the next several quarters before the credit cycle question can be closed with confidence.

The answer to the bear case is the 45% loan-to-cost buffer. The $72 million Boston office loss was a specific situation — a vacant urban redevelopment in one of the most supply-damaged office markets in the country. Even with a 45% LTC ratio, if a property suffers catastrophic demand destruction rather than ordinary cyclical softness, losses occur. But the scale of those losses — concentrated in office and life sciences, contained within a $33 billion portfolio secured at 45% LTC on average — confirms the thesis rather than refuting it. A bank that suffers 0.49% in charge-offs in the worst office and life sciences cycle in a generation, from a portfolio underwritten at half the leverage of the industry, is demonstrating the model working at the margin of failure rather than failing.

What would change this conclusion? If 2026 produces charge-offs above $100 million, suggesting the problem extends beyond the identified office and life sciences credits and into the broader RESG portfolio, the thesis takes longer to play out. If CIB fails to show meaningful portfolio growth by 2027, the RESG-replacement growth story weakens. At a stock price above $65, the multiple moves closer to fair value and the margin of safety narrows; above $90, the business is being valued as an exceptional franchise rather than a cyclically impaired one, and patience becomes the appropriate posture rather than urgency.

George Gleason has compounded Bank OZK's tangible book value at double-digit rates for four decades, through recessions, rate spikes, and credit cycles more severe than the current one. At $47, the market is handing the investor a dollar for sixty-six cents — and declining to charge anything for the 47-year record, the 33% efficiency ratio, or the underwriting discipline that, in the worst office real estate environment in a generation, produced losses of less than one half of one percent of the loan portfolio. The charge-offs were real. The franchise is more real.