International Bancshares is a Laredo-anchored multi-bank holding company whose public identity has long been fortress capital and a cheap Texas-Oklahoma deposit franchise rather than loan growth. That identity now sits under a credit test. In the June quarter the company placed a relationship of several loans secured by interests in affordable housing projects onto nonaccrual, and the provision for credit losses more than doubled versus the year-ago quarter. Spread income still expanded because management redistributed rates paid on deposits. The investment debate is whether the housing relationship is a contained work-out or the first visible crack in a credit book the market has treated as pristine.
First-half net income was essentially unchanged at $198 million. That flat print hides a clean split between spread income and credit cost. Net interest income rose as deposit costs fell. Credit-loss expense absorbed almost the entire spread gain. Nonaccrual loans jumped to $297 million. The year-end stock of those loans was $140 million. Management states that no actual losses on those loans have been realized and that such losses may never occur. The market has to decide whether to believe the work-out language or the classification.
The deposit franchise is still the economic engine. Noninterest-bearing demand still accounts for more than a third of total deposits, and the loan-to-deposit ratio remains in the mid seventies, leaving room to fund loan growth without wholesale dependence. Dennis Nixon, the long-tenured president, frames the second half around the same asset-liability discipline plus an efficiency push that names artificial intelligence as a process tool. The counterargument is that a bank this profitable does not double its provision and nearly double nonaccruals unless something in underwriting or collateral valuation has shifted. Shareholders are not being asked to underwrite a growth story. They are being asked to underwrite the work-out.
Three variables resolve the case. The first is the Affordable Housing Work-Out: whether the classified relationship produces realized charge-offs or is resolved through collateral and specific reserves. The second is the Deposit Cost Floor: whether the rate redistribution that lifted net interest income holds as the bank competes for balances. The third is Provision Intensity: whether credit cost stays elevated as the loan book grows. At a mid-September close near $71 and a trailing multiple near 11 times earnings, the market already prices a high-quality compounder with a credit discount. The open question is which side of that discount is wrong. A contained work-out leaves the compounder multiple intact. A second classified relationship turns the discount into the right price.