The Federal Deposit Insurance Corp.'s Quarterly Banking Profile, released Tuesday, is a document of genuine strength. Banks earned $90.1 billion in the second quarter of 2026, up 12 percent from the first quarter and 28.7 percent from a year earlier. Return on assets reached 1.37 percent. Loan growth was broad, asset quality improved, and 69 percent of institutions reported higher net income. FDIC Chair Travis Hill put it in one sentence: "We had solid earnings this quarter once again."
The same document contains three quiet lines that complicate the celebration, and together they describe a system whose cushions are improving in dollars while shrinking in relevance. All of the quarter's deposit growth, and then some, was money outside the deposit insurance cap. The industry's celebrated loan growth is the reason its capital ratios fell. And the $326.7 billion in unrealized securities losses still sitting on bank balance sheets did not disappear; they went quiet. The strong quarter and the quiet vulnerability are in the same report, sometimes in the same paragraph.
All the quarter's deposit growth was money outside the cap
The deposit line reads well at first pass: domestic deposits rose 0.8 percent, or $142.7 billion, the eighth consecutive quarterly increase. The composition of the growth is where the story is. The FDIC attributes the increase to estimated uninsured domestic deposits, which rose $317.4 billion over the quarter. The arithmetic is not complicated: if total deposits rose by $142.7 billion while uninsured deposits rose by $317.4 billion, then insured deposits actually fell, by something like $175 billion. Every dollar of the quarter's celebrated deposit growth, and then some, was money that sits above the $250,000 insurance limit.
This is not a scandal. Deposit flows move on rates, not on insurance status, and uninsured balances have been a growing share of the funding base for years, concentrated in the largest banks and in operating accounts too big for the cap. But the line deserves to be read for what it says about the system's structure. The safety net that defines the industry's stability guarantee is covering a shrinking share of the industry's funding. The quarter in which deposits grew for the eighth straight time was also a quarter in which the insured portion of the deposit base shrank. The system is not uninsured in any crisis sense of the word. It is, however, growing at its uninsured edge, and the edge is where the 2023 runs began.
The celebrated strength is why the cushions got thinner
The same inversion runs through the capital lines. Loan growth is the headline strength of the quarter, up 1.8 percent from the first quarter and 6.8 percent from a year earlier, and Hill framed the lending rebound as a sign of the economy's strength. But capital ratios moved the other way. The Tier 1 risk-based capital ratio fell 17 basis points to 13.75 percent, and the leverage ratio fell by roughly the same amount to 8.98 percent, because assets grew faster than the unborrowed equity that stands behind them. Regulators lowered the enhanced supplementary leverage ratio for the largest banks last fall, and the quarter shows the direction of travel the industry is on without it.
The mechanics are worth stating plainly, because they are not a contradiction. A bank that grows its loan book is growing the assets against which it must hold capital. When lending accelerates faster than earnings can accumulate, the ratios decline even as the bank gets, in every visible sense, stronger. The quarter's strength and the quarter's thinning cushions are the same phenomenon. The industry did not choose between growing and retaining capital. It grew, and the growth was faster than the retention, and the report records both in adjacent lines.
None of this is a warning of imminent distress. Ratios in the mid-teens are comfortable by historical standards, and the FDIC's own measures of problem banks sit at 47 institutions, 1.1 percent of the industry, inside what the agency calls the normal range. The point is that the buffers are directionally sensitive to the good news. Every strong lending quarter spent without commensurate equity accumulation narrows the margin that a bad quarter would consume, and the report's own numbers show exactly that narrowing happening during the best quarter of the cycle.
The unrealized losses did not leave; they went quiet
The third quiet line is the one the market has learned to skip. Unrealized losses on securities portfolios stood at $326.7 billion at the end of the quarter, up slightly from the first quarter and down 17.4 percent from a year earlier. The losses are the scar tissue of the rate shock that followed 2022: bonds bought at low yields, marked down as rates rose, and held on balance sheets where accounting rules let some of the markdowns sit unrealized indefinitely. The improvement over the year is real. The stock is still enormous.
The earnings of the quarter coexist with that stock. A bank's income statement does not flow through its unrealized securities losses, which is why $90.1 billion of net income and $326.7 billion of embedded losses can appear in the same report without either contradicting the other. The distinction between realized and unrealized is an accounting convention, not a physics. The losses become realized when the securities are sold or when circumstances force the question, and the 2023 failures demonstrated that circumstances can force the question quickly. The quarter's earnings were not built on top of the losses. They were built alongside them, and the two numbers will continue to travel together until rates or time do the work of reconciling them.
The third cushion was spent into the earnings
The cushions are not only on the balance sheet. Some are in the income statement, and the second quarter shows one being spent. Provision expenses fell to $19.3 billion, down 35.8 percent from a year earlier, as asset quality improved and the losses the industry reserved against failed to arrive. The reserves were not wasteful. They were insurance, and insurance that is not used looks, in the quarter after the danger passes, like money the bank can afford to release. So the industry released it, converting yesterday's caution into today's earnings, and the conversion is a large part of why the quarter looked as strong as it did.
The release is defensible, and it is also the pattern by which cushions disappear. Every reserve build is tested by the cycle it anticipated, and the cycle that fails to arrive makes the reserve look like over-preparation. The quarter's strength thus contains a quiet transfer: money that was set aside for losses is now flowing through earnings, which means the industry's future capacity to absorb the losses that did not arrive this quarter is slightly smaller than it was. Noninterest expenses, up 10 percent from a year earlier, moved the other way, the price of the growth itself. The pattern holds across the whole report. The quarter looked strong because the industry spent its margins, its reserves, and its capital growth in roughly equal measure.
There is also a denominator detail in the institution counts. The industry earned its $90.1 billion across 4,238 insured institutions, 41 fewer than a year earlier, the product of 36 mergers, four sales to nonbanks, four new charters, and one failure. The record earnings are spread over a shrinking roster, which is what consolidation does: the survivors get bigger, the averages get stronger, and the headline improves for reasons that have nothing to do with any individual bank's skill. Community banks participated, with net income up 8.2 percent and the share of unprofitable institutions down to 4.4 percent. The strength was real and broad. It was also, in part, arithmetic on a smaller denominator.
The fund is at a record, and the record hides a denominator problem
The Deposit Insurance Fund is, by its own measures, the healthiest it has been. The balance rose $3.7 billion to $161.1 billion, and the reserve ratio climbed 5 basis points to 1.48 percent, well on its way toward the statutory target. The insurance that calmed the 2023 panic is better funded than at any point in that crisis. The awkwardness is the denominator. The fund's adequacy is measured against insured deposits, and the insured deposit base is the part of the system that shrank this quarter. The fund is growing against a shrinking measure of the system it guarantees.
This is the inversion the whole report keeps producing. The industry earned more, lent more, and provisioned less, and each of those strengths touched a cushion. The earnings outpaced equity. The lending outpaced capital. The uninsured funding outpaced the insured. The fund grew while the fraction of the system it covers declined. Nothing in the report suggests fragility, and everything in it suggests a system that grows faster than its safeguards, then counts the safeguards in dollars and finds them ample.
The strongest quarter of the cycle is exactly the quarter in which the cushions should be rebuilt, because cushions are cheapest when nobody needs them. The report shows the opposite pattern: a system spending its strength on growth and letting its margins thin by basis points that no single quarter will notice. The basis points are not alarming. The direction is worth knowing, because the direction is what turns a strong report into a quiet one when the cycle turns.
Primary sources
- The FDIC's press release on the second quarter 2026 Quarterly Banking Profile for the net income, ROA, deposit, uninsured deposit, loan growth, asset quality, capital ratio, unrealized losses, Deposit Insurance Fund, problem bank, and community bank figures, along with Chair Travis Hill's comment.
- The FDIC's Quarterly Banking Profile page for the institution counts and the industry consolidation figures.
- American Banker's report by Ebrima Santos Sanneh for the earnings drivers, the provision expense comparison, and the margin detail.