"Broad-based" loan growth is real for big banks right now — and a coin flip for smaller ones

Direct answer

DBRS Morningstar's September 2026 commentary describes US bank loan growth as broad-based. [C01] That holds up well for the largest banks: the $250B+ cohort has had at least 75% of its members growing loans for four straight quarters, and the $50-250B cohort for three straight quarters. [C05, C06] It holds up less well at the $10-50B cohort, where breadth has flipped between broad and narrow quarter to quarter, including a narrow reading as recently as Q1 2026. [C07] Broad-based growth is also not the historical norm at any cohort — only about half of all quarters since 2016 clear that bar. [C04]

Key findings

What "broad-based" means, tested directly

DBRS's characterization doesn't come with a published definition, so this article uses a specific, falsifiable one: a cohort's 25th percentile of quarter-over-quarter loan growth being positive, meaning at least three in four members grew. By that test, this quarter genuinely qualifies at the two largest cohorts. [C02, C05, C06] But that alone doesn't establish that "broad-based" is a normal, expected state of the world — historically, only 20-21 of the last 41 quarters clear the same bar at any given cohort. [C04]

Roughly half of all quarters since 2016 would also have qualified as "broad-based" by this test. The current streak at the largest banks is real, but it is not evidence that broad growth is now the default.

The streak, cohort by cohort

The $250B+ and $50-250B cohorts both moved from negative breadth readings in early-to-mid 2025 to a consistent, multi-quarter positive streak by Q2 2026. [C05, C06] The $10-50B cohort shows no comparable streak: it flipped negative again in Q1 2026 (-0.52%) before returning positive in Q2 2026 (0.31%). [C07] "Broad-based loan growth" is a more accurate description of what the largest banks are doing right now than what the smallest $10B+ tier is doing.

A headline distortion worth naming before using the summary statistics

Charles Schwab Premier Bank SSB's reported loan book fell 99.97% quarter over quarter — a change consistent with a portfolio transfer or balance-sheet restructuring inside a banking group, not organic loan contraction at a going lending business. [C03] Including this kind of event in an unweighted "banks that shrank" narrative would overstate genuine credit contraction; it is named and excluded from that framing here rather than silently averaged into the population.

Methodology

Reuses OptimaYield's existing, already-validated loan_growth_rate metric contract via the shared deriveValidatedMetrics function, supplying both the current and prior quarter for each institution (the contract correctly abstains when the prior quarter isn't available — confirmed directly, not assumed, before building this analysis). Current-quarter population: every institution in the detailed ($10B+) Call Report layer with a derivable growth rate for Q2 2026 against Q1 2026 (151 institutions). Historical breadth reuses the already-computed percentile series in semantic.peer_statistics (41 quarters with a valid comparison, 2016-06-30 through 2026-06-30); "broad-based" is defined as a positive 25th percentile for that cohort-quarter. Live queries: scripts/analysis-loan-growth-population.ts and scripts/analysis-loan-growth-history.ts.

Limitations and counter-evidence

• This tests one specific, self-defined operationalization of "broad-based," not DBRS's own methodology, which is not published at this level of detail — a different breadth threshold could show a different picture.
• This does not test DBRS's full commentary. [C08] The "improved earnings" half of that commentary, and the broader NIM-trajectory theme this report is narrowed from, are not addressed here — no validated net-interest-margin metric contract exists yet in this codebase, checked directly rather than attempting an incomplete decomposition.
• One quarter's streak is not a trend guarantee. Four consecutive broad quarters at the $250B+ cohort is a real, recent pattern, not a claim about what happens next quarter.
• No cause is claimed for why breadth improved at the two largest cohorts or why the $10-50B cohort lags — this is a description of what happened, not why.

Related on OptimaYield

• Q1 2026 covered-bank growth divergence — the prior quarter's companion piece on asset growth dispersion, the same distributional discipline applied here to loan growth specifically.
• The bank-profile pages for the named institution above (not yet linked pending publication approval).

Official sources

• DBRS Morningstar, commentary on US banking sector conditions, September 2026.