Bank Lending Standards

What the survey asks

Four times a year the Federal Reserve asks senior loan officers at a panel of large domestic banks and foreign bank branches a set of questions about the terms they are offering. For commercial and industrial loans the key question is whether standards tightened, eased, or stayed roughly the same over the past three months. Banks answer separately for large and middle-market borrowers and for small firms, because the two groups are underwritten differently.

The published figures are net percentages: the share that tightened minus the share that eased. A reading of twenty means a fifth more of the panel tightened than eased. This distinction matters more than it looks. If forty percent tighten and forty percent ease, the net is zero and the headline says nothing changed, even though four fifths of the panel moved. The survey's own history suggests that is rare, but it is worth knowing the number is a balance rather than a level.

We average the large and middle-market line with the small firm line, giving each equal weight. Both cover the same quarters since 1990, so the composite has no gap and no splice. Equal weighting is a judgment, not a law: it treats a small firm's credit access as equally informative as a large firm's, which understates how much of the dollar volume sits with large borrowers and overstates how much of the employment sits with small ones.

Why this is the best available read on credit supply

Credit conditions have two sides: price and availability. The high-yield spread, the number most commentary quotes, is a price. It tells you what the market charges for risk, at the margin, for firms that can issue bonds. It does not tell you whether a bank will approve a working capital line for a company with forty employees, because that company never appears in a bond index.

The survey measures availability directly, and it measures intent rather than outcome. A loan officer who says standards tightened is describing a decision already taken, before the loan volume data show it. That is why the survey turns before almost everything else. In the past four cycles it moved up two to three quarters before credit spreads widened and three to five quarters before unemployment peaked.

The 2022 and 2023 episode is a clean illustration. The survey's net tightening jumped in the second half of 2022 as the policy rate rose, peaked through 2023, and then faded. Credit spreads over the same period stayed comparatively calm, because the firms that issue high-yield bonds were earning enough to service debt while smaller borrowers were being quietly rationed out. A reader watching spreads alone would have missed a genuine tightening of supply.

Reading the two legs separately

Large firms have an alternative to the bank. When standards tighten for them, they can issue commercial paper or bonds, and the survey reading usually falls back within a few quarters. Small firms have no such option: their credit access is the bank, and a tightening cycle for them tends to persist longer and cut deeper.

That asymmetry is why we publish both lines in the summary above the chart rather than only the average. When small-firm tightening runs well ahead of large-firm tightening, credit is being allocated by firm size rather than priced by risk. That condition historically precedes a wider gap in hiring between small and large employers, which is a more useful warning than the composite alone.

What the record shows

The composite has a median of roughly zero since 1990, because banks spend about as much time loosening as tightening. The distribution is heavily skewed. The 90th percentile sits near forty, the maximum is seventy-nine, recorded in October 2008, and the minimum is minus twenty-nine, recorded in July 2021 when banks were competing for business in a market flooded with deposits.

That skew is the point. Tightening is abrupt and clustered; loosening is slow and dispersed. A reading of fifty has happened four times in thirty-six years, all within nine months of a recession or a credit event. A reading of minus fifteen has happened dozens of times without consequence. The percentile is therefore more informative on the upside than on the downside, which is why the status line above the chart flags only the top tenth of the distribution.

How we calculate it

We take DRTSCILM, the net percentage of banks tightening standards for commercial and industrial loans to large and middle-market firms, and DRTSCIS, the same question for small firms, both from the Federal Reserve's survey via FRED. We average the two arithmetically and round to one decimal. No smoothing, no seasonal adjustment beyond what the source already applies, and no revision of past values.

The percentile is the share of quarterly observations since April 1990 that sit below the current reading. The change figure is the difference from the prior quarter, in percentage points.

Both input series are Federal Reserve releases in the public domain. The composite, its weighting and its percentile are computed here. That matters for a reason worth stating plainly: the widely quoted credit spread series are licensed commercial products whose terms forbid redistribution, so a site that wants to publish credit conditions commercially has to build its own measure from public inputs. This is ours.

Limits worth knowing

The survey is quarterly, which is slow for a crisis. It is also a survey of intent rather than a record of transactions: banks report what they did to their standards, not what they approved. And the panel is weighted toward large institutions, so it describes the supply of credit to firms that already have banking relationships better than it describes the availability of credit to new borrowers.

Questions

Related

• Bank credit impulse
• Credit market conditions
• Margin debt at US broker-dealers
• St. Louis Fed Financial Stress Index
• Plans and full-data access