Granular monetary policy surprises

Monetary policy surprises are widely used to identify causal effects of monetary policy changes changes. They are conventionally measured as a single number per announcement, common to all banks. We construct bank-specific surprises from banks' executed overnight index swap transactions in euro area regulatory data. Within announcements, banks differ by about as much as the aggregate surprise varies over time. Banks facing larger surprises reprice newly originated loans to small firms by 22 to 38 basis points more, while there is no such response for loans to large firms. Monetary policy transmission is thus not uniform across banks: which bank intermediates a small firm's credit shapes the cost of that credit after the same policy decision.

How strongly monetary policy affects firms largely depends on how banks’ credit provision responds to changes in monetary policy rates. Empirical studies often use the unexpected component of a policy announcement to identify causal effects of monetary policy on banks’ lending decisions. These monetary policy surprises are typically measured as a single shock common to all financial institutions (e.g. Jarociński and Karadi 2020). Such a common measure cannot reveal whether the decision surprised different banks differently. In this empirical paper, we construct bank-specific surprises from banks' own executed overnight index swap (OIS) transactions, contracts whose fixed rate prices the expected path of policy, around ECB Governing Council meetings. We then ask whether the bank-specific surprises generate heterogeneity in banks’ lending decisions to the same borrower.

One decision, many surprises

Our measurement combines two regulatory transaction datasets: the money market statistics (MMSR) provide reliable contract terms and identify the reporting bank, and the derivatives trade reports (EMIR) supply intraday execution timestamps. For the 36 Governing Council meetings between January 2021 and June 2025, we compare each bank's pre-announcement OIS transactions with a common post-announcement benchmark and aggregate these revisions into one surprise per bank and event: 366 in total from 31 banks, spanning 32 of the 36 meetings. Aggregated across banks, the measure closely tracks established euro area surprise series, with a correlation of 0.80 with the interest-rate surprise of Jarociński and Karadi (2020). These surprises matter for the banks themselves: banks with larger absolute surprises experience proprietary equity trading losses of about EUR 1 million per 10 basis points around the same meetings.

The same small firm pays different rates at differently surprised banks

The bank-specific surprises permit an identification that is unavailable with a common policy shock: Firm-month fixed effects control for a firm’s credit demand and allow us to isolate a lender’s specific credit supply, in the spirit of Khwaja and Mian (2008). They also absorb the effect of common policy news and of market conditions shared by all the firm's lenders in the respective month. Matching the surprises to newly originated fixed-rate corporate loans in AnaCredit, the euro area credit registry, we ask whether a bank with a larger tightening surprise charges a higher lending rate for new credit to the same firm.

The incidence differs sharply by firm size (Figure 1). Moving from the 10th to the 90th percentile of the within-event surprise distribution implies 22 to 38 basis points more repricing for small firms in the first and second complete months after the decision, while the response for large firms remains close to zero: the law of one price fails within firm and month. The differential emerges in the first complete month after the decision, is stable after controlling for loan amount, instrument type, and maturity, and operates in both directions. This pattern is consistent with relationship lending: informational advantages and switching costs give incumbent banks more pricing latitude for small firms, while competition and access to market finance limit that latitude for large firms.

Small firms bear the dispersion of their lenders' assessments

It is well established that small firms react more strongly to monetary policy, which is usually attributed to their weaker balance sheets and outside options (Gertler and Gilchrist 1994). Our results show that an important part of this transmission is decided at the level of the individual bank. Small firms therefore bear not only stronger average transmission but also the dispersion of their lenders' assessments, a burden that large firms escape because competition ties their rates to a common benchmark. In the euro area, where small and medium-sized enterprises dominate the firm population and rely primarily on bank credit, this lender-driven dispersion is a dimension of transmission that a common surprise measure cannot reveal, even in loan-level data.

References

Gertler, M and S Gilchrist (1994), "Monetary policy, business cycles, and the behavior of small manufacturing firms", The Quarterly Journal of Economics 109(2): 309‑340.

Jarociński, M and P Karadi (2020), "Deconstructing monetary policy surprises: The role of information shocks", American Economic Journal: Macroeconomics 12(2): 1-43.

Khwaja, A I and A Mian (2008), "Tracing the impact of bank liquidity shocks: Evidence from an emerging market", American Economic Review 98(4): 1413‑1442.

Fecht, F., B. Imbierowicz, M. Kellers (2026), Granular monetary policy surprises, Bundesbank Discussion Paper, No 38/2026