France SME Borrowing Premium: Small vs Large Loan Rates Since 2003

The France SME borrowing premium is an Eco3min composite: the average rate on new corporate loans up to one million euros, minus the rate on loans above that threshold, monthly since January 2003. It measures what a small French borrower pays over a large one for the same act of borrowing. The premium has averaged 0.55 percentage points over 280 months, peaked at 1.86 points in January 2009, and has been negative in 32 of those months — periods when small firms borrowed more cheaply than large ones. No institution publishes this series natively.

Dataset: France SME Borrowing Premium (2003–2026) · Updated 2026-04-01

Source: Banque de France (MIR1) · Eco3min composite

Macro Takeaway

Construction & Components

The composite isolates the price of size in French corporate credit. It removes the level of interest rates from the picture and leaves only the gap between what small and large borrowers pay, which makes episodes of credit rationing visible in a way neither component shows on its own.

Formula:

SME Borrowing Premium = Rate on new loans up to EUR 1m − Rate on new loans over EUR 1m

Components:

• New corporate loans up to EUR 1 million — Banque de France, MIR1 — monthly. The standard proxy for SME borrowing in euro area statistics, since loan size correlates strongly with firm size.
• New corporate loans over EUR 1 million — Banque de France, MIR1 — monthly. The large-corporate and mid-cap segment.

Frequency reconciliation: Both components are natively monthly and published together. The pipeline aligns them with a backward as-of merge on the first day of each month; no interpolation is applied.

Coverage: January 2003 to April 2026, bounded by the start of the MIR series on both legs.

Dataset Overview

Dataset Variables

The CSV, Excel and JSON files contain the following columns.

Column names match the CSV headers exactly.

Download the Complete Dataset

The full SME borrowing premium series, including both components, is available in CSV and Excel formats.

Direct Data Access

No upstream source publishes this difference. Both legs can be retrieved separately from the Banque de France Webstat API, which requires registration. The Eco3min structured dataset returns the composite and both inputs on one grid:

Direct CSV Access — Eco3min Structured Dataset

https://eco3min.fr/dataset/fr/fr-sme-rate-premium.csv

This URL returns the complete dataset in CSV format. It can be used directly in pandas, R, curl, or any data tool. JSON and XLSX are available at the same path with the corresponding extension.

Using the Dataset in Python

import pandas as pd

url = "https://eco3min.fr/dataset/fr/fr-sme-rate-premium.csv"
df = pd.read_csv(url, parse_dates=["date"])

print(df.head())
print(df["sme_premium"].describe())

Using the Dataset in R

library(readr)

url <- "https://eco3min.fr/dataset/fr/fr-sme-rate-premium.csv"
df <- read_csv(url)

head(df)
summary(df$sme_premium)

Both examples load the dataset directly from the URL — no download or API key required.

Methodology

The composite is recomputed daily by an Eco3min pipeline that pulls both components, aligns them on the monthly grid and takes the difference. Both are natively monthly, so no observation is synthesised.

The one-million-euro threshold is the Banque de France’s own reporting cut, not an Eco3min choice, and it has not moved over the period. That stability is what makes the series comparable across twenty-three years, but it also means the threshold has eroded in real terms: one million euros in 2003 bought considerably more than one million in 2026, so the composition of each bucket has drifted.

Data Quality & Provider Notes

Latency follows the Banque de France MIR release, roughly five to six weeks after month end.

Revisions propagate: when the Banque de France restates a past MIR observation, the corresponding months of the composite are rewritten at the next Eco3min run.

No native alternative exists. The Banque de France publishes both legs but not their difference, and no commercial provider offers it as a ready-made series — which is the point of the composite, and also why no external benchmark exists to validate it against.

What This Index Captures (And What It Doesn’t)

The composite is a structural indicator of the pricing gap between small and large corporate borrowers. It is not a measure of credit availability and not a market-timing tool.

What it captures:

• Episodes where small borrowers were priced materially above large ones, which is the classic signature of a credit crunch
• The compression of that gap under negative policy rates, and the periods where it inverted entirely
• The difference between a rate cycle that hit all borrowers equally and one that hit them differently, by reading the two component columns alongside the premium

What it does NOT capture (common misinterpretations):

• Firm size. Loan amount is a proxy, not a measure. A large firm borrowing 800,000 euros lands in the small bucket, and a fast-growing SME borrowing 1.2 million lands in the large one. The composite prices loan size, not company size.
• Credit availability. A narrow premium does not mean small firms are getting credit. Refusals do not appear in a series built from loans actually granted, and the periods of tightest rationing can show a narrow gap simply because the riskiest applications were declined rather than priced.
• The all-in cost. MIR rates exclude arrangement fees, guarantee costs and covenant pricing, all of which weigh more heavily on small borrowers. The true gap is wider than this series shows.
• Public guarantees. The 2020–2021 state-guaranteed loan scheme distorted both legs, compressing the premium for reasons that had nothing to do with market pricing.

The series is most useful for dating and characterising episodes where the cost of credit diverged by borrower size, read alongside its two component columns.

Historical Regimes

• 2003–2007 — a stable narrow gap. The premium started at 0.24 points and stayed under half a point for most of the period. Bank credit was abundant and size mattered little to pricing.
• 2008–2009 — the credit crunch. The premium reached 1.86 points in January 2009, its series high. Impaired bank balance sheets translated directly into a size penalty.
• 2010–2014 — a persistent wide gap. Still 1.146 points in January 2012, at the height of the sovereign crisis. The premium stayed above one point far longer than the crisis itself.
• 2015–2019 — compression. Down to 0.63 points by January 2016 and lower thereafter, as negative policy rates and asset purchases pushed banks to lend across the size spectrum.
• 2020–2024 — inversion. The premium reached 0.026 points in January 2021, turned negative in 2023, and hit its series low of −0.39 points in August 2024. State guarantees and then competition for small-business relationships both contributed.
• 2025–2026 — partial normalisation. Back to 0.28 points by April 2026, still around half the long-run average of 0.55.

Related Macroeconomic Datasets

Related Research

Macroeconomic Dataset Hub

This dataset is part of the Eco3min macro-financial data repository.

Explore the Eco3min Dataset Hub

Sources

• Banque de France — Webstat, MIR1 dataset, new loans to non-financial corporations, split at EUR 1 million. Licence Ouverte / Open Licence 2.0 (Etalab).
• Composite computed by Eco3min from the two series above.

Dataset Reference

Eco3min Research. (2026). France SME Borrowing Premium (2003–2026). Eco3min composite from Banque de France data.<br> https://eco3min.fr/en/france-sme-borrowing-premium-dataset/