ECB Bank Lending Survey

The euro area Bank Lending Survey (BLS) is a quarterly ECB survey for the Eurosystem (the ECB plus euro area national central banks). Launched in 2003, it polls senior loan officers at around 150 to 160 euro area banks.

It covers enterprises, house purchase and consumer credit. Its 22 standard questions span the past and next 3 months, plus 1 open question and ad hoc extras. It supplements loan statistics, reporting lender change only, never loan volumes, rate levels or borrower views.

Each round collects replies over 2 to 3 weeks around quarter-end. Banks report change over the past 3 months and expected change over the next 3 months. Results come out 4 times a year, usually in January, April, July and October at 10:00 Frankfurt time. Each round covers past and next quarters.

Each release brings a press release and a euro area plus five-country overview. It also brings annex tables, the questionnaire, the glossary, the user guide and data series. National central banks publish country detail. The survey has no regular revision cycle.

A positive net means tightening or stronger demand, while a negative net means easing or weaker demand. Euro area nets from -1% to +1% count as broadly unchanged.

Six headline blocks move results across enterprises, house purchase and consumer credit.

• Credit standards: loan approval criteria applied before negotiation, the core supply gauge.
• Credit terms and conditions: actual contract terms: margins on average and riskier loans, non-interest charges, collateral, maturity and lending rates.
• Loan demand: borrowers' financing need apart from normal seasonal swings, including rollovers, whether or not a loan follows.
• Share of rejected applications: realised friction where loan demand meets supply standards.
• Supply-side factors: cost of funds and balance sheet strains, competition pressure, risk perception and risk tolerance.
• Demand-side factors: fixed investment, inventories, mergers, rates and alternative finance, plus household confidence and housing prospects.

The May 2, 2023 round for the first quarter of 2023 shows joint tightening at work. Enterprise standards tightened to net 27%, the fastest pace since the 2011 sovereign debt crisis and stronger than banks had expected. House purchase standards hit net 19% and consumer credit net 10%, with further moderate tightening expected for the next quarter.

Risk perception, low risk tolerance and funding costs drove the squeeze amid rate hikes and thin central liquidity. Firm demand fell more than expected, weakest since the global financial crisis. Housing demand neared its record fall since 2003 on rates, weak investment, housing prospects and confidence.

1. Replies: Banks answer on a 5-point scale from tighten or decrease considerably to ease or increase considerably, measuring change not level.
2. National and euro area aggregation: National replies use equal or loan-book weight, as in France, Malta, the Netherlands and Slovakia. Euro area results weight each country by its share of loans to firms and households.
3. Denominator: Shares cover banks with business in that loan category, including specialists in their own lines. Banks without exposure stay outside, while other not applicable replies stay inside under the April 2018 harmonised rule.
4. Net percentage: Supply net equals share tightening minus share easing, while demand net equals share rising minus share falling. Above zero means tightening or stronger demand, while below zero means easing or weaker demand. For example, 100 banks split 10% and 30% tightening, 40% unchanged and 15% plus 5% easing give net +20% tightening.
5. Diffusion index and mean: Diffusion weights considerably at 1 and somewhat at 0.5 with the same sign logic. The same split gives diffusion 12.5. Mean averages replies 1 to 5, where 3 means unchanged. The same split gives mean 2.75, on the tightening side.
6. Thresholds and seasonal rule: Euro area nets from -1% to +1% count as broadly unchanged. No seasonal adjustment applies, since demand already excludes normal swings.