Residential real estate (RRE) lending standards: determinants and financial stability implications
Published as part of the Macroprudential Bulletin 29, June 2025.
This article looks into residential real estate (RRE) lending standards, focusing on their key determinants and assessing the implications of loose lending standards for financial stability and the real economy. Two key insights emerge. First, lending standards tend to be procyclical – i.e., they become looser during economic upturns and tighter during downturns. Second, loose lending standards amplify the effects of negative housing market shocks on the real economy and heighten financial stability risks via an increase in the probability of default of households.
1 Introduction
Fluctuations in lending standards[1] have a direct impact on the risk profiles of borrowers, lenders and the broader economy, so monitoring these changes is crucial to safeguard financial stability. Tighter lending criteria indicate that financial institutions are applying more selective lending strategies and thereby reducing their exposure to default risk. Conversely, looser lending standards can lead to excessive borrowing and reflect greater risk-taking behaviour among banks. Changes in lending standards can amplify business cycle fluctuations by influencing consumers’ spending and investment decisions. Proper oversight of lending standards needs to avoid a build-up of excessive debt while also allowing for an adequate provision of credit to households.
Lending standards in the euro area had been on a decade-long upward trend during the period of low interest rates, but that trend started to reverse in 2022, at the turn of the monetary policy and housing market cycle. The share of lending with high loan-to-value (LTV) and loan-to-income (LTI) ratios has declined since 2022, while lending with high loan-service-to-income (LSTI) ratios has increased.[2] This underscores the potential link between lending standards and the monetary policy and economic cycles. More recently, the gradual decline in monetary policy rates and indications that the residential real estate (RRE) cycle has turned up again, could prompt renewed attention on the dynamics of lending standards.
Against this background, this article explores lending standards to better understand their driving factors and the potential implications for financial stability. The article presents two main insights: first, in line with previous research, we find that lending standards are procyclical, meaning they tend to loosen during upturns of the cycle and tighten during downturns. This procyclical behaviour can jeopardise financial stability, especially during boom-bust cycles in the real estate market.[3] Second, we show that relatively loose[4] lending standards can have negative implications for the real economy (i.e., GDP, lending, house prices and real estate investment) and financial stability (i.e., the probability of default of households). Implementing targeted macroprudential policies, such as borrower-based measures that seek to enhance the resilience of borrowers and lenders, could avoid excessive risk-taking during periods of economic expansion and help ensure that lending standards are maintained at a prudent level throughout the various phases of the cycle. The crucial role of borrower-based measures is discussed in more detail in Article 2.
2 Lending standards and the economic cycle
A regression analysis indicates that lending standards tend to loosen during the expansionary phase of the cycle and tighten in downturns. Estimates based on a panel of loan-level data for seven euro area countries[5] reveal a positive and statistically significant relationship between lending standards and macroeconomic variables that proxy for the economic and financial cycle (Chart 1, panel a). In particular, a one-standard-deviation increase in GDP growth, RRE price growth and measures of cyclical risks (i.e., the systemic risk indicator (SRI))[6] is associated with an increase in LTVs at origination of 0.08, 0.9 and 3.08 percentage points respectively. Similar results hold for LTIs at origination: a one-standard-deviation increase in GDP growth, RRE price growth and the SRI is associated with an increase in LTIs at origination of 8, 11 and 34 percentage points respectively. The impact of short-term policy rates (proxied by the three-month overnight index swap (OIS) rate) goes in the opposite direction. A one-standard-deviation increase in the OIS rate results in a 1.4 (5.3) percentage-point decrease in LTVs (LTIs) at origination, as the standard balance sheet channel of monetary policy transmission would predict.[7] Consequently, lending standards tend to loosen during upswing phases of the RRE cycle (Chart 1, panel b). On average, during RRE price booms[8], LTVs (LTIs) at origination were around 4 (16) percentage points higher than during periods of lower or moderate RRE price growth.[9]
Chart 1
Lending standards deteriorate in periods of robust GDP growth and RRE price growth.
a) Lending standards are procyclical… | b) … as such, they become a lot looser during RRE price booms. |
|---|---|
(percentage points) | (percentages) |
 |  |
Sources: European Data Warehouse (EDW) and ECB calculations.
Notes: Panel a) shows standardised coefficients and the 95% confidence intervals from a (conditional) linear ordinary least squares (OLS) regression. The dependent variable is a metric of lending standards (LTV and LTI) and the regressors are different macro-financial variables. The period covers 2004 to 2022. The pandemic period (2020-21) is excluded. The regression has country and time-fixed effects and also controls for the unemployment rate and borrower level factors, e.g. loan size, income and employment status. All coefficients are significant at the 1% level. The standard deviation of GDP growth, RRE price growth, the SRI and the OIS rate in the estimation sample is 3.5, 6.2, 0.7 and 1.5 respectively. This means that, economically, a one-standard-deviation increase in GDP growth, RRE price growth and the SRI is associated with around a 0.3%, 6.0% and 2.0% increase in LTV respectively and 28%, 68% and 25% increase in LTI respectively. By contrast, a one-standard-deviation increase in the OIS rate is associated with around a 2% decrease in LTV and an 8% decrease in the LTI. Note that policy rates remained constant (and low) for almost the entire period under consideration. Panel b) shows the standardised linear predictions and the 95% confidence intervals obtained from a conditional OLS regression by introducing a dummy for different states of the RRE cycle. The dummy takes the value 1 (a high RRE price growth state) if the RRE price growth in a given country is above its historical median and 0 (a low RRE price growth state) if it is not.
3 Impact on the real economy and financial stability
Loose lending standards heighten the macroeconomic impact of RRE price shocks on the real economy because they tend to lead to higher indebtedness. When households face higher debt repayment obligations, the effects of a negative housing demand shock on output, lending, house prices and investments tend to be amplified. This is because more indebted households – which typically enter the market during periods of loose lending standards – are more likely to cut spending more sharply when housing values fall, owing to tighter borrowing constraints and worsening macroeconomic conditions. To explore this mechanism, we use two macroeconomic models to simulate a negative RRE demand shock[10]. In the first model, we compare the impact on house prices and residential investment between countries with a historically high debt service ratio (DSR) and countries with a lower DSR. The results show that countries with a higher DSR see a larger fall in house prices and real estate prices and investment (Chart 2, panel a). In the second model, we assess how the same shock affects GDP and lending in the euro area under different levels of cyclical risk, proxied by the two-year change in the DSR. The decline in GDP and lending is significantly larger in a high-risk environment (Chart 2, panel b). Taken together, the results from the two models suggest that economies with more indebted households are more exposed to negative RRE shocks, with potentially more severe consequences for the real economy. This supports the view that more indebted households – those for which debt repayments account for a larger share their income – are more likely to cut consumption owing to reduced household wealth.[11] This amplifies the contraction in GDP and the labour market, raising default risks more broadly. This article further analyses the link between loose lending standards and the probability of default of households.
Chart 2
In high-risk environments, adverse RRE shocks have a stronger impact on RRE market variables, GDP and lending.
a) Impulse response to a negative RRE shock from a panel Bayesian VAR model | b) Impulse response to a negative RRE shock from a non-linear local projection model |
|---|---|
(y-axis: percentage deviation from the starting point, x-axis: quarters) | (y-axis: percentage deviation from the starting point, x-axis: quarters) |
|  |
Sources: ECB and own calculations.
Notes: Panel a) shows impulse responses to a negative housing demand shock from a panel Bayesian VAR model applied to ten euro area countries over a sample period spanning from the first quarter of 2003 to the fourth quarter of 2019. Panel b) shows impulse responses to a negative housing demand shock from a non-linear macroeconomic local projection model. For more details, see Couaillier, C. and Scalone, V. (2024), “Risk-to-buffer: Setting cyclical and structural banks capital requirements through stress tests”, European Central Bank Working Paper Series, No. 2966. For low risk the state variable is at the 20th percentile (blue line), for medium risk it is at the 50th percentile (yellow line) and for high risk it is at the 80th percentile (red line) of the historical distribution of the two-year change in the euro area DSR ratio.
Loose aggregate lending standards can increase borrowers’ probability of default over and above the impact of loose individual lending standards. A loan-level analysis for eight euro area economies reveals that the probability of default of all borrowers increases substantially when many borrowers have high LTI ratios simultaneously (Chart 3).[12] This result also holds when individual loan-level lending standards at origination are used as controls. Our findings are in line with other studies on the topic, which show that weak individual lending standards correlate with a higher subsequent probability of default.[13] Aggregate lending standards can affect borrowers’ probability of default through the following channel. If an adverse shock hits the economy, a large number of borrowers may find it difficult to repay their loans. This can lead to a pronounced decline in consumer spending and economic output, triggering higher unemployment rates. As a result, even borrowers who might not initially appear risky are affected by the broader economic downturn. Their likelihood of default then increases, amplifying the impact of the adverse shock to the economy. Overall, the significance of aggregate lending standards for the probability of mortgage default shows the importance of preventing a large share of borrowers having loose lending standards at the same time.
Chart 3
If a relatively large number of borrowers are granted loans with loose LTIs simultaneously, the probability of default increases for all, on top of the impact of individual risky lending standards.
a) Impact of lending standards (individual and aggregate) on the probability of mortgage default | b) Estimated probability of mortgage default with aggregate LTIs risky compared to non-risky |
|---|---|
(percentage points) | (percentage points) |
 |  |
Sources: European Data Warehouse (EDW) and ECB calculations.
Notes: The charts are based on a loan-level-based logit model on probability of mortgage default within ten years of loan origination, estimated on loans originated in Belgium, Germany, France, Ireland, Italy, Spain, the Netherlands and Portugal between 2004 and 2010. The focus on ten years instead of the full life cycle is due to data availability. Explanatory variables only contain the information available at loan origination, namely maturity of the loan, borrower’s employment status, borrower total income (deviation from the mean), payment type, unemployment rate at country level, borrower lending standards and aggregate lending standards. Loose lending standards are measured by the following dummies: aggregate risky LTI = 1 if the share of LTIs>5 in a given country and year is above the 90th percentile of the distribution of the shares of new loans with LTIs>5 across time and countries. Borrower risky LSTI = 1 if LSTI at origination ≥35 and borrower risky LTV = 1 if LTV at origination ≥ 80.
4 Conclusion
The adverse effects of loose lending standards on financial stability and the real economy underline the importance of well-designed macroprudential policies for mitigating aggregate risks. This article shows that lending standards are procyclical, as they loosen during economic or financial cycle upswings and tighten during downturns. Loose lending standards can have significant macroeconomic implications, as higher debt service burdens amplify the negative effects of RRE shocks on output, lending, RRE prices and investment. In line with these macroeconomic effects, a high share of lending with loose standards increases the probability of default of all borrowers. Taken together, these findings suggest that activating targeted macroprudential policies (i.e. borrower-based measures) early in the cycle and maintaining them throughout (i.e. structural use of the policies) is particularly beneficial, not only for ensuring sustainable lending standards but also for limiting the share of risky lending in the upswing of the cycle (Article 2).
5 References
Couaillier, C., and Scalone, V. (2024), “Risk-to-buffer: Setting cyclical and structural banks capital requirements through stress tests”, European Central Bank Working Paper Series, No. 2966.
Bakker, B. B. et al. (2012), “Policies for macrofinancial stability: How to deal with credit booms”, International Monetary Fund, Staff Discussion Note No. 2012/006.
Guerrieri, L. and Iacoviello, M. (2017), “Collateral constraints and macroeconomic asymmetries”, Journal of Monetary Economics, Vol. 90, October, pp. 28-49.
Jarmulska, B., Perales, C. and Foerster, K. (2023), “Evolution of mortgage lending standards at the turn of the housing market cycle”, Macroprudential Bulletin, No 22, ECB, July.
Kiyotaki, N. and Moore, J. (1997), “Credit cycles”, Journal of Political Economy, Vol. 105, No 2, pp. 211-248.
Lang, J. H. et al. (2019), “Anticipating the bust: a new cyclical systemic risk indicator to assess the likelihood and severity of financial crises”, Occasional Paper Series, No 219, ECB, February.
Lang, J. H. et al. (2020), “Trends in residential real estate lending standards and implications for financial stability”, Financial Stability Review, ECB, May.
Lang, J. H., Behn, M., Jarmulska, B., Lo Duca, M. (2022), “Real estate markets, financial stability and macroprudential policy”, Macroprudential Bulletin, No 19, ECB, October.
Mian, A. and Sufi, A. (2009), “The Consequences of Mortgage Credit Expansion: Evidence from the U.S. Mortgage Default Crisis”, The Quarterly Journal of Economics, Vol. 124, No 4, pp. 1449-1496.
Schularick, M. and Taylor, A.M. (2012), “Credit Booms Gone Bust: Monetary Policy, Leverage Cycles, and Financial Crises, 1870-2008”, American Economic Review, Vol. 102, No 2, pp. 1029-1061.
In this article, the term “lending standards” specifically refers to the standards applicable to residential real estate loans – i.e., those applied to newly originated loans for housing purposes.
See Jarmulska, B., Perales, C. and Foerster, K. (2023), “Evolution of mortgage lending standards at the turn of the housing market cycle”, Macroprudential Bulletin, No 22, ECB, July.
See Mian, A. and Sufi, A. (2009), “The consequences of mortgage credit expansion: Evidence from the US mortgage default crisis”, The Quarterly Journal of Economics, Vol. 124, No 4, pp. 1449-1496 and Schularick, M. and Taylor, A. M. (2012), “Credit booms gone bust: Monetary policy, leverage cycles, and financial crises, 1870-2008”, American Economic Review, Vol. 102, No 2, pp.1029-1061.
Throughout this article, we the terms “risky” and “loose” are used interchangeably to describe lending standards in an environment in which there is a relatively high share of new mortgage loans with high LTVs, L(D)TIs or L(D)STIs in comparison with the distributions of the respective indicators across countries and time.
Data on lending standards are sourced from the European Data Warehouse (EDW), which collects reports on securitised loans. On average, loans reported to the EDW cover 20% of newly originated loans for house purchase. The data for LTV and LTI at origination are taken directly as reported in the EDW dataset (i.e., in numerical form). Data on lending standards can also be sourced from the Bank Lending Survey (BLS). In fact, in the BLS, credit standards are defined as internal guidelines or criteria that a bank adheres to when approving new loans. This definition conceptually matches with the LTV and LTI reported in the EDW dataset. Nevertheless, in the BLS only changes in credit standards can be observed, not the numerical values. The countries considered in the analysis are Belgium, Spain, France, Ireland, Italy, the Netherlands and Portugal.
The systemic risk indicator is a broad-based domestic cyclical risk indicator that captures risks stemming from domestic credit, real estate markets, asset prices and external imbalances (see Lang, J.H. et al. (2019), “Anticipating the bust: a new cyclical systemic risk indicator to assess the likelihood and severity of financial crises”, Occasional Paper Series, No 219, ECB, February).
Overall, these results are in line with previous findings that point to looser lending standards in countries that saw stronger real estate expansions during the pre-pandemic period (see Lang, J.H. et al. (2020), “Trends in residential real estate lending standards and implications for financial stability”, Financial Stability Review, ECB, May).
A country is considered to be in a “high RRE price growth state” in the period when RRE price growth is above its (country-specific) historical median and in a “low RRE price growth state” when the RRE price growth is below the country median.
The relationship between lending standards and economic cycles has been a central topic of financial stability research. The results presented in this paragraph are in line with previous well-establish findings in the literature, indicating that lending standards are typically procyclical. See, for example, Mian, A. and Sufi, A. (2009), “The Consequences of Mortgage Credit Expansion: Evidence from the U.S. Mortgage Default Crisis”, The Quarterly Journal of Economics, Vol. 124, No 4, pp. 1449-1496 and Schularick, M. and Taylor, A.M. (2012), “Credit Booms Gone Bust: Monetary Policy, Leverage Cycles, and Financial Crises, 1870-2008”, American Economic Review, Vol. 102, No 2, pp. 1029-1061.
Simulated as an exogenous negative change in households’ preferences regarding RRE assets. The two models are a Bayesian VAR model and a multivariate non-linear local projection model.
See Kiyotaki, N., and Moore, J. (1997), “Credit cycles”, Journal of Political Economy, Vol. 105, No 2, pp. 211-248 and Guerrieri, L., and Iacoviello, M. (2017), “Collateral constraints and macroeconomic asymmetries”, Journal of Monetary Economics, Vol. 90, October, pp. 28-49.
In particular, the probability of default of all borrowers increases by 0.1 percentage points when the share of borrowers with an LTI above 5 exceeds 31%. This cut-off stems from the fact that we consider the aggregate LTI to be “risky” if the share of LTIs>5 in a given country and year is above the 90th percentile of the distribution of the shares of new loans with LTIs>5 across time and countries.
See for example, Bakker, B. B. et al. (2012), “Policies for macrofinancial stability: How to deal with credit booms”, International Monetary Fund, Staff Discussion Note No. 2012/006