Credit conditions as predictors of housing prices
Marnix Hazelhoff
13 May 2025
Reading time 6 minutes
Van der Drift, R., de Haan, J., & Boelhouwer, P. (2024). Forecasting house prices through credit conditions: A Bayesian approach. Computational Economics.
This study centers on the question of to what extent credit conditions such as mortgage interest rates and borrowing limits (loan-to-income and loan-to-value ratios) contribute to the development of housing prices. The researchers explore whether these variables, which determine household borrowing capacity, can be systematically used to improve the accuracy of housing price forecasts in the Netherlands. Particular attention is given to the added value of Bayesian modeling approaches compared to traditional economic forecasting models.
Method
Because housing market data are limited, typically available only on a quarterly basis and over relatively short periods, traditional econometric models often lack robustness. To address this, the researchers opted for a Bayesian approach, which is particularly well-suited to small samples with multiple interdependent variables.
​Specifically, three models were used: a BVAR in levels, a BVAR in first differences, and a Bayesian Vector Error Correction Model (BVECM). These models analyze the interactions between, among other things, real house prices, mortgage interest rates, and lending ratios (LTI and LTV), using data from the Dutch market.
​
What distinguishes this approach is the use of so-called priors—predefined assumptions about economic relationships. These allow the models to combine theory-driven insights with empirical data. This enhances the stability of the outcomes and helps prevent overfitting, which is essential when data availability is limited.
Calculate bid
Do you want to know what to bid on your dream home? We calculate it for you.
What are the findings?
-
Credit conditions are strong predictors of house prices.
Both mortgage interest rates and borrowing limits prove to be significant explanatory variables for price developments in the housing market. A reduction in interest rates or a relaxation of borrowing constraints leads to higher prices, as buyers are able to bid more. -
Bayesian models outperform classical models
The Bayesian approach yields better forecasting performance compared to conventional non-Bayesian models, particularly for short-term predictions. -
Model choice depends on the type of analysis
-
The BVAR in differences (BVAR-d) provides the most accurate short-term forecasts of house prices.
-
The BVECM is suitable for situations in which long-term relationships between variables are relevant (such as cointegration between income and house prices).
-
The BVAR in levels performs the least well, unless strong economic priors are applied.
-
-
The effects of credit conditions materialize relatively quickly
Changes in mortgage interest rates or borrowing limits tend to influence the house price index within a few quarters, indicating a direct price elasticity on the demand side.
Conclusion
The results emphasize the importance of monetary policy in managing housing market dynamics. Policy changes related to borrowing limits or interest rates have predictable and relatively swift effects on house prices. For policymakers, this implies that credit conditions are not merely boundary constraints, but key policy instruments that can influence the stability and accessibility of the housing market.
​
Moreover, the study demonstrates that Bayesian econometric models are valuable tools for housing market analysis, particularly in contexts with limited data availability or where existing economic theory can guide model structure.
