The Chicken or the Egg?

Whether changes in house prices are caused by changes in credit, or whether credit follows house prices is a long standing debate. In the presented paper, the authors propose a new identification strategy to assess the causal effect of credit on prices, using a matched dataset at zip code level of granular transaction/listing data, loan-level supervisory data, municipality/resister data, price indexes, individual tax records, and macroprudential caps. They use exogenous changes to these debt service to income caps, which are heterogeneous over time and individuals, to identify this effect. This is a more granular strategy than in previous studies, where at most zip code level heterogeneity was used. This strategy is supported by the institutional design of the caps in the Netherlands, where an NGO delivers yearly budgetary computations, upon which the DSTI caps are based. These caps differ by household income and the mortgage interest rate that is assigned depending on credit-risk categories. Despite the enormously large range of possible changes to such credit caps over time, the authors show that the institutional design allows disentangling exogenous changes to maximum indebtedness due to policy,  from changes related to developments in household income, the interest rate or individual choices. This exogenous component is then used to instrument for household debt in a regression explaining prices in the zip codes where this debt is being originated, and where they also observe the quality of the listed property (form the listing data) and the financial situation of the new owner (from tax records). The results show that a reduction of borrowing capacity by 10% of the original principal, would induce a reduction in prices by 4%.