The paper provides an extension to first generation DSGE models with a financial sector -- for which QUEST III would be a typical example -- by explicitly modelling (mortgage) loan demand and supply decisions. We estimate a DSGE model with a housing sector where housing capital is used as collateral against which impatient consumers borrow from more patient lenders. While in existing estimated models with a construction sector the Loan-to-Value (LTV) ratio is imposed exogenously and constant (e.g., Iacoviello and Neri, 2010, In't Veld et al., 2011), we introduce an endogenous LTV ratio by explicitly modelling the riskiness of loans in order to capture changing credit conditions. Using data of the Euro Area, we show that, compared to similar models with an exogenous LTV ratio, the business cycle properties of our model improve. The endogenous default mechanism allows estimating an important amplification mechanism driven by the riskiness of collateral values and propagating, in turn, into the real economy. Housing market-related shocks appear to be the main driver of the pre-crisis growth of mortgage-backed loans and a subsequent reversal of the sentiment on the housing market may have been a trigger that led to a credit crunch, house price bubble burst and a collapse in the construction sector. Shocks on the housing market had also a substantial impact on several demand aggregates, in particular, consumption."--Document home page.