Consumer Mortgage Choice in Spatially Differentiated Housing Markets: An Agent-Based Model of Default and Refinancing

<title>Abstract</title> <p>We develop a spatially explicit agent-based model to examine how household heterogeneity, urban housing-market segmentation, and institutional mortgage constraints jointly shape regional patterns of credit vulnerability in an emerging market economy. Integrating a buffer-stock income process with an option-based mortgage decision framework, the model captures household-level default and refinancing choices within a differentiated urban environment. The framework is calibrated using 11,092 residential listings across ten districts of Ankara, Türkiye, allowing spatial variation in housing values and collateral exposure to be incorporated into the analysis. The simulations reveal three main findings. First, under a high-interest-rate regime, slow mortgage amortization leaves highly leveraged households vulnerable to collateral shocks, with mortgage distress concentrated among borrowers in the upper loan-to-value distribution rather than uniformly across the housing market. Second, spatial differences in housing values and appreciation dynamics generate substantial variation in household balance-sheet resilience, implying that regional location is an important determinant of mortgage vulnerability. Third, institutional features, including full-recourse lending and refinancing eligibility constraints, significantly influence household responses to financial stress by limiting strategic default and restricting access to profitable refinancing opportunities. The results highlight how spatial heterogeneity and mortgage market institutions interact to determine the distribution of housing-related financial risks within urban regions.</p>