In collaboration with a large Belgian bank, we build a unique database of residential mortgage loans with extensive information about the loans and borrowers. With this data we tackle three policy-relevant issues: climate risk (both physical and transition), macroprudential lending restrictions and Covid-moratoria. Regarding the physical risk of mortgage loans, we investigate whether flood risk is reflected in the value of houses and in the pricing of Belgian mortgage loans, as well as whether flood risk affects mortgage performance (probability of default and loss-given-default). Moreover, we assess whether or not the impact is higher for certain social groups, e.g., depending on age or income, exploiting the granularity of the data in combination with the July 2021 floods as exogenous shock. In terms of transition risk, we analyze how energy efficiency affects bank mortgage loan pricing and the default risk of mortgages, and whether banks contribute to climate solutions through pricing of renovation loans. Regarding macroprudential policy, we investigate the effectiveness in curbing bank risk-taking of loan-to-value limits for mortgages recently introduced by the National Bank of Belgium. We also analyze whether temporary Covid-19 moratoria help households in resuming the servicing of their loans without arrears or defaults. With our granular data, we also investigate potential social implications of these policies (e.g., are young/poor households affected differently?)