From Clicks to Credit Risk: Nowcasting Sovereign Asset Quality Through Behavioral Search Analytics

<jats:p>This paper exploits Google Trends metadata to construct composite indices of credit demand, labor distress, and housing cost pressures for Malta between 2009 and 2026. A dynamic autoregressive specification accounts for the persistence of non-performing loans, while a Synthetic Control Method (SCM), using Italy as a counterfactual, isolates domestic from Eurozone-wide influences. The behavioral proxies significantly enhance nowcasting accuracy relative to traditional macroeconomic indicators. The analysis reveals a structural divergence between Malta and its counterfactual, confirming that local search patterns reflect idiosyncratic economic dynamics. These findings provide a scalable, real-time framework for macroprudential surveillance, demonstrating that digital footprints can serve as leading indicators of credit deterioration and facilitate more timely adjustments to capital buffers.</jats:p>