Unit root test results.

This study aims to investigate the time-varying bidirectional causal relationship between household education expenditure and resident credit behavior, as well as the heterogeneous mechanisms under different economic conditions and household characteristics. By constructing a TVP-SV-VAR model and a QVAR-DY model, we analyze urban household data in China from January 2015 to December 2024, unveiling the dynamic relationship between education expenditure and credit behavior, along with their asymmetry and heterogeneity. The findings reveal a significant bidirectional causal relationship between household education expenditure and resident credit behavior, which exhibits heterogeneity across different quantile levels and is influenced by household income, education level, and credit interest rates. Additionally, this study employs static and dynamic window methods to analyze the short-term, medium-term, and long-term spillover effects. Based on these findings, we propose policy recommendations for optimizing household education investment and credit market management under low, medium, and high-risk levels.