Multi-attribute utility and the sign of the cross-derivative: evidence from changes in insurance demand during COVID-19

The COVID-19 pandemic has simultaneously impacted wealth levels and health status. The sign of the cross-derivative of the multi-attribute utility function plays an important role because it captures a trade-off between attributes. This study assesses the sign of the cross-derivative by analyzing data from a representative Japanese survey on demand for income protection and death insurance, comparing figures before and after the pandemic. To estimate preferences, we develop an economic demand model for insurance during the pandemic. We also analyze how the sign of the cross-derivative influences optimal insurance demands as the government raises spending to combat the infection. Additionally, we estimate the sign of cross-derivative using nationally representative repeated cross-sectional survey data by creating prefecture-level pseudo-panels of insurance demand. The empirical results on changes in insurance demand following the onset of the COVID-19 pandemic are consistent with correlation-averse preferences over wealth and health.