Sectoral spillovers and systemic risks: Evidence from China

The recent collapse of Silicon Valley Bank prompts interest in how risks are assessed. We develop a new approach to estimating the systemic importance of sectors, integrating a 'top-down' and 'bottom-up' informed spillover network through component-expected-shortfall and generalizederror-variance-decomposition methodologies. This approach identifies systemically important sectors by combining too interconnected to fail and too big to fail logics. We find that levels of risk spillovers in respective sectors are not typically well correlated with corresponding levels of risk contribution. Consequently, focusing only on connectedness between assets seriously distorts risk estimation. Results are significant for regulatory authorities to accurately identify sector risks.