Risk spillovers between the financial market and macroeconomic sectors under mixed-frequency information: A frequency domain perspective

We construct the Mixed-Frequency Vector Autoregression of Frequency Domain Decomposition model, labeled as MF-VAR-FDD, to investigates the risk spillovers between financial market and the macroeconomic sectors. The MF-VAR-FDD innovatively introduces the spectral density function into the Mixed-Frequency Vector Autoregression (MF-VAR) model, providing a spectral representation of the generalized variance decomposition of the MF-VAR model. It decomposes time domain risk spillovers into high-frequency and low-frequency components, constructs the Mixed-Frequency Frequency Domain Spillover (MFFDS) index, and characterizes risk spillovers between finance and macroeconomic sectors at different frequencies from a frequency domain perspective. The research findings are as follows: First, during major crises, frequency domain risk spillover significantly intensifies, exhibiting notable time-varying characteristics. Moreover, risk spillover is mainly dominated by low-frequency spillover, while high-frequency spillover is more sensitive to crisis events, enabling more agile monitoring of market disturbances and aiding in risk prevention. Second, finance consistently represents a net risk exporter to all macroeconomic sectors, emphasizing that financial risk should remain a key focus for national regulatory authorities. Our conclusions provide additional information for the formulation of macroprudential policies.