Regime-dependent causality between Chinese and U.S. equity markets: Evidence from Markov switching models

Abstract

This study analyses the dynamic interdependence between Chinese and U.S. equity markets using Markov regime-switching vector autoregressive (MS-VAR) models. We cover monthly data from May 2007 to August 2024 within a single-currency framework on the New York Stock Exchange. Two broad-based Exchange-Traded Funds (ETFs)—SPY for the U.S. and GXC for China—serve as proxies for equity returns. The study identifies two regimes: Regime 1, characterised by periods of crisis, and Regime 2, representing stable market conditions. Causality tests based on the MS-VAR model reveal bidirectional causality between the markets in Regime 2, a relationship not detected by conventional Granger tests. In Regime 1, causality is unidirectional from the U.S. to China, indicating that Chinese investors are more exposed to U.S.-driven shocks during periods of market turbulence. The findings suggest the need for collaborative strategies to reduce market risks, address vulnerabilities, and manage spillovers arising from geopolitical tensions.

keywords: Causality, China, Exchange-traded funds, Regime-switching vector autoregressive, The U.S.