U.S. uncertainty and Asian stock prices: Evidence from the asymmetric NARDL model

Abstract

When uncertainty reduces spending among U.S. consumers, it may affect the bottom line stock performance of Asian producers that cater to their needs. Theory predicts that the impact of uncertainty will be asymmetrical: during the two phases of the business cycle, countercyclic shocks will outweigh procyclic shocks, resulting in phase-specific equilibrium price adjustments. We conjecture that relative to recessions, recoveries bring larger long-run price adjustments, a response to pent-up growth potential. This is an extension of existing theories, which predict that recoveries bring overshooting, a transient reaction to pent-up demand. We test for these asymmetric uncertainty effects on 11 Asian stock market indices over the 2000M08 – 2017M02 period. Our independent measures include the economic policy uncertainty index (EPU) of Baker, Bloom, and Davis (2016), the Chicago Board Options Exchange implied volatility index (VIX), and the financial uncertainty indicator (JLN) of Jurado, Ng, and Ludvigson (2015). To characterize asymmetry, we employ the nonlinear autoregressive distributed lag (NARDL) model of Shin, Yu, and Greenwood-Nimmo (2014), in which both short- and long-run nonlinearities are captured through positive and negative partial sum decompositions of the explanatory variable(s). Using the NARDL output, we test three hypotheses. The first, that increases in uncertainty (decreases in uncertainty) result in stock price drops (stock price rises), is broadly supported by our analysis. The second, that equilibrium adjustments following negative countercyclic uncertainty shocks exceed those following positive movements, is supported fully by the EPU analysis and partially by the VIX and JLN analyses. The third hypothesis, that recoveries are characterized by overshooting, is consistent only with the behavior of the Chinese stock responses to EPU and VIX shocks. Our results demonstrate the advantages of the NARDL model in characterizing asymmetry. They suggest that while long-run asymmetry is fairly consistent across countries, short-run asymmetry is more country-specific.