Economic Modelling, cilt.102, 2021 (SSCI)
This paper investigates the predictive content of news-based advanced market, regional, and global economic policy uncertainty (EPU) measures for bond spreads and their volatility in emerging markets (EMs) by extending the higher (k-th) order nonparametric causality-in-quantiles test to a multivariate case. Results show that global and advanced market EPU measures have predictive power for EM bond spreads in the lower and upper quantiles while for volatility, the predictive power is stronger in the upper quantiles and further observes predictability in the mid quantiles. Predictability detected for all EMs is characterized by an inherent heterogeneity leading to an asymmetric pattern over the distribution of EM bond spreads and volatility. The implication for heterogeneity in our results is that when EPU is high in advanced markets, global investors’ appetite for the EM local currency bonds increases due to high yields. However, when EPU is low, global investors move out of EMs because of the perceived unsafe investment environments in EMs.