The causal relationship between economic policy uncertainty and stock indices in OECD and non-OECD countries: evidence from time-varying Granger causality tests on a lag-augmented VAR model

ABSTRACT

This study re-examined the causal relationship between economic policy uncertainty (EPU) and stock indices for Organization for Economic Co-operation and Development (OECD) and non-OECD countries. To this aim, it applied the lag-augmented vector autoregression model with a time-varying Granger causality test . Based on the standard Granger causality test results, only afew countries demonstrated acausal relationship between two variables. However, the empirical findings using of time-varying Granger causality test indicate that the causal relationship between the two variables is not present for the entire period, but it is present for some sub-periods in many countries. Additionally, the results found that causality running from stock indices to EPU (stock price-leading hypothesis) for many periods and countries. The results indicate that the relationship between two variables is not constant, meaning that it changes from one period to the next in many countries.

KEYWORDS:

• Economic policy uncertainty
• stock indices
• time-varying Granger causality
• lag-augmented VAR model
• OECD countries

JEL CLASSIFICATION:

• C32
• C50
• E60
• G10
• G18

Acknowledgments

I would like to thank the anonymous referee and the Editor for helpful comments

Disclosure statement

No potential conflict of interest was reported by the author.

Supplementary material

Supplemental data for this article can be accessed here.

Notes

¹ For detailed results, see supplementary material table A3. Figure A1 in the supplementary material illustrates the Wald statistic sequences and bootstrapped 5% critical values (controlled over aone-year period) for the three algorithms.

Additional information

Funding

This work was supported by JSPS KAKENHI Grant Number JP21K01532.