Effect of uncertainty on U.S. stock returns and volatility: evidence from over eighty years of high-frequency data

ABSTRACT

We explore the asymmetric high-frequency daily response of U.S. equities to financial uncertainty over the 1936–2016 period. We find positive growth of uncertainty reduces stock returns and increases volatility, while, a negative growth of uncertainty primarily reduces variance. More importantly, the impact of uncertainty on volatility is found to be asymmetric. We also model rolling window estimation and find significant time variation in the impact of uncertainty, though the direction largely confirms with the static case. Our study provides new evidence that the response of U.S. equities to uncertainty is intuitively consistent even in the historical and high-frequency context.

KEYWORDS:

• Uncertainty
• stock returns and volatility
• asymmetry
• rolling estimation

JEL CLASSIFICATION:

• C32
• G10

Acknowledgments

We would like to thank two anonymous referee for many helpful comments. However, any remaining errors are solely ours.

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

¹ The data is downloadable from: http://mba.tuck.dartmouth.edu/pages/faculty/ken.french/data_library.html.

² The historical FOMC dates are available at: https://www.federalreserve.gov/monetarypolicy/fomc_historical_year.htm.

³ Our results were qualitatively similar, if we just used the changes in the interest rate on FOMC dates without filtering the recessionary impact. Complete details are available upon request from the authors.

⁴ The data is available for download from: https://www.rbnz.govt.nz/research-and-publications/research-programme/additional-research/measures-of-the-stance-of-united-states-monetary-policy.

⁵ The process allows one to answer the question: ‘What policy rate would generate the observed yield curve if the policy rate could be taken as negative?’ The shadow policy rate generated in this manner, therefore, provides a measure of the monetary policy stance after the actual policy rate reaches zero.

⁶ The data can be downloaded from the website of Professor Jorge M. Uribe at: http://www.ub.edu/rfa/uncertainty-index/.

⁷ To construct the uncertainty shocks, we define a dummy variable which takes the value of one when the growth in uncertainty is positive (negative) and zero otherwise, and then multiply the growth in the uncertainty variable with this dummy to get the corresponding positive (negative) innovation in uncertainty.

⁸ While, FOMC meeting dates are not available prior to 18 March 1936, daily data on interest rates (i.e., the risk-free rate), is indeed available from Professor French’s data library for the period of 6 January 1927 till 17 March 1936, as are the other variables of interest from their respective sources. Given this, we take first-difference of the risk-free rate over this additional period as monetary policy shocks, and merge with our existing data set, and then re-conduct the analysis again. The results have been now reported in Table A1 in the Appendix of the paper. As can be seen, our results are qualitatively similar to those reported in Table 1 (except now mps^pos does not have a significant negative effect on stock returns, but reduces volatility significantly). In other words, the effect of uncertainty shocks is robust, irrespective of how monetary shocks are measured.

⁹ Tests of serial correlation and heteroscedasticity on the standardized residuals for both the short (Table 1)- and long-samples (Table A1) failed to reject (with p-values of all the statistics close to or equal to 1.0000) the null hypothesis of no serial correlation and no heteroscedasticity, suggesting that our model is correctly specified and the results are reliable. Complete details of the Ljung-Box test on the standardized residuals, squared standardized residuals, and the ARCH-LM test are available upon request from the authors.

¹⁰ However, our results are qualitatively similar to a bigger size of the rolling window involving 10 years of data. Complete details of these results are available upon request from the authors.