https://orcid.org/0000-0002-8792-9263
![Sample our Area Studies journals, sign in here to start your access, latest two full volumes FREE to you for 14 days]({"type":"image" , "src" : "/sda/5922/TOC-Subject-Sample-2021-Area-Studies.jpg"})
ABSTRACT
We address in this study the issue of how to proxy sleep and explore sleep’s significance for financial markets. We employ daily Google search activity on sleepiness terms (e.g. sleep deprivation) to develop an index and find that a one-day lagged sleepiness index is related negatively to US stock market returns. When investors lack sleep, stock market returns are relatively low. This pattern could be explained by sleep deprivation causing an increased level of investor anxiety and risk aversion. We find that this relation is most pronounced on days with high uncertainty in the market. Sleep is negatively related to stock market returns even after controlling for sentiment. Overall, our results highlight the application of Google Trend in a new field showing that investors’ sleep patterns influence their investment decisions.
Acknowledgements
I am grateful for valuable comments from two anonymous reviewers, and also for comments received from Betty Wu, Chris Veld, Evangelos Vagenas-nanos, Hao Li, Hue Hwa Au Yong, Jones Edward, Ioannis Tsalavoutas, Ken Klassen, Konstantin Kamp, Matthias Pelster, Stefano Alderighi, Stefan Ruenzi, Valentina Lagasio and the conference and seminar participants at the 2018 Heriot-Watt University, the 2019 Financial Engineering and Banking Society, the 2019 Financial Management Association European Conference, and the 2020 Stirling University.
Disclosure statement
No potential conflict of interest was reported by the author(s).
Notes
1 Studies in asset pricing literature (e.g. Zahirovic and Okicic Citation2016) have previously shown that risk aversion is time varying and that increases in risk aversion are related with relatively low stock returns.
5 They first download Google search activity for a term in monthly frequency that all values during the sample period can be downloaded at once. They then use these monthly values to scale daily Google searches for the same term.
6 https://www.esrs.eu/media/glossary-of-sleep-related-terms.html?no_cache=1&tx_mmdictonary_pi1/%5Bcapital/%5D=A.
7 https://www.quora.com/How-does-Google-suggest-exactly-what-I-want-to-search. Lippi and Torroni (Citation2016) offer a detailed review of available machine-learning techniques.
8 Insomnia, insomniac, jet lag, jet lag calculator, jet lag cure, jet lag pills, jet lag tips, lack of sleep, lack of sleep headache, list of sleep disorders, sleep apnea, sleep apnea causes, sleep apnea devices, sleep apnea machine, sleep apnea mask, sleep apnea mouth guard, sleep apnea surgery, sleep apnea test, sleep apnea treatment, sleep deprivation, sleep deprivation causes, sleep deprivation death, sleep deprivation effects, sleep disorders, sleeping pills, sleeping pills and alcohol, sleeping pills for dogs, sleeping pills names.
9 In untabulated results, we explore the relation between lagged changes in Google search activity for each sleepiness term and stock market returns. We find that the sign of the main relation is negative in 20 (out of the total 28) of these terms.
11 We report later in this study the main result when winsorizing sleep data at the top and bottom 5%.
13 Note that Siganos, Vagenas-Nanos, and Verwijmeren (Citation2014) explore the relation between Facebook’s sentiment and stock market returns for several international markets. This study explores only the US market.
14 http://aa.usno.navy.mil/data/docs/RS_OneYear.php https://en.wikipedia.org/wiki/Geographic_center_of_the_United_States\#/media/File:USCenterPop_Geographic2010.png.
15 We acknowledge that VIX has also been used in the literature as a proxy of market risk and investor sentiment (e.g. Smales Citation2017), not only as a measure for uncertainty.