Last March I summarized 8 years of day versus night stock movements. This note 1) adds another year to that analysis, 2) shows year by year returns rather than just a summary and 3) tests for significance of differences. The day vs night effect was reported by Cliff, Cooper and Gulen's 2008 paper "Return Differences between Trading and Non-trading Hours: Like Night and Day." They reported only on the US. This note shows the effect around the world and finds, surprisingly, that the effect has been much weaker in the US during the last 10 years.
Below we see open and close time for 4 world centers - US, Europe, China and Japan - which represent 75% of the world’s GDP. To estimate day and night returns in these centers I use very liquid US-traded ETFs: SPY for the US; VGK for Europe; EWJ for Japan; FXI for China. Consider China which is open from 9pm to 4am. FXI's return from 4pm_9:45am approximates the China open period while FXI’s change 9:45am_4pm approximates the closed period( 4am_9pm). The same holds with Japan. VGK change from 4pm_11:30am approximates Europe open from 4am_11:30am, while VGK from 11:30am_4pm approximate Europe closed from 1130am_4am.
This ETF-based method finds China,Japan and Europe consistently dropping while open and rising while closed. Specifically in only 1 year, 2018, out of 10 did Europe and Japan do better during the day than overnight. In only 2 years (2014 and 2017) out of 10 did China perform better during the day. Only the US had 4 years (2010, 2012, 2016 and 2019-to-date) better during the day than at night. The first Bar Chart is the most important showing night-day differences. For those wanting details the second chart shows day and night returns.
Except for the US the results are highly significant.
But can we capitalize on this phenomenon? No because of trading costs which are dominated by bid-ask spreads. FXI which represents China trades at 35 dollars with a bid-ask spread of a penny, ie .03%. If we bought at 945 and sold at 4pm there’d be 2500 buys and sells over 10 years costing 75%. Since the net profit from FXI was 100% most of the profit would go to trading costs with a residual return of only 2% per year, ie buying Treasury bills would have been a better play. This strategy fits Eugene Fama’s efficient market qualification which was - Some strategies beat the market but not when trading costs are considered.
Although I don’t currently see how to profit from this I’ll continue searching with filters such as VIX level and day of week. As always suggestions are more than welcome.