VERY interesting question from the thread starter. You'll get some surprises (and great insights into market behaviour) with your research. And you'll stunned to discover that what we call "trends" is often created through overnight moves only. (this is HUGE, and only 1/1000 so-called traders know it). Don't miss the nasdaq in the last 15 years. You might discover NEGATIVE CORRELATION between general trend and intraday trend on quite long periods. Best, Sarah NGuyen
(sorry, just a second post to get the email notifications about this thread, it deserves them) Cheers
The more recent data is just confirmation that overnight returns in bear markets tend to be negative, whereas in bull markets they tend to be positive. No real surprise there. Neutral/negative intraday returns suggest that almost all returns come from overnight "surprises" & news, and that nothing happens intraday to give new & meaningful information to stocks. If you bear in mind that earnings, takeovers, and other meaningful fundamental data are usually released before or after the closing bell, this makes a lot of sense. However, dividends should accrue each day, not just overnight, so this part is a bit of anomaly. Dividends from 1993-2007 have been low, however.
I just recalculate the perf on SPY since 1993 , adjusting for dividends on ex div dateand the total still doesn't make sense,as overnight gains + intraday losses + starting value of spy coming to well over 200. Any idea why ? using the forumlas above, shouldn't we get : overnight performance + intraday performance +SPY starting value = today's SPY value ?
Sorry, can't remember, but it probably had to do with me misusing the spreadsheet. From memory, last results I got (which seemed correct, contrary to the first ones ) pointed at better intraday performance than overnight in the recent past - up to May 2012.
Michael Harris just posted an update on this issue : http://www.priceactionlab.com/Blog/2012/12/spy-profits-due-to-overnight-gaps/