The paper just says Day 0 is earnings announcement day. Are they just using calendar date? Unfortunately this would obfuscate a more precise conclusion since now we’re co-mingling days leading up to post-close announcement and days following pre-open announcement.
The dynamics of earnings are interesting. It could be true that close to open jump vol is over-priced, but once the day move is factored in, realized outpaces implied. One of the things I first noticed trading these is that much less implied vol was reserved for the post-announcement intraday move than I would’ve expected.
Maybe we should start looking a bit closer at jumps that seem close to fair price; buy those instead of passing on them in order to hold through post-earnings intraday move. I’ve been focused mainly on the jumps thus far but this paper will get me thinking a little more about the day moves.