The way this study was conducted :
I have labeled the 23rd of each year, the day before Christmas Eve. In some cases, the labeled day differs because the 23rd was a weekend day. I have measured the gain from the swing low before the 23rd, to the close on the 23rd. On average the swing low is on 17-18 December. The results are:
23 Christmas holidays analysed
1 period ignored in 2008
2 periods resulted in a loss
20 periods resulted in gains
After doing that, I analysed the last trading day before Christmas, from the close of the previous day, to the close of the last day. In order to take advantage of the whole move, you would have to trade futures.
23 days analysed
10 days ended the day with a positive gain
9 days ended without any significant change ~ +-0.10% (7 on plus, 2 on minus)
4 days ended with a loss
Conclusion : Picking the low is impossible, however, going long around 20th before Christmas proved an effective strategy; buy as low as you can around 20th, sell before the close on the last trading day before Christmas (usually 23rd). Going long on the close of the 22nd also proved a winning strategy, however many times there wouldn't have been any gains made.