I need to test whether stock price changed statistically significant after certain event. I got time series of stock price before and after that event. What test or methodology should I use? (not very sophisticated if possible) Is it correct to use t-test?
2
-
$\begingroup$ I realise you state not very sophisticated, but we do need more information to understand what you need. Are you wanting to determine if a short duration spike is different to before and after or are you wanting to see if a shift in local mean is significant, i.e. did the event have a lasting impact? If this is part of training/learning please tag with self-help as this helps alert people to provide answers structured to assist with study. $\endgroup$– ReneBtCommented May 14, 2018 at 8:32
-
$\begingroup$ I think the event has lasting impact, but I do not know how to check it statistically. Thanks, I will add tags $\endgroup$– AlexanderCommented May 14, 2018 at 9:06
Add a comment
|
1 Answer
$\begingroup$
$\endgroup$
The methodology you ask is called event study. Though, it usually incorporates having multiple events so that you can make some sort of inference on abnormal returns.
If you want to check the significance of the price change after just one event, what you can do is to test it against the distribution of daily returns. If it is close to the tails, then you could argue that it is not likely to have happened by chance and it has to have something to do with the event. But you would be on thin ice in this case, since any event happened on that day could have caused it, not just the event you think it did.
