would be interesting to have a look at these papers even if it looks to be grossly oversimplified approach...
The most quick bloomberg news comes from its terminal. one must have that terminal to read these news. the news doesn't come from web. Thanks
I think the paper you mentioned is very useful. I am very interested in the result. when these is regular economic release, as too many people are watching them, the stock price will react very quickly. For example, one can use excel or other program to read bloomberg ecomonic number in realtime. As too many people are competing, there is little profit left. I think the chance comes from these unregular, or surprised news release.
meaning that you need to define in advance what these unexpected news will be.....which can be a hard thing to do because they will be unexpected after all...
Even trading the news with automation has been done for years. The word "buyback" in a story instantaneously generates buy orders. Look at the tick by tick data on story stocks. The stocks are bought (or sold) to higher (or lower) levels before the human mind can comprehend a story.
Here are the papers. It has been a while since I read them, but from what I remember speed is not a huge issue, there is some long term over or underreaction. Also, I think they don't parse the news, they infer the content from the market reaction or something like that.
Thanks, something to think on. If anyone else is willing to post any papers on this topic that would also be very welcome.
No problem. If you look up the Chan paper in Google scholar it will give you a list of all the papers that reference it.