Machine readable news, markets and coronavirus

Sometimes, when I have a burger, it’s the unexpected ingredients which truly make it special. Perhaps it’s a fried egg, perhaps it’s the sauce, maybe the fried onions? Of course, the “core” of a burger, namely the patty and the bun have to be good too, but it’s the extra special touches, which can also make a big difference. 

Financial markets continually try to digest information. Again, it is often the unexpected news which is going to have the biggest impact (ok, I admit that’s a fairly obvious statement!) Sometimes markets can react in odd ways to the information. One thing that I’ve mentioned before is that sometimes you might well forecast an event perfectly, but still still money on it. The market may have ignored that piece of news, or reacted in a very odd way to it. Furthermore, the market may try to digest information from areas where, market practitioners know very little about. The recent market selloff from the coronavirus is an example of such an area. I doubt many folks sitting in markets have a medical background (and I certainly don’t). As a result, market practitioners will try to hoover up as much information as they can about the area. How can an investor make sense of all this information?

There’s an interesting post by Richard Peterson (who does have a medical background, as well as one in finance, so one of the rare few!) looking at news sentiment, measured with his MarketPsych indices (based on news from sources including Reuters) around the recent coronavirus spread, alongside recent market moves. There are many ways of looking at the impact of news, ranging from news sentiment to news volume. If we were to look at news volume which contained the term “coronovirus” that have been published by Bloomberg News, through much of January there was relatively little volume. In February, there was a marked picked up in the volume, another leg higher on 24 February, which was also the time that stocks really started to sell off. Perhaps unsurprisingly, the news volume has increased event more since then. If you have a Bloomberg terminal you can create your own news volume study by going to NT<GO>.

Whilst such a news volume indicator is relatively simple, it does nevertheless show us on aggregate what’s going on, and also against history: indeed a time series is a lot more useful than looking at the single point. Obviously, for discretionary investors, using such news volume (or sentiment analysis) doesn’t mean they don’t ever need to read new themselves. However, having these more automated ways of reading news, can be a useful addition in the investment process, alongside reading news yourself.

Most importantly, hopefully, we’ll win the battle against coronavirus soon.