Context is everything - and we were reminded of that by last Thursday’s ECB interest rate announcement and subsequent press conference when a heap of seemingly bad news was shrugged off by the euro in the space of a few hours.
Why would this happen? Well, I’m no trader, analyst or economist, but I would imagine that as everyone was expecting a swathe of bad news for the euro area, a) they were already short, or b) they knew the market was short and were waiting to see what happened. Either way, even I understand that if everyone’s overly negative, the market can’t go down much further.
But what’s this got to do with news Analytics ? Well, I saw someone tweeting that it would be useful to have a low-latency connection to Mario Draghi’s press conference. As a participant in the news analytics market, I wondered how it could be achieved. And would it help clients?
So, could it be achieved? There are already low latency connections to news vendors like Thomson Reuters, Bloomberg & Dow Jones. These are servers belonging to trading firms co-located with the server delivering text news (not the low latency feeds of economic data that have been subject to much press coverage recently). But that wouldn’t solve the problem, because you still have to get super Mario’s voice from the press conference into text – and although that technology is pretty good, I’m not sure it’s good enough yet.
So you’re probably still better off watching the press conference live if you want to trade the event.
Would A Low Latency Delivery Of The ECB Presser Be Useful To End-Users?
But that brings us onto the second point? Would a low latency delivery of the ECB presser be useful to end-users. Yes, to the small portion of the market that actually trade the economic events. Of course, they would also need a handle on the context – market positioning and orders being the most important – and a good source of that data is pretty hard to come by for the average investor.
But as we know, most traders stay out of the market around key scheduled economic events because it’s a bit of a lottery. So, would news analytics be useful to this latter group of, perhaps not-so-brave - or foolhardy, traders. Absolutely. (But they wouldn’t need the low latency.) Why? A couple of reasons…
First they’ll have a better handle on context – all the news leading up to the event should paint a picture of current market sentiment, a proxy for how the market is likely to be positioned, at the time of the event. Negative news in a negative sentiment environment is likely to have less of a downside impact that negative news in a neutral or positive environment.
Second, a systematic approach means, obviously, there’s no discretion going into the trading decision. Discretion means a likely bias based on a few hyperbolic headlines that capture a disproportionate amount of the market’s attention. News analytics enables the user to see the whole picture – millions of news articles - not just the few with the most sensational headlines.
So, if you knew market sentiment was overly negative going into the ECB’s announcement and press conference, you would ‘ve stayed out of the market during the event and waited for the bounce – or positioned for the bounce beforehand.
But there’s nothing earth-shattering here. Stuff most people know about – short squeezes etc. However, it’s not just one-off events that exhibit this behavior… One of my colleagues did some quant research into FX price moves based on news sentiment on a daily basis and it turns out prices, on average, revert a day later. So, a day full of negative news (and likely price fall) will, on average, be followed by a higher price day. That is, the Forex market is prone to short-term over-reaction to news sentiment. But over the longer term, the market moves in the direction of sentiment.
And that raises an interesting question. Is it worth only trading on yesterday’s news today?
