Financial MarkeTWEETs and BOTS

Communication has changed. That has been said many times. Let’s think about financial markets, these models are founded on information, the faster they can get information the more efficient they could be. Prices and rates are constantly changing because of the flow of new information, which may influence future forecasts about businesses, investments, or countries. Throughout history, many great people, such as Presidents, or CEOs had influenced markets with their statements, reported through traditional media. As gatekeepers, the purpose of media is to drive information to enhance the comprehension of these declarations. 

What about Twitter and direct communication? What if, without any context, Elon Musk wakes up one day and gives an opinion about bitcoin or shares his overrated view of his own company? Or what about the endorsement to some public company of someone in the spotlight (not our beloved Uncle Jimmy and his Twitter account, of course). These are some interesting developments for investors, due to these new forms of “market whisperers.”

At first, let’s distinguish these from “ordinary financial influencers”, who may have their followers, but the impact is surely different. We are talking about companies or political leaders. Some may observe these powerful people on Twitter and Facebook set up trading strategies. And some may point, also, that there is the risk of stock price manipulation, which would be bothering from the point of view of the SEC (Securities and Exchange Commission), a government regulatory agency for markets. 

How can we deal with this flow of “fresh information”? Well, technology is improving every day more and more, reaching the point where one can teach the computer how to learn. This is the so-called machine learning, where there are algorithms that improve automatically through experience as a part of artificial intelligence.

Moreover, there are interesting tools called “bots”, software programmed to do a certain set of tasks on their own, such as surf the Internet. Why are these relevant to this discussion? Well, by running automatically and browsing the Internet they can immediately scrape new information on Twitter and within seconds react to “anticipate” the rest of the operators that are running to the new opportunity due to the statement’s release.

Back when Donald Trump was President of the United States, and he used to type something on his Twitter account every now and then, a bot was created with the purpose to analyze his communication called BOTUS (from POTUS, the acronym of the official Twitter account of the President of the US) and make sound investments based on his own tweets.

Let’s see how such a tool can work. These bots can get into Twitter and read several amounts of information. Making an example, by reading tweets of Elon Musk or the former President Trump. But then you have to tell the machine how to think, which is a step-by-step process.

The first one is called “sentiment analysis”, the capability of the machine to detect words connected to emotions and judgments. Once something interesting comes up, the next step is to understand which company the tweet is referring to. It may appear even easier to get a company’s name instead of sentiments, but it is not always crystal clear, sometimes it is needed to read the context, relying on linking words. An apple can be crunchy and good, while the Apple company can be good as a game-changer, but are they both worth the investment of some hundreds of Euro in stocks? Last but not least, sometimes humans need to step in the automatic process to solve these puzzles, by setting some rules about words that must not be taken into consideration. Speaking about Trump, a decision rule may be not to take into consideration the word “Tiffany”, the name of his daughter and a company name.  

For a deeper dive into twitter bots, i’d recommend listening to Planet Money’s podcast on this topic.

Market gurus are nothing new, as well as statements by powerful people and influencers. By setting up rules and algorithms, teaching the machine how to learn and think, these invisible friends can get money jumping into the rally as soon as it opens, then closing the position a couple of hours later. Here we can assist to another very powerful usage of the combination between new technologies and new communication media.