It is no secret that President Donald Trump likes to tweet. In fact, there have been over 14,000 tweets since he took office in 2016, that’s an average of over 10 per day! He is known to tweet about anything on his mind from “crocked” Hilary’s mails to threatening countries with nuclear war “I too have a Nuclear Button, but it is a much bigger & more powerful one than his, and my Button works!” to more market moving topics targeting specific companies or the ongoing US-China trade war “When a country is losing many billions on trade, trade wars are good, and easy to win. Example, when we are down $100 Billion with a certain country and they get cute, don’t trade any more-we win big”. Have you ever wondered what the actual economic impact of his tweets are? Well, that’s exactly what JP Morgan aims to do with their new Volfefe index, named after another mysterious trump tweet “Cofefe”.
Trump has increasingly been tweeting about forces that have an impact on the market. In an interview by JP Morgan lead analysts Josh Younger and Munier Salem say “Trade and monetary policy have become an increasing focus for the executive branch, and everything from casual sentiments to seemingly formal policy intentions have been disseminated, globally and instantaneously, via this carefully scrutinized social media platform. In response, a broad swath of assets from single name stocks to macro products have found their price dynamics increasingly beholden to a handful of tweets from the commander in chief.”
According to the study, Trump is most likely to tweet between Noon to 2:00 pm US Eastern Time. While he is three times more likely to tweet at 1:00 pm than any other time, the president is also known for his late-night twitter rants. This can be problematic since the depth in the overnight market tends to be fairly thin. The report also concludes that the president is presumably asleep from 5:00 am to 10:00 am, from the lack of social media activity on his accounts during that time. Probably awoken by the smell of his favourite McDonald’s Burger. The index, to begin with, is concerned with US interest rates on the shorter end of the curve, i.e. 2-5-year rates volatility. As the impact of tweets on policy and trade related matters would likely have a higher impact on the shorter duration. However, JP Morgan believes the model can be replicated easily for equity “The stock market is up massively since the election” or currency markets “It’s called currency manipulation”.
For the methodology, the analysts at JP Morgan analysed the database of the 14000 tweets and observed the direct correlation to the bond markets within the first 5 minutes of the tweet. This helped isolate topics that seemed to have an impact on the markets, which included themes or topics related to trade and monetary policy. They subsequently identified words most frequently occurring in these market moving tweets. The results of this showed that that top market moving words focused more on trade, rather than the monetary policy. The top words identified are China, Billion, Products, Inflation, economy and Reserve. “I capitalize certain words only for emphasis”. This led to the construction of the index to quantify and monitor the shifts in markets. They used the information generated above to train a machine learning tool to identify which tweets will move the markets and have flagged 146 of the over 4000 tweets since 2018 as market moving. It should be noted that this index, reflects the change in volatility of US interest rate and not the movement in rates themselves.
Meanwhile analysts at Citi Bank have also created a “Trump Tweet Tracker” which tracks the impact of the President’s tweets on the US-China trade war on the S&P 500 and the Shanghai Composite.
In an era of uncertainty and increased volatility, where Presidents are armed with ammunition of 140 characters, are these innovative solutions the need of the hour, or it is simply a marketing gimmick? We could wait to see how accurately these models are able to predict volatility or we may need a “very stable genius” to figure it out.