Last year, Goldman decided to take a look at what matters for markets when it comes to nuclear war: actual missile launches and nuclear tests by North Korea or Donald Trump tweeting about those missile launches and nuclear tests.
You’ll never guess what they found. To wit:
The regression results are fairly clear: tweets matter, testing does not. Indeed, from a statistical perspective, testing does not appear to matter at all unless President Trump tweets about it.
That’s a testament to something, although I’m not entirely sure what.
While I don’t want to extrapolate too much, the results of that study suggested that the market viewed the President of the United States as less rational than Kim Jong-Un when it comes to pushing the proverbial “button”. As a reminder, Trump’s “button” is “bigger, more powerful” and also “works!”
Well, fast forward six months and Trump claims he’s “solved” the North Korea problem, a contention that’s debatable, to say the least.
But while markets may not have to worry about “button” tweets anymore, Trump’s Twitter feed is still the main source of headline risk for a lot of assets, an absurd state of affairs befitting of the surreal reality in which we’re now forced to live and trade.
So, Goldman decided to conduct a similar study that looks at the effect of Trump’s tweets about trade.
“We have previously shown that when US-North Korea tensions were rising amid continued missile tests last year, President Trump’s tweets seemed to matter more for global risk appetite than the missile launches per se,” the bank writes, rehashing the study mentioned above, before explaining why they need to go through this exercise again:
With trade tensions now in focus, both for markets and President Trump, we can replicate this analysis for tariff concerns. We constructed a daily count of the number of times US President Trump’s Twitter account referenced international trade, this time using only the words “trade” and “tariff”, and once again calculated a 3-day moving average. Exhibit 1 plots this 3-day average of President Trump’s trade-related tweets against the VIX. While both Twitter mentions and the VIX have been relatively elevated this year, on visual inspection, there does not seem to be any obvious relationship.
An accompanying regression analysis confirms the lack of a relationship between implied volatility and Trump’s trade tweets.
Of course the issue here could be that in the case of the nuclear threat, Kim’s actions were seen merely as a continuation of the behavior that markets have been dealing with for years when it comes to Pyongyang, while it was Trump that was the new variable. With trade tensions, on the other hand, it’s the actual imposition of tariffs and the publication of lists associated with those tariffs that matters more than Trump’s editorializing via tweet.
Whatever the case, Goldman does note that there is one asset that’s seemingly correlated to trade bombast emanating from @realDonaldTrump:
The only asset for which Trump’s trade-related tweets appear to have any significance is soybeans. Exhibit 3 shows the results of regressing daily soybean returns on our index of tweets while controlling for the Economic Policy Uncertainty Index and the US Dollar.