Goldman: Trump’s ‘Rocket Man’ Tweets Matter More For VIX Than Actual Missiles

We’ve reached the foreign policy/social media/VIX singularity, folks.

Heisenberg has long bemoaned the fact that Donald Trump’s increasingly unhinged Twitter feed is required reading for market participants. But alas, that’s the world in which we are forced to live and trade.

Whether it’s rants about Amazon not paying taxes, threats to dismantle Obamacare subsidies, feuds with key lawmakers that have the potential to undermine his pro-growth agenda, or outright calls for nuclear war, the President’s social media tantrums have the potential to disrupt markets without warning, leaving bewildered investors to pick up the pieces.

Because there’s no filter and because he’s confined to 280 characters (140 for most of his young presidency), it’s virtually impossible to sort out the implications of any one tweet and the situation is made all the more tenuous by the fact that in the vast majority of cases, Trump is rage-tweeting without regard for the consequences.

Earlier this year, we noted that “tweet risk” had become part of the Wall Street lexicon, a fact that was underscored by the following screengrab from a Morgan Stanley note about defense stocks:

Well on Friday evening, Goldman is out with a truly hilarious piece which takes a deep-dive into the following question. What matters more for the VIX: actual North Korea missile tests or Trump’s Twitter reaction? And yes, I am serious.

“In a year during which the VIX otherwise had the pulse of a rock, North Korean missile launches appeared to be one of the few sources of volatility [and] we find evidence that it is the US reaction to North Korean testing — and President Trump’s Twitter account in particular — rather than the testing per se, that matters most for global risk appetite as reflected by the VIX,” Goldman writes, introducing their study. Here’s an excerpt:

The market’s recent indifference caused us to wonder: what matters most for the market’s reaction? The test itself or President Trump’s reaction on Twitter?

To measure the latter, we constructed a daily count of the number of times Trump’s Twitter feed referenced North Korea using key phrases such as “nkorea”, “north korean”, “rocket man”, “kim jong”, etc. Then to deal with weekend tweets, we calculated a 3-day moving average. The pattern visible in Exhibit 1, which plots this index against the VIX, suggests a strong relationship between Trump’s tweets and the VIX.

And if you’re inclined to say that’s not enough in the way of evidence, Goldman took this several steps further. Specifically, the actually regressed the VIX on its own lag along with:

• dummy variables equal to one if North Korea has tested a missile or nuclear device in the past three days
• the Economic Policy Uncertainty Index (as a crude proxy for policy uncertainty more generally)
• and our constructed index of President Trump’s North Korea Twitter mentions

They even tried a few variations on this using alternative windows and variable subsets. Here are the results:

In case you can’t interpret that, allow Goldman to spell this out in the most unequivocal terms possible:

The regression results reported in Exhibit 2 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. 

For investors trading in the VIX around such events, our results have a straightforward implication. Namely, that to judge how the VIX will react to the next North Korean testing, it will likely be more important to anticipate Trump’s Twitter reaction than to judge the implications of the test itself.

There you go, folks. This is actual, statistical evidence to suggest that the market is far more worried about the possibility that Trump will fly off the handle than it is about North Korea conducting actual ballistic missile tests.

Coming full circle: we’ve reached the foreign policy/social media/VIX singularity.