Behold! JPMorgan Unveils The ‘Volfefe’ Index

“Trade talk, political campaigning and tweets have contributed to volatility, from China to Fed policy to tax policy”, BofA’s Savita Subramanian and Jill Carey Hall wrote earlier this month, in a note discussing policy uncertainty “and the Tweet factor”.

As it turns out, “heavy tweet” days for Donald Trump are associated with negative returns for US equities. Imagine that, right?

“Since 2016, days with more than 35 tweets (90th percentile) by President Trump have seen negative returns (-9bp), whereas days with less than 5 tweets (10th percentile) have seen positive returns (+5bp)”, Subramanian and Hill observed.

Read more: The More Trump Tweets, The Worse Off Stocks Are, BofA Shows

Those results were statistically significant, but in case you somehow weren’t convinced, JPMorgan has taken things to another level entirely.

“A common refrain among investors of late has been the perceived role of the president’s Twitter activity on both shorter-term market dynamics and longer-term expectations”, the bank’s Josh Younger and Munier Salem write, in a September 5 piece called “Introducing the Volfefe Index”.

And yes, you read that correctly. Salem and Younger have constructed an index that pays homage to the president’s infamous “covfefe” tweet.

“The president has maintained a remarkably consistent daily presence on Twitter since taking office, averaging more than 10 tweets a day to his nearly 64 million followers since the start of 2016”, the JPMorgan duo note.

“Remarkably consistent” is probably an understatement. “Incessant”, “obsessive” and/or “bizarre” would also work to describe Trump’s “daily presence” on Twitter, as would “wholly dangerous”.

After noting that Trump’s Twitter activity hit what, for him anyway, was a listless nadir of roughly five tweets per day headed into the inauguration, Salem rather dryly points out that “starting in late 2018, activity has picked up substantially”.

This frenzied – some would just call it crazy – monthslong tweet-a-thon opens new doors for analysts looking to understand how markets interact with policy. As JPMorgan puts it, “such a high volume of policy developments made available instantaneously to financial participants en masse thus provides an unprecedented look into how markets could react to the inner workings of the executive branch”.

Salem and Younger look at the impact of Trump’s tweets on rates, although they acknowledge the obvious, which is that this same exercise is applicable to assets of all stripes, something anyone who trades knows all too well.

Although it’s hard to tell how serious of an exercise this is without hearing it discussed aloud by the authors, the methodology is pretty trenchant. Specifically, Younger and Salem use intraday data from the interdealer UST market and tag each tweet based on the subsequent move in 10-year yields over a 5-minute window.

“Our data reports the tweet’s timing to an accuracy of a second, and we time the measurement to that precision as well”, they proudly write, on the way to plotting the rolling 1-month count of Trump tweets followed immediately by a 0.5bp or greater move in Treasury yields along with rates vol. (They call 12-2 PM the “top hour for a tweet”, but we’d be remiss not to mention that at least anecdotally, 7-9 AM is a particularly dangerous time for market participants.)

(JPMorgan)

As you can see, Trump’s August tweets were associated with the dramatic spike in rates vol.

(Incidentally, it was Younger and Salem who, alongside the bank’s Marko Kolanovic, outlined the factors behind the August tumult, including the mammoth decline in yields.)

The bank goes on to describe how market-moving tweets can be differentiated from the rest of the president’s “covfefe”. Here is a list of Trump’s best wordzes, ranked by how frequently they appear in market-moving tweets (and do try to read this list without laughing):

  1. “China”,
  2. “Billions”,
  3. “Products”,
  4. “Democrats”,
  5. “Great”,
  6. “Dollars”,
  7. “Tariffs”,
  8. “Country”,
  9. “Mueller”,
  10. “Border”,
  11. “President”,
  12. “Congressman”,
  13. “People”,
  14. “Korea”,
  15. “Party”,
  16. “Years”,
  17. “Farmers”,
  18. “Going”,
  19. “Trade”,
  20. “Never”

The only thing slightly disappointing there is that “farmers” is so low on the list. One would have imagined it had a solid claim on a spot in the top five.

Ultimately, Younger and Salem construct a “Volfefe” index with the help of their “trained tweet classifier” which, according to the study, allows the bank to “produce an inferred probability that each tweet is market-moving by 0.5bp+ in the next five minutes”.

(JPMorgan)

Drilling down needlessly into the specifics, the bank “maps [the] market-moving odds to max(0,P-0.5), a so-called ‘move score'” before “finally” taking the rolling 21-day sum of that score across the tweets.

That laborious exercise eventually spits out “Volfefe”.

Judging by Exhibit 9, the index is at an all-time high or, as Trump would put it, it’s “at levels nobody has ever seen before”.


 

Speak your mind

This site uses Akismet to reduce spam. Learn how your comment data is processed.

One thought on “Behold! JPMorgan Unveils The ‘Volfefe’ Index

  1. Re: “… averaging more than 10 tweets a day to his nearly 64 million followers since the start of 2016”

    Fact: SpongeBob SquarePants has more followers than trump!

    Who has the most followers on Facebook?

    1) Cristiano Ronaldo–122.1 million likes. ...
    2) Shakira–104.6 million likes. ...
    3) Vin Diesel–101.6 million likes. ...
    4) Eminem–90.4 million likes. ...
    5) Leo Messi–89 million likes. ...
    6) Rihanna–81 million likes. ...
    7) Justin Bieber–78.7 million likes. ..

    The Most Followed Accounts on Twitter

    Barack Obama (108.2m followers)
    Katy Perry (108m followers) ...
    Justin Bieber (106.4m followers) ...
    Rihanna (93.1m followers) ...
    Taylor Swift (84.6m followers) ...
    Cristiano Ronaldo (79.7m followers) ...
    Lady Gaga (79.3m followers) ...
    Ellen DeGeneres (78.4m followers) .

NEWSROOM crewneck & prints