fed fomc jerome powell Markets

Tasseography And The Shrinking Playground.

Tea leaves, optimal control problems, and inner and outer limits...

Two important developments were lost this week in the fog of the partisan war that raged inside the Senate Judiciary Committee.

The first, as noted earlier, was the implementation of “phase two” in the increasingly contentious trade dispute between the Trump administration and Beijing.

The second was the Fed meeting. Obviously, the Fed hiked and there was some discussion around the removal of the word “accommodative” from the statement.

Normally, everyone would have spent the next 24 (at least) hours parsing Jerome Powell’s post-meeting remarks (which found the Fed chair attempting to play down the FOMC’s decision to nix the “accommodative” language) and just otherwise reading the tea leaves (and there is perhaps no better example of something that is amenable to the sometimes derisive “tea leaves” characterization than the dot plot).

Well, when it comes to the dot plot and the idea that divining anything from it is an example of Tasseography dressed up as statistics and finished off with an attractive coat of academic lacquer, Deutsche Bank’s Aleksandar Kocic is out with some great commentary.

Kocic’s discussions of forward guidance and the channels through which monetary policymakers communicate with markets are always a pleasure to read and discuss. Over the years, he’s developed something that approximates a unique lexicon comprised of unlikely but incisive analogies and apt metaphors that help to contextualize the evolution of the post-crisis monetary policy regime.

In his latest note, dated Friday, Kocic takes on the dots.

“Fed dots has evolved as one of their main communications tools with the market”, he writes, before noting that if you can get past the rather “troubling” fact that we’re all trying to “talk about statistics of a micro sample (16 data points), then the Fed dots, when taken at face value, represent a histogram plot.”

Kocic’s notes very often ask for the reader’s participation and demand actual engagement with the material. That marks a stark (and welcome) contrast to the almost universally dry world of sellside research. His latest is no exception.

Here is Kocic presenting a visual that quite literally requires readers to “rotate [their] head to the right by 90 degrees”:

The Figure shows an alternative visualization of Fed dots in this context. The two lines represent the Fed median (intended to guide the eye) and the OIS forwards (currently priced in by the market). At the end of each year we arranged Fed dots into a histogram (viewed by rotating the head to the right by 90 degrees) by counting the “frequency” (how many dots) of each rate level. To offset the choppiness of such a sparse histogram, we smoothed it across different bins, so that it resembles an actual distribution.


Kocic makes two points about this.

First, he notes that even if we weren’t trying to do statistics with a sample size of n=16 (an absurd endeavor), this would be largely meaningless in terms of forwards. “Even if the dots represented actual distribution, namely, if the Fed had many thousands of members, instead of 16, and the dots could be transformed into a bona-fide histogram, there would still be no fundamental reason that should link the mean or median of these distributions to the market prices”, he says.

But more importantly, there isn’t much in the plot to suggest there’s anything which approximates a consensus outside of 2018, and you can’t really point to that as being especially notable considering how close we already are to the end of the year.

“Distribution of the 2020 dots has very little that lends itself to any statistical interpretation”, Kocic says, adding that “if anything, it resembles a uniform distribution [as] anything in a very wide (100+bp) range in 2019 and 2020 is a fair game.”

In the same note, Kocic also takes you back through the “inner and outer limits”, a discussion that harkens back to the Fed’s “optimal control problem” discussed here in April.

Read more

Inner And Outer Limits: Resolving An Inconsistency And What It Means For Risk

Earlier this year, Kocic and his colleague Steven Zeng broke down the mechanics of the Fed hiking cycle into “two parameters”:

  1. Policy gap or outer limits, and;
  2. Eurodollar spreads or inner limits

The first point there will be familiar to regular readers as it recalls another of Kocic’s trademark analogies – the “shrinking playground”.

The policy gap (the “playground”) defines the scope for rates volatility. As Kocic is fond of putting it, “anything that can happen, happens inside that gap.” Here’s an updated visual which simply illustrates what the gap in fact is, for the uninitiated (i.e., it’s the spread between the long rate and the shadow rate):


(Deutsche Bank)

As alluded to above, ED spreads are a reflection of the playground. Here’s Kocic:

The distance of the Fed funds from the long rate is reflected by the spread between the short and long term projections of the Fed, the money market risk premium like Red/Green or its mirror image, Blue/Green. This is the essence of the optimal control setup of the Fed tightening cycle.

Greens will be, as Deutsche put it earlier this year, a “pivoting point” for the rest of the ED curve re: pricing the Fed path. When the hiking cycle starts, there’s ambiguity about when it will ultimately end, and in that context, the spread between Blues/Greens (3y vs 2y) is basically just how likely it is that Greens are the terminal Fed. Here’s  Blues/Greens plotted with the policy gap:


(Deutsche Bank)

Kocic goes on to note that “while a wide gap does not always guarantee high volatility, it ‘allows’ it [while] a tight gap inhibits (realized) volatility.” Here’s an illustration of that which shows the policy gap plotted against realized vol. in 10Y swaps (3M rolling):


(Deutsche Bank)

Note from the blue shaded regions that Kocic identifies a different dynamic at play currently than that which unfolded previously.

Specifically, he writes that “in the past, rate hikes typically start when short rate is low and long rate is high [and] during the tightening cycle, the Fed was always catching up with the long rate.” The tightening cycle stopped when the gap disappeared.

This time around (i.e., coming out of the QE era), long rates were suppressed when the hiking cycle began. And although there have been periodic episodes where the long end has sold off (yields rise), rates have, to quote Kocic, “struggled to take off materially since [the election], allowing Fed hikes to practically collapse the Policy Gap.”

Regular readers can perhaps anticipate where Kocic is going with this. This is a segue into the discussion of the Fed attempting to normalize the mode of the curve. There’s more on this in “Transparency, Trust And The ‘Fed In Wonderland’,” but the bottom line is that this is an effort to restore the “normal” order of things where shocks arrive at the front end of the curve because given the setup, a vicious bear steepening episode (i.e., a shock in the long end) would likely trigger the bond unwind tail risk. Here’s Kocic:

By systematically maintaining tight policy gap, Fed both extinguishes risk premia and volatility. In this way, they are effectively withdrawing convexity from the front end (by reducing transparency and reintroducing the data dependence) and supplying it to the back end in order to stabilize it.

Going forward, Deutsche Bank’s core view is that the Fed will hike in December on the way to hiking four additional times in 2019 with no hikes in 2020. “In that context,” Kocic concludes, “2019 forwards need to price in another 2 ½ hikes.”



0 comments on “Tasseography And The Shrinking Playground.

Speak On It

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

Skip to toolbar