My FOMC thoughts are in the emails I sent out during the week [The FOMC statement was dovish, but not as dovish as people think]. But I did want to discuss a few things in more detail.
Think about all the dovish FOMC members you heard from the past month or two. Everyone thinks we get at least 1 hike. There is only ONE FOMC member who thinks we get one hike in 2016. Every other FOMC member thinks at least two hikes in 2016. Let’s say the uber-dove is Evans (because that would be my guess). That means – Dudley, Brainard, Tarullo, Kaplan, Rosengren AND EVEN YELLEN all think 2 or more hikes… despite how dovish they have seemed this year (including Wednesday’s statement and the press conference). No one (sane) thinks April. So in only five meetings this year, the Fed thinks 2 or more hikes. That’s not *that* dovish.
I don’t know most of those folks. But a long time ago (in a galaxy far, far away), when I was trading at JPM and Dudley was an economist at Goldman, I used to read his research regularly (since Goldman and JPM were the two best back then). I’ve heard him speak in person multiple times, and I even had dinner with him once or twice when I was just the junior guy on the desk. He is a reasonably bright and thoughtful person. He’s probably always been one of the low dots (because being a low dot made sense) the past few years. So if he is telling me 2 hikes, then I’m going to respect it. Just like if Yellen tells me two hikes, or Fischer tells me two hikes. Could it be a half hike less (or more)? Sure. You never know how things will turn out. But he knows what a statistical “midpoint estimate” is supposed to be, unlike some of his colleagues.
On the other hand, if George tells me 4 hikes, I’m going to say, “it’s time for your medication.” Speaking of 4 dots, there are a total of FOUR “4 hikes in 2016” dots. It almost defies logic that you could be in a senior position in any capacity when you show judgement that poor. These are folks who clearly don’t follow what is going on in the rest of the world, outside of their Taylor Rule models. And there are FOUR of them!
We have all had those business colleagues where you start cringing whenever they speak. They could be the nicest people in the room, but there’s something off when it comes to expressing their view. It is my theory that the senior Fed officials decided to use the “median” dot because they looked at Kocherlakota’s dots vs George’s and Plosser’s dots and said… “how can we bring these people in-line?!?” Just one year ago, after the March 2015 meeting, there was one dot for the end of 2016 that showed a rate of 3.75%. That’s basically a hike every meeting for almost two years. How could this be a “midpoint” estimate? I don’t think they know what they are supposed to be submitting. So there was a chance we could have gone 50s at a few meetings? And just 6 months ago, Kocherlakota’s dots were calling for an ease. If you are trying to build a collegial atmosphere where everyone has an “equal” say, but some of your colleagues are borderline insane, the politically correct thing to do is to use the “median” (and “central tendency ranges”).
But as you all know, I am not politically correct, and I just say things the way they are. I prefer to just cleanly erase the dots of the knuckleheads and calculate a mean of the more reasonable members. If I was a member of the FOMC, I would be a good colleague and team player and use the median. But I have no such constraints. And hence, this is the genesis of my remove “the high 6 and low 2 dots” approach. This works well because Yellen is one of the more dovish dots, and the estimate construction leans dovish. I’m not saying this is “perfect” – but as a simple, singe-number measure, it gives a reasonable estimate. It helps me get a more granular sense as to what the movement in dots actually is from quarter to quarter.
On the right is what the SEP means looks like with the high 6 and low 2 dots removed (“core 9”). If you look at the difference between the Dec and March projections, it’s not that different, other than us taking a break for the first quarter (and part of April). In other words, if you just think of the first one-third of the year being a liftoff hiccup, the prospects for rest of 2016 looks the same to them, as it did in December. Another interesting thing about the table is how the long run FF projection has come down. As late as Dec 2012, the long run projection had a “4” handle (4.0%). In theory the long run estimate is not supposed to change that much, since it is supposed to be in the “long run.” I’ll discuss further in a later post.
You’re probably saying, “Why even pay attention to the dots? They seem to be wrong all the time.” It’s because as we get closer to relevant meetings, the dots are more accurate. I wrote an explanation of it in an email last week (in the Appendix). Consider the movement of dots for the end of 2015 over time. The rows of the table list the meeting dates of the SEP, and the table shows the “core 9” rate for the end of 2015 over time. The Fed was a little off initially (expecting 3.1 hikes in 2015, at the Dec 2014 meeting), but by the middle of the year, they were not that far off of their year-end target of 1 hike.
You’re probably saying, “But that’s a sample size of 1 (hike at the Dec meeting)!” True. But I think the theory is valid. BECAUSE THESE ARE THE PEOPLE WHO SET THE RATES! If the Fed says we will hike rates 25bps at the current meeting, there is a 100% chance the rates will rise by 25bps. It’s a sure thing, because they set the rates. It doesn’t matter if the markets think it’s 50-50. The markets don’t set the rates. If the Fed says they will hike rates 25bps at the next meeting, a lot of things can happen in 1.5 months. But most of the time, probably not enough things will happen to deter the Fed. So maybe the hike is an 80% proposition. In any event, the closer the meeting, the more accurate the Fed’s projections should be. The further away, the less accurate the Fed may be. So it is not as strange for the markets and the Fed to disagree for 2017 and beyond.
The other thing to note is that I believe these projections are sent in over a week before the FOMC meeting. So I do not think the projections capture the uptick in CPI on Wednesday. Granted, it’s only 1 piece of data, but CPI surprised me a little, since I thought last month’s data would be the local high in yoy inflation (with last year’s oil drop getting priced out).
The last point I wanted to make is that the longer run unemployment rate was only lowered a tenth to 4.8%. This is not unreasonable considering the UR is currently at 4.9% and there is still slack in the labor market. My main economic takeaway is that the 2018 unemployment rate projection was lowered two tenths to 4.5%. So in conjunction with being more gradual about rate hikes, the Fed sees allowing growth to run a little more (i.e. allowing the UR to drift further below the long run UR). But who knows? The Fed decided to publish the median (arrgh!) so it’s hard to tell what is going on. The Fed is not going to risk putting the brakes on the recovery, so the path is going to be slow and gradual, and there will be no surprise hikes, at least in the short term. This also should be supportive of curvature further out the curve – there is less of a chance the curve will invert because the Fed is too aggressive with hikes).
 Maybe it’s closer to 1.75 dots but they rounded up.
 One FOM C member thinks the long run FF rate is 5.8%. Eh hem. This model apparently hasn’t been updated since 1976.
 I’m not sure if they changed this policy recently. But there are many pages analytics the staff compiles, so some preparation time is required.