Fed Nonsense And Error Bars

|
Includes: DIA, IWM, QQQ, SPY
by: Michael Ashton

Today’s news was the Employment number. I am not going to talk a lot about the number, since the January jobs number is one of those releases where the seasonal adjustments totally swamp the actual data, and so it has even wider-than-normal error bars. I will discuss error bars more in a moment, but first here is something I do want to point out about the Employment figure. Average Hourly Earnings are now clearly rising. The latest year-on-year number was 2.5%, well above consensus estimates, and last month’s release was revised to 2.7%. So now, the chart of wage growth looks like this.

ahe

Of course, I always point out that wages follow inflation, rather than leading it, but since so many people obsess about the wrong inflation metric this may not be readily apparent. But here is Average Hourly Earnings, y/y, versus Median CPI. I have shown this chart before.

aheandmedian

The salient point is that whether you are looking at core CPI or PCE or Median CPI, and whether you think wages lead prices or follow prices, this number significantly increases the odds that the Fed raises rates again. Yes, there are lots of reasons the Fed’s intended multi-year tightening campaign is unlikely to unfold, and I am one of those who think that they may already be regretting the first one. But a number like this will tend to convince the hawks among them otherwise.

Speaking of the Fed, last night I attended a speech by Cleveland Fed President Loretta Mester, sponsored by Market News International. Every time I hear a Fed speaker speak, afterwards I want to put my head into a vise to squeeze all of the nonsense out – and last night was no different. Now, Dr. Mester is a classically-trained, highly-accomplished economist with a Ph.D. from Princeton, but I don’t hold that against her. Indeed, credit where credit is due: unlike many such speakers I have heard in the past, Dr. Mester seemed to put more error bars around some of her answers and, in one of the best exchanges, she observed that we won’t really know whether the QE tool is worth keeping in the toolkit until after monetary conditions have returned to normal. That’s unusual; most Fed speakers have long been declaring victory. She is certainly a fan and an advocate of QE, but at least recognizes that the chapter on QE cannot yet be written (although I make what I think is a fair attempt at such a chapter in my book, due out in a month or so).

But the problem with the Federal Reserve boils down to two things. First, like any large institution there is massive groupthink going on. There is little true and significant diversity of opinion. For example there are, for all intents and purposes, no true monetarists left at the Fed who have any voice. Daniel Thornton at the St. Louis Fed was the last one who ever published pieces expressing the important role of money in monetary policy, and he retired a little while back. As another example, it is taken as a given that “transparency” is a good thing, despite the fact that many of the questions posed last night to Dr. Mester boiled down to problems that are actually due to too much transparency. I doubt seriously whether there has ever been a formal discussion, internally, of the connection between increased financial leverage and increased Fed transparency. Many of the problems with “too big to fail” institutions boil down to too much leverage, and a transparent Fed that carefully telegraphs its intentions will tend to increase investor confidence in outcomes and, hence, tend to increase leverage. But I cannot imagine that anyone at the Fed has ever seriously raised the question whether they should be giving less, rather than more, information to the market. It is significantly outside of chapter-and-verse.

The second problem is that the denizens of the Fed overestimate their knowledge and their ability to know certain things that may simply not be knowable. Again, Dr. Mester was a mild exception to this – but very mild. When someone says “We think the overnight rate should normalize more slowly than implied by the Taylor Rule,” but then doesn’t follow that up with an explanation of why you think so, I grow wary. Because economics in the real world, practiced honestly, should produce a lot of “I don’t know” answers. It may be boring, but this is how the question-and-answer with Dr. Mester should have gone:

Q: What do you think inflation will do in 2016?

A: I don’t know. I can tell you my point estimate, but it has really wide error bars.

Q: What do you think short rates will do in 2016?

A: I don’t know. I can tell you my point estimate, but it has really wide error bars.

Q: What do you think the Unemployment Rate will do in 2016?

A: I don’t know. I can tell you my point estimate, but it has really wide error bars.

Q: What do you think the Unemployment Rate will do in 2017?

A: I don’t know. I can tell you my point estimate, but it has really, really wide error bars.

Q: What do you think the consensus is at the Fed about the optimal pace of raising rates?

A: I don’t know. Each person on the Committee has a point estimate, each of which has really wide error bars. Collectively, we have an average that has even wider error bars. We cannot therefore usefully characterize what the path of the short rate will look like. At all.

Indeed, this is part of the problem with transparency. If you are going to be transparent, there is going to be pressure to provide “answers.” But a forecast without an error bar is just a guess. The error bars are what cause a guess to become an estimate. So we get a “dot plot” with a bunch of guesses on it. The actual dot plot, from December, looks like this:

dotplot

But the dot plot should look more like this, where the error bars are all included.

betterdotplot

Obviously, we would take the latter chart as meaning…correctly…that the Fed really has very little idea of where the funds rate is going to be in a couple of years and cannot convincingly reject the hypothesis that rates will be basically unchanged from here. That’s simultaneously transparent, and very informational, and colossally unhelpful to fast-twitch traders.

And now I can release the vise on my head. Thank you for letting me get the nonsense out.