E-piphany

March Madness has begun, and as if on cue market volatility is evaporating. Despite a fair amount of economic data, stocks ended the day unchanged. Stocks have had a nice run, but seem extended; a period of time in which the attention of many investors is on the bracket is probably not the best prescription for an extension higher…

Bonds had a bit more action, with the 10y note finishing -8.5/32nds (10y yield at 3.67%), but of course CPI is more directly relevant for fixed-income. It is rather interesting that rates went up, and not down, on a day when inflation fears (the way people conventionally measure them) went into retreat. How much in retreat? Well, 1y inflation swaps fell 10bps (the yield of the April 10s, which is effectively a T-Bill after today’s print, rose 93bps), and longer inflation expectations retreated 2-4bps. Declining inflation expectations are usually associated with declining rates, which means that real rates actually rose more than nominal rates today.

That is a knee-jerk reaction to weak-seeming inflation data, but it isn’t right. Changes in real yields should track changes in the outlook for real economic growth; those expectations certainly shouldn’t have leaped today, with Initial Claims and the Philly Fed both near expectations (the general index of the Philly Fed was up, but the New Orders index fell to 9.3 vs 22.7 and the Employment question drew 8.4 vs 7.4 last month) and Greece looking like it’s out of the frying pan and back in the fire. And, besides, if investors really were marking up their expectations for real growth, then equities would have done better.

No, real yields rose because the knee-jerk response to the inflation data was to dump TIPS. Dumping TIPS in favor of Treasuries (which would be selling breakevens) is the theoretically correct move if your expectations of inflation have declined; in theory, real yields should be roughly unchanged and nominal bonds should improve so that the spread tightens. But investors can’t help but think of TIPS as having inflation “protection,” when a more-accurate statement would be that TIPS are merely neutral on inflation and shouldn’t react to changes in inflation expectations. Remember, the Fisher equation says

(1 + n) = (1 + r) (1 + i) (1 + p)

or, approximately, that nominal yields are the sum of real yields and expected inflation (plus a risk premium we usually ignore because we can’t analyze it independently with 4 unknowns and 2 instruments). TIPS have a real yield, and the Fisher equation tells us they are independent from expected inflation; on the other hand, nominal yields respond to changes in both real yields and expected inflation.

The market isn’t efficient, though, so that when inflation expectations are declining what usually happens is that TIPS and other inflation bonds get killed and nominal markets don’t move much, so that smart investors can get some bargains after such a number.

The funny thing is that CPI, while slightly weaker-than-expected (mostly on the headline figure; I had suggested yesterday that headline needed to come back to core somewhat), wasn’t really all that weak. I am fine with forecasters who, by essentially forecasting rents, tell us that core inflation should continue to decline for a while. I agree. But they should acknowledge that the argument begins and ends with housing.

I read a piece by JP Morgan today which included charts of Median and Trimmed-Mean CPI, among others. All of these measures are declining, and conventional analysis says that this confirms the fact that core inflation is really “nosediving.” In JPM’s words:

“The basic idea behind robust, or ‘hard-core,’ measures of inflation is that while one can always subjectively strip out certain inflation categories to tell whatever story one wants to about what’s happening with inflation, in the robust measures categories are removed systematically to arrive at the underlying trend in inflation.”

But these measures in fact say nothing of the kind. The problem here is not the problem that trimmed-mean and median measures were meant to solve. They were designed to take care of the problem that it is always the case that some measures are rising faster and some slower than core inflation, and sometimes these “outliers” can have meaningful effects on the average inflation number. By “trimming” these outliers, the basic trend is revealed.

But that isn’t at all the problem here. The problem here isn’t that housing, a very large part of the index, is moving violently and skewing the figures. It doesn’t need to move violently, because it is so big. In fact, it is declining, but gently. However, the weight of housing is so big that it (a) dampens the overall volatility of inflation and (b) can change the average even if it isn’t away from the average very much (by dint of its weight).

And we don’t want to “subjectively strip out certain inflation categories to tell whatever story one wants to.” We want to strip out a category that we know does not represent the underlying trend because it is coming off a bubble. Maybe we knew it was bubbling before, maybe not, but surely by now we all know that housing was a bubble, and is in a period of adjustment. The Fed has no power to push housing prices meaningfully higher without starting a very broad inflation, because they’re gonna go down until the bubble is all the way gone. But the Fed can, and is, pushing other prices higher. We take out housing because it is demonstrably something that will not respond in a normal way to monetary policy. There is no systematic CPI measure that adjusts for this.

It is why we refer to the analyst’s “art,” rather than to the analyst’s “science.” A good analyst knows that “subjective” doesn’t always mean “bad.”

So, when you strip out housing, core inflation is running at 2.6%. That’s down a bit from the recent highs, but a far cry from the 1.3% core inflation including shelter and, more importantly, well above both the recent as well as the secular lows (see Chart).

"Core" inflation is getting increasingly deceiving

The models I follow also say that core inflation (including shelter) should continue to decline for another quarter or two. Duh. You don’t have to be a genius to see that rents are going to remain under pressure for a while, and once you know that then you don’t have to worry about core inflation.

But, as a policymaker, you do have to worry about all those things that are not rent. Inflation is rising, and will very likely continue to do so.