Please Note: Blog posts are not selected, edited or screened by Seeking Alpha editors.

Inflation In Food And Energy: The Peripheral Is The Core

|Includes:DBA, GLD, iShares TIPS Bond ETF (TIP), UDN, UUP

I expect to go to grave without understanding why much of the government focuses on the "core CPI" (or in Fed's case, the core equivalent of PCE -- just another way to delay the inflation alarm), meaning CPI excluding food and energy. Sure, I understand the official excuse and the political incentives. I just don't understand how the American people let them get away with it for so long. I guess we're just too happy, or not miserable enough if you're of the half-empty persuasion, to worry about such nerdy details. Fine.

And then again there've been many solid arguments on how the CPI consistently, systematically underestimates inflation. There're many resources on this regard on the web but I'll just link two random good ones: ShadowStats and a recent WSJ blog by Brett Arends. But...fine. CPI is not science, it's not even art, it's politics. I get it.

So let's forget about the Horrendous Hedonics and all the political mathematics. Let's just take it at the face value and see what it tells us beyond the headlines.

Below is the graph of CPI, Core CPI I since 1993 got from BLS website, and the Food and Energy price change I derived from it. It's the CPI-U series, unadjusted, normalized to Jan 1993 by me because that's the date all data streams started. Although BLS provides Food and Energy series separately, I cannot add them up without knowing the weights of the two categories. I back out the Food+Energy price change from the "All items" and "All items less food and energy" series, and assuming a combined weight of Food+Energy of 15% in the CPI calculation.

Chances are the Food+Energy weight is a known number. I'd appreciate your help in eliminating this ignorance of mine and I'll redo the calculation. The higher the weight, the less dramatic the extra inflation in FE, but the dependency drops beyond 10% or so. FE is about twice expensive now as in early 1993 -- this "feels" a bit low but is probably in the right zipcode.

But the red bar series FEELS much more like the real inflation than the smooth CPI and especially smooth Core CPI numbers, doesn't it? A few interesting observations about FE inflation:

  • There was a turning point in early 2002, after which time FE inflation took off. This is a crucial distinction underlying the bull market since 2002 vs the bull market of 1990's. The great boom of 1990's, while ending in a spectacular bubble burst, was nevertheless driven by real growth in productivity, which was in turn driven by innovation and liberation of a vast global pool of talents previously untapped or undertapped. The bull market of 2000's, on the contrary, lacked real, organic growth. It was forced, manufactured, labored via relentless fiscal and monetary stimuli. Many already knew this. Now we know how to use official stats to see it.
  • There was a hysterical surge in FE inflation five months before the 08 crisis. The monthly FE figures are indeed volatile; yet when it surges 2%+, non-annualized, every month, for several months in a row, COME ON DOC, something is probably wrong. Somebody probably should take a closer look rather than throwing it away just because it's inconvenient.
In fact, here's a graph of 3-month moving average of the above estimated FE inflation since Jan 1957.

Again, a few interesting observations:
  • Three recessions, namely those of 1958, 1873-1875, and 1980, were clearly preceded by the 3M MVA of FE inflation > 2% (again, non-annualized). I don't expect such a crude estimate of FE inflation to be of any predictive value; yet it turns out to be quite meaningful. Better quality data should help clarifying this intriguing hypothesis.
  • Two decades with consistently low FE inflation, the 60's and the 90's, correspond to the time when US economy flourished, the middle class either grew larger and/or more prosperous as a whole.
  • Periods with highly volatile FE inflation mark the times of either recession or economic polarization. In particular, FE inflation has been persistently volatile since 2002. It screams of instability.
  • Monthly change in Dec 2010 was 1.5%. Jan 2011 must've been higher.
This is obviously very preliminary and crude. But I hope I have demonstrated the potential value of focusing on the "Peripheral CPI" and ignoring the "Core CPI", which correlates to nothing of any meaning.

While I'm in the rare chart-porn mood, let's look at the components of CPI as defined by BLS. All components are normalized to Jan 1993, sorted roughly by the change from Jan 1993 to Dec 2010, as indicated by the legends from left to right.

Apparel got a little cheaper, obviously thanks to sweat shops in China.
Education and Communication costs "only" increased 55%. Horrendous Hedonics at play, no doubt -- the iPhone (without prison terms) is obviously much cheaper than the old phones because it can run 10,000 apps. Transportation and Food & Beverages all increased ~55% each. HH.
Housing cost increased only ~30% from 1/2002 to 8/2008, and barely changed since. Well this must be a mix-up with some parallel universe in a spatiotemporal anomaly; I'm glad they have managed their affairs so much better over there.
Medical Care and Other costs doubled in eight years.
For the kick, here're two charts on the Purchasing Power of the Consumer Dollar, straight from BLS, but renormalized to Jan 1993. The left one is from Jan 1913; the right from Jan 1993.

Remember this is an optimistic view of your purchasing power decline. Coincidentally, we are almost exactly on par with just before the crisis; no more free extras at Flash Dancers.

Disclosure: I am long GLD, DBA.