The "Measure" of Inflation

I'll make this post a quick one, but I wanted to add a little color to my last installment. Specifically, my criticisms of CPI as a methodology for measuring inflation.

The most oft-stated rebuke is that it excludes food and energy prices from its calculation, due to their volatility. For a host of reasons, I think it's ludicrous. But, it is what it is.

The part that warrants some explanation is the hedonic model, which from my view receives less attention and is probably not as well understood. The gist is as follows: product XYZ costs $100 today. A year from now it costs $103. Intuitively, one would think that the price increase was 3%. In fact, according to CPI, that's not necessarily the case. Rather, the BLS, through the use of some sort of voodoo, err, regression analysis, makes a determination about how much of that $3 increase can be attributed to an improvement/change in product quality - oh, I don't know, say $2. Therefore, on a CPI basis, the rate of inflation would be something more like 1% - even though it still cost the consumer 3 more bucks.

In the end, I think it's clear that the hedonic model could potentially be manipulated or simply wrong, rendering CPI a fiction. And ultimately making the debate about inflation v. deflation, premised on CPI, a fool's game.