Calculating Inflation Accurately
February 22, 2011 3:44 PM   Subscribe

There is a branch of economic theory concerned with the calculation of inflation. It has a name. The name is something like "hexanomics"--but that's not it. Please help.

The branch is basically concerned with this: Television Manufacturer TVCO put a red blinking light on Model 2011A so the it becomes Model 2012A. The call this light a "feature" and charge $50 extra dollars for the TV.

When calculating inflation, the $50 has to be taken into account. Is it a legitimate inflationary expense being passed on to the consumer? Or is TVCO trying to dupe the market with a useless "feature."

There is a name for this branch, and I can't for the life of me remember it. Your help would be greatly appreciated.
posted by jefficator to Work & Money (5 answers total) 1 user marked this as a favorite
 
It has other applications too, but I guess you talking about hedonic regression/ pricing.
posted by hawthorne at 3:59 PM on February 22, 2011


Yes, this is hedonics. I could have sworn Planet Money did a podcast on this, but I can't find it now.

To put a different spin on hedonic adjustments: it's concerned with isolating the improvement made in roughly comparable goods from inflation. You can't buy a car without seatbelts or airbags (I think?) anymore, but the raise on prices shouldn't be calculated as inflation. Similarly, if you could buy a computer from 1991 today, it'd be hilariously cheap. We don't even measure RAM and disk prices in megabytes anymore.

There's been some complaints about this (republished wsj article).
posted by pwnguin at 4:58 PM on February 22, 2011


I assume you're talking about hedonics, though I might quibble with your description of it. The idea is to account for changes in the quality of goods over time. General inflation measures typically are intended to track changes in price of some representative basket of goods over time, but the goods in the basket disappear, and are often replaced with a NEW! IMPROVED! (and more expensive) model. If the new model were simply substituted in for the old with no adjustment, then the inflation measure would show an increase in price. But sometimes those new models actually do have more samoflanges or fips or whatever consumers care about, and at least part of the higher price is due to the added or improved features. So in a sense, the price for the old good, to the extent its features are still available, hasn't gone up as much as a straight substitution would suggest. We discussed this on the blue quite extensively some time ago - see here for a more in-depth discussion. You may want to read that discussion before diving into pwnguin's link, which is a bit misguided for a number of reasons.
posted by dilettanti at 10:03 PM on February 22, 2011


> if you could buy a computer from 1991 today, it'd be hilariously cheap

I looked at that dumbly for a while until I realised it's on a logarithmic scale. What would it look like if it was on a regular scale?
posted by AmbroseChapel at 12:40 AM on February 23, 2011


dilettanti: "You may want to read that discussion before diving into pwnguin's link, which is a bit misguided for a number of reasons."

Well, anything WSJ should be taken with a large grain of salt. And further research suggests this might not even be WSJ, but someone who wished they had that mantle and couldn't get it. Which I take as solid evidence the opposite of whatever they say is true.

What would it look like if it was on a regular scale?

Straight lines on that logaritmic scale means exponentials. So it's a roughly a 10/10^x graph in linear scale. Which will look like you're barely making progress, but the reality is that 0 is a definite, unreachable lower bound. I've been meaning to put his data into a nicer ajax chart.
posted by pwnguin at 8:07 AM on February 23, 2011


« Older Recording loud concerts with minimal gear?   |   Evil care bears Newer »
This thread is closed to new comments.