Monday, August 14, 2006

Gov't to release more accurate numbers

It's about time:
The Bureau of Labor Statistics is contemplating a change in the idely followed consumer-price index that could have a big impact on how markets and policy makers interpret the latest inflation data.

The agency, part of the Department of Labor, is considering publishing the index and its subindexes to three decimal places instead of one, an agency official said. Doing so would greatly reduce the frequency with which rounding produces a misleading inflation rate.
Wall Street and the Fed both attentively follow the core inflation index (which excludes food and energy in addition to having other shortcomings) put out by the Bureau each month. The markets hit a springboard or a precipice depending on the number. Currently, the inflation rate is given as a percentage to the first decimal place. The change reported can be figured by looking at the actual index by month. The problem is, this number is rounded as well. So conceivably a change of .25% is reported as .3%, while a nearly equivalent shift of .249% is reported as .2%, a 50% difference in the magnitude of reported inflation.

Fortuitously, we live in the 21st Century, at a time when extending a number derived from the prices of over 80,000 items (before rounding its decimal extension competes with pi) is easily doable. It's inexplicable that the Bureau wouldn't make this painless and informative change that first-semester business school students would know to institute ASAP.



mping said...

Maybe, but in chemistry you have the concept of significant digits. You only have as many digits as your data allows you to draw.

I wonder if you have a similar situation here. If I remember correctly Greenspan was making the case that the reported inflation numbers are 1% too high (if they report 3% it is actually 2%). If there is the possibility that the number is off by that much, giving an extra digit seems beyond the accuracy of the data.

Inflation is a simple concept, but difficult to measure. How do you handle the case where a $300 40 GB iPod is replaced by a $300 60 GB iPod? How much (negative) inflation is there in that? 50%?

If a house is worth $300,000 one year, and the next the same exact house is worth $450,000, should that be counted as 50% inflation? Or is it different because it is considered an asset?

It also seems to me that wage inflation, commodity inflation and asset inflation are all different beasts. Trying to put them all into one number is really tricky. And giving people an extra digit gives the impression that these numbers can be calculated to a degree of accuracy that I am highly skeptical that they can be.

crush41 said...

But the Labor Bureau reports monthly percentages around .2% or .3%. Was Greenspan referring to annual inflation, or did he suggest it to be off by a tenth of a percent?

The questions you bring up alerted me to the fact that I used a bad link (now fixed). I meant to refer to shortcomings here, which includes hedonic adjustments, although I'm not sure of the methodology used to calculate them. Also, substitution tends to mask inflation, not overstate it. Not that I'm able to cross swords with Alan Greenspan!

If, as you suggest, the number is largerly useless or misleading, why even have it in the first place? Might as well make it more precise if it's going to be used at all.