by Laurent GUERBY
Mon Nov 20th, 2006 at 02:33:50 PM EST
A few monthes ago I asked various economists, including Bernard Salanié (see comments) why the administrations computing inflation didn't make their raw price data available since it would end conspiracy theories, endless debates about the magic "global" inflation number and provide quite useful data to taxpayers who all have different perspectives on inflation : yound, couple with or without children, old or retired face very different inflation, different regions face different prices level and inflation, etc...
The near unanimous answer (see the comments in the link above) was that data can't be released otherwise store owners would change their price behaviour (and that I didn't know anything about economics, etc... nothing unexpected).
After a while, encouraged by Max Sawicki who told me honnestly not knowing the reason and that Bureau of Labor Statistics (BLS) people were nice people, I decided to fire a few emails.
I first asked the BLS contact email listed in their inflation FAQ
. Next morning USA time the answer came in, after an exchange of 7 emails during the day here is what I found out about the BLS way of measuring inflation:
- BLS workers show immediately their official ID to the store owner and ask for permission before collecting price data. The reason is that the store owner could kick them out for suspicious behaviour otherwise (could be an unpleasant experience)
- They come back to the same stores periodically, so store owner is so perfectly aware of the use of her prices, and owners do cooperate seriously with BLS
- Raw data cannot be published since they need the agreement of the store owners, and publication is forbiden by this agreement
Of course, a widely different answer than the one from the so-called experts who-know-everything cited above!
For reference, I tried the same experience with the French INSEE but this went nowhere, to cut it short: data is not released because data is not released.
But in France, citizen can (and do) collect prices in stores (even without hiding it, as long as you don't do something foolish inside the store) and publish them.
In the end no excuse of any kind for not publishing the raw price data (and a few other data needed to build meaningful measures).
Side story: a big store chain "owner" (a bit more complicated) in France, Michel-Edouard Leclerc did try to show that Leclerc stores where cheaper than other store. In a first step he published on the web an index like INSEE and did not release raw data. The site was legally threatened and closed quickly.
But last friday he did put up another site with raw data available for 355 stores all over France, 1536 products, 414469 prices collected: quiestlemoinscher.com.
I quickly wrote a python script to collect all data in a spreadsheet compatible file and did a few experiment with it, eg: on the 30 most widely referenced product, basket prices vary 14% between cheapest and most expensive.
This little story IMHO tells a lot about "experts" and "citizens", hiding data while pretending to be a science ("economics"), and a good target for free our data-like grass-root campaigns.
How's the situation in your country?
Why do economists invent excuses to hide their data from the public scrutiny?
French version on my blog (a bit more data)
Update 20061124: Slightly different version in French accepted at Agoravox, a french citizen-journalist site