Saturday, October 6, 2012

How I Would Manipulate the CPI, If I Ran the BLS

The conspiracy theorists are out in full force charging that the fall in the unemployment rate as reporetd by the Bureau of Labor Statistics to the 7.8% level was a manipulated drop designed to help President Obama in his re-election campaign.

Was there such a manipulation? It would be impossible for an outsider to know. Chances are if there was a manipulation, it didn't come in the form of the number actually coming in at 8.2% and the Labor Secretary ordering that it be reported at 7.8%. The changes would be much more subtle, but just as effective. Assumptions would be changed deep within the BLS that few (even in the BLS would understand) but could move the number significantly.

The same types of changes in assumptions could move the CPI also. I am more familiar with how the CPI is calculated than the unemployment number, so I will use the CPI as an example of how deep level manipulation can occur. Suppose that CPI is going to be up because of climbing oil prices and food prices. Since everyone knows that oil and food, in my example, are climbing, it would be very dangerous for me, if I was at the BLS, to mess with these numbers. If it's, say, late September, I could try something like this:

Few understand it, but the BLS doesn't use the same items all year round to calculate their indexes. For example, they don't measure the price of bathing suits in the winter, but they do in the summer.  On the opposite end, they don't measure the price of winter gloves in the summer, but they do in the winter. One way I could manipulate prices lower, if I was the BLS, is by delaying the switch in measuring between summer goods and winter goods. If I delay the switch in September and I am still measuring the prices of bathing suits, lawn chairs, grills, sun hats, sun tan lotion etc., I can bring the CPI number down, because retailers are dramatically discounting these items to get them out the door before colder weather hits. (Note: I could bring seasonal adjustment factors into my discussion and how they could be manipulated, but my simple example gets the point I want to make across: There are many, many ways to manipulate the data.)

Please keep in mind I am just using this as an example, my point being there are all types of data  where, as Jack Welch has pointed out assumptions can be made and switched at any time, for any set of data the BLS puts out, CPI or unemployment numbers. The only way to determine for sure if manipulation occurred is to haul before Congress the BLS economists in charge and ask them under oath what assumptions, if any, were changed for the September numbers that were released Friday.

1 comment:

  1. Good post. I think the point here is what the BLS is reporting is for the most part a black box because outside the BLS no one really knows what type of methodology they are using to come up with their figures and this methodology changes regularly for all sorts of reasons--some of which may not be legit--and which isn't really clear to those the BLS.