Sunday, January 13, 2008

Different CPI's for Different Income Categories Again

I did a little experiment. I took the consumer expenditure survey data on expenditure shares for 2006 by income quintile, and lined them up as best I could with CPI categories. The match is not perfect, but pretty close: I was able to replicate CPI growth between November of 2006 to November 2007 within about a tenth of a percentage point. It turns out that for that period, those in the lowest quintile had slightly lower than average CPI growth (yes, it surprised me too). The reason is that housing costs over the past year have risen by less than CPI, and low income people spend disproportionately large amounts on housing.

My estimates:

CPI growth by Quintile relative to overall CPI growth, Nov '06-Nov '07.

Lowest 20% -.128%
Second Lowest .174%
Middle Quint .017%
Second Highest -0.017%
Highest 20% -0.094%

Someone out there should do this carefully--I would be happy to share my spreadsheet. It appears to be less of a big deal than I originally thought.


Anonymous said...

The lowest quintile would be insulated from inflation anyway because of their eligibility for gov programs (food stamps, Medicaid, section 8 housing, etc...) Gov programs tend to be inflation adjusted.

The second quintile is the one that is the most vulnerable, especially if they have to pay for their own health insurance (private policies tend to go up at double digit rates).

davelindahl said...

After you get one or two lined up and you start to use them successfully, watch what happens: They will tell their friends, who will tell their friends, and so on. It’s human nature to brag at cocktail parties or at the gym about what a great investment you just made. Before you know it, you will have all the funds you need by the David lindahl scam, and your business will explode!