by Calculated Risk on 7/07/2011 06:13:00 PM

I haven't been following the debate about using Chained-CPI instead of CPI-W for the Cost of Living Adjustment (COLA). (ht Andre).

Menzie Chinn at Econbrowser wrote today: Chained CPI

Recent reports ([WSJ RTE] [Bloomberg] [The Hill]) indicate that under consideration as one approach to curtailing entitlement spending growth is to resort to Chained CPI, as opposed to the current official CPI series, which is based on a quasi-Laspeyres formula.

From the BLS: Frequently Asked Questions about the Chained Consumer Price Index for All Urban Consumers (C-CPI-U)

I haven't been following this, but chained CPI is a relatively new series (started in 2002), and measures inflation at a slightly lower rate than CPI or CPI-W - and over time this would add up both for Social Security payments and also for revenue (tax brackets would increase slower using chained CPI than using currently).

Click on graph for larger image in graph gallery.

The graph shows the year-over-year change in headline CPI, CPI-W, and chained CPI.

There isn't much difference on a year-over-year basis, but notice the blue line is mostly below the other two all the time. Those small differences add up over time as the following table shows.

This table shows the 10 year change in each measure (from May 2001 to May 2011) and the annualized change over that period. If we were using chained CPI instead of CPI-W over the last 10 years, Social Security benefits would be about 3.6% lower than they are now.

	10 Year Increase	Annualized
CPI (headline)	27.2%	2.43%
CPI-W	27.8%	2.49%
CPI (chained)	24.2%	2.19%