Shadowstats CPI calculations
Home › Forums › DISCUSS › General Discussion and Questions › Is ShadowStats’ inflation closer to the truth than CPI? › Shadowstats CPI calculations
I haven't followed Shadowstats much, and I wouldn't pay the money for it, but I read through his April 8, 2013 'Public Comment on Inflation Measurement and the chained CPI'. His calculations look to be way off
He compares the usual CPI-U and an internal measure – CPI-U-RS – calculated as if all the methodological changes since 1980 were in place in 1980. He derives the year-by-year in the rates, sums the rate difference, and voila a 5.1% understatement of inflation for the CPU-I
I think column 1 is with Dec 1967 = 100, but it doesn't make any difference for this purpose.
This summing of year by year rate differences is very suspect. This what the real calculation looks like.
CPI-U-RS price level 2011 330.3
CPI-U-RS price level 1980 127.1
Change in price level 2.5987 (ratio of the two price levels)
average yearly inflation 3.129% (the 31st root of 2.5987 converted to %)
CPI-U price level 2011 224.9
CPI-U price level 1980 82.4
Change in price level 2.7294
yearly inflation 3.292%
difference in average inflation .163%
The difference in average inflation between the two methods is .163%, not his 5.1% . He's wrong by a factor of 31.
So his most important assertion is the result of a simple error in sophomore high school algebra. Doesn't look good for the rest of his comments.