The Boskin Commission found that the CPI is overstated "by as little as .6% and as much as 1.5%", from what I can figure out, the upward bias being probably in the range of "0.65 percent", down from 1.1 percent for the 1995-6 period..
I haven't read Robert Gordon's paper though, and the percentage error margins cited aren't very meaningful unless you can review the whole methodology (classifications, questionnaires, estimation procedures, consumer patterns, and regimen). Robert Gordon discovered that when workers net real incomes decline they go and shop at cheaper stores, and on the other hand, that if supermarket chain monopolies develop, this may also be a factor in the choice of store. Congress decided not to change the CPI methodology. In Western Europe, accelerating price inflation combined with changing consumption patterns during the long postwar boom led some trade union federations to produce "alternative indexes" of price inflation. In general, you could say that statisticians tend to distrust large fluctuations in a data distribution, particularly if they are unprecedented or do not conform to an already existing pattern in the data. The effect of retrospective revisions has a higher likelihood of "smoothing out" data series. http://www.raleightavern.org/wilson.htm http://www.moaa.org/Legislative/Retirement/CPI.asp On quality adjustment, see http://www.boj.or.jp/en/ronbun/01/cwp01e06.htm J. ----- Original Message ----- From: "Michael Perelman" <[EMAIL PROTECTED]> To: <[EMAIL PROTECTED]> Sent: Sunday, July 20, 2003 1:12 AM Subject: Re: [PEN-L] it's over! > Would we be out of the recession without the Boskin inflation adjustments? > -- > Michael Perelman > Economics Department > California State University > Chico, CA 95929 > > Tel. 530-898-5321 > E-Mail [EMAIL PROTECTED] >
