Inflation Forecasts: Are Government Estimates Believable?
28 Mar 2006 in Regulatory Science, Regulatory Economics, Information Quality
Neutral Source recently posted a pair of blogs about how inflation forecasting was like risk assessment, and how it isn't. David Wessel started the discussion in his Capital column in the March 16 edition of the Wall Street Journal. On March 21 he posted a selection of interesting reader responses. (A subscription to wsj.com is required to view them.)
What's interesting is the extent to which some respondents (and probably many others) think the CPI measures one thing but the economists and statisticians responsible for it intend for it to mean something different. The CPI can be used to measure the "cost of living," but as the Bureau of Labor Statistics says, that's really not it's primary purpose. On its FAQ page, BLS tries to explain that a true "cost of living index" would measure the change in income necessary to achieve the same level of personal satisfaction (what economists call "utility"). But that's a tall order, and maybe an impossible one. And it's even more impossible (!) to develop an index that applies to each individual's specific situation. That's not what indices do.
A very important thing to keep in mind is that over time the quality of goods and services changes, mostly for the better. While in college during the late 1970s, when inflation was a serious and growing problem, Neutral Source had a mentor, Prof. Thomas Mayer, who explained it this way.
"Suppose I gave you $1,000 to spend," he would say, "but you have to make a choice. You can choose the 1978 Sears catalog and pay 1978 prices, or you can use the 1958 Sears catalog and pay the much lower 1958 prices. Which would you choose?
Almost everyone in the class said that they would prefer the 1978 catalog and pay 1978 prices, despite the ravages of inflation that had occurred over 20 years. (The rest of the students lied. Hey, it was the '70s, man.) But the lesson he sought to teach was crystal clear: Unless you control for changes in quality, rising prices don't tell us a lot about the rate of inflation.
