How Old Is the CPI?

Looking at the latest press release on the CPI numbers, issued by the Bureau of Labor Statistics (BLS), the idle thought occurred: I wonder how old the CPI actually is? My guess was that it arose, no doubt, either in connection with World War II or in the wake of the Great Depression. I expected the usual quick answer from either the BLS itself or from Wikipedia. No quick answer anywhere. But—finally!— I did find the answer—and it surprised me. I discovered it in The First Hundred Years of the Bureau of Labor Statistics, a 321 page history but issued as Bulletin 2235 by the BLS. It appeared in September 1985 and was written by Joseph P. Goldberg and William T. Moye. The book is available on the web here. I found the relevant event I was after recorded in this paragraph spanning pages 34 and 35:

Two reports prepared by the Bureau for the Aldrich Committee became landmark sources of data on prices and wages. Some wholesale price data were assembled for the preceding half century; for the 28-months preceding September 1891, prices were collected for 218 articles in 7 cities. Retail price collection was limited to the 28-month period, covering 215 commodities, including 67 food items, in 70 localities. Wage data were also assembled for the preceding half century in 22 industries; for the 28-month period, the data covered 20 general occupations in 70 localities and specialized occupations in 32 localities.

The Aldrich Committee was the Senate’s Committee on Finance, chaired by Nelson W. Aldrich on Rhode Island. The Tariff Act of 1890 had been passed. It’s known as the McKinley Act named after then Congressman (later President) William McKinley. (Full disclosure: Quite appropriately for a data maven like me, I live on McKinley Avenue). The McKinley Act was my kind of law. As Wikipedia summarizes it “The tariff raised the average duty on imports to almost fifty percent, an act designed to protect domestic industries from foreign competition.” Hear, hear! Anyway, Senator Aldrich asked the Bureau of Labor Statistics, at that time a mere six-year-old, to collect the right kind of data in order to measure the effectiveness of the McKinley Act.

The fundamental methodologies and ideas behind the Consumer Price Index now in force were already in place 120 years ago. The BLS defines a kind of “representative basket of goods and services”; it sends its agents out to visit actual stores and other institutions to collect current pricing data; and it does so across the country. Mind you. “Basket” here means more than just groceries and clothing. It includes all categories of consumer spending including fuels, rents, mortgage payments, premiums, tuitions, fees, childcare—and, yes, money spent on all kinds of products including groceries and clothing.

The idea of weighting each category within the price indexes—in order to derive a single representative number that stands for prices in general—was already present in 1891. The idea was introduced by Ronald P. Falkner, of the University of Pennsylvania, who had been hired to analyze the data the BLS had collected. Weighting was based on patterns of family expenditures. Thus categories were ranked by the amount of  money families actually spent on them. The greater the spending, the higher the weight. The name of the CPI was initially Cost-of-Living Index. The renaming took place in 1945 and was then called Consumer’s Price Index for Moderate Income Families in Large Cities. Today we call that series CPI for All Urban Consumers (CPI-U). Using the CPI as an official measure to adjust wages, pensions, and so on dates back to 1948.

Having now stared for a while into the well of the past, I will proceed to present the February 2011 numbers.