Consumer Price Index (CPI)

Estimating Inflation

There is a problem in comparing the purchasing power of currency when it changes over time. Prices may rise and fall for different reasons such increases in food prices brought on by drought because of scarcity or decreases brought on by abundance from good weather. These kinds of changes are not "inflation" as we normally use the term. Inflation is the decline in the value of a currency and, as it is often described, there are more dollars to be spent on the same number of goods, the price in those less-valued dollars increases. It is a hard thing to measure. Different people do different things with their currencies in different areas of the country and in different years so the effect of a decline in a currency really varies with each person or company spending money depending on a host of conditions. So, we can try to make an estimate and given that we are starting this analysis with national data on a large group of libraries operating under varying conditions a national estimate of inflation as it affects libraries would be useful.

For libraries, we can use the indexes--such as the Public Library Price Index (PLPI) calculated by Kent Halstead of the Research Associates of Washington and published in the Bowker Annual through 2003. The PLPI begins in 1992, that is, 1992=100. It was a serious effort to get at the real effects of currency changes on public libraries. Because the data in our public library series begin before that date, I used the Consumer Price Index. It seems, alas, that Halstead is no longer compiling the PLPI. It is probable that the CPI does not correctly estimate the effect of inflation on public libraries as closely as PLPI did but it is what we have. After discussing the CPI, I will compare it to the PLPI for the 10 years we have of the two.

Consumer Price Index (CPI) is published by the US Bureau of Labor Statistics. There are any number of series that can be used from that site and I chose "Consumer Price Index; All Urban Consumers - (CPU-U), U.S. city Average, All Items, 1984-84=100. I took the annual average (see column headed "AVG."). The data begin in 1913 and I chose these data because I believe that we can extend the series of public library data back and when we have data that far back, this column is the most useful.

I took the average index by year from that series starting in 1990 and calculated percent changes and calculated changes from 1990 and 1991 because both are needed for different series of library data. These figures are:

                 % 
               Change
               from
Year    CPI    1990
-------------------
2002   179.9   0.73
2001   177.1   0.74
2000   172.2   0.76
1999   166.6   0.78
1998   163.0   0.81
1997   160.5   0.81
1996   156.9   0.83
1995   152.4   0.86
1994   148.2   0.88
1993   144.5   0.90
1992   140.3   0.93
1991   136.2   0.96
1990   130.7   1.00

By multiplying each of these percent changes by the appropriate year and depending on the beginning year of the data, the value of the 1990 constant dollar can be compared with the approximate dollar value in a given year.

What is the relationship between the PLPI and the CPI? Here are the values for the years the two overlap. The PLPI comes from the 2003 Bowker Annual, p. 457. A simple regression equation gives an R² of 98.5 or 98.5% of the variance of one is "explained" by the other. This is a very high number so the CPI is a better predictor of price changes for public libraries than I would have thought. Plotting the residuals by year reveals a very slight U-shaped pattern.

             
Year    PLPI     CPI 
----------------------------
2002            179.9
2001    128.8   177.1
2000    122.6   172.2
1999    118.9   166.6
1998    116.8   163.0
1997    112.7   160.5
1996    109.6   156.9
1995    107.3   152.4
1994    104.1   148.2
1993    101.6   144.5
1992    100.0   140.3

Another way of looking at the effects of inflation can be seen at the US Bureau of Labor Statistics Inflation Calculator.

September 22, 2004