Inflation Indices
Description of Consumer Price Index (Including a Discussion of the Boskin Commission Report)
Change in Consumer Price Index is considered to be an objective measure of price inflation. The greatest financial impact of CPI is to quantify fair adjustments in income payments and wages to offset inflationary erosion of purchasing power. The analytical impact of CPI is most common in translating price series into inflation-free dollars.
CPI is not a direct measure of cost of living. Over-simplified, CPI assumes the same purchase of the same market basket in the same fixed proportions from month to month with a major update about every 10 years based on changes in consumer spending. The 1998 rebenching used the Consumer Expenditure Survey results for 3 years (1993, 94 & 95). About 4,800 families from across the country were interviewed quarterly in each of the 3 years to determine spending habits. To this were added another 4,800 families' comprehensive buying records for a 2-week period in each of the 3 years. All told, the purchasing behavior of some 29,000 individuals and families is used to determine the weight assigned each item in the index structure. The weighted items constitute the "basket" of goods and services purchased. Major groups are food and beverages, housing, apparel, transportation, medical care, recreation, education and communication, and "other goods and services."
The process of identifying a product and the outlet at which the price of the product will be tracked is rigorous. It involves a census of products and sales information from which a specific product is randomly selected. For example, the process of determining which soft drink to track is reported to involve the following steps. The name and package sizes of all soft drinks are recorded. Buyer patterns then determine the likelihood of a specific product and container type being selected and its contribution to the cost of a basket of goods. The randomly selected product is then priced monthly, bimonthly, or semiannually depending upon urban location and its weighted price becomes part of basket of goods and services. Both Kansas City and St. Louis are sampled on a semiannual basis.
The CPI is measured by civil servants within the Department of Labor's division of the Bureau of Labor Statistics (BLS) and is meant to be free of political influence. While politics might have influenced the Senate Finance Committee's formation in December 1996 of the Advisory Committee to Study the Consumer Price Index (Boskin Commission), at least one recommendation was supported by the BLS's own research and is a part of the formula used beginning January 1999. That change is the use of a geometric mean estimator and it will influence the computation of about 60% of consumer spending. The impact of this change is expected to be a reduction of about 0.2 percentage points in the annual rate of increase.
In a nutshell, the Boskin Commission Report (Michael Boskin Chairman) stated that the true cost of living has increased at a lower rate than CPI. In other words, the CPI has overestimated experienced inflation at a rate of about 1.1% per year. The Boskin report identified four types of bias (systematic error) in CPI.
- Substitution bias - Changes in what is purchased. Substitution bias results from the CPI use of fixed products even though consumer behavior changes rapidly. For example, if the price for beef increased very rapidly, consumers might begin to eat more poultry and the price of beef might therefore have little real effect on their cost of living.
- Outlet Substitution Bias - Changes in where goods are purchased. Many consumers now purchase more from discount retailers than was true in the past.
- Quality Change Bias - Quality improvement of products. As products become more efficient, durable, and reliable, the expenditure for their use, repair and replacement decreases.
- New Product Bias - Changes in available products. Products become available at a faster rate than changes in the CPI are made. Home computers, VCRs, microwave ovens, cell phones, cable TV, and many other products are introduced and become a part of the fabric of life more quickly than reflected in CPI changes in purchasing habits.
While the Boskin Commission report has critics, most recognize the existence of these biases and disagree primarily with estimated magnitude of each bias. Whether due to the Commission or its own research, the BLS will use geometric means at elementary aggregates level. Geometric means are commonly used to compute central tendency of index or ratio numbers. To compute a geometric mean, the n values are multiplied and the resulting figure is raised to the 1/nth power. Imagine 6 academic programs with student to faculty ratios of 20, 21, 22, 23, 24 and 45 to 1. The geometric and arithmetic means of this sample are 26.4 and 25.8. Now, imagine that the first school cuts its ratio - from 20:1 to 10:1. The geometric and arithmetic means are now about 23.5 and 24. A second possibility is that the last school cuts its ratio in half - from 45:1 to 22.5:1. The geometric and arithmetic means are now 23.5 and 22. What is most important to note is that the geometric mean was unaffected by which value was halved. This property is one of the reasons that it particularly useful when dealing with ratio, or rate of change computations.
Please note that the Boskin Commission report was not asserting that the BLS never attempted to reflect changes in consumer purchasing behavior or products, only that its rate of change was too slow. There would also be significant disagreement about the size of adjustment associated with changes in quality. Where the Commission and the BLS agreed was in use of geometric means and these will be incorporated in computation of CPI beginning January, 1999.
Reviewed 2022-04-06