No index is perfect, even of consumer prices.
All give useful information, but none gives a perfect indication of changes in the prices faced by any specific household. And given human nature, some people always will argue that such indexes wildly understate true price changes. Others believe the numbers are blatantly manipulated for political purposes. Still, price indexes are important tools for modern economies.
An example of their politicization, however, surfaced recently when negotiations between the Obama administration and Congress supposedly produced agreement on changing the particular version of the Consumer Price Index used to calculate cost-of-living changes in Social Security payments.
The differences between the “Consumer Price Index for Urban Wage Earners and Clerical Workers” and the “Chained Consumer Price Index for all Urban Consumers” are so technical that few economists could correctly describe them, let alone members of Congress. But merely mentioning the fact that the change would reduce Social Security COLA adjustments even slightly was enough to bring Nancy Pelosi and other Democrats out in force. Whether it will be part of any final agreement is unknown.
Just what is all the kerfuffle about?
Let’s review how a consumer price index is constructed. Statisticians first compile a list of all the things people buy: food, clothing, housing, medical care, transportation, education, housewares, appliances, legal services, recreation and so forth. Within each general category, specific items must be listed. This list is called the “market basket” of goods and services on which the index is based.
The relative weight of each item also must be determined. A 5 percent change in the price of table salt should not have the same effect on the overall index as a 5 percent change in the price of gasoline. So each item on the list is assigned a numerical coefficient equal to the proportion of total household spending on it. The coefficients must add up to one.
The next step is to survey many retail establishments to determine the actual prices at which each of the items in the market basket is selling in a specific period, such as a month or quarter.
The index is set at 100 for this base period. The prices for the first period are multiplied by their respective weights. Then the whole process is repeated again in the next period. The weighted total for the second period is compared to the initial one. If it has increased by 1.2 percent, the index now stands at 101.2. This is repeated every period, comparing the weighted total cost of the market basket compared to the corresponding sum in the base period, and adjusting the index accordingly.
Note that the basket is based on national averages and may not represent the actual purchases of any specific household. Some people don’t buy alcohol. Some don’t get a pet groomed or eat tuna fish or go bowling. Some have medical expenses three times the national average and others one-fifth the national average. Some put 1,000 miles a week on their car and others 50. Some have efficiency apartments and others 5,000-square-foot estates.
So if anyone complains their costs went up more than the CPI, they may well be right. Those whose costs are rising less tend to ignore that fact, just as those who are concerned about milk or gas or meat going up tend to not recognize that prices of other things are going down or not going up as fast as the average.
Also note that people’s consumption “basket” changes over time. When the Bureau of Labor Statistics started doing this a century ago, buggy whips were common, but vascular stents were not. Vinyl records were big in the 1960s, but cell phones and pesto sauce were not. And when did you last get the valves ground on your car?
These changes introduce innumerable problems. Some consumption changes are due to changing tastes – less lamb and more turkey over the past 50 years– or to changing technology – Betamaxes out, TiVos in. But changes also stem from changing prices.
People may be eating more hamburger and less steak simply because steak has gotten expensive. People may be riding the bus more and driving less because gas is high. To simply change the weights on items in the basket is to paper over the true effects of rising prices. But to keep beef or pasta consumption weights at the levels they were in the 1960s, for example, would skew the true effects of changes in these prices
The difference between the specific CPI currently used for COLA adjustments and the “chain-weighted” CPI lies in how such changes in what people consume are introduced into the basket. The chain-weighting does so continuously on a sort of moving-average basis. Most statisticians and economists think the newer chain-weighted index is more accurate, but the older one has been enshrined by a third of a century of use.
Moreover, using the chain-weighted index would have increased a Social Security check from $1,000 in 2000 to 1,226.63 in 2011. But that is $33.84 less than the $1,260.47 the check would be using the current index – enough to damn it in the eyes of many congressional Democrats. That is a pity, even though changing the index would do little to solve overall budget problems.
© 2011 Edward Lotterman
Chanarambie Consulting, Inc.