Computing CPI, an Imperfect yet Useful Indicator

Consumer Price Index is an indicator and a measure of the change in the cost of a fixed basket of products and services, including housing, electricity, food, and transportation.

There are five steps in calculating Consumer Price Index. The Bureau of Labor Statistics first survey the consumers which goods are more important to them then weight them accordingly in the basket of goods. Then researchers find prices of each good of this basket at different times. Next, using prices and quantities at different times (note that the quantities are fixed, only prices change), the people at BLS calculate basket's cost. They then set a base year and calculate CPI by dividing one specific year's price by base price then multiply the product by 100. Lastly, they compute inflation rate: if P0 is the current average price level and P−1 is the price level a year ago, inflation rate between two consecutive year would be (P0 - P-1) / P-1 x 100.

The producer price index (PPI), also published by BLS, is a measure of wholesale prices at the producer level for consumer goods and capital equipment (unlike CPI, PPI does not include services). Since PPI is the first inflation indicator published each month, it is often thought as a good predictor of CPI.

CPI is not a perfect measure. Firstly it is not accounting for new goods. When new products enter the market, people have more products to choose from and therefore higher value of a dollar and lower living costs. Since CPI is based on a fixed basket of goods, this change of living costs is not reflected on the CPI until BLS eventually include this good into CPI after some period of time. Secondly, there is substitution bias: The basket of goods remain the same despite people tend to switch from expensive goods to less expensive ones over the years. For example, in base year people might buy more yogurt than ice cream. However, yogurt prices rise and more people buy ice cream now but BLS will still use quantities in the base year to compute CPI, incorrectly reflecting an increase of living costs. Note that both reasons above are due to the inflexibility of the fixed basket of goods. Lastly, there is unmeasured quality change: quality of goods change over the years. Better quality leads to lower cost of living. Worse quality leads to higher cost of living. BLS does adjust prices according to changing quality to make CPI a more accurate measure of cost of living. But since quality is so hard to measure unmeasured quality change is often left out.

Social security recipients receive benefits that are tied to CPI. However, since goods such as drugs that seniors buy have higher inflation than other goods while goods like TVs that seniors don't buy are accounted into CPI, CPI is not a good measure for seniors and social security recipients.

Now we are going to see a comparison between GDP deflator and CPI. Most of the time they are similar but there are differences between the two. First, GDP deflator measures domestic output while CPI measures consumer's products so goods like helicopters bought by government are in GDP but not in CPI. Meanwhile import goods bought by consumers are in CPI but not in GDP. Thus, an increase of oil prices would raise CPI substantially more than GDP deflator. Secondly, CPI measures a fixed basket of goods while GDP deflator measures goods that change in quantity over the years.