| | | | | | | Curiosity 2.3: How Do the GDP Deflator and the CPI Differ? | | | | | | | | | | The GDP deflator and the CPI are both price indices. The CPI is particularly well known because it measures changes in the cost of living and so is of more personal interest to individuals, and for that reason it is the price index normally used to calculate the rate of inflation. There are three major differences between these two indices. First, the CPI reflects prices of only consumer goods and services, whereas the GDP deflator calculation includes prices of all output. Second, the CPI incorporates prices of imports, excluded from the GDP deflator calculation. And third, the GDP deflator allows the output basket to change each year, whereas for the CPI it changes only every two years and with a lag. Two minor differences are of note. First, once published the CPI is never revised, as are other statistics, because it is often used in contracts to measure cost-of-living wage adjustments. And second, these price indices have different base years, currently 1992 for the GDP deflator, and 1982-84 for the CPI. Despite these differences, both indices produce very similar inflation measures. | | | | | | | | | | Here is a technical detail for those really curious about price index calculations. The GDP deflator actually used is a chained index, calculated slightly differently than described in the text (but the resulting numbers are very similar). Real growth from one year to the next, say from 1992 to 1993, is calculated first by measuring and comparing the two years' physical outputs in 1992 prices and then by measuring and comparing the two years' physical outputs in 1993 prices. The geometric average of these two growth rates is used to measure real growth between these two years. This procedure is done for all adjacent years, and then the base-year value of GDP is augmented by these growth rates year (i.e., "chaining" the growth rates together) to produce the real GDP series. This series is then used to produce the corresponding "chained" price index by dividing nominal GDP by real GDP. This procedure reduces bias caused by changes in output bundles from year to year. | | | | | |
| | | | | change, the purchasing power of a dollar has increased because it can buy more computing, but the price index does not decrease to reflect this increase. Quality changes that take the form of introducing entirely new products present similar problems. The Bureau of Labor Statistics makes an effort to deal with these by making imputations for quality changes in high-profile items such as automobiles, apparel, and computers, and by blending new products into the typical bundle, but clearly changes in quality remain a problem. | | | | | | | | | An interesting example is often used to clarify this point. Suppose you were given $10,000 and told that you may spend it on 1990-quality goods and services at 1990 prices, or on 1999-quality goods and services at 1999 prices. If you choose the latter option you are implicitly saying that quality increases between 1990 and 1999 have offset the price increases and so "inflation" between 1990 and 1999 is zero or negative insofar as it affects the purchasing power of your dollar. | | | | | | | | | Another use of the CPI is as a benchmark to measure changes in relative prices. In 1980 the price of a large pepperoni pizza was about 8 dollars, and in 1997 was about 11 dollars, an increase of about 38 percent. Has the price of pepperoni pizza increased or decreased | | | | |
