Author
We all know that some prices change more frequently than others: Digits at the gas pump vary daily, but haircut prices rarely budge. Haircuts, then, are a sticky-price item and gasoline is not.
To draw a systematic picture of price changes and the factors associated with greater or lesser degrees of price stickiness, Bils and Klenow looked at volumes of data compiled by the Bureau of Labor Statistics. Every month, to construct the consumer price index, the BLS collects retail prices from about 22,000 outlets across the country for more than 80,000 items. "Everything from broccoli to brain surgery," noted Bils and Klenow.
In a 2002 National Bureau of Economic Research paper analyzing BLS data from 1995 to 1997, Bils and Klenow observe that prices change quite often, nearly three times a year for the average product—much more frequently than previous research has suggested. Concentrating on nonhousing consumption and weighting for each item's significance in terms of total consumer expenditure, Bils and Klenow find that "the median consumer item ... changes prices every 4.3 months. For 87 percent of consumption prices change more frequently than once a year."
Earlier studies have reported less frequent price changes. For example, one 1995 survey of monthly prices in a dozen mail-order catalogs between 1953 and 1987 found that prices changed, on average, every 14.7 months. A 1998 study of 200 businesses noted that firms reported adjusting prices roughly once a year.
A range of rigidity
The 4.3 month average found by Bils and Klenow obscures a wide range in price stickiness. Services, for instance, have far less price flexibility than do goods. Goods change their prices every 3.2 months on average, but services take over twice as long (7.8 months). "The lower frequency of price changes for services could reflect the lower volatility of consumer demand for them," the economists suggest.
Looking at six broad expenditure categories used by the BLS, they find that transportation prices (new cars and air fares, for example) are quite flexible—changing every 1.9 months—while medical care (drugs and doctors' services) and entertainment (movie tickets, newspapers and books) change prices rarely, every 14.9 months and every 10.2 months, respectively.
Bils and Klenow determine another key distinction by dividing goods by their levels of processing. Products with little value added beyond a primary input, like gasoline or fresh fruit, show much more price flexibility than do more highly processed goods and services: Raw goods change price every 1.6 months, while processed goods change price every 5.7 months.
Duration of Prices by Category of Consumption, 1995-97 |
||
| Category |
Median Duration |
Share of the CPI |
All Items* |
4.3 |
71.2 |
Goods |
3.2 |
30.4 |
Services |
7.8 |
40.8 |
Food |
3.4 |
17.1 |
Home furnishings |
3.5 |
14.9 |
Apparel |
2.8 |
5.3 |
Transportation |
1.9 |
15.4 |
Medical Care |
14.9 |
6.2 |
Entertainment |
10.2 |
3.6 |
Other |
6.4 |
7.2 |
Raw |
1.6 |
12.0 |
Processed |
5.7 |
59.2 |
* The data exclude shelter, which represents the remaining 28.8 percent of CPI. Source: Bureau of Labor Statistics, U.S. Department of Labor, via Bils and Klenow, 2002 |
||
Market structure and flexibility
The raw/processed distinction turns out to be quite significant when Bils and Klenow look into other factors that might be related to price flexibility, especially market structure. Standard economic models of price adjustment predict that more competitive markets will exhibit greater price flexibility because their firms face more elastic demand—that is, customers can easily go elsewhere if they don't like a given firm's price.
Bils and Klenow test the statistical relationship between frequency of price changes and various measures of market competitiveness: Concentration ratios (the share of the market held by its largest four firms), wholesale markups (larger in less competitive markets) and rates of new product substitutions (greater product turnover). All three are related as expected. "Each coefficient has the anticipated sign and is economically and statistically significant," they wrote, meaning that more concentrated markets, higher product markups and lower product substitution rates are all associated with lower price flexibility.
But when Bils and Klenow control for the effect of an item being either raw or processed, the first two variables (market concentration and product markups) lose their statistical significance, implying that those variables simply masked the more important relationship between product processing and price rigidity. Taken together, the raw/processed goods variable and the product turnover variable explain 56 percent of the variation in price flexibility. So, in sum, the less processed a good is, and the more often it's faced with a substitute product, the more often its price is likely to change. Keep that in mind as you're filling your tank on the way to the barber.
