Academic journal article Monthly Labor Review

Interarea Price Levels: An Experimental Methodology: Differences in Relative Price Levels for Areas of the United States Can Be Estimated with a Modified Country-Product-Dummy (CPD) Method Often Used in International Comparisons of the Purchasing Power of Currencies; CPI Observations and CE Weights Are Used to Estimate Experimental Price Level Differentials for 2003 and 2004

Academic journal article Monthly Labor Review

Interarea Price Levels: An Experimental Methodology: Differences in Relative Price Levels for Areas of the United States Can Be Estimated with a Modified Country-Product-Dummy (CPD) Method Often Used in International Comparisons of the Purchasing Power of Currencies; CPI Observations and CE Weights Are Used to Estimate Experimental Price Level Differentials for 2003 and 2004

Article excerpt

Although the Consumer Price Index (CPI) survey is not designed as an interarea survey, it is possible to use its price observations and sampling weights to obtain estimates of area price levels for various categories of consumer expenditure. Combining these estimates across the expenditure categories produces an experimental index of the price level differences for the areas. This was first done some 15 years ago by Mary Kokoski, Patrick Cardiff, Brent Moulton, and Kim Zieschang using 1988--89 prices, and more recently by Bettina Aten using 2003 prices. (1) This article shows a shortcut approach for calculating the 2003 interarea prices and repeats the exercise for 2004. It also describes the methodology, presents the detailed results for 2003, and provides a comparison with the 2004 results.

General methodology and data

The headline CPI (the CPI-U) measures the average price change for urban consumers, who comprise approximately 87 percent of the total U.S. population. The CPI collects prices in selected urban areas throughout the country from about 23,000 retail and service establishments. In addition, data on rents are collected for about 50,000 renter-occupied housing units. (2) Each price observation has a reference date and represents a good or service that is uniquely identified by a set of characteristics, including the geographic area. Not all areas have the same goods and services priced; rather, items are selected within categories to represent those sold in each area. Each observation also has a weight. The weight is an estimate of the amount of consumer expenditure the observation represents. In other contexts this is called the representativity (3) of the price in the framework of the probability sample from which it is drawn.

Because there are multiple quotes for most observations, there are in total more than 1 million price quotes per year. Nonrent items are priced monthly or bimonthly; for rents, there are two quotes per year for each dwelling, taken 6 months apart. (See table 1.) Due to the multiple pricings, there are approximately 245,000 unique annual observations, each identified by outlet, quote code, and version. The price of these unique observations is the geometric average of all of its prices collected over the year.

The CPI is organized in a four-tier system of increasing detail: major group, expenditure class, item stratum, and entry level item (ELI). Many ELI's make use of a fifth tier called a cluster. These observations are organized into eight groups of goods and services: housing, transportation, food and beverages, education, recreation, medical, apparel, and other. Table 1 also shows the number of item strata in each group. An example of an item stratum within the housing group is major appliances. However, the actual price observations are on specific major appliances such as refrigerators, washers, dryers, and so forth, called ELIS. An item stratum corresponds to the lowest level of detail for which expenditure weights are available in all 38 geographic areas and is therefore the target level of the estimation process described in the next paragraph. The two columns in table 1 labeled "Number of regressions" are explained below.

Step one: estimating price parities (item stratum "prices" in each area). The first step of the estimation process consists of obtaining price levels, also known as price parities, for each item stratum in each geographic location, such as flour in Boston or women's shoes in Chicago. Price parities refer to the predicted dollar value of an item stratum with particular characteristics, while price levels are generally expressed relative to one area, or the average of all areas. (4) For example, suppose the estimated price parity for an ounce of white flour in a 1 pound bag sold in a supermarket in Philadelphia is $0.01, and in Honolulu, $0.02. If the average price across all areas is $0.015, the price level for flour in Philadelphia will be 0. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed

Oops!

An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.