Analysis of the Impact of Secondary Source Data on the CPI’s Gasoline Standard Errors

• Activity Number: 525 - Applications of Statistical Measurement to a Range of Social Issues: From Cost of Living to Social Unrest
• Type: Contributed
• Date/Time: Thursday, August 11, 2022 : 8:30 AM to 10:20 AM
• Sponsor: Social Statistics Section
• Author(s): Jenny FitzGerald*
• Companies: Bureau of Labor Statistics
• Keywords: Variance Estimation; Non-Probability Sample; Stratified Random Groups

Abstract

The Consumer Price Index (CPI) estimates the change in prices over time of the goods and services U.S. consumers buy for day-to-day living. In the past, the program estimated its price change estimates for gasoline using about 4,000 price quotes selected from probability samples each month. In June 2021, the CPI replaced its sampled price data for its gasoline index with data from a secondary source. Currently, the secondary source provides the CPI with millions of gasoline prices each month. Given the substantial increase in the number of prices, one would expect to see a significant decrease in the gasoline standard errors (SEs). However, the CPI has not yet seen such a decrease. In this study, we investigate why the gasoline standard errors did not decrease with the increased number of prices from the secondary source. Specifically, the impact of increasing the number of replicate samples (or variance PSUs) assigned to the CPI Index Areas (variance strata) on the gasoline SEs from the secondary source data is investigated.