This repository refers to the SAS computer code that was used to produce the results presented in the Armstrong et al. paper published in the 2022 American Economic Association (AEA) Papers and Processings (forthcoming). To run this code, access to internal BLS data are needed. References to internal BLS data files have been redacted in this code. While some variable names in the code are similar to those appearing in the Consumer Expenditure Surveys (CE) public use data (PUMD), they are not exactly the same. For example, the BLS internal variable for total food expenditures, ZFOODTOT, can be reproduced using the PUMD data by summing the values for the following variables: FOODCQ and FOODPQ. The internal data needed to run the code are stored on a BLS server in the following folder: "/Consumption_Measure/Data_READin/"
The BLS internal data are accessible to researchers with approved projects under the BLS Outside Researcher Program; see: https://www.bls.gov/rda/.