An official website of the United States government
Interim methodology updates
The results linked below for the year 2017 incorporate several proposed methodology changes relative to the December 2022 release of the distributions of PCE. These are summarized below. None of the adjustments below were made in previous publications of the distributional estimates. Pending feedback from experts and stakeholders, future updates will incorporate these revisions.
The distribution of motor vehicle maintenance and repair services, financial services furnished without payment, income loss insurance, and worker's compensation insurance are similar to McCully(2014) in that some of the PCE amount are distributed using a noncomparable CE series (e.g., motor vehicle insurance premiums) as an indicator. They differ from McCully(2014) in that our methods use only published PCE amounts and only CE microdata (McCully also used wage data from the Current Population Survey Annual Social and Economic Supplement).
Category/Topic | Methodological Update |
---|---|
Motor Vehicle Maintenance and Repair Services |
The PCE amount minus the CE amount is distributed based on motor vehicle insurance premiums in the CE. |
Financial Services Furnished Without Payment |
Services provided by Commercial Banks and Other Depository Institutions. PCE is distributed using consumer unit asset values for savings, checking, money market, CD’s, plus value of stocks, bonds, and mutual funds. Since these are only collected in the fourth interview, we impute these values to consumer units who do not have a fourth interview (or who reported having assets but had missing balances) using a statistical match based on before-tax CE-defined income quartile, age group indicators (under 30, 30-60, and over 60), and race indicators (white non-Hispanic, black non-Hispanic, Asian non-Hispanic, Hispanic of any race, and other).
Pensions: PCE is distributed using consumer unit values for pension contributions. |
Income Loss Insurance, Worker’s Compensation Insurance |
PCE amounts are distributed using wages and salaries collected in the first and fourth interviews. Values for the second and third interviews are carried over from the first interview by the BLS in the production of the CE data files. |
Higher Education Services |
The PCE amount minus the CE amount for college tuition is distributed based on the number of students in the CU (using IN_COLL from the member file). |
Pareto Adjustment for Upper Tail |
To mitigate underestimation of inequality, we follow the suggestion in Zwijnenburg, et. al (2022) and construct a type-I Pareto distribution. This distribution is applied to the top 5% of total spending after adjustments and imputations, but before scaling to match the NIPA totals. A shape parameter of 2 was chosen based on Zwijnenburg, et. al (2022) and our judgement. |
Interim Results
PCE distributional results for 2017 (last updated June 2023).
References
McCully, C. 2014. "Integration of Micro and Macro Data On Consumer Income and Expenditures," in Measuring Economic Stability and Progress, D. Jorgenson, J.S. Landefeld, and P. Schreyer, editors, University of Chicago Press - https://www2.census.gov/adrm/fesac/2014-06-13_mccully.pdf.
Zwijnenburg, J., Grilli, J., Engelbrecht, P., 2022. “Pareto Tail Estimation in the Presence of Missing Rich in Compiling Distributional National Accounts.” Paper prepared for the 37th International Association for Research on Income and Wealth General Conference, August 22-26, 2022 - https://iariw.org/wp-content/uploads/2022/07/Jorret-Joseph-Pao-IARIW-2022.pdf.