Improving the Accuracy of Economic Measurement with Multiple Data Sources: The Case of Payroll Employment Data

Working Paper: NBER ID: w26033

Authors: Tomaz Cajner; Leland D. Crane; Ryan A. Decker; Adrian Haminspuertolas; Christopher Kurz

Abstract: This paper combines information from two sources of U.S. private payroll employment to increase the accuracy of real-time measurement of the labor market. The sources are the Current Employment Statistics (CES) from BLS and microdata from the payroll processing firm ADP. We briefly describe the ADP-derived data series, compare it to the BLS data, and describe an exercise that benchmarks the data series to an employment census. The CES and the ADP employment data are each derived from roughly equal-sized samples. We argue that combining CES and ADP data series reduces the measurement error inherent in both data sources. In particular, we infer “true” unobserved payroll employment growth using a state-space model and find that the optimal predictor of the unobserved state puts approximately equal weight on the CES and ADP-derived series. Moreover, the estimated state contains information about future readings of payroll employment.

Keywords: Payroll Employment; Data Sources; Measurement Error; Labor Market

JEL Codes: C53; C55; C81; C82; J11; J2


Causal Claims Network Graph

Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.


Causal Claims

CauseEffect
Combining the CES and ADP data series (E01)Reduces measurement error inherent in both data sources (C83)
Statespace model (C32)Inference of true unobserved payroll employment growth (E24)
Optimal predictor assigns approximately equal weight to both data sources (C51)More accurate estimate of employment changes than either source alone (J60)
Combined approach (C59)Enhances predictive accuracy (C52)
ADP-derived estimates during the Great Recession (E01)More accurate measure of employment declines compared to the CES data (E01)
Combined model (C59)Improves understanding of current and future payroll gains (J39)
Integration of data sets (C55)Provides a more accurate estimate of employment changes (J23)

Back to index