Nowcasting of Advanced Estimates of Quarterly US Private Consumption of Services with Traditional Indicators and Credit Card Payments Data

• Activity Number: 345 - Advances in Macroeconomic Nowcasting and Forecasting: Role of Traditional and Nontraditional Indicators and Big Data
• Type: Topic-Contributed
• Date/Time: Thursday, August 12, 2021 : 10:00 AM to 11:50 AM
• Sponsor: Business and Economic Statistics Section
• Author(s): Baoline Chen* and Kyle Hood
• Companies: U.S. Bureau of Economic Analysis and Bureau of Economic Analysis
• Keywords: Nowcasting; Real time data; Credit card payment information; National accounts estimates

Abstract

This paper assesses whether the inclusion of high frequency credit card payments data might improve the accuracy of nowcasting models for the advance estimates of quarterly U.S. private consumption of services (PCS) at detailed component level. Credit card payments data capture a broad range of spending activities and are available on a very timely basis, making them a suitable indicator for current economic activities especially during sudden sharp short-term economic fluctuations. To investigate the potential usefulness of credit card payments data in nowcasting PCS, we set up a baseline bridge equation model, utilizing information on lagged growth rates of the quarterly target variables and the traditional monthly indicators. Based on the estimated baseline model, one-step-ahead out-of-sample predictions at the detailed level of PCS are generated. The nowcasting exercise is performed again after adding credit card payments variable to the model. Finally, the accuracy of the two models is evaluated according to the root-mean-squared nowcasting errors. The comparative study is conducted using real time data from the U.S. national accounts from 2009Q2 to 2020Q4.