Improving Real-Time Employment Estimates: A State-Space Signal Extraction Approach

Student First Name: 
Matthew
Student Last Name: 
Zahn
Student Picture: 
Matthew Zahn
Project Picture: 
Matthew Zahn at Research Days 2015
Expected Year of Graduation: 
2015
Department/Major: 
Economics
Student Team Members: 
N/A
Mentor(s): 
Professor Tara Sinclair, Associate Professor of Economics and International Affairs
Other Team Members: 
N/A
Fun Fact About Yourself: 
I was an extra in House of Cards Season 3! Also this project won 1st place in the undergraduate Economics category at Research Days 2015!
Project Abstract: 

Over the last several years, real-time estimates of employment have become particularly important for evaluating the state of the macroeconomy. Research has shown, however, that the initial data released by the U.S. Bureau of Labor Statistics (BLS) for payroll employment are heavily revised and may not provide accurate estimates in real-time. At the same time, the BLS releases another estimate of employment based on a separate orthogonal survey of households. This survey has a much smaller sample size, but it is only revised for seasonal adjustments. In this paper, we propose a simple signal-extraction approach to improve upon the real-time estimate of payroll employment related to the method used by Aruoba et al (2013) for GDP. This approach produces a better real-time estimate of the final employment numbers, is simple to implement, and has fewer caveats than when applying a similar method to GDP.