This article is only available to Macro Hive subscribers. Sign-up to receive world-class macro analysis with a daily curated newsletter, podcast, original content from award-winning researchers, cross market strategy, equity insights, trade ideas, crypto flow frameworks, academic paper summaries, explanation and analysis of market-moving events, community investor chat room, and more.
For government agencies and economic policymakers, assessing the performance of the labour market significantly aids their decision making. Every post-war recession in the US has been associated with a year-on-year drop in payroll employment, while outside of recessionary declines, the year-on-year payroll growth has always been positive. It is perhaps because of this that the monthly release of the official US payrolls data is the most market-moving US data series. As you might imagine, then, the accuracy of this measurement has a lot riding on it; and researchers from the New York Fed have recently made proposals for improvement in recently published paper entitled: “Improving the Accuracy of Economic Measurement with Multiple Data Sources: The Case of Payroll Employment Data”.
How US Employment is Officially Measured
The traditional approach to measuring monthly US employment status is the Current Employment Statistics (CES) from the Bureau of Labor Statistics (BLS). It is published a few days after each reference month and is based on a survey of 500,000 private establishments covering 23% of all US private employment.
A more accurate measure of employment, however, is the Quarterly Census of Employment and Wages (QCEW). This has near complete coverage of all establishments and is used for annual benchmark revisions to the CES data. The problem is that the data is quarterly and published with a lag of two quarters.
Meanwhile, the private payroll processing company ADP also provides employment data – both at a weekly and monthly frequency. It is based on actual data, rather than a survey, and captures new firms better than CES. However, it does not have complete coverage of the labour market.
How Close are Monthly Data to Final Revised Employment?
The researchers used the weekly ADP data to construct a monthly ADP-FRB measure of payrolls. This is different from the publicly reported monthly ADP series. They consider QCEW as the ‘truth’, and evaluate the ability of ADP-FRB and CES to make forecasts of the QCEW. They find that CES data does a better job of predicting the QCEW data. However (and most interestingly), they found that the ADP-FRB data did a better job of capturing the ‘true’ employment picture (QCEW) during the Great Recession of 2008-2010 (as shown in the chart below).
Figure 1: Real-Time Vs. Current Vintage Estimates
How Best to Predict the Monthly Payrolls (CES) Data?
The QCEW data is quarterly and lagged, and so policymakers and investors often care more about the higher frequency monthly payrolls data. Running regressions, the researchers found that incorporating the ADP-FRB data generated improvements in the forecasting accuracy of final monthly payrolls data (CES). Additionally, during the 2008 Recession, the regression also suggested that the ADP-FRB index provided a more accurate measure of the employment declines. However, the biggest improvement to forecasting was to include the market expectations of the upcoming payrolls release (see the table below for full model).
Best Approach is to Combine Both Measures
Finally, the authors try to determine the best approach for policymakers to track employment in real-time. They build some state-based models and find that a combined measure of the CES and ADP-FRB data gave the most precise estimate of the current pace of employment growth.
Figure 2: Forecasting Monthly Employment Changes
Reviewed by Cheryl Liu, Research Analyst at Macro Hive. Before coming to Imperial for her MSc finance degree and joining Macro Hive, Cheryl worked in investment banking and commercial pricing actuary whilst pursuing her Economics undergrad at NYU.
Cheryl Liu can be contacted here.
(The commentary contained in the above article does not constitute an offer or a solicitation, or a recommendation to implement or liquidate an investment or to carry out any other transaction. It should not be used as a basis for any investment decision or other decision. Any investment decision should be based on appropriate professional advice specific to your needs.)