MainStreet Macro: Where do jobs numbers come from?

Monthly jobs reports are among the most closely watched estimates of U.S. economic health. Even in normal times, they have the power to move stock prices, influence Fed interest rate policy, and shape spending priorities at the local, state and national level.

There are two important monthly measures of employment.  The official government non-farm payroll report is produced by the Bureau of Labor Statistics, rain or shine, on the first or second Friday of every month.

The ADP National Employment Report comes out two days prior to the government release. ADP’s payroll database captures about a fifth of U.S. private payroll employment, allowing us to provide an early look at the employment situation. We’ve been releasing the report every month for 15 years.

In today’s blog, we provide a peek under the hood of these two critical indicators and discuss their similarities and differences. 

Here are three things to know about how we measured employment during an unprecedented time in modern economic history.

1. The BLS and ADP reports are similar in scale but differ in approach

ADP’s employment report and federal jobs data from the BLS both do a good job of capturing U.S. labor activity by region, employer size, and industry composition. 

The difference is that the BLS estimate is based on a representative survey of establishments, a questionnaire sent to 144,000 business and government offices that encompass nearly 700,000 worksites.

ADP’s NER is based on the actual payrolls of 460,000 businesses representing 26 million workers. Differences in the data mean that the BLS and the NER move in a similar direction but not in lock step.

In the six months leading to the March 2020 lockdown, the BLS and NER measurements of employment were within 9% of each other. To put it differently, the average monthly difference boiled down to just 65,000 jobs in a market of 130 million workers.

2. The pandemic caused massive employment swings

The pandemic upended the labor market, causing firms to shed 20 million workers in April of last year.  Equally unprecedented were the number of workers temporarily laid off.  Nearly 90% of jobs lost that month were due to temporary layoffs, according to the BLS.

As we’ve noted, the NER measures the number of active employees on a company’s books. When companies lay off or furlough workers like they did last April, those employees typically remain on the payroll and count as active even if they’re not actually getting paid.

Employers expected to recall furloughed workers once the pandemic was under control. But as time wore on, more of those temporary layoffs became permanent. Discerning between permanent and temporary job loss became a big challenge.

Though it’s still elevated, the number of companies laying off workers has come down a lot since the pandemic’s early days.

That’s not only a good sign for the jobs recovery and millions of displaced workers. It also makes it much easier for BLS and the NER to measure the employment situation and distinguish between permanent and temporary separations.

3. The BLS and NER show a similar long-run picture of the jobs recovery

Over the last 18 months we’ve seen tremendous volatility in the job market. The economy gained 1.1 million jobs in July before backpedaling significantly to just 235,000 jobs added in August, according to the BLS.

Given that higher-than-usual, month-to-month volatility, it’s useful to take a long view when comparing BLS and the NER.

BLS has reported private sector payroll gains of roughly 4 million this year through the end of August. This comes pretty close to the approximately 3.8 million gain tallied by the NER over the same time period.  That’s a difference of less than 5% over the course of nine months.

My Take

This week, ADP and the BLS will report job numbers for September, data that capture the collective hiring, firing, and churn of U.S. employers.

Data from private sources such as ADP, when used in combination with government data, have never been more important for capturing labor market dynamics.

The job market has become more complex as the workplace evolves. The labor force is smaller than before the pandemic, thanks to an increase in early retirees and an exodus of women. More people are working remotely, and more businesses are investing in technology that is changing the nature of some service sector jobs and reducing demand for others. 

And the pandemic continues to cause economic pain. Last month we got a worrying signal from BLS — the economy created no new leisure and hospitality jobs in August. Economists, myself included, expected leisure and hospitality, one of the pandemic’s hardest-hit sectors, to lead the labor recovery as infections came under control.

That was put on pause by the delta variant’s surge in late July. The sector’s ability to regroup and make gains is a necessary component of returning to pre-pandemic levels of employment.

Employment is an important indicator of the economy’s health. Jobs data shows us if the recovery is helping the people who were put most at economic risk during the pandemic – service sector and frontline workers.

We’ll dive into ADP’s payroll employment data again this Wednesday to help businesses, policymakers and consumers understand what’s happening and what’s next for the U.S. recovery.