How to parse jobs data like a pro
Add Axios as your preferred source to
see more of our stories on Google.
Illustration: Aïda Amer/Axios
One of the great challenges in assessing the economy is separating the signal from the noise. When a surprisingly strong or weak indicator comes out, is it a sign that the economic ground is shifting beneath our feet, or just randomness?
Why it matters: A top economics research shop — the team at Goldman Sachs — issued a report this week on how to discern the underlying message in the jobs data.
- The details of their calculations aside, their process is a great lesson in how any given piece of new data should (and shouldn't) cause us to adjust our view of where things stand.
Driving the news: Friday morning, the Bureau of Labor Statistics will release the March employment report, which forecasters surveyed by Bloomberg expect to show 212,000 jobs added last month and the unemployment rate unchanged at 3.9%.
- The raw numbers will be blasted out in news headlines at 8:30am.
- But if the numbers over- or under-perform those expectations, you probably shouldn't adjust your view of the economy too much, at least if you use the arithmetic Goldman economist Manuel Abecasis recommends based on historical data.
The big picture: To discern underlying job growth, Goldman advises putting a 75% weight on three-month average payroll growth, plus a 25% weight on the last nine-month average of employment growth according to the separate, more volatile, household survey (adjusted to be apples-to-apples comparable to payrolls).
- All that weighting and averaging dilutes the importance of the latest month's number, while making what remains more predictive of the months ahead, according to the analysis.
- For example, if the March job growth number comes in at 312,000, it would be a shockingly high number at first glance — but it would only pull the three-month average from 265,000 to 272,000, as robust December job growth falls out of the three-month average.
Yes, but: At turning points in the economy — when job growth is running below 75,000 a month, signaling the economy is potentially in the vicinity of a recession — the formula changes slightly, putting more weight on the adjusted household survey.
- At those moments, the Goldman analysis puts 60% weight on the three-month average payroll growth and 40% weight on six-month adjusted household employment growth.
- That's because payroll numbers are distorted by estimates of how many new firms are being created or closed — estimates that become less accurate in an economic downturn.
The report also offers other rules of thumb:
- For example, to discern underlying wage growth, they put 85% weight on nine-month total average hourly earnings growth, plus 15% weight on three-month average wage growth for non-supervisory workers.
The bottom line: We in the media tend to be an excitable bunch, so news headlines often overstate the importance of each new data point. To understand what's really going on, you want to look at multiple indicators, averaged over multiple months.
