The Noise Behind the New Jobs Report
James Sherk /
The new jobs report shows employers cranked up hiring in June, creating 287,000 net new jobs. But the last jobs report, for May, showed the economy slowing to a crawl, with employers adding only 11,000 net new jobs in the month.
What explains these large gyrations? Probably statistical noise. Americans shouldn’t read too much into May’s discouraging job numbers or June’s hopeful ones.
Both the payroll survey (which estimates job growth) and the household survey (which estimates the unemployment rate) are large-scale surveys. Like any survey, they have a margin of error.
Nine times out of 10, the payroll survey estimate is within plus or minus 115,000 jobs of the actual job growth rate. Similarly, the household survey is within plus or minus 0.2 percentage points of the true unemployment rate 90 percent of the time.
Consequently, statistical noise can cause the household and payroll surveys to fluctuate considerably month to month, even if job growth and unemployment do not change. And 10 percent of the time, the surveys will report estimates outside this confidence interval.
This statistical noise appears to explain the recent whiplash. Job growth plunged in May and rebounded in June. However, the average across both months was plus 149,000 jobs—very close to average job growth in March and April (plus 165,000 jobs).
Many measures of labor market health showed this pattern: dropping in May and rebounding in June.
For example, the labor force participation rate for 25- to 54-year-olds fell 0.2 points in May. It then increased 0.2 points in June. That increase simply returned the participation rate to its April level.
Similarly, the seasonally adjusted number of Americans working part time for economic reasons increased by approximately 500,000 in May and dropped by approximately 600,000 in June.
The unemployment figures also show this pattern. The number of unemployed Americans fell 484,000 in May and the unemployment rate dropped 0.3 percentage points, mostly driven by Americans leaving the job market and no longer looking for work.
Unemployment increased by 347,000 workers in June and the unemployment rate increased 0.2 percentage points, mostly driven by Americans returning to the labor force and looking for work.
Statistical noise almost certainly explains these swings.
The Bureau of Labor Statistics’ employment reports are surveys. Like all surveys, they have sampling error. Rather than focusing on monthly movements, Americans should look at the long term trend.
These trends show job growth slowed over the past year, but remain far from recessionary levels.