- July payrolls fell an estimated 23,000 and unemployment stood at 4.1%.
- The payroll survey counts jobs while the household survey counts people.
- The May and June payroll gains were revised down by a combined 103,000.
Inside this story 2 perspectives
American employers shed an estimated 23,000 payroll jobs in July, while the unemployment rate eased from 4.2% to 4.1%. Both numbers appeared in the same report on 7 August. On their own they make for an awkward conversation. Did work get harder to find or a little easier? The statisticians who produced them used a less emphatic description of July than either number might suggest. Both changed little.
The US Bureau of Labor Statistics asks employers about jobs and households about people. A position on an employer’s payroll counts in the first exercise. A person with two eligible jobs can therefore appear twice. In the household survey that same person is one employed individual. A self-employed person can appear on the household side without occupying a position counted in the payroll measure.
Who is in the denominator?
The unemployment rate uses the labour force as its denominator. That includes people who are working and those who qualify as unemployed under the survey’s definitions. It does not include every adult who would in principle like a job. Someone who has stopped looking generally drops out of the labour force measure. Under the survey rules, people on temporary layoff can be counted as unemployed without having searched for work.
In a simplified town, imagine that five people in a labour force of 100 are unemployed. If one stops looking and exits the labour force, four are unemployed out of 99. The rate falls, although none of the four still looking has found a job. The example demonstrates one way the arithmetic works. It does not establish that every change in the national rate in July came from people leaving the search.
Look back before looking ahead
Perhaps the most consequential passage in the release concerned two months that had already made headlines. The bureau revised May’s payroll gain from 129,000 to 63,000 and June’s from 57,000 to 20,000. Together that was 103,000 fewer net jobs added than those earlier figures had indicated. An employer considering expansion, or an analyst adjusting a demand forecast, now has a different starting point even before deciding what to make of July.
Conditions also varied by line of work. Retail employment fell by 19,000 in the month while health care added 22,000, according to the bureau. One figure that mixes those industries can conceal the very difference a job applicant cares about. These are net changes across employers. Neither number tells us how many individuals started a job at a particular business.
Financial markets can react to the release immediately. The people producing the estimate will continue receiving information and may revise it. For an overseas reader following US demand, July’s release supplies three different views. They cover households’ participation, employers’ payrolls and a revised account of the spring. Each has something to say about the economy, provided the question fits the survey.



