That’s because the US Department of Labor’s Bureau of Labor Statistics (BLS) is set to unveil four changes to its closely watched May payrolls data. According to a Reuters report, Friday’s payroll figures will incorporate:
- a new industry classification system
- the final phase-in of a new sampling method
- updated seasonal adjustments
- changes to the way government jobs are counted.
Because of the alterations, analysts will find it harder to judge how the employment picture changed in May from April, when the jobless rate rose to 6.1% as payrolls slid 48,000, but the economist overseeing the changes said the report will be more reliable. “We feel pretty ready, but we’re still checking everything. It is a lot of things to be changing all at once,” Pat Getz, BLS division chief for current employment statistics, told Reuters.
The moves will affect only one of the two surveys in the report, the “establishment” survey with data on nonfarm payrolls, hours and earnings. The household survey, used to calculate the unemployment rate won’t be touched. Some analysts expect the changes to result in an upward revision to recent weak jobs data, since the household survey has painted a less bleak picture than the establishment poll.
Among the alterations, the North American Industry Classification System will replace the outdated Standard Industrial Classification system and is expected to provide more detail on the burgeoning services sector. New categories like “Information” will incorporate Internet providers, publishing, telecommunications and broadcasting.
The May data will also boost accuracy by introducing concurrent seasonal adjustment using monthly estimates that include those for the current month, rather than projecting seasonal factors six months ahead.
Finally, the government sample will be changed to include civilian Defense Department employees in the payroll count, a move expected to boost nonfarm employment levels, if only slightly, given the recent run-up in the defense budget. The government series will also be estimated from a sample of federal employers and will be benchmarked annually to counts from tax records, in line with other industries.