Intermoney | The US labour market took centre stage again on Tuesday with the release of the annual revision of payrolls created through March of this year. Given the unreliability of the data in recent months, the revision provides a clearer picture of the labour market. The new figures showed a downward revision of a record 911,000, suggesting that job creation was much less robust in the year to March than previously reported. It should be remembered that government payroll data indicated that employers had added nearly 1.76 million jobs in total for the year, so we are talking about the new figures being cut in half to 847,000, averaging 70,600 jobs per month. By sector, the most affected were leisure and hospitality (176,000 feweer jobs or 1.1 p.p.), followed by retail (126,000 fewer or 0.8 p.p.) and wholesale (110,000 fewer or 1.8 p.p.).
This surprised Bloomberg’s consensus estimates, which pointed to a cut of 680,000, as well as some Fed members such as Waller, who pointed to a cut of 700,000, and Treasury Secretary Bessent (800,000). Although benchmark adjustments are part of a routine data update process, on this occasion they attracted more attention following the White House’s dismissal of the director of the Bureau of Labour Statistics. In addition, recently the variability of the figures has been significantly higher than normal. Historically, the error has been 0.2 percentage points, but this year’s error was 0.6 percentage points, following last year’s error of 0.5 percentage points.
It should be clarified that this is not only due to a low response rate, but also to the limitations of the Bureau of Labour Statistics’ own birth-death model. At times of change in the economic cycle, in the early stages of a recession, they tend to overestimate the jobs created by new businesses and underestimate the jobs lost due to the bankruptcy of small businesses. This occurs because the BLS can monitor information from large companies, but provides less visibility on the small business sector. Therefore, to obtain a more accurate, albeit less frequent, data source, the Quarterly Employment and Wage Survey is used, which is based on state unemployment insurance tax records and covers almost all jobs.
