The estimated number of job vacancies in the United States in September rose from 7.19 million revised from August to 7.523 million, and the number of hiring positions increased from 5.271 million to 5.402 million. However, when the U.S. Bureau of Labor Statistics released the data on November 3rd, it used the same assessment for both national figures: job vacancies at 7.5 million were "not significantly changed," and hiring at 5.4 million was also "not significantly changed." Although there are monthly differences in the tables, the statistical methodology does not support the conclusion that there has been a clear rebound in labor demand.
The point estimate for the job vacancy rate has changed from 4.3% to 4.5%, and the recruitment rate has changed from 3.3% to 3.4%. The job vacancy statistics include positions that are still recruiting at the end of the month, while the recruitment statistics include those who actually ended up on the payroll throughout the month. These two indicators are measured over different time periods. Based solely on the difference in point estimates of 333,000 and 131,000, it cannot be proven that companies 'want to recruit but dare not', nor can it be concluded that there has been a change in the matching of positions.
BLS is more restrained in its summary of industry changes. Regarding job vacancies, officials clearly indicated that there was an increase of 104,000 in state and local governments (excluding education); in terms of recruitment, there was a 158,000 increase in the hospitality and catering services sector. The complete table also lists estimates for other industries, but items that are not identified in the summary as significant changes should not be interpreted individually as signs that those industries have become hot or cold.
The point estimate has changed, but the statistical conclusion may still be that 'there is not much change'.
JOLTS comes from a sample survey, not a detailed count of all job positions in the United States. The sample estimates may be affected by sampling errors, unresponded surveys, seasonal adjustments, and subsequent supplementary reports. BLS will determine whether the monthly difference can be considered an increase or decrease based on the magnitude of the change and the standard error; therefore, "7.523 million is higher than 7.19 million" does not contradict "there is not much statistical change."
The data for August was also revised in this release: the number of job vacancies increased by 111,000 to around 7.2 million, and recruitment increased by 79,000 to around 5.3 million. The starting point for monthly comparisons will be recalculated with the new reports and seasonal factors taken into account. This further illustrates that a single-point estimate is not suitable for drawing strong conclusions about trends.
For the market, this data is more consistent with the view that "there has been no confirmed monthly turnaround in national employment demand." It neither proves a reacceleration in hiring nor indicates a significant collapse in job vacancies. To determine whether the labor market is tightening, it is also necessary to consider non-farm employment, wages, unemployment rates, as well as the continuous changes in subsequent months after JOLTS.
The confirmed changes are concentrated in a few sub-items.
The number of employees leaving their jobs in September increased to about 3.4 million, which BLS considers to be an upward trend; there were approximately 1.6 million layoffs and dismissals, with little change. The number of employees leaving the accommodation and catering services sector increased by 197,000, while recruitment also increased by 158,000, indicating that the labor turnover in this industry was more active that month. This specific statistic is more compelling than the inference of "increased matching friction" derived from the difference between national job vacancies and recruitment figures.
There was an increase of 104,000 job vacancies in state and local governments (excluding education), which is also a change mentioned in the official summary. Public sector positions may be affected by budgetary constraints, approval processes, and the supply of skilled professionals, but this time JOLTS did not provide reasons for these changes. Any explanations regarding approval cycles or skill shortages need to be supported by data on recruitment timelines, job types, or regions; they cannot be directly inferred from the increase in just one industry.
It is indeed possible that companies retaining their recruitment plans, extending interview processes, or waiting for budget approval could lead to a mismatch between job vacancies and actual hiring; similarly, mismatches in job skills can also result in similar situations. However, these can only serve as hypotheses for further verification. Only if there is a continuous increase in job vacancies over the coming months, without corresponding recruitment efforts, and if the time taken to fill these positions becomes longer, will such explanations become more convincing.
For the Federal Reserve, September JOLTS will not determine the interest rate path on its own. What is more worth monitoring is whether the trend of voluntary resignations will continue, and whether wage and service price inflation will strengthen in tandem. The resignation rate is often used to observe workers' willingness or ability to change jobs, but the level for a single month also needs to be considered within a longer time series.
The reliable conclusions drawn from this report are limited, which makes it clearer that there were no significant changes in national job vacancies and recruitment statistics in September. However, some identifiable changes were observed in specific industries and in the subcategories of employee departures. 7.523 million remains a useful estimate, but it is not evidence of "companies significantly accelerating their hiring efforts." Only if the next set of data continues in the same direction will there be a chance to transform this trend assessment from a possibility into a more solid conclusion.
