Divergent surveys were released for both economic activity and labor markets. Is the manufacturing sector expanding or contracting? The answer depends on which PMI survey is accurate. The same is true with the labor market. The Household survey shows job loss while the Establishment survey finds continued growth.
BLS released the May Employment Situation summary (June 7) with the bureau repeatedly commenting data “changed little†in May. Headline employment measured by the Establishment survey increased by 272K net jobs. The U-3 Unemployment rate increased to 4.0% for the first time since January 2022. Establishment survey job gains were, as typical, concentrated in Health Care (+68K), Government (+43k) and Leisure & Hospitality (+42k). Prior months revisions were a -15k adjustment lower.
The Household (HH) survey from which the unemployment rate is calculated painted a different picture. The HH survey found a seasonally adjusted loss in jobs in May of -408k. The unemployed level increased 157k while -250k fell from the work force population. For the 1-year period ending May 2024, the HH survey calculated the economy added just 376k net new jobs compared to the Establishment survey increase of 2.756 million new jobs in the same period.
The difference between the 2 surveys seems material so I investigated how their respective findings compared over various periods of time. BLS provides historical monthly results for both surveys beginning in January 1948. The monthly data over that 76 year period has a statistical correlation coefficient of 0.93. While not exact, the 2 series are highly correlated over their total sets.
With this baseline I reviewed different time periods to determine consistency. I calculated statistical correlation of 0.96 for the period since 2000 and 0.98 since 2016. The correlation between the series seems to strengthen in recent years. When I graphed the data, a clear visual break could be observed over the past 2 ½ years. When I calculated the series correlation since January 2022, the statistical correlation coefficient fell to just 0.12. This represents a very weak correlation in the data that could be described as negligible or nearly random.
The nearby chart shows job growth in the post-lockdown period. The Establishment survey shows 16mm net new jobs with relatively steady monthly increases. The HH survey shows 11mm net new jobs in the period with the peak occurring in November 2023 and the May 2024 level basically flat from June 2023.
Switching to economic activity, the major final May PMI reports were released. The S&P Global PMI data (June 5) accelerated from April and included slight increases from their earlier flash May findings. The S&P Global surveys remain highly correlated when compared to the Institute for Supply Management (ISM) reports. However, in their respective Manufacturing series the ISM levels have mostly been lower since February 2023. ISM Manufacturing surveys have been consistent with contraction in all but 1 month beginning in November 2022. The S&P Global findings in the Manufacturing sector have consistently shown expansion in 2024.