_fred_
More on BLS commissioner issue and data revisions: I've done a lot of pulling of old estimated and actual employment projections. It seems like it's fairly normal to make small revisions, but 2024 and 2025 errors have been extraordinary• Revisions are normal they improve data accuracy over time. • Large revisions (especially downward) are not normal, and when they exceed ~500,000 jobs, they may signal systemic overestimation, economic cooling, or data/modeling issues. • 2024’s and 2025’s revisions are not typical, and they have raised questions about how strong the labor market really was post-COVID and during the inflation/interest rate adjustment period.
As always love your content. I think something you are missing is that the need for revisions is because the initial data points are ESTIMATES. For someone like you probably obvious but not to the average chap. The data gets revised as more data comes in that reflect the actual job market. My opinion is that these initial estimates are political. I think the FED and National banking systems are more political than they would like to admit …
That's absolutely right. Revisions are normal they improve accuracy over time. The reason revisions are normal is just what you say; the BLS publishes estimates. Predicting the future is very hard. However, the most recent 2024 and 2025 estimates have had much more error than has been historically seen in BLS estimates.
If I had to guess there has been such rapid changes in the job market threefold with AI, tariffs, and deportation efforts. It is my knowledge that the BLS uses a “benchmark” in their models that only gets revised once a year. Maybe the change needed is not the model but the frequency the benchmark is changed due to macro volatility. The macro environment and job market fluctuations is almost certainly causing this abnormal deviation. What else could be the solution to more accurate estimates?