The ‘jobs’ number is always wrong

Key Points

1. The BLS jobs reports are based on incomplete data, typically using only about 60% of surveyed employer responses, leading to inherent inaccuracies.

2. Historical examples show political interference with BLS data, such as the Nixon and Kennedy administrations manipulating unemployment figures for political purposes.

3. The BLS adjusts data for seasonality and other factors, which can significantly alter the reported numbers, as seen in the July report where unadjusted data showed a loss of 1,066,000 jobs versus the adjusted 73,000 gain.

4. Statistician George Box's quote 'All models are wrong, but some are useful' is referenced to emphasize that while BLS data is imperfect, it still provides valuable directional insights.

5. The article notes that high-frequency data like non-farm payrolls prioritizes timeliness over absolute accuracy, as 'correct is the enemy of good' in such contexts.