BLS reporting has been in the news lately. I only took one statistics class in college, so I recognize my limitations. Nonetheless, I have some “meta” questions:

Are “good” statistical reports ones that clearly “show my team’s policy to be right”/”show the other team’s policy to be wrong?” Or are “good” reports those that most accurately DESCRIBE economic reality? Given the way humans work, there is ample reason to assume these two approaches to defining a “good” report are not necessarily the same thing.

A follow up question, following closely on the first. Do we WANT “good” reports? If I were in charge of economic policy and could demonstrate some cause and effect relationships between a statistical report and the policies my team was enacting, I would seek to learn from that report so I could adjust our policies appropriately.

Knowing that every policy has near, medium, and long term effects, AND knowing that every report of this large complex system we call “The Economy” (or even the subset called “The Labor Market”) is only an approximation, I know there are always limits on my knowledge.

If I operated on the alternative understanding of what makes a report “good” and cared (publicly at least) only that it made my team look good or the other team look bad, then I if I were an objectivist in the sense of believing that the world is the way it is whether I like it or not, I would have to live with the fact that reality may come back and bite me someday. If I were a subjectivist, i.e., if I believed my beliefs about the world MADE the world the way I believed it to be, then, Hey – I could believe whatever I wanted with no consequences.

Scott Lincicome, an economist who knows way more than I do on the subject, has SOME THOUGHTS. (The image above is from Lincicome’s piece)