When your models are broken what do you do?

Try to force the data to fit your model an do more of the same.

Throw out your model and act on the data you have.

Bloomberg article with out the paywall.

Millions of workers are still missing from the US labor force three years after Covid-19 surfaced, and economists are scratching their heads as to how big the gap actually is and where all these people went.

One estimate found at least 2.1 million who retired earlier than expected. Another calculated a shortfall of 2 million immigrants at the height of the pandemic. Other research pointed to a million or more out of work because of long Covid.

There’s not even an agreement on the overall size of the hole — how many more Americans would be working in 2023 had it not been for the pandemic. That’s a problem because officials at the Federal Reserve need to know if Americans are temporarily or permanently out of the labor force so they can set monetary policy, said Anna Wong, chief US economist at Bloomberg Economics.


How about option #3 - Tweak the model to take into account “adjustments” as needed to do what you can to model as accurately as possible?



It is interesting that the industrialists do not want to publish that higher GDP growth means shortages of labor.


Or option #4 - Throw out your modelers and tweak existing or build new models.

Figures don’t lie, but liars figure.

Meanwhile economists debate which number is correct.

That is all they do. That is their job. Is that productive work? Or a tax on society?


That is not true if you begin to read some of the reports at the CBO or FED. You will get insightful research.


Oops!! There must be a nerve there somewhere. Doesn’t sound like a bot.


Till my unending days, I swear I am a bot.

I am reasonable.

1 Like