I'm curious about your napkin math, care to share?

I got mine from a paper published by the University of Chicago: https://www.journals.uchicago.edu/doi/10.1086/735565.

How do you even estimate that, human productivity is non linear to worktime and often, employees, when given good benefits are likely to be more productive.

also there's the fact of this having a ROI since people generate economic value in the longer run, and hence, more taxes