This seems like the worst possible way of answering the question. I'd be much more interested in looking at this company by company within a vertical. Knowing

This seems like the worst possible way of answering the question. I'd be much more interested in looking at this company by company within a vertical. Knowing that the teleworkability of software engineering is 80% and that software engineering as an industry adds a lot of value per hour from a productivity perspective doesn't seem causal to me. You could for instance just be measuring that physical presence is more correlated with low value work rather than measuring any notion of productivity. Knowing whether the companies within software that allow remote work add more or less value than the companies within software that don't would be far more informative.