There are a number of people I see consistently in the commit logs at work and I am quite confident that they have created more than a handful of new jobs each. So while they didn’t directly remove revenue, these people have an astonishingly bad ROI.

This is true. The performing engineers are underpaid; the underperforming ones should be trained or fired.

How about the good engineer, employed to design the wrong thing?

I feel like I've put good quality engineering work, in to projects that were completely wrong-headed in the bigger picture, and thereby lost the company money.

This seems more of an issue with where the company invests resources than to the engineer themselves.

I agree; my point was just that the "goodness" of an engineer doesn't necessarily relate to their impact on revenue.

In the particular situation I was thinking about above, it actually could've been better from a business perspective had I not been as "good" of an engineer. I took over a project that was simply not working at a technical level - it was a DSP-based thing, and the chosen DSP appeared not to have enough oomph. I managed to optimise the electronics and firmware enough to make it go, so it went on to production and AFAIK (I'm no longer at that company) has been a money pit ever since. The other possibility was a redesign with a burlier DSP, but I'd bet that the project would've been cancelled instead.