I think the prevailing (correct) consensus is that developer productivity is actually very hard to measure, and every time it is attempted the measure is immediately made a target making the whole thing pointless even if it had been a solid measurement- which it wasn't.
IDK where you're getting the idea here that measuring productivity of anyone who isn't a factory worker is easy.