That's a bit like not writing any new code unless you fix every single bug in the existing code. The counter-argument being that newer code will improve the des

That's a bit like not writing any new code unless you fix every single bug in the existing code. The counter-argument being that newer code will improve the design or fix some architectural issues, or handle upcoming feature requests better, etc and indirectly improve upon the flaws in the existing code. Its not always zero-sum. So, why not do both, and prioritize the money accordingly?