You nailed it, Curtis. Our Corp leader established this ridicuously old school strategy of replacing desktops every 4 years and laptops every 3 years based simply on (industry produced) probability curves of when the mean failure rate begins to increase. His logic was that you minimize cost thru reduced downtime for end user by having sudden failures, then having to respond.
But his assumptions were faulty:
1) He assumed you would lease, so you just give back old and get new one. Reality is leasing was too expensive and unnecessary for many of smaller biz units.
2) So without leasing, and implementing this 3/4 year strategy, you owned many spare devices which had nothing wrong with them. So you had plenty of spares, with the enterprise software already installed, ready to deploy in a New York minute if unexpected failure.
3) Since he didn't envision spares existing, he did his downtime cost calculation based on the 1-2 week replacement time for new hardware. No one is leaving a key user without a computer for a week. You'd take one from conf room before you would do that. If user can last 1-2 weeks without computer, I'd question whether you really need that user.
So for me personally, everything you say above is exactly what we consider. It summarizes to: If it ain't broke, don't fix it.
Obviously the scale of your company comes into play, you don't want 25 spare computers. But 5 computers can be spares for 100 users, the probabilities of having that many simultaneous failures is miniscule.