I agree with NEMA's response that old lamps will typically not be re-used when a ballast or fixture is replaced, but nobody ever questions the original premise. They are reaching their conclusion by calculating lamp life based on hours while using a fixed 15 year life for the ballast/fixture. In my opinion, both the ballast and lamp life should be based on hours of operation. If you figure it that way, then the rest is irrelevant. In the residential sector, ballasts and fixtures typically outlast lamps. Of course, ballasts sometimes do fail first, but that's true in commercial use as well.
Here I do not agree. The problem is, all components involved, so fixture, ballast, lamp age due to different reasons and by different mechanisms.
Some are related to juts how long the light operates, some to just plain number of starts, but there are many of those, which are related to just the time, regardless if the fixture is used or not. So after 15 years, the fixture will be full of greasy dust and with many corroded metal parts and similar things making the fixture not operating properly. The fact it light up the lamp is by far not sufficient to call the fixture working. Ff the light e.g. can not pass through dirty refractor or get absorbed on a corroded reflector or when the clogged vents make the thing overheating, or when the fixture breaks apart after just attempting to open it for cleaning and/or inspection, I can not call that differently as just not working, period.).
In a commercial environment the operation is so intense, just the burning hours make the things to fail in the matter of few years, so rather early. SO during that time the other mechanisms do not have chance to progress so much. Moreover the heavy use effectively suppress many of the aging mechanisms (the heat from the system keep it dry, so the metal corrosion can not progress, the need to service it after few years usually mean it gets cleaned as well,...)
But with home installation, where the light is used for just very brief periods, long time passes and the fixture has just few hours on it. Then the 15 years pass without the lamps even making any significant burning time on it, but there is dirt accumulated over the 15 years, hinges and bolts corroded and stuck, so an attempt to even open it frequently ends up breaking it, clogged lamp holders cause the lamp gets damaged with an attempt to remove it (hence new lamp with new ballast and sockets) and so on.
So really just the rating of burning hours is by far insufficient information to even compare two products, mainly when the technology differs a lot, so they wear mechanisms contribute differently. So it is very common one type of light outlast (and I'm speaking about factor of 5, not 10%) the second one in one installation, but it get completely opposite in another installation.
The thing is, maintenance person, has the main experience with cases, where the things do fail (because he is called in to fix it), but has nearly no experience with things working without any failure till they are torn down because of remodeling or so (why would anyone call the service on thing, which works well), so of course, he simply can not have any overview of that.