I'm a bit unclear about what you're asking. When you say, "no matter which method any of us uses for recording the resulting wav is usually 16/44," you are talking about the final mixed-down stereo track that you would play on an audio CD player, right?
If yes, then yeah, your final mix will be 16/44 in stereo, so the size will be identical no matter what went before in the song's history up to this point. But the qualitywill probably differ.
To your question is "wouldn't it then make sense to record a track @96K & convert it to 44 ASAP, if not sooner?", I would answer no - probably not. If that were true then why record at 24/96 at all? The thing with 24 bit files is that there is much more detail and the results of processing -- all the plug-ins for reverb, compression, as well as fades and mixing tracks together and whatever -- can work more accurately with better resolution of the source data. The round-off errors of calculations are smaller.
Think about somebody at the bank calculating your interest on your savings account (as if any musician or home recording enthusiast has any savings -- what a bad analogy I picked here!). You wouldn't want them to immediately start rounding off all the extra decimal points from the first calculation before they go on to the next one. Those round-off errors accumulate as you perform more calculations.