It's an understandable development.
If you have X amount of processing power available and you "throttle" the FPS to a lower value than the maximum your system can do, you have more reserve power for when extra things start happening on the screen.
On the other hand, if the system is already running full blast to maintain as high FPS as possible, then whenever something new occurs on screen there is no reserve processing power to deal with it. So, even though the system should be able to display fluid frame rates, it's already chocked up and out of power at that point.
The difference between the two methods is exactly what you describe. In the first case you may get about 40 FPS but they will be steady, while in the second case you may get 60 FPS but whenever something new happens they will briefly drop to the low 30s before stabilizing around 40-50 FPS again.
The numbers i used are arbitrary, but that's more or less how it works and it's exactly why civilian sim pilots (FSX, X-plane, etc) who run demanding scenery add-ons use FPS limiters. Absolute FPS values mean nothing without the context: how are they achieved, plus how smooth and stable they remain
