No. Collapse transforms is a totally different thing, and shouldn’t be thought of as an effect.
When they fall down, it’s probably just a keyframed animation changing the y position of the smiley face. If you find yourself keyframing every face’s move, you’re going about this the less-efficient way. Each face probably is the exact same comp* popping on screen, staying there for 3 seconds, then sliding down and off. If they pop on at different times, they’ll naturally also do their downward movement at different times as well. You don’t need to animate each one.
* Make one face animating the way you want it, then just duplicate it and replace the expression.
(You may not have a bunch of add-ons at your disposal, but just so you see other workflows: if I were creating this, I would also try using a script or particle generator to randomize their location on screen so I wouldn’t have to manually place them all.)