It's important to note that it's the subject distance that's the primary factor, not the focal length. The focal length is secondary in that it dictates how far you would be to achieve the given framing. If you shoot the picture at 200 mm from the example and then without moving you shoot again at 20, you'll have the same perspective, just way smaller subject in the frame; if you then crop in the picture shot at 20, you'll have the same framing too, just way less pixels.

If you're half a metre away from the dude's nose, you'll be roughly 60 cm away from his ears (20% more distance), but if you're 5 metres away from his nose, you'll be 5.10 m away from his ears (only 2% more distance) - and this is what creates the difference in apparent sizes of the facial features relative to one another.