Why An electrically short vertical has a current maximum at the base that tapers rapidly to zero at the tip, so very little of the structure carries useful current. Adding a capacitance hat or top loading coil at the top forces the current distribution to stay high over more of the radiator, which raises the radiation resistance relative to the fixed ground and conductor loss resistance. More of the power fed to the antenna therefore leaves as radiation instead of heat, which is the definition of improved radiation efficiency. Top loading also puts any loading inductance where the current is smallest, minimizing coil I-squared-R loss compared with base loading.
Watch out Lower Q is tempting because top loading does broaden the usable bandwidth a bit, but the Q of a short vertical stays high regardless, and that is a side effect rather than the principal advantage; higher losses is exactly the opposite of what top loading achieves.
Top loading keeps current high up the whole radiator: more current radiating, less heating the ground.