It is easier to understand with a bit modified logical thinking.
For a electrical short monopole antenna (<0.25 lambda), is it the opposite, antenna size grows if components are added.
A capacitive top-load
increase a monopole antenna length, both electrical and physical.
Measuring a short monopole antenna impedance without any top-cap shows that antenna load at bottom feeding point is somewhat capacitive.
To compensate for that reactive load can a serial inductor be inserted at bottom of antenna.
If compensating instead is done 0.25 lambda away from this inductor location will a inductor become transformed to capacitor.
Top of the somewhat short monopole antenna is almost at that distance from bottom coil location, 0.25 lambda => at this location should a capacitor be placed to act as a bottom coil.
For same reason do a top capacitor not work very well as tuning element for a real short monopole antenna as the short antenna length is too short to transform top capacitor load to a inductive bottom load.
Monopole impedance, borrowed from
https://www.aa5tb.com/efha.html