molecular orbital theory - Why does the PH3 geometry deviate more from the trigonal planar one, than does NH3?

$\ce{PH3}$ has a more bent structure than $\ce{NH3}$. The HOMO-LUMO gap for $\ce{PH3}$ is smaller than for $\ce{NH3}$, and so the distortion from the trigonal planar geometry is said to be larger. However, why is the HOMO-LUMO gap smaller for $\ce{PH3}$?

It should be possible, I suppose, to argue for why this is so by using qualitative molecular orbital theory, and from looking at the construction of the molecular orbital diagrams, but I am not sure how to construct accurate MO diagrams for $\ce{PH3}$ and $\ce{NH3}$.

In the answer to the possible duplicate question, it seems to be taken for granted that some distortion from the planar structure is stabilizing. In my question, the starting point is the planar structure, and then looking at why $\ce{PH3}$ is in a more distorted geometry than $\ce{NH3}$, relative to the planar structures.

This distortion is often explained in terms of the "HOMO-LUMO gap", but I fail to see how the HOMO-LUMO gap can be assessed qualitatively in order to predict which of $\ce{NH3}$ and $\ce{PH3}$ is in a more distorted geometry. So what factor ultimately leads to the smaller HOMO-LUMO gap in $\ce{PH3}$? $\ce{BH3}$ exists in a trigonal planar geometry - why is not this molecule pyramidalized?