A couple of clarifications;
Transformer impedances are often given in PU values. When transformers of equal percent impedance voltages are connected in parallel, the load will divide in proportion to the KVA ratings of the transformers. This relationship will transpose to agree with ScottyUK's post.
However, this assumes that the transformers have equal X/R ratios.
Transformer regulation and transformer impedance are not different expressions of the same effect.
Regulation relates to the voltage drop at full load with a resistive load. The resistance of the load predominates over the transformer impedance and the the voltage drop is mostly dependant on the transformer resistance.
Under short circuit conditions, which are the conditions under which transformer impedance is measured, transformer reactance has more influence and the current is dependant on the impedance.
for transformers with different X/R ratios, the reactive current will divide in inverse proportion to the transformer reactance and the real current will divide in inverse proportion to the transformer resistances.
When paralleling a large transformer with a fairly high percent impedance with a small transformer with a fairly low percent impedance, it may be found that the safe load on both transformers is less than the safe load on the large transformer by itself. This is assuming equal X/R ratios. Consider different X/R ratios for the two transformers and things can get very interesting.
respectfully