I think it was said before, but i too think that the important destinction here is that the total charge changed from toy model 0 to A. If in stead we took away some negative charge out of the two “electrons” and then put these charges in the middle the first term would change in the multipole expansion aswell. So the way i would work with these toy models would be by changing the distribution of the charges not by putting in new charges.