Correction: after reading the other answers, I realise I was completely wrong and that the main complication would be that the calculation of the perturbation (matrix elements) would be extremely hard to calculate without splitting everythin up in the multipole expansion first.
