Bes, isn't it sufficient to just have equal PFs?
Not actually. Imagine three phases all with the same PF and equal currents. The formula would give In=0 and that would be fine for linear loads.
However, if you had third harmonic current in each of the three phase, this component is at three times the supply frequency. A complete cycle of this takes one third of the time for the fundamental which results in the third harmonic current from the three phases adding in the neutral rather than canceling.
I might have a picture which might make the explanation a bit clearer.
And it's the same story for for harmonics that are a multiple of three - sometimes referred to triple
n
In other words, you could have balanced loads in the phases and an overloaded neutral conductor. It's not such an uncommon phenomenon. And will probably get worse as more and more non-linear loads come into use.