So the issue was harmonics, even though you didn't originally state it.
Sorry if I mislead anyone.
Actually, it may or may not be the results of added harmonic Currents.
It is more towards how Linear the Load is - and by "Linear" I am referring to the transferring of True Power, rather than the wave characteristics.
Like an Induction Motor driving an Axial Fan - that would be Linear; whereas an Induction Motor driving a Piston-type Air Compressor would not be very Linear (too many changing amplitudes across the peaks, and valleys which change randomly).
Discharge Lighting is not very Linear, as there will be a considerable level of Capacitive Reactance at the Lamp, and the Lamp's Load will vary greatly with changes in System Voltage + Lamp Temperature & age.
Computers will only be steady when they are at a "stable idle" - meaning there is no Disk Cache work going on, the NIC is not Polling the LAN, and all the "TSRs" are idling.
Otherwise the SMPS (Switch-Mode Power Supply) will be drawing transient bursts of input Volt-Amps, which vary in intensity - making the Load very "Non-Linear"
All the above examples do not take into consideration any Harmonic Currents reflected back against the Supply (Transformer's Secondary) from the Loads which created the Distortion.
If those Currents are large and Zero Sequencing, the extra Load Currents reflected back will become additive in the Circuit Conductors.
On the subject of Harmonic Reflections (Harmonic Currents), I have only seen excessive Currents with very old Technologies.
Back when the first Prototypical "Electronic Ballasts" came into the Market (circa 1988 - 1990), the Hybrid Ballastry first introduced was very noisy.
The total Reactance between the reduced size Reactor Coil and the "Pseudo Inductor" IC + Control Circuitry, resulted in a Tank Circuit which reflected a considerably high level of 3rd, 9th and 27th level Currents back against the Transformer's Secondary Windings.
This only happened in a few rare cases, and the issue was resolved by 1991.
Other than this, the "Harmonics Scare" per "Electronic Ballasts" is more propaganda than fact.
Sure, there will be some levels of reflected Currents, and there certainly is Harmonic Distortion, but there always has been Harmonic Distortion + reflected Currents with Reactive Loads, so it's nothing new!
Motors, Standard "Magnetic" Ballasts, and similar Loads will create Harmonic Distortion, and reflect Currents back against the Supply (Transformer, in most cases).
So, to wrap this up, Loads with Reactance tend to cause the Common Grounded Conductor of a 4 Wire Wye System to carry >=70% of the highest L-N Load Current.
Harmonic Distortion may result in additional Current flowing - however, the overall sum is what Currents are Zero Sequenced across whatever Harmonic Frequencies (orders of the fundamental HZ - such as 3rd, 5th, 7th, 9th order).
If the Zero Sequenced Harmonics tend to resemble something Linear, these Currents may balance across the Ungrounded Conductors more than accumulate in the Common Grounded Conductor.
All this stuff is null & void, when applied to a 1 Phase 3 Wire 120/240V System - either from a single center tapped Transformer, or the "Base" Center Tapped Transformer of a 4 Wire Delta System.
While there will always be Harmonic Distortion produced by the Reactive Loads, the issues of overloading a Common Grounded Conductor does not occur in the 1Ø 3W 120/240V System.