You can make the hole bigger than the magnet but you are then wasting space that could be used to wind with thicker wire to keep resistance down.
It is all a compromise. It is further complicated with axials where you don't have a fixed gap between magnets. Unless you use wedge shaped magnets you have more gap on the outside. You tend to find that you have wasted space on the outside and it is all jammed up at the centre. Many compromise by making the holes a bit small at the centre where you loose little flux. Others go with radial coil sides where the hole is big at the outside. I suspect that it adds more resistance with little return. I strongly suspect that keeping the coil the same shape as the magnet is a good compromise.
Why do people worry so, this is something of a second order effect. Dan has made many good machines with holes a bit small and I am sure many have used radial coil sides with rectangular or circular magnets.
One day someone will write the computer program to sort out these things and all machines will be perfect and still someone will want to do it another way.
The single phase case was sorted out by Ferranti, Mordey and others about 120 years ago, it's a pity they didn't have a go at doing it 3 phase, I would love to know whether they went with overlapped coils or ended up with the single layer winding. They didn't work at variable speed and rectify to charge batteries so the conclusions may have been different anyway.
Flux