Quote by DrGreg
Attached to the bottom of this post is a diagram to help explain things. As was mentioned earlier in this thread, one way to approach the problem is to consider it a variant of the ladder paradox, and consider the different definitions of simultaneity.
But my approach here considers length contraction only. And I am going to consider a complete circuit: not just a single wire with a lefttoright electron flow, but also a return wire with a righttoleft flow. Apart from the ends of the wires, we keep the two wires far apart so they have negligible influence on each other. The diagram is a highly idealised simplification, considering just 16 electrons in the circuit. The ends of the wires should be in contact with each other but I've drawn them as separated to keep the diagram simple.
The top left part of the diagram shows the wires with no current flowing, in the restframe of the wires. 16 electrons equally spread out along the wire.
The top right part of the diagram again shows the wires with no current flowing, but now in a frame moving at the velocity that electrons would flow in the bottom wire if the current were on. We see length contraction as indicated by the yellow arrows. I'm assuming a Lorentz factor γ=2. So far so good.
The two bottom diagrams now show what happens when the current is flowing.
In the bottom left diagram, as we are told the wires remain electrically neutral, there must still be 16 electrons in the wires. There's no reason for the electrons to bunch together anywhere, they will remain spread out around the whole circuit as shown.
Finally, let's look at the bottom right diagram, which I think some people are having difficulty to imagine. We already know what happens to the red positive ions, their separation contracts just as before. The electrons in the lower wire are now stationary, so their separation must be larger than the bottom left diagram as shown. On the other hand, the electrons in the upper wire are moving faster than in bottom left diagram, so their separation must be less than in bottom left diagram. No electrons have escaped so the total number of electrons in circuit must still be 16. But now there are fewer electrons in the lower wire and more in the upper wire. So the lower wire has a positive charge and the upper wire has a negative charge.

I think that using the return wire to prove a point is cheating. It does nothing for the original scenario without a return wire.
The ladder paradox also has some asymmetries that seem to be missing in your example:
Figure 4: Scenario in the garage frame: a length contracted ladder entering and exiting the garage
Figure 5: Scenario in the ladder frame: a length contracted garage passing over the ladder
The two frames do not see the same number of rungs inside the garage in each case.
If we assumed that the protons were represented as tiles on the garage floor, the garage as the wire, and the ladder as the electron current in and out of the wire, then clearly the charge inside the boundary of the garage is
not invariant.
However, considering that the electric field intensity increases by the same amount that the boundary of the garage in the LT frame is length contracted, this would keep the electric flux around that boundary of the garage a constant.