I submit that electricity adheres approximately as closely to the colloquial "water in pipes" model as actual water in pipes does.

Which is to say not very well beyond a certain level of detail. But in neither case is there a more useful first order approximation to use.