I think your question is: what is the advantage of layering a glossy material on a matte material if you can just make one semi-glossy material?
The idea is that the actual car paint also has multiple layers. Some light reflects off the top shiny layer, and some off the rougher bottom layer.
That's why in real life you never see half-sharp reflections on cars, but also you can see some general diffuse shading (i.e. not all cars look like they were made of chrome). You never see a slightly blurred reflection on a car, like you would let's say on a ceramic mug. Instead, you always see a sharp reflection overlayed on a more matte object.
Check this example: http://pinehurstmobiledetailing.com/wp-content/uploads/2014/06/iStock_000003958257Large.jpg - look at the hood of the car. You can see a sharp reflection of the people standing there, but also there is a red streak of highlight.
Very similar approach is used in digital painting, where you first paint a matte object and then overlay reflections. There is a great tutorial here https://www.youtube.com/watch?v=w4uJpHaHZTc
And, as the other anwers says, using the "Mix" node assures that everything stays physically correct, i.e. you don't have an object giving off more energy than it gets. I also agree that people tend to overcomplicate shaders, but I guess this is because you just experiment and finish when the solution looks good, not necessarily when it is pretty in the node editor :)