I think the “official” mounts are tilted back more.
There is going to be “some” drag.
Here’s an engineering paper. The squarer the flag, the more drag it has.
http://arc.aiaa.org/doi/abs/10.2514/1.9754?journalCode=ja
I haven't looked at the PDF, but the abstract says that they were using laminar flow in a wind tunnel for all their testing. This is not what you will find behind the back of the vehicle with a rider on it. This problem would be fun to model one day in a CFD package - if the solution ever converges :-)