In the body of the paper The river model of black holes, the authors do not use the term "drag" anywhere to describe the model. They describe it as a flow, but not a flow of a viscous material that would produce drag. The mathematics of drag do not correspond well to either the river model or to more standard GR in more general spacetimes.
One thing to keep in mind is that the river model is not generalizable. To my knowledge, it applies only for the Kerr spacetime and the Schwarzschild spacetime. The model used in general relativity is curvature, not flow. It only happens that in these specific spacetimes the curvature can be mapped to a flow concept. Even then, the flow concept for the Kerr spacetime is odd and is not really akin to a flowing liquid. So I think it is best to focus on curvature, rather than flow.
The formal explanation for the asymmetry that you mention is the curvature. The "flowing" spacetimes that you mention have curvature. As such, they have the geometric features that you mention, specifically that in the Kerr spacetime initially radial falling geodesics will curve into a tangential path, and any radial timelike path inside the horizon goes inward. The absence of curvature in flat spacetime explains the fact that an inertial object continues at a constant velocity.
The asymmetry is explained by the curvature and the curvature, in turn, is explained by the presence of a large amount of gravitating matter. Symmetries in the laws of physics, in general, need not be preserved in cases where the boundary conditions or sources violate that symmetry.