PBR is a great example of how limitations go a long way to improve something. In forums for games where there are 'sandbox' elements, a lot of people start to whine whenever limitations or restrictions are put on things. They see them as "artificial limitations on their fun" in doing whatever they want to do.
But putting artificial limitations on graphics is exactly what PBR is doing. If a material has a very high reflectivity to it, like a mirror, then the shader makes it impossible for you to give it high levels of sub-surface scattering as well. Likewise, the shader for a rough surface will not allow you to give it high values of reflectivity at perpendicular viewing angles. Single sliders affect multiple attributes of the material where in the past you had to adjust multiple sliders to try and achieve the same effect. The difference is that PBR doesn't let you stray from reality, while before PBR it was hard just to GET TO reality.
These limits aren't artificial in real life. It's the physics of light. But they're artificial in a computer environment because you can write shaders any way that you want to. There's nothing stopping you, computationally-wise, from creating a material that has the translucence and refractiveness of glass, but doesn't reflect at all at glancing angles (the fresnel effect). It MAY not look very good though BECAUSE it doesn't mimic how we see things on a daily basis in the real world.
Implementing PBR, or a similar technique, has at least 2 benefits I can think of for both SC and ED. It will look better because it will closely mimic the way light works in real life, and both games are striving for photo-real high fidelity.
It will speed up production in the long run once there are no more conversions to do. Less choices for the graphics artists means that they work faster. Leather is leather. Aluminum is aluminum. They don't have to spend nearly as much time tweaking the values to get the materials to look right.