Is using multiple cameras, culling masks, and layers expensive on mobile?

I have an endless runner game.

I have my background layer which has a bunch of scrolling art elements. This is culled with a background only camera on its own layer. It has a depth of -1.

I have a foreground layer which has a bunch of scrolling art elements. This is culled with a foreground only camera and its own layer as well. It has a depth of 1.

Then, my main camera is where the gameplay layer is, and it renders all that stuff together. It is set to a depth of 0.

This way, I can have an elegant little background art scrolling system on one layer, an elegant little foreground art layer on another, and worry about the complexities of the game-camera and gameplay on its own layer.

This method seems like an excellent workflow, but I am wondering how much of a performance impact this is on mobile? Should I just find a way to do all of those things using one camera, or is separating it like so the way to go?

Everything comes with their own cost. You have three cameras rendering the same scene. It depends on the complexities of rendering. If you have high quality images all over the place and the scenes are long then yeah the phone can get laggy. You can try optimizing the quality of your gameObjects if you really need to have three cameras. What people usually try to do (at least i do sometimes) is have two cameras, one for the foreground and other for the background (for parallax effect).
What i would suggest is if you can try reducing the number of cameras thats good, if you simply cannot then try optimizing your images and textures.