With how low level a library MonoGame is, it's really vague if you talk about lighting. There's million ways to implement it. But the general concept to have per instance lighting is really straightforward if you understand how the underlying system works.
You have a vertex buffer with model data and one with instance data. Whatever is specific to an instance goes into the vertex buffer with instance data. What that data is differs depending on how your lighting/models work. I explained a bit how the binding works in this thread: http://community.monogame.net/t/how-to-use-multiple-vertex-buffers/10230/2?u=jjagg
There's also an example linked in reply.
So basically if you have effect parameters that you want to change for each instance, consider passing them as vertex data in the instance buffer instead.