I don't have a citation either, but you can empirically determine this to be likely by just comparing the reward per kW spent and noticing that it has become less than the cost per kW of non-renewable energy.