On the conundrum of deriving exact solutions from approximate master equations