This is an aggressive misinformation!
An object suspended in radiative field always tries to get in some kind of equilibrium with that field. Such equilibrium has a thermal nature for wavelengths smaller than the object's size, whilst at greater wavelengths the object becomes more point-like and its motion becomes macroscopical and thermal equilibrium is lost starting from specific scale. The overall picture changes if the the object is a point-like charged free particle; absorption and re-emission becomes impossible and interaction turns elastic and generally aperiodic at broad frequency band. Those aperiodic accelerations deliver a secondary spectrum, and if the system of particles has enough time to interact with radiative field in an effectively closed system, the result contains an equilibrium spectrum, at least as a part of total spectrum.
If you want to have a relic spectrum, you have to ensure zero cross-section for elastic interactions with the free charged particles in your system. That's quite impossible at given IGM plasma density and CMB wavelengths.
Your " here " has got nothing to do with what I'm talking about. The author has got no understanding of possible models for global equilibrium; e.g. it doesn't even cross his mind that global equilibrium may be effectively constructed of a closed chain of local non-equilibria... That's plain stoopid!