The unusual smoothness of the extragalactic unresolved radio background

If the radio background is coming from cosmological sources, there should be some amount of clustering due to the large scale structure in the universe. Simple models for the expected clustering combined with the recent measurement by ARCADE-2 of the mean extragalactic temperature lead to predicted clustering levels that are substantially above upper limits from searches for anisotropy on arcminute scales using ATCA and the VLA. The rms temperature variations in the cosmic radio background appear to be more than a factor of 10 smaller (in temperature) than the fluctuations in the cosmic infrared background. It is therefore extremely unlikely that this background comes from galaxies, galaxy clusters, or any sources that trace dark matter halos at z<5, unless typical sources are smooth on arcminute scales, requiring typical sizes of several Mpc.