The Standard model of cosmology predicts zero circular polarization, but what happens if we consider propagation effects that convert some linear polarization into circular polarization?
Considering propagation effects in regions that have anisotropic index of refraction will lead to circular polarization because photons going in one direction will take longer than other photons, therefore there will be a phase shift between both photons.
Several mechanisms can produce this phase shift, however the strength of the effect is small. For more information please visit my research section, or alternatively, my publication at https://arxiv.org/abs/1803.04505