Supersymmetric Chern-Simons theories in $d=3$

I am reading up on Chern-Simons matter theories in $d=3$. Here is the quote (from http://thesis.library.caltech.edu/7111 page 15) that I am having trouble with:

One could also add a supersymmetric Chern-Simons term for the gauge multiplet. This restricts one to at most N = 3 supersymmetry. However, in the presence of a Chern-Simons term alone (i.e., no Yang-Mills term), the gauge field is nonpropagating, and with a clever choice of matter content and superpotential, one can obtain very large amounts of supersymmetry

Context: The discussion is about possible supersymmetric constructions using vector multiplets. There are two things I do not understand:

1. Is it true that in the presence of a CS term alone (No Yang-Mills, no matter), the supersymmetry is limited to ${\cal N}=3$. If so, why?

2. I understand that when we add matter in a clever way one can obtain extended supersymmetry (as in the ABJM paper http://arxiv.org/abs/0806.1218). What I don't understand is that what does have to do with the gauge field being non-propagating??

I guess that the answer to my second question lies in the answer to the first. Any help?

See that if we have $\mathcal{N}$ supersymmetry generators, then we have $2^\mathcal{N}$ other supersymmetric particles. Given $\mathcal{N}=3$, we see that we must have $8$ superparticles associated to the gauge multiplet in question. A supersymmetry generator is basically a $1$-form (the exterior derivative of a $0$-form, i.e., the "normal derivative" of a function) on a manifold, which "points away" from the manifold.
If the gauge field is not propagating with respect to the background space, then the SUSY can be extended - but only up to a certain limit - this limit depends on the number of particles in the gauge multiplet. In this case, the limit is $3$ because of the aspects of the Chern-Simons.