A communications satellite would have to be in a halo orbit around the Earth-Moon L2 point to have continuous, simultaneous visibility of both the Moon's far side and Earth. The first paper describing this kind of orbit was written in 1968, too late to be used for Apollo. It also described the options for communications satellites in medium lunar orbit.
This was followed up by
SUMMARY of the 1968 paper:
This report investigates the feasibility of communicating with Apollo spacecraft behind the moon and of communications between two such spacecraft. Two satellite geometries are considered for signal relay: a libration or "Hummingbird" satellite anchored 65,000 km behind the moon and a lunar-orbiting relay satellite at 1000-km altitude. The signal transmission frequencies of the Apollo spacecraft are VHF and unified S-band (USB) operating at 300 MHz and 2 GHz, respectively. The three propagation links examined in this report include: lunar far side, i.e., Lunar Module (LM) on lunar surface and/or lunar-orbiting command and service module (CSM), to earth; lunar far side LM to CSM; lunar far side surface to surface.
It is shown that as a result of antenna pointing constraints, only the libration or "Hummingbird" satellite is suitable for relaying USB tracking and communication data to earth. It is also shown that acquisition of a 1000-km altitude lunar satellite by an Apollo spacecraft dictates the use of VHF. Various modes-feasible and unfeasible-of lunar far- side radio relays are listed below, with their respective methods of communication.
(emphasis mine)


So constellations in low orbit were looked at and discarded. Likewise an L2 satellite was not proceeded with.
This tells me that a risk/benefit analysis was made, and the benefit did not justify the effort in designing and building a communications satellite and launching it into an orbit that had never been used before.
A quick search of the NTRS doesn't turn up more information on this though.