The best answer I can come up with is probably not, i.e. the ratings are probably well-established..
Just looking at the Wikipedia source of most listened to radio programs, some of the entries in their list of rankings are self-reported and others are sourced from talkers.com, which is a bit nebulous on their methodology:
This list represents Talkers Magazine's monthly in-house research for December 2017 tabulating cumulative weekly estimates of total unique listeners for the top radio talk show hosts in America. It is based upon the Talkers editorial board's analysis of a combination of on-air, online and on-demand audiences as indicated by a sampling of ratings services and other metrics-gathering sources and methodologies. It includes hosts whose programs are primarily focused on politics, current events and social issues and are carried on stations and platforms described by generally accepted radio industry terminology as "news/talk" or "talk/radio." Note: It does not include several highly listened-to hosts carried exclusively on SiriusXM Satellite Radio, which operates a closed subscriber system and does not make specific host audience numbers available.
The entries that are self-reported mostly report Nielsen Audio as their source. In the case of the #1 spot on the Wikipedia list, Morning Edition, they report "ACT 1 based on Nielsen Audio". ACT 1 is most likely a reference for this software, which describes itself as "the tool of choice for analyzing Nielsen Audio radio audience data."
Nielsen Audio Ratings
It's a hard problem to solve to determine the number of subscribers on a broadcast network. Nielsen acquired Arbitron in 2012 and re-branded as Nielsen Audio. They collect their ratings data by providing audio panelists with "Portable People Meters":
The PPM is a wearable portable device, much like a pager or mobile phone, that electronically gathers subaudible codes that identify the source of a broadcast, such as a radio station.
While wondering if these sub-audible codes is really an effort at mind control by Nielsen and they're using ratings collection as an effective cover story would be off-topic for this question, as a skeptic I would more readily believe results obtained from this objective sampling which is unavailable unless you are willing to subscribe to their service. Even if I was willing, the answer would most likely depend on which geographical market you are asking about, because that's how it is reported. For instance, the Toledo Blade reported ratings for Fall 2016 in their market, and I did not recognize any of those names as a nationally syndicated talk radio host for the top 10 spots (which did not break out news talk radio separately from other formats).
However, even if Nielsen does want to charge money for their service and even if Talkers Magazine doesn't want to share out exactly what their methodology is, I don't see enough of an economic motivation to just dismiss their results out of hand seeing as how they are widely cited by news and other industry publications. That itself isn't exactly evidence, it just means plenty of other people in the industry take them seriously and have for quite awhile.
Absent any other information, the Talkers Magazine source is probably close to reality (and happens to be free), even if its listing has been described as "heavily subjective". The fact that I have to qualify my statement with "probably" is probably what Cenk was getting at, even if he had to resort to hyperbole to make the point.