So, your first thought might be for enhancing clarity using techniques like compression and limiting to give the calls a consistent volume and avoid spikes that might bust an ear drum.

This is partially true; I run all these calls through a compressor and limiter for that reason, though I am not encouraged by my employer to be obsessed with making the calls pristine...after all they are done on regular phone lines over regular phones (viz., not on nice microphone) and as such you can't exactly get Hollywood sound; you actually rarely useful data below 175 hrz and what is audible above 2500 us usually very useful when boosted (it becomes very essy, harsh, and hissy)

As a second consideration, many publicly traded companies, needing to carefully word their situations to their shareholders, will record two versions of their call and which one gets aired is dependent on news or other factors that come between the call and the airing of the call (could be a matter of hours, or a matter of days). This is also true to an extent and happens from time to time.

A third consideration you might have is, throat clearing, coughing, rummaging of papers. I'll tell you....the MFS have the driest mouths and lip smack louder than a firecracker. They also don't seem to realize if they shuffle papers next to the phone it will pick it up.

But no, even that is not the main reason.

The main reason they need to pre-record is because they can't read. They can't read simple sentences. I've picked a sentence out at random, and knowing nothing about their insane vernacular (we had fantastic EBITDA margins that gross outstanding for the coming tailwinds that outshine our core foundation pillars and drivers of growth) I was able to read them without messing up.

And yet they....will frequently have to read the same sentence 2-10 times. I'm not kidding. Most of these CEOs are fucking imbeciles and mean ones at that. They can barely read a sentence without fucking up. It sometimes takes me an hour to edit together a 15 min call.

On rare occasions it's because they care. I'm under NDA but I'll just say I have worked with a certain publicly traded meat-alternative company that has a lot of re-recording and edits but it's because their CEO (seems to me) very passionate about what he's doing and agonizes over the right word choices even up until the moment of recording. Props to him. He's taking pride in what he does and can actually read a full sentence.

Other people on the other side of the spectrum can't even be bothered to read their script before they show up and don't know how to pronounce their own product names.

TL;DR: I am mostly there to make sure I have a clear pronunciation of every line of the script, take notes on where there are errors, and edit the script together to make a coherent whole at the end without any gross factual error. I do a little bit of processing to get rid of throat clearing, make the volume consistent.