My wild assumption is that bot farms need to look like regular users, and therefore keep clicking on things even when they're off the clock. If Facebook has caught up on click farms that click on links but do nothing afterwards, I would assume that filling in some random data goes a long way to make bots harder to spot once your actual clients show up.

When you set up an ad on Facebook (and many other ad services), you can define a "conversion" (I'm more familiar with how google calls them, I'm not sure if that's what facebook calls it.). You can read about it here: https://support.google.com/google-ads/answer/1722022. Then, you can set a target "Cost Per Action", which factors into your bids and therefore how much you spend on advertising and where you advertise. That's why if you're doing ad fraud, you'd want to take it all the way through to whichever action "counts".

That has to be it! That's the only explanation I've heard that really makes sense.

Yes, heard this theory before, think this makes the most sense

This kind of fraud benefits Facebook's competitors in the ad space by increasing the cost per customer acquired.

If someone clicks, signs up, and never pays, the conversion rate on the ad looks worse, and Facebook's targeting looks less valuable.

Facebook serves ads on other apps thru their Facebook Audience Network.

Kill competition? Boost Facebook stock price?