Click fraud would be actual fraud if the the clicks originate from the person getting paid for the advertising. When an unrelated third-party does it, though, it is not.
Though the harms it causes are similar:
* advertiser is paying for clicks/impressions from a bot rather than a user
* publisher/network sending lower quality traffic to advertiser faces reputation and financial risk
Likewise, depending on the situation fake social media profiles also can harm the quality of said network.
In terms of damages caused by the researchers' action... clicks would be the most harmful. Especially against insurance, loans, and mortgages which they seem to be interested in. Those have some of the highest CPCs around. (e.g. http://www.wordstream.com/articles/most-expensive-keywords)