Do not worry, I am sticking to prose for this post. This about a paper with the same title by Hugo Mialon. He looks into how economists choose the titles of their papers and finds that 28% of titles include some form of rhetorical device. He also finds that this leads to four more citations for the paper.
While I agree that a title can tease someone into reading some work, there is a big step to generating a citation. Citations are not for titles, but for content (hopefully). Now, would the citer never have read the paper if it were for the title? Poetic titles often hide the true content of a paper, thus a case can be made for missed citations because of the title.
Then why does Mialon find four more citations for poetic titles? He focuses on article published in the top three journals. These papers are read anyway, poetic title or not. I think his result would be reversed when looking at working papers, where a poetic title is really hit or miss. His sample only has the hits.