A wonderful example of confirmation bias. "Oh, look, box office revenues are down... except for the blockbusters where the big studios earn their money, those got a shot in the arm. But we'll still go with 'hurts revenue'.
That is the effect I'd expect to see (yay more cb). People don't want to take a chance with a small indie film unless they know it's good and with MU down, there's no word-of-mouth information to tell them it is. But they still want to go see a movie. The effect being that they bank on something sure - a well-promoted and known movie. In effect blockbusters with huge marketing get a boost, because smaller films' marketing gets nipped.
From the paper: "Box office revenues of movies shown on the average number of screens and below were affected negatively, but the total effect is not statistically significant. For blockbusters (shown on more than 500 screens) the sign is positive (and significant, depending on the specification)."
How can you claim an effect is present when it is not statistically significant? The paper seems to undermine its own "findings." I'm not sure anyone expects torrentfreak to be concerned with accuracy, but it's pretty blatant to have a misleading article title while still quoting the paper as finding the negative effects "insignificant."
I actually understand the passage as such that the effect is significant for the negative impact on the subset of movies shown on less than 500 screens but the overall effect on all movies is not significant.
Without access to the whole paper it is hard to tell though.