One of things I was wondering about when I was posting about "Jumpergate" (here and here, with a related post here) is why someone had never examined the issue of how important a "star" really is in determining the box office for a film. If one wanted to figure out the importance of casting a specific name for a movie, there are a lot of numbers in the public domain that can be examined to make an accurate determination, so why doesn't Hollywood have its own version of Bill James, or a studio have a top executive like Billy Beane? Since there's so much more cash floating around motion pictures than there is with Major League Baseball, you'd think that this sort of numbers-crunching would be absolutely de rigueur. To put it another way, how can studios justify not performing that sort of analysis to the shareholders of their corporate patrons?

Well, it turns out that someone has looked behind the numbers, according to the New York Times, and come to some conclusions that may be highly unsettling to Hollywood agents:

Anita Elberse, an associate professor at the Harvard Business School, tried to measure the average effect of a star by analyzing casting announcements on the price of stocks on the Hollywood Exchange, a simulated market where hundreds of thousands of users trade stocks in individual movies based on their expected box-office revenue. Prices on this exchange have been found to be fairly good predictors of a film’s box-office success.

Ms. Elberse found, for instance, that the announcement in 2002 that Mr. Cruise had dropped out of Cold Mountain — he had been expected to play the lead — reduced the movie’s expected gross by $10 million. The announcement that Mr. Cruise was in talks to play a leading role in The Last Samurai lifted the movie’s expected gross by $28 million.

Combing through 12,000 casting announcements between November of 2001 and December of 2004, related to 600 movie stars and 500 movies, Ms. Elberse found stars, on average, were worth $3 million in theatrical revenue.

Still, Ms. Elberse and other academics suspect that the box-office power of movie stars might be somewhat of a mirage. Ms. Elberse found that, even when casting announcements had an impact on the expected financial outcome of a given film, they had no discernible effect on the share price of the media companies that owned the movie studio — indicating that the participation of a star had no impact on the expected profitability of the studio.

Moreover, even if a star-studded movie does well, it does not necessarily mean that the stars are causing higher ticket sales. In fact, it seems to move the other way around: stars select what they believe are promising projects. And studios prefer to put stars in movies that they expect to be a success.

“Movies with stars are successful not because of the star, but because the star chooses projects that people tend to like,” said Arthur S. De Vany, a professor emeritus of economics at the University of California, Irvine, who has written extensively about the economics of moviemaking. “It’s a movie that makes a star.” (emphasis added)

Since the $3 million in added revenue is often less than what the star is actually paid, the impact of a Tom Cruise or Julia Roberts is usually a wash. A far less expensive performer, when cast in the same role, will turn out to be much more profitable. Sounds like Moneyball to me....