Unpredictable Blockbusters

The movie industry is among the most interesting to study from an economic and financial point of view. By definition "every movie is different from any other". They also differ from many other products whose demand can be reasonably forecasted in advance - the success of every movie is very uncertain until it's screened. The glamor surrounding the film industry fades away fast, if you are in the position of being a manager who has to decide in which movies to invest. In fact, numerous studies argue that it's practically impossible to predict whether a filmed work will be a hit at the box office or not. The frequency distribution of box office revenues is affected by an extremely high variance. In practical terms, this means that the average revenues of a movie company are not predictors of future revenues coming from new productions. And this because you only need a couple of hugely successful movies (e.g. Avatar in the 00s or E.T. in the 80s) to make a studio profitable.
A recent analysis I wrote ("The dark side of the movie. The difficult balance between risk and return") highlights the extreme financial riskiness and economic uncertainty of the industry. Starting from a sample of 4,178 movies shown in US cinemas from 1988 to 1999, I performed econometric testing on a dataset of 1,636 films, which excludes the numerous outliers which would have caused biased estimates. What is the main message the analysis sends to prospective investors in the industry? That there is a positive correlation between production costs and box office revenues (costlier movies, involving famous actors and celebrated directors, usually sell more tickets in movie theaters). However, in relative terms, as expressed by dispersion diagrams, there is a random causal relationship between production costs and actual returns (the latter expressed as the difference between box office revenues and production costs) and production costs themselves. And this correlation makes managerial decision-making is particularly arduous in the film industry.
Why then has Hollywood expanded so much over the last decades, in spite of the high variance that makes the movie industry so unpredictable? The answer lies in the fact that Hollywood Majors have been able to differentiate their products, by making very different, in terms of cost and content, productions co-exist. Statistically, most of the n films produced by a studio in given a year are unprofitable; however movie companies know that revenue flows from just one or two successful movies can offset the losses of the other movies. It sounds incredible, but it actually happens in the film industry. An example can speak better than a thousand words: in 1997, the film Titanic made $413.2 million in US movie theaters alone, i.e. more than the other 13 productions of that year put together, which grossed $30.6 million on average.

Paradoxically, 20th Century Fox would have still been profitable in 1997 with Titanic only, even if all its other movies had been total box office flops. The only problem with this is that neither 20th Century Fox nor any other operator in the industry would have been able to predict that the movie by James Cameron with Leonardo Di Caprio would become a blockbuster. And this for a simple reason: it's just impossible to forecast the likelihood of a movie hit on the basis of ex ante variables.