Counterfactual decomposition of movie star effects with star selection

We investigate the effects of a movie star on the movie's opening week theater allocations and box office revenue. Because the pairing of a star and a movie involves a bilateral matching process between the studio and the star, the star (hence the nonstar) movie samples are nonrandom and the star variable is potentially endogenous. To assess the star as well as movie characteristics effects, we utilize a switching model to account for endogenous assignment of stars and nonstars into respective movie samples. In addition to controlling for selection biases, the endogenous switching model generates managerially relevant insights into the factors that influence a star's assignment to a movie. Additionally, because the star and nonstar movie characteristics (e.g., movie budget, distribution, genre, etc.) are often systematically different, we counterfactually estimate the theater allocations and revenues that nonstars (stars) would have generated had they acted in movies endowed with the same characteristics as the star (nonstar) movies. The decomposition analysis, conducted at different quantiles of theater and revenue distributions, shows that the presence of a star has a much stronger effect on theater allocations than the movie characteristics have. However, the revenue difference is entirely contributed by the differences in the characteristics of the star and nonstar movies. Thus, the star effects on revenue come indirectly through the theater allocations as well as from the characteristics of the movies in which they participate.