The More You Know … The More You Enjoy? Applying ‘Consumption Capital Theory’ To Motion Picture Franchises

ABSTRACT

In this article, we build on Stigler and Becker’s (1977) “consumption capital theory” and propose a novel conceptualization of film quality for the analysis of motion picture franchises. Generally, this theory predicts that the utility consumers derive from a particular good or service increases with prior consumption. We test our theoretical conjectures by drawing on the population of sequels that were running in the US between 1992 and 2011. The empirical results point to the explanatory power of the proposed framework. Film executives may use our findings to improve the profitability of their sequel productions. From a theoretical point of view, consumption capital theory allows for a more refined analysis of sequel performance along different dimensions. Moreover, it may provide a fruitful basis for the analysis of other serial media content, including books, TV, music, and games.

Notes

¹ Moreover, movie consumption may lead to the formation of consumption capital that is general in the sense that it is not restricted to a particular franchise but provides regular cineasts with an utility gain when watching any kind of movie.

² Such a strategy would also entail lower costs. Compared to upgrades it simply requires cutting out explicit material, instead of having to reshoot certain scenes.

³ In industry jargon, the term reboot is used to describe a new franchise, consisting of a remake and at least one sequel.

⁴ The websites www.the-numbers.com, www.imdb.com, www.metacritic.com, www.rottentomatoes.com, and www.boxofficemojo.com provide complementary information on movies, entertainment individuals and companies. These sources have been widely used in previous film studies.

⁵ Another objection relates to a possible endogeneity problem when using these ratings as a predictor for the commercial outcome of a movie. More recent studies adapt either a simultaneous estimation approach or use independent variable regression (e.g., Basuroy & Ravid, 2013; Chintagunta, Gopinath & Venkataraman, 2010; Duan, Gu & Whinston, 2008).

⁶ Except for the returns from DVD-sales this calculation follows Palia, Ravid and Reisel (2008). We exclude the revenues from home entertainment, because in some cases the DVDs are sold in a box including the original movie and the preceding parts of the franchise, which may distort our results. Data on the returns from merchandising or TV-rights are generally not available.

⁷ As argued above, when producing a movie, the director, producers and writers closely interact, favoring a single variable for the entire team. The considerable correlation between the individual variables comprising the respective functional groups provides another, rather technical argument for this procedure. We further experimented with two variables indicating a re-introduction of central creative and administrative talent. Such returning engagements, however, are very rare in practice and do not alter our results.

⁸ The omitted genre category is thriller and the omitted source is originalscript. To capture possible nonlinear effects we also included the squared terms of number and timelag in our models. The coefficients of both variables, however, turned out to be statistically insignificant.

⁹ The inclusion of this latter variable is inspired by an anonymous reviewer referring to the work of Gil (2009) who found that integrated distributors/exhibitors run their own movies longer than other movies. Unlike in Spain, where Gil’s data originate from, distributors in the United States are not allowed to own theaters. However, and in a similar vein, integrated producers/distributors may arrange with theater owners for a larger number of opening showcases and/or longer timelines for their own films. Film revenues and ROI in turn may be affected by both parameters. Agostini and Saavedra (2011) compare the release strategies of integrated vs. non-integrated distributors/exhibitors in the Chilean market.

¹⁰ The number of observations in these models is reduced to 279 sequels. For n = 9 films the Rottentomatoes website does not report a critics rating; for one additional sequel the respective rating of the parent movie is missing. With a χ²-value of 70.6, a Breusch-Pagan Lagrange multiplier test of independence indicates a significant correlation between the residuals of the two models justifying the use of SUR.

¹¹ To check if the result for the variable number is driven by few franchises with an exceptionally large number of sequels with favorable ratings, we replicated the estimations excluding franchises with more than four installments. The results remained virtually unchanged.

¹² We calculated these figures by entering the average values in the return equation and adjusting the denominator to equal (100 - 28.5) = 71.5% and (100 - 17.6) = 82.4%, respectively. We acknowledge that this approach only allows for a rough approximation, because a possible positive effect of higher costs on a film’s returns, for example, is not considered. Nevertheless, this exercise may illustrate the extent of the underlying effect.

Additional information

Notes on contributors

Christian Opitz

Christian Opitz is a Professor at the Faculty of Business and Economics, Zeppelin University, Friedrichshafen, Germany.

Kay H. Hofmann

Kay H. Hofmann is a Professor at the Faculty of Business Management and Social Sciences, University of Applied Sciences Osnabrück, Osnabrück, Germany.