Non-Parametric Modelling of Heavy Tails and Skewness of Box-Office Revenues
4th International Conference on Computational and Financial Econometrics (CFE'10), University of London & LSE
Posted: 7 Aug 2010
Date Written: August 6, 2010
Using box-office data for movies released in the US market in the 1990s and 1930s, we establish probabilistic statements for the box-office revenues that the market at these instances dictate. Here, we propose a smooth and non-parametric model of heavy tails and skewness using the GAMLSS (Generalized Additive Models for Location Scale and Shape) framework. In doing so, an understanding of the demand dynamics and adaptive supply arrangements of the motion picture industry is presented. The movie market is particularly interesting due to its skewed and kurtotic macro-regularity, which resulted in the hypothesis that "nobody knows what makes a hit or when it will happen". This hypothesis is revised here.
Keywords: BCPE distribution, non-parametric regression model, motion picture industry, GAMLSS
JEL Classification: C1, C14
Suggested Citation: Suggested Citation