UC Santa Cruz

Abstract

FADE IN:

EXPLORING THE EFFECTS OF TECHNOLOGICAL CHANGE ON CONSUMERS AND FIRM REVENUES IN HOME ENTERTAINMENT MARKETS FOR FILM

Gabriel Axarlian

This dissertation is composed of three related works, two of which are similar in data and time span with the third being more independent in this respect. The three works focus on the nature of the film industry’s adaptation to changing technologies in home entertainment markets. My analysis studies the effectiveness of new strategies as well as the effect changing technologies has had on market share of various product types.

The first work is titled, “The Introduction of Infinite Durability to an Information Good and the Decision to Buy or Rent: Evidence from the Film Industry”. Home entertainment markets have seen dramatic declines in revenues over the last ten years due to the digital revolution, making new adaptation strategies of crucial importance to maintaining existing profitability. Recently, the film industry has worked with related firms to develop a standardized ecosystem for the sale and access to digital versions of their films. This system, known as UltraViolet, embodies infinite durability; the perpetual access to a purchased film across computing platforms and time. I make use of exogenous variation in the implementation of this system to study the effect the introduction of this new feature has had on the decision to buy versus rent. My findings show that the inclusion of infinite durability creates enough value to spur increased purchasing behavior. Also, the inclusion of infinite durability creates a high degree of substitutability amongst products that include it in purchase markets. Lastly, I find evidence that the lack of inclusion of infinite durability will spur consumers to rent rather than buy.

The second work is titled, “Early to Sell, Do Revenues Rise? The Effect of the Digital Sell-Through Window on Home Entertainment Markets”. In 2012, several film studios began selling digital versions of their films 1-4 weeks before their release into physical media markets (rental and sell-through). This was an attempt to boost overall waning revenues from home entertainment. I examine the effect this has had on physical media sell-through and rental markets using a nested-logit model approach. The analysis yielded three significant results. One is that the length of the window is inversely related to its effect on the behavior of buying or renting of discs. The second is that this strategy has mainly affected the consumption behavior of disc renters to buy digital. The third result of note is that although the strategy is effective, it is not creating loyal consumers of this new method of distribution. This fact notwithstanding the strategy has increased home entertainment revenues.

The third and final work is titled, “The Long Tail Effect in Home Entertainment Rental Markets”. The long tail, theoretically, is the phenomenon that results from the availability of infinite inventory retailers which allow for niche products to have enough of a ‘shelf life’ to find their audience. In both the academic literature and industry discourse there is a debate about the actual nature of the long tail effect. The counter view is that infinite inventory retailers simply make the more popular ‘superstar’ products even more popular, maintaining the small market share of niche products. I make use of exogenous variation in data availability to provide more insight into this puzzle. The differences in data sources allows for an analysis of the demand for films that are niche (proxied by limited theatrical release) and superstars (proxied by high production value). The main finding is that superstars appear to have gained market share in these markets. Although there is not strong evidence showing that niche consumption has gone down, this is consistent evidence that their market share certainly did not increase. Also, an examination of the relationship of the long tail effect and product substitutability yields results that paint a complex picture that cannot be generalized.