Rethinking rational addictive behaviour and demand for cinema: a study using Japanese panel data

Abstract

In this article, I explore the extent to which the rational addiction model developed by Becker and Murphy (1988) can account for cinema attendance, using panel data from 47 Japanese prefectures for the years 1994–1998. Controlling for unobserved prefecture-specific fixed effects and an endogeneity bias of the lagged and lead-dependent variables (the volume of cinema attendance), by using a fixed effect two-stage least squares estimation, I find that people in Japan are likely to be addictive. This finding is contrary to that found in the United Kingdom (Cameron, 1999). Further results suggest that the younger the people are, the more video is a favoured substitute for the cinema. If other factors are equal, the past accumulated consumption of young people is, of course, smaller than that of older people. Hence, the addictive behaviour is more attenuated for younger people because of their lesser past consumption. The complementary effect of past consumption is more likely to outweigh the substitution effect of preferring video as people become older. These findings are in accord with the Becker and Murphy (1988) model of rational addiction.

Notes

¹ According to Becker et al. (1994), time dummies in part control for the effects of unanticipated growth in wealth.

² The growth of multiplexes was also observed in Europe (Cuadrado and Frasquet, 1999).

³ Becker et al. (1994), Cameron (1999) and Escario and Molina (2001) did not include one lag and one lead of the price. In contrast, Chaloupka (1994), Olekalns and Bardsley (1996) and Waters and Sloan (1995) did include them.

⁴ Becker et al. (1994) incorporated year dummies, which Cameron (1999) did not take into account. The total number of cinema sites and the number of changes of residence within each prefecture were also included in the regression function as independent variables when I conduct the estimation, but due to space limitations, those results are not reported here.

⁵ The price is calculated by the real total revenue divided by the total attendance.

⁶ See Chaloupka (1991), Becker et al. (1994), Cameron (1999), Olekalns and Bardsley (1996) and Escario and Molina (2001).

⁷ I used three lags and three leads of price as instruments. The result is unchanged when four lags and four leads are used.

⁸ Cameron (1999) used the influence of colour television, probably in order to control for substitution. However, during the studied period, colour television was thoroughly established almost everywhere in Japan and therefore, its impact was marginal. The effects of video, rather than colour television, appeared to have a serious impact on cinema attendance.

⁹ Not only films are imported from the USA or Europe but also those from Korea attract interest in Japan (Japan External Trade Organization, 2005; http://www.gaihai.jp/17kunibetu.pdf).

¹⁰ By 2002, the number of multiplex screens has exceeded the total of other types of cinema screen. (http://www.eiren.org/toukei/data.html).