Fighting as a profit maximizing strategy in the National Hockey League: more evidence

ABSTRACT

This article estimates the effect of fighting in hockey games on attendance in the National Hockey League (NHL) over the 1997–1998 through 2009–2010 seasons. After estimating a system of equations developed from a model of a profit-maximizing club owner, it was found that fighting had a small negative effect on attendance implying that encouraging fighting on the ice is not a profit-maximizing strategy. The results are quite robust when incorporating capacity constraints on attendance and exogenous ticket pricing. Other factors that determine club performance and market size were found to significantly affect attendance. The empirical results also suggest that NHL club owners are maximizing profit.

KEYWORDS:

• Fighting
• attendance
• revenue
• NHL

JEL CLASSIFICATION:

• L83
• L21
• Z20

Acknowledgements

Thanks to Kenneth Stewart, Peter Von Allmen and seminar participants at the 37th annual meetings of the Eastern Economics Association, New York, 2011, for useful comments.

Notes

¹ Even with large fixed costs, teams might profit-maximize by pricing in the inelastic portion of the ticket demand curve. Ahn and Lee (2007) suggest that habit-formation could be a profit-maximizing rationale for this; Fort (2004) suggests the importance of local television revenues. These are not features of our model.

² We also estimated the system of regression equations using the Fan Cost Index (FCI) instead of the average ticket price. The FCI is an estimate of the total cost of attending an NHL game for a family of four and includes four tickets (at the average ticket price), four hot dogs, four small soda drinks, two small draft beers, two game programmes, two hats (least cost) and one parking space. The FCI for the 1997–2009 seasons was obtained from the sports data website maintained by Rod Fort noted in Appendix 1. Since the cost of tickets is the largest component of the FCI, the estimation results changed only slightly. The price elasticity estimate increased to −1.138.

³ The Tobit estimator can be used when the dependent variable is censored in the case of a single equation regression model. Using the Tobit estimator for sporting attendance models is becoming fairly commonplace in the literature: Welki and Zlatoper (1999) for the National Football League, Czarnitzki and Stadtmann (2002) for the German premier football league, Leadley and Zygmont (2006) for the National Hockey League, Leadley and Zygmont (2005) for the National Basketball Association and Donihue, Findlay, and Newberry (2007) for major league baseball spring training games are good examples. Unfortunately the Tobit estimator is impractical to use here due to the presence of cross-equation restrictions in the system model that prevent single equation estimation using instrumental variables.