Dyed in the wool? An empirical note on fan loyalty

Abstract

This article attempts to investigate empirically the strength of intracity football club support and seeks to test the commonly held belief among the sports community that supporters will not switch allegiance between city rivals very readily. To test this phenomenon, data on the attendances of five English and two Italian city rival clubs is used to conduct a cointegration analysis. The results imply that a stable relationship does not exist between the attendance series which throws doubt on the proposition that supporters of a team are unlikely to switch their support to a rival club. To test the possibility that this result could be due to fickle supporters ceasing to attend, a sensitivity analysis is conducted between attendances and performance.

Notes

¹ These latter papers make up part of the voluminous literature on the economics of sport. The fact that a recent special issue (Vol. 41, Issue 25) of this journal was devoted to this subject demonstrates its growing importance.

² There are too many instances of this to list. A recent example is the move of the National Football League's (NFL) Cleveland Browns to Baltimore in 1995 to become the Ravens.

³ ‘Brand communities and consumption subcultures share common values, rituals and traditions, they also typically have strong sense of who is an authentic user of a brand and who is not’ King (1998, p. 36).

⁴ The term ‘self-monitoring’ refers to how individuals are motivated by what others think of them (Mahony et al., 1999, p. 148).

⁵ Depken II (2000) suggests that sales of merchandise could be used as an indicator of fan loyalty, but as he says such data are not readily available.

⁶ See Dobson and Goddard (2001) for a survey of attendance demand studies in football.

⁷ A well-known problem in the empirical literature is the fact that football attendance is made up of two types of ticket holders – season tickets and game-specific tickets bought sometimes on the day of the game. Clearly, the factors motivating these two groups will be different, given the ‘sunk cost’ nature of the season ticket (Dobson and Goddard, 2001, pp. 317–8). Historical season ticket data are notoriously difficult to obtain and was not available to us for this study. However, since we are using seasonal average attendances, we can assume that both types of ticket buyers will base their decision to purchase to some extent on the previous season's league position.

⁸ Information on football rivalries was gathered from the following sources: Goodhead (2003), Harvey and Strowger (2004) and www.footballderbies.com.

⁹ Due to the suspension of League football during the Second World War, the attendance data ends after the 1938/39 season and starts again in 1946/47.

¹⁰ Several of the selected cities host some smaller clubs although these are usually based in the suburbs. For example, Tranmere Rover's stadium is technically in the Merseyside area along with Everton and Liverpool, but is actually in the town of Birkenhead, not Liverpool. None of the London clubs are included in the analysis due to their sheer number. At the time of writing, no less than five clubs from the English Premier League are based in London.

¹¹ To test the possibility of a structural break in the attendance data after the end of the Second World War, a dummy variable was included in the cointegrating regressions for the British clubs defined as zero up until 1939 and one afterwards. The dummy was not significant and therefore was not included in the main analysis.

¹² See Dobson and Goddard (2001, Chap. 7) for a comprehensive review.