Attendance demand and core support: evidence from limited-overs cricket

Abstract

In this paper we construct and estimate a model for match attendances in English one-day, limited-overs league cricket. Our dataset separates ‘pay at gate’ spectators from club members, enabling us to distinguish between the attendance decisions of casual spectators and those of the, potentially, more committed ‘core’ supporters. In addition to providing a more direct analysis of the relatively underdeveloped issue of ‘core support’, we also investigate variability within the core. Apart from providing further evidence regarding demand determination in professional team sports we consider some specific policy implications particular to the ‘peculiar’ sport of cricket.

Notes

¹ See Downward and Dawson (2000) for a detailed treatment of the demand for professional team sports and a review of empirical studies as listed in their comprehensive bibliography. The continuing international interest in attendance demand studies has since included: Falter and Perignon (2000, French soccer), Schmidt and Berri (2001, 2002, Major League Baseball),Czarnitzki and Stadtman (2002, German soccer), Garcia and Rodriguez (2002, Spanish soccer), Owen and Weatherston (2004, Southern Hemisphere Super 12 Rugby Union).

² See, for example, Becker and Suls’ (1983) examination of fan behaviour from a psychological viewpoint and, similarly, Wann and Branscombe's (1993) measurement of team identification, and Depken's (2001) study of fan loyalty in the US National Football League.

³ There are many issues involved with measuring a club's population base or potential catchment area for support including boundary definitions, population densities, and the choice of suitably disaggregated data by age and sex. Further problems arise when imputing effective support from potential support. Interest in, and enthusiasm for, a team and its manifestation in attendance need not be strongly correlated with catchment area size (arbitrary or not).

⁴ Paton and Cooke (2005) proxy the core support element in terms of a lagged dependent variable, which they find never to be significant and do not report the specification.

⁵ The initial outcome uncertainty hypothesis can be traced back to Rottenburg (1956) with further development by Neale (1964) and more explicit treatment by El-Hodiri and Quirk (1971). Its relatively persistent employment in econometric studies of team sports is discussed in Downward and Dawson (2000) and Dobson and Goddard (2001). See also Szymanski and Leach (2005).

⁶ There is some evidence of bias in betting markets. See, for example, Vaughan-Williams and Paton's (1997) study of British racetrack betting markets.

⁷ Cairns et al. (1986) note that the Jennett measure uses information unavailable to those agents, the potential attenders at a match, whose behaviour is being modelled, while Cairns (1987) further observes that no role is played by notions of expected relative performances of one club compared to others and that any definition of championship contention (including his own) is unavoidably arbitrary. See also Cairns (1990).

⁸ The large number of studies dedicated to association football, stretching back to Hart et al. (1975), are examined by Dobson and Goddard (2001), while the increasing interest in rugby league includes: Burkitt and Cameron (1992), Baimbridge et al. (1995), Carmichael et al. (1999), Davies et al. (1995), Jones et al. (2000), Dobson et al. (2001), Peel and Thomas (1997). As for studies of other sports, particularly though not exclusively North American, these British studies have covered long time series analysis of aggregate league and club attendances, as well as single season match attendance models.

⁹ See Schofield (1982) for a full economic analysis of the development of first-class cricket in England. The same author also estimates production functions for English county cricket using samples of data for the County Championship and the John Player Sunday League for the seasons 1981–83 (Schofield, 1988). See Bhattacharya and Smyth (2003) for a recent study of the demand for test match cricket in Australia.

¹⁰ A 1-day 60 overs knockout cup competition for the first-class cricketing counties was introduced in 1963 as the Gillette Cup and another limited (55) overs cup competition followed in 1972 as the Benson and Hedges Cup.

¹¹ Full details regarding the rules and regulations determining the conduct of the game can be obtained from the England and Wales Cricket Board (ECB). The league has undergone many sponsorship and name changes over its existence, and in 2000 the competition was restructured to form two divisions linked by a system of relegation and promotion. The initial commitment to Sunday fixtures has now been abandoned with fixtures played on any day with an increasing, though still limited, number of day–night matches part played under floodlights. Over the years, other attempts to make league matches more attractive and spectator friendly have variously involved changing the regulations regarding the maximum number of overs to be bowled (currently 45) and introducing fielding restrictions and over rate penalties, as well as attempting to increase the theatricality of the occasion with coloured clothing (replacing the traditional all white kit) and more emblematic names to identify teams.

¹² The 18 teams are: Derbyshire, Durham, Essex, Glamorgan, Gloucestershire, Hampshire, Kent, Lancashire, Leicestershire, Middlesex, Northamptonshire, Nottinghamshire, Somerset, Surrey, Sussex, Warwicksire, Worcestershire, Yorkshire.

¹³ Paton and Cooke (2005) devised an alternative measure for identifying irrelevant matches, assuming that the match has no likely impact on tournament outcomes if the home team cannot move into a critical promotion, relegation or winner's position even if they win all their remaining games. While the authors allow for a range of demand affecting factors similar to those in our study their specifications and definitions differ. They additionally allow for county fixed effects to capture the fact that a number of factors that are likely to enter the demand function are fixed within each county for any one season, with only limited variation in these factors over two seasons.

¹⁴ Paton and Cooke (2005) proxy closely matched teams by a points difference variable, calculated as the absolute difference in the mean number of points per game between the home and away sides before a game. The variable is adjusted by normalizing for the home-win percentage for all teams, in order to allow for the fact that home teams have a higher win percentage than away teams.

¹⁵ Paton and Cooke (2005) note that these occasional venues (out grounds) are typically located in less populated areas within the county or even in adjoining counties that do not have a first class cricket team of their own. In some cases, however, such matches are played during popular ‘festival’ weeks.

¹⁶ Over the two seasons a small number of matches were rescheduled for midweek when either team were otherwise engaged on the original date, due to playing in a knockout cup final. Additionally, during the 1997 season 3 day–night matches, partly played under floodlights, were experimented within midweek. While remaining an extremely rare and experimental novelty for many seasons, recent evidence (Paton and Cooke, 2005) indicates that attendances are significantly higher at floodlight matches.

¹⁷ Television coverage during the 1996 season was limited to BSkyB satellite coverage of 12 matches on separate Sundays and a small number of Glamorgan matches broadcast in Wales on BBC. During the existence of the Sunday League competition terrestrial television interest has varied considerably with, for an extended period, regular live coverage of matches networked on BBC. The potential TV effect on spectator attendance at sporting fixtures is examined by: Baimbridge et al. (1995, English rugby league; 1996, English soccer) and Carmichael et al. (1999, English rugby league).

¹⁸ Exclusion of all weather affected matches as dummied by CURTMATCH did not materially affect our results.

¹⁹ No H%POINTS or A%POINTS observations were available for matches where one or both teams were playing their first match of a season, while fixed odds were not posted by the bookmakers for a very small number of matches.

²⁰ The reported results are the most parsimonious specifications, obtained by eliminating any variables that were not significant at the 10% level of significance or higher. Results of the full model estimations are available on request from the authors.

²¹ In the continuing search for new and innovative product variants 20/20 cricket was introduced for the 2003 season as an even more action-packed limited overs format, and at the time of writing the ECB have recently announced revised structures for the 1-day limited overs league and cup competitions during the 2006 season. See also Preston et al. (2000) for a ‘blueprint’ for world cricket reform.

²² In addition to its revenue generating aspects large match attendances have potential crowd effects contributing, particularly where there is a large home component, to a home-field advantage (see e.g. Courneya and Carron, 1992).

²³ The issues of membership fee schedules and packages and match day pricing structures offer opportunities for further research requiring additional and disaggregated data for a longer time period.

²⁴ It should be noted that county cricket club season ticket membership covers a full fixture list of County Championship and other competitive and friendly matches (together with various ancillary benefits) as well 1-day league cricket, and that the membership share of total attendance is likely to be greater on a per day basis (albeit of smaller crowds) in 4 day County Championship matches compared with 1-day matches.

²⁵ A recent report by the Octagon sports marketing agency (Octagon, 2005) identifies various insights that help to define cricket fans in terms of conversation, nostalgia, and love of the game, and distinguishes between three distinct fan types based on their emotional relationship with the sport: armchair conservatives, time-out traditionalists, and immersed socializers.

²⁶ Apart from the ECB's attempts to generally increase standards and develop new talent to the general benefit of all clubs, strengthening teams to produce greater quality and more competitive balance is the responsibility of individual clubs. To achieve this clubs require more revenue and, while some counties have a traditionally large membership base, more often than not such revenue generation requires sporting success attracting sponsorship, larger memberships and higher attendances, etc. See Szymanski and Kesenne (2004) and Szymanski and Leach (2005) for a general treatment of competitive balance in the organization of sports leagues.