Game Outcome Uncertainty and the Demand for International Football Games: Evidence From the German TV Market

ABSTRACT

This article adopts a comprehensive 2-step approach to dissolve recent confusion regarding the role of game outcome uncertainty in the TV demand for international team sports. Analyzing the German TV demand for 457 international football games broadcasted during 6 Fédération Internationale de Football Association World Cups and 5 Union des Associations Européennes de Football European Championships and 287 international games played by the Germany national football team, this study reveals that solely the demand for friendly games is affected by increasing uncertainty regarding the expected game outcomes.

Notes

¹ Although outcome uncertainty has been measured on different temporal scales (cf. Cairns, Jennett, & Sloane, 1986; Szymanski, 2003), in this article, we are exclusively concerned with short-term, individual game outcome uncertainty.

² As Cairns et al. (1986) have stated, in sporting contests “the participants need each other, because to produce at all requires a basic interdependence. As a result (sporting) clubs not only compete but also co-operate with one another” (p. 4). The demand for the resulting products, in turn, is, at least from a theoretical perspective, dependent upon several factors (cf. Borland & Macdonald, 2003; Szymanski, 2003) amongst them uncertainty regarding the expected game outcome. Neale (1964), for example, argues that the “first peculiarity of the economics of professional sports is that receipts depend upon competition among … the teams, not upon business competition among the firms running the contenders for the greater the economic collusion and the more the sporting competition the greater the profits” (p. 2). Accordingly, UOH research has been frequently employed as a basis for justifying regulatory intervention in professional sporting leagues (cf., e.g., Forrest, Simmons, & Buraimo, 2005).

³ Please note that throughout this article the term football refers to European football, which in some parts of the world is also known as soccer.

⁴ Although Feddersen and Rott (2011) were the first to assess the German TV demand for international games played by the DFB team, they do not explore the specific role of game outcome uncertainty in shaping such international football game demand.

⁵ On a side note, it is worth mentioning that by using TV ratings as our approximation for consumer demand (rather than, for example, stadium attendance figures), we are able to avoid several significant methodological limitations present in existing studies on game outcome uncertainty’s ambiguous role in domestic team sport competitions (see, e.g., Chiang & Jane, 2013; Forrest et al., 2005).

⁶ For detailed reviews on the role of outcome uncertainty in the demand for sports in general, see Borland and MacDonald (2003) or Szymanski (2003). For a more recent, football-specific review, see also Pawlowski (2013).

⁷ Off the pitch, Rottenberg’s UOH has been tested by exploring the role of game outcome uncertainty in shaping the TV demand for professional sporting products as diverse as, for example, American college football games (Salaga & Tainsky, 2015), baseball (Chung, Lee & Kang, 2016), Formula 1 Grands Prix (Schreyer & Torgler, 2016), National Association for Stock Car Auto Racing (NASCAR) races (Berkowitz, Depken & Wilson, 2012), National Football League games (Paul & Weinbach, 2007; 2015; Tainsky, 2009; Tainsky & McEvoy, 2012), or Tour de France stages (Van Reeth, 2013).

⁸ As Feddersen and Rott (2011), for example, have argued, international friendlies appear at the bottom of an established game hierarchy: “These matches do not take place within a superior competition (i.e., league or tournament) and are not as attractive to viewers” (p. 357).

⁹ Noticeably, the chosen period of investigation captures all major tournaments since the introduction of the FIFA/Coca-Cola World Rating in August 1993 (i.e., inclusive of 2014 FIFA World Cup Brazil).

¹⁰ In general, a FIFA World Cup/UEFA European Championship features 32/16 national teams competing in 64/31 international games. It is, however, worth noting that only 24 international teams competed during 1994 FIFA World Cup USA, i.e., the first tournament considered for our analysis. Therefore, Brazil was crowned 1994 FIFA World Cup champion after only 52 games.

¹¹ Specifically, our dataset omits 70 international games from 1994 FIFA World Cup United States (6), 1996 UEFA European Championship England (4), 1998 FIFA World Cup France (2), 2002 FIFA World Cup South Korea & Japan (38), 2006 FIFA World Cup Germany (8), 2008 UEFA European Championship Austria & Switzerland (4), 2010 FIFA World Cup South Africa (3), 2012 UEFA European Championship Poland & Ukraine (2) and 2014 FIFA World Cup Brazil (3) due to missing data.

¹² In the 2014 FIFA World Cup Brazil final between Germany (2) and Argentina (5), for example, the numerical value of ADSTD was 3. Nüesch and Franck (2009) provide empirical evidence that the FIFA/Coca-Cola World Ranking and the World Football Elo Rating, an alternative approach to capturing an international team’s expected playing strength, can be used interchangeably. Based on additional data collected prior to all 61 games during the 2014 FIFA World Cup Brazil, we observe a strong and significant correlation between ADSTD and both the absolute difference in the Elo rating (r = 0.61, p < 0.001) and the probability of a draw (r = -0.47, p < 0.001). The FIFA/Coca-Cola World Ranking has frequently been used in studies on the economics of soccer (see, e.g., Feddersen & Rott, 2011; Torgler, 2004, 2006, 2008). For a more detailed discussion on the ranking procedure see FIFA (2017).

¹³ The DFB team, for example, plays derbies against the teams from Austria, Belgium, Czech Republic, Denmark, France, Luxembourg, The Netherlands, Poland, and Switzerland.

¹⁴ Since the relative share of foreigners in Germany’s population is comparatively high and well above the European average (Eurostat, 2015), we further included three team dummies representing the three nationalities that are most frequently found in Germany: Turkey, Poland and Italy (Statistisches Bundesamt, 2014). Although we observe a positive relationship between TV demand and Turkey national football team significant at the 0.05 level, including these dummies, however, does not change the overall tone of our results. We therefore refrain from presenting extended estimations (available upon request from the authors).

¹⁵ Data on TEMPERATURE (in °Celsius), RAIN (in millimeters), SUN (in hours), and WIND (in Beaufort) were provided by the German Meteorological Service and reflect the daily average values from all intact weather stations in Germany.

¹⁶ In accordance with other studies exploring the role of game outcome uncertainty in spectator decision-making (e.g., Baimbridge et al., 1996; Buraimo & Simmons, 2002; García & Rodríguez, 2002), we further control for potential nonlinearity by including a quadratic term of ADSTD. Our results, however, suggest that there is no quadratic relationship between game outcome uncertainty and the demand for international tournament games.

¹⁷ It is worth mentioning that in all specifications, variance inflation factors (VIF) are used for assessing collinearity. Except for specification (7) and (9) all VIFs are clearly below the critical value of 10 (Baum, 2006), indicating that multicollinearity was not a problem.

¹⁸ Accordingly, we were also unable to identify a potential interaction between game outcome uncertainty and GERMAN while conducting additional robustness checks. Results from these additional specifications are available upon request.

¹⁹ It is worth noting, that, in line with domestic (football) league competitions, a season runs from end of August or early September to June or July of the following year; that is, depending on whether a tournament such as the FIFA World Cup is played.

²⁰ This data was provided by kicker, which is Germany’s leading sports magazine.

²¹ It is worth noting that (despite officially being tournament games) both the 1999 and the 2005 FIFA Confederations Cup are clear outliers in terms of spectator interest. One reason may be the particularly late kick-off times during the 1999 FIFA Confederations Cup, México, which resulted in rather low ratings for the group stage games against New Zealand (0.91 million viewers) and the United States (1.21 million viewers). Accordingly, in two additional specifications, that is, specification (4) and (8), we present estimations excluding those 8 FIFA Confederations Cup games in order to explore the robustness of our results.

²² As one reviewer has rightly argued, the fact that the comparatively stable FIFA ranking of the DFB team (M = 6.03, SD = 5.03) is used as the reference point to proxy for game outcome uncertainty may, at first glance, make ADSTD appear inferior to alternative proxies derived from factors such as betting odds. Having analyzed archival data on 38 out of 39 DFB friendlies between 2008 and 2014, we have no reason to believe that the observed, small variances in ADSTD may in fact be problematic. On the contrary, although the DFB team was always ranked between #6 and #2, we observe strong and significant correlations between ADSTD and alternative game outcome uncertainty-proxies based on betting odds such as the absolute difference in winning probabilities (cf., Buraimo and Simmons, 2015; Di Domizio, 2010; r = 0.698, p < .001) or the THEIL-measure employed by Benz et al. (2009; r = -0.949, p < .001).

²³ It is worth noting that although this study marks an important contribution to the economic literature analyzing the TV demand for professional sports, it is nevertheless subject to certain limitations that could constrain the generalization of the results presented herein. Specifically, when analyzing specific markets, external validity tends to be lower than in multi market studies. Future demand studies would, therefore, benefit strongly from adopting a similar approach to analyze spectator demand in different cultural settings.

Additional information

Notes on contributors

Dominik Schreyer

Dominik Schreyer, PhD, is an Assistant Professor with the Center for Sports and Management, WHU-Otto Beisheim School of Management in Düsseldorf, Germany.

Sascha L. Schmidt

Sascha L. Schmidt, PhD, is a Professor with the Center for Sports and Management, WHU-Otto Beisheim School of Management in Düsseldorf, Germany and Center for Research in Economics, Management and the Arts in Basel, Switzerland.

Benno Torgler

Benno Torgler, PhD, is a Professor with the Center for Research in Economics, Management and the Arts in Basel, Switzerland and Queensland Behavioural Economics Group, School of Economics and Finance, Queensland University of Technology in Brisbane, Australia.