Should UEFA Introduce an European Major League for European Club Teams?

Should UEFA Introduce an European

Major League for European Club Teams?

Master Thesis IB&M

BY W.J.Tankink Supervisor: Drs. A. Visscher Referent: Dr. R.W. de Vries Groningen, Februari, 1, 2016 University of Groningen

Faculty of Economic and Business

2

Abstract

Nowadays, there exist large differences between European national competitions in terms of television money, which results in large differences between national competitions in terms of talent and the level of soccer players. This is caused by the fact that the labor market is open, since 1996, due to the introduction of the Bosman Ruling. However, the product market is still not open. Subsequently, a separation could be made between the big four competitions in Europe, which are England, Germany, Italy, and Spain and the rest of the European

competitions, which are poor(er) than the big four. According to Kesenne (2007), these differences could be minimized by opening up the product market, by introducing an European Major League. The European top teams, however, are against the introduction of European Super League, as they believe to make more money by staying in their national championship and, on top of that, taking all the money of the champions league. This paper, therefore, examines whether playing at an international stage is beneficial for professional football club teams. European soccer clubs are analyzed, based on underlying factors of revenue instead of revenues itself. Due to the introduction of an European Major League, the competitive balance will change, and according to the literature, this affects the soccer attendance. The soccer attendance consists of tickets sold at a particular match, which is part of the revenue of clubs. Therefore, the competitive balance – soccer attendance relationship is examined here, by focusing on the top European Club Teams matches played internationally and nationally, in order to conclude whether the introduction of an European Super League really harms European top teams in their financial situation. Thereby, the effect on the already existing national competitions is examined, as the top teams will leave their initial national competition and play in the new created European Major League, due to the introduction of an European Major League.

3

Preface

This thesis is made as a completion of the master education in International Business and Management (IB&M). Yours truly has a bachelor degree in the dutch version of Economics and Business Economics (E/BE) from the University of Groningen and this thesis is the product of the master period, which is the last part of the International Business and Management study at Groningen University, Faculty of Economics and Business, MSc. International Business and Management.

This master thesis is written by the contribution of several persons at an academically level, practically level and a supportive level. I would therefore firstly like to thank my head supervisor Ad Visscher and referent Rudi de Vries for their time, valuable input and support throughout the entire period, that I was writing my master thesis. Furthermore I would like to thank Ruud de Koning for his help by brainstorming about a relevant study in the field of sports economics.

4

Table of Content

List of Tables and Figures ... 6

1. Introduction ... 7

1.1 Research Questions ... 8

1.1.1 Participating Club Teams ... 8

1.1.1.2 Home Advantage ... 9 2. Literature Review ... 9 2.1 Soccer attendance ... 10 2.1.1 Fan Behavior ... 10 2.1.2 Competitive Balance ... 11 2.2 Home advantage ... 13

2.2.1 The Negative Effect of Home Advantage on Soccer Attendance ... 13

2.2.2 The Positive Effect of “Superstars” and “Local Heroes” on Soccer Attendance ... 14

3. Methodology ... 15

3.1 Methods to Calculate Competitive Balance ... 15

3.2 Soccer Attendance ... 16

3.3 Competitive Balance ... 16

3.4 How to Test the Hypotheses ... 18

5

Appendix A ... 39

Appendix B ... 40

Appendix C ... 41

Appendix C 1: Competition rules used by national competitions ... 41

C 1.1: Premier League (Eng) ... 41

C 1.2: Ligue 1 (Fra) ... 41

C 1.3: Bundesliga (Ger) ... 42

C 1.4: Serie A (Ita) ... 43

C 1.5: Primeira Liga (Por) ... 44

C 1.6: Primera Division (Spa) ... 45

C 2: Tournament type of rules used by the UEFA Champions League and UEFA Europe League . 46 C 2.1: UEFA Champions League ... 46

6

List of Tables and Figures

Tables

Table 1: ANOVA table of group 1

Table 2: Left out of the sample of Group 2 Table 3: ANOVA table of group 2

Table 4: Data about both regression lines (Group 1 and Group 2) Table 5: Left out of the sample of Group 3

Table 6: ANOVA table of group 3

Table 7: Left out of the sample of Group 4 Table 8: ANOVA table of group 4

Table 9: Data about both regression lines (Group 3 and Group 4) Table 10: UEFA Rankings for Club Competitions: Individual Teams Table 11: UEFA Rankings for Club Competitions: Country

Figures

Figure 1: Group 1: International matches played

Figure 2: Group 2: National matches played by top 18 European Clubs Figure 3: Group 3: Top 6 European competition matches

7

1. Introduction

One person, named Jean-Marc Bosman, has had a profound effect on the transfer of football players within the European Union. In 1990, his contract had expired, and therefore, he wanted to change teams. However, Bosman’s team, RFC Liège, asked a transfer fee, even though, Bosman’s contract had expired. The potential future team, however, refused to meet RFC Liège’s transfer fee demand, causing the refuse of RFC Liège to let him go. Therefore, Bosman took his case to the European Court of Justice as he sued for restraint of trade according to FIFA’s rules regarding football. Eventually, the court of justice judged that the system, as it was constituted, placed a restriction on the free movement of labor, which caused the free transfer of Bosman and all other EU football players after the expiration of their contracts, by the restriction that, the transfer was from a club within one EU Association to a club within another EU Association, nowadays known as the Bosman Ruling. Bosman’s case also caused the prohibition of the imposition of quotas on foreign players regarding domestic football leagues in EU member states, and also UEFA. Still, after the Bosman ruling, quotas could be imposed, but only regarding the number of non-EU players on each team.

Kesenne (2007) examined theoretically the effects of the Bosman ruling within and across countries, regarding club teams. He did research to the consequences of the Bosman verdict of the European Court of Justice in 1995. This verdict deregulated the European player market, whereby the gap between budgets, as well as performances, of the football teams in the rich and poor countries has clearly widened. England, Germany, Italy, and Spain, which are called the big four, could be considered as the rich (football) countries. The remaining countries, such as France, Belgium, and Holland, could be considered as the poor(er) (football) countries.

According to Kesenne (2007), for example, the average budget in the English Premier League (€77 million) is 12 times the average budget in the Belgium Jupiler League (€6.5 million). What causes this budget differences is not a difference in popularity of football in both countries, as the attendance/capita of Belgium (0.31) is higher than in England (0.27) (Kesenne, 2007). The differences in the size of the national market is the main explanation of these budget differences. Namely, the English Premier League (50.000 habitants) is 5 times bigger than the Belgium Jupiler League (10.000 habitants). However, due to these budget differences, it seems obvious that all the best players are leaving the poor(er) countries, and they will run after the big money, which is in line with Kesenne (2007), as his model predicts that more talent will flow to the countries with the bigger markets for club football, which increases the disparity between club teams across countries.

This widening of the gap between football’s rich and poor causes many people in these poor countries to be unhappy. The underlying factor, that causes this gap to widen, is the differences in budget in the European countries. Former European top clubs, such as

Anderlecht, Ajax and PSV, which are top teams in small countries, have to compete with, for example, Real Madrid and Manchester Unit, which are top teams from the big four, to get the best players at an open European labor market, but they are not able to compete with those clubs at an open European product market.

Possible solutions to narrow this gap are to reduce the number of first-division clubs in the small countries or to turn the clock backwards by restricting again the free move of

8

1.1 Research Questions

1.1.1 Participating Club Teams

Going international is the aim of every professional football club. In case of European football clubs, the ultimate goal is to play at the UEFA Champions League, which is the highest international tournament for club teams. Another international tournament for European football clubs is the UEFA Europa League, which is the second international tournament of Europe for club teams. As going international is the aim of every football club, plus the solution provided by Kesenne (2007) to narrow the rich-poor gap between the biggest and smaller national competitions, why does the UEFA not create a European major league? Kesenne (2007) concluded that the European top teams will oppose the introduction of an European major league, as they believe to make more money by staying in their national championship and, on top of that, taking all the money of the champions league. This paper, therefore, will examine whether playing at an international stage is beneficial for professional football club teams.

“Are European top teams actually disadvantaged by acting in an European major league and leaving their national competition, compared to their current situation in which they are participating in their national competition and eventually participating in one of the two European tournaments?”

Multiple researchers, such as Haan, Koning, and van Witteloostuijn (2002), and Solberg and Turner, (2010), have investigated the gap between poor and rich football

countries and recognized that this gap is widening and widening. This gap will even widening more due to the new mega deal sky sports has made to broadcast more than 80% of the English Premier League matches for the coming three years by which they have paid 5.66 billion euros. Therefore, several researches came to the conclusion that a next logical step would be to introduce an European super league (Hoehn and Szymanski, 1999; Kesenne, 2007; Szymanski and Zimbalist, 2005; Vrooman, 2007), which opens up the product market whereby the gap between poor and rich will become smaller.

None of the researches further tried to examine such an European super league, except Kesenne (2007) and Solberg & Gratton (2004). Kesenne (2007) has developed a model by which he analyzes how the competitive balance is affected if one moves from the scenario with nationally protected labor and product markets to a scenario with an internationally open labor market and a closed product market. He also investigated a third scenario, by which he investigates what the possible consequences are of opening up the labor market as well as the product market. However, he calculated and visualized the scenario by making use of a model, he did not make use of the real life world. Solberg and Gratton (2004) did make use of real life data. They analyzed whether European soccer clubs would benefit from joining a ‘Super League’, by which they based their analysis on revenue figures. I will analyze whether European soccer clubs would benefit from joining a ‘Super League’ as well, however, here the analysis is based on underlying factors of revenue instead of revenues itself.

1.1.1.1 Soccer Attendance

In order to answer the research question, answers will be provided to sub questions in which national and international matches will be compared regarding ‘soccer attendance’ and ‘home advantage’.

9

Soccer attendance is a factor that contributes to both the performance and health of a professional football club. Due to increasing soccer attendance, the home advantage will increase which has a positive effect on the performance. Increasing soccer attendance also leads to more tickets sold, which will increase the football club’s health. If the ‘soccer

attendance’ level increases, in case of international games, the top teams will benefit from the introduction of European major league.

1.1.1.2 Home Advantage

Another contribution to the research question, stated above, is to what extent ‘home advantage’ will increase due to the introduction of an European major league.

“Will European top teams experience stronger ‘home advantage’ when performing at an international stage, compared to their performance at a national stage?”

Home advantage has a significantly positive impact on the points scored by a

particular team within a season (Barnett and Hilditch, 1993). If the ‘home advantage’ factor is stronger in international matches, the top teams will profit from this due to more points scored at home and the points scored determine the position on the league table.

1.1.2 Non-Participating Club Teams

By creating an European major league, not only the participants will be influenced but the club teams that stay in the national competition and will not be participating in the

European major league will be influenced as well. The club teams that participate in the European major league will leave the national competition, whereby the overall competitive balance of the national competition will be positively influenced, which increases the ‘soccer attendance’ of the teams that stay in the national competition (Kesenne, 2007). According to Kesenne (2007), however, as the national competitions will maintain similar sizes, the places left by the European major league participants have to be filled up by club teams from the second national competition, whereby the initial increase in competitive balance will be diluted.

According to Pollard (2006b), there is not one single dominant factor that influences ‘home advantage’. Home advantage is the result of a combined effect of individual factors, as the individual factors interact with each other in a manner, which have to be established. One of the factors that have a positive effect on the home advantage factor is crowd size (Nevill, Newell, and Gale, 1996) The ‘home advantage’ factor will slightly change due to

geographical distance differences (Pollard, 2006b), however, this will have no significant effect on the home advantage due to the departure of several club teams out of the national competition, as the distribution of clubs throughout the country will likely stay the same.

“How will the not-participating club teams that stay in their national competition be

influenced by the introduction of an European major league in terms of ‘soccer attendance’ and ‘home advantage’?”

2. Literature Review

Due to the Bosman Ruling, the European football market has become too unfair. The gap between the richest European football countries and the poorer European football

10

Europe, read broadcasting markets, which causes huge differences between different European countries, as the player talents migrate towards the major leagues and teams following the big money.

An example about the product market in European football countries is the differences in television rights paid within a country. Nowadays, absurd amounts of money are paid for television rights, especially in the English Premier League. This causes budget differences between European top teams across Europe. For example, last year’s degradant of the English Premier League, Cardiff, received twice as much TV money as the last year champions of France (PSG) and Germany (Bayern München), and more than 10 times as much as Holland’s champion, Ajax. Cardiff received 84 million, while PSG, Bayern München, and Ajax received 40 million, 40 million, 8 million respectively for TV money. This causes enormous

differences between the money clubs can spent.

The best way to narrow the rich-poor gap is to introduce an European Super League, by which the European product will be opened up. However, the European clubs that will be invited for such an European major league are against this introduction as they believe they become financially less healthier in that case. By focusing on differences in soccer attendance in relation to competitive balance, national and international, and by focusing on differences in home advantage, national and international, this paper contributes to an answer about the prediction of European top clubs that they become financially less healthy due to the introduction of an European super league.

2.1 Soccer attendance

Soccer attendance is a factor that could contribute to both the performance and health of a professional football club. Due to increasing soccer attendance, the home advantage will increase which has a positive effect on the performance (Nevill, Newell, and Gale, 1996). Increasing soccer attendance also leads to more tickets sold, which will increase the football club’s health. Many factors could have an effect on ‘soccer attendance’. Two factors that have an effect on the soccer attendance are fan behavior and competitive balance. Variables that were examined in previous studies to determine the relationship between fan behavior and attendance are; the timing of the games, economy, and certain preferences of the fans (Laverie & Arnett, 2000; Parry, Jones, & Wann, 2014). Certain studies have found competitive balance as a factor that affected the attendance rates (Borooah & Mangan, 2012; Forrest et al., 2005; Lenten, 2009; and Pawlowski & Budzinksi, 2013). When it comes to international soccer attendance, fan behavior and competitive balance are related (Blyth, 2015). However,

previous studies have provided explanations and have helped to communicate why attendance has been fluctuating. According to Blyth (2015), using timing of the games, economy and certain preferences of the fans as variables to determine the relationship between fan behavior and attendance, develop a clearer understanding of the relationship between competitive balance, fan behavior, and attendance. Previous research has examined this relationship of how fan behavior and competitive balance relates to attendance rates of international soccer events (Blyth, 2015). However Blyth’s research has focused on several national leagues (the Barclays Premier League (England), the Liga BBVA (Spain), and the Bundesliga(Germany)) within Europe.

2.1.1 Fan Behavior

11

Sunghyup (2014) focused their research on spectators who attended Iranian Premier League soccer matches and showed that event quality had a significant impact on fan satisfaction. They concluded that their results were important, as higher fan satisfaction means that fans were more inclined to buy tickets and attend these matches.

Another study that was focusing on fan behavior was the research by End, Dietz-Ulher, and Demakakos (2003). They examined how people perceived fans that were basking in reflected glory (BIRG) with a team that they are identified with. End et al. (2003) describe conceptualize BIRGing as, fans that referred more often to their team as ‘we’ when the team was victorious than when their team had suffered a defeat and fans that were more likely to display the insignia of their team on their clothing or wear clothing of their favorite team following a victory than following a loss. According to their results, fans that exhibited constant behavior were perceived as being significantly more of a fan than those that practiced BIRGing. Constant behavior is defined, by End et al. (2003), as someone who maintains a public association with a team regardless of the outcome.

End et al. (2003) have stated that the fans that reacted when their team was performing at a high level were less likely to attend games than those fans that demonstrated a more constant behavior. Similarly, a study conducted by Ware and Kowalski (2012), compared the differences of males and females in sport fan behavior. Ware and Kowalski (2012) found that gender was not really an issue regarding a loss/win. Ware and Kowalski (2012) stated that fans are more likely to attend sporting events if they practiced CORFing or BIRGing because that means they are associated with that team and they are true fans, which means that they found that high involvement in the team was more likely to occur when fans were BIRGing, or CORFing after a loss/win (Ware and Kowalski, 2012). CORFing means cutting of reflected glory and is defined by Campbell Jr. et al. (2004: 151) as, ‘Rather than lessen an internal self

image, or suffer the consequences of a weakened position in the social environment, fans will often dissociate themselves from unsuccessful teams’. This finding, by Ware and Kowalski

(2012), is in contrast to what End et al (2003) stated, as they stated that if fans were BIRGing or CORFing they were less likely to be a true fan and would be less inclined to attend games.

Parker and Fink (2010) conducted a study about the behavior of fans as well. They looked at how team’s sponsors affected fans if there was a negative reputation. Their results showed a positive relationship between the fans’ positive attitude and sponsor. Highly identified fans have significantly more positive attitudes towards the team sponsors in comparison with the fans who identified there selves less with the team (Parker and Fink, 2010). Parker and Fink’s research showed that fans that identified themselves more with one team were more likely to stick by the team, support them, and continue to attend the games. Fans would move away from the team and start following another team that they felt strongly for as the fans were less identified with a team.

2.1.2 Competitive Balance

“Competitive balance is two teams being equal or two teams being a mismatch when they played each other, and if there was uncertainty about the outcome”. (Blyth, 2015: p. 5)

Hogan, Massey, and Massey (2013) were focusing on match attendance within the three main rugby leagues. Their results showed that fans would be less inclined to attend if the team were out of competition, and that the majority of attendees at events were primarily the home teams fans (Hogan et al., 2013). The study by Hogan et al. (2013) gives a clearer understanding of how competitive balance affected the attendance of sporting events in Europe.

12

Budzinksi, 2013). Another result shown by Pawlowski and Budzinksi (2013) was that more than 50% of the respondents is willing to pay for the improvement of the current degree of perceived competitive balance in the corresponding national league or even for the

maintenance of that current degree. For example, the Danish League. Pawlowski and

Budzinksi (2013) found that the average value of going to a Danish League match was around three euros. Fans, however, were willing to pay more than five euros as the perceived

competitive balance of those matches was high (Pawlowski & Budzinksi, 2013). Pawlowski & Budzinski (2013) relevance was that it showed the relationship between competitive balance and fan attendance. Lenten (2009) showed in their study, that there is a positive relationship between competitive balance and match attendance. Forrest, Beaumont, Goddard, and Simmons (2005) also conducted a study about soccer attendance. Their study examined the relationship between attendance and match level uncertainty in the English Football League. Forrest et al. (2005) consider match level uncertainty as individual matches where the outcome is highly uncertain due to the equality of teams in terms of playing ability. They found that a perfectly balanced game had a significantly higher attendance rate than a typical game, as the fans they examined were interested in ex ante match uncertainty. Borooah and Mangan (2012) showed that even when the competitive balance was low, still there was a high level of popularity with the English Premier League. However, this is only the case in the English Premier League.

The European top teams that are eligible for an European major league are the ones that are in the top 18 best performing individual club teams in Europe according to the UEFA ranking for club competitions1, which is visible in Appendix A. Due to this criterion, the talent level and performance level of all eighteen teams is more or less equal. This means that the match outcomes are highly uncertain due to the equality of teams in terms of playing ability. This is in line with the findings of Kesenne (2007), who found that, by the opening of the European labor market and European product market, the competitive balance between the top clubs in the European division is more equal. According to Forrest et al. (2005), highly uncertain match outcomes significantly increase the attendance rate. Taking this into consideration, I propose that the European top clubs, when performing within such an European division, have significantly higher attendance rates compared to their current situation.

H1: European top clubs’ attendance rates are significantly higher when acting in an European division.

If an European division is introduced, the 18, internationally, best performing club teams will leave their national competition. Which means that domestic competitions loses their best performing teams and maintain their less performing teams. This makes me to suggest that the competitive balance, in case of the European domestic leagues, increases, as is concluded by Solberg and Gratton (2004). They state that, if a select group of clubs form an European super league and cease to act in their domestic league, makes those domestic league more balanced. According to Blyth (2015), who did research to such domestic leagues, is competitive balance positively related to soccer attendance, which makes me to propose that the soccer attendance within those domestic leagues will increase as well, in relation to the competitive balance.

H2: The European club teams that stay in their national competition, after the introduction of an European super league, are confronted with significantly higher attendance rates.

13

2.2 Home advantage

Lots of research has been done regarding ‘home-advantage’. Home advantage effects exist in many sports, but the extent varies from sport to sport. According to Courneya and Carron (1992), home advantage appears to be strongest in football. The traditional definition of home advantage is: “a win– loss ratio of greater than 50% under a balanced schedule

where each team plays an equal number of home and away matches” (Poulter, 2009: p. 797).

According to Morris (1981), home advantage in football is, “a basic territorial reaction

caused by the way in which visiting players react to unfamiliar routines in unfamiliar locations.” (Pollard, 2006b: p. 232) Morris (1981) hypothesized that this phenomenon was

not a single-shot event, but it was likely to be worldwide.

Most of the research, however, was at a national level (Clark & Norman, 1995; Pollard, 2006a, 2006b). Only a few studies have examined the home advantage at an

international level (Brown et al., 2002; Poulter, 2009). To contribute to the work of Kesenne (2007), the magnitude of home advantage in an international setting will be compared to the home advantage in a national setting. Additionally, these studies have focused on home advantage in a complete national competition. Only very few studies have focused on home advantage for individual clubs (Barnett and Hilditch, 1993; Clarke and Norman, 1995). However, both studies were focusing on English club teams only. Here, some English club teams will be included as well, but only the ones that are in the top 18 best performing individual club teams in Europe according to the UEFA ranking for club competitions.

Several aspects, such as crowd support, travel effects and accustomization to local playing conditions, were all been considered in attempts to throw light on the cause of

football’s home advantage (Barnett & Hilditch, 1993; Clarke & Norman, 1995; Dowie, 1982; Nevill, Newell, & Gale, 1996; Pollard, 1986). Evidence showed that referee bias may also be a factor that causes home advantage (Nevill, Balmer, & Williams, 2002), as well as

territoriality (Neave & Wolfson, 2003). Territoriality is defined as: “the protective response

to an invasion of one’s perceived territory.” (Pollard, 2006b: p. 231) However, none of the

researchers was able to find an one single dominant factor that influences home advantage. Home advantage is likely to be a result of a combined effect of individual factors, as it is likely that the individual factors interact with each other in a manner, which have to be established.

Pollard and Pollard (2005) provides us with a general review of home advantage in football. Their model describes the mechanisms of the interacting effects of crowd support, travel effects and other possible causes, such as special tactics and psychological factors. Pollard and Pollard (2005) show evidence of considerable variation in home advantage in different national competitions across Europe. For example the Balkan nations of southeastern Europe, show a much higher home advantage than elsewhere in Europe, especially Albania and Bosnia. In general, the northern part of Europe, from the Baltic republics of Estonia, Latvia and Lithuania, throughout Scandinavia and to the British Isles, the home advantage was lower.

2.2.1 The Negative Effect of Home Advantage on Soccer Attendance

14

Simmons’ (2002) study found that match attendance is indeed influenced by the degree of outcome uncertainty. This influence is to such an extent that, if weak teams confront strong teams, however, with the elimination of home advantage, seasonal aggregate attendances would fall. Another study that has taken the home advantage factor into account is the

research by Forrest et al. (2005). Forrest et al. (2005) proposed that fans have a preference for outcome uncertainty, but nevertheless, due to the phenomenon of home advantage, equalizing quality across teams would lower aggregate attendances. Their results indicate that this proposition is not implausible, as they found that the audience size is increased at an increasing rate as the probability ratio is lowered towards one. In other words, the most prospectively balanced contests (i.e. between weak home and strong away teams) attract disproportionate attendances relative to all other fixtures. Therefore Forrest et al. (2005) conclude that fans are interested in ex ante match uncertainty, whereby an perfectly balanced match has an significantly higher attendance rate than a typical match. If the quality

differences between clubs would be removed, Forrest et al. (2005) find a lower aggregate attendances in the English Football League by over 25%.

2.2.2 The Positive Effect of “Superstars” and “Local Heroes” on Soccer Attendance By introducing an European super league, the quality differences between clubs is decreased or in some cases completely disappeared. However, due to the participants of such an European major league, which are the best performing European clubs, most of the outstanding players – called stars (Brandes et al., 2007) – are performing in this European competition. Multiple researchers have found that these ‘superstars’ play an important role in attracting fans (Berri, Schmidt, & Brook, 2004; Berri & Schmidt, 2006; Hausman & Leonard, 1997; Mullin & Dunn, 2002). Berri, Schimdt, & Brook (2004) and Berri & Schmidt (2006) showed in their research that superstars in the NBA (Basketball) attract fans. Brandes et al. (2007) examined a similar relationship, however, they examined this relationship in German soccer. They came to the conclusion that well-known superstars attract fans both at home and on the road due to their outstanding talent. This indicates that talent also has a positive effect on match attendance.

The equalization of club quality and equalization of financial funds between clubs is necessary anyway, as the differences between European clubs are way to large now. Taking this into consideration plus the findings of the studies presented above, I suggest that, in case of an European super league, the negative effect of home advantage on soccer attendance due to the equalization of team quality will be weakened or even cancelled out due to the presence of ‘superstars’, which favors the introduction of an European major league. An more general finding by Koenigstorfer et al. (2010), strengthen the weakening of the negative effect and favors an introduction of an European super league. They found that German fans perceive international success of national clubs as a highly relevant factor to attend matches.

Nevill, Newell, & Gale (1996) did research to the overall home advantage within the eight major divisions of the English and Scottish football leagues. They found significantly variations in the degree of home advantage across the divisions. They found that these divisional difference in home advantage were significantly associated with the mean

attendance of each division. First of all, I, therefore, suggest that the home advantage factor is of less importance in the remaining national competitions by the introduction of an European super league.

15

phenomenon will not weaken or cancel out the negative home advantage effect here. Brandes et al. (2007), however, contributed to these studies by examining not only the effect of well-known superstars on match attendance but also the effect of ‘local heroes’, which are the most valued players of a particular team that has no superstars. They found that ‘local heroes’ attract fans only in home games due to their popularity, while ‘superstars’ attract fans both at home and on the road due to their outstanding talent. Therefore, I propose that this negative home advantage effect, although in weakened form, will be weakened or even cancelled out in the remaining national competitions as well due to these so-called ‘local heroes’.

3. Methodology

The competitive balance and soccer attendance are examined here, for specific clubs in national context and these results are compared to competitive balance and soccer attendance in international context, which are the games played in the UEFA Champions League and the UEFA Europe League, the two international competitions for club teams coming from

European competitions. By comparing the soccer attendance and competitive balance of games played in a national context to the soccer attendance and competitive balance of games played in an international context, it becomes more clear how to answer the research question.

3.1 Methods to Calculate Competitive Balance

Most commonly does competitive balance indicators include league position, winning percentage and points won by clubs. For example, Pawlowski et al. (2010) provides us with multiple, sophisticated types of measurement for competitive balance. However, all his measures, used in his paper ‘Top Clubs’ Performance and the Competitive Situation in

European Domestic Football Competitions’, depend on the points scored in a season and the

team’s league position at the end of a season. One of the measures he used is a modified Hirshman Herfindahl Index, which is called the Hirshman Herfindahl Index of Competitive Balance (HICB), which is explained by Depken (1999) as, the ratio of the Hirshman

Herfindahl Index to the Hirshman Herfindahl Index of a perfectly balanced league. This measure is based upon the sum of the quadratic share of points (𝑆_𝑖2) won by each club in a league with N teams, divided by 1/N (number of teams in competition), times 100.

𝐻𝐼𝐶𝐵 =∑ 𝑆𝑖 2 𝑁 𝑖=1

1/𝑁 ∗ 100

However, this way of measurement is not possible here as this measurement is based upon the points scored within a team with N teams. First of all, there is no ‘league’ yet for the international case, which makes it impossible to get a number for N and second, all clubs play a different number of international matches, which makes it unrealistic to base the

competitive balance measure on points scored here because different clubs are able to score different number of points.

Another measure that is used by Pawlowski et al. (2010) is, the competitive balance ratio (CBR), which is a comprehensive and comparable measure of competitive balance as it is made up of both the average standard deviation of team points and the average standard deviation of league points: 𝐶𝐵𝑅 = 𝑆𝐷𝑇𝑃𝑇

𝑆𝐷𝐿𝑃𝑁, whereby 𝑆𝐷𝑇𝑃𝑇,𝑖 = √

∑𝑇𝑡=1(𝑇𝑃𝑡,𝑖−𝑇𝑃𝑖)2

𝑇 , which is the individual standard deviation of total points won per season (𝑇𝑃_𝑡,𝑖) by team i and the average points won per season (𝑇𝑃𝑖) across a certain number of T seasons. The small letter ‘t’ = a particular season and the capital letter ‘T’ = number of seasons (time span). 𝑆𝐷𝐿𝑃_𝑁,𝑡 = √∑𝑁𝑖=1(𝑇𝑃𝑡,𝑖−𝑇𝑃𝑖)2

16

number of points achieved and the total number of opponents. The international part should be based on matches that were played for the UEFA Champions League and UEFA Europe League, as there are no other possibilities in case of European club teams. However, these competitions use a tournament type of rules. Therefore, every team is playing another number of international matches whereby this method explained by Pawlowski is not applicable here. Appendix C explains the competition rules used by national competitions and the tournament type of rules used by the UEFA Champions League and the UEFA Europe League.

Therefore, another way of measuring competitive balance was required. Winning percentage is, in general, the most widely used indicator in studies of competitive balance. Szymanski (2001) argues that a reliable indicator of success in English football is the winning percentage. As these data [winning percentage] is available for national and international matches, even though, there is no European major league (yet), I have used winning

percentages to calculate the competitive balance for national and international competitions. As is stated by Szymanski (2001), by looking at the winning percentages per match, the competitive balance of teams from different leagues or divisions could be examined, which is not possible by making use of types of measurement in which the points scored in a season and the team’s league position at the end of a season is required. In the international case, the competitive balance had to be calculated for teams from different national leagues (England, France, Germany, Italy, Russia, and Spain), which ask for the method to calculate the

competitive balance by subtracting the winning percentages within a particular match.

3.2 Soccer Attendance

The procedure that is conducted for this study consists of looking at average

attendance and the attendance capacity for each team that is eligible for the European major league(s), in case of the international matches (Appendix A) and for each team that is in the top six European national competitions, in case of the national matches. These teams that are eligible for the European major league are the first 18 teams according to the UEFA ranking list for individual club teams, as these teams are most likely to be the best club teams of Europe. The teams that are in the top six European national competitions, according to the UEFA rankings for club competitions2, are the teams coming from the English Premier League, the French Ligue 1, the German Bundesliga, the Italian Serie A, the Portuguese Primeira Liga, and the Spanish Primera Division. The UEFA ranking list for individual club teams is visible in Appendix B. The statistics, such as soccer attendance of every football match played in all European competitions and European tournaments (UEFA Champions League and UEFA Europe League), are derived from the website www.soccerway.com.

3.3 Competitive Balance

Next, after finding the necessary data about the clubs and matches, the competitive balance was calculated. The competitive balance is measured by subtracting the away team win percentage from the home win percentage of a particular match. The win percentages are calculated by making use of the betting odds of a particular match derived from

www.oddsportal.com. These odds are the result of a rating system. The website:

www.football-data.co.uk has conducted a study in 2003, by which they used matches of the English Premier League and the divisions 1,2, and 3 for seasons 1993/94 to 2000/01 as data. They derived optimal equations, to calculate the home, away, and draw betting odds for every particular match, from this data. The home betting odds equation they found was: (1) 𝑦 = 1.56𝑥 + 46.47, where y = the probability of a home win and x = the match rating. The away betting odds equation they found was: (2) 𝑦 = 0.03𝑥2− 1.27𝑥 + 23.65, where y = the

17

probability of an away win and x = the match rating. They derived their match ratings by making use of the Power Ratings/Rateform, described by Bill Hunter in his book Football Fortunes. These match ratings could be filled in, in the equations, by which the probability of a home win and an away win is calculated, by making use of equation (1) and (2),

respectively. The next step is; dividing 100, by the probability of a home win (y), which resulted from the optimal equation. Whereby, the “Fair home odds”, that are presented for every particular match at www.oddsportal.com, is calculated.

"𝐹𝑎𝑖𝑟 ℎ𝑜𝑚𝑒 𝑜𝑑𝑑" =100 𝑦_ℎ

Where 𝑦ℎ = the probability of a home win. The same calculations apply to the “Fair away odds”, that are presented at www.oddsportal.com.

"𝐹𝑎𝑖𝑟 𝑎𝑤𝑎𝑦 𝑜𝑑𝑑" =100 𝑦_𝑎 Where 𝑦_𝑎 = the probability of an away win.

The “Fair home/away odds” are available at the website www.oddsportal.com. However, in order to calculate the competitive advantage of a particular match, the winning percentages for the home team and away team are needed. As the “Fair home/away odds” are known, and the 𝑦_ℎ and 𝑦_𝑎 are unknown, and are the numbers we need, in order to calculate competitive balance, the “Fair home odd” and “Fair away odd” formulas need to be

rephrased. Resulting in:

𝑦_ℎ = 100

"𝐹𝑎𝑖𝑟 ℎ𝑜𝑚𝑒 𝑜𝑑𝑑"

𝑦_𝑎 = 100

"𝐹𝑎𝑖𝑟 𝑎𝑤𝑎𝑦 𝑜𝑑𝑑"

In order to calculate the competitive balance of a particular match, the winning percentage of the away team should be subtracted from the winning percentage of the home team. By subtracting the win percentages, the closer the result is to zero, the more

competitively balanced the game is.

Such a way of measurement is relatively simple however, due to the fact that a European major league does not exist (yet), it is very hard or actually impossible to use any other type of measurement. By making use of winning percentages, in calculating the

18

3.4 How to Test the Hypotheses

3.4.1 Hypothesis 1

Actually, this measurement consists of four different groups of matches for which I calculated the competitive balance. First, I calculated the competitive balance of all the international matches played between two European top teams, coming from the top 18 European club teams, according to the UEFA ranking list, over the past five years (Group 1). The resulted competitive balance of all those international matches will be related to the soccer attendance of their specific match to see whether there is a relationship between competitive balance and soccer attendance and what kind of relationship for Group 1. The soccer attendance should be calculated in percentage of stadium filled as all the stadiums differ in size and maximum number of spectators. So actually, I examined whether there is a relationship between the percentage of the stadium that is filled and the competitive balance of all the international matches between two top clubs. Subsequently, I calculated the competitive balance of all the national matches played by these European top clubs in their national competition (Group 2), and examined whether and what kind of relationship there is between competitive balance and soccer attendance in Group 2. Thereafter, I compared the relationship of Group 1 with the relationship of Group 2 in order to test hypothesis 1. In order to test hypothesis 1, the p-value of the comparison between regression line 1 and regression line 2 had to be calculated. In order to calculate the p-value of the comparison of the two regression lines, some specific numbers of both regression lines are needed and had to be calculated. The number of observations (n), the slope of the regression line (b), the standard error (𝑆_𝑦∙𝑥), and the standard deviation (𝑆_𝑦∙𝑥), had to be calculated by excel, for both

regression line. By using these specific numbers, the standard deviation of b (𝑆_𝑏), of both regression lines, could be calculated. The formula for the standard deviation of b (𝑆_𝑏) is:

𝑆_𝑏 = 𝑆𝑦·𝑥 𝑆𝑥√(𝑛 − 1)

By now, the numbers to calculate the p-value of the regression lines together, are available, whereby the p-value could be calculated. The p-value needs, the degrees of freedom (df) of the regression lines together and the t-statistic (t) of the regression lines together. However, the t-statisitc needs the standard deviation of the slopes together (𝑆_{𝑏1−𝑏2}). Therefore, three formulas are needed, in order to let excel calculate the p-value of both regression lines together. These formulas are:

Standard deviation of two slopes together  𝑆𝑏1−𝑏2 = √(𝑆𝑏12+ 𝑆𝑏22) = 0,063334

t-statistic  𝑡 =(𝑏1−𝑏2)

𝑆𝑏1−𝑏2 = 1,470681

Degrees of freedom  𝑑𝑓 = (𝑛₁+ 𝑛₂− 4) = 3223

19 3.4.2 Hypothesis 2

Then, I calculated the competitive balance of all the matches played in the top six European competitions (England, France, Germany, Italy, Portugal, and Spain), over the past five years, including the top teams that are eligible for such an European super league (Group 3), and examined whether and what kind of relation there is between the competitive balance of all those matches and soccer attendance. Thereafter, I calculated the competitive balance of the national matches played in those six competitions, without the matches played by the 18 top teams (Group 4), and examined whether and what kind of relationship there is between competitive balance and soccer attendance in Group 4. The reason why I examined the top six European competitions is, because the top 18 European club teams are all coming from those competitions, except two teams: Zenit (Russia: 7th competition of Europe) and Basel

(Switzerland: 12th _{competition of Europe). Finally, I compared the competitive balance-soccer} attendance relationship of Group 3 with the competitive balance-soccer attendance

relationship of Group 4 in order to test hypothesis 2. The same method applied to test Hypothesis 1, is applied in order to test hypothesis 2. This, eventually, results in the p-value of the comparison of regression line 3 and regression line 4, by which this number should be interpreted in order to adopt or reject Hypothesis 2.

4. Analysis

This fourth paragraph “Analysis” describes the examination and findings of the analysis about the effect that competitive balance has on soccer attendance in national and international context. First, the effect that the competitive balance of all the international matches played between two European top teams, coming from the top 18 European club teams, according to the UEFA ranking list, has on the soccer attendance of these matches is described, which is categorized as Group 1 in the Methodology part. Subsequently, the effect that the competitive balance of all the national matches played by these European top clubs in their national competition, has on the soccer attendance of these matches is described, which is categorized as Group 2 in the Methodology part. Then, the relationships of Group 1 and Group 2 are compared in order to test Hypothesis 1; whether European top clubs’ attendance rates are significantly higher when acting in an European division.

Thereafter, the effect that the competitive balance has on soccer attendance in the third group, which is all the matches played in the top six European competitions (England, France, Germany, Italy, Portugal, and Spain), including the top teams that are eligible for such an European super league, is described. Subsequently, the effect that the competitive balance of the national matches played in those six competitions without the matches played by the 18 top teams, has on the soccer attendance of these matches is described, which is categorized as Group 3 in the Methodology part. And finally, the relationships of the third and fourth group are compared in order to test Hypothesis 2; whether the European club teams that stay in their national competition, after the introduction of an European super league, are confronted with significantly higher attendance rates.

4.1 Analysis of Hypothesis 1

4.1.1 Group 1

20

the top 18 best European club teams. However, a match between Juventus and Manchester United, for example, is not included in the Group 1 sample, as not both clubs are in the top 18 best European club teams. Juventus is, but Manchester United is not. This means that only the UEFA Champions League matches and UEFA Europe League matches that include both club teams present in the top 18 European club teams according to the UEFA are included in the Group 1 sample. The sample is about a time period of the last five years, which means that all those matches played, between two club teams coming from the top 18, in the seasons

2010/2011, 2011/2012, 2012/2013, 2013/2014, and 2014/2015 are in the sample. This means that the Group 1 sample consists of 183 international matches. For all these matches, the percentage of stadium filled and the competitive balance is calculated and presented in Figure 1; presented below.

Figure 1: Group 1: International matches played

As you can see, most of these matches were well attended, as most of the stadiums were filled with spectators over the 80%. This is also visible in the linear regression line which is: 𝑦 = 0,094𝑥 + 0,833, whereby y = percentage of stadium filled and x = competitive balance. Most matches are already attended by 83,26%. However, the 𝑅2 _{= 0,013, which} means that this linear regression line only explains 1,25% of the matches, which is a pretty low percentage. Besides that, the most unexpected notification is that number for the independent variable x (competitive balance) is positive: 0,094. The interpretation of this relationship is actually in contrast to the interpretation of a ‘normal’ relationship, as the higher the number of competitive balance is, the lower the competitive balance actually is. In other words, the competitive balance is the highest where the number of competitive balance is 0, and the lowest where the competitive balance is 1. The competitive balance, could never exceed 1, as the competitive balance is calculated by subtracting the win percentages of two teams in a particular match, and a win percentage could never exceed 100%. Therefore, the relationship, visible in figure 1, should be interpreted as: When the competitive balance decreases by 1, the percentage of stadium filled increases by 0,094. However, I expected an increase in spectators if the competitive balance would increase.

Besides that, the significance of F is 0,613, which is way too large (see table 1). In general, the 95% confidence interval is used to test whether there is a relation between two

y = 0,094x + 0,833 R² = 0,013 0,000% 20,000% 40,000% 60,000% 80,000% 100,000% 120,000% 0 0,05 0,1 0,15 0,2 0,25 0,3 0,35 0,4 0,45 0,5 0,55 0,6 0,65 0,7 0,75 0,8 0,85 0,9 Sta d iu m Fille d in % Competitive Balance

21

variables. In that case, the number under the Significance F should be lower than 0,05. If the number is indeed lower than 0,05, you could say with 95% certainty that there is a

relationship between competitive balance and percentage of stadium filled. In my case, you could say that there is a relationship between competitive balance and percentage filled with only 40% certainty. This certainty percentage is too low, to state that there is a relationship between competitive balance and percentage of stadium filled in case of international matches of the top 18 European club teams.

ANOVA

df SS MS F Significance F

Regression 1 0,014 0,014 0,256 0,613

Residual 181 9,846 _0,054

Total 182 9,860

Table 1: ANOVA table of group 1

4.1.2 Group 2

The second group of matches are the matches played by these top 18 European club teams in their national competition. For example, all the matches that Bayern München has played in the Bundesliga in the seasons 2010/2011, 2011/2012, 2012/2013, 2013/2014, and 2014/2015, are in the Group 2 sample. The same applies, for example, to all the matches played by Chelsea in the premier league in the last five seasons. These matches are in the Group 2 sample as well. In total, the Group 2 sample consists of 3044 matches. Actually, the group consisted of 3054 matches, however, due missing or incorrect data, I have deleted 10 matches out of the sample. The matches that are left out are:

Date Home team Away team Result Attendance Capacity

11-09-2010 Athletic de Bilbao Atlético Madrid 1-2 40000 39750 17-04-2011 SSC Napoli Udinese 0-1 65000 60240 06-11-2011 Athletic de Bilbao Barcelona 2-2 40000 39750

27-10-2012 Alania Zenit 2-3 unknown 32464

10-11-2012 Pescara Juventus 1-6 22000 20486

26-11-2012 Zenit CSKA Moskou 1-1 unknown 21838

10-12-2012 Zenit Anzhi 1-1 unknown 21838

04-05-2013 Zenit Alania 4-0 unknown 21838

10-11-2013 Spartak Moskou

Zenit 4-2 unknown 28800

16-08-2014 Zenit Ufa 1-0 unknown 21838

Table 2: Left out of the sample of Group 2

22

Figure 2: Group 2: National matches played by Top 18 European Clubs

As you can see, almost everything is possible with respect to competitive balance and stadium filled. This is visible in the R squared value as well as the number for the independent variable in the linear regression line that is part of Figure 2. The linear regression line that fits the group 2 sample is: 𝑦 = 0,001𝑥 + 0,814. The number for the independent variable is 0,001 which means that the number of spectators will increase by only 0,001 if the

competitive balance decreases by 1. In other words, there is almost no relationship between the competitive balance and percentage of stadium filled, so everything is possible. 𝑅2 = 6𝐸−07 _{which means that this linear regression line explains 0,000% of the matches, which is} even much lower than the group 1 sample. In this sample, the number for the independent variable x (competitive balance) is positive as well: 0,001, what means that, when the competitive balance decreases, the number of spectators increases. However, I expected an increase in spectators if the competitive balance would increase.

Besides that, the significance of F is 0,967, which is way too large (see table 3). The number under the Significance F should be lower than 0,05, in case of a 95% confidence interval. In my case, you could say that there is a relationship between competitive balance and percentage filled with only 4% certainty. This certainty percentage is way too low, to state that there is a relationship between competitive balance and percentage of stadium filled in case of the national matches played by the top 18 European club teams.

ANOVA

df SS MS F Significance F

Regression 1 0,000 0,000 0,002 0,967

Residual 3042 184,617 _0,061

Total 3043 184,617

Table 3: ANOVA table of group 2

y = 0,001x + 0,814 R² = 6E-07 0,000% 20,000% 40,000% 60,000% 80,000% 100,000% 120,000% 0,00 0,05 0,10 0,15 0,20 0,25 0,30 0,35 0,40 0,45 0,50 0,55 0,60 0,65 0,70 0,75 0,80 0,85 0,90 0,95 1,00 1,05 Sta d iu m Fille d in % Competitive Balance

23 4.1.3 Testing Hypothesis 1

Next after finding both regression lines, for group 1 and for group 2, I am able to compare the regression lines in order to examine if those regression lines are significantly different from each other, which causes me to conclude that European top clubs’ attendance rates are significantly higher when acting in an European division, by which hypothesis 1 could be adopted.

Regression line Group 1: 𝑦1 = 0,094𝑥1+ 0,833

Regression Line Group 2: 𝑦2 = 0,001𝑥2+ 0,814

First of all, I had to make some calculations by making use of excel due which I can calculate the p-value of the comparison between regression line of group 1 and the regression line of group 2. First I calculated the number of observations, the slope of the regression line, the standard error of the data and the standard deviation of the data of both groups by making use of excel:

Group 1 Group 2

Number of observations (n) 183 3044

Slope of the regression line (b) 0,094 0,001

Standard Error (𝑆_𝑦·𝑥) 0,172 0,189

Standard Deviation (𝑆_𝑥) 0,206 0,246

Standard Deviation of b (𝑆𝑏) 0,062 0,014

Table 4: Data about both regression lines (Group 1 and Group 2)

and I calculated the Standard Deviation of b, by making use of this formula: 𝑆𝑏 =

𝑆𝑦·𝑥 𝑆_𝑥√(𝑛 − 1)

By making use of these numbers I am able to calculate the standard deviation of the slopes together:

𝑆_{𝑏1−𝑏2} = √(𝑆_𝑏12+ 𝑆_𝑏22) = 0,063

By making use of this standard deviation I am able to calculate the t-statistic which is a ratio of the departure of an estimated parameter from its notional value and its standard error.

𝑡 =(𝑏1− 𝑏2)

𝑆_{𝑏1−𝑏2} = 1,471

The only thing left to calculate, before I am able to calculate the p-value is the degrees of freedom. The formula for calculating the degrees of freedom is:

24

Now, I have all the number to calculate the p-value of those two regression lines together. The p-value is calculated by Excel.

𝑝 − 𝑣𝑎𝑙𝑢𝑒 = 0,141

This p-value is higher as the general accepted alpha of 0,05. As I explained earlier, in some cases an alpha of 0,1 is accepted, however, my p-value is even higher than an alpha of 0,1 which makes me to conclude that Hypothesis 1: European top clubs’ attendance rates are

significantly higher when acting in an European division have to be rejected. Therefore, I

could state that the attendance rates of European top clubs are not significantly higher when they would play their matches in an European division compared to, when those top clubs play their matches in their national competition as they are doing now.

4.2 Analysis of Hypothesis 2

4.2.1 Group 3

Hypothesis 2 consists of two groups of matches that had to be analyzed in order to test the hypothesis. The third group of matches are all the matches played in the six biggest European national competitions (England, France, Germany, Italy, Portugal, and Spain) over the last five years. This means that all matches played within the Premier League, Ligue 1, Bundesliga, Serie A, Primeira Ligue, and Primera Division in the seasons 2010/2011, 2011/2012, 2012/2013, 2013/2014, and 2014/2015, are in the Group 3 sample. In total, the Group 3 sample consists of 10375 matches. Actually, the group consisted of 10396 matches, however, due missing or incorrect data, I have deleted 21 matches out of the sample. The matches that are left out are:

Date Home team Away team Result Attendance Capacity

11-09-2010 Athletic de Bilbao Atlético Madrid 1-2 40000 39750 06-01-2011 Cagliari AC Milan 0-1 23000 21000 17-04-2011 SSC Napoli Udinese 1-2 65000 60240 23-04-2011 Athletic de Bilbao Real Sociedad 2-1 40000 39750 08-05-2011 Portimonense Maritimo 1-0 13836 9544 06-11-2011 Athletic de Bilbao Barcelona 2-2 40000 39750 31-03-2012 Catania AC Milan 1-1 20253 20104

02-05-2012 Genoa Cagliari 2-1 unknown 36703

12-05-2012 União Leiria Nacional Madeira

2-3 489 8378

13-05-2012 Genoa Palermo 2-0 unknown 36703

26-08-2012 Pescara Inter Milan 0-3 22000 20486

16-09-2012 Pescara Sampdoria 2-3 22000 20486

23-09-2012 Cagliari AS Roma 0-3 unknown 16200

10-11-2012 Pescara Juventus 1-6 22000 20486

12-12-2012 Bastia Marseille 1-2 unknown 16480

03-03-2013 Pescara Udinese 0-1 23000 20486

30-03-2013 Cagliari Fiorentina 2-1 unknown 16200

12-05-2013 Parma Bologna 0-2 22000 21473

25 13-03-2015 Nice En Avant Guingamp 1-2 unknown 35624 21-03-2015 Bastia En Avant Guingamp 0-0 unknown 17363

Table 5: Left out of the sample of Group 3

As you can see, most of the matches that are left out of the sample have an unknown number of attendance, and some of the matches have a percentage of the stadium filled over the 100%, which is incorrect. At first sight, nothing looks odd or wrong to the match: 12-05-2012: União Leiria – Nacional Madeira 2-3

However, the betting odds were missing for this particular match. These are the reasons why those matches are left out of the sample. For the remaining 10375 matches, the percentage of stadium filled and the competitive balance is calculated and presented in Figure 3; presented below.

Figure 3: Group 3: Top 6 European competition matches

As is visible here, almost the whole figure is covered in blue, which means that lots is possible with respect to competitive balance and percentage of stadium filled. However, here is a better regression line as the slope of the figure, which is the number for the independent variable, is steeper. Namely, the regression line is: 𝑦 = 0,190𝑥 + 0,634. This means that, when the competitive balance decreases, the number of spectators increases, which is exactly the opposite from what I would expect in case of the relation between competitive balance and percentage of stadium filled. The 𝑅2 _{= 0,028, which means that the regression line that} fits this sample the best, explains 2,8% of the matches.

The significance of F, for Group 3, is only 2,505E-66, which is zero (see table 5). This number is way smaller than 0,05, therefore we could say, with 95% certainty, that there is a relationship between competitive balance and percentage of stadium filled,, in case of the

y = 0,190x + 0,634 R² = 0,028 0,000% 20,000% 40,000% 60,000% 80,000% 100,000% 120,000% 0,00 0,05 0,10 0,15 0,20 0,25 0,30 0,35 0,40 0,45 0,50 0,55 0,60 0,65 0,70 0,75 0,80 0,85 0,90 0,95 1,00 1,05 Sta d iu m Fille d in % Competitive Balance

26

national matches played in the top 6 national European competitions over the last five seasons (2010/2011, 2011/2012, 2012/2013, 2013/2014, and 2014/2015). ANOVA df SS MS F Significance F Regression 1 14,493 14,493 300,233 2,505E-66 Residual 10373 500,720 _0,048 Total 10374 515,213

Table 6: ANOVA table of group 3

4.2.2 Group 4

The fourth group of matches are all the matches played in the six biggest European national competitions (England, France, Germany, Italy, Portugal, and Spain) over the last five years, without the national matches played by the top 18 club teams according to UEFA. This means that all matches played within the Premier League, Ligue 1, Bundesliga, Serie A, Primeira Ligue, and Primera Division in the seasons 2010/2011, 2011/2012, 2012/2013, 2013/2014, and 2014/2015, without the national matches played by, for example, Chelsea (and the other 17 club teams), are in the Group 4 sample. In total, the Group 4 sample consists of 7660 matches. Actually, the group consisted of 7677 matches, however, due missing or incorrect data, I have deleted 21 matches out of the sample. The matches that are left out are:

Date Home team Away team Result Attendance Capacity

06-01-2011 Cagliari AC Milan 0-1 23000 21000 23-04-2011 Athletic de Bilbao Real Sociedad 2-1 40000 39750 08-05-2011 Portimonense Maritimo 1-0 13836 9544 31-03-2012 Catania AC Milan 1-1 20253 20104

02-05-2012 Genoa Cagliari 2-1 unknown 36703

12-05-2012 União Leiria Nacional Madeira

2-3 489 8378

13-05-2012 Genoa Palermo 2-0 unknown 36703

26-08-2012 Pescara Inter Milan 0-3 22000 20486

16-09-2012 Pescara Sampdoria 2-3 22000 20486

23-09-2012 Cagliari AS Roma 0-3 unknown 16200

12-12-2012 Bastia Marseille 1-2 unknown 16480

03-03-2013 Pescara Udinese 0-1 23000 20486

30-03-2013 Cagliari Fiorentina 2-1 unknown 16200

12-05-2013 Parma Bologna 0-2 22000 21473

22-09-2013 Sassuolo Inter Milan 0-7 22001 21584

13-03-2015 Nice En Avant Guingamp 1-2 unknown 35624 21-03-2015 Bastia En Avant Guingamp 0-0 unknown 17363

Table 7: Left out of the sample of Group 4

27

remaining 7660 matches, the percentage of stadium filled and the competitive balance is calculated and presented in Figure 4; presented below.

Figure 4: Group 4: Top 6 European competition matches without Top 18 European Clubs

As is visible here as well, almost the whole figure is covered in blue, which, again, means that lots is possible with respect to competitive balance and percentage of stadium filled. However, here is a good regression line as the slope of the figure, which is the number for the independent variable, is slightly steep. Namely, the regression line is: 𝑦 = 0,095𝑥 + 0,624. This means that, when the competitive balance decreases by 1, the number of

spectators increases by 0,095, which is exactly the opposite from what I would expect in case of the relation between competitive balance and percentage of stadium filled. The 𝑅2 = 0,004, which means that the regression line that fits this sample the best explains 0,4% of the matches.

In case of group 4, the number under Significance F is only 6,501E-09, which is close to zero (see table 7). This number is way lower than 0,05, therefore we could state, with 95% certainty, that there is a relationship between competitive balance and percentage of stadium filled in case of the national matches played in the top 6 national European competitions over the last five seasons (2010/2011, 2011/2012, 2012/2013, 2013/2014, and 2014/2015), without those national matches played by the top 18 European club teams.

ANOVA

df SS MS F Significance F

Regression 1 1,082 1,082 33,755 6,501E-09

Residual 7658 245,483 _0,032

Total 7659 246,565

Table 8: ANOVA table of group 4

y = 0,095x + 0,624 R² = 0,004 0,000% 20,000% 40,000% 60,000% 80,000% 100,000% 120,000% 0,00 0,05 0,10 0,15 0,20 0,25 0,30 0,35 0,40 0,45 0,50 0,55 0,60 0,65 0,70 0,75 0,80 0,85 0,90 Sta d iu m Fille d in % Competitive Balance

28 4.2.3 Testing Hypothesis 2

Next after finding both regression lines, for group 3 and for group 4, I am able to compare the regression lines in order to examine if those regression lines are significantly different from each other, which causes me to conclude that European club teams that stay in their national competition, after the introduction of an European super league, are confronted with significantly higher attendance rates, by which hypothesis 2 could be adopted.

Regression line Group 3: 𝑦3 = 0,190𝑥3+ 0,634

Regression Line Group 4: 𝑦4= 0,094𝑥4+ 0,624

First of all, I had to make some calculations by making use of excel due which I can calculate the p-value of the comparison between regression line of group 1 and the regression line of group 2. First I calculated the number of observations, the slope of the regression line, the standard error of the data and the standard deviation of the data of both groups by making use of excel:

Group 3 Group 4

Number of observations (n) 10375 7660

Slope of the regression line (b) 0,189 0,094

Standard Error (𝑆_𝑦·𝑥) 0,248 0,255

Standard Deviation (𝑆_𝑥) 0,223 0,179

Standard Deviation of b (𝑆𝑏) 0,011 0,016

Table 9: Data about both regression lines (Group 3 and Group 4)

and I calculated the Standard Deviation of b, by making use of this formula: 𝑆𝑏 =

𝑆𝑦·𝑥 𝑆_𝑥√(𝑛 − 1)

By making use of these numbers I am able to calculate the standard deviation of the slopes together:

𝑆_{𝑏3−𝑏4} = √(𝑆_𝑏32+ 𝑆_𝑏42) = 0,020

By making use of this standard deviation I am able to calculate the t-statistic which is a ratio of the departure of an estimated parameter from its notional value and its standard error.

𝑡 =(𝑏3− 𝑏4)

𝑆_{𝑏3−𝑏4} = 4,848

The only thing left to calculate, before I am able to calculate the p-value is the degrees of freedom. The formula for calculating the degrees of freedom is: