Uncorrected Proof

TESG_289.fm Page 73 Wednesday, November 19, 2003 9:03 PM

TESG_289 Pages: 16

RESTRUCTURING THE BELGIAN PROFESSIONAL FOOTBALL LEAGUE: A LOCATION-ALLOCATION SOLUTION

TRUDO DEJONGHE Department of Economy, Lessius Hogeschool Antwerpen, Korte Nieuwstraat 33, 2000 Antwerpen, Belgium. E-mail: [email protected]

Received: July 2002; revised April 2003

ABSTRACT In the 1990s, and even more since the Bosman Case, professional football in Europe has been transformed from a utility maximisation to a more endogenous growth-based profit maximisation consumer-oriented service. In the post-Bosman period Belgian football has declined in the European rankings and has in a more economic-oriented climate problems in surviving. The recent transformation implies that in the long run the economic features and more specific the location of the club is becoming a very important factor. The existing heterogeneous characteristics of the different clubs made it necessary to introduce an empirically-based modified maximum covering location model. The empirical results prove that a restructuring of Belgian professional football to 14 clubs and in some cases a relocation or merger might be a solution for the contemporary problems.

Key words: Location-allocation model, Belgian professional football, consumer-oriented service, market area

The aim of the article is to introduce an eco- a priority. The conflict between the behaviour nomic geographical solution based on em- by the league as a whole and maximisation of pirical data for the economic problems of profits for the most successful clubs required professional football in Belgium. The paper that the league act as a cartel to impose restric- shows that the use of location and a location- tions on for example, number of clubs, number allocation model on an existing situation is of games, salaries paid to players, territorial useful. It can locate the economic-based and exclusivity, and revenue-sharing arrangements market-oriented demand clubs which are in the so that the economic gap between the clubs long term not profitable. The result will be the narrows. These restrictions in the labour construction of a market-oriented league with market and production is accepted by the an exogenous number of clubs that will or can anti-trust competition regulators in the United survive in the long term. States as necessary to maintain the competitive balance for the successful operation of profes- THE TRANSITION OF PROFESSIONAL sional team sports leagues. The US profit- FOOTBALLUNCORRECTEDmaximisation professional PROOF business gets massive subsidies from local governments because the The divergence between the US and the Euro- clubs regenerated multiplication benefits for pean professional team sport model was sig- local enterprises. Another significant effect is nificant until the late 1980s. In the US model the willingness of cities to build new stadiums profit maximisation for clubs and leagues was because the clubs otherwise would move to

Tijdschrift voor Economische en Sociale Geograﬁe – 2004, Vol. 95, No. 1, pp. 73–88. © 2004 by the Royal Dutch Geographical Society KNAG Published by Blackwell Publishing Ltd., 9600 Garsington Road, Oxford OX4 2DQ, UK and 350 Main Street, Malden MA 02148, USA

TESG_289.fm Page 74 Wednesday, November 19, 2003 9:03 PM

74 TRUDO DEJONGHE

another city where the offered facilities are from all over the world will migrate from poor better. The clubs have territorial exclusivity, countries, rich countries with no football tradi- the so-called league-franchises.1 The profit- tion (like the USA and Japan) or the less lucra- maximising club may in contradiction with the tive professional leagues such as Belgium and European clubs not want to maximise playing Holland to the rich European competitions of success. It will try to maximise the difference Spain, Italy, England, Germany and to a lesser between cost and revenue. If this quantity is degree France. maximised by a team of players of lower quality The growing impact of economics on profes- than the club will do so. sional football leads to a transition from utility In Europe, prior to the 1990s, the objective maximisation to a more profit-maximising of the professional football club was to provide approach. The natural tendency for the profes- entertainment by achieving playing success sional leagues is to seek exclusivity. The most while remaining solvent and not to maximise important forms of exclusivity are: profit. Sloane (1971) indicated that profit- • league exclusivity: a single monopoly league making clubs in Europe has been the exception that controls league entrance; and not the rule. Chairmen, directors and others • team exclusivity: a limited number of team in with controlling interests in the football clubs the highest division; have achieved success in other business fields. • territorial exclusivity; a territorial monopoly, Their motives to invest in sport were more a a region wherein a team has no competition desire for prestige, power or sporting enthusi- from other professional teams combined with asm and can be seen as a form of philanthropy. a franchise system; and They get the best seats and a chance to mix • restriction of the freedom of players to with people in similar positions as themselves. transfer to or choose a new club, the ‘reserve Profit seems unlikely to be a motivation factor. clause’ salary cap.3 The pursuit of playing success is the main objective and the clubs can be defined as utility- In their study on Major League Baseball, maximisation oriented subject to some finan- Markham and Teplitz (1981) found that a number cial constraints. of agreements that are violations of the antitrust The development of football in Europe in laws were accepted as necessary in a consumer- the 1990s was from an economic viewpoint oriented professional sport competition. enormous. Global professionalism and the The first and in some way the second form increasing importance of revenues from mer- of exclusivity exists in European football. There chandising, sport sponsorship, broadcasting is only one league in Belgium recognised by rights and in some cases the flotation of clubs the international bodies, UEFA and FIFA, and on the Stock Exchange resulted in a rising com- the number of teams on the highest level is mercialisation of professional football. A deci- limited but the existence of relegation can be sion by the European Court in December 1995, a distortion of the market. The system of relega- known as the ‘Bosman rule’,2 ruled that teams tion undermines territorial exclusivity. Any team and the organising leagues could no longer place with a rich investor can promote to the first restrictions on the free movement of players division. across national boundaries in the EU. The Bosman rule not only eliminated quotas on the PROFESSIONAL FOOTBALL AS AN numberUNCORRECTED of foreign players, but called into ques- ECONOMIC GEOGRAPHICAL PROOF ITEM tion the transfer rules, which was a set of rules and practices that inhibit competition between Top class sports, and in this case, professional clubs for players who seek to sign their first con- football, can be seen as a person-related tract and those whose contracts have expired. immaterial consumer-oriented service with a The result of the Bosman rule in combination short-term utility and an economic function. with the increasing diversification of revenues (Dejonghe, 2001, pp. 289–293). This definition is an increasing concentration and flow of cap- implies that only densely-populated locations ital toward the richest clubs and largest markets. combined with a rate of centrality can sustain A substantial proportion of the best players a professional club. The consequence is that a

TESG_289.fm Page 75 Wednesday, November 19, 2003 9:03 PM

RESTRUCTURING THE BELGIAN PROFESSIONAL FOOTBALL LEAGUE 75

Figure 1. The endogenous growth model (Dejonghe 2001, p. 301).

professional football club can be described in a free market. They have to transform their as a central place function with a high thresh- ‘service’ to maximise the attraction, their total old, the absolute or real number of potential economic value. The economic objective is that consumers. The result in the long run is a very total gate receipts, income from publicity, com- strong market-orientated location of profes- mercial inflow, television and other media have sional football clubs. The revenues generated to be maximised for the economic product ‘the by television, merchandising, stock market league’. The attractiveness of professional team and European competitions are rising expo- sports is the fact that the total monetary inflow nentially and are related to the market poten- depends crucially on the uncertainty of out- tials of the clubs. Bale (1988, p. 509) noticed come of the games played within the league. already that the spatial patterns of the profes- When the probability of either team winning sional football industry had to reflect those of approaches 100%, income from gate receipts or the growing tertiary orientation of the economy. media fall substantially. The most important The evolution of professional football since the difference between the product ‘sport competi- 1990s into a market-oriented service that has to tion’ and other products is that the competitive be sold to its potential consumers emphasised firms, namely the clubs, are obliged to co- this viewpoint. This transition implies that clubs operate to create a sellable product. This in countries and cities with a small population economic impact and reality has already been have a competitive disadvantage. Market orien- introduced in the United States where National tation, as a main causation in the extent of an Football League (NFL), Major League Baseball economic product, leads to a situation of cumu- (MLB), National Basketball Association (NBA) lative cycles. High attendance leads to more and National Hockey League (NHL) are organ- incomeUNCORRECTED from sponsors and media that can be ised according to thesePROOF economic principles of invested in players which in turn leads to better rationalisation and optimisation. results and higher attendance. The long-term The market of a professional football club success of professional football can be seen and its league has to be located and defined in as an endogenous demand-oriented growth the tertiary sector. The clubs can be seen as het- model based on Myrdalls’ cumulative causation erogeneous, this refers to the potential product process. differentiation between the clubs with their These tendencies imply that clubs should be individual club-specific characteristics and the run as a business with a future policy perspective necessity of a long-term budget to sustain the

TESG_289.fm Page 76 Wednesday, November 19, 2003 9:03 PM

76 TRUDO DEJONGHE

club results in the creation of a monopolistic The increasing impact of the economic competitive market structure, which is co- principles on football results in a dichotomic ordinated by one internationally recognised classification of the clubs in (inter)national and national football association. The number of local were for a survival in the long term. The clubs playing in the first league is determined economic spatial potentials have to be analysed. by the central bureaucratic association and is The central topic or the necessary minimal con- not connected to the demand in the national ditions are in this case the presence of potential market. The endogenous limit of the number fans and the willingness of local firms, industry of football clubs in the first division and the and government to invest in the club and its necessity of co-operation result in the creation infrastructure. This concept results in a per- of a producer’s cartel. The purpose of such a ception where professional football and the cartel is the striving for profit-maximisation by urban hierarchical and spatial structure are limiting mutual competition combined with interdependent. supply optimisation. According to De Clercq (1995, p. 281), the formalisation of a cartel is THE ECONOMIC CONSEQUENCES FOR common in a sector with a restricted number BELGIAN FOOTBALL COMPETITION of firms, which are creating analogue products where the total costs made in the production The increasing importance of revenues from process are common knowledge. The substitu- sponsoring, merchandising and broadcasting tion of the term firm by professional football rights, as shown in Table 1, imposes the neces- club and total costs by budget is a rather good sity of an organisation that supports the game description of a top competition. Dejonghe as an economic product that has to be sold (2001, p. 296) described the market of the pro- through marketing to the potential consumers. duct ‘professional football league’ as a coalition The consequences of this trend is the domi- in a near-cartel competition of heterogeneous nation of the above-mentioned ‘market agree- interdependent oligopolic football clubs, ments’, that cannot be guaranteed in the which is organised by the recognised football contemporary structure of football in Belgium. association. Thinking about the economic rationalisation of

Table 1. Revenues ﬁrst division clubs in Europe (2000–2001) (in %).

UEFA Country Ticketing Broadcasting Sponsorship Merchandising, ranking (%) Rights (%) catering and 2001/02 (%) stadium rent (%)

3 England 31 39 30 – 2 Italy 16 54 13 17 1 Spain 25 51 9 15 4 Germany 18 45 22 14 5 France 16 51 18 15 Average 21 48 18 12 9UNCORRECTED Portugal 42 20 18PROOF 20 7 Netherlands 34 12 41 12 12 Scotland 42 23 35 – 21 Denmark 17 4 43 36 17 Norway 40 4 43 13 22 Sweden 33 14 28 25 Average 35 13 35 18

Data for Portugal and Scotland (1999/2000), Spain (1997/98); in England revenues are from merchandising, catering and stadium rents with sponsorship entered as commercial income. Source: Deloitte & Touche (2002) and (15/12/02)

TESG_289.fm Page 77 Wednesday, November 19, 2003 9:03 PM

RESTRUCTURING THE BELGIAN PROFESSIONAL FOOTBALL LEAGUE 77

professional football has been accelerated these clubs have in a demand oriented system, through the Bosman judgement. The com- in the long term, no economic basis to exist. bination of the loss of players as assets, the The Belgian first division will through the use increasing impact of the media and economic of a location-allocation model be transformed principles in professional football and ‘tradi- into a league with an exogenous defined number tionalism’ in Belgian structures resulted in a of clubs with a higher spatial penetration. sudden spectacular decline of Belgium from third place in the UEFA ranking in 1993 to 20th COMPETITION IN SPACE place in 2000 and a recovery to a more normal 14th place in 2002. A good location provides the firm with stra- The question that can be asked: ‘Is there a tegic advantages that competition may find need for a professional football league in difficult to overcome. While other market- Belgium?’ When the answer to this question is ingmix elements may be easily changed in affirmative then an economic analysis is neces- response to a changing environment, loca- sary. Because of the abscence of transparency in tions represent long-term investment that the bookkeeping system the sources of revenue can be changed only at a considerable cost. for the Belgium competition are not known. (Ghosh & Craig 1983, p. 56) The situation today in Belgium is: The search for a location model – The Hotel- • an average attendances of 7,500–9,500 ling and Lösch maxim are examples of eco- spectators in the last few years;4 nomic geographical location forms of two • broadcasting rights of $12.9 million for last competitive suppliers in a uniform market area. season ($15 million for season 2002/03);5 Lösch’s classical locational equilibrium model • lack of international impact and the associ- assumed a market with rational consumers ated income loss from demand-oriented and homogeneous suppliers, operating in the broadcasting rights; simplest possible uniform spatial context. Lösch • total budget of the first division (18 teams) is (1962) argued that firms had to establish them- only $125 million;6,7 selves not in the location with the minimum • no barriers between professional and cost or maximum turnover but in the place amateur football; and where profit maximisation occurs. The demand • teams on the edge of bankruptcy and heavy for a product is a distance-decay function. debts. The distance where the demand is equal to 0, These facts shows us that in an economic envi- the Christaller’s range, is after rotation of the ronment, a restructuring of the Belgian profes- distance-decay function around the Q-axis, a sional football league is necessary for these circle that will be transformed into a hexagon, reasons: by the presence of spatial competition. The volume or demand cone of that market is the sum of •to enter competition with other European the potential consumers in that specific area. countries; Hotelling (1929), on the other hand analysed • the introduction of the licence system by the the location patterns of a linear duopolistic Belgian Football Association in 2000/01; and market for a homogeneous product with no in the near future (2004/05) by UEFA, where price-competition in an area with a spatially a club is not allowed to have debts to the uniform distributed population. The result was social security, taxman, players, etc.; and that the location of the two suppliers would be • to stopUNCORRECTED cases of bankruptcy that have been PROOF next to each other in the centre of the spatial arising since the introduction of the licence market otherwise the market would evolve to a system. Löschian equilibrium. The long-term survival The competition in space, in this case in Bel- of more than one supplier is only possible if gium, will be approached as a competition not the total demand is sufficient to support both between homogenous clubs but between the suppliers. different clubs with their historical and spatial Devletoglou (1965) added the principle of characteristics. The aim is to show that some of the ‘doubtful area’. He argued that the range

TESG_289.fm Page 78 Wednesday, November 19, 2003 9:03 PM

78 TRUDO DEJONGHE

of the market is not an arbitrary border, but a often assigned to the nearest facility. The main transitional zone caused by the heterogeneity objectives of location-allocation models are the in the quality of the sold products. The location- simultaneous and optimal location considering allocation models, where the location of several (Dejonghe 2001, p. 424): competitive facilities in a geographical market • the company objectives (profit maximisation, is analysed, are extensions of classical equilib- utility maximisation); rium models, and can be used to locate the first • consumer characteristics (market potential, division football clubs in an area. The models income, profile); determine the location and the spatial inter- • distance and regional accessibility (public dependence of several competitive facilities in a transport, highways, traffic congestion, parking bounded geographical market. In other words, facilities); and they detect simultaneous localisation of several • competitive environment (competitors, spatial suppliers combined with the allocation of the interaction) potential consumers. Location-allocation refers to two vital basic assumptions necessary in loca- The importance of location strategy is very tion decisions: significant. According to Daskin (1995), the success or failure of both private and public • the optimal location of the supplier or service; sector facilities depends in part on the locations • the allocation to the respective suppliers or chosen for the located facilities. services of the potential consumers with their The combination of spatial location and specific consumer behaviour and mobility optimising the demand is known as the cover- constraints. ing problem in the location-allocation models. A number of important questions, which need (Church & Revelle 1974, Church & Meadows to be solved in location models, are (Daskin 1979; Thomas 1993; Daskin 1995; Daskin & 1995, p. 3): Owen 1999; Golledge & Stimson 1997) The objective of these models is to identify the loca- • How many facilities should be planned? tions that provide potential consumers with • Where should each facility be located? coverage within a specified distance or time • How large should each facility be? constraint. One of the important characteristics • How should the demand for the facilities’ of the covering models of homogeneous serv- services be allocated to the facilities? ices is the existence of a lack of competition. The number and size of facilities to be located The purpose of the model is to find these is often a function of the service-cost trade- locations that will be successful in the long run. offs. In most cases, the quality of total service The long-term equilibrium is an interdepend- improves as the number of located facilities ence matrix of potential locations searching increases, but the marginal cost of providing for a combination that covers the market in an the service can also increase. In the private sec- optimal mode. The covering models have to tor, the decision to create or relocate an existing maximise the utility (public sector) or the profit service has to be related to the profit maximi- (private sector) of the potentials of the market sation intention and the objectives measured in and to optimise the location of the facilities in monetary units. The maximum number of ser- the economic and geographical space. vices implanted in a bounded geographical area One of the main problems of the set of can beUNCORRECTED analysed by estimating the maximum covering models is thatPROOF the number of facilities total profit for the sector. A service has to be that are necessary to cover the total number of closed down or relocated when the long-term potential consumers exceeds the number that perspectives shows that profit will never be real- can, for budgetary reasons, be built. This prob- ised/actualised or that profit in another loca- lem resulted in the creation of some relaxation tion will possibly be higher. In the public sector, in the models. This relaxation of the purposes with a non-monetary and social utility maximi- can be divided into: sation principle, the aim is to maximise coverage of the area. In these models the demand is, • relaxation of the total coverage, a maximum just like in the models of Lösch and Hotelling, covering problem used in the private sector; and

TESG_289.fm Page 79 Wednesday, November 19, 2003 9:03 PM

RESTRUCTURING THE BELGIAN PROFESSIONAL FOOTBALL LEAGUE 79

• relaxation of the coverage distance, a mini- They locate subjective quality differences max or centre problem used in the public and characteristics and make an individual sector. classification of the different clubs. This subjective approach and the introduction of time dis- Professional football is typified as a market- tance assumes that the ‘top football club’ has to oriented service, which means that the location- sell its product to the market. The opening up allocation problem can be analysed as a of an area and the attraction of other clubs leads modified maximum covering location problem to the creation of spatial competition and the of the private sector. This means that a relaxa- existence of unequal service areas. The hetero- tion or modification of the total coverage is nec- geneity of the clubs is projected in their differ- essary to ensure that every club is profitable. ences in local consumption and impact on the In the search for a solution the term ‘market’ region, the centrality, and has an influence on can be used in two different ways: the market and/or the service area. These subjective markets will be used as the basis for the • market or market area is the nominal location-allocation model. aggregation of the potential consumers; and The objective of this new empirical model is • service area is the areal extent of a market – given a fixed number of existing or newly area and refers to the closest territory sur- created heterogeneous clubs – to maximise the rounding the location. It can be expressed covering of potential consumers in Belgium in area measures and it is the region where combined with certain an absolute or relative the probability of the consumption of a threshold for the long term survival of the indi- central function located in a central place is vidual club. higher than zero. The empirical model with its heterogeneity In the case of a consumer-oriented service the relaxation is based on the 45,791 season ticket- market is the determining factor. The range of holders of the different clubs in the first divi- the service area is only an indicator of the spa- sion for the season 1997/98. These data are in tial competition and is in a densely populated their own way a reflection of the subjective area not the main factor. cultural-historical topophilia or pride related to each club. The changes caused by the promo- Professional football: a modified maximum tion and the relegation for the seasons 1998/ covering model – The relaxation in these 99, 1999/2000, 2000/01 and 2001/02 are models is that not all potential consumers have placed in the model. The total number of clubs to be covered but as many as possible in terms that have played in the first division in that of not jeopardising the economic profitability period was 23. Beside these 23 there were or cost effectiveness of the facilities. The basic some ‘theoretical’ regional clubs created in the objective of this relaxation is to introduce or other regional centres such as Namur, Leuven, consider a fixed number of facilities, which Tournai and Verviers. are to be located in service areas combined The season ticketholders are the consumers with maximising the demands covered. The through which the consumption of the service modified maximum covering used, first implies ‘top-football’ is assured, they can be considered the presence of a status quo situation with 18 as the ‘loyal’ or ‘hard-core’ supporters. The clubs. In a second stage an exogenous defined nominal number of season ticket-holders is not number, in this case 14 heterogeneous clubs,8 a reflection of the real consumption of the is locatedUNCORRECTED in a potential Hotelling maxim in product. The nominal PROOF values have to be cor- the most densely populated urban centre or rected to a relative value that indicates the real in a Lösch maxim in the other centre. This attendance. Otherwise, a club with a relatively approach leads to the results that, in contrast high number of season ticket-holders and a rel- with the deductive location-allocation models, atively minor number of floating attendances consumers from the same location do not nec- should have a larger potential market than a essarily prefer the nearest service. The con- club with not as many season ticketholders but sumer is not an ‘optimiser’ but a ‘satisfier’, with a large potential market of floating attendance. personal motives, preferences, and behaviour. The correction is based on the real number of

TESG_289.fm Page 80 Wednesday, November 19, 2003 9:03 PM

80 TRUDO DEJONGHE

= attendants who support the home teams. The Ci local consumption index in location i. c = Belgian Football Association (KBVB) provided Sij corrected number of season ticketholders of these latest data. More formally: club j in location i. = 10 Bi total population in location i. Sn W n j j ×=c n n 100 S j The club specific consumption index, Cij is: St Wt S c n = =×ij S j the nominal number of season ticket- Cij 10 , 000 holders for club j. Bi Sn = the sum of the nominal number of season t The club specific penetration index, P meas- ticket-holders for all clubs in a season. ij ures the relative orientation of a community to W n = the nominal number of home attend- j one or more clubs. More formally: ances for club j. n = c W t the sum of the nominal number of home S =×ij attendances for all clubs in one season. Pij 10 , 000 c = ∑S c S j the corrected number of season ticket- ij holders for club j. An important parameter is the local consump- The empirical determination of the markets tion index Ljj. This is the rate of consumption and service areas implied the formulation of a of professional football in the centre where the model through which the 589 Belgian commu- club is located. A high rate means a strong rel- nities could be assigned to one or more clubs. ative attachment of the inhabitants but can also To do so, a reference index is introduced to be an indicator of local saturation. This means analyse the coverage of these communities. The that a growth of local consumption is not the endogenous reference-index, Ri is the average potential solution for attracting more people. corrected season ticketholders for every 10,000 The only possible growth has to be a rise in cen- inhabitants. trality or in other words, attracting attendances More formally: from outside the centre. The local consumption index, Ljj is: S c t ×= 10 , 000 R c i S jj Bt L =× 10 , 000 jj B c = n j SSt t = c = Bt total population of the total market, in this S jj corrected number of season ticketholders of case Belgium.9 club j in location j, the location of the club. = = Ri endogenous reference index. Bj total population in location j, the location of club j. The empirical determination of the service area and market of the clubs implies the creation of The determination of the markets implies the a model that should make it possible to assign creation of a classification for the allocation of the communities to one of the 18 clubs. The the communities and their potential consumers. model is based on the community consump- This allocation is translated in an allocation tion index, Ci, the club specific consumption rate Aij based on a combination Cij and Pij in the index, UNCORRECTEDCij and the club specific penetration 589 communities in PROOFBelgium. The minimal rate index Pij of ‘professional football’ in those of 10 for Pij is used to eliminate the sporadic communities. consumer from the allocation model. The community consumption index, Ci Pij divides the total market in: measures the rate of consumption of the service professional football in the communities, and is • monopoly area: the area dominated by one calculated as follows: club;

c • monopolistic competition area: the spatial ∑Sij competition is strong but one club still has C =× 10 , 000 i more than 50 per cent market share; Bi

RESTRUCTURING THE BELGIAN PROFESSIONAL FOOTBALL LEAGUE 81

Allocation model.

Pij Cij Aij ≥ Monopoly area 80 Bi ≥ Aa Ri ≤ < Ab 2/3Ri Cij R i ≤ < Ac 1/3R i Cij 2/3R i × Monopolistic 50–80 Pij Bi competition area ≥ Ba Ri ≤ < Bb 2/3R i Cij R i ≤ < Bc 1/3R i Cij 2/3R i Rest area ≤ × Ca 25–50 1/5R i Cij Pij Bi or ≥ ≤ < 50 with 1/5R i Cij 1/3R i 1/4Bi Rest area ≤ × Cb 10–25 1/10R i Cij Pij Bi or ≥ ≤ < 25 with 1/10R i Cij 1/5R i 1/10Bi ≤ Non covered area Cij 1/10R i 0

• rest area: the area with a low rate of number sum of Aij. The non-covered population is there- of potential attendance for a club; and fore the sum of the population of the non- • non-covered area. covered areas and the relative parts of the Bi that are not assigned to the monopolistic competi- The level of Cij results, as shown in Table 2, in tion and rest areas. a further subdivision in an a, b or c. This is an The results for the 23 heterogeneous clubs in indication for the strength of the penetration. the ﬁrst division for the 1997/98 to 2001/02 The subdivision b or c indicates that a ‘new seasons are shown in Table 2. The combination club’ has the possibility of spatial competition of spatial competition, cultural (language) bar- or that the existing club(s) still has the potential riers, country-province borders, time-distance, to gain attendance from this area if they inertia and tradition can be put forward as an improve the quality of their services. The Ab explanation for the contemporary structure and Ac areas are non-saturated monopolistic and impact of the clubs. This empirical out- areas and indicate the direction of growth come shows that the total coverage equals potential for a club. 7,364,354 or 72.5 per cent of the Belgian The monopolistic competition area is the population and 37.4 per cent of the population potential spatial growth zone. These areas can, belongs to the monopolistic area of a club. depending on the quality of the services, be trans- An important parameter is the local con- formed in a monopoly or rest/non covered sumption index Ljj. A high rate means a strong area. TheseUNCORRECTED ﬁndings are summarised in the relative attachment PROOF of the inhabitants but can model. also be an indicator of local saturation. The only possible growth has to be a rise in centrality, this THE RESULTS means attracting attendances from outside the centre. The model locates the communities in one or The local consumption index, Ljj is: several areas of the clubs and it indicates the c non-covered areas in Belgium. The potential S jj L =× 10 , 000 consumers for each area and each club is the jj B j

82 TRUDO DEJONGHE 0 00 0 0 0 4.972 0 0 0 0 74.919 57.410 4.972 229.408 216.443 0 12.965 17.509 88.492 38.052 50.440 76.197 75.877 76.197 90.497 44.533 76.809 90.497 75.639 31.344 76.809 0 75.639 38.067 58.245 0 29.644 17.379 0 0 58.245 0 36.123 0 17.379 89.594 34.884 51.350 36.123 0 26.418 0 34.884 0 33.555 0 30.717 26.418 0 0 0 17.795 8.423 86.412 18.929 0 52.138 9.363 52.919 0 53.961 7.547 0 44.171 0 11.474 80.231 29.643 0 0 13.742 0 4.241 14.282 22.019 13.781 58.296 54.730 0 13.742 0 57.757 10.036 10.734 0 54.730 9.954 33.419 7.951 9.057 7.967 39.495 34.043 23.105 24.877 0 5.830 9.057 42.665 18.262 0 0 30.379 0 21.062 0 10.734 3.664 0 0 23.105 0 21.603 0 34.471 26.029 9.954 14.302 0 0 21.442 47.827 20.169 38.781 18.947 23.600 4.587 28.345 17.939 24.227 10.436 1.008 517.629 872.705 197.272 386.535 34.477 365.511 75.062 207.748 252.919 102.424 245.295 127.547 212.194 252.919 735.804 368.831 112.351 212.194 51.240 0 27565 119.928 56.819 42.338 198.769 274.052 231.810 111.643 55.553 0 79.384 8.930 131.791 91.459 0 256.259 119.424 44.989 94.448 463.595 0 119.424 62.718 0 74.396 240.034 0 0 70.013 5.571 0 182.770 223.561 104298 8.285 18.635 0 0 96.512 97.194 21.896 180.125 5.057 85.972 30.723 37.679 0 70.164 85.576 81.638 33.619 31.995 14.645 0 12.238 16.846 0 98.487 34.405 44.185 0 49685 39.610 35.307 26.616 95.142 18.522 0 30.812 56.922 0 79.649 7.789 9.902 18.873 35.490 5.057 15.093 11.186 21.432 11.968 0 40.060 207327 6.262 25.066 203.864 5.706 14.994 3.463 UNCORRECTED PROOF3.797.141 98.138 93.904 82.722 68.470 Total A Aa Ab Ac B Ba Bb Bc C Ca Cb 768.074 462.634 406.902 403.848 319.678 262.822 248.835 234.380 199.094 185.843 179.252 178.971 173.955 170.442 163.411 126.071 106.458 1.350.055 1.080.395 7.364.354 The A, B and C zones of the 23 football clubs. Table 2. Table Standard Luik FC Brugge SC Charleroi AA Gent RC Genk GBAntwerpen SK Lierse Ex Moeskroen FC Antwerp E Aalst SK Beveren VV St Truiden La Louvière KV Kortrijk RWDMolenbeek KV Oostende V Geel SC Lokeren VC Westerlo RC Harelbeke SC Anderlecht KV Mechelen SK Lommel

RESTRUCTURING THE BELGIAN PROFESSIONAL FOOTBALL LEAGUE 83

Table 3. The local consumption indexes. The combination of the market and Ljj make it possible to divide the heterogeneous clubs in Club Ljj Club Ljj some more homogeneous subgroups. Saey and Lietaer (1980) determined the SC Anderlecht 33.37 SK Beveren 157.41 theoretical range of a regional service area or St Liège 20.15 St Truiden VV 261.80 the accessibility class was set to 15 km. Accord- FC Brugge 74.32 AA La Louvière 133.54 SC Charleroi 32.26 KV Kortrijk 136.60 ing to the ‘traffic principle’ 15 km was the AA Gent 49.04 RWDMolenbeek 8.92 regional impact zone of a regional centre. A RC Genk 208.70 SK Lommel 412.60 football club playing in the first division can GB Antwerpen* 17.79 KV Oostende 103.91 be seen as a hierarchical marginal good that is, SK Lierse 176.31 V Geel 271.69 because of the exogenous limit of 18 clubs, an Exc Moeskroen 214.20 SC Lokeren 203.47 indicator of an urban hierarchical level.11 The FC Antwerp 16.51 VC Westerlo 395.31 combination of the number of first division KV Mechelen 156.20 RC Harelbeke 234.59 clubs, the Belgian urban hierarchical centre and E Aalst 168.80 a professional football club as a hierarchical marginal good lead us to a theoretical range of * The value for GB Antwerpen was calculated on the data from G Ekeren, the first division club that a regional club of 15 km. merged with the third division club Beerschot • The national clubs are monoliths and an Antwerpen obstacle for the extension of other clubs in the nearby region. Their real range extends in a large degree the threshold of

Table 4. The more homogeneous subgroups.

Group Club Location

Major clubs in major or regional cities with a low Ljj SC Anderlecht Brussel FC Brugge Brugge Standard Liège Liège

Regional clubs in major cities with a low Ljj AA Gent Gent SC Charleroi Charleroi

Regional clubs in regional cities with a high Ljj RC Genk Genk AA La Louvière La Louvière KV Mechelen Mechelen E Aalst Aalst KV Oostende KV Oostende KV Kortrijk KV Kortrijk

Smaller clubs in smaller cities with a high Ljj STVV Sint-Truiden Exc Moeskroen Moeskroen SK Lierse Lier SC Lokeren Lokeren UNCORRECTEDSK Beveren PROOFBeveren SK Lommel Lommel V Geel V Geel

Local clubs in communities with a high Ljj RC Harelbeke Harelbeke VC Westerlo Westerlo

Local club in major cities with very low Ljj GBAntwerpen Antwerpen FC Antwerp Antwerpen RWDMolenbeek Brussel

84 TRUDO DEJONGHE

‘professional football’ in Belgium. The the attendances are at their maximum. The service area exceeds the theoretical 15 km. average attendances reach a maximum of They dominate their region and determine 8,000–10,000 in very successful periods endogenously their spatial market zone. FC but drop to 2,000–3,000 in less successful Brugge is located in a regional city and can periods. be seen as a functional substitution of Ghent • The local clubs in major cities have a very low who has only a regional impact. These clubs Ljj and probably an enormous growth have a low Ljj so that even today a potential potential. RWD Molenbeek is located in endogenous growth of attendances is Brussels in a theoretical Hottelling maxim possible. with SC Anderlecht. This monolith tran- • The result for the ‘regional’ clubs in regional sformed Brussels into a Lösch maxim and a and major cities shows a spatial monopolistic second club is thereby in Brussels no option. Lösch maxim for SC Charleroi and RC Genk. The Antwerp situation shows a Hotelling The situation in E Aalst shows the existence maxim of two clubs that are not using the of spatial competition with AA Gent and economic potentials of the city. SC Anderlecht/Brussel. KV Mechelen has The transformation of professional football to a spatial competition with Lier and the a market/consumer-oriented service implies monolith Anderlecht and to a lesser degree the presence of a market that reaches a certain with GB Antwerp, FC Antwerp and VC ‘threshold’. Thirteen of the 23 Belgian clubs Westerlo. KV Oostende has strong spatial have a market of less than 200,000 potential competition for its service area and market consumers, a number that on the long term is with FC Brugge. The club relegated after not enough to support a professional football one year (1998/99) back to the second and club that can compete with Belgium’s top clubs plays today in the third division. KV Kortrijk or with the top European clubs. The empirical relegated also after one year (1998/99) and verification of the markets shows us that the also now in the third division. AA La Louvière allocation-location solution can be applied to promoted after an absence of 21 years to the existing historical-based situation. the first division in 2000 and has a minor service area. AA Gent and SC Charleroi have THE EMPIRICAL ALLOCATION- a low L so that the attendances can rise jj LOCATION OF 14 CLUBS substantially in periods when the clubs are successful. The transformation to 14 clubs – The problem • The ‘small’ clubs in the small cities have, as with professional football in Belgium is not only seen in the values of their B and C zones, a a location-allocation problem, but for interna- limited centrality dependent for a great part tional competition it is also a matter of too many on local consumption. The lack of centrality clubs in a limited market and high international is combined with the psychological barrier thresholds in the most successful football coun- consisting of a refusal to descend in the urban tries. In comparison with the dominant countries hierarchy. To illustrate, the population of the in the Champions League and the UEFA Cup, regional centre Sint-Niklaas is not prepared the average market in Belgium of 550,000 for to consume the service ‘professional foot- every first division club is a lot smaller than ball’ offered by nearby small cities Beveren the averages of the most successful countries. and LokerenUNCORRECTED (Dejonghe 2001, pp. 442–448). In Germany there isPROOF one first division club for Harelbeke is part of the Kortrijk and Westerlo every 4.6 million inhabitants, in France and agglomeration and in the classification of Italy one for every 3 million, England one for Van Der Haegen et al. (1982) a ‘main village’. every 2.45 million, Spain one for every 2 million The club has a very high relative con- and even in The Netherlands only one for every sumption in the local community, but lacks 830,000 inhabitants. centrality and a refusal to descend in the The restructuring of the first division to 14 urban hierarchy. The combination of these clubs implies the elimination, relocation or characteristics results in a small number of merging of several clubs and the creation of potential consumers and indicates that totally new clubs in selected locations. The

RESTRUCTURING THE BELGIAN PROFESSIONAL FOOTBALL LEAGUE 85

Table 5. The theoretical ﬁrst division with 14.

Potential Monopolistic Club Merger Location consumers Area

SC Anderlecht – Brussels 1,619,491 743,926 Standard Liège – Liège 1,064,743 874,955 FC Brugge – Brugge 844,542 575,828 Metropool Antwerpen FC Antwerp – GB Antwerpen Antwerp 634,864 453,030 SC Charleroi – Charleroi 469,110 361,002 Lier-Mechelen Lierse – Mechelen Intercept 459,844 277,033 FC Borinage-Centre La Louvière – Mons12 Intercept 446,577 413,024 RC Genk – Genk 444,840 269,813 AA Gent – Gent 406,985 252,919 FC Southwest Harelbeke-Kortrijk-Waregem Intercept 342,892 152,014 FC Waasland Beveren – Lokeren Sint-Niklaas 310,840 185,222 FC Kempen Westerlo – Geel – Turnhout Intercept 296,174 201,033 FC Haspengouw Sint-Truiden Sint-Truiden 291,743 169,454 Exc Moeskroen Moeskroen 288,916 255,916 Total 7,921,561 5,185,169

UNCORRECTED PROOF Figure 2. The location of the 23 teams.

method that was used to create ‘new’ clubs in petition of the existing clubs. The spatial mar- certain regional centres was to give the poten- kets and the number of potential consumers as tial club the characteristics of the different a result of a theoretical merger of two or more ‘regional clubs in regional cities’. The result was clubs was determined by and deducted from the a minimum and a maximum potential market data of the individual clubs. The result is shown taking into account the potential spatial com- in Table 5 and Figure 2.

86 TRUDO DEJONGHE

Figure 3. The theoretically rationalised ﬁrst division with 14 clubs.

Table 5 shows that despite a reduction by that is larger that those of the regional 4 clubs: cities Oostende (361 km2), Aalst (296 km2), Kortrijk (284 km2), St Niklaas (258 km2) and • the coverage increased an increase to 77.9 per Mechelen (238 km2). The club is located in cent of the total population; a monopolistic Lösch maxim with a high • an expansion of the monopolistic area to centrality in the 15 km zone. The club can be 51.0 per cent of the total population; seen as the historical regional relocation or • an individual consumer potential that rose to functional substitution of Leuven. at least 288,916 combined with a high Ljj for the three clubs with the smallest market; The academic study started with the data of • the elimination of E Aalst, a club located in 1997/98 and the ﬁrst results were published in a regional city on the E40, a corridor for the 2001 and created a storm of protest in the foot- monoliths Anderlecht and Brugge; ball world. Now, we can confront the results of • the potential new location in the regional the research with the existing situation. RWD city Leuven (between Brussels and Liège) has Molenbeek and RC Harelbeke, two clubs with- a monopolisticUNCORRECTED zone of 123,748 and a B and out any future in thePROOF model, went bankrupt C zone of 59,686–95,498 was not selected. and were expelled from the football league. Other potential locations such as the regional KV Kortrijk, according the model had to merge city and capital of Walloon Namur (255,764) with Harelbeke and the second division club and the regional city Verviers (205,524) were Waregem, and E Aalst, a club that on empiri- not selected; cally based data had no future in the ﬁrst • Sint-Truiden is a well-equipped small town division, were bankrupt and relegated to the of level 3 with a strong polarisation in the third division. KV Mechelen, according the urban hierarchy. Van Hecke (1997, p. 106) model had to merge with Lierse, and SK determined a service area of 425 km2, a zone Lommel are bankrupt and will in 2003/04 be

RESTRUCTURING THE BELGIAN PROFESSIONAL FOOTBALL LEAGUE 87

expelled from the competition or placed in The problem with the Belgian football com- the third division. Clubs such as SK Lierse, petition is that in the long run, and as reality GB Antwerpen, VC Westerlo, La Louvière, SC shows us in the short term, the contemporary Charleroi and the at the moment second divi- situation will lead to market failures and elimi- sion club V. Geel have because of their debts to nation of some of clubs. The solution proposed Social Security, taxes and others, problems with in this paper is the creation of a new profes- their licence for next season. The potential sional competition with fewer teams combined revenues from transferring players dried up with a territorial exclusivity with a centre with a because of the implosion of the Belgian and potential number of consumers that reach a European football market. This result shows us certain absolute or relative threshold. The that a theoretical model based on empirical empirical allocation-location model combined data can be of some use to the real world but the maximising of the total coverage and the in the short term can subjective arguments be individual market of the clubs and accounts for used to postpone the inevitable. the existing situation. The result is the creation of a first division with 14 clubs combined with CONCLUSION a larger coverage of the population in Belgium. The 14 allocated and located clubs have a The divergence between the American profit greater coverage and monopolistic area than maximisation leagues and the historically based the 23 existing clubs. The transformation, utility maximisation in Europe was accelerated based on empirical data, implies a merging of through the Bosman Case. The necessity of several clubs sometimes combined with reloca- revenues from broadcasting, merchandising tion or in the case of E Aalst, RWD Molenbeek, and sponsorship resulted in a strong commer- SK Lommel and KV Oostende the elimination cialisation of professional football. Professional of the existing club in the professional league. football was transformed into a consumer- oriented service that could be linked to an endogenous Notes demand-oriented growth model. This transformation showed that the factor ‘market’ became 1. In a professional competition, league franchise very important and had economic consequences means that the national association gives the on the Belgian professional football competi- right to a club to implement a professional struc- tion. The introduction of economic principles ture if the economic and market conditions are to professional football resulted in a decline sufficient and if there is not already another proof Belgian football. The Belgian league reached fessional club in that market segment. a relative maximum in attendances and their 2. The Bosman case was based on the Treaty of market is too small to compete with other Euro- Maastricht. That treaty guarantees free move- pean leagues. The Belgian league has too many ment of labour and questioned the rules for clubs in the first division and many of them player’s transfers among teams. Consequently, have great financial problems with a very small free trade agreements implied a policy that pre- market. The restructuring of this market of vents a domestic monopoly. heterogeneous clubs can be seen as an allocation- 3. Salary cap implies a maximum amount that clubs location problem. The solution of an allocation- can spend on players. The National Basketball location problem in an existing market has to Association (NBA) introduced it in 1980, the take into account the historical and cultural National Football League (NFL) in 1993 and situation.UNCORRECTED The transformation of football from Major League Baseball PROOF (MLB) in the mid 1990s a utility maximising to a more profit maximising (Downward & Dawson, 2000, p. 47) Negotiations consumer-oriented service implies the creation between the top clubs in Europe resulted in the of a model based on empirical data. Tabula rasa introduction of a salary cap. is, because of the heterogeneous characteristics 4. For the season 2001/02: Germany (33,032), of the different football clubs, not an option. Italy (30,747), England (29,619), Spain (24,691), This means that a theoretical maximum- France (22,688) and the Netherlands (15,564). covering solution can be used as a guide for the Belgium has an attendance of 9 people per every restructuring of an existing spatial distribution. 1000 inhabitants a higher relative rate than

88 TRUDO DEJONGHE

Germany (4.0), England (6.1), France (3.8) and in België, doctoraatsverhandeling, Universiteit Italy (5.3). Gent. 5. England $941 million, Italy $486 million, France Deloitte & Touche (2002), Annual Review of $397 million, Germany $384 million, Spain Football and Finance. Manchester: Deloitte & $237 million and Holland $57 million (UEFA, Touche Tohmatsu. 2001, pp. 16–17). Devletoglou, N. (1965), A Dissenting View of 6. 2000/01 season: England $1.5 billion, Italy Duopoly and Spatial Competition, Economica, $850 million, Spain and Germany $700 million London, London School of Economies and (Dejonghe 2001, p. 414). Political Science, 32, pp. 146–160. 7. This means that the broadcasting rights are Downward, P. & A. Dawson (2000), The Economics 11.2 per cent ($15 million) of the total budget. of Professional Team Sports. London: Routledge. 9. The total potential consumers, the total population Ghosh, A. & S. Craig (1983), Formulating Retail of Belgium, was in 2000 10,163,910 (NIS 1997). Location Strategy in a Changing Environment. 10. The number of demand nodes i is the 589 com- Journal of Marketing 47(19), pp. 56–68. munities after the merger of 1977. The popula- Golledge, R. & R. Stimson (1997), Spatial Behaviour: tion of these communities theoretically located A Geographic Perspective. New York: The Guilford in the centre was taken as the number of poten- Press. tial consumers. Hotelling, H. (1929), Stability in Competition, The 11. The urban centres in Belgium are, according to Economic Journal XXXIX(153), pp. 41–57. their centrality, classified in a hierarchical urban Lösch, A. (1962), Die Raumliche Ordnung der system. There is one urban centre in level 0, the Wirtschaft. Stuttgart: G. Fischer. capital city, 3 in level 1, the major cities and 17 Markham, J. & P. Teplitz (1981), Baseball Economics centres in level 2, the regional cities. (Van Der and Public Policy. Lexington: Lexington Books. Haeghen et al. 1982, p. 281). NIS (1997), Bevolking van het Rijksregister 01/01/ 12. AEC Mons promoted to the first division in 97 tabel 70.01 deelgemeenten. 2001/02. Noll, R. (1999), Competition Policy in European Sports after the Bosman Case. In: C. Jeanrenaud REFERENCES & S. Késenne, Competition Policy in Professional Sports. pp. 17–44, Antwerp: Standaard Ed Ltd. Bale, J. (1988), The Place of ‘Place’ in Cultural Saey, P. & M. Lietaer (1980), Consumer Profiles and Studies of Sport. Progress in Human geography 12(4), Central Place Theory. Tijdschrift voor Economische en pp. 507–524. Sociale Geografie 71(2), pp. 180–186. Church, R. & M. Meadows (1979), Location Sloane, J. (1971), The Economics of Professional Modelling Utilising Maximum Service Criteria. Football: The Football Club as Utility Maximiser. Geographical Analysis 11(1), pp. 358–379. Scottish Journal of Political Economy 70, pp. 65–94. Church, R. & C. ReVelle (1974), The Maximal Thomas, I. (1993), Sensibilité du Découpage Spatial Covering Location Problem. The Regional Science Optimal des Services d’urgence à la Définition de Association Papers 32, pp. 101–108. la Demande, L’espace Géographique 22(4), pp. 318– Daskin, M. (1995), Network and Discrete Location 332. Models, Algorithms and Applications. New York: Wiley. UEFA (2001), Report Football and Television 2001: Daskin, M. & S. Owen (1999), 2 New Location Comparative Analysis of Major European Countries. Covering Problems: The Partial P-centre Problem Paris. and theUNCORRECTED Partial Set Covering Problem. Geographical Van Der Haegen, H.,PROOF M. Pattyn & C. Cardyn Analysis 31(3), pp. 217–235. (1982), The Belgian settlement system. In: H. Van De Clercq, M. (1995), Economie Toegelicht, Leuven: Der Haegen, ed., West European Settlement Systems. Garant. Acta Geographica Lovaniensia 22, pp. 252–359. Dejonghe, T. (2001), De Noodzaak aan Exogeen Van Hecke, E. (1997), Hiërarchie van de Stedelijke Oplegde Economisch-geografische Principes bij Kernen in Vlaanderen. Heverlee: instituut voor het Professionaliseren van het door Endogene Sociale en Economische geografie. Factoren Ontstane Wereldsportstelsel. Gevalstudie: www.allesoverstadions.nl, 15/12/02. het Lokalisatievraagstuk van het Topvoetbal www.xs4all.nl, 15/12/02.