El deporte es un fenómeno social y económico en continua expansión, que contribuye a la prosperidad y la solidaridad en los países desarrollados (Aragonés, 2014). Ha sido reconocido por el Comité Olímpico Internacional como una herramienta para el fomento de la paz y las relaciones entre diferentes culturas. En relación a su vertiente económica, sobre la que se centra la presenta tesis doctoral, los ingresos generados fueron de 127 millones de euros en 2015, tal y como habían señalado las previsiones (PWC, 2011) y dio empleo a 1.700 millones de europeos en 2016 (Eurostat, 2018). Además, dada su pujanza y crecimiento en las últimas décadas, la industria del deporte ha sido objeto de interés para académicos y profesionales del marketing, buscando aplicar los principios de esta disciplina a los productos y servicios deportivos (Shank y Lyberger, 2014). Se trata de un sector complejo, en el que participan múltiples actores relacionados con otros sectores: turístico, construcción, telecomunicaciones, educación o el tecnológico entre otros (Laine y Vehman, 2017). Además, el desarrollo de las tecnologías de la información y la comunicación ha permitido que el deporte se haya convertido en un espectáculo de masas (Matheson, 2003) y los deportistas en celebridades seguidas por millones de espectadores (Sawe, 2018), siendo estos a su vez el foco de las principales marcas comerciales (IEG, 2017). Así, el patrocinio deportivo ha resultado ser una de las principales herramientas de comunicación de grandes empresas, alcanzando un volumen de 55.000 millones de euros en 2017 (IEG, 2018). En este contexto, los aficionados o espectadores son vistos como potenciales clientes por las marcas patrocinadoras de eventos deportivos (Davis y Hilbert, 2013). En fútbol, el deporte más popular del mundo (Sawe, 2018), el patrocinio se ha convertido en un factor clave de rendimiento económico y deportivo para los equipos. En las principales divisiones de las ligas más seguidas a nivel mundial, hay clubes de fútbol con ingresos cercanos a los 900 millones de euros, de los cuales más del 50% provienen de acuerdos de patrocinio (KPMG, 2017). Estos clubes se encuentran normalmente entre los más exitosos a nivel deportivo, debido a su gran poder a la hora de contar con los mejores y más valiosos jugadores en sus equipos (Transfermarkt, 2018). Con ello, los jugadores se han convertido en objetivo de las principales entidades deportivas y de las marcas comerciales, que ven en ellos una vía de acceder a su público (Bergkvist y Qiang Zhou, 2016). De hecho, en la actualidad, entre el 14% y el 19% de los anuncios muestran a celebridades, mayoritariamente deportivas, promocionando marcas y productos. Así, el endorsement de celebridades ha emergido como una solución dentro del marco del patrocinio deportivo para que marcas comerciales alcancen sus principales objetivos, como la notoriedad de marca y la asociación de la misma a una imagen y a unos determinados valores (Cantó, 2018). Según el tipo de deporte, los deportistas compiten en solitario o en equipo. En el primer caso, las marcas ponen el punto de mira en el deportista. En el segundo, tanto el deportista como la entidad deportiva son el foco de interés (Sport Business, 2017). Así, en deportes como el que nos compete, el fútbol, los jugadores profesionales llevan endosada una marca deportiva que les proporciona material deportivo además de remunerarles por vestirla en campañas publicitarias y durante la competición (Badenhausen, 2017). Por otro lado, los clubes en los que juegan contraen también acuerdos de patrocinio con marcas deportivas para la obtención de recursos (Cornwell y Maigan, 1998), material deportivo (Pope, 1998), y know-how (Roth, 1990), a cambio de mostrar su marca en sus equipajes, instalaciones deportivas y medios de comunicación (Meenaghan, 1983). De este modo, es común encontrar futbolistas con marcas endosadas diferentes a la marca patrocinadora del club al que pertenecen, situación que no ha sido objeto de estudio en la literatura previa sobre valor del patrocinio deportivo y que denominamos colisión de marcas. Dada la importancia que tienen los aficionados tanto para las entidades deportivas en relación a sus objetivos comerciales, como para las marcas patrocinadoras en cuanto a objetivos de notoriedad y de transferencia de imagen, resulta de gran interés estudiar la percepción del consumidor respecto a las marcas patrocinadoras y las marcas endosadas en las celebridades del deporte (Cantó, 2018). Así, aquellos factores que puedan influir en la percepción del aficionado/consumidor sobre las marcas son de especial interés para la investigación (Hutchinson y Alba, 1991; Cornwell y Humphreys, 2013). Con todo, el objetivo principal que persigue la presente tesis doctoral es el de conocer las consecuencias que tiene la identificación del aficionado en el valor percibido del endorsement (o endoso) de celebridades y en las reacciones de los aficionados a la marca endosada tanto en situaciones en las que existe colisión de marcas como en las que no, para conocer si ejerce un efecto en el resto de las variables del estudio. Objetivos específicos e hipótesis El primero de los objetivos específicos que se pretenden alcanzar se relaciona con dos tipos de identificación del aficionado: la identificación con el equipo de fútbol y la identificación con la celebridad. En concreto, se busca analizar el efecto del primero en el segundo. Según la Teoría de la Identidad Social (Tajfel y Turner, 1979), los individuos sienten la necesidad de aumentar su autoestima y la satisfacen a través de pertenecer o ser socios de una organización, como ocurre con los equipos de fútbol, haciéndoles sentir parte de ella. Factores como la familia y el entorno influyen en tal identificación (McPherson, 1976). Asimismo, Wann y Branscombe (1995) se basaron en la Teoría de la Identidad Social para argumentar que, dado que los individuos tienden a identificarse con un grupo, tenderán igualmente a enfatizar los aspectos positivos del mismo y a minimizar los negativos, abarcando a todos sus miembros. Así, los aficionados con un mayor grado de identificación con el equipo tenderán a desarrollar un mayor grado de identificación con los jugadores del mismo, objetivo que se pretende corroborar. En segundo lugar, se busca estudiar la relación entre la identificación del aficionado con el equipo y con la celebridad y el valor percibido del endorsement en la celebridad. En trabajos anteriores se han estudiado las consecuencias de la identificación del aficionado y las reacciones en el consumidor, tanto emocionales como de compra. En este sentido, se ha demostrado que una mayor identificación con el equipo incrementa el valor percibido de los productos relacionados con este (Kwon et al. 2007), su futura compra (Fisher y Wakefield, 1998), la compra impulsiva (Kwon et al. 2007), o la visualización de los partidos en los que juegue (Funk y James, 2001). La identificación también incrementa el conocimiento de los patrocinadores (Gwinner y Swanson, 2003) y la actitud hacia la marca patrocinadora (Biscaia et al., 2013). A partir de aquí, se persigue confirmar estos resultados en el contexto del endorsement, entendiendo que un mayor grado de identificación con el club y la celebridad provocará un mayor valor percibido del endorsement de dicha celebridad y de la marca endosada. En tercer lugar, se pretende estudiar la relación entre el valor percibido de la marca endosada y las reacciones hacia la misma. Investigaciones previas han analizado las reacciones actitudinales, demostrando una relación positiva entre valor percibido y satisfacción del cliente (Chen, 2013; Pandza Bajs, 2015; Kim y Park, 2017), y entre valor percibido y actitud (Baker et al., 2002; Ruiz-Molina y Gil-Saura, 2008; Wu y Chan, 2011; Im, Bhat y Lee, 2015). Por otro lado, otros autores han trabajado las reacciones relacionadas con la compra, pudiendo comprobar que el valor percibido ejerce un efecto positivo en la intención de compra, a través de la actitud hacia la marca o el producto (Kim y Hunter, 1993; Berger, Ratchford y Haines, 1994; Sirdeshmukh, Singh y Sabol, 2002; Voss, Spangenberg y Grohmann, 2003). Otros estudios, en cambio, han demostrado dicha relación, pero de forma directa, sin mediar la actitud en ella (Dodds, Monroe y Grewal, 1991; Grewal et al. 1998; Petrick y Backman, 2002). El presente trabajo busca trasladar esta relación entre variables establecida en otros campos científicos al contexto del endorsement de celebridades, asociando el valor percibido de la marca endosada con la actitud hacia la marca y la intención de compra. En cuarto y último lugar, se persigue analizar el efecto provocado por la colisión de marcas en las variables de estudio mencionadas.

OBJECTIVES OF THE STUDY All in all, the main objective of the present thesis is to know the consequences that the identification of the fan has on the perceived value of celebrity endorsement situations in which there is brand collision or not, in order to uncover if there is an effect in the rest of the study variables. Hence, this doctoral thesis aims to contribute as follows: • Presenting the state of the art of the sport industry, and more particularly of football´s, as a relevant economic sector in today societies. • Compiling, analysing and comparing the most important academic theories on sponsorship and endorsement as the basis from which to study different marketing constructs. • Setting a model of perceived value in relation to the celebrity endorsement context in sport, analysed from a customer/fan perspective in a non-previously studied situation: brand collision. • Acknowledging how brand awareness influences consumer perceptions about endorsed brands. • Bringing understanding about how customers/fans react towards endorsed brands in terms of perceptions and behavioural intentions. From the previous objectives, we propose a research model compiling the different variables of the study, resulting in seventeen hypotheses that are divided into three groups. The first group is composed by three hypotheses related to the identification of the fan that present the relationship between fan identification with team and fan identification with the celebrity, and the relationship of these two and perceived value of the endorsement. The goal is to measure how the identification of the fan influences the perceived value of an endorsement situation. The second group focuses on the main construct of the model, perceived value, and its outcomes in terms of brand attitude and purchase intentions. Concretely, we aim to analyse the effect of perceived value of the endorsement on perceived value of the endorsed brand, and the effect of these two on attitude towards the endorsed brand. In addition, the effect of brand attitude on purchase intentions is also analysed. The third group of hypotheses studies the particular situation in which the thesis falls within: brand collision. Thus, these hypotheses aim to analyse the moderating effect of brand collision on the rest of the variables of the model. A direct negative effect is studied between brand collision and brand awareness, so as to uncover whether brand collision situations erode brand awareness. Besides, moderating effects of brand awareness are also studied to compare how aware-of-the-endorsed-brand fans behave compared to unaware ones, when it comes to evaluate endorsed brands METHODOLOGY To achieve these goals and to test the presented hypotheses, an empirical descriptive research was undertaken throughout an electronic survey. An online questionnaire was developed and sent through online invitation to football supporters of three Spanish football clubs: Valencia CF, Real Madrid CF and FC Barcelona, in which each one of the constructs were measured, along with sport consumption and practice habits. The selected population enclosed people over 14 years old, living in Spain, who had watched at least a football match in the previous year. No-probabilistic sampling method was used, defining quotas according to age and gender. People were reached online by acceding to several websites related to football and its newsletters. A Sample of 324 valid questionnaires was obtained. Given the study’s interest in predicting the outcome variables and given the not too large sample size and the formative nature of the Perceived Value construct, a Structural Equation Modelling analysis was developed throughout SmartPLS 3. In addition, descriptive analysis has been performed through SPSS 26. Regarding the questionnaire, after completing introductory questions about habits of sport practice and consumption, individuals answered questions about their level of Identification with the team and with the celebrity endorser. Then, questions about the team’s sponsor Awareness and the endorsed Brand Awareness, followed by questions about Perceived Value, Brand Attitude and Purchase Intentions of the two brands: Nike and Adidas. Next, respondents were given the name of the actual endorsed brand on each celebrity and questions measuring Perceived Value of the Endorsement (union of the brand and the celebrity) were presented, along with questions about Brand Attitude and Purchase Intentions, once the union is known. All the questions about brands were randomly presented to respondents. Some of them answered about Nike before Adidas and some about Adidas before Nike. Hence, every respondent answered the same questions about one team and two players (one with the same endorsed brand as the one sponsoring the team, and one with a different brand from the team’s sponsor), so as to compare Brand Collision situations with no Brand Collision situations. The constructs of the model were measured using 7-items Likert reported scales: • Wann & Branscombe (1993)’s scale of Fan Identification. • Sweeney & Soutar (2001) multidimensional scale for Perceived Value, adapting the items to the endorsement context and removing those that made nonsense. • Gwinner & Bennet (2008)’s scales to measure Attitude towards the brand and Purchase Intentions. • Finally, Cornwell et al. (2012) dichotomous scale to measure Brand awareness. A Factorial Confirmatory Analysis was performed to analyse the Psychometric Properties of the measurement instrument. As both first-order and second-order formative constructs were analysed, several tests have been performed. The formative constructs presented multicollineality problems in 7 items, which were consequently removed. All the other parameters were above the threshold regarding reliability, convergent validity and predictivity of the model. In regards to discriminant validity, all the loadings were higher than cross-loadings. Regarding the sampling distribution, almost 90% were Spanish and 10% foreign, and 80% were men and 20% women, consistent with Spanish football followers’ population. RESULTS As for the results obtained from the descriptive analysis, it is possible to affirm that the sample was highly familiarised with sports as practitioners and spectators, as most individuals do or watch sport more than one or two days per week and that Nike and Adidas arose as the two main brands when it comes to buy sport clothes. Concerning the individuals’ identification with their favourite team, significant correlations allow to affirm that the better is one team rated, the worse the main rival will be rated. According to Davis & Hilbert (2013) these results show that the selected sample can be considered, on average, as composed by intensively enthusiastic fans. About Fan Identification with the celebrities of fans’ favourite team, results are diverse, being high with the top performing players and less high or even low with the less performing players. As far as Perceived Value is concerned, results are consistently high, regardless the team and player. However, Perceived Value of the endorsed brand obtains significantly higher results than Perceived Value of the endorsement, showing that individuals give high value to the brands, but that when the brand is associated with a player, the player appears to influence the value. Thus, Perceived value of the endorsement was high in the extent to which the player was one of the popular and top performers. In regard to Attitude towards the endorsed brand and Purchase Intentions, results are high in both brands. Nevertheless, Adidas obtained similar results among the three clubs while Nike obtained higher results from Real Madrid and Barcelona fans. As for Brand Awareness, roughly half of individuals knew the endorsed brand in the celebrities of their favourite team. Separating by players, results show that the most popular and top performers obtained higher degree of awareness. Bivariate t-student analysis comparing aware and unaware individuals showed that the aware ones had a more positive Attitude towards the brand and higher purchase intentions. Furthermore, once everybody was informed about the actual endorsed brand in each player, those that were previously aware showed again higher Attitude and higher Purchase Intentions. As far as the structural model is concerned, we have been able to accept six of the seven hypotheses presented. Thus, fan identification with the team exerts a positive effect on fan identification with the celebrity. Both exert a positive effect on perceived value of the endorsement, and the latter on perceived value of the endorsed brand. In addition, perceived value of the endorsed brand exerts a positive effect on attitude towards the endorsed brand and this one on purchase intentions. The only relationship that could not be accepted was the one linking perceived value of the endorsement and attitude towards the endorsed brand. Regarding the effect of brand collision on the rest of the relationships, a multisampling analysis was performed to compare brand collision situations with no-brand collision ones. Brand collision exerts a negative moderating effect on the relationship between fan identification with the celebrity and perceived value of the endorsement. No significant moderating effects arise between the rest of the variables. Additionally, no significant direct effect has been reported between brand collision and brand awareness. Nevertheless, brand awareness by itself appears to moderate some of the variables of the structural model. In another multisampling analysis, in which aware fans’ answers were compared to non-aware fans’ answers, significant results show that brand awareness exerts a positive moderating effect between perceived value of the of the endorsement and attitude towards the endorsed brand, as well as between attitude towards the brand and purchase intentions. In addition, a negative moderating effect has been reported between perceived value of the endorsed brand and brand attitude, contrary to what was expected as per the previous literature. Hence, all the hypotheses of the structural model have been accepted but one. However, only three hypotheses of the moderating effects have been accepted. CONCLUSIONS AND RECOMMENDATIONS Some theorical and practical conclusions arise from this work. Regarding the first ones, we have seen that endorsement falls within the sponsorship literature as a differentiated promotion tool focusing on individuals rather than institutions. Besides, the goals pursued when endorsing a celebrity are the same as when sponsoring an event, and their factors of success are framed under the same theories of study. The model of Perceived Value can be translated to the endorsement context. In this case, the object of the study is the union of a brand and a celebrity, rather than a product. Regarding Fan Identification, it appeared to be an important antecedent of endorsement evaluations and purchasing decisions, as well as a factor that goes beyond the team, and makes fans to be involved and committed with other elements surrounding the team, such as its members, venues, sponsors etc. As for Perceived Value (a multidimensional formative construct as per the recent literature), it has appeared to be a good predictor of consumer behaviour, not only related to products, but also to brands and endorsement situations. Finally, Brand Collision is a unique situation that deserves to be considered as a moderating factor, given the negative results reported in the sponsorship literature when several brands appear at the same time. Besides, consistently with previous research, it may appear as a factor influencing Brand Awareness, the main goal studied in the sponsorship literature. Regarding the practical conclusions, results allow affirming that fans are more identified with teams than with athletes, and that Fan Identification exerts a positive influence in the performance of sport celebrities’ endorsement campaigns throughout more positive Attitude towards the endorsed brands and higher Purchase Intentions of the endorsed brand’s products. Meanwhile, it is possible to support that Perceived Value of the union of a brand and a celebrity is a key factor explaining the future behavioural intentions towards the endorsed brand. Moreover, although data did not allow demonstrating that Brand Collision negatively affects Brand Awareness of the endorsed brand, we have shown that those who were aware of the brand endorsed in each player, valued the endorsed brand more positively. Additionally, we have been able to demonstrate that when Brand Collision occurs, the effect that Fan Identification with the Celebrity exerts on Perceived Value of the endorsement is eroded. From these conclusions, some managerial implications and recommendations can be extracted. Focusing on the most relevant ones, we would suggest marketers to use endorsement in a long-term basis in order to leverage it and obtain positive evaluations. In addition, we would recommend to focus on the highly identified fans when communicating sponsorship and endorsement campaigns. In the aim of reducing the potential negative effects of Brand Collision, it would be advisable to seek players belonging to teams that are already sponsored by their brand. Finally, it is highly recommended to strongly communicate the endorsement agreements so as to make fans aware of them, given the positive effects on brand evaluations. On the other hand, we would suggest endorsers to reduce the number of brands they are related with. We would recommend to put their efforts into obtaining highly identified fans so as to be of great influence to them. These two measures will help athletes become an appealing platform to sponsors. LIMITS OF THE STUDYAND FUTURE RESEARCH All in all, the present thesis has some limitations. It would be worth to consider the possibility that other factors not included in the study might have an effect on the outcome variables. For example, the sample has shown a high predilection to the two selected brands: Nike and Adidas. In addition, the scale used to measure Perceived Value by Sweeney & Soutar (2001) for durable products has been conceived as reflective, while the present study follows the formative research line, following the indications from Jarvis et al. (2003). Room for further research arises beyond overcoming these limitations. For example, including moderating factors such as congruence between the brand and the endorser would be of great interest in situations of brand collision, seeking whether the latter diminishes its moderating effect when the endorsed brand is not congruent with the player, and so not a competitor of the team sponsor. Finally, given the importance of Fan Identification in the whole model, it would be of importance to analyse the antecedents of this construct, like Football Team’s Brand Equity, in order to uncover how they would influence the whole model of Endorsement Perceived Value.