Does Attractiveness Lead to or Follow From Occupational Success? Findings From German Associational Football

Prior research has provided evidence that attractiveness is associated with work-related advantages. It is less clear, however, whether attractiveness is an antecedent or a consequence of professional success. To answer this question, associational football in Germany is used as an exemplifying case. Portrait pictures of German football players were retrieved, one picture from a very early career stage and one from a very late one. Attractiveness of these portraits was assessed by the “truth of consensus” method. Panel regression models are applied to analyze changes in attractiveness and relate these changes to professional success. Findings show that success as a footballer cannot be predicted with attractiveness at early career stages. Instead, the increase of attractiveness over time is more pronounced among very successful players. It is thus concluded that successful individuals are not more attractive in the very beginning, but improve their appearance throughout their careers.


Introduction
Scholars have increasingly addressed the relevance of physical attractiveness for social interactions and socioeconomic status. Previous research has found, for instance, that more attractive individuals enjoy substantial advantages in almost every domain of social life (Agthe et al., 2011;Eagly et al., 1991;Frevert & Walker, 2014;Langlois et al., 2000). Studies suggest that attractiveness influences grades of pupils in secondary school (Dunkake et al., 2012) and the vote share gained by candidates in political elections (Milazzo & Mattes, 2016). More attractive people report greater happiness, and less distress and depression (Gupta et al., 2016). Moreover, they are regarded in a more favorable way by others, for instance, perceived as healthier, more intelligent and their personality characteristics are perceived more positively (Feingold, 1992;Jackson et al., 1995;Tartaglia & Rollero, 2015). Attractive people are also trusted at higher rates (Wilson & Eckel, 2006;Zhao et al., 2015). Most notably, however, several econometric studies support the idea that physical attractiveness is of importance in the labor market and influences earnings (Anýžová & Matějů, 2018;Benzeval et al., 2013;Fletcher, 2009;French, 2002;Hamermesh & Biddle, 1994;Harper, 2000;Mitra, 2001;Pfeifer, 2012;Sala et al., 2013;Scholz & Sicinski, 2015;Schunk, 2016). A review of experimental studies suggests that attractive people fare better than unattractive individuals in terms of job-related outcomes with an average effect of d = .37 (Hosoda et al., 2003). The biasing effect of attractiveness seems to be particularly high with regard to rather diffuse evaluations of "suitability" or "employment potential," whereas effects are lower when it comes to "performance evaluations." Scholars have also suggested that attractiveness may have a larger effect on economic outcomes in women compared with men (Jaeger, 2011).
However, the mechanisms that lead to such a "beauty premium" are not fully understood by now and researchers thus call for studies that examine the "potential pathway in more detail" (Benzeval et al., 2013, p. 6). In particular, one of the unsolved key questions in this field of research addresses the direction of the effect, that is, whether success in the job market follows from attractiveness or-vice versa-attractiveness follows from (work-related) success. The majority of studies explicitly or implicitly suggest that attractiveness should be understood as an antecedent of occupational success and leads either to positive discrimination or to productivity advantages 903413S GOXXX10.1177/2158244020903413SAGE OpenMeier and Mutz research-article20202020 1 University of Münster, Germany 2 Justus-Liebig-University Gießen, Germany in the labor market. We later refer to this notion as the "success-follows-attractiveness" rationale. Other research, however, argues that attractiveness is a consequence of success, as more successful people invest more money and time into their appearance. In this regard, it is success that helps to make people attractive. We fully elaborate on this "attractivenessfollows-success" rationale later.
Using professional football as an exemplifying case, this article aims to shed light to these divergent explanations. More precisely, this study uses a novel approach and employs panel regression analyses to test whether (a) attractiveness at earlier career stages fosters long-term success as a professional football player and (b) success as professional player leads to an increase in attractiveness in later career stages. Hence, the unique contribution of this analysis lies in the research design that allows for assessing the direction of the dependency.
Moreover, for several reasons, professional football seems to be an ideal domain for testing the relationship between attractiveness and success in detail. First, different from other segments of the job market, sport provides several valid and objective indicators for professional success. Whereas measures for work-related performance are rather crude in many occupational domains, success in football becomes visible in wins and losses (at the team level), and also in the transfer values of players (at individual level). Studies have shown that transfer values can be considered as the monetary reflection of playing ability (Gerhards et al., 2014;Herm et al., 2014). Second, professional football can be considered as part of the entertainment business where attractiveness supposedly matters more than in other domains. There are, for instance, good reasons to assume that physical attractiveness is relevant for athletes' endorser qualities (Cunningham et al., 2008;Fink et al., 2004;Ross et al., 2009) and that selection routines of sports media also favor attractive individuals (Harris & Clayton, 2002). Moreover, it has also been claimed that physical attractiveness matters for athletic achievements, as more attractive players seem to be more likely to get fielded and generate higher transfer values (Rosar et al., 2010(Rosar et al., , 2013(Rosar et al., , 2017. What is less clear, however, is the direction of the dependency, namely, whether success follows beauty or beauty follows success.

Success Follows Attractiveness
Several mechanisms were proposed that possibly can explain the "beauty premium" in earnings (see, for example, Hamermesh & Biddle, 1994). First, earning differentials may stem from sorting effects, that is, a self-selection of more attractive people into those occupations where attractiveness matters for productivity. For instance, more attractive individuals may decide to apply for jobs in the service sector, particularly sales occupations, where good looks may indeed serve as an advantage. Second, it was argued that a beauty effect on income might reflect a productivity advantage. Greater productivity of attractive workers may result from consumer or coworker preferences, who both may prefer to interact and cooperate with better-looking individuals. Moreover, it was assumed that better-looking individuals may have more confidence and self-esteem and that these characteristics may also increase productivity in certain occupations. Third, employer discrimination against less attractive individuals had been put forward as another possible mechanism. This idea is supported by the finding that beauty effects on earnings are present in a broad range of occupational categories (Hamermesh & Biddle, 1994;Scholz & Sicinski, 2015). Benzeval et al. (2013) argue that gatekeepers to occupational positions may judge candidates based on stereotypes. Given that attractive people are stereotyped in a more positive way, that is, usually regarded as more agreeable and more competent, they may ascribe more positive traits to attractive applicants. Hamermesh and Biddle (1994) also conclude that "strongest support is for pure . . . discrimination based on beauty" (p. 1193). What all these mechanisms have in common is that they suggest that success follows from attractiveness. Despite the fact that most study designs do not allow for conclusions regarding the direction of the dependency, this interpretation-due to its theoretical soundness and plausibility-is favored by many scholars.
When it comes to professional football, some of these arguments seem less convincing, as attractiveness seems to be unrelated to football skills and shall thus only be of marginal importance as a productivity factor. Hence, arguments considering sorting effects and productivity effects are less convincing. However, research conducted on the German top-tier football league, the Bundesliga, has claimed that in professional soccer, more attractive players still enjoy a beauty premium (Rosar et al., 2010(Rosar et al., , 2013(Rosar et al., , 2017. Based on an exhaustive survey of all players of the 2007-2008 season, colleagues (2010, 2013) can show advantages for more attractive football players, namely, a higher probability of getting fielded. In their discussion, the authors assume that coaches are biased in their fielding decisions, implicitly giving an advantage to more attractive players. Furthermore, the same authors have shown that a higher transfer fee is ascribed to more attractive players (Rosar et al., 2017). Based on these findings, they conclude that "the physical attractiveness of German professional footballers does indeed exert a substantial influence on their market value, . . . broadly confirming the known findings from other areas of application" (Rosar et al., 2017, pp. 110-111). Hence, the association between attractiveness and professional success also seems to exist in the domain of professional football, and mainstream interpretations favor arguments that imply that successhowsoever-follows from attractiveness. Based on these considerations, we can hypothesize the following: Hypothesis 1 (H1): More attractive young players are more successful in their later professional career.

Attractiveness Follows Success
Despite being less prominent in the ongoing discussions surrounding the link between success and attractiveness, some scholars have suggested that job-related outcomes may be due to investments in physical appearance. In this regard, it was proposed, for instance, that regular exercise increases earnings by 6% to 10% (Kosteas, 2012). However, this association may likely be spurious and caused by a better health status of regular exercisers (Xiao et al., 2015). More narrowly related to the topic of attractiveness, however, is the literature on grooming and "beauty work" (Kwan & Trautner, 2009). Beauty work refers to several practices that aim at improving one's appearance, starting with make-up, hairstyle, and outfits and ending with cosmetic surgery. A recent study suggests that grooming has a beneficial effect on wages and-at the same time-reduces the effect of beauty on wages substantially (Wong & Penner, 2016). There seems also to exist a penalty for poor grooming (Robins et al., 2011). Moreover, a study on attitudes toward job candidates found that neat grooming and business-like clothing had more favorable impacts than mere attractiveness (Ruetzler et al., 2012). Hence, the limited findings on grooming suggest that the beauty premium on earnings is only to a lesser degree due to "natural" beauty but may be better conceived as the effect of "beauty work." This notion is in line with Hakim's (2010) claim that individuals in Western societies can invest into their "erotic capital" (e.g., beauty, fitness, style of dress) and that such investments pay off in the long run, primarily in job and marriage markets.
In the realm of professional sport, it was shown that physical attractiveness matters for the popularity of athletes (Mutz & Meier, 2016). With regard to football players, effects were equally strong for male as well as female players and those athletes, who were successful and attractive alike, were by far the most popular. In professional tennis, however, the attractiveness effect was only significant for female players (Konjer et al., 2019). Popularity and attractiveness, in turn, can be considered as crucially important for athletes' potential to generate revenues from sponsors (Belch & Belch, 2013). Moreover, research has further shown that physical attractiveness by itself can also increase an athlete's endorser quality (Cunningham et al., 2008;Fink et al., 2004;Liu & Brock, 2011;Ross et al., 2009). Hence, in the domain of professional sport, strong incentives exist for athletes to invest into their appearance to increase their endorser qualities and, in turn, their earnings. Therefore, the current study aims to address a second assumption, namely, that successful athletes, who are standing in the spotlight of public and media attention, improve their looks over time: Hypothesis 2 (H2): The greater the success as a football player, the more can a player improve his physical appearance in a later career stage compared with an earlier stage.

Sample
The basic aim of the current study was to provide further insights on the relationship between career outcomes and physical appearance, particularly on the direction of influence. Therefore, a panel design and a sample of male football players from the German Bundesliga were chosen. The sampling strategy proceeded in three steps: First, the rosters from the 18 first league teams of the seasons from 2000-2001 to 2004-2005 (90 teams) were scanned for all players in the age range of 18 to 21 years. The list of players found amounted to 220. Second, head-and-shoulder photographs for these young players, who were all at the beginning of their football careers, were retrieved from the website www.kicker.de, the web portal of Germany's most renowned football magazine. Third, we searched for comparable photographs for each player in the seasons from 2010-2011 to 2014-2015, hence when each player was approximately 30 years old. Hence, about 10 years lie in-between the first picture (t 1 , "young") and the second picture (t 2 , "old"). It was rather simple to find comparable pictures for those players who actually had progressed their career in a top league; however, many players who did not made it as a professional could not be retrieved. For many of these players, we could not find any information on the internet at all, not to mention portrait pictures. It seems that almost two thirds of the young players in 2000-2001 to 2004-2005 dropped out of professional football. Nevertheless, the final sample for which complete material was available consisted of N = 80 football players. The final sample included wellknown German superstars (e.g., Phillip Lahm, Mario Gomez, Bastian Schweinsteiger) as well as players, unknown to a larger public, who spent most of their football careers in the third or fourth German leagues. A full list of the 80 players can be found in the Supplemental Material (see online Appendix).

Indicators for Attractiveness and Professional Success
For measuring physical attractiveness, the "truth of consensus"-rating method was employed, the standard approach in attractiveness research (Hamermesh, 2011). Precisely, five independent raters (two males, three females) evaluated the portrait photographs of the players on a Likerttype scale ranging from 1 = not attractive at all to 7 = highly attractive. A central insight of research on physical attractiveness is that people's perceptions of others' attractiveness are highly similar, so that comparatively small groups of raters-usually between three and eight (Biddle & Hamermesh, 1998;Hamermesh & Parker, 2005;Stelzer et al., 1987;Zakahi et al., 1994)-suffice for generating reliable attractiveness ratings. In this study, interrater agreement was sufficiently high with Cronbach's alpha at .81. Each player was rated two times, with one picture showing him as a young player (<21 years) and one picture showing him in his later career stage (approximately 30 years). However, for the rating, the pictures were divided into two blocs of 80 pictures each, in which none of the players appeared twice. Whether the "young" or the "old" picture appeared in the first or in the second bloc was varied at random. Attractiveness scores range from 1.4 to 5.6 with a mean score of 3.28 (Table 1). Attractiveness scores at t 1 do not significantly correlate with the age (18-21 years) of the player at t 1 .
Professional success at the individual level is hard to determine in a team sport like football. However, following prior research (Gerhards et al., 2014;Rosar et al., 2017), we rely on data on the players' market values. The market value usually equals the transfer fee that clubs have to pay for signing a player before his ongoing contract ends. These transfer fees closely resemble the players' football ability, hence the higher the transfer fee the better the player. Although transfer fees are usually only communicated when players actually switch clubs, proxies based on expert estimates are available from the website www.transfermarkt.de. It was shown that these estimated values and the transfer fees which are actually paid by clubs correlate with r > .90 (Gerhards et al., 2014;Herm et al., 2014). This very high correlation suggests that expert ratings can indeed be used as proxies for the players' market value. In the current study, we have retrieved the market values for each player and use the players' career maximum as a proxy for success as a professional football player. We use the highest career market value and not the market value at the age of 30 because market values are agedependent and tend to decrease when a players' career draws to a close (Kalén et al., 2019). The maximum career market values in the sample range from €130,000 up to €42,000,000, thus players represent a broad range of athletic ability. Moreover, we use a second indicator for professional success which is the number of seasons a player had been under contract in the so-called "Big-Five" leagues (English Premier League, Spanish La Liga, German 1. Bundesliga, Italian Serie A, and French Ligue 1). Having a contract in one of these leagues can be equaled with professional success, given that athletic quality and salary levels in these leagues stand out. Seasons in "Big-Five" leagues are counted for the period 2000-2001 until 2014-2015. The indicator has a range between 1 and 15. Table 1 provides descriptive statistics for the variables retrieved.
Moreover, German Bundesliga teams have become increasingly multiethnic over the last decades (Gerhards et al., 2014), which is also resembled in the player sample used here. Previous research has shown, however, that race and ethnicity influence perceptions of attractiveness as well as social judgments (Blair et al., 2004;Parks & Kennedy, 2007;Sims, 2012). Therefore, a control for possible ethnic bias is used in our analysis. Nationality and citizenship are inadequate as measures, because both features are not recognizable by looking at portrait pictures. However, among the 80 players in the sample are 62 White and 18 non-White players. Hence, we use this difference as a crude proxy for ethnicity.

Results
According to H1, more attractive young players supposedly have a higher probability to make it as a professional in the European top football leagues. The data, however, do not provide any support for this notion: Attractiveness (at t 1 ) and professional success, measured either with the player's highest career market value (r = −.09, p = .41) or with the number of seasons the player was under contract in "Big-Five" leagues (r = −.03, p = .81), are uncorrelated. Attractive young players thus have no advantage in their football career and the hypothesis suggesting that professional success follows from attractiveness must be rejected.
According to the second hypothesis, football players who enjoy more successful careers can invest into their appearance and thus improve their attractiveness over time. As panel data with repeated observations are used here, that is, attractiveness ratings for the same individual players, panel regressions were conducted. As Hausman tests suggest that fixed and random regressions yield very similar results, random effects regressions are reported ( Table 2). The table shows regression estimates for models with market values (Models 1 and 2) and seasons in "Big-Five" leagues (Models 3 and 4) as a proxy for career success, including interaction effects for Time × Market Value (Model 2) as well as Time × Seasons in "Big-Five" leagues (Model 4). In addition, all models control for ethnicity, that is, players with a non-German appearance.
First, the results confirm that football players become significantly more attractive over time. Hence, players around the age of 30 are considered more attractive compared with younger players below 21 years. Second, support is found for the hypothesis that more successful football players enjoy a higher increase in attractiveness over time compared with less successful ones. The significant and positive Time × Market Value interaction indicates that an increase in the market value by 1 million is associated with 0.030 points increase in attractiveness. Keeping in mind that some players in the sample had market values around €40 million, one would estimate that these players (compared with their colleagues who never were rated with a market value above €1 million) could increase their attractiveness by one additional point. The finding that success can increase attractiveness is further buttressed in the models that use seasons in "Big-Five" leagues as a measure for success: The positive Time × Big-Five league interaction shows that each additional year in one of Europe's leading football leagues is associated with 0.043 points increase in attractiveness. Finally, ethnicity is not significantly associated with attractiveness, although tentatively non-German appearing players are rated marginally less attractive compared with German-looking players.
To illustrate the main finding that male football players tend to become more attractive over time and-in accordance with H2-this gain in attractiveness is even greater among more successful players, we simply divided the sample into three groups, representing those with the most success, the least success, and those in-between (Figure 1). It is shown that all three groups of players start at rather similar levels regarding their attractiveness at younger ages. Mean scores at t 1 vary only slightly between 2.88 and 3.01. However, at later career stages, the least successful group of players increased their attractiveness by 0.42 points ( M t 2 = 3.30) and the medium group of players increased their attractiveness by 0.62 points ( M t 2 = 3.63), whereas the most successful group of players increased their looks by 1.01 points ( M t 2 = 3.91).

Discussion and Conclusion
The role of physical attractiveness for job-related interactions and outcomes is intensely debated. Previous research has pointed to the existence of a beauty premium in the labor market, but scholars have recently emphasized that the causal mechanisms behind this beauty effect are not completely understood. The objective of this study was to provide some clues on the direction of the dependency, whether attractiveness leads to or follows from success. The first notion that attractiveness fosters professional success in associational football was clearly rejected (H1). At the same time, it was shown that more successful football players markedly improve their physical appearance over time, lending support to the second idea that attractiveness follows from success (H2). Hence, it can be concluded from the findings that attractiveness is less an antecedent, but more a consequence of success. Hence, beauty is not a stable characteristic of a football player, but something modified by "beauty work." Large cross-sectional studies on football in Germany had shown that attractiveness and success are correlated (Rosar et al., 2010(Rosar et al., , 2013(Rosar et al., , 2017. In the interpretation of this association, it was claimed that coaches may give attractive footballers an advantage in fielding decisions which may help attractive players to become successful. In particular, the interpretation that coaches favor more attractive players was put forward by Rosar and colleagues (2017). However, bearing in mind that football is one of the few professional domains where attractiveness has particularly no relevance as a productivity factor, this interpretation comes as a surprise. Our results lend more support to the notion that players who are fielded more often (and are thus more often in the public spotlight) invest more into their beauty. Although this needs to be tested in future research more explicitly (including measures for grooming), the findings presented here Note. Random effects panel regressions using the xtreg command in Stata. 160 observations in total, 80 groups (players) and two observations (attractiveness ratings). *p < .05. **p < .01. ***p < .001.
suggest that the beauty premium in sport is probably more accurately interpreted as a by-product of beauty work and not as a form of discrimination against less attractive players.
If this line of reasoning is correct, it is still unclear what motivates this beauty work: On one hand, professional athletes are offered huge financial rewards for attractiveness and popularity, because these qualities are valued by media and the sport industry. For an athlete, beauty work can thus be a form of strategic investment to reach a broader public beyond the narrow scope of regular football fans and, in doing so, increase his endorser qualities. David Beckham or Cristiano Ronaldo may be considered textbook examples of this strategy (Coad, 2005). In forms of sponsorship and marketing deals, beauty work may thus pay-off for athletes and lead to higher revenues. However, Hamermesh et al. (2002) have also contested the idea that additional earnings due to investments in physical appearance recover costs (e.g., for clothing and cosmetics). However, this study was not conducted in the realm of professional sport and may thus not hold true in this particular context. On the other hand, beauty work must not necessarily represent an investment strategy, but may simply be a form of "conspicuous consumption" (Veblen, 1899(Veblen, /2007. Conspicuous consumption refers to the acquisition of luxury goods, including expensive clothing, to publicly demonstrate wealth and a high social status. Hence, in this line of interpretation, the "returns" of beauty work do not tend to a monetary but to a symbolic level, aiming at distinction and prestige. Moreover, it was also claimed that showy spending increases sex appeal among men (Sundie et al., 2011). Hence, beauty work among high-class football players, who stand in the limelight of a huge TV audience each weekend, may simply represent a form of impression management to showcase oneself in a positive way and generate symbolic capital.
This finding comes with strong implications for future research on the role of physical attractiveness in professional sport: Future research has to go beyond correlational analysis and needs to employ longitudinal research designs to be able to discriminate between different mechanisms at stake. Simple correlational analysis does not suffice for making conclusive inferences on the impact of attractiveness on football players' careers. Moreover, as the current study leaves unclear why successful football players improve their physical appearance, future research should address beauty work and its financial and symbolic returns.
One limitation of this study is that it measured beauty solely based on facial attractiveness. According to Hakim (2010), beauty, sexual attractiveness, physical fitness, liveliness, charm, and style are distinctive features that can make a person attractive for others. Although some of these characteristics are hard to measure as they are not assessable with pictures (e.g., charm) or change quickly (e.g., style), it should be kept in mind that this study (as with many previous studies) reduces beauty to facial attractiveness while ignoring other (body) characteristics. Moreover, as an alternative to Note. N = 80 players rated at t 1 and t 2 , attractiveness ratings represent the mean of five independently conducted ratings. Error bars indicate the standard error of the mean. the "truth of consensus"-rating method, scholars have suggested a software-based approach, analyzing facial geometry, for instance, horizontal symmetry, ratio of nose to ear length or ratio of face width to face height (Hoegele et al., 2015). This is a promising approach so that future studies would do well to integrate rater-based as well as softwarebased methods for assessing facial attractiveness. Finally, this study solely focused on male athletes so that it remains uncertain whether these findings would also hold for female athletes. Previous studies on attractiveness and occupational success found stronger effects for women compared with men (Jaeger, 2011). Similar findings were reported for female professional tennis players, whose popularity is much more driven by their attractiveness compared with male players (Konjer et al., 2019). However, in view of the fact that women's football is less professionalized and commercialized as a sport in Germany (e.g., with regard to media coverage, salary levels, or endorsement deals), the incentives to invest into beauty and appearance may not be as high as in men's football. Hence, replications of this study in women's football, in other fields of professional sport, and in different domains of the entertainment industry would be helpful to assess whether the findings presented here are generalizable or an expression of peculiarities of European associational men's football.

Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.

Supplemental Material
Supplemental material for this article is available online.