Human capital, formal qualifications, and income of elite sport coaches.

Wicker, Pamela ; Orlowski, Johannes ; Breuer, Christoph 等

Introduction

The development of elite sport is a key policy concern in many Western countries including the United Kingdom (Green, 2004), Australia, Canada (Green, 2007), and Germany (German Federal Government, 2014). Consequently, governments allocate large amounts of public funds to elite sport development (Green, 2007; Grix & Carmichael, 2012). For example, the German government has provided approximately 1 billion [euro] for the promotion of sport between 2010 and 2013 with a large part of the money being attributed to the promotion of elite sport (German Federal Government, 2014).

Within elite sport systems coaches are situated at critical positions because they represent the link between government policies and investments, respectively, and elite sport achievements (Liston, Gregg, & Lowther, 2013). In addition to coaches, there are more critical factors because elite sport success is a combination of several factors as conceptualized in the SPLISS model (i.e., sports policy factors leading to international sporting success) by De Bosscher, De Knop, Van Bottenburg, and Shibli (2006). This model states that nine factors influence international sporting success. These pillars are: financial support; governance, organization, and structure of sport policies; foundation and participation (e.g., in clubs and schools); performance (e.g., talent identification and development); excellence (e.g., athletic career support); training facilities; (inter)national competition; scientific research and innovation; and coaching provision and coach development (De Bosscher et al., 2006).

The focus of this study is on the second facet of the last pillar (coach development) and more specifically, coach education and the returns to education. Generally speaking, coach education is a complex topic because the job of a coach is characterized by various roles and responsibilities. For example, in addition to the organization of the actual sport practice, coaches are responsible for selecting talent (Inoue, PlehnDujowich, Kent, & Swanson, 2012), have administrative (Laios, 1995) and managerial responsibilities (Inoue et al., 2012), fulfil parental roles (Burke & Johnson, 1992), need pedagogical skills (Jones, 2007), and serve as psychologists and mental coaches (Gucciardi, Gordon, Dimmock, & Mallet, 2009). These skills should also be reflected in coach education; yet, given the variety of coaches' responsibilities, there is no specific type of coach education or degree that covers all these skills.

In an effort to acquire the relevant skills mentioned, many coaches now hold various qualifications such as academic degrees, coaching licenses offered by (inter)national sport associations, and various types of additional coaching formations and certificates. However, it is questionable if all of the available qualifications are equally significant in terms of obtaining the relevant coaching knowledge and generating income. While the content of coach education has already been examined in previous research (e.g., Piggott, 2012, 2015), the effect of different formal coaching qualifications on income has been largely neglected. Since the working conditions of many elite sport coaches are characterized by high weekly workloads and relatively low income (Digel, Thiel, Schreiner, & Waigel, 2010), the question of what coaching qualifications pay off is a relevant one.

The purpose of this study is to examine the relationship between different formal coaching qualifications and income of elite sport coaches in less commercialized sports. Previous research almost exclusively looked at intercollegiate athletics when examining the determinants of coaching salaries (e.g., Byrd, Mixon, & Wright, 2013; Grant, Leadley, & Zygmont, 2013), probably because information about coaching salaries in other sports are hardly publicly available. Therefore, primary data were collected using an online survey of elite sport coaches (n=186). Coaches were asked to state all the formal qualifications they have, allowing a detailed analysis of the role of qualifications. This study contributes to the body of knowledge on coaching salaries and labor market research in elite sport.

Research Context

The research context of this study is Germany, where the working conditions and specifically the salaries of elite sport coaches in less commercialized sports are on the political agenda (German Parliament, 2014). This study uses the definition of elite sport suggested by Hong (2011): "Elite sport can be defined... as a competition in sport at the highest international level with a priority put on sports in the Olympic Games programme, and on those sports with regular world championships" (p. 977). In Germany, elite sport is funded by the federal government, while community sport is mainly supported by state and local governments. This is why the federal government and the German Parliament discuss and set the regulatory frame and financial means of elite sport coaches in less commercialized sports. In these sports, elite sport coaches are financially supported by the government; coaching salaries are only partially determined by the market. Having said that, this study excludes more commercialized sports like football, tennis, and boxing.

Since coaches in less commercialized sports have complained about their salaries for several years (Suddeutsche Zeitung, 2013) and coaching migration is a concern (Gienger, 2008), the federal government took measures to improve the financial compensation of elite sport coaches. Generally speaking, there is a directive that people employed in publicly funded jobs are not allowed to earn more than other employees in the public sector in comparable jobs (Federal Office of Administration, 2014). Since elite sport coaches are also publicly financed, this regulation would also be applicable to them. However, it was decided that elite sport coaches are excluded from this regulation to ensure the competitiveness of German elite sport. Up to 104,000 [euro] in funding is available for the yearly gross salary of national coaches (Federal Ministry of the Interior, 2015). Yet, the decision about the salary level is at the discretion of the national sport association. Thus, national coaches do not automatically receive this gross salary because the association can decide to pay a coach less or use this money to hire several coaches.

Theoretical Framework

The relationship between coach qualifications and income is rooted in the theory of human capital (e.g., Becker, 1962; Mincer, 1974; Schulz, 1960). Following Becker (1962), "activities that influence future real income through the embedding of resources in people... is called investing in human capital" (p. 9). The focus here is not on physical resources, but on less tangible (i.e., intangible) resources like knowledge. Investment in human capital includes, for example, schooling and on-the-job training (Becker, 1962) and is associated with gains in information, knowledge, skills, capabilities, and competencies (Becker, 1962; James, 2000; Schulz, 1960). In addition to schooling and on-the-job training, there are further activities that "raise real income primarily by increasing the knowledge at a person's command" (Becker, 1962, p. 26). Investments in human capital can lead to a competitive advantage when the individual's competitors have not made such investments (James, 2000).

The human capital theory assumes that an individual's level of human capital is positively associated with income (Becker, 1962; Mincer, 1974). However, the amount of resources invested and the monetary returns differ between the different ways of investing in human capital (Becker, 1962). Notably, investment in human capital is associated with costs; foregone earnings are costs of human capital as well as resources that are invested in training rather than in producing current output (Becker, 1962; Schulz, 1960). Thus, the typical relationship between age and earnings (i.e., earnings increase with age at a decreasing rate) can also be explained with human capital theory; earnings are lower during the investment period and greater afterwards (Becker, 1962).

Applying the concept of human capital to this study, on-the-job training is reflected by the number of years a person has worked as a coach and gained coaching experience. The formal qualifications that are available to elite sport coaches reflect different types of investment in human capital; while academic degrees reflect schooling (i.e., an investment in human capital made in an institution that specializes in teaching; Becker, 1962), the various coaching licenses and certificates can be considered further activities that increase the coaches' (sport-specific) knowledge base.

Human capital theory is often discussed together with social capital theory (e.g., Barros & Barros, 2005; Sagas & Cunningham, 2005). Following Lin (2001), social capital "consists of resources embedded in social relations and social structure" (p. 24). From a professional perspective, it includes an individual's social network and relationships with peers, colleagues, subordinates, and superiors (James, 2000). Research has shown that both human capital (e.g., education, experience) and social capital (e.g., network, weak, and other ties) have a positive effect on the earnings of sport administrators (Barros & Barros, 2005). For coaches, social networks were found to be especially relevant to the reception of job offers (Taylor, 2010). While it may be interesting to examine the role of social capital in coaching income, the focus of this research is on human capital.

Literature Review

The majority of studies examining the effect of human capital on coaching salaries were conducted in intercollegiate athletics, particularly in college football (Byrd et al., 2013; Fogarty, Soebbing, & Agyemang, 2015; Grant et al., 2013; Humphreys, Soebbing, & Watanabe, 2011; Soebbing, Wicker, & Watanabe, 2016) and basketball (Brewer, McEvoy, & Popp, 2015; Humphreys, 2000). A few studies looked at coaches in professional team sports (e.g., Kahn, 2006). The main reason for this research focus is the availability of salary data, which can be retrieved from public data bases (e.g., Fogarty et al., 2015; Humphreys et al., 2011; Soebbing et al., 2016).

Within these previous studies, a coach's human capital has been measured with age (Fogarty et al., 2015; Kahn, 2006), number of years on the job reflecting experience (Byrd et al., 2013; Grant et al., 2013; Kahn, 2006), and number of years employed in the organization reflecting tenure (Fogarty et al., 2015). Yet, in most previous studies human capital was only used as a control variable, since the focus was more on onfield and off-field performance (Byrd et al., 2013; Fogarty et al., 2015; Grant et al., 2013). A set of formal qualifications (i.e., undergraduate varsity athletic status, type of undergraduate institution, major in physical education, and years of higher education) was only considered by Knoppers, Bedker Meyer, Ewing, and Forrest (1989). Since intercollegiate athletics head coaches in revenue-generating sports share the job characteristics of chief executive officers and command relatively high salaries (Soebbing & Washington, 2011), their salary determinants may be less comparable to those of elite sport coaches in less commercialized sports.

At least two shortcomings can be observed when looking at the body of research examining the relationship between human capital and coaching income. First, research has focused on intercollegiate athletics and--to a smaller extent--on professional team sports, while less commercialized sports including various sports that are at the core of Olympic Summer and Winter Games have not yet been examined. Second, the existing studies predominantly measured human capital with age, experience, and tenure (Byrd et al., 2013; Grant et al., 2013) with one exception (Knoppers et al., 1989), while formal qualifications have been largely neglected. The present study attempts to increase the knowledge base by taking these shortcomings into account.

Methods

Data Collection

Since data on coaching salaries in elite sports are not publicly available--unlike in intercollegiate athletics--primary data had to be collected. An online survey was used for the data collection, which was online from July 17 to August 17, 2015. Since the support of elite sport is taken care of at the federal level in Germany, all elite sport coaches are at least partially funded by the federal government (i.e., national coaches, federal state coaches, and coaches at Olympic training bases). Formal ethics approval for this study was obtained by the university's ethics committee (approval number: 96/2015). This research is part of a larger study examining the location factors of elite sport coaches in Germany.

Due to data privacy issues, emails of coaches could not be made available. Thus, coaches had to be invited by umbrella organizations to complete the survey. An invitation email including a description of the project, the guarantee of anonymity, and the link to the online questionnaire was sent to the Professional Association of Coaches in German Sport (BVTDS) and the German Olympic Sports Confederation (DOSB)--the head organization for organized sport in Germany. While the BVTDS forwarded the invitation email directly to coaches, the DOSB sent an email to the sporting directors of the national sport associations and the directors of the Olympic training bases, who then forwarded the invitation email to the respective coaches within their organization. This sampling procedure ensured that coaches from a variety of sports, regions, and affiliations were invited.

Given the high workloads and relatively low salaries of elite sport coaches (Digel et al., 2010) an incentive of 50 [euro] was provided for taking the time to complete the survey. In light of the incentive, it seemed acceptable to program the survey in a way that respondents were forced to answer all questions, allowing a complete case analysis. Information about income is usually sensitive and, therefore, less likely to be declared; yet, this information is required for the current analysis. Altogether, 233 elite sport coaches participated in the survey. For the empirical analysis, 47 cases had to be removed because of incomplete responses resulting in a final sample size of 186.

Given the sampling procedure where coaches were (had to be) invited via various sport organizations rather than by the university leading this research project, we do not know how many coaches received an invitation email to the survey and, thus, it is difficult to report a response rate. As noted previously, elite sport coaches in Germany are at least partially funded by the federal government and the Federal Ministry of the Interior (BMI), respectively. According to the Federal Office of Administration (2015), a total of 687 elite sport coaches received a (full or partial) salary from the BMI in 2014--this figure represents the total population of elite sport coaches in Germany. This information allows us to report that 33.9% of these 687 coaches clicked on the link and started the survey and 27.0% completed the survey. The completion rate of 79.8% is relatively high, indicating that most coaches who started the survey also finished it.

The number of coaches in this sample is similar to previous studies examining the determinants of coaching salaries (n=185 coaches in Byrd et al., 2013; n=172 in Fogarty et al., 2015; n=184 in Inoue et al., 2012). Yet, previous studies were able to collect panel data because salary data of college football coaches are publicly available. This study shares the challenges of other survey-based studies facing a trade-off between guaranteeing anonymity to the survey respondents and collecting panel data. The latter requires surveying individuals more than once and matching the data sets using a key variable (e.g., name) that allows for identifying the respondents. This key variable requires personal information that would compromise the coach's anonymity. In the present study, collecting panel data was not possible because questions about income are highly sensitive and must guarantee anonymity to the survey respondents.

Measures and Variables

An overview of the variables used in this study is provided in Table 1. In line with previous research (Fogarty et al., 2015; Inoue et al., 2012), it is assumed that the income of coaches is determined by human capital, performance, and organizational characteristics. In the survey, the coaches' personal monthly net income was assessed. As can be seen in Figure 1, the income distribution is highly skewed. Therefore, it is common to use the natural logarithm of income [Ln(Income)], shown in Figure 2, which is closer to the normal distribution (Mincer, 1974).

[FIGURE 1 OMITTED]

[FIGURE 2 OMITTED]

The coaches were asked to list all the formal coaching qualifications they have. Given the variety of existing coaching qualifications, an open question format was used and space was provided for eight different qualifications (only needed by one respondent). On average, coaches claimed 2.3 formal qualifications. Altogether, coaches reported 65 different types of formal qualifications, which could be summarized into the following 11 categories.

The coaching A-License, B-License, and C-License are the standard licenses for coaches in Germany, which are provided by the sport associations (A is higher than B; B is higher than C). Usually, holding a C-License is a requirement for participating in a training course for a B-License, and holding a B-License is a precondition for being eligible to participate in an A-License training course. However, not every coach has the opportunity to enroll in a training course for a B- or A-License; some sport associations not only require coaches to possess the respective lower license, but also to have other requirements related to, for example, years of coaching experience, performance level of coached athletes, etc. Moreover, the number of participants at a training course is limited. Given the limitations in terms of eligibility and limited participant numbers of training courses, particularly the higher licenses can become bottle necks for coaching jobs at sport associations. Typically, specific licenses are required for specific coaching jobs.

Nevertheless, there are exceptions to these rules, which are at the discretion of each sport association. For example, former elite athletes do not necessarily have to obtain all licenses from the bottom up (i.e., first obtaining a C-License, then a B-License, and then an A-License). Individual arrangements are made that allow reducing the period needed to obtain the necessary qualifications, resulting in a "fast track coach qualification for former athletes" (De Bosscher, Shibli, Westerbeek, & van Bottenburg, 2015, p. 295). Consequently, research shows that higher-level coaches are more likely to have international experience as an athlete rather than having completed a higher-level coaching qualification (De Bosscher et al., 2015).

The category Fed_license includes all other licenses that are provided by national sport associations (e.g., skiing coach, tennis coach). Int_License summarizes all coaching licenses issued by international sport associations. At the Coaching Academy of the DOSB, a specific coaching diploma can be obtained (DOSB_Diploma). It represents a job-integrated degree that, however, is not yet considered equivalent to a university degree. All other coaching-related licenses issued by the DOSB are summarized under DOSB_License (e.g., instructor, physical fitness, trainer certificate).

SportSci_Degree measures whether the coach has a university degree in sport sciences. It includes different types of degrees in sport sciences such as undergraduate, postgraduate, PhD, and the previous diploma degree (i.e., a recognized four-year degree before bachelor and master programs were established in Germany). A more detailed examination of these qualifications would be interesting because they differ in terms of the time spent at a university. However, such a distinction is not possible because in the open question many respondents did not specify what type of degree they have; they simply wrote "university degree in sport sciences."

Other_Degree summarizes stated university degrees in other subjects (e.g., medicine, psychology, pedagogy, biochemistry, molecular biology). While the various coaching licenses are qualifications that are only valid in the sport field, university degrees (including those in sport sciences) are also recognized in other fields.

The variable Certificate captures the various types of coaching-related certificates, formations, and vocational trainings that are provided by other organizations (e.g., certified performance specialist, mental coach, barbell coach, life kinetics, neuro-linguistic programming coach, systemic coach, wing wave coach, back therapy training, functional training).

The various non-coaching related formal qualifications are included in the category Other_Qual (e.g., club manager, sport marketing manager, fully qualified groom, referee, nutrition consultant, sport organization manager, educator). In this context, "non-coaching related" means that the reported qualifications are not directly related to sport practice and talent development, but may nevertheless be relevant to the job of a coach as explained earlier (Inoue et al., 2012; Laios, 1995; Martens, 1990). All qualification variables are dummy variables since one coach typically possesses more than one qualification.

In previous research on college football and basketball (Fogarty et al., 2015; Humphreys, 2000; Kahn, 2006; Soebbing, Tutka, & Seifried, 2015), performance was typically measured by career winning percentage. Since the present study includes various types of sports and not only team sports (e.g., alpine skiing, judo, track and field, biathlon, rowing, cycling, handball, basketball, swimming), performance is measured by whether the coach's athletes or teams belong to the Top5, Top10, or Top15 in the world. These categories are mutually exclusive: Top15 means that the athletes are among the top 15, but not among the top 10 or top 5 in the world; Top10 means that the coach's athletes are among the top 10, but not among the top 5 in the world.

This study also includes age (Age) and the number of years working as a coach (Exp) as well as their squared terms (Age_sq, Exp_sq) to control for non-linear relationships. Having previously migrated to another country (Migration) may also be a form of experience and, thus, adds to a coach's stock of human capital. Moreover, this study includes the number of years in the current position, reflecting organization-specific human capital (Years_pos), nationality (German), and gender (Male). Weekly working time (Work_hours) can also affect income; some coaches in the sample do not work full-time. To better reflect the coaching reality, the actual weekly working time was assessed rather than the working hours specified in the contract. We also control for marital status (Married) and the presence of children (Children) because we examine net income, and people who are married and/or have children pay fewer taxes.

Descriptive Statistics

The summary statistics (see Table 2) show 79.0% of the surveyed coaches are males, reflecting the common gender distribution among elite sport coaches (Greenhill, Auld, Cuskelly, & Hooper, 2009). On average, coaches were 43.0 years old and have worked as a coach for 17.3 years, including 8.0 years with their current organization. Most of the surveyed coaches are German (95.7%) and 12.9% have already worked as a coach in another country. Altogether, 83.3% of the coaches are married and 56.5% have at least one child. The high weekly workloads of 48.9 hours on average are similar to previous research (Digel et al., 2010). On average, coaches have a monthly net income of 2,786 [euro]. The relatively high standard deviation (SD=2,557) and the median of 2,200 [euro] indicate that the mean value is biased by some outliers who earn substantially higher incomes (see Figure 1).

With respect to formal qualifications, the results show that 76.3% of the respondents hold an A-License, 24.7% a B-License, and 13.4% a C-License. Typically, coaches only report their highest license. For example, when a coach has an A-License, he would not say that he also holds a B- and a C-License. And, as described earlier, holding a higher license does not necessarily mean that the training courses for all lower-level licenses have been completed. Moreover, it is likely that some elite sport coaches possess licenses from several sports. For example, a triathlon coach can also hold a coaching license in swimming or cycling. This possibility also explains why the proportions of coaches reporting these three licenses exceed 100%.

Fewer coaches have another license provided by a national (4.3%) or international sport association (6.5%). Approximately one third of the coaches possess a coaching diploma (30.6%) issued by the Coaching Academy of the DOSB; fewer coaches hold another coaching-related license (4.8%) issued by the DOSB or have completed other types of certificates, formations, and vocational trainings (7.5%). Altogether, 41.9% of the respondents have a university degree in sport sciences, while 3.8% hold a university degree in another subject. Formal qualifications not directly related to training practice are held by 6.5% of the coaches.

Empirical Analysis

Regression analysis is used to examine the effect of formal qualifications on coaching income while controlling for other potential influencing factors. Regression diagnostics were performed before the analysis. First, the model was checked for the presence of heteroscedasticity by plotting a residual-versus-fitted plot as well as by applying a Breusch-Pagan test. Neither the plot nor the Breusch-Pagan test ([chi square]=2.05; p=0.153) showed evidence of heteroscedasticity. Second, the regression model was checked for multicollinearity using variance inflation factors (VIFs) and correlation analyses. The highest VIF was 2.34 and all correlation coefficients were below 0.6 (with the exception of Age, Age_sq, Exp, and Exp_sq, which naturally show high correlations). Following Hair, Black, Babin, and Anderson (2010), multicollinearity should not be an issue when correlation coefficients are below 0.7 and VIFs below 10.

Altogether, three log-linear models were estimated using ordinary least squares (OLS) with Ln (Income) as the dependent variable. In Model 1, the remaining variables from Table 1 were entered as independent variables. Models 2 and 3 take into account that income levels may differ among sports and associations, respectively. The sample includes coaches from 45 different sports that belong to 36 different national sport associations (e.g., alpine skiing, cross-country skiing, ski jumping, and biathlon belong to the German Skiing Association). To consider sport-specific differences, sport association dummies were included in Model 2. Since the ratio between the number of observations and the number of independent variables must be taken into account in regression analysis (Hair et al., 2006), standard errors were clustered by sport association in Model 3.

Results and Discussion

Table 3 displays the results of the regression analyses. Models 1 and 3 explain 46% of the variation in the dependent variable, while Model 2 explains 56%, supporting the fact that some variation in income can be attributed to the type of sports and sport association, respectively. Overall, the results can be considered relatively robust in the sense that the signs on the coefficients and significant effects are similar across models. The number of weekly working hours has a positive effect on income. The effect of nationality (German) is insignificant--similar to insignificant effects of race and visible minority in previous research (Fogerty et al., 2015; Kahn, 2006). Contrary to previous research reporting an earnings gap between males and females in intercollegiate athletics (Humphreys, 2000; Knoppers et al., 1989), the gender effect is insignificant in this study. Age has a positive effect and age squared a negative effect. Thus, the typical relationship that earnings increase with age at a decreasing rate (Becker, 1962) was also found for elite sport coaches in less commercialized sports.

With respect to formal qualifications, the results reveal that only a university degree in sport sciences has a statistically significant and positive effect on income. The effects of all other formal qualifications such as the coaching diploma and licenses issued by the DOSB; other certificates, formations, and vocational trainings; other university degrees; and all coaching licenses issued by international and national sport associations are insignificant (with the exception of a B-License, which has a significant negative effect in two out of three models).

The negative effect of the B-License may be explained by the bottle neck phenomenon noted earlier. The B-License is the second highest formal coaching license issued by the national sport associations. However, this license may not be sufficient because for some higher-level coaching jobs such as national coaches (which are also associated with higher salaries) an A-License may be required. It is likely that coaches pursue obtaining the higher license, but may be hindered by the limitations in terms of eligibility and participant numbers at training courses.

Several explanations can be advanced for the positive effect of the sport sciences degree and the insignificant effects of most other formal qualifications. First, a university degree is a general qualification that is also valid and recognized in other fields, while coaching licenses, diplomas, and certificates are only valid in the sport field. Possessing a degree in sport sciences may provide coaches with a competitive advantage. Second, coaches who are busy collecting certificates may have less time for their athletes since investments in human capital also require time and energy in addition to monetary resources.

Third, it is likely that the various certificates, licenses, and diplomas are not expected to improve the coaching performance and are, therefore, not reflected in coaching income. Previous research outside of the sporting industry has also documented weak returns to certificates and diplomas (Liu, Belfield, & Trimble, 2015). Thus, the value of these qualifications may be relatively low. Fourth, other university degrees as well as other certificates, formations, and vocational training might indicate that the person is a career changer and has less experience as a coach, which is reflected in the insignificance of these qualifications.

The negative experience effect and the positive effect of the squared term indicate that a coach needs a certain level of experience before experience pays off and gains in income can be expected. This effect may be explained by investments in human capital and associated costs and foregone earnings, respectively. At the beginning of their career coaches may accept lower-level coaching jobs with lower pay to gain experience and invest in their human capital. This may especially apply to career changers who must gain coaching experience at the beginning of their coaching career and may accept a lower income. For example, experience could be gained in assistant coaching jobs through on-the-job training and learning from more experienced head coaches. In line with human capital theory, an investment period with lower earnings is followed by a period with higher earnings.

The experience effect could also be explained by the need of a track record that can reduce uncertainty for potential employers. Elite sport coaches have to prove their coaching abilities through successful athletes. At the beginning of their career, coaches typically train younger or grassroots athletes rather than top international athletes. Such an investment in a track record is necessary to reduce uncertainty for potential employers. While formal qualifications reflect stated coaching knowledge, a track record may be a better signal because it reflects revealed coaching quality. Moreover, the better the track record and reputation of the coach, the higher may be his bargaining power over employers.

Conclusion

This study examined the effect of various formal qualifications on the income of elite sport coaches in less commercialized sports. The results provide evidence that only a university degree in sport sciences has a positive effect on monthly net income, while other formal qualifications including various coaching licenses, diplomas, and certificates issued by national and international sport associations and other organizations have no significant effect. The findings indicate that schooling (i.e., degree in sport sciences) and learning on the job (i.e., experience) are more relevant than further activities that increase the knowledge base (i.e., certificates, diplomas, formations, vocational trainings). The contribution of this study lies in a detailed analysis of formal qualifications and their relationship with coaching income, which has not yet been examined in previous research.

This research has implications for (prospective) coaches. In light of these findings, coaches should invest in a university degree in sport sciences if they want to earn a higher income. The variety of formal qualifications reported in this study indicates that elite sport coaches have invested in different types of licenses, formations, vocational trainings, and certificates that are available; however, they do not pay off and, therefore, it cannot be recommended to obtain these various qualifications if they are not required by the coaching position.

The findings also have policy implications in the sense that sport officials and policy makers should reconsider why various formal qualifications provided, promoted, and requested by sport associations are not reflected in coaching salaries. Given the diversity of skills needed for high performance coaching and the critical role of elite sport coaches for the achievement of international sporting success and related policy goals, the compensation of coaches should reflect their investment in human capital to a greater extent, particularly when some qualifications are necessary for specific positions.

This study has some limitations that can guide future research. First, it is only based on cross-sectional data. Future research should try collecting panel data that allow tracking the development of coaching salaries and their determinants. Second, the present research design should be extended taking the inherent limitations into account. In future research, data allowing a more detailed examination of sport-specific differences that goes beyond the inclusion of sports dummies in regression models should be collected. It would be interesting to see if the determinants of coaching income differ between sports. Moreover, a more detailed analysis of the role of different degrees in sport sciences (i.e., undergraduate, postgraduate, PhD, etc.), which was not possible in this study, should be conducted in future studies. Furthermore, the relationship between social capital and coaching income should be examined for elite sport coaches in less commercialized sports. Third, the present research design should be applied to other labor markets within the sport sector such as personal coaches who can also have various formal qualifications, but also other coaching purposes such as health or weight management.

References

Barros, P. C., & Barros, C. (2005). The role of human and social capital in the earnings of sports administrators: a case study of Madeira Island. European Sport Management Quarterly, 5, 47-62.

Becker, G. S. (1962). Investment in human capital: A theoretical analysis. Journal of Political Economy, 70, 9-49.

Brewer, R. M., McEvoy, C. D., & Popp, N. (2015). Predicting intrinsic value of NCAA Division I men's basketball coaching salaries. Journal of Issues in Intercollegiate Athletics, 8, 74-91.

Burke, K. L., & Johnson, J. J. (1992). The sport psychologist-coach dual role position: A rebuttal to Ellickson and Brown (1990). Journal of Applied Sport Psychology, 4, 51-55.

Byrd, W. J., Mixon, P. A., & Wright, A. (2013). Compensation of college football's head coaches: A case study in firm size's effects on pay. International Journal of Sport Finance, 8, 224-235.

De Bosscher, V., De Knop, P., Van Bottenburg, M., & Shibli, S. (2006). A conceptual framework for analysing sports policy factors leading to international sporting success. European Sport Management Quarterly, 6, 185-215.

De Bosscher, V., Shibli, S., Westerbeek, H., & van Bottenburg, M. (2015). Successful elite sport policies. Maidenhead, UK: Meyer & Meyer Sport.

Digel, H., Thiel, A., Schreiner, R., & Waigel, S. (2010). Berufsfeld Trainer im Spitzensport [The profession of a coach in elite sports]. Schorndorf, Germany: Hofmann.

Federal Ministry of the Interior. (2015). Richtlinien des Bundesministeriums des Innern uber die Gewahrung von Zuwendungen zur Forderung von Bundessportfachverbanden [Guidelines of the Federal Ministry of the Interior for granting benefits for the support of national sport associations]. Retrieved from http://www.bmi.bund.de/SharedDocs/Downloads/DE/Themen/ Gesellschaft-Verfassung/Sport/foerderrichtline_novellierung.html

Federal Office of Administration. (2014). Allgemeine Nebenbestimmungen fur Zuwendungen zur Projektforderung [General rider for granting benefits for project support]. Retrieved from http://www.bva.bund.de/SharedDocs/Downloads/DE/BVA/Zuwendungen/ANBest-P.html

Federal Office of Administration. (2015). Ubersicht bundesfinanzierter Stellen fur Trainerinnen und Trainer bei den Bundessportfachverbanden mit olympischen Sommersportarten, Wintersportarten und nichtolympischen Sportarten [ Overview of federally funded female and male coaching positions in national sport associations]. File provided upon request by the Federal Office of Administration.

Fogarty, K., Soebbing, B. P., & Agyemang, K. J. A. (2015). NCAA Division I-FBS salary determinants: A look at new and amended contracts. Journal of Issues in Intercollegiate Athletics, 8, 123-138.

German Federal Government. (2014). 13. Sportbericht der Bundesregierung [13th sport report of the federal government]. Retrieved from http://www.bmi.bund.de/SharedDocs/Downloads/ DE/Broschueren/2014/13_sportbericht.pdf?_blob=publicationFile

German Parliament. (2014). Forderung des Leistungssportpersonals - Situation der Trainerinnen und Trainer [Support of the elite sport personnel - situation of female and male coaches]. Retrieved from http://dip21.bundestag.de/dip21/btd/18/033/1803394.pdf

Grant, R. R., Leadley, J. C., & Zygmont, Z. X. (2013). Just win baby? Determinants of NCAA Football Bowl Subdivision coaching compensation. International Journal of Sport Finance, 8, 61-74.

Green, M. (2004). Changing policy priorities for sport in England: The emergence of elite sport development as a key policy concern. Leisure Studies, 23, 365-385.

Green, M. (2007). Olympic glory or grassroots development? Sport policy priorities in Australia, Canada and the United Kingdom, 1960-2006. The International Journal of the History of Sport, 24, 921-953.

Greenhill, J., Auld, C., Cuskelly, G., & Hooper, S. (2009). The impact of organisational factors on career pathways for female coaches. Sport Management Review, 12, 229-240.

Gienger, E. (2008). Zur Zukunftsfahigkeit der Spitzensportforderung in Deutschland [The sustainability of elite sport funding in Germany]. Leistungssport, 7(3), 5-12.

Grix, J., & Carmichael, F. (2012). Why do governments invest in elite sport? A polemic. International Journal of Sport Policy and Politics, 4, 73-90.

Gucciardi, D. F., Gordon, S., Dimmock, J. A., & Mallet, C. J. (2009). Understanding the coach's role in the development of mental toughness: Perspectives of elite Australian football coaches. Journal of Sports Sciences, 27, 1483-1496.

Hair, J., Black, W., Babin, B., & Anderson, R. E. (2010). Multivariate data analysis. (7th ed.). Upper Saddle River, NJ: Pearson Prentice Hall.

Hong, E. (2011). Elite sport and nation-building in South Korea: South Korea as the dark horse in global elite sport. The International Journal of the History of Sport, 28, 977-989.

Humphreys, B. R. (2000). Equal pay on the hardwood: The earnings gap between male and female NCAA Division I basketball coaches. Journal of Sports Economics, 1, 299-307.

Humphreys, B. R., Soebbing, B. P., & Watanabe, N. M. (2011). Determinants of FBS head coaches' salaries: An economic perspective. In M. Leach & F. Guven (Eds.), Sports for dorks college football (pp. 145-167). Houston, TX: Sports for Dorks CFB LLC.

Inoue, Y., Plehn-Dujowich, J. M., Kent, A., & Swanson, S. (2012). Roles of performance and human capital in college football coaches' compensation. Journal of Sport Management, 26, 73-83.

James, E. H. (2000). Race-related differences in promotions and support: Underlying effects of human and social capital. Organization Science, 11, 493-508.

Jones, R. (2007). Coaching redefined: An everyday pedagogical endeavour. Sport, Education and Society, 12, 159-173.

Kahn, L. M. (2006). Race, performance, pay, and retention among National Basketball Association head coaches. Journal of Sports Economics, 7, 119-149.

Knoppers, A., Bedker Meyer, B., Ewing, M., & Forrest, L. (1989). Gender and the salaries of coaches. Sociology of Sport Journal, 6, 348-361.

Laios, A. (1995). Administrative responsibilities of the coach. International Journal of Educational Management, 9, 10-12.

Lin, N. (2001). Social capital. A theory of social structure and action. Cambridge, UK: Cambridge University Press.

Liston, K., Gregg, R., & Lowther, J. (2013). Elite sports policy and coaching at the coalface. International Journal of Sport Policy and Politics, 5, 341-362.

Liu, V. Y. T., Belfield, C. R., & Trimble, M. J. (2015). The medium-term labor market returns to community college awards: Evidence from North Carolina. Economics of Education Review, 44, 42-55.

Martens, R. (1990). Successful coaching. Champaign, IL: Leisure Press.

Mincer, J. (1974). Schooling, experience, and earnings. New York, NY: Columbia University Press.

Pigott, D. (2012). Coaches' experiences of formal coach education: A critical sociological investigation. Sport, Education and Society, 17, 535-554.

Pigott, D. (2015). The Open Society and coach education: A philosophical agenda for policy reform and future sociological research. Physical Education and Sport Pedagogy, 20, 283-298.

Sagas, M., & Cunningham, G. B. (2005). Racial differences in the career success of assistant football coaches: The role of discrimination, human capital, and social capital. Journal of Applied Social Psychology, 35, 773-797.

Schulz, T. W. (1960). Capital formation by education. Scottish Journal of Political Economy, 68, 571-583.

Soebbing, B. P., Tutka, P., & Seifried, C. S. (2015). The effect of sanctions on external job promotions of NCAA Division I-FBS college football head coaches. International Journal of Sport Finance, 10, 375-388.

Soebbing, B. P., & Washington, M. (2011). Leadership succession and organizational performance: Football coaches and organizational issues. Journal of Sport Management, 25, 550-561.

Soebbing, B. P., Wicker, P., & Watanabe, N. M. (2016). The effect of performance expectations on total compensation of Division I-FBS head coaches. Journal of Sport Management, 30, 70-81.

Suddeutsche Zeitung. (2013). Deutsche Trainer fordern mehr Gehalt und Anerkennung [German coaches want higher salaries and more recognition]. Retrieved from http://www.sueddeutsche.de/sport/sportfoerderung- deutsche-trainer-fordern-mehr-gehalt-und-anerken- nung-1.1819684

Taylor, M. (2010). Football's engineers? British football coaches, migration and intercultural transfer, c. 1910-c. 1950s. Sport in History, 30, 138-163.

Authors' Note

This research was funded by the German Federal Ministry of the Interior following a decision by the German Bundestag (grant number: IIA-071805/14-16).

Pamela Wicker [1], Johannes Orlowski [1], and Christoph Breuer [1]

[1] German Sport University Cologne

Pamela Wicker, PhD, is a senior lecturer in the Department of Sport Economics and Sport Management. Her research interests include labor economics, non-profit economics, economics of physical activity, health economics, and economics of sport consumption.

Johannes Orlowski, MSc, is a research associate and PhD student in the Department of Sport Economics and Sport Management. His research interests include labor economics, economics of physical activity, health economics, and monetary valuation of intangibles.

Christoph Breuer is a professor of sport management in the Department of Sport Economics and Sport Management. His research interests include labor economics, non-profit economics, economics of physical activity, and sport sponsorship.

Table 1. Overview of Variables

Name              Description

Income            Individual monthly net income (in [euro])
Ln (Income)       Natural log of income
A-License         Coaching A-License (1=yes)
B-License         Coaching B-License (1=yes)
C-License         Coaching C-License (1=yes)
Fed_License       Coaching license issued by a national sport
                  association (1=yes)
Int_License       Coaching license issued by an international
                  sport association (1=yes)
DOSB_Diploma      Coaching diploma issued by the DOSB (1=yes)
DOSB_License      Other coaching license issued by the DOSB (1=yes)
Certificate       Coaching-related certificate/formation (1=yes)
SportSci_Degree   University degree in sports sciences (1=yes)
Other_Degree      University degree in another subject (1=yes)
Other_Qual        Non-training related formal qualification (1=yes)
Age               Age
Age_sq            Age squared
Exp               Number of years employed as a coach
Exp_sq            Experience squared
Migration         Coach has previously worked in another country
                  (1=yes)
German            Nationality (1=German; 0=other nationality)
Male              Gender (1=male)
Years_pos         Number of years in current position
Top5              Coach's athletes are among the top 5 in the world
                  (1=yes)
Top10             Coach's athletes are among the top 10 in the
                  world, but not among the top 5 (1=yes)
Top15             Coach's athletes are among the top 15 in the
                  world, but not among the top 10 or top 5 (1=yes)
Work_hours        Number of working hours per week
Married           Marital status (1=married; 0=other marital status)
Children          Coach has at least one child (1=yes)

Table 2. Summary Statistics (n=186)

Variable          Mean     SD      Min      Max

Income            2,786   2,557    120     22,299
Ln (Income)       7.733   0.624   4.787    10.01
A-License         0.763   0.426     0        1
B-License         0.247   0.433     0        1
C-License         0.134   0.342     0        1
Fed_License       0.043   0.203     0        1
Int_License       0.065   0.246     0        1
DOSB_Diploma      0.306   0.462     0        1
DOSB_License      0.048   0.215     0        1
Certificate       0.075   0.265     0        1
SportSci_Degree   0.419   0.495     0        1
Other_Degree      0.038   0.191     0        1
Other_Qual        0.065   0.246     0        1
Age               43.01   10.63    18        65
Age_sq            1,962   943.0    324     4,225
Exp               17.27   10.06     2        43
Exp_sq            398.9   420.6     4      1,849
Migration         0.129   0.336     0        1
German            0.957   0.203     0        1
Male              0.790   0.408     0        1
Years_pos         8.040   7.578    0.5       42
Top5              0.570   0.496     0        1
Top10             0.237   0.426     0        1
Top15             0.193   0.396     0        1
Work_hours        48.88   14.48     4        80
Married           0.833   0.374     0        1
Children          0.565   0.497     0        1

Table 3. Summary of Regression Results for Ln (Income)

                          Model 1                  Model 2
Variables                  Coef.      Std. Err.     Coef.

A-License                  0.088        0.103       0.105
B-License                 -0.206 *      0.122       -0.084
C-License                  0.114        0.155       0.019
Fed_License                0.268        0.189       0.181
Int_License                0.067        0.177       0.331
DOSB_Diploma               -0.047       0.088       0.035
DOSB_License               -0.025       0.186       -0.115
Certificate                -0.057       0.150       -0.070
SportSci_Degree           0.171 **      0.082     0.256 ***
Other_Degree               -0.042       0.211       -0.246
Other_Qual                 0.124        0.166       0.217
Age                      0.162 ***      0.038     0.158 ***
Age_sq                   -0.002 ***     0.000     -0.002 ***
Exp                      -0.074 ***     0.020     -0.078 ***
Exp_sq                   0.002 ***      0.000     0.002 ***
Migration                  -0.040       0.131       0.023
German                     -0.026       0.209       0.170
Male                       0.086        0.105       0.149
Years_pos                  0.009        0.007       0.013
Top5                       0.145        0.108       0.142
Top10                      0.019        0.119       0.058
Top15                       REF                      REF
Work_hours               0.015 ***      0.003     0.013 ***
Married                    0.156        0.110       0.051
Children                   0.005        0.088       0.073
Constant                 3.542 ***      0.762     4.027 ***
Sport association            No                      Yes
  dummies included
Std. Err. clustered          No                       No
  by sport association

n                           186                      186
[R.sup.2]                  0.463                    0.558
[R.sup.2] adj F            0.384                    0.390
                         5.795 ***                3.316 ***

                                      Model 3
Variables                Std. Err.     Coef.     Std. Err.

A-License                  0.112       0.088       0.103
B-License                  0.139     -0.206 *      0.119
C-License                  0.172       0.114       0.196
Fed_License                0.223       0.268       0.158
Int_License                0.242       0.067       0.142
DOSB_Diploma               0.100      -0.047       0.073
DOSB_License               0.219      -0.025       0.179
Certificate                0.161      -0.057       0.101
SportSci_Degree            0.096      0.171 *      0.100
Other_Degree               0.256      -0.042       0.153
Other_Qual                 0.177       0.124       0.126
Age                        0.041     0.162 **      0.062
Age_sq                     0.000     -0.002 **     0.001
Exp                        0.022     -0.074 **     0.029
Exp_sq                     0.001     0.002 **      0.001
Migration                  0.145      -0.040       0.118
German                     0.250      -0.026       0.147
Male                       0.120       0.086       0.132
Years_pos                  0.008       0.009       0.008
Top5                       0.122       0.145       0.101
Top10                      0.123       0.019       0.093
Top15                                   REF
Work_hours                 0.003     0.015 ***     0.004
Married                    0.126       0.156       0.108
Children                   0.096       0.005       0.109
Constant                   0.866     3.542 ***     1.206
Sport association                       No
  dummies included
Std. Err. clustered                     Yes
  by sport association

n                                       186
[R.sup.2]                              0.463
[R.sup.2] adj F                        0.384
                                     141.1 ***

Note: *** p<0.01; ** p<0.05; * p<0.1; reference category for sport
association is German Canoe Association.