Introduction

The number of overweight adults worldwide has almost tripled since 1975 (WHO 2018). The World Health Organization (WHO) has estimated that by 2016 more than 1.9 billion adults were overweight, of which 650 million were considered obese (WHO 2018). Thus, 39 percent of the adult world population are overweight and 13 percent are obese. In Germany, 54 percent of adults are overweight and 18 percent suffer from obesity (Schienkiewitz et al. 2017).

Socio-economic environmental factors in general, income, and level of education are important for the development, consequences, and management of obesity (Lampert et al. 2018). Obesity and its secondary diseases, which in turn cause the majority of disease costs, seem to follow an inverse social gradient (Marmot 2005): People with high socio-economic factors are less likely to develop obesity or secondary diseases than people whose income or school-leaving qualifications are low. For this reason, in recent years, socio-economic status (SES) has often been the subject of empirical studies on the occurrence and consequences of obesity. The SES usually consists of the individual factors, education and income, whereby the parameter school leaving certificate is used to explain level of education and different income groups are created in order to compare to an average income.

The problem with many existing studies on the links between obesity, income, and educational level is that they only deal with the prevalence of the disease without considering the consequences. Moreover, various studies are based on inadequate data collection or involve only a specific population group. A further problem is the partial lack of differentiation between obesity and 'normal' overweight, which leads to elementary differences in the results. Against this background, the following questions will be addressed in this paper: (1) is it possible to identify a correlation between BMI, income, and a person´s educational background? and (2) do income and years of education, as parameters for professional success, differ depending on the BMI? The study uses data from the Socio-Economic Panel of the reference year 2016.

Theoretical background on obesity and labour market success

Obesity: demarcation and economic consequences

Obesity is defined as an increase in body fat above the normal level, and complex interaction of different genetic and environmental factors is characteristic. Thus, obesity is primarily seen as a risk factor for the development of further diseases (WHO n.d.). The classification of obesity is usually based on the body mass index (BMI), since this allows a simple determination, no dependence on gender and age, but in return a correlation with body fat mass is given (CDC n.d.). Furthermore, a differentiation between normal and underweight as well as between different forms of obesity can be made according to the WHO classification based on the BMI:

The consequences of obesity, like its causes, are multifactorial (Han et al. 2011). They can be of physical, psychological, social, and economic nature, and have negative effects on the quality of life (Jiang et al. 2013). Overall, obesity can be associated with various secondary diseases and a high level of health impairment. From an economic perspective, it can also increase the costs of the country's health system (Klein et al. 2016; Biener et al. 2017). Effertz et al. (2016) estimate the total operationalisable costs at 63 billion EUR per year, of which 33.65 billion EUR are attributed to indirect and 29.39 billion EUR to direct costs. Furthermore, each insured person with a BMI of over 30 kg/m2 causes additional costs of EUR 166,911 (men) and EUR 206,526 (women) on average in Germany (Effertz et al. 2016). Overall, studies (e.g., Withrow and Alter 2011; Dee et al. 2012; Koenig et al. 2015) show that the costs for medical care are higher for obese people and that their state of health is worse than that of people of normal weight.

Determinants of labour market success

Long-term effects of physical activity on health, subjective well-being, and various labour market variables were examined by Lechner (2009) in a comprehensive study. Over a period of 16 years, the results show that high physical activity increases income by about 1200 euros per year compared to very low activity. Based on these results, Lechner (2015) developed a model with the variables health, physical appearance, human capital and soft skills of a person, which influence success on the job market (Fig. 1):

  • A high level of physical activity has a positive effect on a person's state of health (Murray and Lopez 1997). Consequently, low physical activity is associated with negative health effects. Poorer health with corresponding complaints leads to an increase in health care costs (Katzmarzyk and Janssen 2004). It can be assumed that a person who is frequently and regularly physically active will have a better state of health than those who exercise little. As a logical consequence, an employee's state of health also influences his or her productivity at work, which can be measured by a lower number of days of absence or higher performance (Lechner 2009).

  • In addition to the positive effects on health, physical activity also influences the physical appearance of a person. According to Sierminska (2015) and Harper (2000), employers expect higher productivity from more attractive people, which leads to higher employee compensation even with the same productivity. This phenomenon is known as Thorndike‘s halo effect. Otherwise, especially in professions where regular customer contact is necessary, it can be seen that employees who are perceived as attractive by the customer are more productive than others (Pfann et al. 2000). Accordingly, it can be expected that the productivity and also success of a person on the labour market are greater if he or she is perceived as attractive by his or her environment.

  • Soft skills describe non-specialised competences that describe qualifications that are not typical for the profession. Soft skills are to be understood as a collective term for various characteristics and abilities that contribute to professional success (Schulz 2008), and include personal, social, or methodical competencies such as the ability to work in a team or organizational competence (Schulz 2008; Robles 2012). These soft skills are significantly influenced by physical activity and exercise (Lechner 2009; Bailey 2015). Soft skills also have a positive influence on a successful career and in life in general (Heckman and Kautz 2012).

  • Human capital is to some extent related to the soft skills and health status of a person. Human capital, as the performance potential available in trained and qualified persons (Swart 2006), is determined by physical and intellectual factors. Thus, soft skills and state of health fall under human capital just as much as education, physical performance, and the (professional) experience of an individual. Due to the increasing importance of intellectual work, human capital is increasingly defined as the personal knowledge of the individual employees of a company and thus focuses more on the intellectual factors (Swart 2006).

Fig. 1
figure 1

Sports activity and success in the labour market (Lechner 2015)

Full size image

Various studies show that higher physical activity can have positive effects on the job market. Kari et al. (2016) show in this context that those who were already frequently and regularly active in sports during their childhood earn more in later life. Ewing (2007) also shows that former high school athletes earn more in working life than people who did little or no sport. Barron et al. (2000) provide similar results, which also show that the educational level of people with higher physical activity is higher. In addition to better wage prospects and a higher level of education, athletic people also have greater chances of finding a job through application procedures than less athletic applicants (Rooth 2011).

Combining these findings with the basics of obesity, the answer to the question whether obese people have poorer chances of success in the labour market is obvious, among other things because of their lower level of physical activity and the associated possibly poorer state of health. Based on income studies, this assumption can be proven in part for England and the USA. Sundmacher and Morris (2010) and Harper (2000) show that obesity significantly reduces the wages of female employees, while the same cannot be said for men. The reasons for these economic disadvantages of obese people are primarily based on the described discrimination by employees by employers and customers, as well as a lower human capital (Hammermesh and Biddle 1994; Koenig et al. 2015). Whether this income disadvantage of obese people compared to normal-weight persons can be proven in Germany is not conclusively clarified. For this reason, the focus below is on a relationship between health based on a person's BMI and income and years of education as measurable factors of labour market success.

State of research

Income and health

With the proven and quantifiable burdens in the form of direct and indirect costs (see theoretical background), it should be noted that different health conditions cause varying costs, among other things due to different utilisation of medical services and different care needs, which in turn depend to a large extent on a person's income (Baltagi and Moscone 2010). From a health and socio-economic perspective, income therefore plays a central role in the development of obesity, in coping with its consequences and the associated costs, and in a person's general state of health. Among other things, income forms the basis for social integration and quality of life. Income disadvantage affects a person's ability to consume, which in turn can have a negative impact on other areas, such as health (Baltagi and Moscone 2010).

In low-income strata, the incidence of diseases and complaints and the risk of premature death is higher, and the quality of life in terms of health is worse (Schiller et al. 2012). Various studies also show a negative correlation between BMI and income (Kuntz and Lampert 2010; Lampert et al. 2013; Koenig et al. 2015; Kjellberg et al. 2017). Instead of personal income, existing studies tend to use income groups or household income across several individuals. Overall, individual income has hardly been taken into account so far, and it is usually the correlation between prevalence and income that is concerned.

Obesity in the educational context

In the literature, the factor income is often linked to the factor education in order to investigate the connection between the social environment and the economic circumstances of a child and its health. With regard to children and adolescents, the SES is usually calculated by researching the educational level of the parents, their occupational status and their income, in order to be able to draw conclusions about the social environment in which the children grow up in (Lampert et al. 2018). In addition, the literature shows that children's health behaviour is primarily determined by their family and social environment (Gadsden et al. 2016) This could also be the reason why obesity often occurs in childhood and adolescence, but is not diagnosed until later in life, when the consequences manifest themselves (Güngör 2014). Insofar as obesity already develops in childhood and socioeconomic factors play an important role, it seems reasonable to include educational status in the further course.

Kuntz et al. (2018) prove an increased probability that a child will receive fresh fruit daily if the SES is also high. On the contrary, the probability that a child will consume sugar-sweetened refreshment drinks every day increases six times for children with low SES compared to the group with high SES. Thus, not only is the nutritional behaviour dependent on the parents' education and income, but also the children's leisure-time physical behaviour (Kuntz et al. 2018; Sanyaolu et al. 2019). Thus, the prevalence of overweight and obesity in children and adolescents might be traced back to the school-leaving qualifications of their parents: The higher the school-leaving certificate, the lower the probability of an excessive BMI in children and vice versa (Devaux and Sassi 2013; Amis et al. 2014; Ruiz et al. 2016). According to Hruby and Hu (2015), the prevalence of obesity in adults follows an educational gradient in all age groups and regardless of gender. In most cases, school-leaving qualifications are used as a marker for the level of education. It has been shown that the educational background of parents has a strong influence on a child's school career.

Gender-specific differences

Mensink et al. (2013) show clear differences between men and women in the prevalence of overweight. 43.8% of men have a BMI that is equivalent to being overweight, but only 29.1% of women. This ratio is reversed when considering the prevalence of obesity: here the percentage of women is higher than that of men. Furthermore, the frequency of the level of obesity varies between the sexes. In the male gender, only 3.9% can be assigned to class II obesity and 1.2% to class III obesity. For women, however, these values are 5.2% and 2.8% respectively. Thus, more women suffer from massive obesity, whereas more men are overweight (Mensink et al. 2013). There are also fundamental differences depending on income regarding the link between obesity and graduation, while more men with an above-average income develop obesity than high-income women (Kuntz and Lampert 2010; Devaux and Sassi 2013; Aitsi-Selmi et al. 2014). In terms of educational level, women show a higher prevalence of obesity than men in every age group in the lower educational groups (Schienkiewitz et al. 2017). In addition, according to a Federal Health Cost Account for the reference year 2008, the direct costs caused by obesity and other over-nutrition in Germany are higher for women than for men (Destatis 2015). Gender differences can also be demonstrated in the additional health care costs incurred by obese individuals over the lifespan compared with normal-weight individuals, e.g., for medical care (Effertz et al. 2016, see Theoretical background).

Based on these results, gender-specific income should also be considered. In 2016, a woman with a gross hourly wage of 16.26 euros earned on average around a quarter less than a man (20.71 euros) (Destatis 2017). In addition, income disadvantages for women in other countries can be documented in relation to their BMI (Harper 2000). The factor obesity does not seem to play a role for men, so that a differentiation by gender is also appropriate in this study. Conversely, when comparing educational levels in terms of gender differences (Destatis 2014), the proportion of women who aspire to and achieve a higher level of education is higher than the proportion of men. However, the proportion of men in schools with a lower level of education is significantly higher.

Methodical approach

Research questions and hypotheses

Based on the existing findings, the present study aims to determine the correlation between income and BMI for Germany. By looking at income, a comparative statement regarding the success on the labour market in different BMI groups is possible. By means of the SES, the socio-economic factors of educational background and gender are taken into account in addition to income, since hardly any study deals with the educational background without considering the respective school graduation. Therefore, the total number of years of education is used as a variable in this study. A fundamental problem of the study situation is the partial lack of differentiation between obesity and 'normal' overweight, which leads to differences in the results. This is because obesity cannot be equated with 'normal overweight' from a medical and economic perspective due to resulting health consequences and health costs. Furthermore, the consideration of individual income in connection with obesity is missing.

  • H1: BMI, income, and years of education correlate with each other: based on the study results in the state of research, there is a negative correlation between the factors BMI and years of education, and BMI and income.

  • H2: Within the BMI group, BMI, income, and years of education correlate. As in hypothesis 1, there is a negative correlation between the factors BMI and years of education and BMI and income.

  • H3: Income and the number of years of education differ significantly depending on the BMI group affiliation: in line with the study results in the state of research, the number of years of education and the associated income decreases as the BMI increases.

Methodology

Data basis, inclusion criteria, and relevant parameters

The survey is based on data from the German Institute for Economic Research (DIW) from 2016. Since 1984, the socio-economic panel has been conducting annual surveys on topics such as vocational and school education, physical activity and health behaviour, family status, origin, and net income. Both Lavie et al. (2019) and Lechner (2015) provide evidence of a link between physical activity and the possible development of overweight and obesity, as well as an influence on labour market success. However, the socioeconomic panel does not collect data on sporting activity and therefore cannot be included in the analysis. The DIW provided the data in SPSS format.

The data set included the respondents from all participating households. The further statistical analysis included only people of working age (between 18 and 65 years) at the time of the survey who were employed in monetary gainful employment. Since success in the labour market can be primarily depicted by income, people without their own earned income were not taken into account, in order to avoid a distortion of the results. Data lacking information on required parameters were also not considered. These criteria reduced the study population from 47,998 to 14,196 subjects.

The parameters BMI, income, and years of education relevant for the statistical analysis could not be used without restrictions as stored in the data set:

  • For the calculation of the BMI, the height was converted into meters. In the next step, the calculation of the BMI was carried out with the inclusion of the weight and height of the participants as a further variable.

  • The questionnaire allowed a division into income from self-employment, from a part-time job, and from the main job. Since total income is of interest in this analysis, all types of income were added to form a new variable total income.

  • The variable years of education of individuals refers to the number of years of education and includes years of schooling, years of training, and years of study, depending on the individual’s career.

  • The classification of the groups into obese, normal weight, and overweight was carried out according to WHO guidelines (see Table 1). The further subdivision according to the severity of obesity is not considered in this study.

Table 1 Weight classification based on the BMI of adults (WHO 2000)
Full size table

Data analysis

The statistical analysis was performed using SPSS25 (IBM Corporation). Initially, an explorative data analysis and a test for normal distribution was performed. Subsequently, correlation calculations were performed between the parameters BMI, total income, and years of education for the collective and the three groups. After analysis of variance, t-tests were performed on independent samples to detect differences in salary and/or years of education between the BMI groups.

Results

Sampling and descriptive data analysis

The 14,196 test participants are composed of 6,791 men and 7,405 women. The average age is 44.14 years. In both groups, the youngest participants are 18 and the oldest 65 years old. Descriptive statistics for the variables income, years of education, and BMI are determined with the help of explorative data analysis: The mean values of the relevant factors were 26.23 for BMI, € 31,339.41 for annual income and 12.85 years of educationFootnote 1. The Shapiro–Wilk test showed no normal distribution in any of the cases. Consequently, non-parametric tests had to be selected for further statistical calculations.

BMI, income, and years of education

Correlations are calculated using the Spearman Rho test (see Appendix Tables 12, 13, 14 and 15) . Relative to the total population, a highly significant, low positive correlation can be found for the variables BMI and income [r (14,196) = .079, p = .00]. A significant, slightly negative correlation can be found for the variables BMI and number of years of education [r (14,196) = − .159, p = .00].

Correlation calculations within the obese group show highly significant, weakly negative correlations between the BMI and the income of individuals [r (2672) = − .085, p = .00] and between the BMI and years of education [r (2672) = − .067, p = .00]. Similar results were obtained in the group of overweight subjects: The BMI and the number of years of education of an overweight subject have a slightly negative correlation [r (5043) = − .061, p = .00]. Finally, the group of normal weight people is checked for correlation: In this subcollective, the BMI and the number of years of education are weakly negative [r (6481) = − .048, p = .00] and the BMI and income are weakly positively correlated [r (6481) = .16, p = .00].

Evaluation according to BMI groups

The Kruskal–Wallis test shows that significant differences (p = .00) exist between the groups obesity, overweight, and normal weight (see Table 17) for the two variables income and the number of years of education. The Mann–Whitney U test was used for the further calculations, as there was no normal distribution. The individual BMI gradations show a significantly (p = .00) higher income in the overweight (MRank = 3,994.35) group than in the obese group (MRank = 3,600.65) (see Tables 18 and 19). According to Cohen (1992) there is a weak effect. When comparing the normal weight group (MRank = 4,534.56) with the obese group (MRank = 4,679.94), the latter show a significantly (p = .02) higher income than the normal weight group, with only a very small effect (see Tables 20 and 21). The overweight (MRank = 6,189.8), on the other hand, have a significantly (p = .00) higher income than the normal-weight (MRank = 5,430.01), but with a likewise only weak effect (see Tables 22 and 23). The corresponding mean values of annual income are shown in the following Fig. 2.

Fig. 2
figure 2

Mean values of total income by BMI groups

Full size image

The years of education are significantly (p = .00) lower in the obese group (MRank = 3,526.95) than in the overweight group (MRank = 4,033.4) (see appendix Table 18). According to Cohen (1992), this is also a weak effect. Compared to the normal weight group (MRank = 4,891.93), a significantly lower (p = .00) number of years of education in the obese group can also be demonstrated (see appendix Table 21). In contrast, the number of years of education of the normal-weight group (MRank = 6,027.71) is significantly (p = .00) higher than that of the overweight group (MRanK = 5,421.67) (see appendix Table 22). The following Fig. 3 shows the corresponding mean values of the individual groups:

Fig. 3
figure 3

Mean values for number of years of education by BMI groups

Full size image

Gender-specific differences

A Mann–Whitney U test without weight-related categorisation shows a significantly (p = .00) higher income among men (see Table 2). According to Cohen (1992), this corresponds to a medium effect.

Table 2 Gender-specific differences (own analysis of SOEP data)
Full size table

In contrast, women have significantly (p = .00) more years of education than men (Table 3) :

Table 3 Statistical analysis of gender-specific differences (own analysis of SOEP data)
Full size table
  1. a)

    Gender-specific differences in the normal-weight group (Tables 4 and 5)

Table 4 Years of education and income normal-weight people (own analysis of SOEP data)
Full size table
Table 5 Statistical analysis normal-weight people (own analysis of SOEP data)
Full size table

According to the results, men of normal weight earn a significantly (p = .00) higher income in their jobs than women in this group, although they have significantly (p = .00) fewer years of education.

  1. b)

    Gender-specific differences in the overweight group (Tables 6 and 7)

Table 6 years of education and income overweight people (own analysis of SOEP data)
Full size table
Table 7 Statistical analysis overweight people (own analysis of SOEP data)
Full size table

A similar picture as with normal-weight people is also seen with overweight ones:

Accordingly, overweight women have a significantly (p = .00) lower income than men, despite a significantly (p = .00) higher number of years of education.

  1. iii)

    Gender-specific differences in the group of obese people

Gender-specific differences also exist in the obese group (Tables 8 and 9) :

Table 8 Educational years and income of obese people (own analysis of SOEP data)
Full size table
Table 9 Statistical analysis of obese people (own analysis of SOEP data)
Full size table

Obese men achieve a significantly (p = .00) higher income than women in this BMI group. There is no significant difference in the years of education (p = .23).

The analysis of the BMI groups thus confirms the existence of gender-specific differences. Overall, the BMI classification produces the following results on income and the number of years of education:

Fig. 4 shows that in the present evaluation the group of overweight persons has the highest monetary success in the labour market, given the mean income and the associated statistical significance. Normal-weight individuals have the lowest mean significant differences in gender earnings.

Fig. 4
figure 4

Income depending on gender and BMI

Full size image

There is a negative correlation between the number of years of education and the allocation to BMI classes. In the normal-weight group, there is an above-average number of years of education compared to the obese and overweight classes. Consequently, the results in Figs. 4 and 5 are not confirmed, since a higher average number of years of education would also be expected to result in a higher income among the normal-weight individuals. Overall, on the basis of the present evaluation, no statement can be made about a causal relationship between a person's BMI, income, and years of education.

Fig. 5
figure 5

Years of education depending on gender and BMI

Full size image

Discussion

BMI and years of education

With regard to the number of years of education, it is apparent that the number of years invested in education decreases as the BMI rises. This is confirmed both when looking at the test person selection and when comparing the BMI groups: People of normal weight have more years of education than overweight people, who in turn have more years of education than obese people. This means that there is not only a difference between normal and overweight people, but also between overweight and obese people. These results are basically consistent with various studies on the assumption that the frequency of obesity increases with a decreasing level of education, even if the graduation was used as a benchmark (Kuntz and Lampert 2010; Lampert et al. 2013; Schienkiewitz et al. 2017). This correlation raises the question of causality: is it because of a high BMI that fewer years of education are decisive or is a high BMI developed because of a low number of years of education? In view of a possible primary genetic cause of obesity, which manifests itself already in childhood, the first approach seems to be correct. If the argumentation of current literature that proves an inverse association between BMI and learning performance (Cournot et al. 2006) is followed, the assumption that a person has fewer years of education in the course of his life due to his already early above-average BMI is confirmed. If, on the other hand, socioeconomic environmental factors are seen as the cause, it is more likely that people have different chances of having the same number of years of education because of their social position, environment, and material resources, and therefore develop obesity (Woessmann 2004). Overall, it can be said that the number of years of education is closely related to a person's BMI. The results show that obese people are at a disadvantage compared to people of normal weight.

BMI and income

The results show a correlation between income and BMI for obese people: with increasing obesity the level of income decreases, as Kjellberg et al. (2017) showed. With regard to the entire population and the group of normal-weight people; however, it can be determined that people with a higher BMI earn more. This correlation is surprising but could be explained by the group size: among the normal- and overweight group, there were almost twice as many participants as in the obese group. The negative correlation among the obese could therefore have had no influence on the population. However, for the present evaluation, in summary, the (partial) hypotheses regarding a negative correlation between BMI and income earned must be rejected.

With regard to average income, the overweight have an advantage over the obese. In contrast, the income of the normal-weight group is on average lower than that of the other two groups. This result confirms numerous studies that linked an increasing prevalence of obesity with a decreasing income (Kuntz and Lampert 2010; Lampert et al. 2013; Devaux and Sassi 2013; Aitsi-Selmi et al. 2014). However, it cannot be concluded that people of normal weight earn better than those who are overweight or obese. Possible explanations could lie in the development of the labour market structure: The steady growth of the IT sector and the general digitalisation of work are associated with the possibility of earning a high income, for example, regardless of social skills, physical appearance, or physical ability. In addition, in many professions with high physical demands, such as tradesmen or nursing staff, income is rather low. Overall, the sedentary lifestyle, the constant availability of food, and the enormous progress in basic medical care (Freese et al. 2017) appear to have had a major influence on the health behaviour of the population in recent years. If this trend continues, an increase in the prevalence of overweight and obesity would be expected: increasing proportions of the population with overweight and obesity would presumably be accompanied by a further increase in demand for medical services and an overall increase in health-care expenditure.

Otherwise, it would have to be clarified whether someone already is obese and then starts his or her professional career or whether obesity develops over the course of one's working life. On the one hand, greater financial freedom goes hand in hand with higher income: the more monetary resources are available, the more of them can be spent on one's own health without triggering restrictions in other individual areas of expenditure. This could suggest that people are still of normal weight when they start work and that obesity only develops during the course of their working life. On the other hand, it has already been explained that obesity can be genetically determined. This in turn would mean that these people were already overweight or obese at the beginning of their working lives. Since obesity is a multifactorial disease, the exact time of its manifestation is difficult to determine and a statement regarding causality is limited. In order to be able to further investigate these two factors, it would be interesting in the context of follow-up studies to find out in which professions the prevalence of obesity is particularly high.

Further decisive factors in the discussion of income and BMI are the occupational prerequisites and the determinants of labour market success according to Lechner’s model. In most of the studies presented here, success on the labour market is determined by the income of a person (Lechner 2015). According to the available results, overweight and obese people are more successful than those of normal weight. A possible consideration here is that the factors soft skills and human capital are only conditionally dependent on a person's BMI and physical activity. The logical consequence would be that both factors have a much greater influence on the success on the labour market than the physical appearance and state of health. This would once again confirm the influence of the changes in the professional world in recent decades, according to which intellectual human capital is increasingly taking on greater significance than physical capital. However, even this idea is only justifiable to the extent that a person's state of health does not impair cognitive abilities and soft skills. This factor would probably be the boundary between overweight and obese people. The relevance of soft skills is confirmed by the results, and will become even more important for future generations in order to distinguish themselves from other workers (Dean and East 2019). The higher income of the obese and overweight compared to those of normal weight can also be discussed in terms of time, which is not considered by Lechner. Due to the excess of calories, it can be assumed that obese and overweight people have less physical activity than normal-weight people. Thus, they invest a smaller part of their limited resource of time in exercise, sports and physical activities. On the one hand, this is time that more inactive people could invest elsewhere, such as in working time, and on the other hand, it is time that can have a negative impact on productivity at work due to excessive training. The question also arises as to where the health capital gained through physical activity is reinvested. Is it invested in more working time, which would be beneficial from a purely economic point of view, or is it invested in leisure time, which may result in a loss of productivity? Lechner's model thus presents the determinants of labour market success in a highly simplified theoretical form. In contrast, the interplay between physical activity and professional success appears to be much more complex in relation to practice and the present study.

Finally, the time horizon should be considered. Although the results show an income advantage of overweight and obese people compared to normal-weight people, it has already been worked out that obesity and its secondary diseases can have a long history of illness. In the case of obese people, this is possibly accompanied by an increase in the number of days absent from work. Overall, the number of working years would therefore be reduced. When considering the productivity of this group in relation to the entire working life, a decrease in income for those affected would have to be expected (Goettler et al. 2017).

Gender-specific differences

This paper also demonstrated significant differences depending on the gender of a person in terms of income and number of years of education: if a differentiation is made according to the BMI groups normal weight, overweight and obesity, it is shown that in all groups the male gender receives on average almost twice as much income as the female. This is confirmed by data from the Federal Statistical Office (Destatis 2014), according to which women in Germany earn less than men per hour. The existence of the "gender pay gap" is also proven for the year 2016. A potential explanation for the high income differences could be the employment relationship of the present respondents, which was not differentiated in more detail in this paper. Depending on the type of employment and the amount of work, basic income differences can be assumed. For example, in Germany in 2018, 7.2 million women and only 1.9 million men worked part-time (Bundesagentur für Arbeit 1992).

At the same time, the number of years of education within the BMI groups is exactly the opposite: women have more years of education in the normal-weight and overweight groups than men. This education-related effect is also evident regardless of BMI classification. Only within the obese group do the years of education of men exceed those of women. This is in line with data from the Federal Statistical Office (Destatis 2014) and shows that the number of years of education and school-leaving qualifications depend on gender, among other things. Developments in adolescence and early adulthood and differences in specific cognitive functions between the two sexes appear to be among the decisive factors (Güngör 2014). It is questionable why gender is the only category not playing any role in the number of years of education in the obese group. One possible explanation would be that the obese already have fewer years of education than the other two, which means that less variation is possible within the groups and therefore less apparent effects being demonstrated.

In addition, it should be mentioned that gender differentiation has a greater effect on income than on a person's years of education. This can be explained by the fact that both genders have identical access to educational opportunities. Otherwise, there is evidence from the Federal Discrimination Office (2018) that women earn lower wages than men across all industries in Germany (Antidiskriminierungsstelle des Bundes 2018). Furthermore, Zucco (2019), among others, show that women generally receive lower wages than men with the same education and work experience.

Socio-economic disadvantage of obese people

People with obesity have fewer years of education than the other two groups and a lower income than overweight people. Consequently, the (partial) hypotheses regarding a negative correlation between BMI and the number of years of education can be confirmed. This monetary disadvantage for obese people can be reflected in different access to and use of medical care (Alberga et al. 2019). If the salary is lower, more attention should be paid to spending, possibly in the areas of health, nutrition, or sports. In addition, the BMI can be a barrier to entry for occupations: a further disadvantage for obese persons may be their appearance. Here, depending on their perceived attractiveness, fundamental differences in income can be demonstrated, and a potential bias on the part of the employer and his associated expectations of the employee's productivity also play a major role (Pfeifer 2002). It is also worth reiterating that obese people are more likely to have fewer years of work than people of normal weight, which can only have long-term negative effects. It is possible that a lower number of years of education in the obese group means a certain knowledge gap in terms of strengthening health resources and the ability to act in health-related areas. This disadvantage becomes more serious when applied to Lechner's model: if health suffers from the lower number of years of education, this also results in a lower productivity of those affected, which may be decisive in the labour market (Decker and Decker 2014). This factor can be decisive for career opportunities and advancement compared to overweight people.

Many disadvantages can be attributed to obesity, at least in theory, although practical evidence would still have to confirm this. According to the results of this study and of Kjellberg et al. (2017), it is shown that especially with a BMI above 30 kg/m2, further gradual gradation is decisive, as this is reflected in income and years of education.

Limitations

Due to the essential relevance of multiple causal and diagnostic factors, the diagnostic approach in particular limits the informative value. Even if BMI is considered a standard classification, it is not a guarantee for an error-free diagnosis of obesity. This is demonstrated by Davillas and Benzeval (2016), who state that the results of a study are strongly dependent on the underlying classification. Thus, a combination of BMI and a measurement of fat distribution pattern, for example, would be a sensible alternative for future investigations. In this way, classifying athletes with a high weight and a high muscle percentage wrongly among obese people can be avoided. Furthermore, the fact that the number of subjects in the obesity group was only half as big as in the other two groups could distort some of the results presented. These include the positive correlation between BMI and income for the entire sample and the non-significant gender-specific number of years of education in the obese group. The high-income gap between men and women could be due to women´s higher share of part-time employment, as mentioned above. This limits the meaningfulness of the results and suggests that an additional subdivision should be created after the partial and fulltime employment relationship for the future. Further, the question of individual income is problematic, from which also the quality criteria can suffer. The result depends both on the level of knowledge about one's own income and on the validity of the data, since the information is usually based on personal data of the individuals questioned (Andersen and Mayerl 2019). This makes it possible for the information to be based to a certain extent on social desirability and a better reputation.

Furthermore, in view of the proven correlation between physical activity and health (Lavie et al. 2019, among others) and between physical activity and labour market success (Lechner 2015), physical activity should also be taken into account in the future in order to be able to make statements about occupational success. Contrary to existing assumptions on a negative correlation between BMI, the number of years of education, and thus income (similar Devaux and Sassi 2013; Amis et al. 2014; Ruiz et al. 2016), this evaluation proves a better income level in the overweight group. However, this is a cross-sectional design, which means that it is not possible to make statements about causality, but rather that all statements are to be understood merely as correlations. Rather, it seems possible that BMI rises due to a higher income, since, for example, less free time remains due to the occupation in a higher position, or also more financial resources are available for care. In addition, other factors such as professional experience, additional training, and the type of job could have a decisive influence on the income earned. In this context, it should also be mentioned that the age of the participants was not taken into account. This, too, could be evaluated as a relevant factor for the income level with a view to increasing human capital. Overall, an assessment of labour market success was primarily based on the monetary criterion of income, with qualitative aspects not being relevant in the socioeconomic panel.

Conclusion

To analyze the influence of the determinant health on labour market success, correlations between BMI and the number of years of education as measurable factors on individual earnings were examined. In summary, this study was able to show the income and education differences between men and women, which decrease with increasing BMI. In addition, correlative relationships between BMI, income, and number of years of education were demonstrated and the socioeconomic disadvantage of obese people in terms of income and education was shown. If success in the labour market is measured on the basis of income, the group of overweight people achieves the highest income compared to normal-weight and obese people. However, the results should not be interpreted as causalities, but in the future other factors will be included in the analysis of labour market success.

The inconsistent results between normal and overweight people require further investigation. In addition, a longitudinal study would be of interest in this context, which could be used to make a statement with regard to the causal relationship. In the present case, a longitudinal study was initially dispensed with, since no reference to individuals was possible on the basis of the data. This circumstance precluded an analysis of concrete developments and associated causalities. By comparing socio-economic panel data from previous years, it would only have been possible to make a comparison showing changes over time. In this context, however, an increase in the wage level must also be taken into account. Rather, in an industrial society with an average increase in BMI in Germany, an evaluation of current data appeared to be relevant.

Furthermore, Kjellberg et al. (2017) already reported a dependence on the degree of obesity. With reference to the year 2016, more detailed results could be presented by further differentiating the severity of obesity.

The strain on one's own health associated with obesity can have far-reaching consequences affecting one's daily life, social interaction, and professional career, and could thus be a major socio-economic disadvantage. Therefore, further preventive and curative approaches to obesity need to be developed. For many countries and their health-care systems, obesity already is a serious, costly, and momentous problem that requires even greater attention if long-term social and economic consequences for the (national) economy are to be contained.