Economic Report Proofreading and Editing

I. INTRODUCTION

‘Beauty Pays!’ Hamermesh (2011) reaffirmed in his writing a seemingly clear truth: appearance affect income, employees that have a relatively better appearance can enjoy higher wages, more extra allowances and better special treatment than their mediocre colleagues. In another research that was done by Rhode (2010) also highlighted the importance of appearance in personal job search and promotion in her writings. In addition to gender discrimination, census discrimination, disability discrimination and racial discrimination, there is external decimation exist in the labour market. Hamermesh (2011) refer it as the ‘Economics of Beauty’.

Since the study of Hamermesh and Biddle (1994), the appearance characteristics has influence on individual employment opportunities and income levels has received the attention with regards to the labour economics. In order to develop it into a more objective approach, height and weight will be the two important and easy to measure the appearance of the characteristics of a particular person. What kind of role they are playing in Australian labour market? This report will attempt to proceed from these two variables to study the ‘beauty economics’ in the Australian labour market.

With the increasingly fierce market competition, appearance discrimination in Australian labour market is gradually highlighted. Often, candidates have to attach their photos together with their resume. Some companies said that peer competition is too intense, so they had to give priority to the job seekers with a good image to maintain a good corporate image. A recent report said that some college graduates even went to the extreme and undertake plastic surgery to enhance their appearance to increase their competitive advantage in the labour market (Thomson, 2014).

Coincidentally, in 2013 August, The Victorian Equal Opportunity Commission reported that many workers have claimed discrimination on their apperence, including weight, tattoos, even hair style. According to their record in the past five years, 96 workers have alleged discrimination on the grounds of appearance. At the same time, there are 107 people lodged discrimination claims on the grounds of obesity with 27 person on being underweight or satisfaction on their height. Nevertheless, body odour was the grounds for two discrimination claims, 22 claims related to for tattoos and piercings and 38 claims on hairstyles. In additional, according to the first Australian study of the financial return for physical attractiveness, a man with good-looking command an $81,750 salary, compared to for men with below-average looks has only $49,600. Some employers even receive warning form job placement agencies that would ‘look you up and down’ and make a hiring decision before listening to ‘what comes out of your mouth’. The recruitment conditions of enterprises reflect the importance of appearance characteristics to employment. A good appearance may be able to enhance the candidate’s competitive advantage, improve job opportunities and lead to a higher salary.

There are two possible reasons for the above problems: first, Australia has not established a very clear legal provision with regards to the appearance discriminatory behavior; second, the appearance discrimination has not yet attracted enough attention from the general public as well as government’s attention, or even acquiesced. Compared with the gender discrimination, census discrimination and disability discrimination, the discrimination of the individual appearance characteristics is less of a concern to the society. Discrimination in the labour market is detrimental to the core concept of social justices and fair, and it may also aggravate income inequality and lead to social incoherent. Such factors could lead to some serious consequences to the society such as crime or bad reputation for Australia as a country.

This report attempts to verify the important impact of the two forms of factors that determines appearance that is height and weight, on employment, and provide empirical evidence for the appearance discrimination in Australian labour market.

This report will be reported in the following three aspects: First, start from the perspective of a personal figure, we study the influence of height, weight and by using the body mass coefficient (BMI) on income. Second, use of quantile regression to analysis the importance of height and weight in different income classes. The OLS regression results only reflect the effect of the appearance feature on the average income. However, the effect of appearance characteristics on income in difference income classes may be different. Therefore, it is necessary to explore the impact of the body on the people with different income. Thirdly, personal weight and height are two inseparable subjective variables that affect individual income. So considering only one variable will lead to errors in the estimation result. Therefore, this report’s study controls the effect of both at the same time.

The structure of this report is as follows: the second part is a simple outline of the existing relevant literature. The third part introduces the measurement model and the selected data. The fourth part is the empirical part, which discusses the influence of height and weight on personal income. The fifth part is the conclusion of this report.

II. Background literature review

The appearance discrimination in the early labour market has been an area of interest, to both sociologists and psychologists. At the same time, economists on the other hand usually think that the characteristics of appearance and personal production capacity are not related (Loh, 1993). However, for more than a decade, economists have gradually focused on the features of appearance, such as the impact of the body figure on wage, and think stature factors can explain why there are differences in wages of employees (Register and Williams, 1990; Loh, 1993). Harmermesh and Biddle (1994) further demonstrate the facets into the wage equation and explored the role of body figure and appearance in the labour market. They found that the nice appearance has a positive effect on the employee’s income, which is also called the beauty premium, and this conclusion was confirmed later by subsequent studies done by Harper (2000) and Johnston (2010).

So, why do the body figure feature cause wage differentials and inequality of employment? One of the explanation is that body characteristics indicate the differences in individual health and ability of a man. Different from the traditional point of view nearly a decade of researches believe that the body is related to the individual’s ability to work, cognition and non-cognitive ability (Bockerman et al., 2010). The wage differential caused by height can be explained by the difference in labour capacity (Steckel, 1995; Lundborget al., 2009). Averett and Korenman (1999) also argue that weight-related wage differences may be related to individual abilities, but they did not test and verify this conjecture due to data limitations. In addition to being related to production capacity, the body may also be related to the individual’s non-cognitive ability, for example, Persico et al. (2004) found that the wage gap caused by adult height can be largely explained by the height of adolescence. The explanation they gave is that the height difference in adolescence will affect the individual’s ability to communicate and then affect the accumulation of human capital. In some recent studies they have shown that height is also associated with cognitive function. Case and Paxson (2008) argue that differences in the standard of living before childhood lead to differences in cognitive function and height between individuals. Therefore, the height of adults reflects the impact of differences in early life on human capital accumulation and show in the labour market in the form of ‘height premium’. However, the results of these studies show that the control of childhood living environment differences still cannot fully explain the wage differences caused by body figure. Even studies have shown that childhood living conditions cannot explain wage differences cause by height (Behrman and Rosenzweig, 2001).

There is another explanation for wage differential and employment inequality. The discrimination in the labour market leads to wage difference and inequality of employment opportunities. Since competency factors cannot fully explain the wage differences caused by the body figure, more scholars tend to believe that discrimination in the labour market behavior is more reasonable (Register and Williams, 1990; Loh, 1993). The results of these researches show that discrimination in the labour market is more pronounced in female employees. The impact of body figure on the wage has gender differences may be due to the mobility of male between different occupations, while female is often limited to a small range of specific industries. In this way, labour market’s ‘punishment’ toward overweight male and female showing respectively in the form of occupational differences and wage differences (Pagan and Davila, 1997). Case et al. (2009) also argue that the wage difference caused by body figure is distorted by the movement of workers between different industries. In order to obtain a more credible conclusion, Rooth (2009) used experimental data further confirmed that the wage difference varies by the body figure was due to discrimination.

8 in every 10 of the existing literature believe that discriminatory behavior exist in the labour market can be originate from firms itself, but it could also come from consumers, or both, and corporate discrimination is often the main reason (Harper, 2000). Another evidence of corporate stature discrimination is the inequality of employment opportunities caused by the body figure. Studies have shown that overweight has a significant negative impact on employment opportunities and moreover it will have a greater impact on female’s employment (Morris, 2007) Mobius and Rosenblatt (2006) explored the problem of discrimination with experimental data. In their experiments, the ability of the ‘staff’ to perform a job would not be related to the appearance characteristics, but the appearance feature still significantly caused the ‘wage’ difference. This conclusion further confirms that the discriminatory behavior of employers is the main source of discrimination against employees.

In recent years, the contradiction between supply and demand in the labour market has provided a realistic basis for the various acts of discrimination in the Australian labour market. Professor Richard Hall, professor of work and organizational studies at the University of Sydney business school, claimed that the appearance of a man was even more prized than performance in the hospitality, and retail sectors. Meanwhile, director Russell Zimmerman of Australian Retailers’ Assocaition believe that stores wanted staff to fit their image, he stated ‘Personality and looks are seen to be much more important than previous experience or their qualifications to do the job’. Many scholars have discussed gender discrimination, census discrimination, and disability discrimination. The problem of discrimination has always been a controversial subject all over the mass media, but less concerned with the academic community. Although, Thomson (2014) explores the plastic surgery popular problem prevalent in current graduated from the perspective of jurisprudence and Case, Paxson and Islam (2008) explored the relationship between appearance and income using British Household Panel Survey data, confirming the existence of ‘height premium’ in British labour market. There are still three insufficient aspects of these research on Australian domestic labour market: First, it is still relatively lacking from the perspective of economic research. Due to the lack of relevant laws and regulations, Australian external discrimination has attracted the attention of the mass media and the legal profession. However, the relevant economic discussion is relatively small.

Second, only consider the height characteristics as the single factor cannot reflect the true representation of the body figure. From the existing research, personal weight and height are two inseparable characteristic variables that affect individual employment and income, which means by not taking into account of the weight factors can lead to estimates errors.

Third, at different levels of income, the impact of the body figure on wages may also be different. Therefore, the conclusions based on the mean regression results are still very limited. Also, if there is discrimination in the labour market, differences in body size may also be manifested as inequality in employment opportunities. Given this, there is a possibility of further improvement on related research.

Appearance discrimination is a kind of implicit discriminatory behavior. However, social contradictions which are due to this kind of discrimination cannot be ignored. On the one hand, appearance discrimination usually brings psychological trauma on discriminated against and even cause some tragic events happen. On the other hand, discrimination in labour market will lead to inequality of income, damage enthusiasm of people’s investment in human capital and hardworking, and even reduce the welfare of society as a whole.

III. DATA AND ANALYSIS FRAMEWORK

1. Model Specification

We will use the two important appearance characteristics which are height and weight to explore the appearance of discrimination in Australia's labor market. In order to avoid discrimination due to gender differences, male and female samples were studied separately to give full consideration to the differences in the effects of different variables on male and female income. Since the appearance discrimination in the labor market can be manifested in the form of wage discrimination based on height and weight, we use the following econometric model to analyze the impact of height and weight on personal income:

Among the model, indicates the logarithmic of income, is a variable that represents personal physical measurement, , , , and are dummy variables respectively representing the individual’s educational level (lower number indicates higher education level), marital status, current living state (when the questionnaire took place), whether smoker or not and whether drink alcohol or not. is the other variables that affects the income, including age, health rate, occupation, etc. The existing literature usually uses two methods to describe the individual’s physical body: one is directly used weight and higher as control variables; the second is use dummy variable based on BMI. The results obtained using these two indicators may be slightly different. This article uses both two indicators to discuss. By using BMI it can reflect whether the individual body is in a normal range, overweight or underweight. According to the department of health of Australian government (Department of Health, 2009), this article defines BMI below 18.5 for underweight, above 25 for overweight, the rest is normal range.

We use the OLS and QR methods to estimate the model. Han et al. (2009) argue that the impact of the figure on income varies with age, gender and career changes. Johar and Katayama (2012) explored the differences in income from different income group. These studies show that the OLS method cannot adequately reflect the distribution relationship between height, weight and income. Therefore, we also adapt the QR method to analyze the distribution of physical body and income in a more comprehensively.

Another way to test the existence of discriminatory behavior in the labour market is to estimate the employment equation. When analyzing the impact of height and weight pm personal employment, we consider the following employment equation:

Where is a binary variable representing the individual working state (when Y equals to 0 that mean it does not work). represents the other control variables, including age and how many children under the age of 7 need to take care, the remaining variables are same as income equation. According to the study of Morries (2007), we used the probit model to estimate the employment of male and female, respectively.

2. Data and Variables

This article uses the Household, Income and Labour Dynamics in Australia (HILDA) wave 15 (2015) survey data, which is the currently available and most recent cross-sectional data of household-based panel study. HILDA began their survey in 2001, covering more than 17000 Australians who lives within all eight states of Australia each year, and has been conducted 15 times so far. The survey collected information on respondents’ demographics, socioeconomic and health levels. An important feature of this set of data is that each survey has professional physicals examination and record the height and weight of the respondents and also other important information, such as mental state evaluation. These provide relatively accurate information for our study on the relationship between individuals’ body and personal income (HILDA, 2017).

A total of 17606 individuals were interviewed in the wave 15. According to Fair Work of Australian Government (Fair Work Ombudsman, 2017) an employee who are ages under 21 years old is a junior and should treat with junior pay rates. Thus we removed samples who are aged under 21 years old (1743 observations). Considering retirement, the difference between pension income and wage income, probability hump and height decrease cause by osteopathy of the elderly, we will also exclude male individuals older than 59 years old and female individuals older than 52 years old (5629 observations) based on Australian 2014-15 average retirement age published by ABS (ABS, 2017). Individual height, weight are the two main variables in this research, so we will remove the samples who are missing their variables (1855 observations). We also removed samples that did not provide employment status information (1584 observations). The final sample in this research contains 6795 observations, of which 2983 female and 3912 male.

The main income information collected by HILDA includes “Main job, current weekly gross wages & salary ($)”, “All jobs, current weekly gross wages & salary ($)” and “Gross financial year wages and salaries ($)”, here we use “All jobs, current weekly gross wages & salary ($)” as income variable. Information of height (cm) and weight (kg) is measure by professional medical staff and are therefore high accuracy. Other demographic and socioeconomic characteristics include age, self-assessed health, highest education level achieved, marital status, occupation, industry, country of birth – brief, etc. In addition, HILDA also collected important life and mental variables such as hours spending on different activities, relationship status and psychological tendencies of personal control. Table 1 gives a descriptive statistic of the relevant variables.

The average weekly wage income of the male samples is about $1422. Female weekly wage income is much lower than male, which is about $962. The average yearly wage income of the male samples is about $75927. Female yearly wage income is also much lower than male, which is about $49227. Considering that the income is usually a lognormal distribution, we use the logarithm of weekly wages. The average height of male and female respondents in the sample are approximately 179 cm and 166 cm, respectively, with an average weight of 88 kg and 72 kg, respectively. There is no big difference in the average age and highest education level achieved in average for both genders.

Figure 1 shows the height and the corresponding weekly wage distribution, from which we can generally see the relationship between height and wage income. From the scatter chart, male’s height and the corresponding average income distribution is more dispersed, while the females is relatively concentrated. This may be due to the greater mobility and variety of occupations choices available for male in different industries, while female may choose from a fewer variety of occupations compare to male. From the fitting point of view, male’s height and wage income showed a positive correlation; female’s height and wage income slightly positive related, but this trend is not very obvious. Figure 2 depicts the relationship between the BMI and corresponding average weekly income. Without regard to other influencing factors, male’s wage income is slightly positively related to BMI, but female’s wage income seems unrelated with BMI. Compare with Figure 1, it can be found that both relationships between height and income or the relationship between BMI and income, male employees in the sub-samples have a more obvious trend.

IV. MODEL ESTIMATION

1. The Impact of The Body Figure on Income

a) OLS regression estimation

We limited the sample to all observations which were not missing either income, height or weight information, including 3432 males and 2783 female. Table 2 is the estimated result of the female income equation. In model 1 of Panel A, we only control individual education level, lifestyle, marital status, and states dummy variables, and in addition, since as individual’s age increases, the body’s metabolic rate will decline, which cause in general, the weight will increase with age, therefore we consider the interaction between age and BMI index. Model 1 results show that for every 1 cm increase in height, on average the weekly wage income will increase 0.3%. The coefficients of ‘underweight’ and ‘overweight’ variables are all insignificant at any level, which indicate there is no enough evidence to support the relationship between abnormal weight and female personal wage income. Another factor to consider is that wage differences due to height or weight may be caused by health differences (Steckel, 1995). Therefore, Model 2 considers the self-assessed health status of respondents, the estimated result shows a negative relationship between health status and weekly wage income, which means compared with self-rate ‘fair’ and ‘poor’, ‘excellent’, ‘very good’ or ‘good’ health condition significantly increased wage income. Compared with Model 1, the estimated coefficient of height in Model 2 is almost unchanged, while the coefficient of underweight and overweight are still insignificant. This shows that in Australian labour market, the female individual wage difference caused by height is not due to health differences. Another possible explanation is that individual physical characteristics may distort personal career choices, industry and occupational differences may also explain ‘height premium’ or ‘punishment on underweight or overweight’ (Pagan and Davila, 1997). After controlling the dummy variables of occupation and industry in Model 3(8 types and 19 types respectively), although the estimated coefficients of height, overweight and underweight are just slight variation from previous model, these three coefficients are all now insignificant, which represents there is no clear evidence to say height or abnormal weight can lead female’s wage difference. In Model 4, we further control the individual’s activities time, which can reflect the individual’s ability to work. In addition, height and weight are related to the individual’s cognitive ability, which is usually determined by the living environment of the early years (especially childhood) (Case and Paxson, 2008). In order to further control these influencing factors, we control the dummy variables of the birth country in Model 5, which can control the difference reflected in the living environment at a different birth group. In Model 6, we added psychological tendencies of personal control as dummy variables which can reflect the individual’s self-cognitive. Finally, people with high levels of education may be more conscious of their own body, the reason being health to them would rank higher than other activities that might be unhealthy. In addition, the employer’s attitude towards employee with different education level may also be different. In order to examine whether there is a difference in the impact of the body on the individual at difference levels of education, we consider the interaction between overweight and education in Model 7.Even after added a lot of variables with fully considered, the estimation of Model 7 still shows insignificant on all three main variables’ (height, overweight and underweight) coefficients. Different from the clear positive relationship of height with weekly wage income in previous Model 1 and Model 2, Model 7 indicates that the data does not support our external discrimination hypothesis. In simple terms, based on our regression, we cannot find any relationship between height or abnormal weight and female’s wage income.

In Panel B of Table 2, we replaced the BMI with weight, and the control of those remaining variables is exactly the same as the variables control in the corresponding model in Panel A. Model 7 shows that in the case of other factors unchanged, female’s weight for each additional 1 kg, the wage income will increase by 0.5%; and there is in sufficient evidence to say that height variation will cause female’s wage income difference.

Our estimation results show that under different models (expect Model 1) using weight and height variables, female’s body weight is an important factor in explaining the difference of female wage, while female’s height only affects income independently under Model 1 (with BMI variable) and Model 2 (with BMI variable). This shows that there is an obvious preference for the body weight of female employees exist in the Australian labour market. However this does not mean female employees with abnormal weight will be treated by discriminatory wages, according to above estimations analyses, our data and models does not show any relationship between abnormal weight and wages difference.

Table 3 is the result of male income equation estimate, and its structure is the same expressed in Table 2. According to Table 3 Panel A, in the Australian labour market, male’s height and overweight have more effect on income (‘height’ coefficients are all highly significant across all 7 model using BMI variable; ‘overweight’ coefficients are highly significant within first 6 model using BMI variable). Based on Table 3 Panel B, similar as female’s height, the impact of male’s height is not clearly significant as well (only significant in Model 3 to Model 6), indicating that the male’s height cannot show a clear ‘height premium’. There was a significant positive correlation between male’s weight and income, which indicate slim male will have a lower income. Model 7 shows that control of health, education, occupation, birth groups and other variables, there is not enough evidence that height affect the income of male, but weight still has a positive effect on wage income. The above shows that Australian labour market has an obvious preferences on male’s appearance.

By comparing Table 2 and Table 3, we can find out that the labour market in Australia is more demanding on the weight for male than female. Male’s height and weight-related wage differences can be explained primarily by the health and ability factors. While the differences in wage income caused by female’s weight cannot be explained entirely by these factors, it indicates that in additional to these factors, the discrimination in the labour market on female’s appearance (especially weight) is also the cause of female wage income differences (Loh, 1993).

b) Quantile regression estimation

The OLS estimation results reflect the impact of height and weight on income at average income levels. However, if the impact of height and weight on income is different at different income levels, the information that can be presented by the average income is very limited. Johar and Katayama (2012) found that there were significant differences presented in the effect of appearance characteristics on income at different sub-sites. In order to reflect the distribution of income on height and weight more accurately and comprehensively, we use the quantile regression method to reanalyze that income equation of male and female. For simplicity reason, the quantile regression only takes into account the Model 7 in Table 2 and Table 3. Table 4 shows the result of the quantile regression. The table contains only the height, weight and BMI’s effect on income while the regression results of the rest of the control variables are not reported. We selected 5%, 10%…, 95% of the 19 points to return, Table 4 reported regression results of seven of them (respectively, five odd points and two even points). These seven results are sufficient to reflect the trend of regression.

Table 4 Panel A is the effect of female height and BMI on income. All coefficients of the three main variables (overweight, underweight and height) across the seven points are insignificant. These means there is insufficient evidence to say abnormal weight or height variation can cause female’s wage difference in any income level. The Panel B is the effect of female height and weight on the wage income. Different from Panel A, Panel B shows that weight does has impacts on female wage income in middle-income and higher-income classes. With the rise of income level, the impact of female’s weight on wage income showed wave curve trend. When the wage income is in middle, the coefficients of weight is 0.3% and relatively slightly significant (only significant at 10%). After the wage income climbed to 70% point, the weight has a higher impact on wage, which is 0.6% and highly significant. For higher-level income class, weight significant leads 0.7% and 0.5% increment in female’s wage for who earn wage in 80% and 90% income level. These results show there is not enough evidence to say that female wage in the lower-income class is affected by the personal figure, while the female wage in the middle-income class is influenced by the weight significant and the female wage in the higher-income class affected by personal weight most. Which means higher weight can help female who in the upper layer gain more in wages. This may be due to the fact that female at higher-income level may engage in executive labour, which require a stable image for female because they are the representative of the firm. Meanwhile female in the middle-income level are often engaged in physical labour or technical labour, such occupations may have more emphasis on physical and professional skills and the appearance requirments are relatively low compare to executive labour. However, the coefficient is lower than 1% and therefore have little contribution to our conclusion. Panel C and Panel D are male income regression results. There are two points that we should mainly focus on: First, based on Panel C, underweight only affects male’s wage at 10% level of income, males with underweight in 10% income level bear wage punishment with 34.3% less than normal male’s wage income; overweight only has impact on male’s wage at 30% income level and cause a 11.1% increment in wage; while, height just makes slight change to males’ wage if they are in 20%, 50%, 80% and 90% income level (0.3%, 0.2%, 0.3% and 0.3% influence respectively). Second, according to Panel D, male’s weight has significant positive impact across all level of income and these effects do not change much. The male’s quantile regression result analyze shows that in the low-income class, robust physique is one of the factors that increase male income, therefore, male with underweight suffering wage punishment, while in the middle-income class, having a moderate size weight may be more favorable to raise male’s salaries. This may be due to the low-income class of male often engaged in physical labour since weight to a certain extant reflects the male’s ability to work, or there may be due to low-income industries preference on male’s weight.

c) Corporate Discrimination and Consumer Discrimination

Discrimination that exist in the labour market may arise from businesses, but may also arise from consumers (Harper, 2000). In order to further illustrate the discrimination behavior in Australian labour market arise from consumers rather than businesses themselves. We consider the following wage equation:

is a 0-1 dummy variable, and is 1 when the occupant’s occupation need to be in direct contact with the consumer. If discrimination comes from the consumer, employees who works in the front line that have contact with the consumer are more likely to be discriminated against, and other employees that works in the backend are less likely to be discriminated against, or even do not suffer such discrimination at all, at this time, is significantly less than 0 and is equal to 0. HILDA does not contain information on whether respondents are in direct contact with consumers at work. As service workers are often more likely to be in direct contact with consumer, we use whether the respondents’ occupations belong to the service industry (salespeople, attendants and community, etc.) as a proxy variables that determine if they have direct contact with consumers. Table 5 is the estimated result of the above equation, all of the interactive variables’ confidents are insignificant from 0. This is no clear evidence that discrimination in Australian labour market comes from consumers, therefore, any discrimination against both gender is mainly due to the behavior of the firm itself.

2. The Impact of Body Figure on Employment

If discriminatory behavior comes from the enterprise, then the physical characteristics will also lead to unequal employment opportunities. Thus, the employment equation can also be used to test the appearance discrimination in the labour market. Table 6 and Table 7 are the results of estimating the employment formula using the Probit model. The Model 5 regression result in Table 6 shows there is insufficient evidence to say that female’s appearance has any influence on female employment though any aspect. For the Model 5 regression result in Table 7, it can be seen that male’s height has a significant positive impact on individual’s employment, however, we still cannot find clear relationship between male’s weight and employment possibility. Model 5 in both table also take into account the interaction between overweight body figure and education level. Unfortunately, we cannot find any evidence to support our hypothesis that with the improvement of education level, overweight may has some impact on the possibility of female or male employment. The impact results of other variables on employment in Table 6 and Table 7 are basically in line with our expected. For example, a higher level of education increases the employment possibility for female; female’s birth country has significant impact on employment which may due to gender and racial discrimination; female who got married lose employment opportunities at 74.1% level than female in other marital status; male full of self-confidence has 16.3% more chance to find a job, yet if male is overconfidence or even arrogant, they will lose nearly one-third of the chance for catching a job.

3. Further Discussion and Robustness Check

a) Income adjustment

Weekly wage income may not adequately reflect corporate discrimination against height and weight. Corporate discrimination may also be expressed in the form of bonuses. HILDA also investigated respondents’ gross financial year wages and salary, which combines basic weekly wage and yearly bonues. We re-estimate the Model 7 of Table 2, Table 3 and Table 4 with the logarithm of the gross financial year wage income and list the result in Table 8. Compared with Table 4, the basic results of Table 8 did not change significantly, but there are still some minor changes. In Panel B, female weight is significantly positive at almost all income level expected 10% level (compared with Table 4 Panel B, weight only significant at 50%, 70%, 80% and 90%). For the male, Table 8 Panel C shows more significant large negative impact (54.9%, 37.3%, 17.6% and 34.7% at 20%, 30%, 50% and 70% respectively) from underweight on male’s gross financial year wage and income (Table 4 Panel C only indicates there is negative relationship in 10% income level). Male’s height’s positive impact also becomes significant in nearly all gross financial year income level (expected 10% and 90% level). The OLS results are almost same from the Model 7 in Table 2 and Table 3, only the underweight in Panel C for male change to significant in Table 8. In short, from Table 8 we can get a similar conclusion, that is, businesses have significant discriminatory behavior against both male and female employees’ weight, meantime, male in lower middle income level is facing ‘height premium’ and ‘overweight punishing’.

b) Age group test

There may be a discrepancy between discriminatory behaviors among different age groups (Han et al., 2009). In order to examine this, we divided the sample into three parts by age (21 to 29 as low age, 30 to 44 as medium age and 45 plus as old age), and progress the OLS regression, respectively. Table 9 shows the result of the regression of each age group. Panel A indicates only the medium age’s male’s height has slight (0.5%) significant impact on income, while there is not enough evidence to show that underweight, overweight and height in any other female or male age group can cause income difference. Panel B shows height and weight impact on income, their impact on income has become very weak after we treated these two variables independently. The result shows only the low age’s males’ weight has impact on the income as one kg increase in their weight leads to 1.5% increment in wage income. Height and weight did not significantly affect the income of both genders in all other age groups. All in all, corporate discrimination is more pronounced in the male’s low and middle age groups. This may be due to males in these two age groups is facing a more intense occupational competition and mobility, the appearance characteristics are more likely to become the bargaining chip.

V. Conclusion

The problem of appearance discrimination in the labour market in Australia has not yet received enough attention from the general public as it should. Discrimination in appearance seems to be more accessible to the public than gender discrimination, census discrimination, disability discrimination and racial discrimination. However, in any way, discriminatory behavior in the labour market can lead to a decline in the overall level of social welfare and a less desirable outcome for the entire economy. If the appearance discrimination is not fully addressed formally, or even acquiesced, the appearance discrimination in Australian labour market may be become more and more intense over time as the competition in the labour market intensified.

This report explores the existence of appearance discrimination in the labour market in Australia from the two characteristics of body figure, which are personal height and weight. We found that Australian labour market is more demanding on male’s appearance characteristics, on the other hand female is relatively tolerant on appearance characteristics. For both gender overweight and underweight does not clearly affect the individuals employment opportunities or the income level (since main coefficients are insignificant, we cannot find enough evidence to support our abnormal weight discrimination hypothesis). However, when we look at personal height, it is clearly that male’s height has significant positive influence on either their income or employment, but this relationship does not exist on female’s height. After controlling for health, education, occupation and other factors’ variables, only middle-upper income level’s female’s weight has weak effect on their income, and at different levels of income, the effect of weight on income is not variable much. The case for male is similar, for all income level, there is only obviously positive relationship between male’s weight and their income (the impact amount also very slightly and lack of variation). In additional, for male employees in 10% income level, underweight person bear 34.3% lower income than employee with normal weight and male in 30% income level can gain 11.1% more under overweight status. This shows that business do have some preferences on individual’s appearance characteristics across both gender.

This preference for appearance in the labour market means that the competition in labour market is unfair and hence it will leads to negative outcome to the market and society. In order to resolve the current situation, we can start working from the following two aspects: the first is to strengthen market regulation, eliminate gender discrimination. More rigid regulations should be put in place to ensure equality, as well as encourage victim to report such act by firms or individuals. Often victims do not know how to deal with such treatment by the others. In some instance victim did not know that they can report such unfair treatment, the reason being the publicity is not enough. By doing so it can improve the liquidity between industries for both gender. More advertising or campaign could be run by government to get public’s attention to such event. The second way is to enhance the society overall education level, by doing so it can have several benefits. Firstly, by increasing the general education level the impact of personal technical capacity can replace the impact brought from appearance differences. The reason being now that people are more educated and equip with higher skill level appearance will not be the most important factor. On the other hand, with a higher level of education can also make society better recognize and aware that appearance discrimination is a wrong behavior. This is because generally as people are more educated they can better understand the moral values and any kind of discrimination is not adequate. Certainly, these solutions cannot be put in place overnight and the problem with discrimination cannot be gone overnight as a result. It needs a long-term process to improve the market’s fair competition mechanism, improve the society’s perception and finally change the business preferences of appearance.

The study of this report is only a preliminary discussion based on Australian labour market in the appearance discrimination, there are some issues that still need to be further explored and addressed in details. For example, why the labour market in Australia is more demanding on the appearance of male, what is the reason behind it? In addition, due to the limitations of the data, there are still some shortcomings exist in this report. For instance, the role of social capital in the process of employment and income determine cannot be ignored. For data reasons, this report does not control social capital variables, which can be the direction for further research topics in the future.

19 / 20