Looks and Labor - Do Attractive People Work More?I

Looks and Labor - Do Attractive People Work More?$

Markus Gehrsitza,b,

aThe City University of New York, 365 5th Avenue, New York, NY, 10016, USA bCologne Graduate School in Management, Economics and Social Sciences, Richard-Strauss-Str. 2,

50931 Cologne, Germany

Abstract

Using the German General Social Survey (ALLBUS) 2008, I investigate how looks affect an individual's labor supply decision. My results are, by and large, in line with predictions derived from the neoclassical model of labor supply. Applying regular probit, bivariate probit and craggit regression models, I find that good looks go hand in hand with higher employment probabilities and more hours of market work. Furthermore, physical attractiveness is positively associated with spousal income and spousal employment. Hence, beauty appears to affect labor supply decisions both directly and through the marriage market.

Keywords: Beauty, Labor Force Participation, Working Hours, Gender Differences

1. Introduction

Social and evolutionary psychologists have long known that beauty matters. They argue that physical appearance is one of the most important determinants of human mating preferences (Buss, 1989). Economists have also discovered the subject and shed some light on the economics of beauty. They have found that physically attractive persons are perceived as more trustworthy (Wilson and Eckel, 2006) and are granted loans on better terms (Ravina, 2008). They are also more likely to receive an invitation to a job interview (Rooth, 2009), are regarded as more competent by employers (Mobius and Rosenblat, 2006), and consequently enjoy higher incomes (Judge et al., 2009, among others). All their relative economic advantages appear to culminate in a relatively higher life satisfaction (Hamermesh and Abrevaya, 2011). The homely, on the other hand, lack preferential treatment. They have lower incomes and tend to marry spouses with less educational attainment if they find a mate in the first place (Harper, 2000).

In the labor market, it has been shown that physically attractive men and women on average earn higher wages in the United States (Hamermesh and Biddle, 1994), Australia

$I am deeply indebted to Dan S. Hamermesh who provided the impetus for this paper and gave great advice throughout the writing process. I would also like to thank David A. Jaeger for valuable comments and suggestions. I wish to thank Philipp Doerrenberg, Christian Weyand, and the participants of the CGS dissertation seminar for comments on an earlier version of this paper.

Corresponding author Email address: gehrsitz@wiso.uni-koeln.de or mgehrsitz@gc.cuny.edu (Markus Gehrsitz)

Preprint submitted to Elsevier

July 24, 2012

(Leigh and Borland, 2007), and the UK (Harper, 2000) among other countries. Two major explanations have been provided for this wage premium. First, employers might simply have a taste for beautiful workers (Becker, 1971). That is, an employer might derive more utility from a good-looking employee than from an equally productive but plain employee. Such taste-based employer discrimination will result in a wage gap between attractive and unattractive workers. In a perfectly competitive market such differences cannot prevail. Companies employing homely, but equally productive workers should be able to squeeze rivals that are staffed with attractive employees out of the market. Hence, it is likely that beauty has a productivity-enhancing effect that justifies pay differentials. This is obvious in occupations, such as modeling, where beauty is an integral part of the product or service itself. If customers prefer to interact with physically attractive employees, this also explains part of the earnings differential. For instance, a goodlooking salesperson might be more convincing. The fact that good-looking professors receive higher evaluation ratings by students could also reflect customer discrimination (Hamermesh and Parker, 2005). In a similar vein, coworkers might prefer to interact with pretty colleagues and handsome supervisors. For example, Pfann et al. (2000) find that revenues of Dutch advertising firms rise commensurate with the attractiveness of their CEOs.

Discrimination, be it statistical, taste-based, or based on customers' and coworkers' preferences, is a phenomenon that predominantly affects labor demand. There has been remarkably little research, however, with respect to the role of beauty in labor supply decisions. This paper intends to fill this gap in the literature. Economic theory predicts that physical attractiveness is associated with higher employment rates, that attractive workers supply more hours of work to the labor market, and that beauty should affect men and women, and married and single individuals, in a different manner. My empirical analysis is based on the German General Social Survey (ALLBUS) of 2008 (for details see Wasmer et al., 2010). The survey provides the first nationally- representative data set that contains beauty measures for all respondents since the 1981 Canadian Quality of Life study. Using this unique data source, I find that physical attractiveness is indeed associated with higher employment probabilities and a larger labor supply, especially for men. Beauty also appears to affect labor supply via the marriage market. Husbands of physically attractive women tend to be more likely to work fulltime jobs, and earn higher incomes. Spousal income is generally negatively associated with women's labor supply. Yet, this negative income effect is offset by better employment prospects for physically attractive women. Overall, beauty is positively associated with labor supply in many different ways.

The structure of this paper is as follows. Section 2 discusses whether universal beauty standards exist, what makes a beautiful person and how beauty is measured. Section 3 derives theoretical predictions about the relationship between beauty and labor supply. After giving a brief overview on the dataset and the variables at hand, I will empirically test these predictions in Section 4. Section 5 is concerned with the robustness of the results while the concluding Section 6 discusses implications of my findings.

2. Beauty Standards and Measures

Unless there are standards of beauty that are commonly agreed upon, any analysis of the economic effects of beauty is futile. Yet, as Hamermesh (2011) points out, the notion

2

that beauty is "in the eye of the beholder" turns out to be both widespread and false. Admittedly, beauty standards differ across cultures. While Western celebrity magazines glorify female slimness, force-feeding of young girls is still common in Polynesia and some Arab countries (Harter, 2004). Within the Western hemisphere, beauty standards have changed dramatically over time. The voluptuous beauties that were ever-present on the theater stages of the 19th century would probably not stand a chance of landing the lead in most of today's Hollywood productions (Stearns, 2002). However, beauty standards change very slowly over time. Catherine Tramell, the seductive serial killer played by Sharon Stone in the 1992 movie "Basic Instinct" is probably still deemed attractive today.

Table 1 - Distribution of Ratings...

...by Rating Method: Beginning of Interview End of Interview

Self-Assessed

Above Average Looks Average Looks Below Average Looks

862 34.04% 1382 54.65% 285 11.27%

921 36.42% 1348 53.30%

260 10.28%

591 23.37% 1514 59.87% 424 16.77%

...by Sex of Interviewer: Male Interviewer

Above Average Looks Average Looks Below Average Looks

466 31.59% 849 57.56% 160 10.85%

Female Interviewer

396 37.57% 533 50.57% 125 11.86%

...by Sex of Respondent: Male Respondent

Above Average Looks Average Looks Below Average Looks

382 30.44% 715 56.97% 158 12.59%

Female Respondent

480 37.68% 667 52.35% 127 9.96%

...by Age of Respondent: Respondent < 40 Years Respondent 40 Years

Above Average Looks Average Looks Below Average Looks

398 42.31% 457 49.73% 64 6.94%

464 28.82% 925 57.45% 221 13.73%

All ratings had originally been submitted on a scale of 1 (=unattractive) to 11 (=attractive). The average rating at the start of the interview is 7.71 with a standard deviation of 1.89. The category Above Average Looks comprises all respondents with a beauty rating of 9 or higher. The category Below Average Looks comprises all respondents with a beauty rating of 5 or lower. Consequently, the category Average Looks contains all respondents with a rating larger than 5 and smaller than 9.

Data support this claim. Patzer (1985) is among the first to find that individuals tend to have similar judgments about what makes a beautiful person. Ritts et al. (1992) support this hypothesis in their analysis of student perception by teachers. Hamermesh and Biddle (1994) compare ratings of respondents' attractiveness in different survey

3

waves by different interviewers, and find that the interviewers mostly agree with each other. In a different study, four different observers had to rate law students' pictures on an attractiveness scale of 1 to 5 (Biddle and Hamermesh, 1998). Again, the interviewer ratings were highly correlated. In their meta-analysis Langlois et al. (2000) conclude that raters agree on a universal beauty standard, not just within cultures, but also between cultures. Although observers do not always exactly agree, there is rarely large disagreement about looks. That is, interviewers have the same standards and it is mostly differences in looks that lead to differences in ratings, rather than differences across the raters.

This has important implications for this study. When standards of attractiveness are more or less universal, and persistent over a sufficient time horizon, they can also affect labor supply decisions. Hence, the mere existence of such standards is a prerequisite for physical appearance to influence labor supply decisions. Of course, the context in which beauty is assessed, affects how beauty is rated by different beholders. In the ALLBUS 2008, interviewers were asked to rate participants twice on a scale from 1 (unattractive) to 11 (attractive). I deem those who received a rating of 9 or higher to have aboveaverage looks and categorize those with a beauty rating of 5 or lower as below-average looking.1 It is important to note that the interviewers were not asked to focus on facial attractiveness, but took the overall physical appearance into account. They had to render their first rating based on their first impression, that is before they had even started the actual interview. At the end of the session, the interviewers had to submit another rating which tended to be slightly more flattering (see Table 1).

One reason for this pattern is the "halo effect". That is, the perception of a respondent's attractiveness might be influenced by other impressions the interviewer gains over the course of an interview. For example, nice and polite participants might be rewarded a higher score at the end of an interview, while rude respondents are subject to penalties. In order to account for endogeneity bias arising from the halo effect, I use the initially assigned score for all analyses in this paper. Generally, the initial and second assessment are highly correlated with a correlation coefficient of 0.838. Furthermore, people were asked to rate their own looks during the course of the interview. On average, they were much harder on themselves than a third person, namely the interviewer (see Table 1). I also find differences in ratings depending on the interviewer's and the respondent's sex and age. Female interviewers tend to give slightly more generous ratings. While this is consistent with Hamermesh (2011), I cannot confirm his finding that women's looks are rated more extreme. Instead I find that women are attributed above-average looks more frequently than men. Women also seem to receive low ratings less often, although a simple t-test provides only weak evidence for this hypothesis (p-value: 0.037). Generally, women appear to be rated more favorably than men. Moreover, I find evidence for the impact of the ravages of time on beauty. Respondents who are more than 40 years old, are significantly more likely to be attributed below-average looks and significantly less likely to be categorized as above-average looking (see Table 1). Youth and beauty are clearly positively correlated.

Overall, it is safe to assume that there is a universal beauty standard that most interviewers agree on. These standards are very stable, even over long periods of time.

1The average rating is 7.71 with a standard deviation of 1.89 (see Table 1). Thus, a person whose rating comes under more than 1 standard deviation from the mean is categorized as unattractive.

4

Variations in beauty assessment, for instance in terms of interviewer or respondent gender and age, can easily be accounted for by stratification and including appropriate control variables.

3. Theoretical Predictions

In a standard neoclassical model of labor supply (e.g. Bryant and Zick, 2006) individuals divide their time between market work, household work and leisure depending on their marginal productivity and the real hourly wage. Due to productivity or discrimination, an attractive individual will be offered higher wages than an otherwise identical peer. This has been repeatedly shown (Fletcher, 2009, among others) and also holds true for the data set at hand.2 As with any conventional wage increase, this will induce three effects. First, there will be a production-substitution effect. Holding the total working time constant, an increase in the wage rate provides an incentive to substitute house work for market work. This, of course, assumes that beauty does not affect the nonmarket productivity. Second, leisure becomes relatively more expensive. Therefore, the individual is induced to substitute leisure time for work time in order to buy additional goods. This is the consumption-substitution effect. Both substitution effects lead to an increase in market labor supply. Third, a wage raise for physically attractive individuals will also raise their income. Since both leisure and consumption goods, are normal goods, the demand for either will increase. Consequently, the income effect counteracts the two substitution effects. Only if the substitution effects dominate the income effect, physically attractive people will supply more labor than unattractive individuals.

So, theoretically, higher wage rates, induced by an attractive physical appearance, can affect labor supply either way. Yet, empirical studies tend to find positive wage elasticities for men and even more for women (Heckman and Killingsworth, 1986). As a result, I expect attractive individuals to have higher employment rates and - conditional on employment - to supply more work hours. I also expect the effects of attractiveness and labor supply to differ across married and single individuals, and between men and women. The literature in evolutionary psychology concurs that individuals prefer attractive mates and value good financial prospects as well as traits that signal earnings capability, such as intelligence. The latter qualities appear to be of particular importance for female mate selection while men attach more value to physical attractiveness (Fisman et al., 2006, Hitsch et al., 2010). Put differently, good looks can be traded for income in the marriage market. If attractive individuals are more likely to marry affluent spouses, this induces an additional (negative) income effect. As a result, I expect the effect of beauty on labor supply to be less pronounced for married individuals than for single individuals, and less pronounced for women than for men. In the following section, I will test these predictions empirically.

2An increase of two standard deviations in the beauty rating is associated with an increase in net hourly wages of about 14 percent for women in fulltime employment and about 11 percent for men. In line with Hamermesh and Biddle (1994), I find that it is particularly attractive women who are rewarded and unattractive men who are punished in terms of their hourly wage rate. The results of the wage regressions are available from the author on request and will be published in a different paper.

5

Table 2 - Impact of Looks on Employment Probability: Any Employment

Women

Men

(1)

(2)

(3)

(4)

(5)

(6)

Above Average Looks Below Average Looks Attractiveness (Scale 1-11)

0.028*** (0.009)

0.039*** (0.011)

0.091** (0.039) -0.119* (0.070)

0.024*** (0.007)

0.016** (0.007)

0.015 (0.024) -0.080 (0.051)

Observations

1105

797

797

1096

Interviewer-Effects

Yes

Yes

Yes

Yes

Married Respondents Only

No

Yes

Yes

No

Partner's Employment Status No

Yes

Yes

No

Pseudo-R-Squared

0.172

0.189 0.186 0.274

Robust standard errors in parentheses

*** p ................
................

In order to avoid copyright disputes, this page is only a partial summary.

Google Online Preview   Download