Blonde women earn more, have husbands who earn more

David W. Johnston. Physical Appearance and Wages: Do Blondes Have More Fun? (Economic Letters 108, s. 10-12):
Our data source is the 1979 cohort of the National Longitudinal Survey of Youth (NLSY79). [. . .] We limit our sample to Caucasian women aged 25 and over. [. . .] Regression results indicate that blonde women receive a wage premium equivalent in size to the return for an extra year of schooling. A significant blondeness effect is also evident in the marriage market. Blonde women are no more or less likely to be married; but, their spouse's wages are around 6% higher than the wages of other spouse's.

7 comments:

Hail said...

Earn more than who?

Does it mean blondes earn more than brunette white women?

Anonymous said...

Did you miss this part?

We limit our sample to Caucasian women aged 25 and over.

Anonymous said...

Population structure in the distribution of blondness amongst "Caucasians" in the USA seems to be a likely reason for this.

Hail said...

Thanks Anon-#2. Don't know how I missed that!

This is 1979, too, so the effect of Hispanics in the white sample will be minimal.

Hail said...

Next question: Is the blonde advantage a product of better education or something like that? (I'm not sure what Anon-#3 means by "Population structure in the distribution of blondness").

In other words, "all else being equal", do blondes still earn more?

Connected? --
1.) Pigmentation and intelligence
2.) IQ by hair-color and nationality
3.) IQ by hair and eye color

All from the 1920s USA. [Note: Most of the sample sizes in #2 are too small to be of use].

Anonymous said...

There's also a correlation with taller people earning more. Blondes and Northern Europeans in general are taller on average. Women also generally choose men who are taller than themselves.

n/a said...

Anonymous,

"Population structure in the distribution of blondness amongst "Caucasians" in the USA seems to be a likely reason for this."

According to the authors: A possibility, however, is that blonde hair is more common in certain ethnicities (e.g. Nordic) and that these ethnicities are favoured in the labour market (e.g. because of a high cultural work ethic). Fortunately, the NLSY contains detailed information on ethnicity and so we include in each specification indictors of whether the individual classifies themselves as English, French, German, Irish, Italian, Polish, Scottish or American.

Hail,

The authors say: Table 1 presents sample means of log hourly wages, spouse's log hourly wages and selected covariates. These statistics show that blonde women have higher log wages than other women, and that their spouse's have higher log wages than the spouse's of other women. There are, however, few clear differences between women in their characteristics. Blonde women are lighter than brunette women, but there are no systematic differences in education, Armed Forces Qualification Test (AFQT) scores, immigrant status or marital status. From these raw statistics it appears that differences in wages by hair colour are not driven by differences in productivity related characteristics.

Looking at the actual data, women with blonde hair have slightly higher AFQT percentiles and years of education compared to women with other hair colors except light brown, though the difference is not statistically significant at the 0.05 level. Light brown-haired women also have comparable years of education and slightly higher AFQT scores than blondes, but don't show a boost in wages.


Anonymous,

"There's also a correlation with taller people earning more. Blondes and Northern Europeans in general are taller on average. Women also generally choose men who are taller than themselves."

Blondes are slightly taller than women of other hair colors, but looking at the results in this paper hair color has a much larger effect on wages than height.

Some numbers from table 1:

Blonde Light Brown Brown All others
Height (in.) 65.00 64.68 64.82 64.39*
Weight (lb) 146.0 143.6 150.4* 150.2
Education (years) 13.89 13.90 13.61 13.66
AFQT percentile 0.587 0.602 0.565 0.566
Sample size 4288 4573 10,373 1606