Hobson's Choice

free script provided by

Income Inequality-5: Occupations

January 29, 2005

[ 1 | 2 | 3 | 4 | 5 | 6 ]

The Evolution of U.S. Earnings Inequality: 1961–2002 (PDF; Zvi Eckstein & Éva Nagypál, Minneapolis Fed, Dec '04)

It is orthodox thinking in economics to assume that workers, given their endowments of skills, experience, and education, can move from occupation to occupation based on opportunities, pay, and so on.

The division of workers by occupation has not been a standard practice of economists who emphasize competition and mobility in the labor market. Search theory introduces frictions into the labor market, while theory of speciﬁc human capital introduces mobility costs. Thereby, both restrict mobility and, hence, point to the importance of the division of the labor market by occupation.
[Eckstein & Nagypál, p.19]

In fact, anyone with ample experience looking for work knows that job flexibility is quite restricted: in nearly all job descriptions, the eliminating requirement is that the prospective employee have at least five years of experience doing almost precisely the same job. Despite decades of hearing how the new paradigm of employment is constant changing of careers, employers, skills, and the like, what we have actually seen is an almost Luddite demand on the part of firms for workers to do exactly what they have done in the past.

Unfortunately, Eckstein & Nagypál stick to the three categories of workers (blue-collar, white collar, and professional). This is disappointing because it fails to capture segmentation between occupations. Because it is difficult for employees to be accepted in new fields, it is difficult for them to emigrate out of moribund ones. For all that, they make some interesting observations:

The share of male professionals grew somewhat until 1983 from 24 percent in 1970 to 30 percent in 1983, with the blue-collar group making up for most of the decline within the other two groups. It is striking to see that, despite the change in the educational composition and in wage inequality that took place at the same time, the employment shares among men of various occupations have stayed roughly constant from 1983 to 2001. (The share of professional workers increased slightly from 30 percent to 33 percent, with the share of the other two groups each declining 1.5 percentage points.)
The occupational composition changed much more markedly for women, with the share of professionals showing a secular increase from 8 percent in 1970 to 28 percent in 2001, which was compensated by a roughly 10 percentage point decline in both the white collar and the blue-collar group. In 2001 about 50 percent of men are BC, 17 percent are WC, and 33 percent are PRO. Among women 23 percent are BC, 48 percent are WC, and 28 percent are PRO.
[p.19b-20a]

(Definitions of the categories "professional," "white collar" (WC), "blue-collar" (BC) are found in the appendix).

Eckstein & Nagypál observe that, while men holding postgraduate degrees are likely to be professionals, this is less true for women: they are somewhat more likely to be white collar instead. Moreover, within occupational brackets, the gender gap shows some surprising resiliency:

In terms of the closing of the female/male wage gap, it is worthwhile to point out that while blue-collar women have closed 31 percent of the wage gap with respect to their male counterparts, professional and white collar women closed only 15 percent and 5 percent of their wage gap, respectively. Inequality within the occupation groups started increasing for men in the mid ’80s, with similar increases for all groups. For women, inequality within each group started to grow earlier: in the mid ’70s for the professional group (within which inequality grew the most), in the late ’70s for the blue-collar group, and in the early ’80s for the white collar group. Again, the trend of mean and median wages by occupation is similar to the trend by education both in pattern and in magnitude.
[p.20a-20b]

The premium for professionals has grown strikingly (wages for professional men grew 33% faster than those for blue-collar workers, and 8% faster than for white collar workers; for women, professional wages grew 16% faster than for blue-collar, and 15% faster than for white collar. In part this was because white and blue-collar wages grew quite rapidly).

Eckstein & Nagypál ran regressions of earnings on the dummy variables for different educational and occupational categories, and, as one should expect, found them to be multiplicative; e.g., in 1995, men having a postgraduate enjoyed a wage premium of 1.6 times the mean; that of being a professional, 1.33 times the mean; professional postgraduate men presumably enjoyed a wage premium of 2.128. For women, this was 1.67 x 1.42. suggesting that, professional postgraduate women enjoyed a 2.38-fold premium relative to the mean for women (fig. 14, p.23; p.16, p.24; all premium-ratios are logs). In my opinion, however, the "smoking gun" for wage discrimination was their discovery that men enjoy a 60% higher premium for experience than women do, regardless of profession (fig. 15, p.24). I had previously assumed that women were paid less than men because they had obstacles in reaching certain positions, or that employers had a problem transferring female employees to higher-paid positions. These might be so, but it turns out that a giant share of the wage differential is pretty obviously sexism. The paper doesn't delve into this, however, and the authors don't make any judgments.

These are the components of a typical Mincerian equation; and in my next post, I hope to reveal a startling admission from the Fed Reserve researchers.