Database Reference
In-Depth Information
observable characteristics, such as sex and race, as proxies for the expected produc-
tivity, estimated by their prior knowledge on the average productivity of the group
the worker belongs to. Wages are then set on the basis of the expected productiv-
ity of the group, not on the basis of the person's productivity. We refer the reader to
(Altonji & Blank, 1999) for a comprehensive mathematical introduction to both the-
ories of economic discrimination, as well as for past empirical approaches to show
direct evidence. More recent or comprehensive reviews of theories and empirics in
labor market are available in (Cain, 1987; Charles & Guryan, 2011; Kunze, 2008;
Lang & Lehmann, 2011). (Weichselbaumer & Winter-Ebmer, 2005) conduct a meta
regression analysis of the works on gender wage differentials, where each point of
data is not an individual but a research study. (Neal & Johnson, 1996) observed that,
after controlling for the ability of a worker, the racial wage gap greatly reduces. In
such a study, ability was measured through the controversial Armed Forces Qual-
ifying Test (AFQT) score, a test of cognitive skills taken by male adolescents and
available from the National Longitudinal Survey of Youth. In the following, we
briefly review the most recent lines of research and extensions of the two economic
models.
Approaches on taste-based discrimination. The additional cost of minority work-
ers in presence of taste-based discrimination leads to an equilibrium wage differen-
tial and to segregation of minority workers in less discriminating firms or for specific
occupations. Lower earning for discriminatory firms implies that discrimination oc-
curs mainly in low competitive markets. This is known as the static implication of
the Becker's model 5 . Influential papers are (Charles & Guryan, 2008), which com-
bine GSS data (to measure racial prejudice) with CPS data (to measure differences
in wages), and (Hellerstein et al., 2002), which relate firm profitability to the pro-
portion of female workers both in low competition and high competition markets.
Recent approaches using survey data include (Sano, 2009; Tsao & Pearlman, 2010;
Zhang & Dong, 2008). On the basis of the identity of the discriminator, Becker's
model distinguishes employer discrimination (taste in hiring), customer discrimina-
tion (taste in buying), and co-employee discrimination (taste in co-operating). Re-
cent analyses of consumer discrimination have been conducted on data from restau-
rants (Parrett, 2011; Myers, 2007), contact jobs (Combes et al., 2011), retail stores
(Leonard et al., 2010), Major League Baseball (Coyne et al., 2010) and taxicab
drivers (Ayres et al., 2005). Evidence of correlation between the predominant race
of customers and the race of the marginal hired worker has been shown in (Holzer
& Ihlanfeldt, 1998).
5
The dynamic implication of the Becker's model predicts that non-discriminating employers
earn higher profits by hiring members of the protected group, and, in the long run and in
a competitive market, discriminatory firms will be driven out of the market. The dynamic
implication has been investigated in the context of banking deregulation (Black & Strahan,
2001; Levine et al., 2008), globalization (Black & Brainerd, 2004; Neumayer & Soysa,
2007; Oostendorp, 2009) and in the adoption of equality laws worldwide (Weichselbaumer
& Winter-Ebmer, 2007).
Search WWH ::




Custom Search