and act on it, there will be an increase in demand for well-qualified
women, eventually reversing the effects of the initial discrimination.
In the third case, the cause of the discrimination is the preferences of the
customers. Unlike the second case where the discrimination reflects the
views of the firms’ managers, there are no economic forces working to
offset this customer-driven discrimination. As long as such discriminatory
views persist, the discriminatory wage differentials will be sustained—
only legislation or fundamental changes in customers’ attitudes will
reverse the situation. For many years, this was an important aspect of
racial discrimination in the southern U.S. states, where many white
customers were not prepared to have some services, such as haircuts,
provided to them by blacks.
1 Figure 14-2 shows only the correlation between education and earnings; these raw data are not
themselves evidence of a causal relationship. More advanced statistical studies, however, show
that there is a clear financial payoff from additional years of education.