<<

. 24
( 32 .)



>>

confound highly disparate reasons for differences in outcomes
among groups. Even if one regards differences in life chances as
more important than discrimination, any serious attempt to deal
with either social phenomenon must distinguish them from one
another.


Life Chances
In one sense, life chances are easier to determine than discrimina-
tion. For the past, the probabilities that a black child would be-
come a physician or a member of Congress can be determined
mathematically from statistics on the size of the black population
and the numbers of blacks who ended up in those occupations.
Even in this simplest case, however, there are complications.
Do we mean by life chances mere statistical probabilities in gen-
eral or do we mean the likelihood that a given attempt, with a
given effort, will succeed? What if very few black boys want to be-
come ballet dancers or are prepared to endure peer disapproval if
they persist in their efforts in this direction? In short, even the ap-
parently simple issue of determining life chances raises complicat-
ing questions as to the source of the differences in probabilities and
outcomes among groups. Whether that source is internal or exter-
nal becomes even more of a complication when seeking to deter-
mine the existence and magnitude of discrimination.
180 APPLIED ECONOMICS


Where discrimination is distinguished from differences in life
chances, the empirical question is whether individuals of similar
qualifications have similar prospects of employment, college admis-
sion, and other benefits when they come from different groups.
Where there are substantial differences in qualifying characteris-
tics among groups, as there often are, the question then becomes:
What of those particular individuals who have the same qualifying
characteristics as members of other groups? Do they have the same
prospects or results?


Comparability
Gross comparisons between groups may tell a radically different
story than comparisons of comparable individuals. For example,
black income has never been as high as white income in the
United States and, in years past, the disparities were even greater
than they are today. Yet, as far back as 1969, young black males
who came from homes where there were newspapers, magazines,
and library cards earned the same incomes as young white males
with the same things in their homes and with the same education.
As of 1989, black, white, and Hispanic Americans of the same age
(29) and with the same IQ_(100) all averaged between $25,000
and $26,000 in income when they worked year around.
American women's incomes have never been as high as the in-
comes of American men but, as early as 1971, women in their thir-
ties who had never married and who had worked continuously
since high school earned slightly higher incomes than men of the
same description. In Canada, women who never married have
earned more than 99 percent of the income of men who never
married. Among college faculty members, American women who
had never married earned substantially higher incomes than men
who had never married, as far back as 1969.
The Economics of Discrimination 181

None of this says anything about life chances, however relevant
it may be to questions about employer discrimination. The per-
centage of black males who came from the kinds of homes de-
scribed was very different from the percentage of white males who
came from such homes. IQ_scores are likewise distributed very dif-
ferently among American blacks, whites, and Hispanics. It is not
uncommon, both in the United States and in other countries, for
one racial or ethnic group to differ in age from another by a
decade or more”and age makes a large difference in income. So
does marriage. While married men tend to earn more than single
men and married men with children still more, the exact opposite
is the case for women. Nor are the reasons particularly obscure,
since marriage and parenthood create very different incentives and
constraints for women than for men.
The justice or social desirability of traditional sex roles in mar-
riage can be debated. But that debate is not about employer dis-
crimination, much less about the presumed predispositions of
employers. Questions about the existence or magnitude of em-
ployer discrimination are questions about whether individuals who
show up at the workplace with the same developed capabilities
and liabilities are treated the same there. It is not a question about
whether life has been fair to them before they reached the em-
ployer, however important that question might be when dealing
with other issues. Yet, increasingly, laws and policies have defined
as employer discrimination a failure to pay the extra costs associ-
ated with those workers who are pregnant, suffering from mental
illness, or are more of a risk because of a prison record or the
health problems associated with old age. When hiring, employers
have incentives to think beyond stage one, even if those who make
or advocate laws and government policies do not.
Whatever the merits or demerits of having such additional costs
paid by the employer or by the government or by others, a failure
182 APPLIED ECONOMICS


to do so is not the same thing as treating equally valuable workers
differently because they happen to come from different groups.
When workers whose net value to the enterprise are in fact differ-
ent are hired, paid and promoted according to those differences,
that is not employer discrimination, whatever else it might be.
This is not even a question about desirable social policy. It is a
question about using words that have some specific meaning,
rather than as mere sounds to evoke particular emotions. Only
when terms have some specific meaning, understood by those on
both sides of an issue, can we at least engage in rational discussions
of our differences of opinion about what is or is not desirable social
policy. Otherwise, we are simply talking past each other.
While isolated data can demonstrate graphically the distinction
between differing life chances, on the one hand, and discrimina-
tion on the other, the more general problem of determining which
phenomenon exists in particular situations requires more complex
and often uncertain methods. For example, it is common to "con-
trol" statistically for certain factors and then see if individuals from
different groups have similar results when those factors have been
taken into consideration. Yet this is much easier to accept in prin-
ciple than to apply in practice.
Nothing is easier than to imagine that one has taken all relevant
factors into account, when in reality there may be other factors
which have influence, even if no data have been collected on them
or they are not quantifiable. Where there are very significant dif-
ferences in known factors between one group and another, it would
be reckless to assume that all remaining unknown factors are the
same. Yet that is done again and again in discussions of discrimi-
nation.
A scholar in India, for example, attempted to compare Bombay
medical school students with such comparable social characteris-
tics as father's income and occupation, but later discovered in the
The Economics of Discrimination 183

course of interviewing these students that those who were the
same in these respects, but who were also untouchables, were not
comparable on other variables, such as the quality of the schools
they had come from, the number of books in the homes in which
they had grown up, and the literacy rates among their fathers and
grandfathers.
Similarly, a study of black and white faculty members in the
United States found that those who were comparable on some
variables were nevertheless very different on other variables. Even
black professors with Ph.D.s usually received those Ph.D.s at a
later age, from institutions with lower rankings in their respective
fields, and these professors had published fewer articles. When
such variables as published articles, the ranking of the institutions
from which Ph.D.s were earned, and years of academic experience
were all held constant, black professors with Ph.D.s earned more
than white professors with Ph.D.s, as far back as 1970”a year be-
fore affirmative action policies were mandated by the federal gov-
ernment.
None of this is unusual. It is commonplace in countries around
the world for groups who differ in the quantity of their education
to differ also in the quality of their education. This means that
comparisons of individuals with ostensibly the "same" education
from different groups can amount to comparisons of apples and
oranges. Therefore equating differences in pay between members
of different groups who have the "same" education with discrimi-
nation is a non sequitur.
Educational quality may differ along many dimensions, includ-
ing institutional ranking, individual performance, or fields of spe-
cialization. Such differences have existed between the Chinese
and the Malays in Malaysia, Tamils and Sinhalese in Sri Lanka,
and European and American Jews versus Middle Eastern Jews in
Israel, as well as between caste Hindus and untouchables in India
184 APPLIED ECONOMICS


and among blacks, whites, and Hispanics in the United States. The
education of American college-educated women and men likewise
differs in fields of specialization and in the postgraduate education
added on to that of these women and men whose education is con-
sidered the "same" because they have all graduated from college.
Education is not unique in having such differences among ap-
parently similar individuals from different groups. A highly publi-
cized study which concluded that there was racial discrimination
between minority and white applicants for mortgage loans in the
United States failed to note that the minority applicants had larger
debts, poorer credit histories, sought loans covering a higher per-
centage of the value of the properties in question, and were also
more likely to seek to finance multiple dwelling units rather than
single-family homes.
When these factors were taken into account, the differences in
mortgage loan approval rates shrank dramatically, though the re-
maining difference of 6 percentage points was nevertheless attrib-
uted to discrimination, as if there could not possibly be any other
differences in the remaining unstudied factors, despite large differ-
ences in the variables that were already studied. For example, other
data from the U.S. Census show that blacks and whites with the
same incomes average very different net worths”and net worth is
also a major consideration in decisions to approve or disapprove
mortgage loan applications.
Much statistical data does not permit such a fine breakdown as
the data comparing black and white faculty members in 1970 or
the follow-up study on mortgage lending. For much of the me-
dia”and often even in academia”it is sufficient to find inter-
group differences in outcomes to conclude that there has been
discrimination. This happens, however, only when the conclusion
fits existing preconceptions. Statistics showing Asian-Americans
to have higher incomes, lower infant mortality rates, and higher
The Economics of Discrimination 185

rates of mortgage application approval than white Americans are
never taken to show discrimination against whites, even though
identical data”often from the same studies”are taken as proof of
discrimination against blacks.
Gross statistical differences between the "representation" of dif-
ferent groups in particular occupations, institutions, or income
levels are likewise usually considered sufficient for concluding that
the "under-represented" groups were "excluded" by discriminatory
policies or by individual or institutional biases or prejudices, how-
ever subtle, covert or unconscious these might be. Yet nothing is
more common, in countries around the world, than huge differ-
ences in representation between groups, even when the more for-
tunate group is in no position to discriminate against the less
fortunate group. These include situations where the more fortu-
nate group is a minority with no institutional or political power
over the majority.
In Malaysia, for example, for the entire decade of the 1960s,
members of the Chinese minority in Malaysia received more than
400 degrees in engineering, while members of the Malay majority
received just four. This was in a country where the universities
were controlled by the Malays, who also controlled the govern-
ment that controlled the universities.
Nor was this situation unique. In czarist Russia, the German
minority”about one percent of the population”was an absolute
majority in the St. Petersburg Academy of Science. Jewish physi-
cians were an absolute majority of all physicians in Poland and
Hungary in the period between the two World Wars, though Jews
were less than 12 percent of the population in both countries.
Southern Nigerians were an absolute majority of the people in a
number of professions in northern Nigeria when that country
gained its independence. The list could go on and on, filling a
book much larger than this one.
186 APPLIED ECONOMICS


None of this disproves the existence of discrimination or mini-
mizes its magnitude. It simply asks that those who assert its exis-
tence demonstrate its existence”by some evidence that will stand
up to scrutiny.


ANTI-DISCRIMINATION LAWS

Both discrimination and affirmative action involve costs. It could
hardly be otherwise, since they are both essentially preferential
policies, though the preferences are for different groups. Whatever
the rationales or goals of these policies, the economic question is:
What are the actual consequences of such policies”and not just in
stage one? In principle, anti-discrimination laws are a third policy,
designed to preclude preferential treatment of either a majority or
minority. But this policy as well cannot be defined by its inten-
tions, goals or rationales. The real question is: What are its specific
characteristics and their consequences”and, again, not just in
stage one?
The likely consequences of laws and policies against discrimina-
tion depend on how discrimination is defined and determined by
those with the power to enforce such laws and policies. The incen-
tives and constraints created by these definitions and methods of de-
termining whether or not discrimination exists provide more clues
as to what to expect than do the goals or rationales. What is then
decisive is the empirical evidence as to what actually happens.


Legal Definitions
In the United States, "discrimination" had a very simple and
straightforward legal meaning at the beginning of the federal
civil rights legislation of the 1960s. It meant treating individuals
differently according to the racial or ethnic group in which they
187
The Economics of Discrimination

<<

. 24
( 32 .)



>>