CHAPTER 6 |
Error is never so difficult to be destroyed as when
it has roots in language.
—Jeremy Bentham, On Evidence
WITHOUT A DOUBT, part of the confusion in understanding racial issues lies in the imprecise and ambiguous language used by scholars and laymen alike in discussing race. Words can, and usually do, have more than one meaning, and therefore can be used ambiguously. In analytical usage, not only is it necessary to separate the connotative from the literal content of words, but precise and operationally useful distinctions and definitions also must be made.
An example of ambiguous language is found in the use of the phrase “racial segregation.” Consider the following observations. Blacks represent about 65 percent of the Washington, D.C., population. Reagan National Airport serves the Washington area, and like every such facility, it has water fountains. At no time has the writer observed anything close to blacks being 65 percent of water fountain users; a wild guess would place their usage at 5 or 10 percent at most. To the extent that this observation approximates reality, would anyone move to declare that Reagan airport water fountains are racially segregated?
Casual observation of ice hockey games would suggest that blacks attend them far below their percentages in the general population. A similar observation can be made about operas, dressage performances, and wine tastings. The population statistics of states such as South Dakota, Iowa, Maine, Montana, and Vermont show that not even 1 percent of their populations is black. On the other hand, in states such as Georgia, Alabama, and Mississippi, blacks are overrepresented in terms of their percentage in the general population. Would anyone suggest that racial segregation accounts for any of those observations?
Just because blacks are not proportionately represented in some activity, according to their numbers in the general population, how analytically useful is it to assert that the activity is racially segregated? A more useful test is whether, for example, a black person at Reagan airport is free to drink from any water fountain he chooses. If the answer is in the affirmative, then the water fountains are not racially segregated, and that would be true even if no black person ever uses the water fountains.
If the average American were asked whether the country’s public schools are segregated, a consensus would be virtually impossible. Some would argue, as has Harvard University’s Civil Rights Project, that schools are racially segregated and becoming more so:
Civil rights goals have not been accomplished. The country has been going backward toward greater segregation in all parts of the country for more than a decade. Since the end of the Civil Rights era, there has been no significant leadership towards the goal of creating a successfully integrated society built on integrated schools and neighborhoods.[1]
A little reflection on the matter shows that people give the term racial segregation one meaning when applied to water fountains, operas, and ice hockey games—and an entirely different meaning when applied to schools and neighborhoods. The test used to determine whether Reagan airport water fountains were segregated or not should also be used to answer the question of whether schools were segregated or not. If a black student lives within a particular school district, is he free to attend that school? If the answer is affirmative, then the schools are not segregated, even if not a single black attends the school. In contrast to the past, there are today no legal or extra-legal barriers to keep blacks who reside within a school district from attending its schools.
When an activity is not racially mixed today, a better word is racially homogeneous, which does not mean that it is racially segregated, at least in the historic usage of the term. It would surely be deemed ridiculous, foolhardy, and a gross abuse of government power if, for example, one were to conclude that since blacks are “underrepresented” at Reagan airport water fountains, there ought to be a policy to bus blacks to such fountains. Similarly, I doubt whether one would propose compelling blacks to move from Georgia to Iowa, and the reverse for whites, until those actions satisfied some sort of preconceived notion of what constitutes racial integration across states. Why blacks are “underrepresented” in some activities and “overrepresented” in others may reflect personal preferences, history, cultural influences, income differences, and discrimination.
Those who advocate and litigate for school desegregation today are not fighting against state and local laws that mandate racial separation. Their argument rests solely on the fact that black attendance at some schools is not proportional to or representative of the numbers of blacks in the population. For those advocates and litigators, the school does not have a pleasing racial mix. The fact that many of today’s large-city school systems are predominantly black is mostly a result of residential housing patterns and not legislated school-segregation policy. That fact makes racial heterogeneity virtually impossible. For example, in Manhattan, public schools are nearly 90 percent black or Hispanic, while private schools are 80 percent white.
While there is a smaller overall percentage of blacks in private schools, they are somewhat more racially heterogeneous nowadays. Those who see racial heterogeneity in schools as desirable should support measures such as education vouchers or tuition tax credits to strengthen the private-education sector. A majority of black parents support educational vouchers that would allow them educational choice.[2]
Other terms and concepts used in the racial literature and debate are just as misleading and confusing. Among the aims of this chapter is to discuss those ambiguities, suggest operational definitions, and maybe shed more light on racial phenomena we observe.
Discrimination Operationally Defined
More generally, and inclusive of legal attempts to define the term, discrimination might be operationally defined as the act of choice or selection. All selection necessarily and simultaneously requires non-selection. Choice requires discrimination. When one chooses to attend the University of Chicago, he non-selects or discriminates against Harvard University as well as every other university. When one selects a Bordeaux wine, he non-selects a Burgundy wine. We might call these cases university discrimination and wine discrimination. Similarly, when the term discrimination is modified with the nouns race and sex, we merely specify the criterion upon which the choice is made.
At this juncture, we might ask if there is any conceptual distinction between discriminating for or against particular universities, wines, and other goods and services and discriminating for or against particular races and sexes. Or should one discriminate at all? Can one make a case for indifference or random choice among universities, wine, or people? Indifference or random choice is hardly ever the case. Our lives are largely spent discriminating for or against prospective selected activities, objects, and people. Some of us discriminate against those who have criminal records, who bathe infrequently, who use vulgar speech and have improper social graces. Most of us choose mates within our own race, ethnic group, or religion, hence discriminating against mates who, save for those characteristics, might be just as suitable. According to the 1992 census, only 2.2 percent of Americans were married to someone outside their own race.[3] There is also evidence of discrimination based on physical attributes in politics: not many short men have been elected president of the United States. In fact, twenty-two out of forty-two presidents have been five feet, eleven inches and taller, well above the population’s average height.[4] That is not an outcome that would be expected if height had nothing to do with choices.
Furthermore, discrimination is not consistent. Sometimes people discriminate against theater entertainment in favor of parties or against women in favor of men; and at other times and circumstances the same people do the opposite.
One might be tempted to argue that racial discrimination in marriage is trivial and does not have important social consequences requiring a legal or political remedy, as do other forms of racial discrimination. But it does have important social consequences. When there is assortive (non-random) mate selection, it heightens whatever group differences there are in the population.[5] When high-IQ people marry other high-IQ people, when high-income people marry other high-income people, and to the extent there is a racial correlation between these characteristics, racial discrimination in mate selection enhances the inequality in the population’s intelligence and income distribution. There would be greater income equality if high-IQ and high-income people mated with their low-IQ and low-income counterparts. But I imagine that most of us would be horrified by the suggestion of a mandate to require such mating.[6]
It would appear that the term discrimination, defined simply as the act of choice, is morally neutral in the sense that there are no unambiguous standards that permit us to argue that the choice to attend the University of Chicago or to purchase a Bordeaux wine is more righteous than the choice to attend Harvard University and purchase a Burgundy wine. More important, no argument can be made for government forcing a person to select one university or wine over another. And none can be produced to force people to grant equal opportunity when choosing universities, wines, and marriage partners.
If people are free to discriminate in favor of, or against, a university or wine, what argument can be made against their having that same right with respect to choosing based on the race or sex characteristics of their mates, employees, tenants, or club members? If one shares the value of freedom of association, why should some associations be permitted and others denied? If a man is not permitted to bring a court action against a woman who refuses to deal with him, in the form of a dating or marital relationship or for any arbitrary reason she chooses, what is the case for bringing court action for other refusals to “deal,” such as in employment, renting or selling a house, or club memberships?
Nobel Laureate Kenneth Arrow argues that “There are many varieties of liberalism, which draw the boundaries between social and individual action in different places, but all agree in rejecting racial discrimination, by which is meant allowing racial identification to have a place in an individual’s life chances.”[7] If “allowing racial identification to have a place in an individual’s life chances” means refusal to deal, what policy recommendations emerge? Refusal to deal can apply to any setting, including activities like marriage, friendship, invitations to social gatherings and golf games, all of which might affect one’s “life chances.” If refusal to deal is permitted in one activity, for any arbitrary reason, what case can be made for not permitting it in others? The practical answer to this question has more to do with the threat of government violence against people who refuse to deal in prohibited ways than with any kind of internally consistent logic.
Prejudice
In much of the racial literature, prejudice is usually seen as suspicion, intolerance, or an irrational hatred of other races. Sometimes it is seen as oppression, as suggested by law Professor Khiara M. Bridges, when he says, “Therefore, if racial prejudice, the subordination of people of color, and White supremacy persist, they do so largely because the legal system sanctions them.”[8] Other times, prejudice is understood to mean racial preferences as implied by Justice Sandra Day O’Connor, when she wrote for the majority in Adarand Contractors, Inc. v. Pena (1995), striking down a government set-aside: “[B]ecause that perception especially when fostered by the Congress of the United States can only exacerbate rather than reduce racial prejudice, it will delay the time when race will become a truly irrelevant, or at least insignificant, factor.”[9]
These visions of prejudice expose analysts to the pitfalls of making ambiguous statements and advancing faulty arguments. A useful operational definition of prejudice can be found by examining its Latin root—praejudicium, meaning “to judge before the facts are known.” Thus, we might define prejudicial acts as decision-making on the basis of incomplete information.
That kind of decision-making, before facts are known, is necessary and to be expected in a world of scarcity, uncertainty, complexity, costly information—and often erroneous interpretation of that information. Furthermore, different individuals might arrive at different interpretations even if confronted with the same information. Also, different people reach different decisions on just what constitutes the optimal quantity of information to gather prior to making decisions.
Consider a simple, yet intuitively appealing, example of how decisions might be made on the basis of incomplete information (and possibly erroneous interpretation of evidence). Suppose a fully grown tiger suddenly appeared in a room. A reliable prediction is that most individuals would endeavor to leave the room with great dispatch. Such a response to the tiger’s presence is not likely to be based on any detailed information about the behavioral characteristics of that particular tiger. The response is more likely to be based upon one’s stock of information about tigers as a class, through folklore and observations of other tigers. The individual prejudges the tiger; we might say he employs stereotypes. He simply ascribes known or surmised group characteristics to the individual tiger. He is not likely to seek additional information, because he calculates that the expected cost of acquiring an additional unit of information about that tiger, such as talking to or petting it, is likely to exceed the expected benefit. People seldom undertake an activity when they surmise that its expected cost will exceed its expected benefit.
Sometimes when people use the terms prejudice and stereotype, they are used pejoratively to refer to those whose chosen quantity of information for decision-making the observer deems too small. However, what constitutes the optimal quantity of information collected before decisions are made is subjectively determined by the individual’s calculation of his own costs and benefits.
Information is not costless. To acquire an additional unit of it requires a sacrifice of time, effort, or other resources. People therefore seek to economize on information cost. In doing so, they tend to substitute less expensive forms of information for more expensive forms. Physical attributes are “cheap” to observe. If a particular physical attribute is perceived as correlated with a more costly-to-observe one, the observer might use that attribute as an estimator or proxy for the costly-to-observe attribute. The cheaply observed fact that an individual is short, an amputee, a black, or a woman provides what some people deem to be sufficient information for decision-making or predicting the presence of some other more costly-to-observe attribute. For example, if asked to identify individuals with doctorate degrees in physics only by observing race and sex, most of us would assign a higher probability that white or Asian males would have such degrees than black males or women. Such behavior is what decision theory expects: an unobservable attribute must be estimated from an observable one.
Stereotyping and prejudging can be independent of preferences. Observing a person’s decision-making behavior does not allow us to say anything unambiguous about that person’s personal preferences with regard to race, sex, ethnicity, and nationality.
A simple example can demonstrate this. Imagine the reader is on a particular university campus. He is asked to pick a five-person basketball team from a group of twenty students; if the team he selects wins the game, he wins $10,000. The group from which he is to choose consists of five black males, five white males, five black females, and five white females. He has zero information about their basketball proficiency, and they appear otherwise indistinguishable except by race and sex. That is, they are identical in terms of other physical characteristics, such as weight and height. Assuming that the person’s objective is to maximize his chances of winning, his selection would probably be dominated by black males. That choice would reflect the real-world associations he has seen between basketball proficiency and race and sex.
What can an observer, watching that person’s choices, say about his race or sex preferences? Absolutely nothing unambiguous simply by observing choices based on race and sex. Moreover, a person bearing antipathy toward blacks would select in the identical fashion as everyone else—if maximizing winnings dominated his objectives. Furthermore, given the high correlation between race, sex, and basketball proficiency, would anyone care if the selector indulged a racial preference for white males or women? He would personally bear the cost of preference indulgence.
A vast number of decisions must be made during our lives. Some of them, such as deciding whether to greet a passerby in the morning, require relatively small amounts of information. Others, such as selecting a marriage partner, require large amounts. A person is not “prejudiced” or “unprejudiced.” Rather, he always exhibits prejudiced behavior to the extent that he substitutes general information (prejudgment or stereotypes) that is less costly in exchange for specific information that’s more costly. What distinguishes one person from another are their comparative degrees of prejudiced behavior when facing similar situations. For any given decision, some people will search for more information than will others.
Prejudging People
Physical characteristics can be used as proxies for other costly-to-observe characteristics. Some racial and ethnic groups have higher incidences of and mortality from various diseases than the national average. In 1998, rates of death from cardiovascular diseases were about 30 percent higher among black adults than among white adults. Cervical cancer rates were five times higher among Vietnamese women in the United States than among white women there. Pima Indians of Arizona have the highest-known diabetes rates in the world.[10] Prostate cancer is nearly twice as common among black men as white men.[11]
Approximately 80,000 Americans have sickle-cell disease. About 9 percent of blacks have the trait. One in every 1,000 to 1,400 American Hispanic children is born with sickle-cell disease itself. The high incidence of the sickle-cell gene in these and other specific populations is due to their ability to make red blood cells resistant to the malaria parasite.
Whether genetics, environment, or other factors account for the association between race and some diseases, it is undeniable that such an association exists. Thus, a physical characteristic such as race can be used as an indicator of the higher probability of some other characteristic, such as prostate cancer and cervical cancer. Health-care providers can therefore better assess patient screening needs; for example, they can prescribe more frequent blood tests for the prostate-specific antigen (the well-known PSA test) for black males. They can employ what is derisively called racial profiling. One might take the position that while it is legitimate for doctors to use race or ethnicity as indicators of the higher probability of certain diseases, it is not legitimate to use those two characteristics as indicators for worker productivity, criminal behavior, or basketball proficiency. Other than simply stating that it is acceptable to use race or ethnicity as a predictor in medicine but not in other areas of life, is there really a difference? Surely race and ethnicity are not perfect indicators of the risk of prostate cancer or hypertension; neither are they perfect indicators of SAT scores, criminal behavior, or basketball/football/sprinting skill.
However, there are concrete factual data that indicate such associations. For example: In 2002, the average black score on the combined math and verbal portions of the SAT was 857, the average white score 23 percent higher—at 1060.[12] While blacks are 13 percent of the population, they constitute 80 percent of professional basketball players and 65 percent of professional football players. Blacks who trace their ancestry to West Africa, including black Americans, have achieved more than 95 percent of the best times in sprinting.[13] For the crime of homicide, over the years 1976–2000, blacks, while 13 percent of the general population, made up 51.5 percent of the offenders; whites were 46.4 percent, and others 2 percent.[14]
Racial Profiling
Using race as an indicator does not necessarily tell us anything about the chooser’s racial preferences. The Washington Lawyers Committee filed a lawsuit in April 2001 on behalf of Bryan Greene, a black man, against Your Way Taxicab Company for violations of 42 U.S.C. sec. 1981 and the District of Columbia’s Human Rights Act, both of which prohibit discrimination in the making of contracts. As Greene approached a hotel entrance, the doorman was assisting a customer out of a Your Way Taxicab. The doorman saw Greene and attempted to hold the cab for him; however, when the driver spotted Greene, he sped away. After mediation, Your Way Taxicab Company reached an out-of-court settlement.[15]
In a number of cities, there have been complaints by blacks of similar behavior by cab drivers. The question we might ask is this: are the drivers’ decisions based on racial preferences, or might they fear being asked to go into a neighborhood where there is a high probability of being robbed, assaulted, or murdered? By simply knowing that a driver refused a black passenger, we cannot make an unambiguous statement about whether the decision was motivated by racial preferences.
Evidence that the taxi driver’s decisions might very well be based on criteria other than racial preferences is seen in a 1999 story written by James Owens, titled “Capital Cabbies Salute Race Profiling.” Owens said:
If racial profiling is “racism,” then the cab drivers of Washington, D.C., they themselves mainly blacks and Hispanics, are all for it. A District taxicab commissioner, Sandra Seegars, who is black, issued a safety-advice statement urging D.C.’s 6,800 cabbies to refuse to pick up “dangerous looking” passengers. She described “dangerous looking” as a “young black guy . . . with shirttail hanging down longer than his coat, baggy pants, unlaced tennis shoes,” etc. That’s one typical description—but the cabbies know, from fear-filled experience, about many other “looks” of black-male threat, especially at night. She also warned cabbies to stay away from low-income black neighborhoods (which comprise much of Washington, D.C.). Her action was triggered by the most recent murder of a cabbie in Southeast Washington.[16]
Another example of race as an indicator stems from a case in which residents in Southwest Washington filed suit in a U.S. district court after Domino’s Pizza repeatedly refused to make home deliveries in certain neighborhoods and instead made customers meet drivers at curbside to pay for and receive their orders. The lawsuit alleged racial discrimination by Domino’s Pizza and Team Washington, a company that operates more than 50 Domino’s stores. According to the plaintiffs, Domino’s delivers to the door in Georgetown and other mostly white areas of Northwest Washington. The suit also alleged that deliverymen engaged in similar delivery decisions in Southeast Washington’s Potomac Gardens, where another customer filed a bias lawsuit. Again, the question is whether the drivers were acting on the basis of racial preference or out of fear of assault or robbery?
According to Pizza Marketing Quarterly, similar charges of racial discrimination were filed in St. Louis against Papa John’s pizza delivery. Cathy Juengel, a St. Louis Papa John’s district manager, said she could not and would not ask her drivers to put their lives on the line. She added that the racial discrimination accusation was false, because 75 to 85 percent of the drivers servicing the complaining neighborhood were black and, moreover, most of those drivers lived in the very neighborhood being denied full delivery service.[17] One doubts that black pizza deliverers can be charged with disliking black people.
The Reverend Jesse Jackson once said, “There is nothing more painful for me at this stage in my life than to walk down the street and hear footsteps and start thinking about robbery—then look around and see somebody white and feel relieved.”[18] A former Los Angeles police chief, Bernard Parks, defending racial profiling, said, “It’s not the fault of the police when they stop minority males or put them in jail. It’s the fault of the minority males for committing the crime. In my mind it is not a great revelation that if officers are looking for criminal activity, they’re going to look at the kind of people who are listed on crime reports.”[19] Charleston, South Carolina, Chief of Police Reuben Greenberg argued that the problem facing black America is not racial profiling. He said, “the greatest problem in the black community is the tolerance for high levels of criminality.”
The percentages of black arrests for selected crimes, found in Crime in the United States, 2006: Uniform Crime Reports, bear out the disproportionate criminal activity by blacks: murder and non-negligent manslaughter (51 percent), robbery (56 percent), aggravated assault (35 percent), and vehicle theft (34 percent). For blacks under the age of 18, the percentages are even more startling: murder and non-negligent manslaughter (60), robbery (67), aggravated assault (42), and vehicle theft (43).[20] Although racial profiling is often seen as racism, I doubt that anyone could find much evidence for a claim that Jesse Jackson, black police chiefs Bernard Parks and Reuben Greenberg, black taxi drivers, and black pizza deliverers are racists.
The law-abiding black citizen who is passed up by a taxi, refused pizza delivery, or stopped by the police can rightfully feel a sense of injustice and resentment. But the bulk of those feelings should be directed at those who have made race synonymous with higher rates of criminal activity rather than the taxi driver or pizza deliverer who is trying to earn a living and avoid being a crime victim.
Racial Preferences
People have likes and dislikes for many things. In everyday language, as well as in economic analysis, an individual is said to prefer object A to object B if, being free to choose, he chooses A rather than B. There are no objective criteria by which we can judge whether one set of preferences is better or more righteous than another. There are no commonly accepted standards to prove, for example, that it is better for a person to prefer Bordeaux wines to those of Burgundy or to prefer three-piece suits to blue jeans. For a person to say that a preference for Bordeaux over Burgundy is sensible or more righteous is plain nonsense. To conclude that he should be indifferent to the distinctions between them is also nonsense. To say that Bordeaux wine is better than Burgundy wine is simply a value judgment in a debate that can go on forever; it boils down to a matter of tastes and opinion.
The absence of standards for evaluating statements dealing with better and best is found in all sciences. For example, it is meaningless, and in fact nonsense, for a physicist to suggest or even entertain a discussion about whether a gaseous state is better than a liquid or solid state or whether they are equally good. So far as human choice is concerned, the most we can ever say objectively is whether, given a set of preferences and the constraints of prices and income, an individual is or is not doing the best he can.
The same holds true when it comes to individual preferences for human physical attributes, such as height, weight, a “richly” endowed body, and hair color: these are solely matters of individual taste. To the extent that individuals have preferences for or against human physical attributes in general, we also expect people to exhibit preferences for or against racial attributes. From an objective analytical point of view, there are no conceptual distinctions between racial preferences and preferences for other objects of desire.
It might be argued that racial preferences are not comparable to other kinds in terms of the consequences they have for society and for individuals. Although the indulgence of racial preferences has specific effects that the indulgence for, say, certain wines does not, are they basically different? If so, how? The preference for Bordeaux wines “harms” Burgundy producers by reducing the value of resources held in Burgundy production. If the consequences of preferences are generally thought to reduce the value of some resources and increase the value of others, then it can be said that preferences for physical attributes have effects similar to those of other preferences. The essential difference—by no means small—between preferences for racial features and those for wines is that the latter are not as specialized as the former. In other words, if Burgundy producers see that consumers prefer Bordeaux, they theoretically would be able to convert their resources into Bordeaux production. On the other hand, people who are black cannot become white.[21]
The fact that racial characteristics are specialized or unchangeable does not place them in a class by themselves. Persons with average IQs are generally preferred to those with below-average IQs, and persons who are not physically disabled are preferred to those who are. In each of these cases, and many others, the less-preferred physical attribute is unchangeable; and in each case, the less-preferred person suffers a disadvantage in some competitive arenas. This disadvantage is to be expected. Disadvantage and advantage are inevitable consequences of differences in individual tastes, abilities, and traits on the one hand, and freedom of choice in an open society on the other.
Human preferences—whether for such physical attributes as race, food, child-rearing practices, or entertainment—can have a moral dimension. There may be a moral consensus condemning preferences for forms of entertainment such as pornographic movies and a consensus that condemns indulgence of racial preferences. The fact of a consensus on what constitutes moral or immoral preferences does not alter the facts: people do exhibit preference; and there is no commonly agreed-upon standard for assessing which preferences are better than others. The best we can hope for, in terms of objective analysis, is to take human preferences as a given and then use economic theory to analyze how people choose in the face of objective constraints of price and income.
If there can be a moral dimension to preference indulgence, it should be to oblige people to pay for their preferences. Open markets make this possible, and as a result there is less tendency to indulge preferences. For example, at one time there were no black professional basketball and football players. Today, black players dominate both sports. That came about without anti-discrimination litigation and racial-employment preferences. It resulted from the natural actions of markets and competition among teams that could not ignore a huge pool of black talent.
Suppose a particular basketball or football team owner decides that he is going to indulge his racial preferences against black players by not hiring them. Many fans, both black and white, might find his actions reprehensible and attendance would fall. The team’s winning record is likely to be lower. Plus, he would have to raise the salaries he was willing to pay in order to acquire highly productive white players. His racial-preference indulgence could cost him tens of millions of dollars in the forms of lost revenue and higher costs. So we would not expect a basketball or football team owner to indulge any preferences he might have against blacks.
While the basketball and football example is easy to visualize, the same forces are at work in other markets. If an employer chooses to discriminate against blacks, he will have to pay higher wages for white workers. The black person he shunned will be available to employers who might be more interested in profits than in indulging their racial preferences at wages less than the discriminating employer has to pay.
When analyzing racial phenomena, we must be careful to distinguish among the three determinants of choice: preference, prejudice, and real differences. If we assume that racial preferences are the root of the problem we are addressing—when prejudice or real differences are a better explanation—we will derive erroneous conclusions, and promote policy that is ineffectual and possibly harmful to its intended beneficiaries. Let us briefly examine several areas where racial preferences are generally assumed to be the villain.
Hiring and Employment Discrimination
Many recruitment and hiring practices are often said to reflect racial preferences, but an alternative explanation might be drawn from our previous discussion about seeking information. When a company sets out to hire workers, it must discover how productive those seeking jobs are and whether they are suitable for training. Since this process costs money, an employer has an incentive to select a recruitment strategy with a high probability of success. For example, if there is thought to be a correlation between the candidate’s performance and the quality of the high school he attended, some recruitment costs can be reduced just by knowing that simple detail. Research by Professor Abigail Thernstrom found that:
[I]n 1998, the average seventeen-year-old African American could only read as well as the typical white child who had not yet reached age thirteen. In 1992, just 18 percent of black students in twelfth grade were rated ‘proficient’ or ‘advanced’ in reading, as compared with 47 percent of whites. As of 1998, those numbers were unchanged. . . . The most recent data thus show black students in twelfth grade dealing with scientific problems at the level of whites in the sixth grade and writing about as well as whites in the eighth grade.[22] This poor primary and secondary school academic achievement is reflected in the performance of college-bound black seniors who took the Scholastic Aptitude Test (SAT). Nationally the combined average SAT for blacks in 2008 was 1276 out of a possible 2400; whites and Asians scored respectively 1581 and 1623.[23]
Given those statistics, race might be a fairly good, though not perfect, indicator of expected worker productivity. One should therefore not be surprised that employers take race into account when assessing worker productivity. To observe a process that selects in part by race and to attribute the selection to preferences (in this case, to employer “racism”) may be misleading. One example of this misunderstanding occurs when researchers perform “paired testing” as a means to discover racial discrimination in employment, housing, or insurance. Paired testing matches individuals in terms of such qualifications as age, education, experience, credit record, and other factors deemed important to an employer, landlord, or insurance agent. They differ only by race or sex, and researchers get to see whether they are treated differently. Authors Michael Fix and Margery Turner say that “Paired testing is an excellent vehicle for understanding and measuring actual discrimination understood here simply as the practice of treating people differently because of their membership in a protected group.”[24]
Objective qualifications are only part of the information an employer or landlord wants to know prior to making a decision about whether to employ or rent to an individual. An employer or landlord might deem other information—such as trustworthiness, promptness, congeniality, respect for property, and other personal attributes—an important part of the decision. Whether a person is white or black, objective qualifications do not adequately convey information about those attributes. Prior employment or rental references might not be reliable, because previous employers or renters might fear legal reprisals for giving candid references. Employers or landlords could hire private detectives to get such information, but might deem that too costly and therefore resort to cheaper methods: hunches, race or sex, word-of-mouth opinion, and existing references from employers.
There are demographic differences between races that increase the likelihood that race can be used as a proxy for more costly-to-observe attributes. To assume that because a white and a black both have a high school diploma or a bachelor’s degree, they have equal levels of academic proficiency ignores reality. There are significant differences by race as seen in SAT scores for admittance to college and on tests such as GRE, LSAT, and MCAT that are used in admissions to graduate and professional schools. The racial gap in scores on these tests suggests that on the average a high school diploma or a BA degree held by a black is not the same as one held by a white. None of this is to say that every black with a high school diploma or a BA degree shares the group characteristics, but how does an employer know?
The auto insurance industry provides a nonracial example of the use of physical attributes, Drivers under twenty-five years of age are charged routinely higher premiums. But one doubts that auto insurance companies can be charged with disliking young people. Life insurance companies charge women lower premiums. Can we reasonably assert that life insurance companies like women and dislike men? In both cases, a physical attribute acts as a proxy for an unobservable attribute. In the case of driver age, there is a higher probability of accident claims and, in the case of the insured’s sex, longer life expectancy.[25]
Suppose an employer has racially neutral preferences and is uncertain about average black worker productivity relative to whites. What would encourage or discourage him from seeking more information and experimenting with and perhaps revising his beliefs? Laws or practices that require him to pay all workers identical wages and laws that make it very costly to fire a worker reduce employer incentive to experiment. A nonracial example might be helpful.
Imagine that a new supermarket, unfamiliar to local customers, locates in an area of established markets. The latter have considerable customer loyalty. How can the new arrival induce customers to try it out? The method typically employed is to offer sales, perhaps also prizes, coupons, and other amenities, all of which are effective means of attracting customers.
Just as easily, customers can be given an inducement not to experiment. Imagine there was legislation requiring that all supermarkets charge identical prices and sell their goods on identical terms. That being the case, one cannot think of much inducement for customers, who already have loyalty to established supermarkets, to experiment with the newcomer. Imagine there was also a law stipulating that once a customer chose a market, he had to stick with it or go through very costly procedures to change. In other words, if he was dissatisfied with the supermarket, he could not cheaply “fire” it by taking his business elsewhere. It is very easy to predict the response of customers to the new supermarket: why experiment and incur the risk?
Continuing with this scenario, one should not be surprised if incumbent supermarket owners invested in lobbying for the creation of a law mandating sales on identical terms. It would be to their financial advantage to reduce customer incentive to experiment by making experimentation more costly. Moreover, political justification for the promotion of these laws could be couched in glowing, civic-minded terms like the prevention of cutthroat competition and encouraging level playing fields.
In some respects, this example can be applied to labor markets. Labor laws or collective bargaining agreements that mandate equal starting wages reduce employer incentives to experiment. Similarly, anti-discrimination regulations make experimentation costly. Once a black employee is hired, it may be very expensive to fire him if the employer deems he has made a mistake. The employee’s lawyer could tie up much of the company’s resources in litigation. Whatever increases the cost of firing an employee also has the effect of increasing the cost to hire.
Home Mortgage Discrimination
A study of mortgage-bank lending practices reported that black and Hispanic applicants were denied credit in greater proportions than white, even when income was comparable. Moreover, the study reported similar disparities in application denials among neighborhoods classified by race and income.[26] A staff writer for the The Wall Street Journal interpreted the report as follows: “[I]f you’re black, it’s twice as likely your mortgage application will be rejected as it is if you’re white. And if you live in a low-income neighborhood, chances are that many lenders have little interest in mortgage-lending in your community anyway.”[27] Referring to the study, economist Julianne Malveaux wrote that blacks continue to be segregated “because of well-documented discrimination in mortgage lending.”[28] Jesse Jackson said it is criminal that banks “systematically discriminate against African-Americans and Latinos in making mortgage loans.”[29]
While rejections of loan applications differ significantly between the groups discussed above, we cannot unambiguously assign the difference to racial preferences; there might be some real differences. The Home Mortgage Disclosure Act data used by the study has serious limitations. It does not contain information about applicant assets and creditworthiness. An important element in assessing the creditworthiness of individuals is their net worth. Defined as the difference between the value of assets and debts owed, net worth differs significantly between blacks and whites. As of 1984, white households had a median net worth of $39,140, while that for black households was $3,400.
Net-worth differences between blacks and whites vary a great deal regardless of income. In the $2,000 to $3,999 monthly income category, white householder net worth was $50,529; for blacks, $15,977. In the $4,000 income per month and more, white net worth was $128,237, and for blacks, it was $58,758.[30] In 2000, white households had a median net worth of $79,400, black households $7,500.[31] In the third (middle) income quintile, white net worth was $59,500, while that for blacks was $11,500. In the fifth (highest) income quintile, white net worth was $208, 023, and for blacks it was $65,141.[32]
Given the history of blacks in the United States, it should not be surprising that they have not accumulated, through work or inheritance, as many assets as have whites. But that fact in no way implies racial discrimination by banks as an explanation for racial differences in loan approval rates. Since net worth is one of the determinants of creditworthiness, it suggests that blacks as a group will qualify for fewer loans, even in a society with neutral racial preferences.
Racial preferences as an explanation for the higher incidence of loan denial to blacks and Hispanics loses some of its appeal when we examine the data further. The following are the loan denials by percent for conventional loans, according to a 1992 Federal Reserve Bulletin: blacks (38), American Indian (27), Hispanics (27), whites (17), and Asians (15).[33] Ignored in the debate over the mortgage application-approval gap between whites and blacks is whether white-owned mortgage banks also discriminate against whites, since the banks approved a greater percentage of mortgage applications of Asian-Americans than whites. No doubt the explanation is not racial preferences against whites at all. Japanese and Chinese-Americans are on the average financially more successful and have better credit ratings than do whites on average. When Alicia Munnell, author of the Federal Reserve study, was interviewed by two Forbes reporters, she conceded that the report did not sufficiently take into account differences in creditworthiness, adding: “I do not have evidence . . . [n]o one has evidence” of lending bias.[34]
Another factor to consider is that if banks maintained a policy of stiffer loan approval requirements for blacks, the default rates among blacks would be lower than those for whites. The Federal Reserve study found that black and white default rates are roughly equivalent.
In 1999, Freddie Mac, the huge mortgage lender, conducted a consumer credit survey. In addition, Freddie Mac and the historically black colleges and universities convened more than fifty focus groups. The survey found that nearly twice as many blacks as whites have bad credit. Specifically, 47 percent of blacks and 34 percent of Hispanics had bad credit, compared to 27 percent of whites. Among people with incomes under $25,000, the figures were 48 percent for blacks and 31 percent for whites; among higher income-earners ($65,000 to $75,000), 34 percent for blacks and 20 percent for whites.[35]
Racial discrimination as an explanation for differences in loan rejections loses more credibility when assessed by the findings of a 2002 study, “A Test of Cultural Affinity in Home Mortgage Lending,” conducted by Raphael W. Bostic, former economist at the Federal Reserve Board in Washington.[36] Professor Bostic surveyed forty minority-owned and 106 white-owned banks. He found that, in 1995, minority-owned banks rejected black loan applicants at a rate double that of their white-owned counterparts: 35 percent of black loan applications were denied at minority-owned banks and 17 percent at white-owned banks. The denial rate for Latino applicants at minority-owned banks was 16 percent versus 13 percent, respectively. In the case of Asian loan applicants, bank ownership had little or no effect on denial rates. (Black-owned banks historically have invested a greater percentage of their loan portfolio outside the community in which they are located.)[37]
Subprime Mortgage Problem
Between 1994 and 2003, overall home ownership grew from 64 to 68 percent of the population, as the residents of 9 million more households became homeowners. The greatest gains in home ownership occurred among blacks (1.2 million) and Hispanics (1.9 million). Nearly one-half of black and Hispanic households are homeowners compared to the white rate of 71 percent.[38]
A major stimulus to minority home ownership has been increased access to credit through the 1980 Depository Institutions Deregulatory and Monetary Control Act’s elimination of usury controls on first-lien mortgage rates and the 1990s’ strengthening of the provisions of the Community Reinvestment Act of 1977; the latter is a federal law that requires banks to offer credit throughout their entire market and discourages them from restricting their credit services to high-income areas. The Federal Housing Administration, which guarantees mortgage loans, liberalized its rules for guaranteeing mortgages. These and other factors combined to give lenders financial inducement to make mortgage loans to those who posed an elevated credit risk and had weak credit histories. In contrast to the past, instead of a lender denying riskier applicants a mortgage, he made the loan at a higher interest rate.
Blacks and Hispanics pose higher credit risks. Thus, one should not be surprised to see them disproportionately represented among subprime borrowers. Blacks make up 27 percent of the number of subprime home-purchase loans, and 21 percent of subprime home-equity loans, while Hispanics, respectively, make up 20 and 15 percent; those figures compare to whites, who respectively make up 7 and 6 percent. As of 2003, the foreclosure status and serious delinquency on prime loans were, respectively, .48 and 1.1 percent; those for subprime loans, 3.4 and 7.4.[39]
According to an article in the Atlanta Journal-Constitution, “Black Atlantans often snared by subprime loans,” a national study of credit scores—not just mortgage loan applicants—found that 52 percent of blacks have scores that would classify them as subprime borrowers compared with 16 percent of whites.[40] Such a finding contradicts the suggestion of racial discrimination in a New York Times editorial, which said that “blacks and Hispanic borrowers were far more likely to be steered into high-cost subprime loans than other borrowers, even after controlling for factors such as income, loan size and property location.”[41]
Prior to the emergence of the subprime market for home mortgages, lending institutions rationed credit by denying risky loans and those involving some property locations, a practice that became known as redlining and was attacked as racial discrimination. Interestingly, now that legal ceilings on home mortgage interests have been relaxed, enabling less creditworthy people to secure loans, the racial discrimination charges focus on higher loan-default rates experienced by people who would have otherwise not been able to secure a loan.
Public policy directed at mortgage-lender racial discrimination will miss its mark and may, like affirmative action in lending, possibly exacerbate the credit problems of blacks. Banks will simply move away. A more effective policy would attempt to reduce the high cost and risks of lending in inner-city neighborhoods. When what is prejudice (as we have defined it) and response to real differences are misdiagnosed as racial preferences, misguided policy will emerge.
Discrimination against Low-Income Shoppers
During the mid-1960s, it was widely alleged that white merchants in ghetto areas exploited their customers by charging higher prices and selling lower-quality merchandise. The merchants, it was said, were seeking to earn above-normal profits as a way of acting out their racial hostility toward blacks.[42] But it turns out that racial hostility by merchants could not adequately explain the higher prices.
They were indeed higher in black neighborhoods, and several studies showed that retail food chains followed different pricing policies between them and white neighborhoods. With these findings in hand, the Federal Trade Commission (FTC), along with consumer advocacy groups and public-interest lawyers, attempted through public pressure to oblige merchants to offer black and white customers the same terms.
To view the merchants’ behavior as exploitative or racist ignores the fact that black neighborhoods tend to be high-cost business environments. Losses from business-related crime are higher there than elsewhere as a percentage of total sales; business, fire, and theft insurance premiums are also higher; and it is riskier to extend credit. In addition, because of the low income of many black residents and the effect of that on sales mix and volume, merchandising techniques used to lower sales costs in higher-income neighborhoods are not as readily adaptable to low-income areas.
High crime rates figure in another, unappreciated way. One of the goals of a supermarket manager is to maximize the rate of merchandise turnover per square foot of leased or owned space. When theft is relatively low, the supermarket can make use of space along the exit aisles and entryways. Often merchandise such as plants, fertilizer, and other home and garden items is placed outside of the store and sometimes left overnight. Therefore, a merchant can use all of the space he pays for, which raises his profit potential. That merchandising technique is unavailable in localities where there is less overall honesty. Merchandise cannot be located near exits and out of sight of store employees. If it is placed outside, the merchant must bear the expense of having it guarded, and leaving it out overnight is out of the question. Unlike low-crime neighborhoods, the merchant must pay for square footage he cannot use, thus lowering his profit potential. In addition, it is not unusual to see uniformed guards on duty at supermarkets in high-crime areas.
Much of the behavior that critics have condemned is therefore merely an economic response to an environment that has a higher cost of doing business. If products and services are to be provided in higher-cost neighborhoods, prices must reflect that cost. Evidence substantiates this explanation of merchant behavior. The FTC has shown that while gross margins were higher in black neighborhoods, the difference was more than accounted for by higher operating costs and a lower return on equity.[43] The assertion that above-normal profits were earned becomes even less credible when we recognize that retailing is characterized by relative ease of entry, so that if above-normal profits existed, one would expect to see merchants opening new businesses until profits in black and white neighborhood areas were equal. The opposite of this has occurred in urban areas: businesses have left without being replaced.
The crusade that blamed the problem of the black consumer on the greed and racism of whites may well have had a negative impact on consumer welfare. The adverse publicity and boycotts (and other actions) against merchants in black communities gave merchants increased incentive to move out. The result is fewer neighborhood stores, with shoppers being forced to travel longer distances or pay even higher local prices than in the past.
High crime takes a toll on consumers in another way. In low-crime neighborhoods, FedEx, UPS, and other delivery companies often leave merchandise at the door if no one answers. In high-crime neighborhoods, that is not an option. The companies must bear the costs of making return trips, or the customer has to bear the cost of traveling to pick up his goods.
Conclusions
In today’s America, there is a broad consensus that race-based discrimination in many activities is morally offensive and in many cases rightfully illegal—as it should be when there is taxpayer-based provision of such goods and services as public schools and universities, libraries, and social services. Even though people should be free to deal with, or refuse to deal with, anyone in private matters, there is little evidence that race-based discrimination is widespread in today’s America. After all, there is a difference between what people can do and what they find it in their interest to do. That conclusion is suggested by laws that once codified racial discrimination in the United States and elsewhere. In this country, there were antimiscegenation laws and restrictive covenants. During South Africa’s apartheid era, there were also job-reservation statutes and others that reserved amenities such as theaters, restaurants, and hotels for whites only.
One of the first implications of the existence of a law is that not everyone will voluntarily behave according to its specifications. Otherwise, there would be no need for the law. After all, to the writer’s knowledge, there is no legal requirement that people eat or avoid tossing their weekly earnings onto the street. Yet we need not worry because most people will not find it in their private interest to do either.