While their discussion emphasized the “unwanted child” theory, Donohue and Levitt never separated it from the eugenics approach, which was left without refutation in their work.27 Indeed, Donahue and Levitt seem to have deliberately avoided the racial implications of their own theory, as their inquiry is one of the few academic studies on crime by economists that doesn’t account for these demographics. Perhaps this was just too explosive to mention; after all, who would dare to state that abortion lowers crime rates by reducing the population of poor African Americans?28
The relationship between abortion and crime is complex. We must begin by considering the circumstances and expectations under which a woman becomes pregnant. Remember the fundamental principle of economics: if something becomes more costly, people do less of it. If abortion is illegal, the “cost” of sex is relatively high due to the possibility of pregnancy. No method of birth control is 100 percent effective, and without the option to have an abortion as a last-ditch safeguard, having sex carries a risk. When contemplating having pre-marital sex, women know that they might have to bear and raise a child, possibly on their own. Likewise, men know that they might end up having to support a child, and both know that having a baby could create pressures on them to marry even if they don’t want to do so. Consequently, both men and women tend to be more reluctant to engage in casual sex, especially unprotected sex, when abortion is illegal.29
In contrast, if abortion is legal, the incentives are different. Knowing that the abortion option is there to save them from raising an unexpected child, women who are willing to have an abortion—as well as men in general—become less concerned with contraceptives and more likely to engage in premarital sex. As more women have premarital sex, social mores become more accommodating to the practice. This creates social pressure on other women to have premarital sex, including women who would never have an abortion. Increasing rates of premarital sex among these latter women leads to higher pregnancy rates. The result is rising numbers of women who are single, pregnant, and unwilling to have an abortion.
Indeed, multiple studies have shown that legalized abortion, by raising the rate of unprotected premarital sex, increases the number of unplanned births, even outweighing the reduction in unplanned births due to abortion.30 From the early 1970s, when abortion was liberalized, through the late 1980s, there was a tremendous increase in the rate of out-of-wedlock births, rising from an average of 5 percent in 1965-69 to over 16 percent twenty years later (1985-1989). For African Americans, the numbers jumped from 35 percent to 62 percent. While not all of this rise can be attributed to liberalized abortion rules, it was nevertheless a key contributing factor.
Let’s return to the personal level. In an environment of legal abortion, a man might well expect his partner to have an abortion if a sexual encounter results in an unplanned pregnancy. But what happens if the woman refuses? Maybe she is morally opposed to abortion, or perhaps she thought she could have an abortion, but upon becoming pregnant, she decides that she can’t go through with it. What happens then?
Many men, feeling tricked into unwanted fatherhood, will likely wash their hands of the affair altogether, thinking “I never wanted a baby; it’s her choice, so let her raise the baby herself.” What is expected of men in this position has changed dramatically in the last four decades. The evidence shows that the greater availability of abortion largely ended “shot-gun” marriages, where men reacted to an unplanned pregnancy by doing the honorable thing—marrying their partner. But with abortion as a legal option, men became more reluctant to stay with a woman who refuses to have one.31
What happens to these babies of reluctant fathers? The mothers often end up raising the child on their own, as single mothers have become much less likely to give their children up for adoption. Even as out-of-wedlock births have surged, adoption rates have plummeted. In the two decades before Roe, 19.3 percent of babies born to unwed white mothers were placed for adoption and 1.5 percent for unwed black mothers. From 1973 to 1981, those percentages fell to 7.6 and 0.2 percent, respectively, and continued to decline thereafter.32 How can we explain this? As stated above, after Roe, a rising number of out-of-wedlock children were born to women who opposed abortion, or at least would not have one themselves. These women have also proved less willing to give up their children for adoption.
With work and other demands on their time, single parents, no matter how “wanted” their child may be, tend to devote less attention to their children than do married couples; after all, it’s difficult for one person to spend as much time with a child as two people can. Grand-parents and other relatives may sometimes help out, but on average, the children of single parents still receive less care.
The children of unmarried, cohabitating couples also receive less care and attention than married couples’ children. This is partly because cohabiting partners face much higher separation rates than married couples do.33 Many cohabitating couples are not married because they are not yet ready to make a lifetime commitment to each other. And if partners believe their relationship may not be permanent, they are less likely to make career adjustments or other sacrifices for the sake of their family. Thus, single parents and unmarried couples are less likely than married parents to read to their children or take them on excursions, and more likely to feel angry at their children or feel that their children are burdensome. Children raised outside of wedlock experience a higher rate of social problems in nearly every area than children of married couples. Unsurprisingly, children from unmarried families are also more likely to grow up to commit crimes.34
Percentage of Children with Different Types of Problems by Marital Status of Parents35
So the opposing arguments are clear—one stresses that abortion eliminates “unwanted” children, while the other emphasizes that abortion increases out-of-wedlock births. Both effects, conceivably, could be occurring at the same time. The question is: which one has the bigger impact on crime?
This must be answered empirically. Unfortunately for advocates of the “abortion decreases crime” theory, Donahue and Levitt’s data were undermined by methodological flaws. As The Economist magazine noted in an article entitled “Oops-onomics,” “Donohue and Levitt did not run the test that they thought they had.”36 Work by two economists at the Boston Federal Reserve, Christopher Foote and Christopher Goetz, found that when the tests were run correctly, they indicated that abortion actually increases violent crime.37 I co-authored a study with John Whitley that found a similar connection between abortion and murder—namely, that legalizing abortion raised the murder rate, on average, by about 7 percent.38
We find particularly troublesome problems with the “abortion decreases crime” theory when we analyze the population according to age group. Suppose that liberalizing abortion in the early 1970s can indeed explain up to 80 percent of the drop in murder during the 1990s, as Donohue and Levitt claim. Then the impact of deregulating abortion, undoubtedly, would first reduce criminality among age groups born after the abortion law was changed, when the “unwanted,” crime-prone elements of these groups began to be weeded out through abortion. Yet, looking at the declining murder rate during the 1990s, Whitley and I found that this is not the case at all. Instead, the rate of committing murder began falling first among an older generation—those twenty-six and older—who were born before the Roe v. Wade decision.39 It was only later that criminality among those born after Roe began to decline as well.
Comparing the Rate of Murders Committed by Those Born in the Four Years Before and After Roe v. Wade for the 45 States and the District of Columbia Affected by the Ruling
This pattern is more consistent with the theory that legalizing abortion led to a rise in crime. In fact, those born in the four years after Roe were much more likely to commit murder than those born in the four years prior to Roe. This was especially the case when they were in their teens—in other words, in their “criminal prime.”
And that’s not all. The “abortion decrease
s crime” argument encounters further inconsistencies when we compare U.S. crime and abortion trends to those in Canada. While crime rates in both the United States and Canada began declining at the same time, Canada liberalized its abortion laws much later than the U.S. did. Although the province of Quebec effectively legalized abortion in late 1976, it wasn’t until 1988, in a case originating in Ontario, that the Canadian Supreme Court struck down limits on abortion nationwide.40 If the legalization of abortion in the U.S. caused crime to begin dropping eighteen years later, why did the crime rate begin falling just three years after the comparable legal change in Canada?
In sum, even if one effect of abortion were to lower crime by culling out “unwanted” children—a conclusion derived from flawed statistics—the effect is greatly outweighed by the rise in crime that abortion causes by increasing out-of-wedlock births. It should be noted that African Americans are disproportionately harmed by the crime stemming from legalized abortion. That population has seen the biggest increases in abortion, premarital sex, and out-of-wedlock birth rates, resulting in more African Americans being raised by single parents and eventually committing crimes, mostly against other African Americans.41
Thus, “legalized” abortion actually served to increase crime since the 1970s. However this effect—both among African Americans and the population at large—was more than offset by other factors that caused the massive drop in crime of the 1990s. Before discussing these aspects, let’s look at one other factor that proved counter-productive in fighting crime.
What Increased Crime? Part II
Affirmative Action Hiring in the Police Force
Many police departments implemented affirmative action policies in the 1990s, just as the crime rate was entering a steep decline. But a close study reveals that crime fell despite these programs, not because of them. Perhaps most unexpectedly, these policies not only resulted in the hiring of less qualified women and minorities, but also of less qualified white applicants as well.
We must specify at the outset that crime rates did not rise due to the hiring of more women and minorities per se. Rather, the fault lies with the particular affirmative action rules that were adopted. There is little doubt that adding women and minorities to an all-male, all-white police force carries substantial benefits. Minority police officers often function more effectively than whites in minority areas. Since minority residents tend to vest more trust in minority officers, they are often more forthcoming with information that leads to arrests and convictions of criminals. Especially if the minority police officers grew up in the community they are patrolling, they may also be better at predicting the behavior of criminals in those areas. Furthermore, minorities are essential for various kinds of undercover operations, such as infiltrating race-based gangs. Likewise, female officers tend to elicit more information and honest reports of rape and spousal abuse from female victims, and they are much more able than men to carry out certain operations such as prostitution stings.
But affirmative action programs are not simply focused on achieving these benefits of a diverse police force. Instead, the trend has been toward boosting minority representation in a police department so that it reflects the demographic ratio of the surrounding community. The two goals are not necessarily the same, and the methods used to achieve the latter objective, unfortunately, have reduced the effectiveness of police in stopping crime.
Affirmative action policies have sought to transform traditional police hiring standards that rely on intelligence exams, strength tests, and criminal background checks. Because, on average, women are less likely to pass strength tests than are males, while African Americans have lower passing rates on intelligence exams and criminal background checks than whites, many police departments adopted new standards in an effort to increase minority hiring.
Police affirmative action programs have entailed two main approaches. The first is to lower testing standards across the board until they produce equal pass rates among minority and non-minority groups. This approach has become increasingly common over the last few decades. A leading method here has been to replace cognitive or intelligence exams with more nebulous psychological tests that aim to gauge a candidate’s temperament. Asking questions such as what is an applicant’s favorite color and whether he watches much television, the tests are designed to produce equal pass rates across different groups of applicants.42 According to a 1993 survey of twenty-three large police and sheriff’s departments, twenty departments had reduced their emphasis on cognitive skill testing due to the tests’ “adverse impact” on minority hiring.43 The other three departments had completely eliminated cognitive testing in hopes of increasing minority recruitment. An example of the meager new hiring standards is that to pass the reading test, “applicants had to score only as well as the bottom 1 percent of current police officers.”44
Louisiana provides a particularly instructive example. There, the Police Department discarded entrance exams in response to a lawsuit filed by the Department of Justice. The objections were that 66 percent of whites passed the test compared to just 25 percent of African Americans, and that the test did not relate to the abilities required for the job. The lawsuit was weak, failing even to explain what parts of the test were unrelated to the job. Moreover, a federal judge had previously ruled that the test did not discriminate against minority applicants to the police and fire departments in another city. Nevertheless, rather than incur the cost of litigation, the Louisiana State Police dropped the test and agreed to pay $1 million to African Americans who had failed the exam. The department even vowed to hire eighteen new officers from among this group 45
In some cities, such changes have resulted in myriad problems caused by a less-educated police force. For example, between 1986 and 1990, 311 of the 938 murder cases that the Washington, D.C. police brought to the U.S. attorney’s office were dismissed. One local prosecutor commented that “many D.C. cases were thrown out because prosecutors couldn’t read or understand the arrest reports [written by the police].” The officers simply lacked the ability to write comprehensible English.
But lower standards do not guarantee the desired pass rates among all races. In Chicago, the city paid $5.1 million for consultants to develop “unbiased” exams, only to have unacceptable numbers of minorities once again fail the tests. The city then moved to a heavily weighted seniority system for promoting police officers and a lottery system for hiring firefighters in order to ensure the correct racial composition of new classes of recruits.46
The second method for implementing affirmative action programs in police departments is “norming.” This assigns different standards to different groups of candidates in order to ensure similar pass rates. Norming is frequently used for female recruits, both to the police and to the military. For example, in the military, women are required to run two miles in eighteen minutes and fifty-four seconds, while men have sixteen minutes to run the same distance. Women have two minutes to do eighteen push-ups and two minutes for fifty sit-ups, while men must do forty-two push-ups and fifty-two sit-ups within the same time.47
Norming inevitably leads to the hiring of women who are, on average, physically weaker than men. Physical strength testing of public safety employees consistently finds large differences between men and women; women’s upper body strength ranges from 44 to 68 percent of men’s, while their lower body strength is typically 55 to 82 percent of men’s.48 Weaker officers face some obvious disadvantages: it is more difficult for them to chase and catch fleeing suspects or to control a resisting suspect without resorting to a weapon. Furthermore, an influx of weaker officers can affect police procedures. For example, police departments come under pressure to end patrols by single officers, as well as to reduce foot and bicycle patrols in favor of car patrols.
The risk inherent in hiring weaker officers is demonstrated by the case of Brian Nichols, a thirty-three-year-old, 196-pound rape defendant. In an Atlanta courthouse, Nichols overpowered his guard, seized her gun
, and used it to kill a judge, a court reporter, a police officer, and a federal agent. Nichols’ guard, a sheriff’s deputy, was a five-foot-two, fifty-one-year-old woman. “Why was a tiny woman, or any woman, given such a job?” asked Mary Ellen Synon, a columnist for the Mail on Sunday. “Because the Atlanta police force, like many others in America, has been subjected for years to government demands for ‘gender and minority’ balance; changing hiring rules and lowering standards so more women and people from ethnic minorities can join up.”49
Are such occurrences merely isolated instances, or do they speak to a larger problem created by affirmative action policies? To answer this question, I gathered statistics to analyze how changes in hiring rules and the demographic composition of police departments affect crime, arrest, and conviction rates.50 I included detailed demographic, income, and socio-economic information to help explain changes in crime and arrest rates. I also considered related factors such as illegal drug prices, gun laws, and various policing policies.
The results were dramatic: crime rates jumped in cities using affirmative action policies that lowered testing standards. Interestingly, however, the use of norming had much less harmful results.
The implementation of consent decrees—agreements by local police departments to use affirmative action in hiring and promotions—increases the rate of murder, other violent crimes, and property crimes. Overall, using affirmative action to achieve a one percentage point increase in African American officers on the force is associated with an increase in murders of at least 2 percent, violent crime of almost 5 percent, and property crimes of 4 percent.51
Freedomnomics: Why the Free Market Works and Other Half-Baked Theories Don't Page 13