Volume 2  Number 5
May 2000

    Bias-motivated crime has unique characteristics. As in heterosexual rape, victims and offenders come from different groups. Unlike rape, however, hate crime is reciprocal. Each group can prey upon the other. Though not obvious, these singular aspects incline the data in a unique way. The sizes of victim and offender groups influence victimization rates in a way that is often more significant than intrinsic group bias. Methods are developed for interpreting hate-crime statistics. They are applied to recent FBI data.

    Would you believe that a black in the US is about 20 times more likely to be a victim of hate crime than a white? This is not the claim of a left-wing crazy. It comes from data in the 1998 hate-crime report of the FBI. Data from the two previous reports yield about the same odds. We will try to put the facts in perspective.

In 1990, Congress passed the Hate Crime Statistics Act, requiring the Justice Department to collect and publish annual statistics on crimes that "manifest prejudice based on race, religion, sexual orientation, or ethnicity." To comply, the FBI collects data submitted voluntarily by local law-enforcement agencies, and assembles them into an annual summary report. Most of the analysis you will find here pertains to the numbers found in these reports.

Anyone who has pored over Government documents, knows first-hand how agencies can manipulate data to make a point. Crime statistics are a good example, the treatment of hate-crime data being especially egregious. The Justice department has wide latitude in how they comply with the Hate Crime Statistics Act. Accordingly, it has bent the data toward its own ends by omitting categories for ethnic offenders. Thus, Hispanics cannot be hate criminals, only hate victims. When a Hispanic commits a hate crime, he is counted as white. When he is a victim, he becomes Hispanic. In this way the FBI pads the number of white offenders. Despite this baggage we can learn much from the FBI data. By focusing on victims, we can sidestep Justice Department attempts at obfuscation.

Minorities suffer simply because they are minorities.

A blueprint for hate-crime data
    In a nation of more than 270 million inhabitants, composed of various ethnic, racial and religious groups, targets for hate crime abound. Anyone inclined to commit a hate offense will have no trouble finding a victim. The Census Bureau reports that as of July 1, 1998 there were approximately 223.0 million whites residing in the U.S., 34.4 million blacks, 30.4 million Hispanics, 10.5 million Asians and Pacific Islanders, and 2.4 million Native Americans, not to mention Protestants, Catholics, Jews and Moslems. The target is enormous. For practical purposes it is infinite.

If there are 1,000 hate criminals in an offender group, there will be 1,000 or more victims in the victim group. Regardless of the victim group's size, 1 million, 10 million or 50 million, the 1,000 hate criminals in the offender group will commit the same number of offenses. For a given hate-crime proclivity, the size of the offender group determines the number of its offenses. All things equal, a majority group produces more offenders than a minority group. Differential group tendencies not withstanding, a majority group simply has more members.

The size of the victim group is also important. It determines how well the group absorbs the crime directed toward it. If a thousand crimes are perpetrated against a group of one million and a like number against a group of ten million, each member of the smaller group suffers a tenfold greater risk. Consequently, with respect to hate crime, minorities are at a double disadvantage having nothing to do with differential bias. A large dominant group produces many offenders, whose crimes must be borne by relatively few in a minority victim group. In other words, minorities suffer simply because they are minorities. This is a mathematical reality having nothing to do with differential group bias. One should always view hate-crime evidence against this backdrop.

Inverse square risk
    For simplicity, assume a black and white universe. This is not too bad a model for the US, because blacks and whites combine to form more than 95 percent of the population. From our previous discussion, we know the probability of a white being hate-victimized in a given year is proportional to the black population and inversely proportional to the white population. An analogous relation exists for the probability of a black being hate-victimized.

Quadratic dependence on group size makes it the major determinant of the black to white risk ratio.

Suppose our black and white universe contains NW whites and NB blacks. Let pW and pB be the probabilities, respectively, of a white or black being victimized in a given year. Assuming the average number of victims per offender is constant across the two groups, we can write,


The quantities, kB and kW , are constants closely related to the respective probabilities that a black or a white is a hate-criminal. If there were one victim per offender, the constants would be precisely these probabilities. We call kB and kW the black and white hate proclivities, respectively. They are intrinsic group properties. From (1) we can write the (per capita) risk ratio, pB /pW :



And we have the interesting result that the per capita risk ratio varies inversely with the hate-proclivity ratio and also inversely with the square of the group size ratio. Quadratic dependence on group size makes it the major determinant of the black to white risk ratio. As the disparity in group size grows, minority group members rapidly become more vulnerable. At the same time, members of the dominant group become safer.

In the special case where both groups have equal proclivities for committing hate crime, kW = kB, and



That is, assuming equal hate proclivity, the per capita risk ratio is the inverse square of the group population ratio.


We need to ask . . . not why the per capita risk ratio of blacks to whites is so high, but rather why is it so low?

Equal hate proclivity hypothesis
    We began by noting that a black in the US is about 20 times more likely to be a victim of hate crime than a white. The actual figure from the 1998 FBI data is 21.8. We can now correctly interpret this observation. Because the Justice department skews the data by excluding Hispanics from offender status, we cannot achieve the level of accuracy we would like. However, we can come pretty close by considering only victims and excluding Hispanics from the analysis. The FBI includes hate crime against property in their tabulation, and counts "offended" properties as victims. We consider here only crimes against persons.

If you were one of the 195.4 million non-Hispanic whites in the U.S. in 1998, there were 32.7 million non-Hispanic blacks potentially ready to abuse you in some way and vice versa. For the moment assume that racial and ethnic groups share equal tendencies to abuse members of other groups. That is, if one in every thousand whites has this proclivity, then also one in every thousand blacks or Native Americans or Asians has it as well. We call this the equal hate-proclivity hypothesis. Under it, we seek the relative hate-crime risk to groups of various sizes. According to (3), under the equal hate proclivity hypothesis, a black should have been (195.4/32.7)2 or 35.7 times more likely than a white to be a victim of hate crime in 1998. Put in this light, we see that in reality the relative risk of blacks (21.8 times that of whites) was less than expected assuming no differential group bias. In fact, the black risk was only 61 percent of that expected from group size considerations alone. We need to ask, therefore, not why the per capita risk ratio of blacks to whites is so high, but rather why is it so low? The answer is simple: The equal hate proclivity hypothesis is false.

Differential hate proclivities
    We can calculate the hate-proclivity ratio from the FBI data. In 1998, excluding Hispanics, we find an observed per capita risk ratio of 21.8, and from the census, a population ratio square, (NW
/NB )2, of 35.7. Then from (3), the ratio of hate proclivities, kB /kW is 35.7/21.8 = 1.6. That is, if blacks were 1.6 times more likely to commit hate crimes than whites, the FBI data are explained.

If we also know the number of black and white victims, NVB and NVW , respectively, we can evaluate the individual hate proclivities. From (1), the probability of a black being victimized by a white is kWNW /NB . This quantity is also the rate of black victimization, or NVB /NB . Consequently,




The FBI reports the number of hate-offenders by race, though the white offender entry is inflated by the inclusion of Hispanics. For 1998, the report lists 1303 suspected black offenders and 2084 suspected white offenders. This yields a per capita offender ratio (B/W) of 1.7, in reasonable agreement with the hate-proclivity ratio of 1.6. Table 1 summarizes these calculations for the three most recent FBI reports.


Hate-proclivity ratio
kB /kW )

Per capita offender ratio
(black / white)










Table 1. The hate-proclivity ratio compared with the per capita offender ratio.


Church burnings

    Between January 1995 and November 1996 the burning of black churches in the South created a sustained news frenzy. By the summer of 1996, Americans had been treated to over 2,000 articles in major newspapers, many on the front page. Not to be outdone, church arson lead all other stories in the TV nightly news.

Political types of various stripes chimed in. Ralph Reed, then Director of the Christian Coalition, termed the arsons, "the greatest outbreak of violence against the black church since the height of the civil rights movement." Deval Patrick, assistant attorney general for civil rights, proclaimed the fires to be "an epidemic of terror." President Clinton, in one of his weekly radio addresses, recalled in a now famous evocation, "vivid and painful memories of black churches being burned in my own state when I was a child." (It was later discovered that no black church had been burned in Arkansas during his childhood.) Clinton called a conference of Southern governors to deal with the burnings. He toured burned-out churches, once on his 50th birthday. The press tagged along. Congress, not to be upstaged, passed the Church Fire Prevention Act of 1996, making church burning a federal crime.

Not everyone bought in. Michael Fumento, writing in the Wall Street Journal, analyzed the data and found that much of it was spurious. He showed that there had been no increase in church arson in the South from 1990 to 1995. Fumento noted that in 1995, USA Today reported 45 arsons against white churches, compared to 27 against black churches. He also observed that the 1996 figures were inflated by copycat crimes.

Eventually, numbers began to roll in indicating that more white than black churches had been torched. It did not make much difference to some in the press. Paula Walker, vice president and news director of WNBC-TV, reacted to the reports while attending a National Association of Black Journalists convention in August 1996. She concluded that, "There didn't seem to be much substantiation other than raw numbers." 

. . . while whites were being blamed for burning black churches, and were drowning in their own guilt, the facts reveal a black was 5 times more likely than a white to commit bias-motivated church arson.

We can put the Church-burning data under the microscope. Our foregoing analysis, with some modification, is well suited to the task. Faced with the fact that bias-motivated crime is only a small fraction of total crime, hate-crime activists fall back on the position that these offenses are underreported. But the charred ruins of a burned-out church cannot go unreported. We include all the burnings in our analysis, including copycat arsons. We also make the worst-case assumptions, ascribing bias-motivation to an arson whenever there is doubt.

We want to calculate, from available data, the hate-proclivity ratio, kB / kW , as it applies to church burning. For this case it is the ratio of proclivities of blacks and whites, respectively, to torch each other's churches. People are the perpetrators, but victims now are churches. We use asterisks to denote torched churches and write from (4),

where N*W and N*B are the numbers of white and black churches, respectively, burned between January 1995 and November 1996. Of the 298 incidents during this period, federal investigations found 43 percent involved black churches. About 2/3 of the arsons were determined to be bias motivated. If the same fraction applies to both black and white churches, its precise value is not important.

The states where most of the burnings occurred were Texas, Tennessee, South Carolina and Florida. Blacks make up about 20 percent of the population in these states. Putting all the numbers into (5), yields kB / kW = 5.3. That is, while whites were being blamed for burning black churches, and were drowning in their own guilt, the facts reveal a black was 5 times more likely than a white to commit bias-motivated church arson.

    Obsession with hate crime has spawned official subcategories of offenses: anti-white, anti-black, anti-American Indian/Alaskan Native, anti-Asian/Pacific Islander, anti-multi-racial group, anti-Hispanic, anti-other ethnicity/national origin, anti-Jewish, anti-Catholic, anti-Protestant, anti-Islamic (Moslem), anti-other religion, anti-multi-religious group, anti-atheism/agnosticism/etc., anti-male homosexual, anti-female homosexual, anti-homosexual, anti-heterosexual, anti-bisexual, and multiple bias. Each year when the FBI publishes its hate-crime report, it lists about 8,000 of these offenses. To put this in perspective, violent crime alone numbers about 8 million annually. The overriding reality is that hate crime is a tiny fraction of total crime. A 1960's observer propelled into the future would surely think we have gone berserk.

Each year, when the FBI releases its hate crime report, the press and other media take notice. Race bias is the favorite theme, though anti-homosexual acts are making a run for first place. Here are some typical headlines.

  • (AP) FBI: Most Hate Crimes Racially Motivated and Directed at Blacks
  • (CNN) Whites commit most hate crimes
  • (Washington Post) Reported hate crimes are on the rise in the United States, and African Americans continue to be victims of such acts more often than any other group, experts said today
  • (Los Angeles Times) Racial Bias Tied to Most Hate Crimes in 1998, FBI Says

Both FBI press releases and the media omit the singularly compelling fact that hate crime is a minute fraction of total crime. Selective exaggeration is not restricted to hate crime. An analogous pattern emerges in the reporting of AIDS. AIDS accounts for under 1 percent of deaths in the US, less than from the respiratory diseases, bronchitis, emphysema and asthma. When did you last see a bronchitis headline?

In its last complete National Criminal Victimization Survey (1994), the Justice Department revealed blacks to have committed 1,600,951 violent crimes against whites. Only 15 percent of these had robbery as a motive. We can safely infer that most of the rest had race as at least a partial motive. Eighty-five percent of the attacks were assaults and rapes. While blacks were committing these 1.6 million crimes against whites, whites were reciprocating with 165,345 violent offenses against blacks. Blacks, representing thirteen percent of the nation, committed more than 90 percent of the violent interracial crime. Fifty-seven percent of the violent crime committed by blacks had white victims. Less than 3 percent of violence committed by whites had black victims. In 1994, a black was 64 times more likely to attack a white than vice versa. This is the real story of hate in America. It is the media's well-kept secret.


HOME   complete contents of La Griffe du Lion  

Problems for the Heterodox   

Write to  
La Griffe du Lion