Wow, you really have trouble with statistics. Are you REALLY trying to use those stats you cite to show whites are higher risk than blacks? Here, I can help.
OK, your fist error is that you are mistaking aggregate and per capita rates of crime. Whites outnumber blacks by 5.6 to 1 (2010 census)
http://www.census.gov/2010census/data So, if the per capita rate of crime by race was the same, which is to say the RISK of any individual man, on average, was the same for all races, then we would expect the White percentile to be 5.6 x the black one. BUT, the white total percentage of crimes is only 2.4 x the black rate... hmm, so by your statistic, the RISK of any individual black being a criminal is higher. hmmm
But, that's all crime. Let's look at violent crime. Here Whites average 54%, and Blacks 43%. How about murder and robbery, which have the most accurate reporting? Whites 44%, Blacks 54%.
But wait. That FBI table and the Census conflates Whites and Latinos together racially, but other data, such as imprisonment figures and school discipline figures, and victim surveys show Latinos commit crime at about 2.5x the non-latino White rate. So, about 42% of "white" crimes are committed by Latinos.
So, only 25.4% of violent crime(M&R) is committed by non-latino Whites, who are 56% of the population, and 54% of violent crime is committed by blacks who are only 12% of the population. In other words, Blacks commit murder and robbery combined at 9.96 x the White rate.
So, the risk of violent crime of any individual black is 9.96 x the risk of any individual white, on average. Blacks are per capita 8.7 times the risk of anyone of any other race.
Let me then rephrase my original more precisely. A lady, to lower her risk per customer of being Murdered or Robbed, can reduce that risk by 89.6% by excluding all black men (89.6% = per capita black rate / total per capita rate all races). If a woman saw men strictly at population demographic ratios, then she would only reduce her total risk of murder and robbery by 54.3% by excluding all black men, only 12% of her customer base. Her remaining non-black customers, 88% of her income, would present a total risk to her of only 45.7% of her risk if she saw all races proportionately.
Conclusion, screening by race is a rational and effective way to heavily reduce the risk of violent crime while modestly reducing the loss of income. Individual screening could be more effective, but racial screeing is still effective and rational.