The US is unusual in providing racial breakdowns of crime data, including the race of the perpetrators of violent crimes. The US Bureau of Justice Statistics are an excellent source of real data. They do surveys of victimization in the community (National Crime Victimization Survey) which are un-affected by any presumed Police biases, and include crimes reported and not reported to the Police. Uniform Crime Reporting records the actual crimes reported to the police. The findings of both are often interpreted in the light of US history, particularly since the beginning of African slavery in 1500, and US immigration history, with White Americans now being 60% of the population.
What about another country, with a different history? Specifically, Britain, which had virtually no African population until 1948, and in which (2011 estimate) White British are 87% of the population. These different histories should make a difference in crime figures. In Britain the racial aspects of the crime statistics are less clear. Some results have come out through Freedom of Information requests, but proper data of the US sort is less readily available.
So, it was interesting to hear what the Commissioner of the Metropolitan Police said in evidence to the Parliamentary Home Affairs Committee in July 2020. For background, the Metropolitan Police deal with London, and are the largest police force in Britain. London has the greatest immigrant population. Crime is dealt with by the Home Office, a government department. The most recent figures for London show 15,590 knife crimes in the past year.
Parliamentary enquiries are investigations similar to the US model, and combine close questioning with political point scoring. Now, a further complication: men with a knighthood are addressed as Sir, women as Dame.
Dame Cressida Dick fielded questions about how the police were dealing with the lockdown, and with knife crime and crime generally, with a focus on racial differences, particularly on the numbers of people stopped and searched.
If I go to violent crime, in London last year 72% of homicide victims under 25 were black. Nationally—you probably know the figures—you are four times more likely to be a victim of homicide if you are black and eight times more likely to be a perpetrator. The overlap with my key metric, which is knife injuries for under 25s, which we have been reducing for the last two years and into this year, shows enormous disproportionality in the way it affects our young black men as victims and, I am sorry to say, as perpetrators. That is horrible. For knife robbery, gangs, county lines, line holders: hugely disproportionate.
“County lines” and “line holders” refer to those running drug delivery networks.
The “8 times more likely to be a perpetrator” figure for Black British was not generally known, so it was a surprise to hear it in a public arena. The figure is roughly the same as that for homicide for African Americans, despite their different histories. US Police and UK Police differ considerably, in that the latter are not usually armed. That does not appear to make a difference in the ratio of black to white offenders. Black British were originally from the West Indies, and after 70 years in the UK, whatever their experiences, they were not subjected to possibly 500 years of life in the US. It is either a coincidence that both homicide rates are equally raised, or it suggests that the US experiences are not a unique cause of Black homicide rates.
Although the Commissioner did not give a reference, the homicide data is available, though in a format which does not make the ratio immediately obvious.
The relevant results are in Table 27, showing 800 convictions for homicide. (The last three rows of numbers are the population expectations, which I have added in).
These results are startling. Although Blacks are 3.3% of the population, they account for 17.6% of all murders, 5.34 times more than expected.
Although Whites are 86% of the population, they account for 69.8% of all murders, only 0.81 times those expected, so less than their population numbers. Putting the two together means that Blacks are 6.6 times more homicidal than Whites.
As a result, although there are 26 times more Whites than Blacks in England and Wales, Blacks murder more Whites (45), than Whites murder Blacks (11), a 4-fold difference. Proportional to their numbers in England and Wales, it is far more likely that a Black person murders a White person than the other way round.
The likelihood of a Black person murdering a White person is 45/(56,100,000*.033) = .0000243072
The likelihood of a White person murdering a Black person is 11/(56,100,000*.86) = .000000227998
Moving now from relative rates of offending to the disputed matter of “Stop and Search”, the next issue is how the UK Police operate this system.
Question 19 reveals that 35% of those handcuffed are black.
In reply to Question 21 on whether stop and search yields positive results (meaning that the Police find a weapon, drugs or stolen property), the Commissioner replies:
I have not had rigorous academic research on this, but I am utterly sure there is a link. I have said so on many occasions. There is a link between the presence of officers on the streets and the level of violent crime. There is a link between the use of stop and search and the level of violent crime. You can draw a graph in the last few years that pretty much shows exactly that.
Question 25, asked by the Chairwoman, is whether black men are being picked on disproportionately.
Chair: Yes. I am looking particularly at the Met figures for May alone for young black men in London. It looks like—and clarify if we have misread the figures—10,000 young black men aged 15 to 24 were stopped and searched in that one month of May and over 8,000 of them were not found to be carrying anything or doing anything that required any further action. That is 8,000 people in one month and there was no further action required. The young black population of London, from the figures I have, is only between 70,000 and 80,000 people. That suggests that in one month alone more than one in 10 young black men in London were stopped, searched and found to be carrying nothing and found not to be doing anything that required further action. That is just in one month and also at a time when most people would have been at home during lockdown. Does it alarm you that you have so many people from that cohort being stopped and searched and nothing further found, especially when the chances of them being in that group where there is no further action needed is five times higher than for young white men during the same period?
Dame Cressida Dick: I am not alarmed. I have said before that I am alert and I remain alert. First, the positive outcome rate is the same, whatever ethnic group you come from. The disproportionality in London has dropped a bit. It has become better or less bad in the last year or so. It is very much lower than many other forces, as you know. In London you are 3.8 times more likely to be stopped and searched if you are black, whereas in many places it would be seven or nine times more likely.
Q26 Chair: But your likelihood of being stopped and searched and having nothing found and no further action taken, that is the group I am talking about. I am not talking about the cases where you have found something and somebody was carrying a knife or there is some further action to be taken. I am talking about the sheer number of people being stopped and searched and nothing is found at a time when you had a very big increase in stops and searches—in May, when most people were at home. You had this big increase in stops and searches and also a drop in the proportion of positive outcomes. You were stopping more people but you, therefore, as a result, had a lot more people being searched without anything being found and a lot of people who may well be feeling, “Why am I being stopped and searched? Why am I going through a potentially humiliating experience?” and so on. Does it worry you that you have had such a big increase in the number of people stopped without finding anything and the community impact that that is likely to have?
Dame Cressida Dick: I am alert to that. Of course, I am concerned about that and I am talking to people about that. To be clear, it came down from 28% to about 20% slowly during the last many months. It is now heading slightly back up. I want the positive outcome rate to be high and higher, if you like. That is what I am talking to the teams about all the time.
This is a very interesting exchange. Assume no political posturing and that the questions reveal an honest misunderstanding about Police work. The Commissioner points out that the rate of stopping and searching Blacks is 3.8 times as high as for Whites, and that in other places it will be 7-9 times more frequent. Given that she has already said that Blacks are 8 times more likely to be perpetrators, it seems that other Police forces have got it right, and London is going off at half cock. A reasonable supposition is that London needs to double its searches of young Blacks. It would prevent injury to other citizens, particularly other Black citizens.
Here is an interesting technical question: what hit rate makes the search policy worth-while? The Chair of the Parliamentary Commission is concentrating on those who were stopped, searched and found not to be carrying anything, as if this invalidated the procedure, but no search technique can be based on absolute foresight, only reasonable suspicion. Searching people at airports comes to mind. The hit rate is very low, but the harm avoided is worthwhile.
According to the figures, 10,000 out of (a maximum of) 80,000 black men have been stopped and searched with a hit rate of 20-28%. The Government actuary values a UK life at £1,800,000 so in purely financial terms preventing even one murder seems worth it. Police wages must be paid anyway, so there is little marginal cost. Stop and search procedures have to be explained and documented for each person searched, so they will take up some time for police and person searched, which is a real but hidden cost for citizens.
A search does not have to be interpreted as a humiliation, but it certainly implies suspicion, and can be a cause of resentment if it happens very often. However, it also protects even those who are irritated by it.
One question to which we don’t have a precise answer, even though it is highly relevant, is what the chance is of finding weapons, drugs or stolen property on an average citizen, or even on an average young Londoner. We can estimate it thus:
The number of 15-24 year-olds in London is 1,033,030. The most recent figures for London show 15,590 knife crimes were carried out in the past year. Assuming that each knife crime was perpetrated by one 15-24 year-olds carrying a knife, then 1.5% of young men in London are carrying a knife at some time. If the Police stop young men at random, for every 10,000 stopped they will get a positive result in 150 cases. For every 10,000 Black Londoners, assuming an 8 times higher rate of offending, they would find 1,200 offenders.
In fact, they found 2,000 offenders. The Police did a good job. They chose well, probably because they not only noticed suspicious behaviours, but probably also because they recognized frequent offenders. The Police have noticed some things about criminals. It is hard to stop people noticing things, unless of course you punish them for it, or pay them not to notice.
I think we have proved the case once, but why not prove it again with another approach?
On average, by searching 5 young Blacks you prevent 1 offence, possibly a very serious one. In the process, you might irritate 4. Is it worth it? As you will recognize, we are looking at Numbers Needed to Treat, and Numbers Needed to Harm. No medicine or procedure cures every single patient, so NNT = 1 is usually a dream.
Stop and Search with a 20% success rate has an NNT of 5, which is very good.
If the hit rate was 28% the NNT would be 3.5 (usually rounded up to 4) which is better than 5.
The NNT for statins to prevent one heart attack is 104.
(Only worth it if you are at high risk, and if they cause harm at a very low rate, and certainly less than 1/104).
NNT for anti-hypertensives to prevent one heart attack is 100.
Flu shot, to prevent flu, healthy adults NNT of 12-37, for seniors, 40.
Early use of anti-biotics in open fractures, NNT 16.
Anti-epileptics for first unprovoked seizure, NNT 10.
Migraine relief in 2 hours with sumatriptan, NNT 2.6
So, sumatriptan is worth taking for headaches. If you do not have cardiac symptoms or predisposing factors, it is probably not worth taking statins prophylactically.
Traditionally, having calculated Numbers Needed to Treat, we should now calculate the Numbers Needed to Harm. Ideally, we want to treat hundreds of people without causing any harm, however slight. For vaccinations, we might want any harm (severe reaction) to happen to less than 1 in 750,000. For example, if Police officers sometimes stabbed the person they stop and search, then the NNH would be 100, or 1000 depending how often they stabbed. As it is, the harm they are accused of is taking up time, and by implication casting aspersions on those they search.
In summary, Black British rates of offending are very similar to African American rates, despite very different generational histories, and different Police procedures. Stop and Search is a useful procedure, and saves lives.
A final point about mutative effects: if the Police did far more stopping and searching of youths, then they would eventually reassure all youths that other youths were not carrying knives, which would lead to a drop in knife crime. There would be far less need to carry a knife in self-defense. At that stage the hit rate for stop and search would begin to plummet, and there would be protests that it was inconveniencing too many innocent people, and should be abandoned, which it probably would be, until crime rose again.