So were Piers Morgan and Christiane Amanpour dishonest about crime rates last night?

CNN has made quite a habit of outrageous statements on gun issues.  You can see previous interviews here, here, herehere, and here.  From my appearance last night on CNN.

MORGAN: . . . How do you justify the claim more guns makes more safe people in America? I don't -- don't get it. 
LOTT: Every place that guns have been banned, murder rates have gone up. You cannot point to one place, whether it's Chicago or whether it's D.C. or whether it's been England of whether it's been Jamaica or Ireland. 
MORGAN: I'm sorry, but that's just a complete lie. It's a complete lie. The gun murder rate in Britain is 35 a year, average. You need to stop repeating a blatant lie, about what happens in other countries.  [cross talk]  No, you're not going to get away with this. You lied about it the other day. Thirty-five gun murders a year in Britain, eleven to twelve thousand in America. Stop lying, because what you say drives Americans to defend themselves. 
After Morgan then claimed multiple times that I had lied, the video then shows that I tried to explain that there is a difference between levels and changes.  In an obvious setup, Christiane Amanpour claimed that the murder rate in the UK had initially been flat after the ban and then fell.
Amanpour: After Dunblane, they put in these bans, they put in these punishments, fines, jail sentences, etc. and its true that straight afterwards there wasn't a huge change, but 2002/2003 until 2011 the rate plummeted by 44%.
Morgan and Amanpour were clearly taking about the number of homicides so initially here is a chart for that (source here see Table 1.01 and the column marked "Number of offences currently recorded as homicide").  She is right that there were substantial increases in law enforcement activity (for the original data see here and here), which one suspects should have been associated with reduced crime rates, but, even with that, how can she make the claims that she did about homicides?  (Note that it often takes a couple of years after a person becomes a police officer before they become very effective.)

Two clear points can be seen from the next two figures.  First, after the ban, clearly homicide rates bounce around over time, but there is not one single year during the 15 years after the ban where the number of homicides is lower than it was immediately prior to the ban in 1996.  By the way, the average yearly homicides from 1990 to 1996 was 601.  For the time period after the ban started it was 707, an 18 percent increase.  Second, the number of homicides remained higher than the immediate pre-ban rate despite a large increase in the number of police officers during 2003 and 2004.

If you look at the percent changes, the change from 1996 to 2003 was bigger than the drop since then.  If she says that there is no "huge change" between 1996 and 2003, how can she say that there is a "plummet" after that (when it fell by 32%, not the claimed 44%)?    

Note also that Morgan must have misspoken about the number of gun homicides a year.  Indeed, at least since 1990, the average has been twice that high and has never even got as low as the average of 35 a year that he claimed.  I think that total homicides are the most important concern, rather than how a homicide was committed, but if that is what some would rather focus on, it is still hard to see that even firearm homicides fell after the ban.  The averages in the pre- and post-ban periods are virtually identical (61 pre-ban and 62 post-ban), and there are only two years that the number of firearm homicides fell below what the number was in 1996 (2009 and 2010).

Obviously, guns are involved in more than just homicides or murders.  One thing that is clear has been the huge increase in gun crime generally in England and Wales since the gun ban (for similar discussions see here, here, and here).  Firearm Offenses involving handguns, rifles and shotguns were falling from 1991 to 1997.  At that point, they stopped falling and kept increasing until 2006.  The number of firearm offenses in 2011 was still 16 percent higher than in 1996 and the average for 1997 to 2011 was 8,326 or 31 percent higher than in the 1990 to 1996 period.  If Piers means to include Scotland in Britain, that would raise the number of gun homicides in 60 to 73.

As an aside, homicides in England and Wales are not counted the same as in other countries.  Their homicide numbers "exclude any cases which do not result in conviction, or where the person is not prosecuted on grounds of self defence or otherwise" (Report to Parliament).  If correct, the problem isn't just that it reduces the recorded homicide rate in England and Wales, but what would a similar reduction mean for the US.

More information on the adjustment for England and Wales is below for "Update 4."  If taken literally and as I discuss below, that isn't clear, a simple comparison can be made.  In 2012, the US murder rate was 4.7 per 100,000, a total of 14,827.  Arrests amounted to only 7,133.  Using only people who were arrested (not just convicted) would lower the US murder rate to 2.26 per 100,000.

Gun crimes have apparently also been seriously underreported in the UK.  From the UK Telegraph in 2008:
The internal memo, written by a senior officer, says there has been significant under reporting of serious crime and warns of "serious concerns" that confidence in the police and Government will be knocked when the true levels are revealed.
It was drawn up in response to a briefing paper given to the Metropolitan Police Authority outlining Home Office changes to the definition of crimes.
Under the changes, police have been told to classify all offences as gun or knife crime when there is a threat with a weapon. Previously, this did not happen if the weapons were hidden.
Similarly, more assaults are to be classified as grievous bodily harm rather than the less serious actual bodily harm when a victim is injured.
In the memo, Det Chief Superintendent Peter Barron said: "The potential increase could be a rise in recorded GBH of 58 per cent, a rise in gun crime of 20 per cent and a rise in knife crime of 15 per cent." . . .
The UK Telegraph had this story in 2010 of how this bias has increased over time in Nottinghamshire.
"Their crime figures are totally inaccurate. Unless a gun gets discharged, it often doesn't get reported. A lot of people are brandishing guns out there and they are not put in the figures because a gun isn't discharged.
"If the figures from hospitals, of people coming in with gunshot and knife wounds, were used, the figures would be a lot greater. The hospital figures should be taken far more seriously." . . .

I am more concerned about total murders than just firearm murders, but firearm murders have also risen after the 1997 handgun ban.  Indeed, there are only two years after the ban where the number of murders were below what it was before the ban.  In 2011, there were 60 murders, up from 49 in 1996.

The discussion on CNN was supposed to be a Townhall where people from different views were in the audience.  Instead the people that they brought in from Arizona and Wisconsin and other places were all on the same side. I asked the people in my section if anyone opposed increased gun control regulations and no one said that they did.  Several shouted that they wanted to ban all semi-automatic guns.

So here are previous figures that I put together. These figures are from the third edition of More Guns, Less Crime from the University of Chicago Press (2010). Click to make the figures larger. The numbers for the UK are available here in Table 1.01 (see column marked "Offences currently recorded as homicide per million population").  There is only one year (2010) where the homicide rate is lower than it was in 1996.

The most recent violent crime data for England and Wales is available here and for the US here.  One important thing to note first is that the rate that crimes are reported to police is much higher in the US than England and Wales, and that difference will make England and Wales look relatively better than they actually are.  But still if one uses violent crimes reported to police, US in 2011 there were 1,203,564 and in England and Wales 821,957.  The US doesn't differentiate Violent crime with and without injuries, but in England and Wales 368,647 violent crimes involve injuries and  453,310 do not.  Given that the US has about 314 million people and England and Wales 56 million people (a ratio of 5.6 to 1), 368,647*5.6 = 2,064,423, or still about 72 percent greater than the number for the US.  I have often pointed to the International Crime Victimization Survey as a better comparison because it deals with the different rates that crimes are reported and it tries to make sure that crimes are defined the same way across countries.  As expected, that measure makes England and Wales look even worse compared to the US.

Other information for Ireland and Jamaica.

How about for DC and Chicago (Figures taken from More Guns, Less Crime)?
The raw data for DC over a long period of time is available here (the crime rates are available on the bottom half of the screen).  My books The Bias Against Guns looks at crime data for other places including even a police state such as the former Soviet Union and other countries.

Now Australia didn't have a complete ban on guns, they didn't even ban all semi-automatic guns, but I have a discussion on the changes in their crime rates here (see also here).

Does it look like murder rates fell in any of these places after a ban was enacted?

Here is some other information that might be useful on the two places available here and here.

The International Crime Victimization survey also provides some interesting comparisons on overall violent crime rates across countries.  To roughly get the violent crime totals add robbery, sexual incidents, and assaults.

Further notes on England and Wales homicide numbers:

In the 2002/2003 reporting year, 172 homicides were attributed to the serial killer Harold Shipman. 

In 2005/2006 reporting, 52 homicides were from the 7 July London bombing. 
In 2000, 58 homicides were due to a group of chinese immigrants found suffocated in a lorrie crossing into Britain.  The person who perpetrated the crime was in England.

Two points: 1) Even if these numbers are subtracted from the previous figures, nothing substantial is changed.  None of the values previously showing a higher number of homicides or a higher homicide rate after the ban are changed.  2) There are also similar unusual events in years prior to the ban.

Finally, the definition of homicides have changed over time in England and Wales.  If case it wasn't obvious earlier, because of this change and in order to insure that the same definition is used over time, I used the measure of homicides provided here, see Table 1.01 and the column marked "Number of offences currently recorded as homicide"

It is hard to do a similar comparison for Japan because their gun laws have been essentially the same for hundreds of years.  Even prior to WWII only the Samurai could own guns for anything but hunting and hunting is still allowed now.

UPDATE:  Jason Riley has one of his typically interesting pieces in the Wall Street Journal.  After recounting the exchange that I noted at the beginning of this post, Mr. Riley writes:

. . . In his book, "A Brief History of Crime: The Decline of Order, Justice and Liberty in England," Peter Hitchens cites a 2001 study that found "the use of handguns in crime rose by 40 percent in the two years after such weapons were banned in the U.K." The study is not an outlier. Joyce Lee Malcolm, the author of "Guns and Violence: The English Experience," reports that "armed robberies in London rose from 4 in 1954, when there were no controls on shotguns and double the number of licensed pistol owners, to 1,400 in 1981 and to 1,600 in 1991." She adds: "In 1998, a year after a ban on virtually all handguns, gun crime was up another 10 percent." . . .
UPDATE: Discussion on CNN International from December 18, 2012 where there were similar attacks on the data that I discuss above.

UPDATE2: Here is an amusing clip of Piers Morgan preaching about the importance of civil discourse (thanks to Breitbart TV).

UPDATE 3: After Piers had on conspiracy theorist Alex Jones last night.  What is amazing is how Piers called up Politico reporter Dylan Byers to complain that Jones demonstrated "vitriol, hatred, and zealotry [that] is really quite scary."

UPDATE 4:  There is a much more serious problem in comparing homicides in the US and England and Wales.  Homicides in England and Wales are adjusted based on the outcomes of trials.  This adjustment reduced the number of homicides in 1997 by about 12 percent.  There is no similar adjustment for the US data.

To compare conviction rates in the US and England and Wales, the Bureau of Justice Statistics has some numbers here (other related information is available here).


  1. Thanks John for such exhaustive and thorough research. What most people in the media don't understand is that teachers don't necessarily have to be armed to prevent these awful massacres. If teachers with concealed carry permits were simply allowed to carry in schools, that may be enough for these crazy lunatics not to try to go in and shoot. Just knowing that a teacher MAY be armed could be enough. You've called that the "freeloader effect," of which my family and I are beneficiaries in our home.

  2. Your graph for the UK shows a time series that seems quite at odds with what I have seen elsewhere, including here:

    And also, why do you conveniently leave aside from your discussion the case of Australia (which was mentioned in the CNN segment you are referring to) and its shocking drop in gun homicide rates?

    1. He links to Australia above if you look and read again

  3. Piers Morgan is a liberal totalitarian who has no problem with armed security protecting his ass.....but protecting average Americans is "hogwash".

  4. John,

    What I find amazing is how number-blind people are and how unwilling they are to look more deeply.

    In New Orleans, the murder + non-negligent manslaughter rate was 72.8 per 100,000 in 2010. More than 70x some other cities of similar size. Why?

    In the US, the non-gun homicide rate is 8x that of the UK. Clearly, the US is a different place, and too many people are comparing apples with oranges, without any sophistication.

    What matters now is not where you are, but where you are going. In that regard, your statistics (on the impact, after the imposition of restrictive laws) look both compelling and very worrying.

    Why people assume this mass-media debate is neither dangerous nor requires skilled academic and intellectual intervention, I do not know. But with such an obviously wide range of influences, the important role of gun laws at this time needs skilled consideration, not mass hysteria.

    Ideally, the US will gradually become less murderous, but changes should not be experimental and need the utmost caution. The media and politicians will happily experiment with people's lives, through ignorance, self interest, or lack of intellect.

    I truly hope this debate runs its full course and more measured programmes will surface in the media to allow people to make their minds up in an informed manner.

    Rob UK.

    P.S. for anyone reading the emboldened text is by me, using HTML tags, not by anyone else.

  5. Dr. Lott: I just want to thank you so much for your tireless dedication to trying to speak the truth about these issues in the face of such relentless and uncivil opposition. Your research & books are a godsend to those of us arguing in our own sphere of influence.

  6. I just wanted to say I think it's brave of you to do interviews like this knowing that Morgan will refuse to let you state your case.

    I will share the graphs you have posted here with others. It's shameful of CNN to make overblown claims about your motives or about so-called "Assault" weapons in general.

    The fact that these commentators - who have little or no knowledge or expertise of these issues - continue to act as experts is saddening.

    But nothing is absolute. It may take years for the public to learn the distinctions between various firearms and the effects of allowing concealed carry, but I don't think all is lost.

  7. Hey, Nicolas Vincent, why don't you CHECK his sources? He has them all posted right there in front of your face. But, then, I guess blind ideology beats investigation, right?

  8. John, have you tried submitting these graphs or datasets ahead of appearances? Might help ease their skepticism a bit - why not bring Piers a copy of More Guns, Less Crime and challenge him to read it? Its not like your making this stuff up, so perhaps stuffing this info in front of their face is the best way of informing people like Piers (who seems to be incapable.of understanding the difference between an automatic and semi-automatic weapon).

  9. Nicolas, John's graphs appear to be referencing different data from the graphs at your link. Homicide rates and gun crimes vs. gun homicides. Also, yours says UK while John's say England and Wales. You can find the graphs for Australia here:

  10. My understanding is that the UK or at least England does not include a gun or violent crime in it's official statistics until it is "a closed" or "solved case" or perhaps "adjudicated". That approach keeps the stats down. Is this true

  11. The number of firearm offenses in 2011 was still 16 percent higher than in 1996...
    What I would dearly like to see is clarification of whether the "firearm offenses" statistic includes non-violent (administrative) violations of firearm possession laws.

    If such offenses are included in this statistic, it is very difficult to draw any meaningful conclusions about the effects of firearms laws on the incidence of violent crimes involving a firearm.

    Are such non-violent offenses included in this statistic?

    @Nicholas Vincent: Are you ignoring the steady increase in Australia's overall rate of violent crime since the ban was instituted as irrelevant to this discussion? Are assaults, rapes and murders OK as long as they aren't committed using a gun?

  12. I think people are missing an important point in this gun ban histeria. The gun ban people wanted to ban guns on school property and they got their wish. So what happened?

    There was a massacre in Colimbine High School. Then a massacre at Virginia Tech. Then a massacre at the theater in Colorado. Now the massacre at Sand Nook elementary school.

    All the massacre's have taken place at schools or where the gun free zone sign's have been posted. Common sense tells me the gun ban people are responsible for all of the massacre's, not the gun's used.

    I've began refering to the places where I see a free gun zone sign as Free Kill Zones, which seems to be completely true.

    I've been keeping tract of the areas around the country with conceled carry permit holders as opposed to the places which ban firearms. John, your research for your books, more guns, less crime, is very true. I hope more people will see the connection between the free kill zones and the location ofthe massacre's.

  13. Thanks Mr. Lott for all you do. Not sure how you put up with jerks like Piers Morgan.

  14. I refer to "gun-free zones" as victim-disarmament zones, and to "assault weapons" as homeland defense rifles.

  15. Hey everybody. I am watching CNN. Piers Morgan, who is obviously very moved by the Newton shooting, is leaving no chance to John Lott JR to even answer the questions he (Morgan) is asking. For the 20 seconds he was allowed to talk, Lott got a point.You can't point one case, as gruesome as it is,and use it to ban guns!

  16. John,

    I was so disgusted--but not surprised--at the way Piers Morgan treated you. You could see how thoroughly he had front-loaded his audience and guests to only spout the liberal position on gun control. Any time you were about to clarify a key point concerning the facts and the statistics, he would simply call you a liar, cut you off, change the subject, express his shock at your obvious stupidity, or conveniently declare it was time to move to another guest or a break.

    Thank you for being willing to put up with all this bunk to get the message of truth out there.

    By the way, I did catch your NPR interview within that same timeframe I believe. In that interview you had what I thought was a powerful but simple argument that you might try with the Piers Morgans in the future. You simply asked the question (I'm paraphrasing) "If you knew someone wanted to harm you, would you put a sign on your door saying THERE ARE NO GUNS IN THIS HOME? Probably not; you would never do that. Yet we do it every day all across the country in these supposedly gun-free zones."

    Thank you, John!

  17. Many economic issues are usually counterintuititive such as the minimum wage, and gun control. That is why it is always an uphill battle to try to explain them no matter how rationally you try to explain them. That "forum" with Morgan is another frustrating example. All you need to show is the "voice of reason" of a Deepak Chopra and call Mr. Lott a liar in a sea of antigunners and the debate is over. Damn be the statistics, the graphs, and the research.
    In the eyes of many watching, Mr Lott ends up as just another nut!

  18. Dear Nicolas Vincent:
    You will notice that the links that you posted don't have the crime data going back before the ban so they miss the initial increase in homicides. As to my source, I provide the links to the latest UK data sources. They have also gone back and fixed some earlier crime numbers and the links that you are pointing to don't seem to have realized that.

  19. I find it strange that despite having a PhD in Economics, a very statistics heavy degree, you made a very freshman analysis of the numbers without digging into the facts behind them.

    There are several problems drawing conclusions using your numbers.
    1) you use raw values and don't normalize to population. Population in England increased from 57.2 million (1990) to 61.9 million (2011) ( Normalization alone would decrease the trends you see in homicides and gun-related violence/homicides per capita.
    2) Even with the raw numbers, fitting a trend line to the gun homicides for England and Wales from 1990 to 2011 shows a negative trend. Not huge, but negative.
    3) In terms of England and Wales homicides and gun related violence, you ignore the fact that reporting changed significantly in 1998 and very significantly in 2002/2003 (see sharply increasing the counting records in the last 10 years. Your comparison of raw numbers across this divide is highly flawed. The 'Violence against the person' statistic increased by a 118% because of this rule change. There was another 23% increase from the National Crime Recording Standard (NCRS) 202 change resulting in a 23% increase. If there was an increase in actual counts, it was a few %, not the tens you show. If you normalized the % increase attributable to reporting changes, the number of homicides likely has been flat or decreasing since 1998.
    4) Your raw homicide numbers also ignore the fact that some homicides are reported at the time of their judicial decisions in England, not when they occurred. For example, in the 2002/2003 reporting year, the number of homicides included 172 homicides attributed to Harold Shipman, a serial killer whose murders extended over decades from before 2000. 2005/2006 reporting included 52 homicides from the 7 July London bombing. In 2000, 58 homicides were recorded that were attributed to chinese immigrants found suffocated in a lorrie crossing into Britain, but their deaths occured outside the country. Removing the Shipman anomaly alone would show the trend decreasing since roughly that year.

    I would appreciate your thoughts on my above comments.

  20. Dear Unknown:
    I also reported it as a rate in a figure later on. However, I wanted to report it both ways because Piers and Christiane were both talking in terms of raw values, not rates, and I didn't want people to think that I was hiding anything by only reporting rates.
    2) You would fit two separate trend lines from 1990 to 1996 and from 1996 to 2011. If so, I don't think that the second trend line would show any improvement from the law.
    3) If you noticed what numbers I used, I used the ones where they went back and redefined homicides. See my note: "Number of offences currently recorded as homicide"
    4) Removing either of those two events would not change the conclusion that if you look at homicide rates there was still only one year after 1996 where the rate was as low or lower than it was in 1996.
    Thanks very much for your note, but after correcting your mistakes, I don't see how anything that I put up was incorrect.

  21. Why is it that all your Home Office links are 404 Not Found? Perhaps this is the graphic you were looking for:

    I wonder where you got the little graphs you did, as they bear no relation to the statistics published.

  22. Dear Paul:
    I just went through and double checked my links to all the Home Office documents and all the links are working perfectly fine. So is the link to the UK Telegraph article. Here are the main links again.

  23. Dear John:

    In July of 2009 the Daily Mail published an article showing Britain was the most violent country in europe, more violence per capita than the U.S.. You let Morgan turn "murder rate' into "gun murder rate.' The bare murder rate, by guns or anything, is much different and much higher than he said to you on air. You should have said, "I'm not talking about the _gun_ murder rate, but murder period."

    Martin Rice

  24. How do you explain how DC has no crime increase for a LONG time after the ban, and why crime fell in the UK in 2006? And isn't choosing 2011 as the end, and not 2010, or any other date where crime fell cherry picking?

  25. Dear Alexander Adams:
    No increase in DC's murder rate after the ban? DC's rate had been falling prior to the ban. It stopped falling and increased from being about 13 percent above the average in the other cities to over 40 percent above the average. That is a very dramatic change in just one year. How much more of an increase do you need to have?

  26. Thanks for the response, Dr. Lott. I see the increase, but it seems as though after the ban there is a few year interval where crime flat-lined, and there was a substantial lag before crime increased. Am I reading he graph wrong?

    Also, for the UK, it seems that 1 year before the ban crime was increasing, and it fell a little after, then rose. Is that due to other factors? Or is it because it took a year before they began enforcing it?


    P.S. I think guns should not be banned, but I am undecided on the MGLC hypothesis, though your numbers on CCW are very interesting.

  27. Dr. Lott,

    I've read that England only reports a death as a homicide if someone is convicted of the offense. The supposed reasoning is that it would be unfair to any accused to prejudge a potentially criminal offense. Here's the link:

    Is this true?

    If so, wouldn't it serve to under report actual homicides?

    Thanks for your efforts to interject reason and fact into the debate.

    Steve Moss

  28. Just a quick note of "THANKS" for all your time and effort in presenting the facts.
    The push for ever-increasing gun control will be high on Obama's "to do list" as his last year in power is upon us.
    Education of the public is key as the battle over the second amendment intensifies.
    You are a valuable resource in this cause.

    Best wishes to you and yours...


  29. I would hope you are unaware rather than outright lying and manipulating the truth that when you show the change the screenshot of what is recorded as homicide is not accurate since homicide in the UK at current includes murder, killing with the intent to murder but where a partial defence applies (killing someone in self defence applies here), conduct that was grossly negligent given the risk of death that resulted in death or conduct that took the form of an unlawful act involving danger of some harm, resulting in death. Hence your evidence is inaccurate.

  30. Dear Loyal Stormtrooper:
    Of course homicides include murder. Who says that homicides don't include murder? The problem is that you really just want to look at murders, not justifiable homicides, which is the difference between the two numbers. Why would anyone want to include justifiable homicides?