May 8, 2009

Homicide rates over the past 800 years

I recently got this comically smug comment in response to my claim that most people who long to live in the golden past are gay arts majors who have no clue what it was like:

Straight, history major, know lots about the 19th Century, long to have lived when the ideal of monogamy ruled, and crime rates were orders of magnitude below what they are today.

To each his own.

Not to pick on this guy in particular, since flapdoodle like this is everywhere. Still, let's put this nonsense to rest. Here is a graph from this review that shows homicide rates (per 100,000 population) for England over the past 800 years:

The review has similar graphs for other European countries, but the overall picture is about the same, with the exception that violence started to decline much later in southern Italy compared to northern Europe. The scale of the homicide rates is logarithmic, so each line going across represents an order of magnitude. Also note that the graph doesn't show the steady decline from the early-mid 1990s up through 2008. As we see, in the 19th C homicide rates, which correlate with other violent crime, were not "orders of magnitude below what they are today" -- if anything, higher. So much for the guy who "knows lots about the 19th Century" -- fictional novels don't count, only history.

Where do we observe order-of-magnitude differences? Between roughly the Industrial-and-after era compared to the Elizabethan era, when a typical day of amusement would have included a stop to see bear-baiting in specially dedicated arenas, or paying to stand on the rooftops near the gallows to see criminals hanged in front of a mob.

The next order-of-magnitude difference takes us back to the late Middle Ages, about 1350. Ah, the good ol' days of repeated famines, Mongol raids, and the Black Death. The climate was warmer back then too -- and you know how rowdy people get when it gets hot out.

Ideals, shmideals. Get real.


  1. What about illegitimacy rates?

  2. "Straight, history major, know lots about the 19th Century, long to have lived when the ideal of monogamy ruled, and crime rates were orders of magnitude below what they are today."

    To each his own.
    Not to pick on this guy in particular, since flapdoodle like this is everywhere. Still, let's put this nonsense to rest. Here is a graph from this review that shows homicide rates (per 100,000 population) for England over the past 800 years
    I don't think that anyone disagrees that murder had declined over the years. Crime seems to have increased quite a bit, although how much of that is due to the increased body of law and increased vigilance in enforcement is not a question I can answer.

    Check out page 14

  3. Thank you for another honest and data based post.

    I remember elementary school, when the question would be posed about would we wish to live in some bygone era. First thing I thought, hmm, no toilets, no band aids and even royalty had lice. Can I choose the future instead?

  4. Law Enforcement is so much better now because of scientific tools and the evolution of techniques to ascertain guilt and prove it in court.

    Getting away with a murder of someone that you have motive to kill ain't easy. Once the detectives are on to you these days, the forensics, the street cameras that can put your car in the area, the techniques in questioning that make you think your alibi doesn't hold up, dogs sniffing your stuff for blood or DNA not at the crime would be just too hard to kill someone you were known to hate or were seen by others in an altercation with.

    In 1620 though, this wouldn't have been so. One could have gotten away with it back then. I for one, sure am glad I lived in the late 20th to 21st century. Above 1920 would be as far back as Id ever have wanted to have lived. Post WW2 was probably pretty nice. Im way too spoiled to "rough it" now. I mean can you -imagine- how tough life would have been to have been a warrior with Ghengis Khan's horde, or Charlemagne or Alexander? The filth of the middle ages in soggy Europe? Ick.

  5. Smug Flapdoodle5/9/09, 4:04 AM

    I was comparing crime in general in the Victorian Era to crime in general now, not homicide many centuries ago to homicide now. The 1900-1997 increase in indictable offenses was 3712%. Less than a hundredfold increase, so I guess I could be said to have exaggerated with "orders of magnitude". Mea culpa.

    I may be biased. One category of crime that has never affected my life is murder. Burglary, rape, assault & battery have all made their mark. I do not consider myself unusual; I know no one who knows anyone who has been murdered.

    But the study I linked indicates homicide was about 40% higher in the 1990s than it was in 1900 (p. 14).

    I have never read a novel from the era, though I have seen some non-crime-related movie adaptations like The Four Feathers (worth a look for the wonderful costumes and actors, not the plot or realism).

    (BTW, looking at the comments it looks like Johnny Abacus has the same source as me. To give credit where credit is due, I found the source in a Mencius Moldbug post from the summer of 2007.)

  6. Indictable offenses are a poor long-term indicator of crime; aside from better enforcement, under democracy laws multiply like rabbits, as I documented a while ago:

    Michael Mann would approve of the graph.

    How many of today's indictable offenses were perfectly legal back then? I'll throw out drunk driving as an example.

  7. "Indictable offenses" are the more serious crimes, so there may have been a proliferation of these since 1900 as Martin suggests.

    However, the bigger thing is that that graph looks nothing like any other crime graph, so we know right away that it can't show what you think it does.

    It says in the caption that the crime must be reported to the police, and then recorded by the police. Otherwise it doesn't show up. A change in either or both of reporting and recording will produce a change when no change in the crime rate has happened.

    The caption says a greater under-reporting in the past alone accounts for "some" of the increase -- probably most of it. It's become trivial to report purse-snatching to the police, and for them to note that in their computer. Murdered bodies don't lie and are hard to miss, though -- one reason why homicide is preferred for studying overall crime.

    I'll probably put up a post at GNXP just on this topic, since it warrants more discussion than a comments section will allow.

    To see for yourself, though, you can easily get spreadsheets (and so, graphs) from the Bureau of Justice Statistics for all sorts of violent and property crime rates, from 1960 to 2007.

    They all look the same, suggesting an underlying "level of criminality," which tells us that the indictable crimes graph must be missing a lot since it doesn't record the stark fall-rise pattern in the homicide rate of the first half of the 20th C. Crime rates are correlated.

    (Like the US, homicide in the UK rose steadily from roughly 1905 to the mid-1930s, then fell until the late '50s, then rose until the mid-1990s, and has fallen since.)

    And even when they're rising or falling, there's visible year-to-year wobbling around the main trend, even in the composite measures which should smooth some of that out.

    There's almost none visible in that indictable crimes graph.

  8. i always figured, a time when getting caught doing anything was difficult meant that a lot more crime was probably committed.

  9. Smug Flapdoodle5/9/09, 4:18 PM

    Okay, so the recorded thirty-seven hundred percent increase per capita in indictable crime, per the national legislature of the United Kingdom, doesn't mean much to you. On that we differ.

    What I want to know is, do I still come across as a novel fan? Faggot wasn't too bad, but man, "fictional novels"? That's just cold.

    S. Flapdoodle

  10. No, we differ on how to arrive at conclusions -- you take an obviously misleading graph at face value, which is dumb for three reasons:

    1) Change in number of indictable offenses

    2) Change in reporting patterns

    3) Change in recording patterns

    What honest people do is examine all the evidence and consider the weight of it. If one graph pops out as completely anomalous, there's clearly something fishy about it -- not the entirety of the data.

    I don't expect the innumerate to understand any of this. Just cherry-pick a single graph, and pretend it's the whole truth.

  11. PS, the faux-flip use of "toodles" confirms you're gay, one of the hyperactive kinds.

  12. "Murdered bodies don't lie and are hard to miss, though -- one reason why homicide is preferred for studying overall crime."

    Murdered bodies also don't wind up as murdered bodies if medical science advances sufficiently to save lives that would have been lost in the past.

    There are confounding variables on both sides. Possible under-reporting of crime in the past versus easier death from violence in the past. People also adapt to high crime and disorder. There are no-go zones in cities today. Don't go there and you're at much lower risk. Don't go out after dark and you're at much lower risk. No countermeasures show up in the costs of crime but they are real.

    Ultimately if you want to measure the effect of crime you can't go with objective measures such as the crime rate since the data composition is so inconsistent over the period of a century but subjective measures. What did newspaper reports look like in 1900? Did a purse-snatching make the paper (i.e., was it unusual enough to be considered news)? Did people fear to travel in certain areas? There's evidence for these propositions but it's not quite as neat as stuff you can graph. In Harlem there are beautiful buildings with intricate carved moldings that are boarded up and graffitied. Drive along the Grand Concourse in the Bronx and you can see the same thing.

    -Steve Johnson

  13. I would hypothesise that crimes in previous time were far more likely to go unreported or undiscovered by authorities than today, so I wonder how accurate those numbers are.

    Also, when people think of the past and claim to want to live there they are romanticising it. The person in this case was probably imagining life in the upper class, where everything was comfortable and people waited on you hand and foot. As opposed to the unwashed, short, hard lives lived by the majority of people in that time.

    Secondly, they probably never stop to consider what it would mean to consider one bath a week a luxury, where most people didn't have a change of clothes for everyday of the week, where lice was rampant, infant mortality high, the common cold could kill you, etc.

  14. Agnostic and Martin-

    If anything, crime was probably over reported back in 1900. Here is an excerpt from Christopher Hitchen's Brief History of Crime:

    -start quote-
    So, before showing that something enormous and damaging has happened to English society since the Second World War, it is worth quoting an astonishing footnote from Jose Harris's social history of Britain before the First World War, Private Lives, Public Spirit: Britain 1870-1914: "A very high proportion of Edwardian convicts were in prison for offences that would have been much more lightly treated or wholly disregarded by law enforcers in the late twentieth century. In 1912-13, for example, one quarter of males aged 16 to 21 who were imprisoned in the metropolitan area of London were serving seven-day sentences for offences which included drunkenness, 'playing games in the street,' riding a bicycle without lights, gaming, obscene language and sleeping rough. If late twentieth-century standards of policing and sentencing had been applied in Edwardian Britain, the prisons would have been virtually empty; conversely, if Edwardian standards were applied in the 1990s then most of the youth of Britain would be in gaol."

    As it happens, 1913 was a fairly bad year by the standards of the time, with 98,000 serious offences recorded. This level would not be surpassed again until 1920 when the total rose to 101,000 after a wartime truce during which annual crime tallies sank to as low as 78,000 in 1915. Even convicts were reported to be showing patriotic zeal as they broke their rocks. Measure this against the figure of 2,521,000 recorded in 1980 and, even when you grant that the population had risen from 36 million to 49 million, the figures could be from different planets as well as from different eras."
    -end quote-

    I'd also note that when conservatives complain about crime, they are mainly focused on the complete break down of the social order in the big cities. Cities like Baltimore, Philadelphia, and Washington DC all had an order of magnitude increase in homicide rate over the course of the last century. If hicks in small towns are battering each other over the head with beer bottles in bar fights, it does not matter so much for the future of Western civilization. But it's very disturbing when large swaths of once great major cities become nearly unlivable by civilized people due to high crime rates.

    The fact that Baltimore has levels of crime unseen since the 13th century should be pretty alarming to anyone.

  15. Agnostic, the study you cite confirms a sharp rise in crime in the 20th century in all of those countries. Notice the wiggle and upturn at the end of the plot. It is compatible with the Met figure, not contradictory.

  16. There's no "sharp upturn during the 20th C" -- there is an oscillation around the sharp downward trend.

  17. the common cold could kill youIs that so.

  18. Gay? of the hyperactive kinds?

    What are you, 14? What's up with your attitude toward homosexuals?

  19. Agree with Finbarr re crime since Victorian times.

    One of the favourite events of the 'moral panic' school of criminologists in the UK is the garrotting panic of 1862, which shows that unreasoning fears of non-existent crime waves have always been with us ;-)

    Alas, when you look at the figures, you see this :

    "Magistrates also tended to redefine minor crimes – such as pickpocketing – as garottings, and to send them on to the major courts. The result was a rise in reported violent street robberies, which in turn fuelled further panic. According to the metropolitan returns, the figure rose threefold from an average of 32.5 robberies with violence in 1860-61 to 97 in 1862. The increase came only after Pilkington was attacked on 17 July."

    The current Metropolitan returns for Robbery, Person are running around 2,500 - per month !

    (and re murder - this piece posits that murder rates would be five times higher in the absence of current medical technology)


You MUST enter a nickname with the "Name/URL" option if you're not signed in. We can't follow who is saying what if everyone is "Anonymous."