The Decline and Fall of Violence

Our latest Freakonomics Radio segment on Marketplace concerns a topic we’ve been writing about for a long time: violent crime — and especially why it rises and falls. In this segment, Levitt and I discuss the fact that overall crime and violence are likely at a historic low these days, and not by a little bit either. The conversation builds off the fascinating new book by Steven Pinker called The Better Angels of Our Nature: Why Violence Has Declined.

Pinker has just completed a very good Q&A on our blog, and you’ll hear him in the Marketplace segment as well. Even though many people are convinced that the world today is more violent than ever (can you say “media effect”?), Pinker lays out the facts of the decline and fall of violence in a way that is hard to dispute:

In tribal societies, hunter-gatherers and hunter-horticulturalists, an average of about 15 percent of people met their ends through violence. In the 20th century, if you try to come up with the highest estimate you can, combining all the wars, all the genocides, all the man-made famines, you get to about 3 percent.

Here’s where you can find Marketplace on the radio near you.

Audio Transcript

Kai Ryssdal: Time now for a little Freakonomics Radio. It's that moment every couple of weeks where we talk to Stephen Dubner, the co-author of the books and the blog of the same name. This time, though, a special treat, the other co-author Steven Levitt. Gentlemen, welcome to the broadcast.

Stephen Dubner: Hey Kai

Steven Levitt: Great to be here.

Ryssdal: Dubner, you are not in New York. You have gone to the city of big shoulders, Chicago, with your comrade there.

Dubner: Yeah, I've come out to Chicago to try to persuade Levitt to run for president cause I thought that would be fun.

Ryssdal: We have a guy from Chicago already, you know?

Dubner: I know but we've talked on this show about how the president of the United States doesn't really matter that much. So, I figure that even if Levitt stinks it wouldn't be any great loss. And unlike most economists he'll admit that the president can't control or even predict the economy very well. But there are some useful things that he actually does know about. For instance, what causes crime to rise and fall?

Ryssdal: All right Levitt, do tell. What do you know about crime?

Levitt: Let me start by saying I absolutely, categorically am not running for president. If nominated, I will not run. If elected, I will not serve. When it comes to crime, one thing that I think I know that many people seem not to know is that crime has gone down, way down. And while many people are haunted by the thought that crime is high and rising, it categorically is not doing that.

Dubner: Contrary to popular belief the world is a more peaceful place today than literally at any time in history by a long, long shot.

Ryssdal: I'm going to need something better than "long, long shot." Quantify for me, Dubner.

Dubner: So the Harvard psychologist Steven Pinker has just published a new book called"Better Angels of Our Nature: Why Violence Has Declined." And he argues that our rosy view of history is pretty much entirely wrong.

Steven Pinker: In tribal societies, hunter-gatherers and hunter-horticulturalists, an average of about 15 percent of people met their ends through violence. In the 20th century, if you try to come up with a highest estimate you can -- combining all the wars, all the genocides, all the man-made famines -- you get to 3 percent.

Ryssdal: So that's 3 percent of non-natural. I mean people who met their deaths by other means.

Dubner: Yeah, violent. And he's including, there, man-made famines -- famines caused by civil war and what not. And keep in mind also in the 20th century we had mechanized weapons. So killing becomes easier. That early violence was carried out in much more rudimentary ways. Here's Pinker again talking about 16th and 17th century England.

Pinker: Samuel Pepys, in his diary, talks about going to the town square and watching a general being drawn and quartered. And in the next sentence he talks about going to a pub and getting some oysters. It was such a routine part of his day. In the era of Henry VIII I think 10 people were executed a week.

Ryssdal: Ouch. So Levitt, what's the trend though? I mean, in America in the 20th and 21st century, how are we doing?

Levitt: Well, we've been doing pretty well lately. There was a very inauspicious trend from the '60s to roughly '90 with crime going up and up and up. But, over the last two decades crime in the U.S. has fallen in half.

Ryssdal: How come?

Levitt: I think there are three factors that explain most of it. The first is we've dramatically increased the prison population. There are now 3 million people behind bars in the United States up from maybe 500,000. We've hired a lot of police as well. That's another factor as well that matters. The third factor people think a lot about that seems to be important is the rise and fall of the crack cocaine markets and the violence associated with that. That the rise in the late '80s led to a lot of increase in crime and then the decline with the falling apart of those markets. I have my own pet theory. Not that many people agree with me but I do think the data support it. The legalization of abortion in the 1970s led many of the people most at risk for crime never to have been born. And so when they would have been grown up 20 years later they just weren't here to do the crimes.

Ryssdal: All right, but Dubner make the connection for me between a lousy economy and crime. Cause you would think that it would go up.

Dubner: You would think. And even though the evidence argues against it over and over again. Smart people still think that it will. Crime actually fell during the Great Depression when the economy was in the toilet. Now, that was in part because prohibition had ended. Crime spiked during the 1960s when the economy was booming. And this most recently recession, despite a lot of predictions to the contrary, violent crime has continue to fall. Against, really, all the smart money bet against it.

Levitt: If there is one bet you know you will win. It is the bet that the next time that crime falls when the economy is bad, the headline of the New York Times article that talks about crime falling will say: 'In spite of everyone's prediction because of the poor economy that crime would rise, crime has fallen.' You can see the same sentence written 14 times by The New York Times.

Ryssdal: Levitt, let me this question though because if you go out there and ask people today -- ask 10 people on the corner of the street outside my studio here in Los Angeles today -- they will probably say that year over year crime is getting worse. And yet your data shows its not. What's that about?

Levitt: I think people in general are terrible at assessing risk -- that we very naturally over estimate dramatic things. So come to death, as an example, things like airplane crashes or terrorist attacks are very vivid and attract a lot of attention, whereas, we systematically understate the likelihood of dying from those kinds of deaths which are slow and silent and involve wasting away. Those don't make the headlines and so we don't really think of the risks associated with, say, eating way too many cheese burgers, the same way we do as being shot out of the sky by missiles hitting airplanes.

Ryssdal: is the website. Stephen Dubner and Steven Levitt, guys thanks a lot.

Dubner: Talk to you soon Kai.

Levitt: Thank you Kai.

Leave A Comment

Comments are moderated and generally will be posted if they are on-topic and not abusive.



  1. Ryan says:

    Steven Pinker is awesome. You guys really need to find some psychologists in a similar vein to post on your blog. Us social sciences have to stick together.

    Thumb up 4 Thumb down 0
  2. MW says:

    I haven’t had the opportunity to listen to your radio segment, so perhaps you discuss this.

    Crime is at an historic low. Imprisonment rates are at an historic high. Therefore the imprisonment-to-crime ratio is hugely greater than historical norms.

    One extreme interpretation: We’re obsessed by the idea that crime is a huge problem and imprisonment is the way to deal with it. We’re wasting fortunes and the primes of people’s lives on imprisoning people who could be productive members of society.

    The other extreme interpretation: Crime is low because imprisonment is high.

    Where does the truth lie? (No answers without statistical evidence, please.)

    Well-loved. Like or Dislike: Thumb up 14 Thumb down 1
  3. Michel says:

    This is surely an interesting topic. Is there somewhere a valid study about imprisonment rates in different countries compared to the crime rate? I don’t think imprisonment rates are at an historic high in all countries.

    A second thing I wonder about is if the definition of violent crime is broadened in more civilised societies?

    Well-loved. Like or Dislike: Thumb up 9 Thumb down 0
  4. Fitty Stim says:

    @MW: You’re only thinking in terms of the US – which has the world’s highest percentage of the population in prison and some of the harshest prison sentencing.

    Take a look at Scandinavia. The Vikings were some of the most violent “businessmen” the world has ever experienced. Now Scandinavia has very low violent crime rates – even though prison sentencing is a joke (especially if the violent perpetrator is a woman).

    I would suggest that any decrease in violence has everything to do with raising the standard of living for people.

    Well-loved. Like or Dislike: Thumb up 13 Thumb down 1
  5. A.R.Yngve says:

    It has got to be the standard of living. The more you’ve got to lose, the more careful you are.

    Or how about boredom (or the lack of it)? Having nothing to do is a great incentive to mischief. The more distractions a society has to offer (especially to restless, sullen young males), the less energy will be left over for violent behavior.

    Thumb up 3 Thumb down 0
  6. Ben says:

    Joshua Goldstein has written a related book, called “Winning the War on War” that’s just been published.

    Thumb up 0 Thumb down 0
  7. rhjeiqwjnpwankakn pipi says:


    Thumb up 0 Thumb down 3
  8. Nils says:

    I will allow myself to be extremely skeptical of the accuracy of violent death rates of hunter/gatherers being 15%. Where does that number come from? How reliable is it? How large is the variation?
    It would seem to me that any such number would be attached with enormous uncertainty.

    Thumb up 2 Thumb down 0
    • Rew says:

      The 15% number probably comes from (relatively, 150-200 years ago) recent observations of tribal hunter-gatherer societies – which could reasonably be extrapolated backwards.

      I would speculate a wide variation, of course. In resource-poor environments, tribal conflict and violence would be far more prevalent than in resource-rich environments, where there is little competition for food.

      Thumb up 0 Thumb down 0