The Great American Crime Drop and the Case for Modesty in Public Policy

Steve Novick

In the past twenty years, during which many bad things have happened in America, one spectacularly good thing has happened: Crime is down. Way, way down since 1990 (with the biggest drop coming between 1990 and 2000). I’ve been curious as to why, and have read assorted articles over the years offering one explanation or another. Recently (having time on my hands) I ordered a whole book on the subject: The Great American Crime Decline, by Franklin Zimring.

Zimring’s book is terrific: well-written, thorough and chock full of statistics. The book rather reminded me of baseball writer Bill James’ old “Baseball Abstracts” (though with less humor and ego.) It should be read by every policymaker, regardless of whether crime is part of their portfolio, for two reasons: (1) it is a model for the kind of research we should be doing in every policy area, and (2) it will serve as an ongoing reminder that modesty is often a virtue in public policy. A major lesson of the book (although the author doesn’t put it this way) is “don’t be too quick to think you have all the answers. Human behavior is complex, correlation doesn’t mean causation, and things aren’t always what they seem.”

Because the fact is that we don’t really know why crime is down. Zimring takes on each of the most popular explanations, and convincingly knocks them down. He doesn’t offer any pet theory of his own. Americans – and Canadians, as it turns out – are being a lot nicer to each other, for no identifiable reason.

Here are some of Zimring’s main points:

Crime really is down. Down all over the country, since 1990, by nearly 40% in most major categories, and much more than that in New York City. It’s not a matter of police departments cooking the books. Homicides and car theft have collapsed, and those books aren’t really cookable; it’s hard to hide bodies, and people report car theft to insurance companies.

It’s awfully hard to argue that increased incarceration is the key reason. Yes, we have been locking more people up for longer periods of time lately. But we started doing that in the ‘70’s and ‘80’s – and crime kept going up. New York, which had the biggest drop in crime, had a smaller increase in the incarceration rate than most places. And Canada, which did not have a prison boom, saw a drop in crime in most categories of at least 70% as big as that in the U.S. (starting from a lower base) during the same period.

“The economy” is not a satisfactory stand-alone explanation. Yes, there was a boom in the mid-to-late ‘90’s. But there was also a boom in the late ‘80’s – and crime kept going up. For that matter, the economy roared in the late ‘60’s – and crime kept going up.

It’s not all about the crack epidemic. Crime kept on dropping continuously after the intense crack epidemic had faded into history. Hospitalization for drug overdoses in New York was almost exactly the same in 2000 as it had been in 1990, but crime had collapsed.

Economist Steven Levitt’s eye-catching argument about legalized abortion doesn’t seem to hold water. Levitt created a buzz a few years ago arguing that the drop in crime followed the legalization of abortion; fewer unwanted babies, fewer juvenile lawbreakers 15 to 20 years later. Zimring points out that if Levitt’s argument were true, then the advent of the Pill years earlier should have prompted a big drop in crime – doesn’t any new form of birth control prevent ‘unwanted babies’? He points out that other major indicators of ‘unwantedness’ (babies born to unwed teenage mothers for example) did not change.

Changed policing practices can’t really explain it (with a caveat). New York, of course, pioneered community policing, the ‘broken windows’ theory, etc., and had a big fat drop in crime. But there was not a uniform change in policing methods everywhere in America, and there was a nationwide crime drop. And again, Canada, which had no major changes in policing practices and did not add police, saw a huge crime drop. Zimring also says that previous studies don’t suggest that adding more police in itself reduces crime.

Why the caveat? Well, New York had a ferocious focus in changing policing practices, and also increased the number of cops on the street by 35% - and it saw the biggest drop, by far. Zimring can’t find any other explanation for the New York difference, so he can’t dismiss the idea that in the specific case of New York, more cops doing different things made a difference. One interesting observation he offers is that since New York is so densely populated, each additional cop walking a beat ‘covers’ a hell of a lot more people than in most places; that might make a difference. (This part of the book reminds me of Bill James’ observation in the 1988 Abstract that while he usually doesn’t buy the argument that the running game has all sorts of hard-to-measure side benefits, in the specific case of the 1987 St. Louis Cardinals, one of the great base-stealing teams of all time, he thought there was something to it.)

Demographics can’t explain it all (with another caveat). The percentage of the population in the high-crime 15 to 25 year old demographic did decline. Zimring thinks this may be part of the answer. (Canada also saw a decline in that demographic.) But there wasn’t a 40% drop in the number of 15 – to 25-year olds; crime rates WITHIN that demographic fell.

Now, I don’t want you think that we should abandon any thought that public policy can reduce crime. I’ve been on a bit of a crime spree lately: I also bought a book called “When Brute Force Fails: How to Have Less Crime and Less Punishment,” which is full of opinions and suggestions, many of which seem sensible, though the evidence offered for the suggestions is often anecdotal. One of my favorites is his suggestion that high school classes should start later in the morning, and end later, so there are fewer kids on the streets between 3 and 6 tempted to break into people’s houses before they come home from work. The reason that’s my favorite is that it also tracks with research showing that teenagers, biologically, just can’t go to sleep early, so they go through high school classes exhausted, and don’t learn as much as they could; and research also shows that tired people eat a lot more than well-slept people. So – although I understand it might be complicated in terms of after-school activities, parents’ schedules, etc. – this is a potential win-win-win: make more teenagers lighter, lawful learners!

But Zimring’s book makes a good case for modesty in public policy. Sometimes, when you look carefully, things we thought made a lot of difference may not have been that big a deal. Sometimes the flashy new idea from a smart economist evaporates under scrutiny. And sometimes really good things happen for no identifiable reason.

I wish I knew exactly what we need to do to keep crime dropping. But I also kind of like the idea that starting around 1990, Americans, for no particular reason, just got nicer.

  • (Show?)

    Is there any evidence to suggest that crime rates are cyclical and we are just in a down period?

    • (Show?)

      Michael -- Zimring speculates there may be a cyclical aspect to it but this is an historically unusual drop and 'cyclical' sort of begs the question - what causes the cycle?

  • (Show?)

    This is fascinating. I've always assumed the decline in the crime rate in the 1990s was a combination of age demographics and the fact that there was a large public commitment to get tough on crime.

    It isn't so much that any particular measure (mandatory sentences, longer sentences, more cops on the street, etc.) is the answer as the fact that when the public at large focuses on something, it tends to make a difference. Judges sentence differently, law enforcement officers behave differently, parents pay more attention, neighbors are more attentive, etc. It all adds up and makes a difference.

    I sort of compare it to the various crazy diets people keep coming up with. It turns out just about anything works as long as you believe in it, focus on it and are determined to stick to it. (Or, at least, so I've been told.)

    After all, that seemed to work with the federal deficit in the 1990s, too. :-)

  • (Show?)

    Do you think Zimring's analysis and data suggest that we may be over investing in police and prisons: that we could have less police and less incarcerations and still have low crime rates? If we cut one or both of those, at what point might our crime rates go up?

    Or do we fund police and prisons for other reasons.

    Anyway, Steve, thanks for bring Zimring's book and its insights to our attention.

  • (Show?)

    This is a fascinating and vexing subject. It's bad when bad things happen and we don't know how public policy affected them; but it may be worse when good things happen and we don't understand them.

    What we're seeing must be a combination of factors--perhaps not all known to us. Levitt's theory was satisfying because it was a single-variable solution: abortion law. But almost nothing in the social sciences has a single-variable cause. I did research about the child welfare system for 14 years, and one thing we learned was that the effect of any single variable was almost certainly going to be weakly correlated with outcomes.

    It could be that the reasons won't ever be clear.

    • (Show?)

      Good point (re: multiple and varying factors). And one I concur with. I think the natural human reaction to want to distill systems into dichotomous ones is certainly a component of both the push for direct causation explanations, as well as the countering of 'discounting' said proposed single cause factors.

  • (Show?)

    I'd usually buy Jack's argument; here, though, Canada doesn't seem to have done much of anything about crime in the '90's but had a big crime drop anyway. David - the other book I referenced, "When Brute Force Fails," argues strongly that we can have less incarceration and less crime. On police, I don't think a comparable case has been made. Zimring cites contradictory evidence on the effect of increasing police numbers.

    • (Show?)

      Interesting write-up and discussion. However there are larger sociological events that differentiate the social climates in say the 1960s vs. 199s economic booms.

      The 60s had everything form race riots to the Kent State shootings over Vietnam, etc. combined with a vast and raging youth culture revolt. Not so in the 1990s. So I have a suspicion that discounting the economic factor by comparing 1990s to 1960s booms (and their divergent crime rates) is an apples to orange proposition.

  • (Show?)

    I share Jack and Jeff's view; it is the combination of efforts and demographic changes. Different approaches in different cities made an impact and they all might have impacts beyond their borders. Who know, parents might have had some impact on their children or the change in family size and focus on children's activities. All of these efforts were society's focus on a problem and each in their way helped.

    What really irks me is that while this was going on over the past 20 years, local news every where in the country has communicated to the public just the opposite. There are very few people who think that crime is down based upon what they see on TV every night.

    • (Show?)

      That (local news reporting) has far more to do with the news programing directives of the past 20 years. Which literally uses a formula on what types of stories are reported in the time-slots during local newscasts. This is why you get sex crime news reporting even if the local news channel has to go out of state to get it. To simply follow the programing models that are satirically shown to increase and retain viewership over the broadcast length.

      News program is almost entirely formula driven by what types of stores are shown, regardless of the how reflective it is of actual relevance to the geographical viewing audience.

      It is ratings formula, not actual reportage that drives local news programming.

    • (Show?)

      The news reporting comment here is spot on, IMO.

      Local news in the Portland market (which is really the only one I've noticed with any consistency in the last 10 years) is virtually unwatchable unless you enjoy a regurgitation of the most recent police blotter. It's a ticking off the list of that day's criminal activity..from petty thefts to car wrecks to the rare murder case.

      Perhaps Mitch is right that this is simply the way the ratings formula plays out. If this is the case, then I think we have a substantive argument for making news into a not-for-profit entity. The news is a service for the public good (or at least that's what my moral compass and reasoning tell me). If it's simply a ratings driven attempt to make money--then it's ability to provide that service is drastically curtailed or perhaps impossible.

      • (Show?)

        There is a body of writing on this. The piece entitled " Canned news: what does it mean when local TV news isn''t local?" by the Columbia Journalism Review is a good introduction to the problem, though not sure if it is still available in something other than in an online archive.

        I have had numerous conversations over the years with State Rep. Lew Frederick, who for quite awhile earlier in his career, was a reporter and he has some good insights into this and is the one who has made me first aware of this issue.

  • (Show?)

    "I wish I knew exactly what we need to do to keep crime dropping."

    We can learn from the western european societal models like Denmark and Sweden which have driven down crime and poverty with bold progressive public policy. They decided long ago to invest in people.

    Phil Zuckerman, author of "Society w/ out God" lived in Denmark for about a year writing his book. My recollection is that he was there for nearly a month before seeing a cop.

    • (Show?)

      JW's post seems to contradict Steve's point (2) "it will serve as an ongoing reminder that modesty is often a virtue in public policy."

      I happen to agree with Steve on this one as a kind of general principle and increasingly worry about the tendency to call for "bold" policy action.

      But that could be just my age showing - I seem to remember being like JW in the early days of JFK's administration.

      Steve, is there any evidence or even good speculation on this point?

  • (Show?)

    Rex -- What I don't know is whether we know what specific aspects of Danish society are correlated with a lower crime rate - and how much of it is public policy and how much is culture. Joshua?

    • (Show?)

      My basic premise is that if we want to make the United States a happier and healthier place to live, we should look at the advanced democracies which have created the happiest and healthiest societies. No nation is the same and no nation has the perfect formula, but the fact is many are producing better results than we are. One area being crime. The United States has one of the highest violent crime rates amongst advance democracies.

      One of the most important answers Kitz delivered in his debate the other night was regarding education. He emphasized the value of early childhood investment. He said that his plan starts at birth, instead of kindergarten. In nations like Denmark, mothers receive mandatory paid maternity leave, usually around a full year. In the U.S. mothers receive exactly zero mandatory paid maternity days. Mandatory maternity leave is progressive public policy…… is family-based public policy.

      Things like universal healthcare matter. When people aren’t worried about paying for their medication, they are less likely to commit crimes. Nations like Denmark have much stronger social safety net systems to support people. They have more mandatory vacation time and not surprisingly lower rates of depression. Happier people commit fewer crimes.

      Marijuana prohibition is completely ridiculous and makes America less safe handing over the market to Mexican drug cartels and throwing non-violent productive citizens into the “rehabilitation” system. Kitz answer to the legalization question may have been politically beneficial but was

      Of course nations like Denmark, Sweden, France etc. don’t spend billions on the military industrial complex that we do. They are clearly prioritizing their people.

      I could go o and on with more examples. The point is we have similar and generally very liberal/progressive societal models that are producing much better results than we are. I would like to give more extensive and specific commentary but I’m simply pressed for time.

      How much is public policy and how much is culture…….I don’t know. Those are kind of hard to separate.

      I recommend Zuckerman’s book, which I have read, but also recommend “Europe’s Promise” which I haven’t read. I have listened to a couple interviews w/ the author who spent ten years in Europe and seems to have a lot of helpful insight on this matter.

    • (Show?)

      We can’t leave religion out of this discussion. In Zuckerman’s book, “Society w/ out God” he demonstrates that the least religious democracies have created the happiest and healthiest societies. These nations concentrated in Western Europe dominate categories like low crime, low poverty, low infant mortality, healthcare, education, gender equity, social justice, etc. I don’t believe this is coincidental. I made a satirical attempt to discuss this issue in a previous BO post:

      There’s a price to pay when huge chunks of the population think the end of times is near. There’s a price to pay when tens of millions of our citizens are focused on stopping gay people getting married instead of eradicating poverty. There’s a price to pay for the mass acceptance and perpetuation of unjustified and unsupported faith-based ideas.

      Religion is often treated like biology instead of ideology in this country and across the globe. To criticize religion, is like criticizing race or sexual orientation for many. Criticism of religion is often met w/ accusations of bigotry. It’s nearly impossible for a “non-believer” who is open and honest about their crazy evidence-based beliefs, to get elected to public office in most places throughout the U.S. Atheists/agnostics etc. remain one of the most despised groups in America.

      Regardless of our religious beliefs, if we want to make our home a better place to live we must drastically change the anti-reason sentiment/attitude/environment in this country. We need to destroy the shield which protects faith-based ideology from criticism. We must make sure all ideas/belief systems are open to the highest levels of scrutiny, particularly those which form the world views for so many of us. One conversation at a time.

  • (Show?)

    I think "When Brute Force Fails" is a must-read for anyone seriously interested in public policy and public safety. The author, Mark Kleiman, is a professor of public policy at UCLA and worked in the DOJ for the Clinton administration. He's purposely written a very readable and accessible book but I think you'll find there's significant research and experience behind the anecdotes.

    He doesn't claim to have all the answers. On the contrary, he proposes building incrementally on solutions that have worked in the past in specific instances, continually improving as results are available.

    I think the most important thing about his book is the way he lays out the economics (both social and monetary) of how we deal with crime.

    • (Show?)

      Thanks, All

      Over the last 60 years, I have supported, worked and/or argued for the U.S. modeling itself after the "advanced democracies" on almost every one of the policy fronts Joshua lists - and I still support that. But as Steve said and Joshua repeated, the precise interaction between public policy and culture is not at all well understood.

      There is no doubt that the cultural context in the U.S. is dramatically different from those in the "advanced democracies."

      And while we may not understand the precise interaction, we do know that culture can impact both the ability to enact policies and their effectiveness once they are in place.

      So moving the culture with policy becomes a very dicey business at best. Culture can push back very hard and sometimes in ways which have significant social costs, and or may actually delay the arrival of a desired state

      Prohibition is probably the classic case in that regard, with the movement away from smoking providing another interesting and very different one.

      So on balance, I have come to more often like think like the author Doretta quotes -specific, strategic and incremental policy action coupled with persistence in regard to improvement is often best.

      Which is why Steve's point that modesty is often (but NOT always!) a virtue in public policy pursuits resonated with me as sound advice. I see It as appropriately cautionary, but certainly not a call to abandon activism in pursuit of progress.

  • (Show?)

    What were the trends in crime from, say, the 40s to the 90s, and how do they correlate to public policy. Surely there's something to be learned there.

    Perhaps Britain and France don't count as "advanced democracies" but France has levels of violent social unrest that would be shocking to Americans and Britain has seen an enormous rise in violent crime, I believe surpassing that of the United States from the 1960s to about 2000.

  • (Show?)

    Professor Zimring is a leading opponent of "over investing in police and prosecution" and while he's right that there is no one answer there is no way getting around that there is a distinct correlation between "highER rates of incarceration" and "all-time crime lows." Oregon still jails its citizens at a rate significantly less than most of the country. We are about #30, up from #42 when our crime state was really bad. Most felons don't go to prison (less than 30%) It would be really great to see Democrats embrace the constituency of victims, who tend to be - out of proportion to the general population, women, children, and people of color. There have been far too many :estimates" that balanced crime measures like Measure 57 were going to break the bank and cost $175 million for the 09-11 biennium when it ended up costing less than $10 million for the one year it was allowed to exist. Measure 57 is currently suspended because claims were made it would cost too much. The poster who brings up marijuana is chasing ghosts. There is NOBODY in Oregon's prisons for simple possession of pot and over 60% of all prison inmates are there for having committed violent crime. We pay judges among the lowest in Oregon and the Attorney General is among the worst paid in the entire country. Public safety is infrastructure as much as roads. If you look at what part of state government seems to be working, public safety would clearly come out first. So will the response be to de-invest in public safety?

  • (Show?)

    But Josh, how do you explain the fact that New York did not have as great an increase in incarceration as the rest of the nation but had a much greater drop in crime?

  • (Show?)

    Also I do not think it is fair to say Zimring opposes investments in police. He says the evidence is mixed but does not dispute that NYC's huge increase in police may have played a significant role there.

  • (Show?)

    Zimring opposes increasing funding for enforcement and prosecution. I'm sure he supports prevention programs but the bigger debate here is the one about whether we should dis-invest in the public safety infrastructure as some in Oregon have advocated, usually on the basis that we supposedly cannot afford it.

connect with blueoregon