Criminal Justice
 

City’s murder rate drops: A mere blip or the start of a long-term trend?

The city's sky-high murder rate declined in 2013, but not for the first time.

Mark VanLandingham

The city's sky-high murder rate declined in 2013, but not for the first time in recent years.

OK. The numbers are in. Our murder rate for 2013 appears to have dropped by 22 percent since this time last year. Good news, indeed.

But does it mean the long reign of violence is finally loosening its grip on New Orleans?

Or is the plunge just another of the bounces and dips we’ve seen over the past decade? To determine which it is — random flux or a real drop — my doctoral student Mengxi Zhang and I pursued a number of statistical analyses, consulting with several highly-qualified statisticians along the way. Here’s how we framed our query:  Given the recent “bounciness” of our murder rate, what are the odds that last year’s dip was a result of random flux?

The short answer: The odds on it being mere flux are far from trivial.

The varying statistical approaches we tried yielded consistent results: Given the level, trend and variability of our murder rate over the recent past (we looked at different periods but focused on the last 10 years), the dip in 2013 was not significantly different from what has happened in other recent years.

That’s not to say 2013 was entirely unremarkable. “Not statistically significant” in this case just means that the jury is out. But while random flux cannot be eliminated as an explanation, I think that other factors may also be in play and that it’s just possible — but only possible — that we are on the cusp of a sea change in the level of carnage that this city has endured during the past decade.  I’ll explain why after reviewing where we are and how we got here.

With 155 murders — and assuming the same slow-but-steady population increase we’ve been experiencing in recent years — our 2013 murder rate was 41 per 100,000, down from 52 per 100,000 in 2012. Merely awful is a long way from good, but the abrupt change in trajectory — that whopping 22 percent plunge — offers a glimmer of hope about our ability to eventually turn the corner on what many have long-considered the most glaring failure of the city’s post-Katrina recovery.

U.S. cities have much higher murder rates than cities in other highly developed countries, and New Orleans’ rate is regularly the worst among our country’s worst — indeed, several times higher than U.S. cities of comparable size.

Why? Some social scientists point to proximate causes such as a poorly managed police force, a robust traffic in illegal drugs, and high unemployment among young black men — a group most likely to be both the victims and perpetrators of violent crime. Others look at factors further “upstream,” such as our historically underperforming schools, widespread poverty, sub-cultures that emphasize honor and revenge, institutionalized racism, and a lack of political will.

If 2013 turns out to be the turning point leading to a sustained decline in our city’s perennially high murder rate — and this is a question that can’t be effectively addressed until this time next year — what factor or set of factors might have led to it? The answers can only be speculative at this point, but are worth a try, even before definitive results are in. Because if we have begun to do something right, we should do more of it at once.

When I look at the murder rates by year in the graph, I see a gradual downward trend. So does a statistical program that we used to draw the line that best describes these data. The trend is consistent with what we know about outliers. It is the nature of statistical outliers — and the New Orleans murder rate is an extreme outlier — to fall back into line over time. Statisticians call it “regression towards the mean.” No action is necessary to achieve this; it just happens.

Moreover, as epidemiologists tell us, an affliction like our city’s murder rate should diminish over time as the individuals who are most integrally caught up in it — either as victims or perpetrators — die off or are imprisoned or otherwise sidelined by severe injury or advancing age. This “natural” epidemiological pressure helps explain the gradual decline we’ve seen over the past decade, but does not explain the much sharper decline in 2013. If we are on the cusp of a sea change in our murder rate, we need to look at possible explanatory factors which also have changed recently.

Might there be some interventions that are playing a significant role? It’s a tough question to answer because first, we don’t yet know whether a real shift has actually occurred; and second, as with most matters of broad public policy, calculations of cause and effect involve a complex weave of variables.

Sophisticated evaluation of interventions currently under way in New Orleans and elsewhere will improve our understanding of what works best, but at this point consensus remains a distant goal. In the meantime, we have to make assessments based on the information we have, imperfect as that information may be.

For example, what about improved care for the victims of gunshot wounds? Could that be a significant factor in  the declining murder rate? I think we can effectively rule it out, since non-fatal shootings are sharply down, too. Poverty, to look at another possible factor, has also abated post-Katrina, but the biggest shifts occurred almost 10 years ago. As for the seismic changes in our school system post-Katrina: They may soon pay dividends towards reducing our murder rate, but it still seems awfully early for them to be reshaping it. Similarly, if the city’s consent decree with the federal government to reform the police department is to have an impact on violent crime, those benefits likewise will occur down the road.

What’s new in the past couple of years is the city’s investment in a number of anti-violence programs that have proven effective elsewhere. They’re being implemented under the city’s umbrella program NOLA FOR LIFE: Group Violence Reduction Strategy (an intervention approach); CeaseFire (a prevention and intervention approach, variants of which have been implemented successfully in both Chicago and Boston); and the Multi-Agency Gang (MAG) unit (a law-enforcement approach).

Might these programs be having an impact? My intuition is that the right answer is “maybe.” There are several related reasons why. First, the dip in murder took place soon after these programs were implemented here in a serious way. Second, a large body of research indicates that focused attention on a problem often brings results, even if the intervention is imperfect. Third, something about our long run of highly elevated murder rates — powerful local sub-cultures of violence? a revolving-door prison system? lack of local political will? — has long constrained the natural regression to more normal numbers mentioned above. It seems unlikely to me that these processes would kick in all of a sudden without an additional spur in the city’s flank.

One thing we do know for certain is that other cities have been successful in bringing down very high rates of murder and other types of violent crime, and then maintaining these low rates. This brings us back to the issue of political will.

In the other cities that have cut violent crime, the declines have followed sustained interventions. Significantly, in some instances the declines have taken root in districts where the interventions have been implemented and not in districts where they have not.

While we’re nowhere close to being able to claim that our current interventions are working, we’re also not in a position to “wait and see” if the current level of effort will lead to a sustained decline. Our murder epidemic is much too dire for that luxury; plus, I sense that if this is a moment of opportunity, it will be a fleeting one. I propose that we follow our hunch and double down on what may be working. Expanded investment in the set of programs mentioned above might well lead to a second year of reduced murder rates.

Decisions about how that expanded investment might be allocated should be determined by an advisory panel of our nation’s best criminologists and program evaluation experts. Such a determination won’t be easy. A decision to go in one direction alienates advocates of other approaches. Trying a variety of approaches simultaneously dilutes resources and muddies the water when it comes to determining which are truly effective. And a focus on violent crime takes money and other resources away from other important initiatives.

But at this moment there is a lot of room for optimism about the potential impact of an expanded investment in interventions that have preceded the one-year dip in our city’s murder rate. To repeat, the evidence does not support a conclusion of cause and effect between the city’s efforts and the lower murder rate or even prove that the decline is anything other than a random dip. But, there are intriguing features of the 2013 decline that should make us hopeful this might be for real.

In summary, here is what I make of the 2013 decline in our city’s murder rate:

  • A 22 percent decline is an impressive and hopeful drop.
  • That the one-year decline was due to chance cannot be ruled out at this point. Also, one year does not make a trend.
  • But it’s possible that the decline was due to new-found political will and the implementation of promising programs that have worked elsewhere.
  • Speaking of political will, I’d like to hear the mayor and police chief say: “A continued decline in our city’s murder rate is our top priority for 2014.”
  • Let’s hold the applause until this time next year.  The jury is still out.

Mark VanLandingham, Ph.D., is the Thomas C. Keller Professor at Tulane University.  He is on leave this year at the Russell Sage Foundation in New York City, where he is working on a book about New Orleans’ Vietnamese community. Helpful statistical advice from Mr. Alejandro Falchettore, Dr. Tianhua Niu, Dr. Sudesh Srivastav, and Dr. Hongyun Fu is gratefully acknowledged. 

Help us report this story     Report an error    
The Lens' donors and partners may be mentioned or have a stake in the stories we cover.
  • scotchirish

    First, about the statistics, levels of levels of statistical significance are arbitrary. If this change is not significant at, say, a P level of 0.05, what actually is the probability that it occurred at random? 0.10? 0.20? Thanks.

    Second, one can almost sympathize with social and behavioral scientists and engineers when this change coincides occur with decreases in the numbers on the police force. This appears to LSU’s dramatically improved graduation rates over more than a decade as the budget is cut.

  • scotchirish

    … appears similar to…

  • scotchirish

    Similar “nice day, if it doesn’t rain” story at the _Advocate_:

    http://theadvocate.com/home/8152937-125/some-fear-homicide-drop-in

  • Ezra Boyd

    Looking just at the graph, the murder rate appears to be on a general downward trend of approximately 2% per year. Out of eleven years of data, the murder rate drops in all but four. This tend appears significant (both statistically and substantially) but unrelated to Landrieu.

    This trend precedes Landrieu’s initiatives and the graph provides little evidence that the trend can be attributed to his policies. From the graph, it appears that the drops observed during his term are just the continuation of pre-existing trends. While the author focuses on the 22% drop in 2013, the biggest percent decrease (approximately 35%) occurred in 2008 before Landrieu was mayor.

  • scotchirish

    I struggle with statistics, but “up” or “down” appears analogous to a coin flip. Seven heads (or seven tails) out of eleven tries will occur about 89% of the time, when I plug the numbers into

    http://www.vassarstats.net/textbook/ch5apx.html

    and four out of seven flips (post Katrina) about 77 percent.

    A followup study might be how many statisticians does it take to accept a null hypothesis (precedent would be Aggies and light bulbs).