Is there a problem or not? The article does not say.

A frustrating piece of reporting in the New York Times yesterday, Murder Rates Rising Sharply in Many U.S. Cities by Monica Davey and Mitch Smith.

It is interesting as an exercise in trying to understand what is going on in the author's mind. Information is put in to the article that is unclear why it is relevant, information is left out that you would think to be critical, opinions are advanced without support and opinions are rejected without evidence. You reach the end of the article and you wonder, what does what I just read mean?

They open with dramatic evidence.

Milwaukee - Cities across the nation are seeing a startling rise in murders after years of declines, and few places have witnessed a shift as precipitous as this city. With the summer not yet over, 104 people have been killed this year — after 86 homicides in all of 2014.

More than 30 other cities have also reported increases in violence from a year ago. In New Orleans, 120 people had been killed by late August, compared with 98 during the same period a year earlier. In Baltimore, homicides had hit 215, up from 138 at the same point in 2014. In Washington, the toll was 105, compared with 73 people a year ago. And in St. Louis, 136 people had been killed this year, a 60 percent rise from the 85 murders the city had by the same time last year.

But is this just noise in the system? What is the overall trend in the murder rate for all US cities? If it is flat, then these surges are balanced by dramatic dips in other cities leading one to conclude that perhaps this is just routine noise in the system and that there likely have been similar rises and falls among cities in past years. The article does not say.

And how does the trend in murder rates in major cities relate to that of the country overall? If major cities are seeing a rise but overall the murder rate continues trending down, then there is a different set of issues to address than if the overall rate is rising. The article does not say.

Resorting to google, it appears that there was a recent release of information in mid-July, likely from the FBI, of crime rates in major cities. There are half a dozen reports in major papers, all following the same template as the NYT: Murder up in select cities, down in others, no context in terms of major city trend, and sotto voce indication that the overall national crime rate appears to continue to decline.

Since nobody seems to be willing to establish the context to understand the selected numbers being reported, one might naturally conclude that perhaps the FBI released a press release summary of the data which all the papers are using as their template for reporting rather than analyze the data themselves and come up with their own conclusions. This scenario makes sense as press release journalism is a common cost saving practice among media companies and because all crime enforcement agencies have an incentive to accentuate the negative in order to shore up their budgets.

Scanning several of these articles in addition to the NYT reporting, it appears, but I can’t be certain, that the national murder rate continues to drop but perhaps more slowly than in the past. It appears that the murder rate for all the major cities is either flat or showing a small rise.

Definitions: What constitutes a major city? The article indicates that there are 35 cities with increases in the murder rate but it does not define what constitutes a major city. All the cities in the article make the top fifty list by population except St. Louis. If the list includes St. Louis, then they are probably looking at all cities with populations greater than 200,000 or about 100 cities. If that train or reasoning is correct, then we can conclude that there have been sharp spikes in the murder rate but only for a third of major cities in the US.

Big increases among some major cities are being balanced by continued declines in other cities (such as Los Angeles, Phoenix, San Diego, and Indianapolis.) What sets the surge cities apart? What do they have in common with one another and different from all the rest?

It appears that what is being reported is not so much a changed trend in the murder rate at the national level or even among major cities. What is being reported is that there are sharp surges in select cities. So is the murder rate highly varied among cities from year to year? Again, the article does not say. My general impression is that there is normally a 10% noise factor in any given year around the long term trend lines and that what is being called out is that these few cities are having surges outside the normal noise level.

If that is true, then there are different issues at play than the headline and original reporting indicate. IF (a big If) all the above conclusions are ballpark correct, then these 35 cities are seeing an abnormal surge at the same time that the national rate continues to fall and the rate for all cities remains flat or up slightly. What do these 35 cities have in common with one another and that is different from all the other major cities that might explain these spikes. The NYT only names 10 cities with increases in the murder rate (in declining order they are Milwaukee, St. Louis, Baltimore, Washington, New Orleans, Chicago, Kansas City, Dallas, New York, and Philadelphia). The only thing that stands out at a high level are that they are all older cities (no west coast representatives or New South cities), they are old Democrat machine cities, and they all have high African-American populations. Indeed, USA Today indicates the surge in violence is concentrated African-American and Latino neighborhoods.

Other distinctive aspects of the surge cities include low clearance rates, randomness of the killings, concentration of killings among individuals known to one another, and the youth of the killers and victims. If I were to construct a narrative of what is going on, based on the fragmentary evidence in the NYT reporting and trying to link known recent events and changes to the surges in 2015, it might look something like this:

There is a surprisingly sharp increase in the murder rate among African-Americans in old Democrat dominated cities. The violence is being committed primarily by younger offenders against their peers. Those committing the crimes have already committed other violent crimes indicating an escalation in their behavior. The cities experiencing the surges in violent crime have responded to the cessation of Stop-and-Frisk in New York City, to the Ferguson shooting, and to the Black Lives Matter advocacy group by encouraging police officers to moderate their policing in African-American neighborhoods. In addition, efforts to reduce the population of the incarcerated has exacerbated the situation by increasing the percentage of the population at large who have violent histories.

Based on this interpretation and based on the data in the NYT which indicates that the 10 cities have experienced an excess of 322 deaths over the same time last year, then the logical conclusion is that moderating policing in African-American neighborhoods and deinstitutionalizing criminals has come at the expense of 322 primarily African-American lives. I don't think that is what the NYT wants us to believe but that is what their reporting is implying.

Despite the reasonable conclusion above, I have no confidence that it is reflective of reality. At the end of this puzzling and frustrating article, I have some tentative conclusions but the NYT has served its readers ill. They have created cognitive pollution without shedding light. This could be an important national issue warranting attention. Or this could be ephemeral local noise. The article does not say.

UPDATE: David French goes where the NYT fears to tread, tracing a straight arrow from reduced policing to increased crime, #BlackLivesMatter Costs Black Lives by David French.