Is the City of St. Louis a Safer Place in 2011 than in 1971? How Are We Supposed to Know?


The question seems simple: is the city of St. Louis becoming a safer place? Recent reports would have you believe it is. "Police and crime-fighting organizations around St. Louis are claiming a big victory," begins a story by KMOV. For 2011 the number of homicides in the city decreased by 22% from 144 in 2010 to 113. (no explanation has been provided for why the story stated a 30% drop. We're guessing a mathematical error.). This is a good thing. But take a step back, glance at the larger trends, ask a few questions, look for additional context, and the story becomes much less clear.

The number of homicides in St. Louis has been trending lower for decades. The urban crack cocaine epidemic of the early 1990's being the greatest outlier, there's significant consistency in the trend. Has the city of St. Louis gradually become better at fighting crime? Every drop in crime is met with claims of better, smarter crime-fighting. Again from the Post-Dispatch regarding 2011, "a new approach – using all facets of the justice system – is showing serious success." When crime goes up? From coverage of a spike in homicides in 2008, "Homicides do tend to go up in an economic downturn," stated a criminologist. One resident said he was told, "there was nothing (the district police commander) could do to protect us and the community … that he didn’t have the manpower." At the time, a North Side alderman implored, "the community has to be ready to defend itself."

Are we to believe that an increase in crime is the product of forces outside our control and that of the police? Are we to believe that a decrease in crime is the result of smarter crime-fighting? Using the homicide rate as a benchmark, St. Louis is basically the same city it was in 1971 when the city had 602,600 residents and 277 homicides. Would it be a mistake to conclude that the most significant factor in the number of homicides in St. Louis has been the number of residents?

Shouldn't the homicide rate, and that of other serious crimes, not just the raw numbers, have fallen farther? The crime rate for serious crimes has fallen dramatically across the country over the past two decades and is at its lowest rate since 1963. The murder rate has dropped by almost half, from 9.8/100,000 in 1991 to 5/100,000 in 2009. Over the same time period, the rate has dropped from 65 to 40.3 in St. Louis. Have other communities simply been quicker to use "all facets of the justice system"? 

Another question may be whether the City of St. Louis has the right number of police officers. A quick look reveals some variation in the number of police officers per resident across American cities. St. Louis employs 1,372 police officers, or 430/100K residents. Baltimore: 3,100 – 499/100K. Philadelphia: 435/100K. Boston: 333/100K. Memphis: 317/100K. Detroit: 388/100K. Charlotte: 230/100K (these numbers were updated 1/19). The more interesting question would be how the number of officers per resident has, or has not, changed over time.

While the long assumed direct correlation between poverty and crime has not been confirmed in the current economic downturn as many had anticipated (see 2008 comment above), evidence still suggests a positive relative relationship between economically depressed neighborhoods and high crime rates. Can the City of St. Louis, with a poverty rate of 27.8% be expected to have the same homicide rate while employing the same number of police officers as Charlotte, a city with a poverty rate of about 15%? What about Baltimore with a poverty rate of 21% and nearly twice the number of officers per resident?

For much of the past decade, it appeared that the homicide rate was in constant decline. Homicides were relatively low by historic standards and we were consistently told that the 2000 Census count of 348K residents was the low-water mark for the city. St. Louis was growing. You use decreasing homicides and increasing population and voila – a safer city. In reality, the city was still losing residents. A recalculation shows that 2008 and 2010 offered the highest homicide rates in St. Louis since the height of the crack epidemic in the early 1990s.

24/7 Wall St., a link-bait site we won't link to, recently ranked "The 10 Worst Run Cities in America". St. Louis appeared at no. 7, but was then described in a somewhat favorable light. "St. Louis has had a hard time controlling violent crime. With 17.47 incidents per 1,000 residents in 2010, the city has the second highest rate of violent crime in the country. This is due in part to the city's high poverty rate of 27.8 percent and its median income of $32,688, which is the 10th lowest out of the 100 largest cities. Additionally, nearly 20 percent of housing units in the city are vacant. All of these measures influence government revenues. Despite this, St. Louis has managed its finances fairly well." Perhaps St. Louis is doing ok given the challenges faced?

Of course there are other issues at play in St. Louis, where the city has not controlled its own police force since the Civil War. Besides being a bit of odd trivia, there are practical implications to crime-fighting. The Post-Dispatch recently found that St. Louis police officers fire their weapons up to three times more frequently per violent crime than officers in many other big-city police departments. Given the higher rate of violent crimes, that's a lot of shots – 117 incidents over the past five years, according to the P-D. An opaque, internal process has cleared 112 of those as being justified. If police officers firing their guns more often resulted in less violent crime, why hasn't there been more success in St. Louis? Is an officer's weapon considered a facet of the justice system?

We'll leave the unanswerable question of whether St. Louis is one of the worst run cities alone for now and return to the equally unanswerable "why has crime decreased in St. Louis and why less than in other cities"? Fewer homicides is good. If the current rate of decrease can be maintained, if St. Louis experiences fewer than 100 in 2012 and a new smarter, community supported, crime fighting culture can emerge, then we'll have a victory.

Looking at the national trend of decreasing violent crime rates, criminologists cite the following as primary factors: 

  • Increased incarceration, including longer sentences, that keeps more criminals off the streets.
  • Improved law enforcement strategies, including advances in computer analysis and innovative technology.
  • The waning of the crack cocaine epidemic that soared from 1984 to 1990, which made cocaine cheaply available in cities across the US.
  • The graying of America characterized by the fastest-growing segment of the US population – baby boomers – passing the age of 50.

It's tempting to list all the crime-fighting efforts employed when the violent crime rate drops. Police are targeting "hot spots", judges are taking a tough stance on crime, homeowners are spearheading local efforts. All are true and everyone engaged in combating crime, from police officers to someone dialing 911 should be encouraged and applauded. Nothing is so public and private as crime-fighting. Yet few cause-and-effect relationships are so cloudy as crime and crime-fighting. Are we to hope for better economic times? A gentrified city? Ever more effective crime-fighting? Police firing their weapons even more often?

Is the City of St. Louis doing an effective job reducing violent crime? When indexed to national trends: no. When we attempt to put it in context: who knows. But as an empirical question, "is the city a safer place than four decades ago?", it appears that little has changed.

STL homicide table

Pin It
  • Raydonn


    • matimal

      That includes wages, net workth skills, age distribution…everything. If St. Louis has no increase in population, but has increases in average wages or in the skills of its residents, and workers who do not live in St. Louis City, it will have advanced in important ways.

      • Alex Ihnen

        At some point, we have to believe that St. Louis will increase in population, that the current number is artificially low. I know popular places like Shaw are decreasing in density as smaller families move in and multi-family units are downsized, but there’s so much vacancy that if just about anything at all is built, more people will be added to the city. Of course there are big residential projects in the central corridor as well that will have an impact.

  • matimal

    Maps showing the changes in WHERE crime is happening might be more revealing. I’m told by my police detective brother-in-law that the police pay more attention to the geographic patterns of crime than they ever have before. St. Louis is divided in many ways, including in the concentration of crime.

  • Reverend Bacon

    I think one really needs to look at the black population over time, and the homicides per black. It’s essentially the black (and to a lesser extent, Hispanic) population of a city that tends to be the main driver of homicide rate. The non-black population contributes far less to the murder rate. Even Hispanics, with murder rates well above Whites who are well above Asians, commit fewer murders per capita.

    In 1970, blacks made up 41% of the population, so there were about 255,000 blacks and 309 homicides. In 2010, blacks made up 49% of the population; so there were about 155,000 black and 144 homicides. Based on that, I would conclude that the blacks in St. Louis have become, on average, far less violent over the past 40 years.The murder rate per black went from about 121 per 100,000 blacks to about 93 per 100,000 blacks.

    • STLEnginerd

      Put aside the inherent racism, that’s just bad analysis.

      You can’t take a general homicide rate, attribute it entirely to one ethnic group and then make conclusions based on that.

      If you want to talk about crime rates among blacks (a perfectly valid line of analysis), you have to start with statistics that can be linked directly to that community.

      Just because you can take one number and divide it by another doesn’t mean it means anything. Your numbers are only as good as your assumptions and yours are inherently flawed.

      • Reverend Bacon

        If you ran a linear regression with all the homicide data, where the city’s murder rate is the dependent variable, you’d find that the percent of blacks would have the highest coefficient, i.e., it would be the most important independent variable. Second would be the number of Hispanics. Third would be, I’m told, a dummy variable (0 or 1) about whether the mayor supports gun control. If you look at who’s doing the murdering across the country, blacks account for over half of them. That’s with an overall population share of about 12%; in cities, that’s generally far far higher. An exception is San Francisco, with only 8% blacks (who still commit most of the murders). White homicide rates are pretty constant, and about 6-10 times lower than blacks. In St. Louis, there may have been one or two arrests of whites for murder last year.

        St. Louis’ blacks have gotten less homicidal over the past 40 years, it seems. But that’s part of a national trend, too. They’ve gotten more homicidal over the past 10 or so years, which I believe bucks the national trend.

        I did some quick searching to get rough confirmation of these statements, and they look fairly accurate; one of the first pieces that showed up was from this site and this author- just put this /2013/01/understanding-st-louis-homicides-2005-2012/ after the nextstl dot com.

        • STLEnginerd

          You may well have a valid conclusion but that doesn’t mean you have a valid analysis. You put numbers to several factors but you said because blacks are well known to commit the majority of murders, let’s just assume in both cases they commit ALL murders. The majority could mean anything from 51% to 100%. Whose to say the percentage hasn’t changed in the last 40 years. That’s what I mean by bad analysis. If you have credible numbers for % homicides committed by blacks in 1970, and 2010 then you could salvage the credibility of your argument.

          Total # murders x “blacks are basically more violent that all other races” / # of black in STL= cringe worthy offensive stereotyping

          Total # murder x % murders committed by blacks / # of blacks in STL= blunt but potentially worth discussion

          Also national number might be worth SOMETHING but it is hardly considered accurate when applied locally especially when you use the same data to point out how ST. Louis is bucking national trends. As I said homicide rate in black communities are a well known and well documented problem. It’s not that the topic is untouchable. You should just avoid applying a false level of numerical accuracy to your conclusions.

          • Reverend Bacon

            Let’s see if I can post this, with the dots removed:

            wmbriggs (dot) com/post/7168/

            See figure 1. It’s showing what I’ve been saying, which is that the murder rate for whites barely ripples, while the murder rate for blacks fluctuates dramatically over time.

            So, for the purpose of most analyses, one can assume a constant (albeit non-zero) homicide rate for whites. So, yes, the percentage changes a bit over time, but Blacks are always (in the period studied) between 5 and 10 times more violent than whites. This graph shows that blacks in the US have lowered their propensity to murder. The previous link (by the same author) showed that blacks in St. Louis have been responsible for over 90% of the murders for the past 5 years. I think it showed that, of 100+ murders in the year prior to the article (2012??), 1 or 2 are alleged to have been committed by whites.

            I applaud you for actually putting your head into this, and not simply shouting “racist.” And you are spot-on about the level of numerical accuracy; I’ve tried not to state things as if there is any precision in this. The analysis is quantitative, but the conclusions must be qualitative due to all the noise in the data. Had we a national database of murders, with each datum containing all the known attributes of the murder, precision would be more feasible. It’s shameful that, in an era where such a once-daunting database could be transmitted in less than a second, the BJS is actually becoming less granular, rather than more. They haven’t published even the aggregated white/hispanic vs. black homicide data since 2008.

        • Alex Ihnen

          Here’s that direct link: We know the color of victim to a greater certainty than perpetrators, but we found of 567 homicides from 2008 to 2011, for which the race of the victim is available in the SLMPD annual reports, 502 (89%) are listed as black, while 64 were white.