The Violence Conundrum

Violent crime, homicide in particular, has been declining for the past five years, both in large cities and across the country. Given that many of the underlying problems related to violence, such as a proliferation of highly efficient guns, have not gone away, this decline is nearly as surprising as it is welcome.

The national homicide rate has moved from a record high of 10.5 murders per 100,000 persons in 1993 to eight in 1995 (the most recent national data available from the FBI) and will probably be far lower for 1996 and 1997. In L.A. County, rates have been declining since 1992 and are now down to levels not seen since the mid-1970s. The exact timing and size of the downward shift differs from state to state and from city to city, making the phenomenon seem, on the one hand, to be local and, on the other, to be loosely national. But it can’t be both. It is hard to see why Missouri, for example, should follow the pattern of New York or California.

This good news has caused various experts and policymakers to offer widely different explanations. They haven’t yet debated the issue, perhaps because as long as rates decline, everyone is happy. But why do these experts disagree?

There are at least five reasonable explanations being offered for the decline. Each is, in many ways, intuitively satisfying. The explanations include recent changes in policing strategies emphasizing community policing; shifts in drug sales, especially crack, that make it less public and less likely to cause violence; subtle demographic decreases both in thenumbers of young men and in their propensity to commit crimes; rising employment, which takes people into less risky lives, and increased imprisonment resulting from mandatory sentencing, such as California’s “three strikes” law.

Yet, no one answer is definitive. There have been dramatic downturns before, ones that defy sensible explanation. In late-19th-century New York, for example, homicides declined at a time of corrupt and inefficient policing, terrible crowding and poverty, and when the likelihood of a murderer getting caught and going to prison was very low. At the time, few social observers realized their good fortune. Today, we look at Jacob Riis’ famous slum photographs of that era, unaware that he portrayed poverty and chaos from a peaceful world--the city’s homicide rates were less than four per 100,000 through the 1890s. The still-unexplained low violence rates of that bursting city should caution anyone offering comprehensive theories today.

At the same time, the young were not nearly as violent as today. It is not that there were not young murderers then, for there were: boys of 13 and younger stabbing or stoning each other to death. But the rates per age group were low.

During the 20th century, the national homicide rate, as well as that of big cities like New York and Los Angeles, fell from the early years of the Depression through the early 1950s, when they flattened, then began the long increase that we hope has finally ended.

Only a blurry message may be discerned in the big picture. After each major war--the Civil War, World War I and II--homicide rates declined. So much for the notion that war gets men’s blood boiling. Cities have grown steadily since the 19th century, while homicide rates have gone both up and down. So it is not pure urbanization at fault. Capital punishment steadily increased from the 19th century to the Depression, then dropped, only returning to popularity recently. These rates, too, marched to a different drummer, simply because relatively few murderers are executed.

This historical background leads one to be cautious about easy explanations. It also creates tolerance for our current ignorance. Understanding personal violence, its ebbs and flows, is a complex challenge. But it is an urgent challenge, and one that is particularly American, not only because individuals suffer but because it hurts our economy.

Politicians, including our president and many mayors, claim that community policing has led to the decline. The problem is, violent crimes have declined in cities without community policing. It is hard to argue that a change in policing in New York has reduced violence in Chicago.

New York had no uniformed police in the early 1840s; homicide rates decreased for two years after the city got its first police, then seesawed back to new highs. The same thing happened in the mid-1850s, when the police went through a dramatic breakup and recreation. Is it not possible that policing innovations came just as things had begun to change for other reasons?

Recent research by a group of social scientists associated with the National Consortium on Violence Research implies this might be the case. Working in several major cities, they discovered that the decrease in violence in these cities is for specific offenses, those related to crack. The drug now has a bad reputation; crack addicts are disdained by youth. At the same time, the crack markets have become less chaotic. Street-corner battles are no longer the cost of doing business. Just as with legal enterprises, the market shake-up has left a more orderly group of suppliers and buyers, who often work in less risky locations, using pagers and cell phones. Hence, the drop in drug-related crimes.

But there are other pretty good accounts for the decline in homicide, for example, a drop in the number of young males. Allan Abrahamse, at Rand Corp., has identified a group of young offenders in California who account for a high proportion of violent offenses. Small changes in the numbers and actions of these young men can dramatically affect homicide rates. Nationally, the proportion of young males aged 15-24 has declined slightly. However, the decline started at least two years before the drop in violence, and stopped two years ago. It is not purely demographics.

And, what about unemployment and high rates of imprisonment? For some reason, these two ideas have taken on ideological shading, as though they are mutually exclusive. Just getting 140 potential L.A. County shooters into prison or a job would reduce homicide rates by 10%. Has this happened? Maybe, but we can hardly find out if someone would have murdered if they were not working, or were out of prison. So we cannot really test these hypotheses.

We have difficulty explaining homicides because they are relatively rare and because we have too many good explanations and too few actual tests of them. Careful research on homicide and personal violence is done by a only small number of people across the United States. It is characteristic of a field about which we don’t know much to expect big solutions, big breakthroughs and for the experts to never say, “I don’t know.” For homicide, a national problem as well as a local problem, the first step toward knowledge is to allow experts to say how much they do not know, to allow them to posit partial explanations and then to be flexible in their policy applications.

Picking one explanation is not a good idea, yet. The systematic study of violence is young. Accepting that understanding may come slowly and with enormous effort is an important step toward knowledge. Meanwhile, expect a diverse range of accounts for the decline in violence, but ask for some quality research to back them up.

If we wish to harken back to the “good old days,” we must be careful in saying which ones. The mid-19th century would be a bad choice. We would be appalled at the level of violence then. The 1890s would be better, or maybe the early 1950s. Let us hope that the next decade will usher in continually falling rates, and that we will consider reducing violence as a serious public goal and not just at election time.*