Problems with Crime Statistics

About Crime
By John Roman
Special to washingtonpost.com's
Think Tank Town
Monday, July 10, 2006; 12:00 AM

New national FBI statistics show violent crime increased in 2005 -- the first jump in 14 years. For a public accustomed to the annual reassurance that the nation is becoming safer, this news is troubling.

Does it signal another epidemic of violence, rivaling the worst of the 1980s? Or is it just a blip in a long trend toward safer streets? It's a situation easy to politicize but hard to explain.

For those who study crime, the answer is simple: We don't know. Even though information costs are dropping quickly, criminal justice data are harder than ever to find. Key sources, such as those tracking drug use patterns or describing each criminal incident, have been discontinued or underutilized.

If this crime spike does herald a long-term trend, steps to strengthen data resources, anti-crime initiatives, and other tools need to be taken now, before negative momentum precipitates a crisis. It would help if we knew where to start: Drugs? Gangs? Demographics? Economics? Domestic violence? Again, policymakers and those who deal in policy information don't know.

This sorrowful and potentially dangerous lack of information aside, available data do allow some informed speculation. The statistics are straightforward: Murder and aggravated assault increased significantly between 2004 and 2005, particularly in middle- and large-size urban areas. Importantly, crime was down in the largest cities, the smallest cities, and rural areas. Property crimes also generally declined, as did rape.

The story of a suddenly more dangerous America is troubling. Many interpretations of the new statistics are suspect, however, and may be overshadowing more cogent explanations.

Criminology and law enforcement experts warn against reading too much into a single year's data. A one-year change may be nothing more than noise, background static. Data can swerve for any number of reasons, they argue, and it takes time to establish a trend. However, in the past thirty years, with the exception of 1987, every time the crime rate changed direction, a new trend followed.

Some believe that rates may have fallen as low as they can go. If so, the latest statistics simply represent a return to a higher, but more stable, level of crime. But crime rates were higher in 2005 than they were in 1970, and both those rates were much higher than rates in the 1950s and '60s. Nothing in the data suggests that crime rates couldn't go lower.

One plausible explanation is simply that 2005 was a very hot year. The "temperature aggression theory" posits that heat itself can spark violence. Last year ranked among the ten warmest years ever in 17 states, and among the top 15 in seven others. Nearly every Midwestern state made this list. However, in the South, where violent crime rates jumped, temperatures were moderate.
If, in the age of air conditioning, the weather thesis doesn't convince you, consider this: Each year since 1992, the percentage of all crime reported to the police has increased. For that reason alone, crime should be increasing annually. However, closer inspection of the 2005 statistics suggest that only about 25 percent of the crime increase can be explained by the rise in reporting.

Although all these factors could be at work, for my money, two competing explanations for the increase are much more compelling and troubling.

First, most experts agree that jumps in crime in the 1970s and 1980s were driven largely by increases in the drug trade, culminating in the crack epidemic. Two areas with large increases in violent crimes -- the Midwest and the South -- are where methamphetamine use is growing fast, and the drug has yet to penetrate markets in the populous northeast corridor, where violent crime increases were smallest last year.

Meanwhile, funding has increased for police and corrections, but not for community supervision. So while the nation's jails and prisons return four times as many ex-inmates to the streets than 25 years ago, there are only marginally more parole and probation officers to supervise them. Budgets for programs to ready inmates for life back home have also been cut over the past decade. As a result, ex-prisoners who received fewer services while incarcerated come home poorly supervised and less prepared to reintegrate -- not a recipe for crime-busting.

Nor can we keep locking up more and more people as we have for the past 25 years. Until 2005, crime had fallen by more than 20 percent over the past 15 years, while incarceration rates tripled. If last year's crime increase represents the start of a trend and more incarceration is the only available response, by 2014, one in ten Americans would be locked up every year.

Unfortunately, none of the theories described here can be easily tested until the federal government gets serious about understanding what causes crime in this country. And without that understanding, little can be done to stop crime.

John Roman is a senior research associate in the Justice Policy Center at the nonpartisan Urban Institute. The opinions are those of the author and do not necessarily reflect those of the Urban Institute, its trustees, or its sponsors.
© 2006 Washingtonpost.Newsweek Interactive