A new analysis of 25 years of crime statistics
By Dan Schroeder
Ogden’s crime statistics have been in the news again lately, mentioned in connection with the departures of Mayor Godfrey and Chief Greiner.
When Greiner was dismissed, the city issued a press release giving him credit for reducing Ogden’s crime rate by 33% between 1999 and 2009, “almost 50% more than the national average for the same period.” The Standard-Examiner had already printed the 33% statistic on November 27 in a list of Godfrey’s accomplishments, and printed it again on December 29 in an article about Greiner. The Salt Lake Tribune similarly reported that there was a 33% drop in Ogden’s crime rate between 1997 and 2009.
It’s been more than three years since this blog took a thorough look at Ogden’s crime statistics, so an update is in order. How has Ogden’s crime rate changed during the tenure of Godfrey and Greiner, and over the longer term? Is the 33% statistic correct?
These questions might seem relevant only to historians who want to assess the Godfrey-Greiner legacy. But the answers could also be of interest to potential clients of Godfrey’s new consulting firm, who have a legitimate interest in knowing his actual crime-fighting record. It’s also important that we not hold the new mayor, and the new police chief, to an unrealistic standard.
Fortunately, there’s quite a bit more data available now than there was three years ago. Of course, we now have data for three additional years: 2008, 2009, and 2010. (Data for 2011 won’t be available for several more months.) In addition, the FBI recently put some older data on its web site, available through its new UCR Data Tool. Finally, with the 2010 Census behind us, we now have a more reliable figure for Ogden’s population (which comes into calculating the crime rate, e.g., the number of crimes per 1000 residents).
Before getting into the numbers, I should explain that when we read about crime statistics in the U.S., those statistics almost always come from the FBI’s Uniform Crime Reporting (UCR) program, which dates back to 1929. The advantage of the UCR system is that it is well defined and widely used. But unfortunately, it counts only certain types of crimes: so-called “violent crimes” (murder, nonnegligent manslaughter, forcible rape, robbery, and aggravated assault), and “property crimes” (burglary, larceny, and motor vehicle theft). The UCR statistics omit all other crimes including simple assault, fraud, drug crimes, traffic crimes, and so on. (Arson is sometimes included as a property crime in UCR statistics, but is not included in the data used for this article. This omission has very little effect on the overall numbers and trends.)
The UCR data can also be inaccurate or misleading in other ways. Local law enforcement agencies gather and report the data, and these agencies may not always interpret the FBI guidelines consistently. Some agencies might even misreport the data for political reasons. Even if the agencies’ reporting practices are perfect, they cannot report crimes that citizens never report to the police. And of course, the data tell us nothing about the causes of crime and of crime trends, which are extremely complex.
With these limitations in mind, here is a graph of Ogden’s UCR crime rate that goes back to the earliest available online data (1985). The rates for Utah and for the entire U.S. are shown for comparison.
The most striking feature of the graph is Ogden’s steadily decreasing crime rate over the entire time period shown. The occasional increases in the crime rate appear to be mere short-term fluctuations, superimposed on the strong decreasing trend. The Utah and U.S. crime rates have also been decreasing steadily, though only since about 1995 and 1991, respectively.
But the decrease has not been as steep as Mayor Godfrey and the newspapers have implied. If we average over the short-term fluctuations, the decrease from 1999 to 2009 was only about 24%, not 33%. Similarly, the decrease from 1999 to 2006 was only 17%, not 23% as Godfrey claimed during his 2007 reelection campaign. You can, of course, obtain either higher or lower numbers by cherry-picking the starting and ending dates to take advantage of short-term fluctuations. Perhaps coincidentally, if you look at the long-term trend since Greiner became police chief in 1995, the decrease is almost exactly 33%.
Furthermore, the decrease since 1995 has been entirely consistent with state and national trends. For example, between 1999 and 2009 the Utah and U.S. crime rates decreased by 30% and 19%, respectively, compared to 24% for Ogden (according to the long-term trend). While the Ogden Police Department obviously deserves a great deal of credit for working hard to keep crime rates down, there is no evidence that Ogden is anomalous in any way or that the mayor and police chief have somehow worked miracles. The only possibly significant discrepancy between Ogden’s crime rate and the state and national trends was actually in the late 1980s, when Ogden’s crime rate decreased while the Utah and U.S. crime rates were increasing.
The graph above also shows that Ogden’s crime rate has remained significantly higher than the state and national rates. But this difference is entirely expected, because crime rates are almost always higher in cities than in suburbs or rural areas. There are far more opportunities to commit crimes in cities, where many suburban residents regularly go for work, shopping, and entertainment.
The vast majority of UCR crimes are thefts of various types. Here is a breakdown of Ogden’s UCR crimes:
To a first approximation, the UCR “crime” rate is really just the larceny rate—with significant additional contributions from burglary and motor vehicle theft.
To get beyond this focus on theft, analysts often quote just the “violent” crime rate. This statistic is dominated by aggravated assaults and robberies, with a smaller contribution from forcible rapes and a still smaller contribution from murders and nonnegligent manslaughters. Here is a graph of Ogden’s violent crime data since 1985, again showing the Utah and U.S. data for comparison:
The most striking aspect of the Ogden data is its increasing volatility in recent years, with large upward and downward fluctuations. Any long-term trend is hard to discern: One could argue either that the overall trend has been flat throughout this time period, or that there was a gradual upward trend for the first decade followed by an even more gradual downward trend since the mid-1990s. The state-wide data do show a gentle rising and falling pattern of this type, while the national violent crime rate has risen and fallen rather dramatically (with a somewhat earlier peak).
While Ogden’s violent crime rate was below the national average through 1997, it has been above the national average most years since 1998. But this is because the national average has fallen—not because Ogden’s violent crime rate has risen.
In any case, we again find no evidence of any miraculous decrease in Ogden’s crime rate. The sharp decline in violent crime since 2007 has mostly just canceled out a sharp increase during the preceding three years.
There is one more law enforcement statistic that Ogden residents have occasionally heard over the years: the number of new police officers recently hired. For example, in 2007 Mayor Godfrey claimed that Ogden had hired 18 new police offers during his first two terms, and promised to soon hire six more. But data from the FBI web site (which ultimately come from the Ogden Police Department) do not support these numbers. Instead, the size of Ogden’s police force has grown at the same average rate as the total population, rising from 120 in 1999 to 131 in 2010, with a peak of 135 in 2008.
Finally, a technical note on population: Different agencies have used different estimates of Ogden’s population in non-Census years, and some of these estimates have shown unrealistic year-to-year fluctuations. Instead of using these contemporaneous estimates, I have retroactively estimated the annual population changes by making linear interpolations between Census years. As a result, I have eliminated a few spurious fluctuations in apparent crime rate statistics that were caused by fluctuating population estimates. The following graph compares the population estimates used here to the more erratic population estimates that are found on the FBI UCR web site:
For those who would like to see all the numbers that went into the graphs in this article, you can download the full spreadsheet here. Meanwhile, comments are invited from one and all.
