Louisiana auditor: NOPD crime statistics flawed

The New Orleans Police Department may have substantially underreported the amount of serious crime that occurred in the city last year, according to a report released Monday by the state’s legislative auditor.

The report was requested by members of the Legislature after a Times-Picayune article published in May raised questions about the validity of some of the city’s crime data, in particular, its relatively low rate of aggravated assaults compared with its sky-high murder rate.

Of 1,000 incidents reviewed, 319 should have been reported as serious crimes to the state and to the FBI but weren’t, the auditors wrote, though they said they purposely picked a group of dispatched calls to police they thought were at “high risk” of being misclassified.

In a four-page written response to the report and a news conference Monday morning, Police Superintendent Ronal Serpas took issue with some of the auditors’ methodology, while also saying the audit generally supports his repeated assertions the NOPD’s crime statistics are accurate on the whole, if imperfect.

Like most big-city police departments, the NOPD keeps track of “Part I” crimes in various standard categories — murder, rape, robbery, aggravated assault, arson, theft, burglary and car theft — and sends the data to the FBI for the bureau’s annual “Crime in the United States” report.

Nearly a third of the cases the auditors looked at involved crimes that should have been reported to the FBI, but were not.

The authors cautioned their work should not be used to estimate an overall “error rate” for the NOPD’s crime statistics, because the auditors purposely examined incidents they considered more likely to be wrongly classified.

Among the incidents that auditors said were not reported to the FBI, but should have been: a case in which a woman was robbed at gunpoint at an ATM, at least four assaults, including one in which a gun was fired, the theft of various items from a person’s back porch and 11 kidnappings that included “Part I” crimes such as assault. Kidnapping itself is not a Part I crime.

In more than half of the cases the auditors flagged, police officials agreed the incidents should have been reported to the FBI, based on what the officer wrote in the report. The NOPD blamed almost 40 percent of the misclassifications on a “computer programming error” Serpas insisted has been remedied.

In the other cases, there was no report to thoroughly double check the NOPD’s classification because the incident was categorized as “(victim) gone on arrival,” “unfounded” or “necessary action taken.”

That means an officer determined the crime originally reported to the dispatcher did not occur or could not be investigated.

Auditors said in the cases they highlighted, there was “documentation that suggested the incidents were Part I ... reportable crimes.”

Auditor Lauren Whatley said she based her determination on notes taken by the dispatcher.

Serpas, in his written response, argued the auditors were wrong to assume the dispatcher had it right.

“Only a trained police officer on the scene, talking to witnesses, and collecting evidence, can really get the facts,” he wrote.

He added the department does its own follow-ups on randomly selected reports, calling victims and visiting crime scenes, to ensure the record reflects what complainants told police.

Serpas also repeatedly emphasized the audit found no evidence of intentional “downgrading” of crimes by police to make the city appear safer — something he says critics have accused the NOPD of doing.

Moreover, Serpas added auditors “cherry-picked” the calls they chose to review, having the effect of magnifying the department’s errors in crime classification.

He said the NOPD conducted its own, more thorough audit of 28,000 incidents that were dispatched as “Part I” crimes, and found an overall error rate of less than 2 percent.

Finally, Serpas underscored the imperfect nature of crime reporting, and ticked off various safeguards the department has implemented to get it right.

He also listed seven steps the NOPD will take to correct the glitches the auditor found.

In his news conference, Serpas said police chiefs often are forced to answer questions about the validity of their crime statistics, particularly when they claim crime is on the wane.

While Serpas claimed an error rate of less than 2 percent, and the auditors cautioned against applying the error rate they found more broadly, the report’s findings suggest crime in New Orleans could be far higher than the official rates.

Of the 1,000 incidents the auditors reviewed, 650 were called in to dispatchers as “Part I” crimes but were ultimately not reported to the FBI.

The other 350 were randomly selected incidents that were neither called in as “Part I” crimes, nor reported as such.

Of the first group, nearly half were misclassified, according to the auditors.

But because the overall number of such incidents is relatively small, even applying the error rate the auditors found would nudge the city’s crime rate up by less than 10 percent.

The second group involves a huge number of cases: There were nearly 475,000 calls to police dispatched in 2012 that were never reported as “Part I” crimes.

Auditors only found 12 errors in the 350 incidents of that type they examined, translating to an error rate of more than 3 percent.

Although the 350 reviewed incidents are a small sample, if a similar percentage of the 475,000 incidents were misclassified, the city’s true crime rate could be close to 100 percent higher than what was reported to the FBI.

The legislative auditor’s report did not attempt to assess whether certain types of crimes, for whatever reason, were underreported.

The Times-Picayune article in May that prompted the review found New Orleans’ rate of violent assaults was abnormally low compared with other “peer” cities with high murder rates.

Criminologists generally agree murder is the crime most accurately recorded by police departments, because there’s less wiggle room about what constitutes a murder.

Several criminologists said in that May report the low ratio of assaults to murders reported by the NOPD hinted at undercounting of assaults.

As part of their analysis, the authors of the May report examined 66 assault reports from a random week in 2012, and found that six of them were wrongly classified as simple assaults rather than aggravated assaults.

Police agreed that three were misclassified, but stood by the others.

Experts said that underreporting of violent assaults could be attributable to a variety of causes, such as reluctant witnesses, downgrading by police or even slow response times.

The state audit, for instance, found in several of the incidents the NOPD wrote up as “gone on arrival” and thus not reportable to the FBI, police took more than three hours to arrive at the scene.

“If you’re a tourist who’s assaulted or robbed at gunpoint, you’re probably not going to wait at the scene for three hours,” said state Sen. J.P. Morrell, D-New Orleans, who authored the resolution asking for the audit.

“That doesn’t mean a crime didn’t take place. It could be a staffing issue.”

Morrell, who has two brothers in the NOPD, said that aspect of the report might present an argument for hiring more police.

He also said he was puzzled by Serpas’ response, which quibbled with many of the report’s findings but also promised to take various actions to correct the problems the auditors said they found.

“There’s sort of a complete denial that anything is wrong, and then here’s how we’re going to fix it,” Morrell said. “It can’t be both.”

The state audit is the first in a series of reports on the accuracy of New Orleans crime data that are expected to become public in the near future.

When Morrell authored his resolution in May, the city’s Office of Inspector General announced that it, too, was engaged in an audit of New Orleans crime statistics. Inspector General Ed Quatrevaux said by email Monday that the office is actually writing a series of reports.

The first report is on “improper classification of certain crimes in a certain (police) district,” Quatrevaux said.

The subsequent reports will each focus on a certain type of crime and how accurately the NOPD is recording it. The first of those reports focuses on rape, Quatrevaux said, and will likely be published in mid-December.

The legislative auditor’s report on the New Orleans Police Department’s crime statistics can be viewed here.