Investigators with the body of a woman in Mexico City in mid-December.
By DAMIEN CAVE
The Mexican government updated its drug war death toll on Wednesday, reporting that 47,515 people had been killed in drug-related violence since President Felipe Calderón began a military assault on criminal cartels in late 2006.
The new official tally provided by the attorney general’s office included data only through September, and it showed that drug-related killings increased 11 percent, to 12,903, compared with the same nine-month period in 2010. Still, a government statement sought to find a silver lining, asserting that it was the first year since 2006 “that the homicide rate increase has been lower compared to the previous years.”
But that will hardly calm a public scared by the recent arrival of grisly violence in once-safe cities like Guadalajara, nor will Wednesday’s limited data release silence the increasingly loud call for better, more transparent government record keeping.
The Mexican government has failed to create the tracking system it needs to understand criminal trends and improve security, experts say, even as it has become more secretive with the limited information it has.
“Our frustration is that they have some information and some numbers, something that would be valuable and they are not releasing them,” said Eric Olson, a security expert at the Woodrow Wilson International Center for Scholars in Washington. “And there is a whole bunch of other things that are not well defined and can lead to erroneous conclusions.”
The number of drug-related deaths is the subject of much dispute. Government officials last gave a figure — 34,612 — at the end of 2010, promising to update their tally regularly. They did not follow through. A group of Mexican and American academics, including Mr. Olson, began pleading with the Calderón administration for death figures, along with other data known to be collected, including violent episodes involving the military. But members of the group say they were ignored.
Pressure began to mount late last year as the government received several public records requests seeking information on crime-related deaths nationwide. The Calderón administration initially said that the data was confidential for reasons of national security, then last week the government said that the figures would be published after further study. The release on Wednesday came after Mexico’s freedom of information agency said it would ask for an investigation if the data was not released.
Now, the question is whether the report accurately reflects the reality on the ground. Some Mexican news organizations have arrived independently at similar totals, whereas others have found that the government regularly undercounts the number of drug-related deaths.
“Since there are very few actual investigations, those are approximations at best,” Mr. Olson said. “They’re hunches. There is not really a way of knowing precisely if it was caused by organized crime or a drug trafficker or not.” Molly Molloy, a librarian at New Mexico State University who closely tracks deaths in Ciudad Juárez and other parts of the country, said that given the investigative failures, the most reliable figures come from the Mexican census agency, which identified 67,050 homicides from 2007 through 2010, nearly double the government’s count of drug-related deaths for that period.
Inconsistent responses to The New York Times for requested public records from every Mexican state suggest that the data problems begin at the local level. The federal tally of drug-related homicides is at least partly a compilation of figures from state prosecutors’ offices, but these offices do not appear to have consistent systems for recording or releasing such data.
While some state prosecutors’ offices complied with requests for information on drug-related deaths, others denied tracking the most basic information, arguing it was the job of the federal government to do so. Veracruz, which has grown increasingly violent in the past few months, said in a letter that it “does not generate the statistic you request,” and furthermore, that the information was “reserved and/or confidential.”
Karla Zabludovsky contributed reporting.