Americas view | Crime in Mexico

The governor’s miraculous achievement

Mexico state's governor plays fast and loose with the murder rate

By T.W. | MEXICO CITY

IT IS the season of informes in Mexico, when the president and state governors deliver state-of-the-union-type reports on what they have been up to. This year Enrique Peña Nieto, the outgoing governor of Mexico state and the early front-runner in next year's presidential election, used his September 5th informe to mark the unofficial launch of his presidential bid.

Prominent in his address was the stunning claim that the murder rate in Mexico state had fallen by more than half during his six-year term. “One of the most illustrative achievements that we have is the reduction in murders per 100,000 people, from 16.5 in 2005 to 7.6 in 2010,” he said (you can watch it here at about 1:05). This was particularly amazing given that the national murder rate more than doubled during the same period. Anyone doubting Mr Peña's word could see the numbers for themselves in the print version of his informe (here, on page 222, under “homicidios dolosos”).

The claim is absolutely false. The numbers listed in Mr Peña's informe are indeed the official figures, but they make no mention of a statistical revision in 2007 which saw the murder rate halve overnight. You can see a month-by-month breakdown on the website of the National System of Public Security (SNSP), a federal body. Compare December 2006 with January 2007, when the new methodology was introduced, and you will see that the number of murders in Mexico state magically fell by 62% in the space of a single month.

I noticed this when writing a story about crime in Mexico state a few months ago. Mr Peña's aides explained to me that the revision involved striking out deaths which had previously been mistakenly classified as murders (suicides, accidents and so on). They told me at the time that they were confident that the new methodology was more accurate, and they may well be right. What no one can doubt is that it is misleading nonsense to compare murders after 2007 with those from before, because they were measured in completely different ways.

So what happens if we look at murder trends in the state from 2007 until now, using consistent methodology? According to the SNSP, in 2007 there were 1,127 murders, against 1,153 in 2010. Taking population growth into account, this means that the per-capita murder rate stayed pretty flat—a decent result given Mexico's well-documented problems with public safety.

Lately, however, things have taken a turn for the worse. If you look at the figures for the first seven months of 2011—which are listed by the SNSP but, curiously, are not in Mr Peña's informe—you see an alarming recent increase in murders in Mexico state. January to July of this year saw 837 murders, some 40% more than the 597 committed during the same period in 2007, at the beginning of Mr Peña's term. Even taking population growth into account, this means that the murder rate in Mexico state has risen substantially under Mr Peña. Will he continue to claim the opposite?

UPDATE, SEPTEMBER 27th: Mr Peña has today published a response to the above criticisms in an online article (in Spanish). To his credit, he accepts that the comparison he made in his informe was wrong. “The figures underwent a methodological modification in 2007...The criticism that The Economist makes is that it is technically inconsistent to compare figures derived using different methodologies. I share this view. In a democratic society, we politicians have to accept criticism, especially when it is founded,” he says.

He goes on to quote different figures—this time from INEGI, the national statistics-agency—which show that between 2005 and 2010 the murder rate in Mexico state was fairly stable, falling at first and then rising somewhat. He drops the claim that the murder rate fell by more than half, now writing only that it was “contained”, which is true, at least as far as 2010. INEGI does not yet have any figures for 2011 (during which time the murder rate in Mexico state rose, according to the SNSP, as quoted above). INEGI is likely to publish preliminary figures for 2011 at the end of July—just after the presidential election, which might turn out to be fortunate for Mr Peña.

One worrying thing to take from this episode is that although the numbers of the SNSP and INEGI show broadly similar trends, they are radically different in gross terms. In Mexico state, for instance, Inegi registered 2,096 murders last year, whereas the SNSP (which gets its figures from state prosecutors' offices) registered only 1,153. Nationwide, INEGI registered 24,374 murders, whereas the SNSP registered 20,585. That's a difference of 3,789 missing victims—roughly five years' worth of murders in Britain, for instance. As the election campaign heats up, expect plenty of arguments about which of these data sets is the more reliable.

Discover more

Business backlash

A weakened Enrique Peña Nieto faces calls to roll back his tax reform

Back to the table

The FARC's kidnapping of a Colombian general last month did not kill the peace process


The new brooms

Dilma Rousseff's new economic team talk about their plans