Tracking Mexico's cartels with Google

For the vast majority of us, who don't have access to high-level law enforcement data, the best way to keep track of Mexican drug cartel activity is by reading the newspaper. But with information often dispersed in local sources and reporting on the ground becoming increasingly dangerous, it can be difficult to get a big picture view of how the drug war is progressing. But a new tool developed by two Harvard graduate students could help provide such a broad view.

Viridiana Rios and Michele Coscia have created an algorithm they call MOGO (Making Order Using Google As an Oracle) which processes Google data to track cartel activity. "MOGO does the jobs we could never do," Rios told me in a phone interview today. "It reads all of the newspapers that have ever been published in the last 20 years and extracts information about whether and when a particular cartel is mentioned and where that cartel is mentioned. We get the organization, the municipality in which it is supposedly operating, and the year in which the note was published."

That raw data can then be used to show how a particular cartels area of operation has changed over time. For instance, here's how the Juarez cartel grew over the last decade:


And the Zetas:


And here's a picture of how Mexico's various crime organizations have compared in prominence over the last two decades, with the Zetas, Beltran Levya, and Familia groups bursting onto the scene, and Juarez seeming to fade in comparison:


Rios believes these maps give a more nuanced view of cartel activity than existing ones created by private intelligence firms. "Cartels are not operating all around Mexico," she says. On Stratfor's maps it shows every part of Mexico with dominance by one cartel. That is not true. We analyzed about 2400 municipalities and in those 2400 we only found presence of drug cartels in 730."

While the data MOGO gathered may not completely overturn conventional wisdom on cartel activity -- after all, it's drawn from previously published reports -- it seems like a powerful new tool for analysis and it's easy to think of applications well beyond Mexico.


Will U.S. military fatalities sway voters on Election Day?

Time and again, we've been told that the economy is the most important issue in this year's campaign. But political scientist Douglas Hibbs believes economic indicators aren't the sole predictors of election outcomes. His "Bread and Peace" model, which forecasts the winner of presidential races based on growth in personal, after-tax income and American fatalities in unprovoked wars, is the only forecasting model I've come across that takes foreign policy into account in divining whether Barack Obama or Mitt Romney will triumph next week.

And unlike several competing models that focus on economic measures, Hibbs's has Romney beating Obama -- by a comfortable margin of 53-47.

So, is it the Peace component that's tipping the scale for Romney? Not so much. "There's not much action, I don't think, from Afghanistan on Obama's vote share," Hibbs told Foreign Policy, adding that there have been roughly 1,500 U.S. military fatalities in Afghanistan under the president's watch. "Proportional to U.S. population, that's just way too small to have great electoral effect." It's paltry income growth and the sluggish economic recovery, he argues, that could dash Obama's quest for a second term.

Hibbs, who cites John Mueller's 1973 study War, Presidents, and Public Opinion as a major influence on his work, says that since World War II, troop fatalities have only played a decisive role in two elections: 1952 and 1968. In those instances, the bloody Korean and Vietnam wars torpedoed the campaigns of Adlai Stevenson and Hubert Humphrey despite favorable economic conditions.

"Absent America's interventions in the Korean and Vietnamese civil wars, the strong real income growth record prior to those elections (particularly in 1968) should easily have kept the Democrats in the White House," Hibbs wrote in a recent article for the journal PS: Political Science and Politics. His graph below shows the extent to which the 1952 and 1968 election results were outliers when you plot income growth against the incumbent party's share of the vote. As Hibbs sees it, the deeply unpopular wars raging during those two years produced the anomalous outcomes:

For some perspective on how today's post-draft, high-tech wars differ from Korea and Vietnam, consider this: There were 29,260 U.S. military deaths (190 per millions of U.S. population) in Korea at the time of the 1952 election and 28,900 (146 per millions of U.S. population) in Vietnam at the time of the 1968 election. The 1,500 fatalities under Obama, by contrast, amount to roughly five deaths per millions of U.S. population. 

In his model, Hibbs distinguishes between "provoked" and "unprovoked" conflicts, with the implicit assumption that voters are more willing to stomach fatalities when the United States has been attacked. And he posits that when voters are unhappy about military casualties, they punish the party that initiated the deployment of U.S. forces. According to this logic, Americans won't hold fatalities in Iraq against Obama this year because he inherited the war from George W. Bush. But they will attribute fatalities in Afghanistan to the president, since Bush invaded Afghanistan in response to al Qaeda's provocation on 9/11 while Obama recast the conflict as a "war of necessity" and ordered a troop surge in the country. 

It's unclear from polling whether voters make these distinctions and consider the Afghan conflict Obama's war, but Americans have clearly soured on the military engagement. According to a Pew Research Center poll last month, 60 percent of Americans want to remove troops from the country as soon as possible regardless of whether the situation there is stabile, and more than half think the military effort isn't going well. When pollsters ask voters what the most important issue facing the country is, the war in Afghanistan typically garners no more than 5 percent of responses. 

I asked Hibbs whether military fatalities would have a more pronounced impact on the electorate if the media covered the Afghan war more extensively. "In an open and democratic society with quite a free and inquisitive and aggressive press, the press follows the reality, it doesn't create it," he responded. "If we had 1,000 guys coming home in body bags a month in Afghanistan like we did in Vietnam ... the press would be all over it."