Comments on: Murders in Mexico

Murders in Mexico

daksya — Wed, 07 Jul 2010 04:07:19 -0800

Statistical Analysis and Visualization of the Drug War in Mexico

By: daksya

daksya — Wed, 07 Jul 2010 04:11:13 -0800

The 2009 homicide numbers by the SNSP are out.

By: rmmcclay

rmmcclay — Wed, 07 Jul 2010 05:10:00 -0800

Well...the world gets the governments it deserves. The US has set the gold standard for waging war on drugs. Good fucking luck!

By: dragonsi55

dragonsi55 — Wed, 07 Jul 2010 05:37:31 -0800

I like that the post holds homicides as the implicit cost of the drug war. Showing a map helps elucidate this comment: The heartland of Mexico is to the south, far from the country's northern tier. The north is largely a sparsely populated highland desert region seen from Mexico City as an alien borderland intertwined with the United States as much as it is part of Mexico. Accordingly, the war raging there doesn't represent a direct threat to the survival of the Mexican regime. [FabiusMaximus] Focusing on murders could help alter one of the worst parts of a situation that cannot be 'won.' But what people have to understand is Mexico would collapse without drug money. Our agencies estimate Mexico earns $30 billion to $50 billion a year in foreign currency from selling drugs. [Charles Bowden]

By: Bathtub Bobsled

Bathtub Bobsled — Wed, 07 Jul 2010 05:39:31 -0800

Research - A Journalism - A- Style/Form - C+ To explain the last one, it's stuff like this: Like Finland in 1944, the Gulf cartel in 2010 saw the writing on the wall and decided to switch sides and ally itself with its former mortal enemy, the Sinaloa Cartel, the guys they stopped in Nuevo Laredo. Wow, this drug war goes back farther than I thought! Seriously, I thougth it was well done. The blogger will be happy to know NewsWeek will have something very very similar in next week's issue. Quite possibly, if the stars align, they'll cite this... again, if the stars align.

By: shothotbot

shothotbot — Wed, 07 Jul 2010 05:43:43 -0800

I know about Benford's law, but have never used it so I could be off base here, but doesn't this plot he puts up show that the homicide data do, in fact, follow Benford's law?

By: a womble is an active kind of sloth

a womble is an active kind of sloth — Wed, 07 Jul 2010 05:56:58 -0800

As the author mentions, these graphs are made in R using ggplot2 (I think it is ggplot2 eather than ggplot). He also makes the code available which is cool!

By: Bathtub Bobsled

Bathtub Bobsled — Wed, 07 Jul 2010 06:00:44 -0800

I had to look that up. I remember seeing something like that in college, but I don't think it was called that. Maybe I'm wrong. Anyhow, it may be beneficial to direct people to a quick explanation of Benford's Law.

By: brokkr

brokkr — Wed, 07 Jul 2010 06:10:18 -0800

shothotbot: "doesn't this plot he puts up show that the homicide data do, in fact, follow Benford's law?" You don't draw statistical conclusions from looking at plots and waving hands, but from χ²-tests. Which the plot fails, according to the article.

By: shothotbot

shothotbot — Wed, 07 Jul 2010 06:25:38 -0800

You don't draw statistical conclusions from looking at plots and waving hands, but from χ²-tests. Which the plot fails, according to the article. Agreed, I am wondering if he is doing it right because the points follow the plot very closely, considering the number of months he has to work with.

By: hat

hat — Wed, 07 Jul 2010 06:54:13 -0800

There is an overwhelming amount of data here. Thanks for the post. One thing I've always wondered about but never investigated is the specific contribution of United States consumption of (illegal) recreational drugs, broken out by drug, to some generalized measure of south-of-US crime or cartel strength. Are there numbers on this? Where I live right now, the general public believes that heroin and meth are the two big drivers of local drug-related crime, including gang turf disputes. In the Midwest where I lived while growing up, we would have blamed various forms of cocaine into the late 1990s, when it seemed like meth, heroin, and prescription drugs started to mentally displace coke as the agreed-upon local scourge. I don't think marijuana enters people's thoughts much in either place as a serious source of local organized crime, in part because we all know someone who uses it benignly or we use it ourselves. So all of this anecdotal local experience in both places leads to my (and I assume the general public's) hazy assumption that heroin, cocaine, and now meth are also the big contributors to south-of-the-border crime and cartel growth. How much truth is there to this idea? I'm mentally drawing parallels to controversial sourcing of other things we consume, like CAFO meats and conflict minerals and diamonds and sweatshop goods. Are smokers who don't know the ultimate source of their pot contributing as much to the chaos in Ciudad Juárez as heroin/cocaine users? What about meth, now that it seems like an increasing amount of production is happening in Mexico and elsewhere? There is a whole other side aspect of this, related to crime and environmental destruction within the United States caused by organized pot grows embedded in places like public redwood forests, but I'm mainly interested in the impact on Central and South America.

By: rodmandirect

rodmandirect — Wed, 07 Jul 2010 07:38:43 -0800

Nowhere in the article is it mentioned how the Juarez boys are going to kill Heisenberg.

By: brokkr

brokkr — Wed, 07 Jul 2010 07:44:14 -0800

shothotbot: "I am wondering if he is doing it right ..." Luckily you can download the data and check for yourself.

By: seanmpuckett

seanmpuckett — Wed, 07 Jul 2010 07:45:16 -0800

I opened the article, glanced at the first chart, looked closer, and then screamed. The murder rate graph looks really dramatic. Until you look at the Y scale: 8-18. Why the fuck would anyone without an agenda do that sort of visual misinterpretation? I could see a reason if you are trying to represent data from 1000 to 1020. But there's really no reason at all to skew the origin in this chart where the data should have been plotted from 0-20 and done. I don't care how honest you are in your text, how much raw data you supply or what analysis you've done -- offset chart origins can make you look dishonest and should never be used without a damn good reason.

By: vacapinta

vacapinta — Wed, 07 Jul 2010 08:04:16 -0800

He uses a zero-origin later when he compares the same data to homicide rates in the US, Brazil etc. All plotted on the same graph. And it still looks dramatic.

By: Philosopher Dirtbike

Philosopher Dirtbike — Wed, 07 Jul 2010 08:12:17 -0800

I'm suspicious of the use of Benford's law here, especially since it is one of those laws that CAN'T be exactly true - that is, with large amounts of data, we'd expect it to fail. The Chi-square test will be sensitive to arbitrarily small deviations from Benford's law as the sample size increases. It would be interesting to also do the analysis for several other countries, and compare deviations from Benford's law in several countries to one another.

By: Philosopher Dirtbike

Philosopher Dirtbike — Wed, 07 Jul 2010 08:22:09 -0800

The murder rate graph looks really dramatic. Until you look at the Y scale: 8-18. Why the fuck would anyone without an agenda do that sort of visual misinterpretation? Most statistical programs have defaults for axis limits that are the range of the data. Don't attribute something to an "agenda" when it is probably just laziness or software incompetence.

By: luciaisdada

lucia__is__dada — Wed, 07 Jul 2010 08:46:20 -0800

He says Mexico has a first world rate of cocaine consumption but that's not true. He mentions the Encuesta Nacional de Adicciones where it does say that 2.4% of the mexican population has tried cocaine but only 0.44% has used it in the last year (what he calls the number of addicts). If you look at the World drug report and the SAMHSA data, the US had a 2.2% rate of cocaine users in the past year (2007-2008). Same thing for marijuana: 1.03% of mexicans vs. 10.2% of americans consumed it in the last year. Mexico is no way a big market for these cartels as otherwise it would be implied.

By: luciaisdada

lucia__is__dada — Wed, 07 Jul 2010 09:00:45 -0800

Also, according to the World Drug Report, Mexico is the country leading the statistics of amount of cannabis herb seized so it's obvious where all that pot was going.

By: warbaby

warbaby — Wed, 07 Jul 2010 10:09:53 -0800

Beyond the statistics, I found the discussion of North et al's concept of natural states very interesting. The first link is to a review and the original papers are here and here. The point being that Mexico is stalled in a "limited access" system, with the cartels being major competitors with the state for control of armed force and large parts of the economy. In very simple terms, North's concept of a "limited access" system shares many characteristics of feudal orders where power is invested in personalities rather than institutions and rights are sharply curtailed by a class system which favors elites so strongly that non-revolutionary change is virtually impossible. The side view here is that most reactionary politics normally has the direction of trying to move society from an "open access" system to a "limited access" system. The corollary for atavistic endeavors like the Tea Party is the desire to lurch backward into a fantasy past is relatively widely held, but unlikely to ever succeed in a major way - but also unlikely to get politically resolved. Hence, they aren't going to succeed but they have been, are and will be a pain in the ass for a very long time. This dovetails nicely with some speculations on revolutionary change written by Chalmers Johnson back in the 1960s.

By: dongolier

dongolier — Wed, 07 Jul 2010 10:29:48 -0800

I dont know why the author backpeddles from a "failed-state" synopsis when the security of government participants is nil. However the author is correct in his assessment that the federal government has merely become an unwitting participant in the narcotrafico cartel strategies and turf battles. As more and more police, judges, prosecutors, witnesses and other "good guys" die; the remaining survivors are presented with an increasing moral hazard to defect---which is exactly what happened in the late 1990s where mexico's highly trained special forces defected to the Sinaloa/Gulf Cartel to become the Zetas. The death toll simply follows the money; the financial stakes are in the billions of dollars. If the US is genuinely interested in stemming the violence in mexico, they have no choice but to lower the incentives and the street price of these narcotics north of the border. And by that i mean lower demand for them (impossible) or raise the domestic supply (legalize and regulate). warbaby: to modify your point slightly: mexico has always had less power invested in personalities and much, much more in monopolies (ie. cartel is a misnomer---in this discussion we really mean monopoly). This country is home to monopolies on corn, beer, oil, cement and telecommunications---its no wonder drug trafficking seems like an open business opportunity.

By: dongolier

dongolier — Wed, 07 Jul 2010 10:38:57 -0800

I think you may get a certain Cult of Personality necessary to create and maintain a monopoly; a true "limited access" mexican oligarchy exists either with and without charismatic or powerful personalities---North is right on the money.

By: Maias

Maias — Wed, 07 Jul 2010 20:28:25 -0800

Drugs themselves as pharmacological agents aren't a big cause violence-- and in fact, the drug most strongly correlated with violence is alcohol and that's not even true in all cultures, oddly enough. Drugs simply magnify mental states and cultural beliefs-- so if you believe a drug will make you violent, it may do so but this won't happen most of the time unless you have underlying violent intent or predisposition. That's not to say that drugs have no pharmacological action, but the reaction to cocaine-induced paranoia by a shy, anxious person will usually be to hide while the reaction of a tough, thuggish person to it may be to kill someone. So, linking particular drugs with violence is somewhat silly: what's most directly linked with violence is the monetary value of the drugs, the unemployment rate and instability of the drug trade and the number of transactions involved (this is why crack was especially problematic: shorter lasting high, sold in small doses = more transactions = more chances for the transaction to go wrong and result in a fight). A study I read that tried to quantify the link between violence and cocaine found that something like 80% of the violence was straight-up trade disputes which were related to the drug's illegality (you can take someone to court if they rip off a legal product and have insurance on it), and the rest was either pharmacological (ie, someone got high and paranoid and hurt someone) or there was one other category which I now fail to recall. Anyway, point being that illegality is the biggest cause of drug-related violence, not drugs themselves.

Comments on: Murders in Mexico

Murders in Mexico

By: daksya

By: rmmcclay

By: dragonsi55

By: Bathtub Bobsled

By: shothotbot

By: a womble is an active kind of sloth

By: Bathtub Bobsled

By: brokkr

By: shothotbot

By: hat

By: rodmandirect

By: brokkr

By: seanmpuckett

By: vacapinta

By: Philosopher Dirtbike

By: Philosopher Dirtbike

By: lucia__is__dada

By: lucia__is__dada

By: warbaby

By: dongolier

By: dongolier

By: Maias

By: luciaisdada

By: luciaisdada