Cross-border Spillover: U.S. Gun Laws and Violence in Mexico

[DanTex]

Here is a study on the effects of the expiration of the assault weapons ban on violence in Mexico, presented at an NBER conference this summer. By comparing the differential increases in homicides and gun crimes in areas near TX and AZ, versus the areas near CA which still has a state-level assault weapons ban, the authors find, among other things, that at least 158 additional deaths each year in Mexican border regions can be attributed to the expiration of AWB.

Do more guns cause more violence? We exploit a natural experiment induced by the 2004 expiration of the U.S. federal assault weapons ban to examine how the subsequent exogenous increase in gun supply affected violence in Mexico. The expiration relaxed the permissiveness of gun sales in border states such as Texas and Arizona, but not California, which retained a pre-existing state-level ban. Using data from mortality statistics and criminal prosecutions over 2002-2006, we show that homicides, gun-related homicides and gun crimes increased differentially in Mexican municipios located closer to Texas and Arizona ports of entry, versus California ports of entry. Our estimates suggest that the U.S. policy change caused at least 158 additional deaths each year in municipios near the border during the post-2004 period. Notably, gun seizures also increased differentially, and solely for the gun category that includes assault weapons. The results are robust to controls for drug trafficking, policing, unauthorized immigration, and economic conditions in U.S. border ports, as well as drug eradication, trends by income and education, and military and legal enforcement efforts in Mexican municipios. Our findings suggest that U.S. gun laws have exerted an unanticipated spillover on gun supply in Mexico, and this increase in gun supply has contributed to rising violence south of the border

http://www.nber.org/public_html/confer/2011/SI2011/CRI/Dube_Dube_Garcia_Ponce.pdf

[Remmah2]

1. Is it possible the ATF lied about the gun traces?

They've broken the rules otherwise? Is the ATF credible? Is the ATF reliable? Is the ATF honest?

[iverglas]

2. oh, c'mon

"Science"?

:eyes:

That calls for multiple unrec-ing.

Just wanted to note that the study period largely excludes the period of Gunrunner, which began in 2005 in Texas and only in 2006 nationally, while the effects of the repeal of the AWB were found to be immediate in 2004.


One almost wonders ...

http://www.statcan.gc.ca/pub/85-002-x/2011001/article/11561-eng.htm#a4





Interesting little post-2004 peaks in Canada there.

[aikoaiko]

3. How odd since one could still buy AR 15 and AK style semi auto rifles before 2004

Do you think the addition of collapsible stocks, flash hiders and bayonet lugs were what finally made these rifles useful to criminals?

:shrug:

[gejohnston]

4. Could it be that California is kind of a backwater?

If reality operated in the same vacuum as the economic model, it might be useful. The real world is not so linier. One reason why I am not a fan of economists playing amatur intelligence anylist or criminologist.
A better question to ask would be "why is the violence less at the California border than the rest?" It allows for more factors instead of concentrating on one possible reason, one that is most likely a minor factor if it is at all. For example:
http://geo-mexico.com/?p=3536

Not much of a trade route to fight over. You also have a smaller and probably more secure border. Texas and Arizona's borders are more direct from the cocaine fields and meth factories.

This one shows where the disputed territories are. The most relevent one is closest to.....Texas. The gang that controls Baja, seems to be on its last legs. Their top management is in jail.
http://geo-mexico.com/?p=1166

As the conflict in Mexico among various drug cartels, and between the cartels and the Mexican
government has intensified, tracking of automatic rifles, rocket propelled grenade launchers
and high caliber machine guns have increased dramatically. Guns are typically not sent via
tunnels in the desert or across the Rio Grande by boats, but through commercial ports of entry.
In particular, "firearms are generally tracked along major U.S. highways and interstates and
through border crossings into Mexico. The firearms are normally transported across the border
by personal or commercial vehicle because, according to U.S. and Mexican government officials,
the drug cartels have found these methods to have a high likelihood of success

How is this relevent? Machine guns and rocket launchers going from the US? These weapons had nothing to do with the AWB. Do the writers think you can buy these at Wal Mart or gun shows in the US? Not that the AWB really was a ban anyway. Take the same gun, change a few cosmetic features, you now have a non "AW" that functions exactly the same as before.

It repeats the "90 percent of guns traced by ATF" meme, but does not mention the guns not given to the ATF for tracing. That only means that the guns were manuctured or imported in the US. That includes their own military weapons. The vast majority of the weapons, not given to the ATF, is not even addressed.

http://en.wikipedia.org/wiki/Mexican_Drug_War#Ju.C3.A1rez_Cartel

So far my conclusion is be that it is simply a coincidence at most.

[slackmaster]

5. LOL@ complete gobbledy-gook fake "study"

:lol:

[Euromutt]

6. I don't see it

Compare this chart, showing homicides and cartel activity areas; the state worst affected is Chihuahua, with Sinaloa coming a consistent second. But note that Baja California, which shares shares almost all of its international border with California, has a worse record than Sonora, which shares all of its international border with Arizona.

Note moreover that in 2010, Tamaulipas (bordering on Texas and the Gulf) saw a massive increase in homicides, following the dissolution of the alliance between the Gulf Cartel and Los Zetas, and their subsequent turning on each other. Chihuahua is the battleground between the Sinaloa and Juarez cartels for control of the smuggling conduits around Ciudad Juarez and El Paso. But Coahuila, despite being wedged between Tamaulipas and Chihuahua and bordering Texas, has not seen a comparable increase in violence, because it's uncontested Zetas turf. Similarly, the Sinaloas' control over the routes into Arizona (primarily via Nogales and Agua Prieta) also seems mostly uncontested.

Very simply, the violence is heaviest where various cartels are contesting each others' control over the smuggling routes, which is why it's consistently been the case since 2007 that ~70% of organized crime-related homicides occur in 3-4% of Mexico's municipalities. If the Dubes' and Garcia-Ponce's claims were correct, we should be seeing more violence in Sonora (bordering Arizona) and Coahuila (bordering Texas) than in Baja California (bordering California). But we don't.

[DanTex]

7. Of course, there are other factors besides gun availability at play...

I'm not sure how closely you followed the study, but the authors are well aware of the fact that changes in drug cartel activity are also a major driver of increases in violence. This is precisely the reason that the study was subjected to so many robustness checks and controls, to isolate the effect of the policy change rather than such other factors.

One important thing to realize is that there can be (and, in fact, there are) multiple drivers to violence. The fact that one major contributor is the geography of drug cartel activity obviously does not imply that gun availability is not also a factor. This is a fallacy that I find frequently with people attempting to deny the many studies finding that guns have a positive influence on homicide. As an analogy, if a study finds that exercise is a factor in heart disease, pointing out the diet is also a factor does not refute the study.

Next, this study examined differential changes in homicide rates, not absolute rates. Obviously, if there is more violence in one region than another both before and after the policy change, that doesn't tell us much about the effect of the "natural experiment" which is the AWB expiration.

Also, the study looked at the data at a much more granular level than that economist map, at the municipal level rather than state level. The kind of analysis they performed is impossible to either validate or refute using such state-level data, as they found statistically significant differential increases in homicide specifically in municipios close to border crossings which are major ports for gun flow in TX and AZ (see figure 5 for a visual representation). And, as a test for robustness against general geographic trends they re-ran the tests first excluding all areas near the AZ border, and then excluding all areas near the TX border, and in both cases the statistical significance held up.

And then there are all of the different control variables used, including controls for "drug trafficking, policing, unauthorized immigration, and economic conditions in U.S. border ports, as well as drug eradication, trends by income and education, and military and legal enforcement efforts in Mexican municipios". As the authors point out, some of the robustness checks actually are over-controlling, because one would expect that the increased gun flow would produce a general increase in criminal and drug cartel activity, rather than just an increase in homicide specifically. However, the fact that the increases in homicides are still significant even when potentially over-controlled in this way makes for a "particularly tough hurdle" in terms of a test of robustness.

A last thing I'd point out is that, since this study focused on localized increases in violence in border areas, it is likely that it underestimates by a significant amount the overall increase in violence attributable to the AWB expiration. Obviously, weapons from the US make it further into Mexico than just 100 miles from major highway border crossings, and this would result in an overall, non-localized increase in violence. The portion of the overall increase in violence in Mexico attributable to the AWB is more difficult to measure than the localized increases, because with data points and it is more difficult to isolate the effect of the AWB from other factors. Thus, studying localized effects is useful for demonstrating that the expiry of AWB did have a significant effect, but one should not assume that the effects are limited to just these localized increases in violence, because it does not account for the general increases due to diffuse gun flow throughout the rest of Mexico.

[Euromutt]

15. Hold on, there's another difference I missed

The analysis specifically excludes the time period after 2006, when the government of Felipe Calderon sent the Mexican army into numerous states initiating a major escalation in the drug war.

So I was mistakenly comparing data from 2008-2010 to those from 2002-2006. My bad.

But that does mean that, if we take the authors' conclusions as read, that the absence of an "assault weapon" ban played a role in 1% of the organized crime-related homicides that occurred in Mexico last year.

[Atypical Liberal]

8. The AWB could have had no effect on gun crime.

The AWB did not do anything to change the functionality or availability of assault weapons. It made them slightly more expensive.

All manufacturers and importers of these weapons did was make some minor cosmetic changes and alter the US-made parts count to the firearms to make them comply with the law.

So for an AK-47, they ground off the lug for a bayonet, ground off the threads on the end of the barrel, and replaced the grip and a few internal parts with US-made parts.

None of this would make any difference to a criminal in Mexico.

The firearms used the same old magazines, and fired the same old ammunition in the same old way.

[jeepnstein]

9. Hasn't been a problem for Canada.

If the presence of guns instantly kicks off an orgy of shootings, beheadings, and unspeakable acts involving 55 gallon drums then why isn't Canada in flames as well? I'm thinking it has more to do with street drug smuggling routes than American gun laws. Why don't we have large Canadian gangs spilling over into American territory? We don't, and for good reason. It's the drug supply that is fueling the violence.

[DanTex]

10. Canada didn't have violent drug gangs before the AWB expired...

Mexico did. But there was a statistically significant differential increase in violence in areas near border crossings in AZ and TX, relative to areas near border crossings in CA, where a state-level AWB continued.

Nobody is denying that there are other factors besides gun availability, nor is anyone suggesting that "the presence of guns instantly kicks off an orgy of shootings, beheadings, and unspeakable acts..." As I mentioned to Euromutt, the inability to understand that there can be more that one factor that determines levels of gun violence is probably the single most common fallacy I've come across when it comes to pro-gunners attempting to deny all the evidence that gun availability contributes to gun violence. Yes, the geography of drug cartels and other things are obviously important factors, but the sudden availability of a huge supply of weapons particularly suited for drug cartel violence at certain locations -- near AZ and TX border crossings -- as compared to regions near CA border crossings, where there is still plenty of drug supply, but an AWB is still in place, makes for a good natural experiment of the effect of the AWB on gun violence in Mexico.

BTW, lax US gun laws do contribute to gun violence in Canada, but there, it is the ease of purchasing and trafficking handguns (e.g. no background checks in the secondary market, no registration), rather than AK's and AR's that are the problem. Drug cartels have more use for AK's, while street criminals favor handguns.

[jeepnstein]

11. Drug gangs fund private submarines.

I'm thinking that my eliminating my right to own a firearm is not really going to make much difference to their business model. There are plenty of sources for cheap AK's, real ones sporting full-auto. Any organization that already is used to dealing with international smuggling and that owns it's fair share of politicians is not going to worry the least bit about a source for small arms.

The violence has more to do with the fact that smuggling routes get restricted along the border and as groups compete for that scarce resource, well, head will roll. That area along the border is vital for business and they protect that turf any way they can.

[DanTex]

12. Yes they do.

As I said before, the violence has to do with a lot of things. The fact that the growth of violence was significantly greater near gun trafficking crossings in states where the expiration of AWB produced a change of policy, versus crossings in CA, is pretty good evidence that gun availability is a contributing factor. And there is plenty of other evidence of the ill effects of lax US gun laws in Mexico, as well as in Canada and other nations.

I would find your argument about RKBA more persuasive if you were willing to accept the facts, rather than attempting to deny them. That loose gun laws do have negative consequences has been confirmed repeatedly in many different kinds of empirical studies. Simply ignoring all of the evidence does not make it go away. And, acknowledging the truth that repealing the AWB cost lives doesn't mean you have to give up supporting RKBA. If you supported RKBA as a fundamental right in earnest, you wouldn't be afraid to acknowledge this reality -- you could instead concede that, yes, lax gun laws do cost lives, but you believe that the rights issue is more important. That would be an honest position.

But to deny the evidence and the science, just to make RKBA appear painless, to pretend that there are no negative consequences to lax gun laws, this is disingenuous.

[jeepnstein]

13. My Constitutional Rights...

do not in any way shape or form contribute to anything other than a pretty good way of life for my family. I'm responsible, law abiding, and active in my community. You are advocating some kind of collective punishment against folks like me because the drug cartels are blood thirsty. I can't help that. I don't buy illegal street drugs and I certainly don't go around arming anyone. It's just not my fault. I'm not contributing to the drug violence problem in Mexico or anywhere else.

Would you be willing to accept giving up your speech rights, or your protections against unlawful search and seizure if it were for the greater good of "reducing violence"? I'm not giving up any of my rights thank you very much. None. It's just going to happen.

That's my honest position. I would suggest you open a dialogue with the drug gangs and see how they can work to reduce violence in their communities and work places. I'm sure that will be a productive use of your time.

[iverglas]

14. forgive me for contradicting you

but it very much is a problem for Canada.

In order to get the cocaine and other "hard" drugs to market in Canada from other sources via the US, would-be distributors here have to have something to trade. The loonie is not in high demand. Pot is, in the US.

So pot is grown in Canada for export to the US. And the organized crime element that grows most of that pot is just like organized crime everywhere: vicious and violent.

The stakes aren't as high and the society and economy aren't as chaotic in Canada as in Mexico, so we don't have all-out war. But we have a significant problem.

http://www.statcan.gc.ca/pub/85-002-x/2011001/article/11561-eng.htm

Gang-related homicides, Canada, 1991 to 2010


Overall, organized crime and/or gang activity is related to fewer than 1 in 5 homicides in Canada each year. According to the Criminal Intelligence Service Canada (CISC), extreme violence is generally counter-productive for organized crime groups as it both distracts from profit-oriented activities and attracts the attention of law enforcement (CISC 2010). In the Homicide Survey, incidents are classified as “gang-related” when police believe the homicide occurred as a consequence of activities involving an organized crime group or street gang. Homicides of innocent bystanders who are killed as a result of gang-related activity are also considered to be gang-related.

In 2010, 94 homicides were considered by police to be gang-related, accounting for 17% of all homicides reported to police. This represented a 25% drop and the second annual decline, following a high in 2008 when 138 homicides were reported by police as gang-related (Table 6). Despite these recent declines, the rate of gang-related homicide has generally been increasing in all provinces since the Homicide Survey began recording this information in 1991 (Chart 8). The only exception is in Quebec, where gang-related homicide was at its highest in 2000.

... The characteristics of gang-related homicides tend to differ from other types of homicides in a number of ways. Compared to homicides that were not gang-related, gang-related homicides in 2010 were more likely to have been committed by more than one accused person (66% versus 13%), to have involved the use of firearms (76% versus 18%) and to have been related to the illegal drug trade (such as trafficking or settling of drug-related accounts) (62% versus 9%). The most common drugs identified in gang-related homicides involving drugs were cocaine (51%) and cannabis (31%).

Victims of gang-related homicides, like persons accused in these incidents, are usually male, relatively young and are often involved in criminal activities themselves. More specifically, in 2010, close to 93% of gang-related homicides involved a male victim, compared to 66% of other homicide victims. Victims of gang-related homicides were also younger on average than other homicide victims (31 and 36 years, respectively), though not as young as persons accused in gang-related incidents (24 years on average).

Close to 7 in 10 victims in gang-related homicides (68%) had a criminal record, lower than the proportion among persons accused in gang-related incidents (88%). Victims of these homicides were also more likely to be involved in criminal activities themselves. Police respondents recorded illegal activities as the main source of “employment” for 7 in 10 victims of gang-related homicide (71%), six times higher than for other homicide victims (12%). The most common motive3 recorded by police for gang-related homicide was the settling of accounts (61%).

Gang-related homicides in Canada have notable non-gang-involved victims: mistaken identity, bystanders including children, and simply terrorizing of neighbourhoods by random shootings are seen.

Now, consider the effect on overall homicide numbers and rates:

Following a decade of relative stability, homicides decreased substantially in 2010. There were 554 police-reported homicides in 2010, 56 fewer than the year before ... . The 2010 homicide rate fell to 1.62 per 100,000 population, its lowest level since 1966 ... .

Canada's homicide rate is ridiculously low already (with the sharp drop in 2010): 1.62/100,000; consider the 20-year decline, and whether the decline observed in the US has been proportionately equivalent, even not considering the high starting point and thus more room for improvement in the US:

Homicides, Canada, 1961 to 2010

Were it not for the extent to which gang-related homicides have driven that figure (see the chart at the beginning), that rate would have been more like 1.4/100,000.

It might be interesting to see the US breakdown.


If the presence of guns instantly kicks off an orgy of shootings, beheadings, and unspeakable acts involving 55 gallon drums then why isn't Canada in flames as well?

Well, if pigs could fly, you'd have yourself a question there.

As for Canada, the obvious fact that the numbers of firearms and particularly handguns in criminal hands just doesn't begin to approach the situation in the US or Mexico just might be a contributing factor, but there are obviously a load of others, including, as I mentioned, a strongly functioning civil society with all its defences against lawlessness.