Friday, October 7, 2011

Downplaying the Role of Non-Trafficking Income

Alejandro Hope, in a rather lengthy comment to the Keith Humphreys post I mentioned here, thinks that the low-ball estimates for non-drug organized crime revenue in Mexico are closer to the mark:
The 50%-plus estimate for non-drug revenues of DTO’s is absolutely ludicrous. Let’s go quickly through the potential sources of income: a)human trafficking: according to the latest National Employment and Occupation Survey (ENOE) done by the Mexican equivalent of the Census Bureau (INEGI), 150,000 Mexicans emigrated to (mostly) the US in 2010; out of those, a portion were legal migrants and another portion were visa overstays, i.e., they did not need a trafficker to get to the US. So let’s put at 100,000 the number of illegal Mexican migrants to the US in 2010. Migrants from third countries have always been a fraction of Mexican migrants, but for the sake of argument, let’s say they were the same number. So you have around 200,000 potential clients for human trafficker, which charge about USD 3,000-5,000 a pop (and that includes about 3-4 attempts). That brings the total to somewhere between USD 600 million and US one billion, not all of which is accruing to DTO’s because there are a lot (probably a majority) of independent gangs of coyotes; b) piracy: according to data from the Mexican Film and Music Protection Association (guys that would tend to exaggerate the problem), in 2010 music piracy caused losses of 436 million dollars to the industry. Such potential losses are not equivalent to the pirates’ earnings: they imply that 30 to 40 million pirate CDs are sold every year in Mexico at a price of not more than two dollars each. Therefore, the earnings from CD piracy would stand at levels of 60 to 80 million dollars. The pirate DVD market is predictably smaller and that of clothes and other articles somewhat larger. In any event there is a high probability that the total amount of income produced by piracy is in the range of the low hundreds of millions of dollars (and there are a lot of non-DTO players); c) fuel theft: according to Pemex, about one million barrels of fuel (mostly gasoline) were stolen from its network in the first four months of 2011; extrapolated, that would translate to three million barrels per year. A barrel is equivalent to about 80 liters of gasoline. Thus, some 240 million liters are stolen. In both the US and Mexico, a liter of unleaded gasoline is sold in a gas station for something less than a one dollar, but since the thieves probably have to offer a discount given the illegality. So the total amount accruing from this is likely less than USD 200 million (again, with many independent players); d) kidnapping: according to data from the Ministry of Public Safety (SSP), the average paid ransom in Mexico is approximately USD60,000; in 2010, there were 1264 recorded kidnappings; thus the income from “official” kidnappings would be around USD 72 million. Let’s assume (correctly) that that figure is a wild underestimate and multiply by 10: you get to USD 720 million (with a lot of that money going to specialist kidnapping gangs, not DTO’s); e) extortion: this is the tricky one. No one as far as I know has a minimally reliable estimate of the size of the phenomenon. But according to some anecdotal evidence, the average extorted firm is paying about USD10,000 per year (NB: since extortion is decentralized, the death, imprisonment, or relocation of a specific henchman can lead for a firm to the suspension of payments for a few months or even permanently, so that estimate seems not that wildly off the mark). If that number is more or less correct, you would need about 100,000 firms paying protection money to get to one billion: that would be one in forty Mexican economic units, according to the latest economic census; to get to half of what RAND estimated as drug export revenues, one in thirteen economic units in the country would have to be extorted. That would seem to me pretty much unprecedented in human history (maybe someone has an example). There are, of course, other types of extortion (e.g., phone extortion), but that is mostly the work of con men posing as kidnappers or Zetas or whatever. So, in summary, I would be massively surprised if non-drug revenues of Mexican DTO’s were much more than USD 1-2 billion (and that range could be on the high side), i.e., 15-25% of total revenue at most.
This is easily among the most well articulated breakdowns of drug income that I've read, but I think a couple of points could be made in response. One obvious one is that back-of-the-envelope calculations about hidden industries, even when guided by a logical set of assumptions, lend themselves to rather distorted figures. Of course, that goes both ways: Hope could also be overestimating the total proportion of non-drug revenue. In any event, a few of the assumptions jump out at me as being debatable. The first is immigration: according to the Washington Post, Mexico caught more than 100,000 Central Americans illegally in Mexico in 2010, while US authorities caught 50,000. That doesn't make his figure of 100,000 wrong, since their could be a large number of migrants getting caught multiple times, and it's not clear that the Central Americans caught in Mexico were migrants who had paid coyotes; however, it does suggest that it could be an underestimation. So, for instance, if 200,000 Central Americans are paying Mexican coyotes $4,000 a head instead of 100,000, then we are talking about an extra $400 million. It's also worth mentioning that the UN, using a methodology that I know nothing about, estimates the size of the coyote industry in Mexico $6.6 billion.

As far as kidnapping and extortion, reliable figures are, as Hope mentions, rather hard to come by, because few people involved are interested in reporting their activities. Given that constraint, the key step in Hope's kidnapping figures is this: "Let’s assume (correctly) that that figure is a wild underestimate and multiply by 10". But that seems a pretty casual jump, and if the proper multiplier is 15, we are talking about another several hundred million dollars. (Also, if the average payout is $60,000, does that mean we are not talking about express kidnappings?)

Furthermore, Mexico's home-grown drug market implies a domestic source of organized crime revenue that is largely independent of American policy decisions or the US drug market, so in that sense it is more like non-drug revenue. Conadic estimates that there are 600,000 addicts in Mexico; federal officials say the entire domestic market is worth about a billion dollars.

With just those three changed assumptions (and you could continue tinkering for a while without drifting into the land of fantasy numbers) we've added almost $1.8 billion. Again, my assumptions are probably no more reliable than Hope's (particularly on immigration, other evidence suggests that an aggregate of 200,000 Mexicans and Central Americans paying coyotes may be too high), and what this illustrates most indisputably is the need to take all these statistics with a large grain of salt. However, though I agree that this non-drug revenue isn't going straight to Chapo and his counterparts at the top of the drug-trafficking industry, I think that Hope's statement that he would be "massively surprised" if the proportion were higher than 15-25 percent is a bit too certain.

Finally, check out Hope's new blog; it's full of interesting commentary and is completely worth your time if you speak Spanish. 


Anonymous said...


First of all, thank you very much for taking the time to read and write carefully about my stuff.

I think we pretty much agree that no one knows for sure how much money is accruing to the cartels from non-drug sources. And of course, my assumptions could be very wrong, particularly on the issue of kidnapping and extortion. However, my two larger points were the following:

1. You need to make really heroic assumptions about non-drug income accruing to drug gangs to get to the 6.6 billion that RAND estimates as drug export revenues. For instance, UNODC's estimate of human trafficking revenues would require that somewhere between 1-2 million foreigners crisscross Mexico every year. That population is about the size of Tijuana or Ciudad Juárez or about three to five times the size of Nuevo Laredo. If it were true, border towns would be overrun with foreign migrants at all times; that's clearly not the case.

2. It is a mistake to confuse illegal income with income accruing to the drug gangs. There are a whole lot of gangs out there, specializing in kidnapping or human trafficking or piracy or fuel theft, that are not controlled by the cartels nor pay tribute to them. For instance, most piracy in Mexico City (probably the largest market i Mexico for counterfeit or pirate goods) is in all likelihood not controlled by the cartels, but by an array of local gangs. Or, another example, if you go to Nogales (one of the most active illegal immigration corridors), most of the coyotes tend to be independent players, not employees of the cartels. So, yes indeed, illegal markets could be much larger than my back-of-the envelope estimates, but that does not mean that drug gangs are getting all or even most of that income.

As for the domestic market, I actually wrote a piece in Nexos about that (, showing how ludicrous SSP's numbers are. Using super-wild assumptions, you get to 1.4 billion USD; using more reasonable ones, you get to a range of 300 to 900 million USD. And again, not all of that income is going back to cartels, unless you suppose that all narcomenudistas are slaves that get no income whatsoever from their rather risky profession.

Thank you again and let's keep in touch.

pc said...

Hi Alejandro, thanks for the thoughtful comment. And for the article, I read Nexos regularly but somehow I'd missed it. With regard to point 1, yeah it's hard to see how 6.6 billion is credible. There seems to be a relatively minor flow of non-Central American foreigners using Mexico as a passageway, and they might add some more still to the total, but 6.6 seems like an overshot.

With point 2, I also agree, and I think that's potentially the most interesting development with regard to Mexican crime. I'd say that one of the fundamental changes over the past few years is the blurring of the lines between organized crime and petty crime, though one that I don't think I have been able to quite my brain all the way around. The theories that the Zetas have taken over people-smuggling whole hog are certainly a bit off (although would your point from Nogales also be true in northeastern border towns, or would those guys be paying quotas?), but it certainly seems as though there is a more fluid and interconnected relationship between petty crime and the traditional DTOs today. I'm not quite sure what that means from the standpoint of law enforcement, but it would seem like a more difficult challenge today. Definitely worse from the standpoint of citizen security as well. I'm also not sure what that means in terms of our estimates with regard to the breakdown of the revenue. Are we better off just estimating the size of the industries, with the understanding that this money is now being spread around a larger number of actors/organizations than before, or should we be trying to estimate how much specific gangs make from a given activity?

Anonymous said...

I am coming very late to this, but I wanted to say thank you both for this thoughtful exchange about an interesting topic.
--Keith Humphreys

pc said...

Hi Keith,

Sorry for my tardiness in seeing your note, thanks for the comment.