YT YouTube2Text

YouTube2Text → Transcripts → BREAD-IGC virtual PhD-level course 2026 | Lecture 9 Crime and Organised Crime

BREAD-IGC virtual PhD-level course 2026 | Lecture 9 Crime and Organised Crime — Transcript

by International Growth Centre · 15,862 words · 2,568 segments · language en · Watch on YouTube

Open in transcriber Summarize with AI Download SRT/VTT

Full transcript

0:19Well, hello everyone.
0:21Really warm welcome to you on this
0:23November Tuesday morning.
0:26We are delighted to have Chris and Mika
0:30here to share a little bit on crime and
0:33organized crime.
0:35Um
0:36just by way of introduction, Chris
0:38Blattman is a professor of global
0:41conflict studies at the University of
0:43Chicago and Mika Sviatachi is a
0:47professor of economics and public
0:48affairs at Princeton University. So
0:51we're really happy to have both of them
0:53and they are uh
0:55as you've seen with with our our past
0:57weeks, quite specialized in this field
1:00and so we're really excited to learn
1:01from them.
1:02Hopefully most of you were able to look
1:04at the the reading list so you have a
1:06little uh
1:08kind of background but
1:10um
1:11we'll we'll get some kind of theoretical
1:13grounding and then more specific cases
1:16as well
1:18from from hopefully over the next hour.
1:21And then as you know, we'll go into our
1:22Q&A. So those that would like to be live
1:25panelists and ask questions live, please
1:28just raise your hand and I'll randomly
1:30select 10 to 15. And if you could have
1:32your video on during this and your mic
1:34muted, that would be great. Um I'll also
1:37just quickly mention that
1:40if you missed past lectures or
1:41interested in today's slides or
1:43recordings, those will still be on the
1:45same IGC website and I'll put that in
1:47the chat.
1:49Uh
1:51that being said, I will pass over to
1:53Chris. Thank you.
1:57>> Great, thanks so much.
1:59You can see the slides all right and in
2:01advance though
2:02>> I can't see them.
2:03>> You you can't you cannot see them.
2:05>> No.
2:06>> Let's try that again then.
2:08I'm going to share my whole desktop at
2:09the risk of
2:12>> All right.
2:13>> How does that work?
2:15>> Now we can see them.
2:16>> Okay, great. Okay, perfect.
2:20>> [clears throat]
2:21>> So there's this famous
2:23saying in economics
2:25I think it comes from Lucas. Once you
2:27start thinking about economic growth
2:28it's hard about to think about
2:29anything else and about 25 years ago
2:33someone by accident I found myself
2:35>> [snorts]
2:36>> starting dissertation research or or
2:38actually
2:39even before starting dissertation
2:41research working in an active conflict
2:42zone in Africa and I used to revise
2:44that. I said once you start to see some
2:46of the horrors of violence and just how
2:49you know impossible or it just how
2:51incredibly damaging it is and how how
2:53little work was done 25 years ago it was
2:57hard to think about anything else. And
2:59then I'd say about 10 to 12 years ago I
3:02had the same revelation with respect to
3:05organized crime and and so really the
3:07last 10 or 15 years has been about me
3:09shifting my agenda over towards trying
3:11to understand organized crime and we
3:13could talk in abstract numbers you know
3:16um
3:17this is just a huge sector of the
3:18economy. It's probably there's probably
3:19a million
3:21members of gangs organized criminal
3:23groups in the US alone. We don't know
3:25about the rest of the world. If we had
3:26to think about Latin America if we just
3:27took this as a proportion of population
3:29it could be 6 million.
3:31Um the burden of violence is massive. So
3:36just to put it in stark terms more
3:38people are killed through urban
3:39homicides
3:41every year than in all wars combined.
3:44And and the proportion of those urban
3:46homicides that are related to gangs and
3:48organized crime also probably are large
3:50enough that they exceed all wars
3:51combined. And so this is the largest
3:53effect of
3:55of essentially violence in the world.
3:59And if we think about
4:01the cost of this violence and this is
4:03actually I'm going to try to talk a
4:04little bit of a research frontiers today
4:06and then dive into a few examples of my
4:07own work.
4:08But a huge research frontier and one
4:11that I'm trying to probably not do but
4:14I'll talk a little bit about an
4:15institution that I'm building along with
4:17Mika and others to help support. But
4:19really understanding the cost of
4:20violence is massively important. Uh
4:23because not only is there like that
4:24death which itself is tragic and the and
4:27and and the effects on the family, but
4:29you can think of concentric circles.
4:31There are impacts on
4:33uh the justice system and the cost of
4:36imprisoning people, the cost of
4:37investigations, etc. That's a burden on
4:39society. Uh but the the bigger and more
4:42important circle is what happens to the
4:44economy. So here in Chicago where we
4:46have occasionally had severe violent
4:48spikes in the last decade,
4:50um we had uh Boeing move its
4:52headquarters to Virginia. Uh Amazon HQ2
4:56passed over us. Ken Griffin, one of the
4:59wealthiest people in America and with
5:00one of the wealthiest sort of hedge
5:02funds, Citadel, moved to Miami. And and
5:04and all of these were influenced by
5:07violence and then a million smaller
5:09investment decisions and a million
5:11smaller
5:12uh population and movement decisions
5:15were made. And so what's the
5:16counterfactual Chicago without two
5:19violence spikes since 2016? It's perhaps
5:22a much wealthier, more prosperous place.
5:25And so we could I think magnify that
5:26even further for many many uh many many
5:30cities in the many sorry, many countries
5:32in this hemisphere, especially Latin
5:34America, many parts of Europe. And I
5:36think my guess is organized crime is
5:38coming for for West Africa as well. Uh
5:42and that my fear is that uh as a transit
5:44route for cocaine and as an an area for
5:46local extortion, retail drug markets,
5:49West Africa looks much sadly like
5:51Guayaquil or Rio
5:54uh does today.
5:55Um
5:57And so so to me this is really maybe one
5:59of the most important national security
6:01and development issues
6:03in the world. And really there's just a
6:05tiny handful of economists and
6:07quantitative social scientists studying
6:09it. There's actually a large number of
6:11criminologists and other, you know,
6:12sociologists and things who tackle this,
6:14but I really believe uh quantitative
6:16social scientists, especially
6:17economists, have a lot to bring. There's
6:19not many there, and so this is just a
6:21huge wide-open area for exciting
6:23research.
6:24Um
6:26And so
6:28you know, and you know, I've talked a
6:30little bit about this, but you know, you
6:31can just think of the varieties. It's
6:33not just these sort of cities like
6:34Chicago
6:36or or Cali or uh elsewhere where we have
6:39cities with crystallized criminal gangs.
6:42Uh if you go to Brazil, if you go to uh
6:45Italy, uh if you go to parts of the
6:47West, we've had entire prison systems.
6:50Uh we've had uh entire city public
6:52services systems captured by syndicate
6:55syndicates and cartels.
6:57Um in places like Rio or São Paulo,
6:59Medellín, San Salvador, northern Mexico,
7:02they they challenge the state.
7:04Um we could draw close parallels, as my
7:07colleague Raúl Zibechi Sierra does, to
7:10to armed groups in Africa that are that
7:12are essentially operating
7:14uh like organized criminals uh outside
7:17of the state with territorial control.
7:19Um and then we have an increasing
7:21suddenly number of of narco states,
7:23Guinea-Bissau, Sierra Leone. We can
7:25probably add to this list recently. Uh
7:28many northern Mexico states, many Afghan
7:30provinces. Uh even Venezuela to a large
7:34degree, though not as large a degree as
7:36as perhaps a year ago, but still uh very
7:38high levels of state capture. So, these
7:40are just sort of tremendous sort of
7:41challenges and and really understudied.
7:44And there's I think a huge number of
7:46ways in which we can tackle this.
7:49Uh why these groups arise in some places
7:51and not others, I don't think we have a
7:52good sense. What markets they enter, how
7:55that shapes behavior. Uh I think drug
7:57and extortion markets are are quite
8:00prevalent and and obvious, but why
8:03illicit mining, uh why avocados,
8:07why uh various consumer markets,
8:10uh etc.
8:12Um
8:13you know, which many of these groups are
8:15essentially moving laterally into.
8:18Um
8:18why do so many provide security and
8:20other public goods to civilians?
8:23Um that's probably the best developed as
8:25I'll talk about it in a moment. Um we
8:27know very little about the industrial
8:28organization and the organizational
8:30economics. So, I'm going to spend some
8:31time talking today about that because
8:33that's a passion of mine.
8:35But I think that's a huge area for both
8:37theory and empirics.
8:40I talked about how we don't really
8:41understand the economic and political
8:43costs,
8:44um which is really going to motivate, I
8:46think, public investment in
8:48interventions, right? If we just focus
8:50on the impacts of violence on the
8:52individual deaths
8:53of a bunch of people we consider
8:55criminal, maybe it's not a huge public
8:56policy issue. On the other hand, if it's
8:58the biggest driver of economic
8:59underdevelopment and the biggest
9:02driver of the growth trajectory of
9:04Brazil or Mexico or Nigeria in future,
9:08maybe it is a first-order driver of
9:10development. Uh and then of course, you
9:12know, what kinds of interventions work
9:15and which do not
9:16uh is is a is a huge area. And so, um
9:20and and and you know, as as Miko will
9:22attest, there's maybe a a
9:25maybe maybe 10 or 12 economists in the
9:29top, you know, 100 econ departments who
9:31are focused on this. There's maybe a few
9:33dozen
9:34PhD students. Um and that's it. So, it's
9:37pretty it's pretty nascent and it's
9:39pretty wide open. And so, uh and and so,
9:42I'll talk a bit a little bit about at
9:43the end about how we're trying to crowd
9:44in students into this area.
9:47Probably the best developed um
9:49literature is on criminal governance.
9:52And there's sort of two dimensions of
9:53criminal governance. So, you can sort of
9:55think of, you know, the state governs
9:57society, all right? There there's a huge
9:58political economy literature on that.
10:01Um but a lot of gangs and a lot of local
10:04criminal groups
10:06uh govern civilians as well. So, if we
10:08think about just Latin America, my
10:10colleague Ben Lessing
10:12um based on Latinobarómetro surveys and
10:14and related surveys estimates that one
10:16in 10
10:18Latin American Latin Americans can look
10:20look to a gang to provide at least some
10:22protection or dispute resolution or
10:24related services in their neighborhood.
10:25So, that's massive.
10:27Um and so there's a tremendous
10:28qualitative literature there, but then
10:30there's also, as I'll talk about,
10:32there's criminals' governance of other
10:33criminals. So, in the absence of the
10:35state enforcing property rights or the
10:38and and and and given the need for lots
10:40of collect of of small criminal
10:43organizations to solve collective action
10:44problems, we develop larger, you know,
10:47criminal hierarchies and organizations.
10:49And these until recently have been
10:51mostly qualitative and comparative
10:52literatures in which economists have
10:54only dabbled, and I think there's just a
10:56huge amount of room for
10:58um for for uh
11:01for both theory and empirics here. So,
11:02I'll talk really briefly about a little
11:03bit of work I've done in this area.
11:05Papers I didn't assign, just to breeze
11:07through so you can go and look at them
11:08to see examples of what I think are just
11:11two examples of of efforts to understand
11:13and tackle
11:15uh this civilian-level criminal rule.
11:17And then I'll spend a little bit more
11:18time talking about the macro criminal
11:22rule and what I think are some of the
11:23theoretical and empirical directions.
11:25So, I have one paper um recently
11:28published in in in ReStud where what we
11:32tried to do is we said, "Look, let's
11:34actually We found in national We We I'll
11:37talk a little bit about Medellín,
11:39I'll I'll I'll talk a little bit later.
11:40Medellín is a city It's essentially a in
11:43prosperous [clears throat] industrial
11:44and commercial metropolis of more than 4
11:47million people in Colombia. It's the
11:49birthplace of the international cocaine
11:50trade.
11:52It's mostly now a place where you have
11:54domestic uh basically the criminal
11:56groups, which are very well-organized,
11:58are mostly no longer focused on the
12:00international trade, but are focused on
12:02on local retail drugs and extortion.
12:05Um and they There's a tremendous amount
12:08of gang rule. Basically, there are
12:11I'd say in in about a third of the
12:13neighborhoods of Medellin, when it comes
12:16to security protection dispute
12:18resolution,
12:19um the state is present and the state is
12:21providing a whole bunch of services both
12:23policing and all sorts of civilian
12:25security work. There's a lot of
12:27non-policing security provided
12:29throughout Latin America, especially
12:30Colombia, including dispute resolution.
12:33Uh and
12:35and [clears throat] basically some
12:37people are more proximate to those
12:39police and and non-policing security
12:42services than others.
12:43And we and and how does that influence
12:46the degree to which criminals rule? And
12:48our initial hypothesis was that
12:51uh this criminals enter where the state
12:54is absent and that criminals are filling
12:56a vacuum.
12:57And we stumbled upon what we thought was
13:00a plausible natural experiment. You can
13:01sort of see these internal borders of
13:03Medellin. Well, what happened 30 years
13:05ago or or actually now 35 years ago I
13:08would say is is that the these borders
13:11were introduced in a way that would
13:12organize all of these policing and
13:15non-policing services. And once these
13:17internal borders were introduced, what
13:19it did is it meant that previously, you
13:21know, if you were this block over here,
13:24you might receive some of your services
13:26from from this station. All right, but
13:28as soon as this border was introduced,
13:30it meant that you were in you might
13:31receive your your your service your
13:32security services from one of these
13:34other blocks that's further away. Um
13:37much more stark. You can imagine that if
13:40you're you know, one of these blocks,
13:41previously you would have received
13:43policing services and and other security
13:45services from here. Now they're much
13:46further away. So this introduced these
13:48geographic discontinuities that on
13:50average pushed some blocks that were
13:54almost identical to the block, you know,
13:56across this boundary
13:57about 300 m 400 m further from their
14:01security services. And what we found as
14:04a result of that was not that
14:06the further away you got from the sorry,
14:08the closer you got to your security
14:10services, the fewer the gang rule you
14:12had. It was actually the opposite. We
14:14actually found that the closer you ended
14:15up getting to your security services,
14:18the more that over the next 30 years
14:20gangs responded with increased gang rule
14:22and civilian rule. And why? Uh
14:25qualitatively and in some of our data,
14:26it suggests it's because
14:29the uh because they were trying to
14:30protect their drug markets. More police
14:32around means we need to sort of keep
14:34order in the neighborhood so that they
14:36don't interrupt our drug sales. And so
14:38this really kind of in some ways
14:39overturned some of our expectations of
14:41gang rule.
14:43Around the same time we were running an
14:45intervention, we helped sort of persuade
14:47the city government to run one of the
14:49first um
14:51randomized experiments designed to
14:53increase the quality of state
14:54governance. And so we essentially chose
14:5780 neighborhoods with varying levels of
14:59criminal governance
15:00and we persuaded the city to sort of
15:03devote personnel to 40 of those to
15:05basically govern the heck
15:07out of those neighborhoods with
15:09non-policing
15:10services.
15:12Um and uh and and and and so you're
15:15welcome to go look at this. It's a
15:16recently published in the EPSR
15:19uh as an experiment in in essentially
15:22not what happens when you introduce
15:24non-policing services both to legitimacy
15:26of the state, the quality of state
15:28services, and legitimacy of the combo
15:30and of the of the gangs which are called
15:31combos. And um and we found sort of very
15:35interesting patterns. We found in places
15:37where the state already had some
15:38capacity
15:39and was well-established, generally
15:41people felt state services improved and
15:43the legitimacy of the state improved.
15:45And in places where the state initially
15:47had little capacity and was
15:49underperforming, um the opposite
15:51happened. Generally the state failed to
15:53deliver. People raised their They raised
15:55people's expectations, they disappointed
15:57them, and state legitimacy and
15:59assessments of state performance fell.
16:01Sort of showing how complex it is to
16:03sort of think about how do we actually
16:05do state building in the city.
16:08Um
16:09and and uh and and what the criminal
16:11responses are going to be. So I mean
16:14those are just a couple of examples of
16:15ways in which one can, you know, uh
16:18I mean, experiments on this scale, not
16:20things I would
16:22recommend as a as a PhD student, but
16:25quasi-experiments like this I think are
16:26completely um attainable for PhD
16:29students and and junior professors as
16:31you sort of try to explore this work.
16:34But I actually want to point people a
16:35little bit away from all this literature
16:39on criminal governance. I think there's
16:40lots more to explore, but I think
16:42there's a whole bunch of, you know, the
16:43risk as a PhD student is you end up
16:45chasing the same questions that people
16:47have been asking. And so I want to point
16:49you to questions that I think people
16:50have not been asking very much. And this
16:53has to do with the industrial
16:54organization and the organizational
16:55economics of crime. So how are these
16:57criminal firms
16:59and these and this and and how are
17:01criminal markets structured? And based
17:04on that, what does this suggest about
17:07what policy instruments are going to be
17:08functional or not?
17:10Um and
17:12I gave you guys a reading which is which
17:14is more of an overview that's that's
17:16built for for policy makers who are
17:17trying to think about, you know, what
17:19works and what doesn't. And this whole
17:21idea of what works is in some ways kind
17:23of a nonsense point of view because it's
17:25not like just like there's many diseases
17:28and every treatment has to fit the
17:30disease, there many different types of
17:31organized crime. And so I won't go into
17:33detail here because we don't have time,
17:36but I want to draw a distinction between
17:38cities like New York or Bogota or
17:41Santiago de Chile that have you know,
17:43basically lots of individual criminals.
17:46I want to just draw a distinction
17:47between that and cities like Chicago
17:50or Cali, Colombia or Monterrey, Mexico
17:53or uh
17:55Lagos, Nigeria where you have lots of
17:58sort of freelance groups that are flat,
18:00informal. They're they're conducting
18:02illicit activities but not as a
18:04structured firm with a
18:06uh unlike unlike when they become more
18:08formal gangs where you have a gang
18:10leader who's the residual claimant on
18:12profits and they're hierarchical. These
18:14are flatter and people are engaged in
18:16their extortion or their drug selling or
18:18whatever they're doing more
18:19entrepreneurially in subgroups.
18:22But very seldom do you see a city with
18:24like 400 firms. And that's because the
18:27sort of the competition
18:29uh and I'm going to talk a little bit
18:30about the drivers there's there's a
18:32whole bunch of collective action
18:33problems, not least of which is violent
18:35conflict that leads very quickly to most
18:39cities if they don't look like they're
18:40atomized or a bunch of freelance
18:42cliques, they actually look more like
18:44confederations of gangs. And I'll talk a
18:48little bit about, you know, some of some
18:49of the or at least this is my hypothesis
18:51and something I've been working on. And
18:54only in a handful of cities like New
18:56York under the Cosa Nostra, the the the
18:58Italian Mafia um in in in much of the
19:0120th century or Medellin or Sao Paulo
19:05today, do you see them actually forming
19:07citywide cartel institutions?
19:10Um and and actually where all of the
19:12confederations themselves
19:15form into cartels and and and
19:17essentially try to keep peace amongst
19:18themselves. Basically, trying to solve
19:20the collective action problems through
19:23informal institution building.
19:26Um so so let me tell you just a little
19:28bit about this. Um so
19:31contrary to I think what we learned from
19:34One Gang in Chicago by, you know, my
19:36colleague Steve Levitt and Sudhir
19:38Venkatesh uh where they found a single
19:40Chicago gang where people were
19:41surprisingly very poorly paid at the
19:43bottom level, that's not true.
19:45Um most gangs are very well paid. What
19:47was true about that gang and is true
19:49elsewhere is most criminal organizations
19:50are very small and labor intensive.
19:53Um
19:54and large organizations of the
19:56Sinaloans, the MS, you know, 13
20:01uh Barrio 18 in in El Salvador, the
20:02Maras in El Salvador,
20:04uh the Comando Vermelho in Rio um
20:08you know, the Black Disciples or the
20:10Latin Kings uh or the Bloods or the
20:13Crips in in Los Angeles, they sound like
20:16they're these huge organizations of
20:17thousands of members, but in fact that's
20:20false. These are confederations of many
20:22many many groups that are often only
20:24between 20 and 50 people.
20:26Uh and so they simply look bigger and
20:28more organized than they are. And then
20:30occasionally those confederations will
20:33form coalitions or governing boards, be
20:36that, you know, the Mafia Commissions in
20:38the United States or the peace treaties
20:41that the Mara factions formed in in El
20:44Salvador or something called La Oficina
20:47in Medellín or uh or the PCC in São
20:51Paulo. Occasionally they form coalition
20:53governing boards to sort of solve their
20:54collective action problems.
20:57Um so this is true of all sorts of
21:00criminal organizations. So like there's
21:01some famous work by Annelise Anderson,
21:04who's an economist writing in the 1970s,
21:06or Peter uh Reuter, who was another
21:08economist writing in the 1970s, who were
21:10looking at these groups uh almost
21:13ethnographically.
21:15And they would find even in a single
21:17Mafia family, whether it's in the New
21:20York, which is where Peter Reuter was
21:21studying, or a a Midwestern city, which
21:24uh was not named but I my guess best
21:26guess is Philadelphia, is like even a a
21:28Mafia family itself was a small
21:30confederation, that each capo, each
21:32chief with their own little mini
21:34business in their part of the city, was
21:36a residual claimant on profits. And
21:38there was no transaction really with the
21:40boss or the underboss or that Mafia
21:43core. What the boss or the underboss was
21:45doing was essentially
21:47basically carving up the city for these
21:49different capos to have different
21:51uh uh areas either defined by territory
21:54or defined by
21:56uh
21:56market. And there might be some tribute
22:00um or they might sell services and
22:02transact on price. But therefore, you
22:04know, if we think of the definition of a
22:06of of a firm as organizations that don't
22:08transact on price, then then this is not
22:11an entire firm. A Mafia family is in
22:13some sense a confederation where this
22:15small little group is essentially
22:16providing conflict management, market
22:19collusion by division of territory and
22:21markets, and various club goods. So the
22:23boss buys off uh you know,
22:27but pays for the lawyers or buys off the
22:28judges or buys off the police and
22:31provides those services in return for
22:34tribute or other, uh, or status or
22:36other, you know, pecuniary non-pecuniary
22:38factors. Okay, so that's a a great
22:40example of a confederation at the
22:42smallest level within a city. Um, this
22:44was also true of Chicago. So, the
22:47Almighty Vice Lord Nation, which was one
22:49of the sort of the six gang nations that
22:51grew to sort of expand across Chicago,
22:54uh, again, each each local area had its
22:57own set,
22:59um, and then there were many different
23:02branches of the Vice Lords,
23:05um, but each set was essentially its own
23:07residual claimant on profits. It had its
23:09own supply chain, uh, it had its own
23:12connections, and the role of the
23:15branches and the role of the or or the
23:18overall councils were not to be the
23:20suppliers, were not to There was no
23:22transacting based on price. There was
23:23very little tribute or any other money
23:25through that made its way up. Um,
23:27basically they were there over there to
23:28solve collective action problems amongst
23:31the different sets.
23:34Okay, and so basically, so why do we
23:37tend to see this appearing not just in
23:39Chicago? You know, I could go through
23:41examples throughout Latin America as
23:43well. Uh, we've recently seen, you know,
23:46the Sinaloa Cartel, you know, this
23:47people think of the Sinaloa Cartel as,
23:50uh, whoops, as a as a as a as like a as
23:53a uniform whole with its current civil
23:56war between
23:58you know, the Mayo faction, the the the
24:00El Chapo faction, you know, we can see
24:03now some of the some of the the
24:05breakdown, uh, and the the
24:06confederation, the conflict management
24:08that was happening before and is and is
24:10failing today. So, the basic fact Why do
24:13we see lots Why are Why are these firms
24:15tend to be very small? Why do most of
24:16these little sets or
24:18or factions or whatnot, why are they
24:20very seldom larger than 30 or 50 people?
24:23It's because basically there's
24:24diseconomies of scale. There's special
24:26diseconomies of scale that are coming
24:28from state predation.
24:30And so more workers mean more potential
24:32informants. Um you can't write things
24:34down, so you don't have basic, you know,
24:36management technologies such as
24:39uh as as as record keeping and finances
24:43and things. So this is limiting what's
24:44going on. Nobody's enforcing your
24:46property rights. Um violence and the
24:50ability to use violence is extremely
24:52costly to you and your organization. And
24:54then there are all these potential
24:55benefits from from collusion.
24:58I'm going to jump through cuz I know I'm
25:00probably only have like five or 10 more
25:01minutes.
25:02Um
25:04And so what you tend to see is the
25:05formation of
25:08uh
25:09these sort of criminal confederations
25:12uh within the US you see some of the the
25:14formation of things like within cities
25:16you see the Latin Kings, the the the
25:19the um
25:21the the Vice Lords I gave an example of
25:24the different mafia families are all
25:25different examples, but then they sort
25:26they they tend to form super
25:28confederations. And so around, you know,
25:30you you sort of see failed attempts in
25:33the United States to form super
25:34confederations over time.
25:36Um
25:38There are successful examples such as
25:41for a long time the Sinaloa Cartel for a
25:43period of time before the crackdown in
25:46El Salvador
25:47uh
25:48successful sort of attempts to sort of
25:50create peace and governance and bring
25:52down the violence rate. Another example
25:54is Medellin where I spent a lot of time
25:56working I mentioned before where you
25:58have more than 400 street gangs called
26:00combos organized under about 18 19
26:05mafia-like organizations called
26:06Personas. They're organized into two
26:09alliances.
26:11Um and then atop themselves they sort of
26:13govern themselves
26:14by uh
26:16having created this institution called
26:17La Oficina. And so here you can sort of
26:20see the different alliances in purple
26:22and blue are the two main alliances.
26:25Um and [clears throat] why are they and
26:28and they managed to sort of go from
26:30being the most violent city in the world
26:32with you know, 395 deaths per 100,000
26:35which made it maybe the most violent
26:37place in all of human history. If we're
26:39actually if I had sent this to 2024,
26:41they're actually down to nine this year.
26:43So, they're only slightly more violent
26:45than New York despite having one of the
26:47most hierarchical forms of organized
26:49crime in the world.
26:51Um
26:52why are they doing so? Well,
26:54[clears throat] essentially, you know,
26:56firms these criminal firms are facing
26:57three coordination problems. So, given
26:59that the state is keeping you small,
27:01this equilibrium where you have hundreds
27:03and hundreds of like capos in New York
27:05or hundreds and hundreds of modas in El
27:07Salvador or hundreds and hundreds of
27:09combos in Medellin is unstable. Um
27:12because they want they want to be able
27:14to collude
27:16uh in order to get monopolistic prices.
27:18They don't want to kill one another
27:19because that's extremely costly to them.
27:21And then there's all sorts of club goods
27:23that that could be provided if they
27:25cooperate like buying off police and
27:27officials or getting access to lawyers
27:29or favorable prison treatment or access
27:31to bulk purchasing and wholesale supply.
27:34And so, these confederations can't grow
27:37and become super vertically integrated
27:39firms because then they have to start
27:40writing things down or then they have to
27:43they become a they they become easier to
27:45dismantle. And so, you have to find a
27:47way against state predation to solve
27:49these coordination problems and they do
27:51so via these loose confederations.
27:57Um and so, you know, one of the you
27:59know, areas I and some others have been
28:01pushing the sort of like okay, this is
28:04very different than our sort of standard
28:06cartel theory and our standard IO theory
28:09and our standard theory of the firm
28:11where we have imperfect information
28:14about where we're trying to sort of
28:16price collude amidst imperfect
28:19information and an ability to sort of an
28:22inability to know where the demand or
28:23supply shocks are coming and why whether
28:25or not our adversaries are defecting.
28:28This is actually probably not true. We
28:30are in a repeated game scenario sort of
28:32with our other firms um but
28:36but but the fact is is that there's
28:38there's not just a
28:39a market collusion and price collusion
28:41problem to solve, but there's also
28:43conflict management and a club good
28:45problem to solve.
28:47Uh against the backdrop of these
28:50diminishing returns to scale. And so
28:53this is what sort of creates a a
28:56a role for for essentially criminal
28:59institutions akin to cartel institutions
29:01in the legal cartel literature that are
29:03trying to either create facilitating
29:05devices or or or ways to sort of
29:07monitor, punish defection, and stabilize
29:09prices in a way that doesn't invite
29:11state attention.
29:13Um and so we're sort of looking at
29:16different ways to sort of you know, what
29:18are you know, how do we actually take
29:20traditional IO theory
29:23um
29:24about legal cartels and how do we move
29:27that to understand this new criminal
29:29market? And I think that there's
29:30actually a lot of different case studies
29:33and a lot a lot of both theoretical and
29:34empirical traction that could be
29:37could be driven here. So I think this is
29:38just one example of a frontier that
29:40we're driving along and I think I sent
29:42you guys a
29:43uh uh uh
29:44an example of a of a paper that we're
29:46currently working on.
29:47Do I have about 5 minutes left or are we
29:50wrapping up?
29:51>> I think we just hit time, but if Mika's
29:54flexible then it's it's really the hour
29:56is is your guys's.
29:57>> Yes, no problem. No problem.
29:58>> Okay, well let me just really briefly
30:00just talk about you know, uh so
30:02naturally I have too many slides and
30:04talking fast
30:05uh didn't quite solve the problem. So I
30:07I'll just sort of say for I think
30:09another I think basically a a really
30:11important direction to go is on
30:13recruitment on understanding criminal
30:15business lines other than including some
30:17of the obvious ones and some of the
30:19non-obvious ones and internal
30:20organization of criminal firm. So I sent
30:22you some slides
30:23um
30:25but I would say like I think there's
30:26just a tremendous amount of opportunity
30:28to understand how extortion works and
30:31how to counter it, how retail drug
30:32markets work and how how it. Um why they
30:36enter consumer markets, there's a
30:38tremendous recent
30:40job market paper called on green gold,
30:42which is the avocado markets and why the
30:44Sinaloa cartel entered avocado markets.
30:46There's a lot of promising work
30:48happening on illicit gold mining and
30:49other types of mining and so forth. So
30:51there's all sorts of ways I think to
30:52understand these rents.
30:54Um as well as sort of internal
30:56organization incentives and recruitment.
30:59And so I talked a little bit about um
31:01Medellin uh
31:04and I if we had more time I could tell
31:06you about the internal structure of how
31:09uh gangs are organized. But what one of
31:12the things we're very much interested in
31:13and I know Mika's interested in is like
31:15what is the recruitment pathway look
31:17into these organizations and and could
31:19you actually tackle labor supply? And so
31:22I I won't go into detail, but I sent you
31:25a um
31:27you know, I I essentially we don't just
31:28we know a little we don't we just don't
31:30know that much about criminal
31:31occupational choice. Mika has some
31:32tremendous work which you might talk
31:34about on how early dropout and early
31:36criminal involvement have have are
31:39likely to be sticky. So the question is
31:41is like but why are people dropping out?
31:42Why are they getting involved? So
31:44there's a bunch of pre-existing social
31:46behavioral risk factors that that
31:48sociologists um emphasize, but I think
31:50as economists we can bring occupational
31:52choice. We can sort of understand what
31:54are adolescent beliefs about legal
31:56versus criminal careers. What
31:58information asymmetries, misperceptions,
31:59biases do they possess? And so I sent
32:02you some slides where I talked about how
32:03we've been following a panel of of more
32:06than 10,000 youth where we're trying to
32:09understand basically we we present them
32:12with different career paths that are
32:13low-skill, medium-skill, high-skill, and
32:15criminal and we try to get them we
32:17elicit their beliefs about pecuniary
32:20non-pecuniary benefits and also, you
32:23know, we sort of ask them to think about
32:24themselves as sort of 10 10 versions of
32:26themselves when they're 25 and where
32:29they're going to end up and why.
32:31Um and we we we sort of we find that you
32:33know, in Medellin, you know, depending
32:35on some of the
32:37uh scenarios we present to them,
32:40kids put basically a 5 to 15%
32:42probability of entering these criminal
32:44careers, which is uh,
32:47which is consistent with the number that
32:48actually enter over time. And then what
32:51we've begun to do is sort of look at the
32:53extent to which
32:55non-pecuniary factors like perceived
32:57status, enjoyment, success with women,
33:00stress, the risk of death, the risk of
33:02incarceration actually
33:05at 13 or 14 influence the degree to
33:08which they put, you know, their future
33:10selves in one of these different
33:12categories.
33:14Um, we're beginning to estimate
33:16structural models. So, I think there's
33:18an opportunity to sort of understand uh,
33:21how these different um,
33:23whether it's disutility of schooling,
33:24perceptions about pecuniary returns, or
33:26all these non-pecuniary benefits shape.
33:29Uh, we can begin to estimate
33:30elasticities
33:32of about how much these different types
33:34of pecuniary non-pecuniary factors
33:36matter for the choice between these
33:38different careers.
33:40Um, and so forth. So, that's just sort
33:42of a quick overview of I think some of
33:44the one direction one could go on
33:46recruitment, but then I think there's a
33:48whole other set of directions where
33:49we're taking is in what types of
33:50interventions might change these
33:52beliefs, might change some of the
33:54behavioral factors or misperceptions
33:56that I think is is wide open. I have
33:58more than one PhD student who's working
34:00on that.
34:02Um, so anyways, I just want to end just
34:04sort of pointing out how many research
34:07questions remain unanswered that I would
34:09encourage anyone who wants to work on
34:11this not to sort of pick up an existing
34:12research question and try to marginally
34:14improve it, but actually to blast into
34:16one of these new areas. And to
34:18facilitate this, Mika uh, Rodrigo
34:21Canales, who's a management sociologist
34:23Sancho Lythgoe at EAFIT, uh, Ben Lessing
34:27here at UChicago, myself, uh, Joana
34:29Montero at um,
34:32uh, uh,
34:33in uh, uh, in Rio, are basically
34:36creating a new network for research on
34:37organized crime. And this is going to be
34:39an intellectual network of economists,
34:41political scientists, criminologists who
34:43are working globally to understand these
34:45questions. We're building field labs
34:47across right now we have them in Mexico,
34:50Medellin, Rio. We're hoping to establish
34:53them elsewhere in Latin America as well
34:56as in starting to
34:58push them in in Africa potentially in
35:00Nigeria. Um starting this year we're
35:02going to be launching soon a graduate
35:04student fellowship. So we'll be taking
35:0610 fellows per year. Uh which will be
35:09essentially an advisory and mentorship
35:11network by
35:12this group of scholars plus other
35:14scholars working on organized crime plus
35:16dissertation research funding so that
35:18those who are interested in this can
35:20actually we can actually build this
35:21pipeline. And so this mentorship network
35:24um which is sort of going to be led by
35:26Mika and I are basically
35:28trying to sort of turn this agenda into
35:30action saying we realize this is hard.
35:32This is a new area. This is a
35:34challenging area to do research. It
35:36needs guidance and it needs resources.
35:39And so you should be looking in the
35:40coming uh weeks or months for an
35:43official announcement of this of this
35:46fellowship and this mentorship network
35:47that we're going to be running for PhD
35:49students.
35:59>> Okay. I wish uh thank you Caleb you're
36:02there.
36:04>> Yep. It's over to you. Thanks.
36:06>> Okay.
36:07Thank you Chris. Yes, please I hope that
36:09everyone got very excited about this
36:11network that we are creating with Chris
36:13and others. So please reach out stay
36:15they are looking to the calls and
36:17everything because we are super excited
36:19to be building a greater network. As
36:21Chris said, a lot of the things that we
36:23have in styling there are a lot of
36:24research gaps and I think would be very
36:27interesting for all of you. So let me
36:29start. The nice thing is that a lot of
36:31things
36:32this topic doesn't need motivation
36:35because Chris motivated very well. Let
36:37me there.
36:39Okay.
36:41Let me see.
36:42Can you see my screen?
36:45>> Yeah, we can.
36:47>> But I want to put it here. Let's hope
36:49that works.
36:52It's black in my computer, so I don't
36:54know if you're seeing it. Are you?
36:57Yes?
37:00Can you see well?
37:01>> Um the format's a little off. If you go
37:04back to how it was, we could see the
37:05whole
37:05>> Because I did uh the full screen.
37:09Okay, let me see.
37:11There?
37:12Is better like this?
37:14If I do it like this, is better?
37:16I guess.
37:17>> Yeah, we can see more of the screen.
37:18Yeah, this is great.
37:19>> okay, perfect. It's not full screen, but
37:21okay, perfect. So
37:22>> You're you're showing your age now.
37:24>> What?
37:25>> I said you're showing your age.
37:27>> Uh why? Why I'm showing my
37:29>> So I'm just saying because we're we
37:30become less and less technically
37:31proficient. I can speak for myself.
37:32>> Uh I know. I know, but I have always
37:34been like this.
37:36Very bad. Okay. So So I tried to It's
37:40funny because when I was preparing this
37:41this slide, I tried to put like a
37:43two days classes of two hours, like four
37:46hours class in 30 minutes because I was
37:48very excited. So but the motivation is
37:50very similar as what Chris said.
37:52Something that we have been observing.
37:53I'm from Argentina in Latin America, and
37:56I have worked before the PhD years in
38:00the Inter-American Development Bank, so
38:01I visited a lot of these countries in
38:03the social sector. And something that it
38:05really called attention was it's very
38:08common to see armed groups, criminal
38:10groups to take control over this state
38:12territory. And this is actually I can
38:14give you some numbers about this. In the
38:15case of Mexico, it is said that 20% of
38:18the territory is under the control of
38:19drug trafficking groups. 60% can be in
38:22Rio de Janeiro, Brazil. And also in El
38:24Salvador, where I have worked a lot. In
38:252015, there were 50% of the
38:27municipalities that have gang presence.
38:30So it's not only as Chris was saying,
38:31there is a lot of um
38:33of violence that they create, but even
38:35when they have the monopoly of certain
38:37territories or they get organized, they
38:39can have other consequence entering into
38:41new businesses, even in consumer markets
38:44as they as he was mentioning with the
38:46avocado market, [clears throat]
38:47but also into extortion, human
38:49trafficking, recruitment.
38:51So, big things that they they are very a
38:54lot of distortion economic distortions
38:55created by these economic by these
38:58criminal organizations. So, let me skip
39:00the following slides because I think
39:01this is something that Chris mentioned.
39:04But the question is that when I started
39:06working on this, which was actually when
39:07I was doing the PhD in 2013 at Columbia,
39:11the first thing I was like very curious
39:13to understand was more from a micro
39:15perspective. First like, how do they
39:17search? But in particular, why is that
39:19kids start becoming part of criminal
39:22organization? Something that I was
39:23always observing is that some common
39:26aspect of these criminal organizations
39:29is that the members is start at a very
39:31young age. So, this was something that
39:33was really calling my attention about
39:35how like common this was across many
39:37different organizations, whether it was,
39:39you know, the maras in in San Salvador
39:42to also drug trafficking groups in in
39:45Peru in rural areas.
39:46And then the other thing that also
39:48called the attention was that
39:50even in places, and this is something
39:52that also Chris mentioned, in places
39:53where you uh have also So, the main
39:56consequence that always the newspapers
39:58and media and everyone knows is the
40:00effects that they have in violence, in
40:02homicides. These criminal groups fight
40:04over territory, they compete, they also
40:06force
40:07contracts in violence, and this has a
40:09lot of negative consequences
40:11for the population. But beyond that,
40:13even when they get the monopoly, there
40:15are other consequences that go beyond
40:17violence. And that's what I we are going
40:18to talk a bit about with the paper that
40:20I assigned you on labor mobility. And if
40:22we have time, just to talk about
40:24potential ways to break the cycle.
40:27So, but going when you start the working
40:30on this,
40:31why do criminal organizations emerge in
40:32some places but not others? Most of the
40:34literature that we see in economics um
40:38and others comes from Italy. And
40:41actually it's related to also something
40:43[clears throat] that we resonate what
40:45Chris was mentioning. The first part is
40:47that
40:48uh there is some some for weak
40:49institutions create markets for
40:51protections. Basically, when the state
40:53cannot uh
40:54reliably enforce property rights and
40:56contracts, then you have these uh
40:58private protection from criminal
40:59organizations that become valuable. And
41:01this there is evidence on in Italy.
41:04The same case is for economic
41:05opportunities that may increase also the
41:08returns for these organized violent
41:10violence. And also political conflict,
41:13there is a very nice paper
41:15by Acemoglu and Dell and Dell and Dell
41:17which basically shows when there is a
41:19conflict and elites may rely on criminal
41:21organizations to defend their economic
41:23and political interests, so they will
41:25demand. So, most of this is from the
41:27magic mafia origins, but again, we don't
41:29know much as Chris said about like why
41:32do they appear in certain places beyond
41:35uh some few cases. And also, there is
41:37this other thing which is that most of
41:39these explanations focus on the demand
41:41for organized crime.
41:43Economic opportunities, weak
41:44institutions, political conflict that
41:46create demand for these criminal
41:48organizations. But there is this other
41:50aspect that I'm a bit obsessed about who
41:52supplies this labor needed to sustain
41:54this criminal organization. As you will
41:55see, it's mostly children that start.
41:59So,
42:00uh the first thing I was trying to
42:02understand is comes
42:04with economic incentives.
42:06Similar as um
42:08uh as other legal businesses or formal
42:11opportunities in in labor markets, also
42:15the returns in the illegal sector
42:18also can push kids into a criminal
42:21career in criminal organization. And
42:23this is what I was trying to uh show
42:26with the case study that I sent you in
42:27Peru.
42:28In Peru is uh
42:30in Latin America, you have only three
42:32countries that basically produce coca,
42:34which is this input to transform into
42:36cocaine. And these are Colombia, Peru,
42:39Bolivia.
42:41And the idea here in this paper is what
42:43I was trying to understand is what
42:44happens when the return to this uh to
42:46the returns to uh any legal
42:50uh in the sector increase, what happens
42:53with these kids?
42:54And basically, this is difficult to know
42:56because as we were mentioning before,
42:58one of the main difficulties of studying
42:59this question is
43:00first, how do I know which who are
43:03involved in the business? In this case,
43:05was the cocaine industry as I call it in
43:07the paper.
43:08It's very difficult to know, no? Because
43:10these are like you can use So, what I
43:12was what we can use as proxy here was,
43:14okay, I have administrative records on
43:16incarcerations. I also knew school
43:19records where kids schools were located,
43:21geocoded.
43:22And also, the nice thing with coca is
43:24that you have satellite images where you
43:26can um
43:27see where it's produced. And a lot of
43:28the labs that transform coca into
43:30cocaine paste are located in the same
43:32places where there is coca production.
43:34And coca production is in places where
43:36it is suitable. So, it's In this way,
43:38you can predict where this illegal
43:39sector can boom or not. So, the first
43:42thing that I do is try to see what
43:43happens when you have an increase in the
43:45prices of coca that is driven by some
43:48exogenous variation that comes from
43:50Colombia, from the war on drugs in
43:51Colombia, that push production to Peru,
43:54and then this return to
43:56um
43:57produce coca increase. So, the coca
43:58prices I can see had increased. And then
44:00I try to see what happens with kids that
44:02grew up in those areas suitable for coca
44:04production.
44:05And the first thing that I see in the
44:07data is that the kids start working in
44:09coca. Of course, this is again, it can
44:11be like any legal sector, the returns of
44:14of cultivating coca increase. So, more
44:17kids get involved.
44:19But what it is very interesting, in
44:20particular in the case in Peru,
44:22is that
44:23while the returns to any
44:25legal sector in of of production in any
44:27legal sector
44:29will increase dropout, what you have in
44:31the case of coca that is produced to for
44:35the cocaine industry is that the kids
44:37not only drop out from school just to be
44:39and have lower human capital, but also
44:42they drop out from school and they enter
44:44this illegal business. That goes beyond
44:46collecting the coca leaves.
44:48And this how I got I could prove this,
44:50basically, I looked these incarceration
44:52records and I try see I start seeing,
44:54"Okay, what happens with the kids at
44:55different ages that grew up, that were
44:58in these suitable coca areas, and what
45:01is the probability of being incarcerated
45:03for um
45:04for crimes related to the drug
45:06industry?"
45:08And what I observe is that especially uh
45:11if you were in your early teens, your
45:13probability of being incarcerated is
45:15much larger than if you were younger and
45:18if you were of the same ages growing in
45:21places that are not suitable for coca.
45:24So, and these are the ages exactly where
45:26kids drop out. So, and in the paper I
45:28tried to argue that what happens with
45:29these kids is that yes, economic
45:31incentives matter, but the thing is that
45:33once you enter, even if it is the early
45:35stages of production, is you have
45:38contact with other uh stations. For
45:40instance, the transformation of cocaine
45:41into cocaine. And this is because once
45:43the kids were collecting the coca
45:45leaves, then they can go to these
45:46maceration pits where they transform and
45:48have access to these other stations of
45:49production, also networks, and
45:53and that's how they start getting
45:54involved.
45:55So, two things you have to take out from
45:56this uh paper. First is that organized
45:59crime feeds from uh children and
46:01adolescents, and that once they enter,
46:03it's very difficult in certain way to
46:05leave because they have a a lot of
46:07information. That's why when we are
46:08adults, we see that they are more likely
46:10to be incarcerated.
46:12And the second thing is that most of
46:14like in this pattern you can see that
46:15most of the effects are concentrated in
46:18drug-related crimes, drug trafficking,
46:20and violent part.
46:21And another interesting fact, which in
46:23general we tend to think that women are
46:25not involved in into the in these
46:27criminal organizations, actually they
46:29are. But in general are not in the
46:31violent part, but yes in other parts.
46:34And here what we tried to do is just to
46:35divide and show you that
46:37what is what happens is actually the
46:40legal sector that is generating this
46:41criminal skill. It's not like any shock
46:43to
46:44commodities in legal sector that would
46:46create this. And actually there is
46:47something nice in Peru that actually
46:49there is an illegal sector for coca
46:51because uh traditionally Peru it had it
46:54was used to drink tea for medicine
46:56purpose. So, we look what happened in
46:58those areas and you see that in those
47:00areas where coca the coca leaf collected
47:02doesn't go for this illegal sector, most
47:04of the effects mitigate.
47:06So, economic incentives matter, key ages
47:09of exposure.
47:10And there are also other examples that
47:12are known from my own paper that
47:14which which is one in Mexico that also
47:16shows how these labor market conditions
47:19matter, but not only in the illegal
47:21sector, but also in the illegal sector.
47:23And here this is a very nice paper from
47:26Dale Feigenbaum and Teshima, which
47:28actually tried to see what happens in
47:30Mexico in places that were affected by
47:34Chinese exports when they increase
47:36basically this reduce the jobs in
47:38Mexican manufacturing industry. So, in
47:42these places where you have this
47:43industry and also you already have
47:45criminal organizations, what happens
47:48in terms of drug-related violence when a
47:49lot of people lose their job in this
47:51manufacturing business.
47:54And what they find is actually that the
47:56the job losses due to this Chinese trade
47:58competition actually increase
48:00drug-related homicides. And the effects
48:02are mostly stronger in these places that
48:04already have drug trafficking
48:05organizations.
48:07And they also find some measure that
48:09drug trafficking activity increase they
48:11use as a proxy cocaine seizures.
48:14And the mechanism here is very similar
48:16to
48:17an opportunity cost mechanism.
48:18Basically, the job loss now lower the
48:20opportunity cost of you of enter
48:22criminal employment and now, you know,
48:24drug trafficking which is in constant
48:26fighting in this location is more
48:28attractive.
48:29>> [clears throat]
48:29>> And that's why we see this increase in
48:31homicides.
48:33So, this is another example about how
48:34the returns in local labor markets
48:37matter.
48:38Actually, I should have put also another
48:40nice
48:41paper of a job market candidate that is
48:43a student of mine, Maria Montoya, which
48:45she also studies how the return
48:48to synthetic drugs, fentanyl, actually
48:51when they increase also this demand
48:54young people. But what is interesting
48:56that case is that the demand the returns
48:58to a skill since it is a synthetic
49:01drugs, the demand for a chemist
49:03increase. So, this actually more kids
49:06stay in school and then they enter.
49:09And there is more evidence about how
49:10labor markets matter. Also in Colombia
49:13machine they show that when formal
49:16employment a reduction in formal
49:18employment benefits actually is linked
49:20with an increase in arrest for crimes
49:22linked to criminal organizations.
49:25And another factor about remember I was
49:27studying, okay, how do they discriminate
49:28organizations expand that we don't know
49:30much? We have some evidence actually for
49:33the case of El Salvador, which I have
49:35worked on others,
49:37which is that the movement of criminal
49:39leaders can also bring crime to new
49:41areas. And this is something that I was
49:43also a bit obsessed when I was visiting
49:45these drug trafficking areas in Peru,
49:48which was okay, once the kids acquire
49:50these skills and once they move, do they
49:52use it? But of course they are the
49:54movement is endogenous, like who moves
49:56to other locations. So, but here in the
49:58case of El Salvador we have a very nice
50:00natural experiment, which was basically
50:02that
50:03>> [clears throat]
50:03>> in 1996 there was a deportation policy
50:06in the US that moved criminal leaders
50:09who grew up in the US to El Salvador.
50:12So, and I know which were the places
50:13where they were born, so I can look into
50:16those places because I know that those
50:17are the places that is a high
50:19correlation in the places where they
50:20will go where they were born, and I
50:22tried to see what happens before the
50:24policy 1996 in terms of homicides,
50:27extortion, and also drug-related
50:30crimes. And you can see that most of
50:32these and most of these homicides
50:34increased, and this is mainly because
50:36they actually they deported leaders for
50:38of
50:39to rival gangs.
50:41And
50:42this also brought extortion into the
50:44country.
50:45And again similar as we saw in the case
50:48of Peru, how is the case that they
50:50expanded so fast? Because basically you
50:52see that there were mostly there were
50:54the leaders were about like 500, but
50:57over time when you start looking at
50:58incarceration rates of these
51:01gangs with US origin in El Salvador, you
51:03start seeing that they expand. And by
51:05data of the El Nacional Police, by 1987,
51:08there were 20,000 people that joined.
51:10So, one question is like, how did they
51:11expand in this 500 so fast?
51:14And again, similar as the case of Peru,
51:15they did because by recruiting children.
51:18Early exposure again is a determinant.
51:20Children seem to be particularly
51:22affected by these criminal leaders that
51:24they arrive.
51:25>> [clears throat]
51:25>> And we observe a similar pattern that if
51:28you were younger, and you were in
51:29schooling ages and younger than 16 by
51:32the time that these leaders arrive, your
51:33probability of being incarcerated for
51:36gang-related crimes, which were gangs
51:39with an explicit name of the gangs that
51:41were formed in the US, you see that they
51:43this probability increases by a lot.
51:46And again here, something that is very
51:48interesting, and there is a lot of
51:49qualitative evidence, different than the
51:51case of Peru where they used economic
51:53incentives, what they were mostly using
51:55was a
51:57social capital. Basically, they were
51:58providing a social status to these kids.
52:01If you know a bit about the history of
52:02El Salvador, in 1980, they went through
52:05a long civil war, so a lot of kids were
52:07not having parents, so they saw in these
52:09gang leaders like a family. So, they
52:10also used that.
52:12>> [clears throat]
52:12>> So again, the takeaway here is that
52:15exposure during childhood matters.
52:19There is something I'm going to skip
52:20because I think I have very little time.
52:22Probably I have 15 minutes, but
52:25and this I think that we need a bit much
52:27more of research as well, which is the
52:29role of prisons. When you look at all
52:31the criminal organizations, at least in
52:32Latin America, a lot of them of the big
52:35ones were created in prisons. And the
52:37story is kind of similar is kind of
52:39similar of like the a bit of like the
52:41mafia, like in these prisons there was
52:43no control, these criminal organizations
52:45provided control, they acquired some
52:47criminal governance that then expanded
52:49to different neighborhoods.
52:51So,
52:52one question we wanted to understand is
52:53how
52:55how these leaders, once you get into a
52:58prison, and if you are next to someone
53:00that was a leader of the gang, what
53:02which is your probability that if you
53:03enter there and you didn't have any
53:05relation to gangs, what is the chances
53:07that you will end up being part of these
53:09gangs? And actually
53:10there is very little research on this
53:12about how they expand at least
53:15coastal. There is a lot of qualitative
53:17work. Yes, for sure, but we don't know
53:19much how is that the the system works.
53:22And this has to do as I said is because
53:24most of the prisons in Latin America are
53:25overcrowded and terrible conditions. So
53:27they would mix people that have low
53:30offenses with also criminal leaders. And
53:32this is what we try to do here. We
53:34exploited uh transfers of gang leaders
53:36that were in maximum security in El
53:38Salvador to normal prisons and we try to
53:41see what happens to those inmates once
53:43they were exposed to these leaders. And
53:44we see that they are more likely to
53:47um
53:48commit again a crime and
53:50um
53:50when they were more when they were
53:52exposed to these leaders.
53:54Again, there are a lot of research
53:56questions and gaps. I think that we
53:58don't know still like I I mentioned
54:00economic incentives a bit of social
54:02status, but we don't know very well how
54:05this recruitment is is shaped by
54:09these expectations about future labor
54:10market returns. It's something that
54:11Chris mentioned by the end. Also how
54:14identity, status, motives, the peer
54:15effect, and also even like how
54:18interventions can, you know, reduce
54:20recruitment that target automatic
54:22behavior, the CBTs. I will present you
54:24if we have time one example, but these
54:26are mostly individual level
54:27interventions, but there is no link
54:29about how these individual level
54:31interventions can have an aggregate
54:32effect at the community level.
54:36Uh okay, this is about the just a little
54:39bit just to for you to get a flavor of
54:41what I have been working in terms of
54:43like kids joining gangs. In terms of
54:45consequences, I'm going to focus on this
54:47paper that you also I send you to read,
54:50which is the one on gangs labor
54:52mobility, that uses the exactly kind of
54:55natural experiment that I mentioned in
54:57the other paper on deportation. But the
54:59main point we wanted to see here was
55:02okay, what happens in territories once
55:05um
55:06once they have defined that you know the
55:07territories of control, this is El
55:09Salvador how it was uh before um
55:12in like this was we measured this in
55:142015
55:16um and the blue lines are the blue areas
55:18are the ones that were the territories
55:20controlled by gang. As you can see, a
55:22large uh
55:23share of the territory of El Salvador in
55:252015 was controlled by terri- by
55:28criminal organization by gangs. So, we
55:30want to know, okay, these were the areas
55:33that were in a stable uh territories,
55:36what happens to individuals that live
55:38inside versus individuals that just live
55:39outside? Of course, we know that these
55:41gangs have presence in many places, but
55:43these are the territories that they
55:44control, that they actually decide who
55:46enters or not. So, we try to do
55:49we exploit this setting to do a spatial
55:51regression discontinuity design to try
55:52to answer these questions.
55:55So, as I said, why they were
55:57controlling? And this is something that
55:58again you observe in many places that
56:00the territorial control that having like
56:02some area where they control uh can be
56:05very small, but it's very important. In
56:07the case of El Salvador, this meant that
56:10the gangs exert near total control of
56:12the territory. Basically, to prevent the
56:14rival gang or the police or
56:17>> [clears throat]
56:17>> or the state to enter, they would check,
56:19you know, your place of residence and
56:21your ID. So, you would need to have the
56:23permission of the gang member to enter
56:25and sometimes even pay a small fee to
56:28walk in this territory.
56:29And most of the revenue of the gangs in
56:31El Salvador, and this I should have said
56:32earlier, comes from was coming from
56:35extorting local businesses. There was a
56:36little bit of drug, but not that much,
56:38only in the north, but mostly the main
56:40business was extortion. So, the
56:42territory not only allowed them to
56:45extort people that live there and um
56:48transportation companies that would pass
56:50and extort them, but also nearby places
56:52because then they could hide in these
56:53territories.
56:58And this is how this is an example of
57:00how they were mapping all the areas like
57:02the different neighborhoods, zones, um
57:05and corridors that they control. This is
57:07from El Diario de Hoy that they were
57:09coming like examples about him how these
57:12gangs precisely knew the whole
57:13territory, who lived there, how much
57:16they were earning. They have a precise
57:19knowledge about all the people that live
57:21in their territories.
57:23And as I said,
57:24the main business was extortion. It's
57:26difficult to measure extortion.
57:29People in these neighborhoods, we
57:30couldn't ask it in a in a survey
57:32question because people were not wanted
57:34to
57:34to answer. So, what we did, and this is
57:36something that we using this other paper
57:38that I also sent you to read, is we
57:41collaborate with one of the main
57:42distribution companies there, and we got
57:45information about all the places that
57:47they were delivering and the extortion
57:48that they paid. So, the triangles is
57:51where the the buses deliver these goods,
57:54and um
57:56in orange and yellow, you have there the
57:58amount of extortions that they have to
57:59pay, all the stops that they have to do.
58:01And as you can see, they stop in many
58:03places, these drivers, and they pay a
58:06lot in a lot of areas.
58:08But I'm going to focus on this paper in
58:10just a couple of slides.
58:13Uh so, what we do first is ready to see,
58:16okay, how economic conditions vary if
58:18you live inside the boundaries of gangs
58:20versus outside. And what we observe is
58:23that they have much worse material
58:25well-being,
58:27lower income, and lower education. And
58:29even when we compare to people that is
58:31just 50 m away. So, the discontinuities
58:33are very stark. And when you look to the
58:35data in 1996, you don't have these
58:37discontinuities. So, the first thing we
58:39were wondering is like, how is this
58:40happening? Is it that they are extorting
58:42more these places, but then we learned
58:44that they were with the data of the
58:45companies, we could see that they were
58:46also extorting nearby places. So,
58:48extortion couldn't be what could be
58:50explaining at full, you know, these
58:51discontinuities.
58:53So, we went to these places, and we
58:54tried to collect household surveys. As I
58:56said, collecting data on these actors is
58:58difficult, and also Chris mentioned a
59:00bit that. So, the only things that we
59:02could
59:02ask a bit was mostly about labor
59:04mobility, and also some questions about
59:07um
59:07public goods, or who they resort when
59:10they have a problem, they resort to the
59:11criminal actor, but we couldn't ask
59:13anything related to um
59:16to extortion.
59:18So, the first thing that we saw in that
59:20same survey is that if you live in an
59:22area that is under the control of gangs,
59:24basically in those blue areas, versus if
59:26you live just in front of the street, 15
59:28m away, you are much more likely to be
59:31working also in that same area.
59:34So, it seems that you have restrictions
59:36to mobility that that not allow you to
59:38go and work in this location on or or
59:41increase the cost to work
59:44um further away.
59:46Which if you think about it, if you can
59:47only look for jobs in where you live,
59:49this limits your labor options and
59:51create some uh frictions.
59:54Uh
59:55Okay, so this is uh in terms of
59:57mobility, we also saw there are other
59:59restrictions. This was not only in terms
1:00:00of the labor market, also we see that
1:00:02they are much less likely to go to the
1:00:04beach that is just 20 minutes uh away
1:00:07from here and here, so it's it's a lot
1:00:10is very small. So, the restrictions to
1:00:12mobility apply beyond the labor
1:00:15um sector.
1:00:17In terms of aggregate, so I show you
1:00:19just now that there are these
1:00:21islands that they control, but they also
1:00:23have aggregate effects. In terms at the
1:00:25more aggregate level, you can think that
1:00:27the effects uh the negative effects are
1:00:28even larger because if you are a firm
1:00:31and you decide where to open, I'm not
1:00:33going to open in the blue areas, but I
1:00:35will not also open in the nearby areas.
1:00:36I will open further away, and this is
1:00:38what we see at the more aggregate uh
1:00:40levels that in areas that have some gang
1:00:42presence, firms are less likely to open,
1:00:45lower luminosity growth, less household
1:00:47income per capita.
1:00:49And as I And if you remember what I
1:00:51showed you before, so if now I live in a
1:00:54gang controlled territory,
1:00:56it seems that um
1:00:57it's much more important for me to be
1:00:59able to commute because I know that the
1:01:01probability of me of having a good firm
1:01:03nearby or like an an employment
1:01:05opportunity declines in those locations.
1:01:07Okay?
1:01:09So, this is in terms as you can see
1:01:10here, I was not talking anything about
1:01:12violence. This was a territory that we
1:01:14could think that was a stable control.
1:01:17Actually, most of the homicides, when we
1:01:19look at the data, it happens when when
1:01:20someone was to enter just in the
1:01:23in this in these areas.
1:01:26But beyond that, you still have all
1:01:28these negative economic consequences.
1:01:31Uh
1:01:32and also there there are other work by
1:01:34other um
1:01:35colleagues that are working also on how
1:01:37these criminal organizations
1:01:39can also beyond violence can also
1:01:40affect, you know, um support for
1:01:43hardline security policies. For example,
1:01:45they show that in
1:01:47if you're exposed to homicides from
1:01:49organized crime in imaging, you are much
1:01:50more likely to endorse extrajudicial
1:01:53solutions to crime and violence, and
1:01:55your position, you know, on authorities
1:01:57acting outside the law, which can have
1:01:59also other um consequences um
1:02:03for institutions.
1:02:05So, I think we have 5 minutes. So, how
1:02:07can we prevent the expansion of
1:02:08organized crime and violence? This is an
1:02:10open question. I don't think that we
1:02:12know the answer of this, but
1:02:14[clears throat] I'm going just to review
1:02:16a bit of the literature so that you
1:02:17know, and tell you the gaps there, but
1:02:19we have very little. In general, we have
1:02:22a we have some literature about like
1:02:24crackdowns, and that can be costly.
1:02:27I have a one slide that I left um
1:02:30with clickable, so I will not cover here
1:02:32that you can also look, and you can
1:02:34email me if you have questions. But
1:02:35basically, what we know the events of
1:02:37crackdowns is that in general, um it's
1:02:40when you target one locations, these are
1:02:42strategic actors, so you will have a
1:02:44spillover effect.
1:02:45And I we want to discuss now a bit more
1:02:48about other options that we can have.
1:02:50So, and if I have time to cover this of
1:02:53the trusses that I know that you read.
1:02:55Okay, going back, if you remember the
1:02:57paper on Peru on economic incentives.
1:02:59So, if we know that kids join because of
1:03:01this higher returns to the illegal
1:03:03sector, so
1:03:05when I was doing that paper, I I found
1:03:07out that Peru also have these
1:03:09conditional cash transfers, which
1:03:10basically they provide you a transfer
1:03:12with a condition that you have to have
1:03:13to stay school. And the conditional cash
1:03:15transfer
1:03:16roll out was independent on where
1:03:18coca was produced or not. So, I tried to
1:03:21see what happened with those kids that
1:03:23experienced a boom in the illegal
1:03:25sector, but also receive a transfer with
1:03:27a condition to stay school. And we see
1:03:29that they are less likely to work, and
1:03:30then I also see that they are less
1:03:32likely to be exposed to illegal
1:03:33industries looking incarceration data.
1:03:36So, here, I remember when I did this
1:03:38paper, I was very hopeful. I was, "Okay,
1:03:41for cases that these economic
1:03:42incentives, we have to go to the Becker
1:03:44model of human capital and school, then
1:03:48we can have also an incapacitation
1:03:50effect or an increase in the return to
1:03:52human capital because they stay school,
1:03:53and this can reduce criminal expulsion."
1:03:56But this actually, and this is why I put
1:03:58if this would work in other contexts,
1:03:59unfortunately, what I'm observing in
1:04:01other places is that, um,
1:04:05what before schools would be like a safe
1:04:07place where, uh,
1:04:09where you would like learn, now
1:04:12increasingly they're having more
1:04:13criminal actors inside the school.
1:04:15Via students,
1:04:17principals, or teachers. So, all this
1:04:20model of all, you know, the school as
1:04:22protector kind of breaking those other
1:04:24contexts where I have been working in
1:04:25Peru and in Ecuador where kids do not
1:04:28drop out from school. Everyone stay, but
1:04:30that has also discovered that even like
1:04:33people that would be recruiting inside
1:04:34school would stay.
1:04:37So, and this came to this other paper
1:04:39that we have. This is not published and
1:04:41we're still working on the revisions of
1:04:43this paper, but the main thing that we
1:04:45wanted to understand is like is it the
1:04:47case that if I prevent gang recruitment
1:04:50at young ages, this will affect, you
1:04:52know, the business of these gangs at an
1:04:54aggregate level? And of course, we go to
1:04:56micro to a bit more macro, and as you
1:04:57can see, as we go a bit more macro, the
1:04:59evidence is weaker because it's more
1:05:01difficult to observe. But basically,
1:05:03again, in the case of El Salvador, what
1:05:04we were observing, high territorial
1:05:06control. Here, the
1:05:08the recruitment
1:05:10and kids were mostly exposed not inside
1:05:12the school, but actually,
1:05:14in the entrance and outside, they were
1:05:16exposed in their way home, when they
1:05:18were in the neighborhood. So, in 2015,
1:05:21>> [clears throat]
1:05:21>> the government decided to do this uh,
1:05:23Escuela Seguras, which was basically to
1:05:25send preventive police patrolling around
1:05:27the schools at these times of entrance
1:05:29and exit.
1:05:30The idea was to reduce the contacts with
1:05:34gang members that could
1:05:36recruit kids.
1:05:38So, the first thing that we see is that
1:05:40if you were in the key ages of exposure
1:05:42like in in these grades, we see that
1:05:44they are less likely to drop out
1:05:46compared to
1:05:48to
1:05:49other cohorts that
1:05:50sorry
1:05:51to compare to other people that were not
1:05:54in the key ages when the program
1:05:55started.
1:05:57And then we also look at recruitment
1:05:58data. This is from the national
1:05:59intelligence, which basically they had
1:06:01people in these neighborhoods
1:06:04that could take notes whenever someone
1:06:07was joining the gang. So, of course it
1:06:08has a lot of measurement error and this
1:06:09is a sample of the real
1:06:13gang registry, but we use it as a proxy
1:06:15to try to see in those schools where you
1:06:17have this preventive police patrolling,
1:06:19is it the case that we start seeing less
1:06:20recruitment compared to other schools?
1:06:22And that's what we observe. Then we were
1:06:24worried that maybe what is happening is
1:06:26that you prevent recruitment or maybe
1:06:28this preventive police is arresting a
1:06:30bunch of kids and this is not what we
1:06:31want in this case. So, we look at
1:06:34incarceration data and we actually see
1:06:36that if anything then the young
1:06:38captures, if you were in in younger
1:06:41ages, you are less likely to be
1:06:43incarcerated and no effects for older
1:06:45ones.
1:06:46And then we look at extortion data and
1:06:48this is the part that is like the I
1:06:49would always say that it's the weakest
1:06:51because we try to see, okay, how is it
1:06:52the case that how recruitment can
1:06:54translate into the business of these
1:06:55gangs? And we see that in the long term
1:06:57it seems that there is a reduction in
1:06:58extortion.
1:07:00Why do we think of course here you will
1:07:02be thinking, oh Mika, before you just
1:07:04said that gangsters are strategic, that
1:07:05they can move. But there is something
1:07:07very interesting about this um
1:07:10this program is that
1:07:13strategic behavior
1:07:14two things that can happen, no, that
1:07:15gangster recruiting older individuals.
1:07:18It's still you increase the cost, no,
1:07:20because if the optimal was to recruit
1:07:21these
1:07:22young kids, even if you increase the
1:07:24cost and they still do it with older
1:07:26years increasing the cost of them. And
1:07:27the second thing is that they will go to
1:07:30a schools that are were not targeted.
1:07:32And as Chris mentioned before, this is
1:07:35very important and it goes against what
1:07:36we think when we think about
1:07:38decriminalization, the territories they
1:07:40control are very small, are small
1:07:42islands. So, and they have and they are
1:07:44a bit independent. It's not that for
1:07:46example, you do someone does a policy in
1:07:48my territory I'm a gang leader, it's not
1:07:50that I can go and recruit in the
1:07:53nearby territory. If you need the
1:07:54permission of the other ones too. So,
1:07:56it's not that obvious that they would be
1:07:58able to move. So, it's very important to
1:08:00try to understand how these gangs are
1:08:03distributed. And this is also important
1:08:06in terms of research because sometimes
1:08:07we start a research project,
1:08:10but we don't understand how is the
1:08:11environment and that's why I always
1:08:12think it's important what Chris was
1:08:14mentioning, understanding how these
1:08:15gangs work, whether they are
1:08:17confederations, whether they are
1:08:18fragmented, where they are located, is
1:08:20important to understand also these
1:08:21spillover effects.
1:08:24Okay, I think I passed the time, but I
1:08:26want to talk. The last thing I want to
1:08:28mention is about
1:08:30extortion.
1:08:32As I As If you remember well, I started
1:08:34saying that there are other consequences
1:08:36beyond violence,
1:08:38but in general most of the governments
1:08:40try to target violent behavior. And one
1:08:42of the policies that has been done a lot
1:08:45is truces. It's like you negotiate with
1:08:47these leaders to reduce
1:08:49the violent actions.
1:08:51But we are But when we were starting
1:08:53this paper, I think we know very little
1:08:55about how the structure of these gangs,
1:08:56whether they were competing or having a
1:08:58monopoly can affect these other
1:09:00the main business in this case that was
1:09:02extortion. And that's what we try to do.
1:09:05In general, as I said, extortion is
1:09:07difficult to measure. So, what we did
1:09:09was to try to get administrative data
1:09:11from the private sector on how much
1:09:14extortion they paid. So, we could like
1:09:15answer questions like, how do gangs
1:09:17determine extortion rates? How does gang
1:09:20competition matter for extortion? And
1:09:22what are the downstream effects on
1:09:23consumers?
1:09:24Again, these questions I put here, but I
1:09:27don't think that is um
1:09:29that we have the the answer. It's only
1:09:31for the context of El Salvador and it's
1:09:33only uh for extortion. But, there are
1:09:36other businesses that we don't know
1:09:37anything that is going on.
1:09:39But, focusing on this case, so basically
1:09:41what we do is to to understand how the
1:09:44structure of gangs matter for extortion,
1:09:46we exploit this 2016 non-aggression
1:09:49pact, which basically move from
1:09:51competition between the gangs to uh
1:09:53collusion. So, now they have divided the
1:09:55territory. And we want to see, okay,
1:09:57what happens? Yes, homicide declines,
1:09:59but what happens with extortion?
1:10:01Uh and what we see is actually that the
1:10:03amount that is collected by extortion
1:10:06that this uh increases
1:10:08by about 20% uh in places that
1:10:11previously have gang competition.
1:10:15And then we also have data on
1:10:17uh on prices from pharmaceuticals and we
1:10:19see that the prices of these goods that
1:10:21the distribution distribution company
1:10:23delivers and face higher extortion,
1:10:25increase the prices that they charge to
1:10:27these uh
1:10:28convenience stores and pharmacy
1:10:30pharmacies, and then the pharmacies also
1:10:32increase their prices to consumers.
1:10:34And what it is worse is that when we
1:10:36looked in hospital visits for chronic
1:10:38conditions that are affected by these
1:10:40increasing prices of
1:10:42uh drugs, we see that hospital
1:10:44admissions increase by 9%.
1:10:48So, although truces can reduce violence
1:10:50in the short term, higher extortion can
1:10:52also affect households in other measures
1:10:54like higher prices and uh health.
1:10:58Uh just to conclude, I'm going to the
1:11:00conclusion. So, I want you to get from
1:11:03this talk is first, we don't know much
1:11:05about these criminal organizations, so
1:11:07we are very happy to uh start that
1:11:09network with Chris.
1:11:11But, something that we don't know is um
1:11:14that they can affect through the very
1:11:15different channels, not only violence.
1:11:17Here I show you just some like less
1:11:19human mobility, less human capital,
1:11:22extortion.
1:11:23Um
1:11:25also the the same thing that Chris was
1:11:26mentioning on the article that you read
1:11:28of that he wrote, interventions must
1:11:31match this local criminal structure.
1:11:32It's very important. And something
1:11:35important is that they recruit children
1:11:36and adolescents, so we have to try to
1:11:38think more about interventions about to
1:11:40to reduce their exposure. And that's it.
1:11:42Thank you.
1:11:47>> Thanks so much, Milena and Chris.
1:11:50That was really, really interesting.
1:11:52We've had very many questions come come
1:11:55in, so we'll take some time to answer a
1:11:57few of those. And um
1:12:00maybe I'll start with one in the chat
1:12:02from Masateru,
1:12:04uh who's essentially asking
1:12:07uh about how you collect data on illicit
1:12:11activity. So, maybe could you guys both
1:12:13share a little bit of your the
1:12:15challenges and some successes of good
1:12:17ways to collect data um on criminal
1:12:20organizations and the informal sectors
1:12:22where they have incentive to hide or
1:12:24misrepresent their activities.
1:12:28>> Chris, do you want to go next because I
1:12:30think you were more successful and you
1:12:31know this better.
1:12:33>> I mean, it's very hard. It takes a lot
1:12:34of experimentation. Um one is I think it
1:12:37helps to have local collaborators or
1:12:39partners,
1:12:40even if they're not academics, but
1:12:42people who can help you understand
1:12:43what's safe or not. So, I think we're
1:12:45able to work in Chicago or Medellín or
1:12:48there are things I would do there that I
1:12:49wouldn't do in um
1:12:51northern Mexico. Just like when we were
1:12:54doing conflict research, there were
1:12:55things you could do in Liberia or Uganda
1:12:58or much of Africa that you could never
1:12:59do in Iraq or Afghanistan, and you saw
1:13:02conflict researchers um
1:13:05behave accordingly according to local
1:13:06security conditions.
1:13:08Um and this is where advising is very
1:13:10important, and this is part of the
1:13:12reason for our mentorship network.
1:13:14Uh I would say but at the end of the
1:13:16day, it's just, you know,
1:13:18any kind of new measuring something
1:13:20that's informal, not just clandestine
1:13:22activities, but just measuring new
1:13:23things is just a big investment, and
1:13:26it's and it's uh it's a big contribution
1:13:28in itself. I think journals tend to
1:13:29reward that.
1:13:31Um and and and there's a lot of trial
1:13:33and error. So, in terms of measuring
1:13:35extortion or or measuring proclivity or
1:13:38interest in in entering gangs like our
1:13:40our approach to clones, it was it was
1:13:42months of trial and error.
1:13:44Uh and I think interventions are months
1:13:46of trial and error. And I think the
1:13:48advantage that the young scholars have
1:13:49is time. So, so I think, you know,
1:13:52people like me can have lots of
1:13:53experience, um
1:13:55but we don't have a lot of time. And and
1:13:57so I think there's a lot of time for
1:13:59trial and error to sort of get a sense
1:14:01of what's working. That's partly based
1:14:03on intuition and then eventually
1:14:04confirmed by results. But but I
1:14:07But just big investments in really
1:14:09learning how to measure things is is
1:14:10something development economics has
1:14:12always done well
1:14:13and has always pioneered. And I think um
1:14:16this is no exception.
1:14:21>> Great. Thank you. Um
1:14:23maybe we'll go to one of our our live
1:14:25participants. Upamanyu, would you like
1:14:28to ask your question?
1:14:29>> So, when you are talking about this
1:14:31criminal governance, so you're mostly
1:14:33talking about them outside the purview
1:14:35of formal elections or political
1:14:37contest. So, in India and other
1:14:39developing countries, we are seeing
1:14:40these criminals enter into the official
1:14:42contest and then protecting either their
1:14:45interest or solving some of voters'
1:14:47problems. So, in these Latin American
1:14:49countries, do do you see this sort of
1:14:52phenomenon where they're entering
1:14:54elections and contest and then in which
1:14:56constituencies are precincts they
1:14:59generally run from? And like what do
1:15:01they do after entering election if they
1:15:03do so?
1:15:04>> Well, I would make you speak about Latin
1:15:05America, but but I mean I mean India, I
1:15:08don't know if you're from India or not,
1:15:09but India's unusual to my mind to in the
1:15:12degree to which politics is
1:15:13criminalized. The most important fact in
1:15:15the world that almost nobody knows is
1:15:17something like 50% of parliamentarians
1:15:19in India have a felony criminal charge
1:15:21outstanding against them. Maybe a
1:15:23quarter like a quarter of
1:15:24parliamentarians have a homicide or
1:15:26attempted homicide charge against them.
1:15:28Like that's crazy. And nobody studies
1:15:30it. There's one political scientist
1:15:33uh huge huge huge area for study.
1:15:36Um,
1:15:37nobody's touching it.
1:15:38Completely insane.
1:15:40Mika, I don't know if you want to speak
1:15:41to Latin America.
1:15:42>> Uh, we have a lot of like presidents in
1:15:44Latin America that went to [laughter]
1:15:46prison, but for other may No. So, I
1:15:49let's see.
1:15:51Quantit- I think I have Cordova has nice
1:15:53work on this uh,
1:15:55for El Salvador, but it's mostly about
1:15:56how they collude, but not that much that
1:15:59they enter criminal actors. At least of
1:16:02the what I'm having like this thinking
1:16:04fast.
1:16:05Uh, as the case of India, yeah, we don't
1:16:07have that, but
1:16:09it doesn't mean that that doesn't exist.
1:16:11Uh,
1:16:12that's the thing. I I could ask I I
1:16:14should ask I know I there was a
1:16:17a colleague of mine that was at Colombia
1:16:19studying this in Peru about the links
1:16:21um,
1:16:22from um,
1:16:24municipal mayors and links with criminal
1:16:26actors in these areas of drug
1:16:28trafficking. Which qualitative they
1:16:30would talk about, but you wouldn't see
1:16:32it like in the big cities, you know?
1:16:34Like it's it's something that some
1:16:36people
1:16:37talk a little bit about these like
1:16:38agreements that they would have with
1:16:40criminal organizations, but not like a
1:16:41criminal member being, you know,
1:16:44as part of.
1:16:50>> Really interesting. I I I think that
1:16:53leads into a couple other questions in
1:16:55the chat that are kind of asking about
1:16:57state collusion with with criminal
1:16:59groups. Um,
1:17:00and I think there's a lot of first-hand
1:17:02accounts that are being shared that are
1:17:03really really interesting to read. Um,
1:17:06but maybe if you guys could touch on a
1:17:08little bit about um,
1:17:12maybe under what conditions state
1:17:15collusion with criminal groups produces
1:17:17stability and when does it ends
1:17:19when is it source of instability and
1:17:21violence and how how have you guys um,
1:17:25seen or what
1:17:26research is there on how to tackle that?
1:17:36>> Do you want to go crazy? I think
1:17:38I'm very negative on that.
1:17:41Okay, so like the only the one that I
1:17:44have studied with more is the one that I
1:17:45explained briefly before about like
1:17:48So the it was actually the one that we
1:17:49focused on extortion but it was a
1:17:50collusion among the gangs, but this was
1:17:54possible actually because in 2012 they
1:17:57have these truths with government. Where
1:17:59basically they said they sat down with
1:18:02the leaders in the table and they were
1:18:03like okay Let's negotiate. I want you to
1:18:06reduce
1:18:07this homicide rates. What do you need?
1:18:09How what we can do for you to reduce?
1:18:12Um
1:18:14But of course this was not
1:18:16uh
1:18:17didn't have the legitimacy. A lot of
1:18:18people were not not knowing about these
1:18:20truths were secret and then of course
1:18:23this broke and then homicides went even
1:18:25higher.
1:18:27Um
1:18:28But from there it's what motivated us to
1:18:31start the pact in 2016 because uh
1:18:34qualitatively what they were
1:18:35talking a lot in the communities were
1:18:37okay yes violence is lower, but we're
1:18:40still trapped in these under like these
1:18:44restrictions that criminal organizations
1:18:45impose on us in different ways like
1:18:47extortion in this case, but I can
1:18:48imagine that other margins might work in
1:18:51other countries.
1:18:53So how to prevent him? I don't know.
1:18:57I I don't know how you can prevent maybe
1:18:59Chris would know better about this like
1:19:01how you prevent like
1:19:02a government because incentives to
1:19:04reduce violence are very high.
1:19:06Politically is pays a lot, no? Like if
1:19:10you reduce homicide, but I don't know
1:19:11how you can prevent this from to not for
1:19:14them to not have this.
1:19:15>> Yeah.
1:19:16>> process. I mean I think um
1:19:20I think that there I think there's a big
1:19:22open
1:19:23I could talk for 20 minutes about this.
1:19:24I think here's a few big open research
1:19:26questions.
1:19:28Colombia was completely captured at the
1:19:30national congressional presidential
1:19:32level by criminal cartels
1:19:3430 years ago and today it's a much less
1:19:37captured much more functioning
1:19:38democracy. How is that possible?
1:19:40Completely unanswered.
1:19:42Um to what extent does um so Medellin is
1:19:46is not very captured by criminals. Some
1:19:48surrounding municipalities are, but has
1:19:51you know do mayoral terms, does does
1:19:53does electoral turnover, does the
1:19:55shifting of police chiefs every few
1:19:57years prevent criminal capture? So so
1:20:00you could imagine large and municipal
1:20:02analyses across the hemisphere
1:20:05uh trying to answer questions like that.
1:20:08Um with the
1:20:09or there's enough countries that are
1:20:11large and have very presence like Mexico
1:20:13and Colombia or Brazil that you could
1:20:15you could do these within country
1:20:17analyses as well. So
1:20:19I think it's actually a big question
1:20:21with that's a frontier. We don't have
1:20:23good answers.
1:20:24Um I think I don't think it would be
1:20:26actually that hard to to start to push
1:20:28down that as a as a research avenue.
1:20:36>> Thanks, guys. Sorry. Um
1:20:39I I think
1:20:40we're getting lots of good ideas for
1:20:42potential
1:20:44uh dissertations and theses uh from all
1:20:46these these open research questions.
1:20:48>> That's our goal.
1:20:49>> Everyone take notes.
1:20:50Um
1:20:52Amalia, I I see you next on my list.
1:20:54Would you like to ask your question?
1:20:56>> Sure.
1:20:57So more from a
1:20:59Sorry,
1:20:59>> [clears throat]
1:20:59>> more from a practitioner perspective,
1:21:02how well does your message that there's
1:21:04no cookie-cutter approach to organized
1:21:06to tackling organized crime? How well
1:21:08does that sit with funders at a time
1:21:11when cost-effectiveness is seems to be
1:21:13getting the premium?
1:21:15And [clears throat] from what I know and
1:21:17what I've observed across continents,
1:21:19there is no easy answer as you flagged
1:21:21in one of the papers on the reading
1:21:23list. When you put that forward for
1:21:26funders, how do they react?
1:21:29>> I mean I'd say there's varied. I mean
1:21:33what like
1:21:34the IDB is a big lender in this area. I
1:21:36think they have been receptive to this.
1:21:40Um
1:21:41I I think I think I think the more
1:21:43sophisticated the
1:21:45like the mayor or the police chief we
1:21:46deal with, I think the more receptive
1:21:48they are to this.
1:21:49Um some of them are also
1:21:53happy to
1:21:55understand like that
1:21:58Anyway, I I think some of them are also
1:22:00like the like the fact that it gives
1:22:02them some defense against other people
1:22:04who are pushing them to go do these
1:22:06cookie-cutter solutions that cost lots
1:22:08of policing.
1:22:09Um and they get it, but but I I think
1:22:12it's it's it's not universal. So, it's
1:22:14been very heterogeneous so far.
1:22:17I Time will tell.
1:22:23>> Thanks. Let's have one more from the the
1:22:25panel. Andrea, would you like to ask
1:22:26yours?
1:22:28>> Uh hi. Thank you very much for the
1:22:30presentation. It was amazing.
1:22:32Uh first, I was wondering if we could
1:22:33have the paper about the mayors in Peru
1:22:36uh listed somewhere that uh
1:22:40Miguel Adam mentioned. And also I I had
1:22:43a question about Peru specifically in
1:22:44Peru. We've been uh studying alternative
1:22:46development in Peru. And I noticed a
1:22:49different between difference between
1:22:52both two places that both produce coca,
1:22:55but some of them have also uh informal
1:23:00or internal ways of solving
1:23:02uh crime or other things that are going
1:23:04on inside a community. So, I was
1:23:06wondering in the framework that we have
1:23:08heard about criminal organizations, how
1:23:10would these different coca-producing
1:23:14uh places would be characterized?
1:23:17Thank you.
1:23:19>> Yes, actually Andrea, there is another
1:23:21paper that I have here that is um she
1:23:24was a former job market candidate from
1:23:26Minnesota that worked with me, Natalia
1:23:27Pieragallo, who is in who she's Peruvian
1:23:30as well. And she has worked a lot of
1:23:32these of alternative uh the role of
1:23:34Ronda Campesinas. And actually that they
1:23:37are like very uh good because they were
1:23:39they it's it's a very nice paper. I will
1:23:41put it that one and the one and other
1:23:43one in Peru. I will send it to you.
1:23:45Um
1:23:46how they uh move more this restorative
1:23:48justice and they could prevent, you
1:23:50know, entrance of criminal actors. Uh
1:23:53that's a one that is is actually you're
1:23:55right. In those places you also have
1:23:57this other factor that could be
1:23:58explaining why I don't see in the I
1:24:01never thought it in that way like uh for
1:24:03the other paper that I have that I
1:24:04presented today, but you're right.
1:24:06That's another mechanism that can go in
1:24:08on why in those legal places I'm seeing
1:24:10less of the crime because they also have
1:24:12these other institutions.
1:24:17>> Amazing. Thank you. Um
1:24:19and maybe this that that's a good uh
1:24:22note. Could you guys just talk a little
1:24:24bit more about this mentorship um
1:24:27or organization that that you've
1:24:29started. There's a couple questions
1:24:31about who is eligible, what uh country
1:24:36of origins or country of residence this
1:24:38might be uh eligible and um more about
1:24:42the scholarship potentially
1:24:44um and maybe just just give it a a
1:24:47little more plug.
1:24:49And how they might like
1:24:50>> Well, yeah, to be we're still defining
1:24:52it because we've just received our
1:24:54funding, but um Mika, do you want to say
1:24:56a little bit?
1:24:57>> Yeah, but the idea like what we're
1:24:59trying to do is to have
1:25:01great coverage, not only, you know, uh
1:25:04the usual places because we try like to
1:25:06like for it try mostly like to value
1:25:08like people that are very interested on
1:25:10this topic and that they want to, you
1:25:12know, understand and have like good
1:25:14research and not just only based on I
1:25:16don't know I'm here in Princeton on my
1:25:18on the students that I have here. So, we
1:25:19would try to do it broad. The idea is to
1:25:22have like um
1:25:23we we have we are thinking about like
1:25:25different like program one that is like
1:25:26for early um
1:25:29career early like PhD when you're trying
1:25:31to think about an idea and then for
1:25:33people that already have advanced uh
1:25:36into the proposal, but then they need
1:25:37more of this network to provide also
1:25:39feedback and have monthly meetings.
1:25:42Uh and we also would have a bit of
1:25:43funding so to start, you know, a
1:25:45project.
1:25:46I also some pilot. Not like for a big
1:25:48RCT, but also when you need to do some
1:25:50scoping to try to understand, I don't
1:25:52know, you want to work on criminal
1:25:54organizations Nigeria, but you we don't
1:25:56know anything even qualitative. Okay, we
1:25:58can give you a pilot, you know, to go
1:25:59and start trying to see which data you
1:26:01can collect. But the idea is to do it
1:26:03broad, like we will try to do as much as
1:26:05we can with the funding. The but the
1:26:07open will be called like and then we
1:26:09will try to select based on
1:26:11on the research question and mostly
1:26:13about like how interested you are on
1:26:14this topic.
1:26:16>> Yeah, we have 2 years of pilot funding,
1:26:18which we hope successfully turns into
1:26:20like sustained pilot fund sustained
1:26:21funding from this donor. Um
1:26:24it the idea is I think we'll probably
1:26:27have a competitive call where there will
1:26:29probably be an equilibrium, maybe
1:26:31hopefully 10 per year.
1:26:33So it's not a lot. Um
1:26:35if there's more qualified applicants,
1:26:36we're going to try to create a broader
1:26:38community of practice and community of
1:26:39scholarship rather. Um it's mostly going
1:26:42to be PhD students, maybe some postdocs
1:26:45at different stages.
1:26:47Uh our our executive director Ben Krause
1:26:49is probably going to help to sort of
1:26:51help the broader group of people connect
1:26:53cuz if we're but but you know, Mika and
1:26:55I and the small group of people, we have
1:26:56very limited time.
1:26:58And we can't advise everybody. We want
1:27:01to try to advise more than we're
1:27:02currently advising and so we're going to
1:27:04start with 10 and um and but I I guess I
1:27:08could just say it's going to be like a
1:27:10competitive meritocracy and with the
1:27:12best ideas and we expect there to be,
1:27:14you know, global eligibility.
1:27:17Um but but and and and we'll figure and
1:27:21and by year two or three it hope it will
1:27:23probably look different than it does now
1:27:25because we'll have learned.
1:27:27>> Yeah, perfect. And where can people
1:27:30um
1:27:31see this when the calls open or who can
1:27:33they contact about questions?
1:27:35>> good question.
1:27:37>> We will share it to you. Probably we
1:27:38will share it to all these networks so
1:27:40that then you will receive
1:27:41>> Yeah, you you can push it out to your
1:27:42network. I think I mean, I'll be posting
1:27:45it on X.
1:27:48And we'll probably Yeah, it's a good
1:27:50question. We don't know. We haven't
1:27:51figured this out yet. We just got the
1:27:53money like 3 weeks ago. So
1:27:56I don't know.
1:27:57>> Yeah, I mean that's very very recent. So
1:27:59I guess everyone follow Chris on Twitter
1:28:02X and he'll keep you posted.
1:28:05And then I'm sure other things will come
1:28:07out through through the the network that
1:28:10comes comes to fruition.
1:28:12Well
1:28:14I think this is probably a good place to
1:28:16wrap.
1:28:18I think that this is a huge area for for
1:28:21research and it's been really exciting
1:28:23to hear
1:28:24both of your perspectives today. Thank
1:28:26you everyone else for your questions and
1:28:29your your insights. I always enjoy
1:28:31reading people's experiences as well and
1:28:33how they've seen what has been talked
1:28:36about in their hometowns and home
1:28:39countries and the perspectives that you
1:28:40bring.
1:28:43We will wrap up for today. We will
1:28:46obviously be back next week and I just
1:28:48want to clarify with the time change
1:28:50that happened with the UK and then with
1:28:52the US.
1:28:55The the time is set now for 3:30 p.m.
1:28:58GMT London time. So I apologize the past
1:29:01couple weeks that people have showed up
1:29:03earlier late because their countries
1:29:05didn't have it. But it's 3:30 p.m. UK
1:29:07time. So just plan accordingly for next
1:29:09week.
1:29:12We will have an exciting lecture next
1:29:14week by James Robinson
1:29:17on the political economy of Africa. So
1:29:20if you're interested in hearing a Nobel
1:29:22Nobel laureate
1:29:24come by at 3:30 p.m.
1:29:26So thanks again for for Meagan Chris and
1:29:29I hope everyone has a great rest of your
1:29:31day.
1:29:33>> Thank you everyone for connecting.
1:29:38>> Thank you.

About this transcript

This page contains the full transcript of BREAD-IGC virtual PhD-level course 2026 | Lecture 9 Crime and Organised Crime by International Growth Centre, generated from the public captions YouTube serves with the video. The transcript has 15,862 words across 2,568 segments, with the original timestamps preserved so you can click any line to jump to that moment in the embedded player.

What you can do with it

Use the transcript to take notes, quote the speaker, build a study guide, generate a summary with ChatGPT or Claude via the YouTube Summary tool, or export it as a timed subtitle file with YouTube to SRT. You can also re-open it in the transcriber to choose a caption track available on the video.

Free YouTube transcript tool

YouTube2Text is a free YouTube transcript generator — no signup, no daily limit. Paste any YouTube link and get the full transcript instantly, with timestamps, click-to-jump, caption language selection, AI prompts for ChatGPT, Claude, and Gemini, and exports to TXT, SRT, VTT, or Markdown.