Transcript political

Political Determinants of Violence
in the Metropolitan Area of Buenos Aires
October 2014
Alberto Föhrig
UdeSA
Research Questions, Hypothesis, Methods
This paper intends to provide some evidence and analysis about the links between
politics, the police, and crime in the metropolitan area of Buenos Aires. It intends to
provide criteria to explain significant variability in municipal crime rates.
Why violence? Dependent variable: crime against individuals excluding car accidents
Are political factors related to increases in violent crime?
What is the relationship between stability –measured as re-election rate for mayor´s –
and fragmentation –measured as increased effective number of parties and intra-party
fragmentation – with violence?
Hypothesis: The increasing number of political and drug trafficking groups competing for
territory produce unstable agreements and tend to increase violence.
Mixed method approach. Panel data model with fixed effects and clustered errors
combined with the qualitative study of court cases involving relationships between
2
politics, police and crime.
Theoretical background: crime, the police and politics
 Different authors (Saín, 2002; Tokatlian, 2011; Auyero 2012) have underscored the
increasing links between certain political actors with criminal organizations in
Argentina.
 Gambetta (1996), Villareal (2002), Wilkinson (2004), Garay (2013), Osorio (2012),
studied the relationship between politics and crime in different contexts. Snyder and
Duran Martinez (2009) theorize under what conditions criminal groups are able to use
state sponsored protection rackets to develop their activities.
 Fajnzylber et.al. (1998) produced a classic study on the determinants of crime in
Latin America in which they concluded that inequality more than poverty as well as
GDP per capita rates had a significant impact on crime rates.
3
Theoretical background: Fragmentation as a Multilevel Game
•
Politics and crime are both territorially defined and structured in multi level layers
•
1.
2.
3.
Consensus on fragmentation of the Argentine political system:
Decreasing levels of party nationalization (Jones and Mainwaring 2003, Leiras 2006)
Increased ENP and Territorialization (Calvo and Escolar 2005, Leiras 2006)
Lack of Congruence between the national and provincial party systems (Gibson and Suarez
Cao 2010)
4. Intra-party fragmentation (Föhrig 2011, Föhrig and Post 2007)
•
Mayor´s linked to the police in a variety of informal dimensions:
– Influence police officers’ careers: they lobby the governor and may in fact veto the appointment of
police authorities in their districts given their previous records.
– Influence their promotions and exonerations.
– In operative terms they provide police with money and equipment. Operationally influence the
allocation of police resources given their monitoring capacities provided by surveillance cameras.
– Mayors authorize commercial ventures to operate within the boundaries of their municipalities
– Have privileged access to a key political asset: information. Bridge informational gaps.
4
Context
•
The Metropolitan Area of Buenos Aires concentrates a quarter of the country´s population
.
and is the second most violent area in the country (Lodola and Seligson, 2012: 128)
•
Significant increases in crime rates, concentration of crime, and organized crime activities.
• The context in which this paper tests it empirical hypothesis is one in which the police
informal regulation of criminal activity started to crumble as a consequence of the
expansion of the drug market. The new market and institutional incentives in place
generated the emergence of new organized crime organizations on the ground.
•
Party system change: simultaneous influence of fragmentation and party predominance.
5
Drug´s Market
Cocaine Seizures in Argentina (kg. per year)
Source: UNODC, various years.
Drug Consumption by School Children
Source: OAS, 2013
Local Processing: 80 facilities producing different phases of drugs were shut down by
enforcement agents between 2000 and 2006 (Sedronar, 2011).
Sinthetic Drugs: 600.000 pills production facility discovered in Mar de Ajo (2013) doubled total
seizures in Ezeiza Airport since 2004.
6
Almirante Brown
Avellaneda
Berazategui
Esteban Echeverria
Ezeiza
Florencio Varela
General San Martin
Hurlingham
Ituzaingo
Jose C Paz
La Matanza
Lanus
Lomas de Zamora
Malvinas Argentinas
Merlo
Moreno
Moron
Quilmes
San Fernando
San Isidro
0
500
1000
0
500
1000
0
500
1000
0
500
1000
Crime rates against individuals
1995
Tigre
Tres de Febrero
2005
2010
Vicente Lopez
0
500
1000
San Miguel
2000
1995
2000
2005
2010 1995
2000
2005
2010 1995
2000
2005
2010 1995
2000
2005
2010
year
Graphs by municipality
7
Intra-party fragmentation
Avellaneda
Berazategui
Esteban Echeverria
Ezeiza
Florencio Varela
General San Martin
Hurlingham
Ituzaingo
Jose C Paz
La Matanza
Lanus
Lomas de Zamora
Malvinas Argentinas
Merlo
Moreno
Moron
Quilmes
San Fernando
San Isidro
1
2
3
4
1
2
3
4
1
2
3
4
1
2
3
4
Almirante Brown
1995
Tigre
Tres de Febrero
2005
2010
Vicente Lopez
1
2
3
4
San Miguel
2000
1995
2000
2005
2010 1995
2000
2005
2010 1995
2000
2005
2010 1995
2000
2005
2010
Year
Graphs by municipality
8
Reelection
Avellaneda
Berazategui
Esteban Echeverria
Ezeiza
Florencio Varela
General San Martin
Hurlingham
Ituzaingo
Jose C Paz
La Matanza
Lanus
Lomas de Zamora
Malvinas Argentinas
Merlo
Moreno
Moron
Quilmes
San Fernando
San Isidro
0
2
4
6
0
2
4
6
0
2
4
6
0
2
4
6
Almirante Brown
1995
Tigre
Tres de Febrero
2005
2010
Vicente Lopez
0
2
4
6
San Miguel
2000
1995
2000
2005
2010 1995
2000
2005
2010 1995
2000
2005
2010 1995
2000
2005
2010
Year
Graphs by municipality
9
Crime rate and effective number of parties
Graph. Distribution of Crime Against Individuals
.6
.4
0
0
.2
.002
.004
kdensity nep
.006
.8
1
.008
Graph. Distribution of the Effective Number of Parties.
200
600
x
400
Year 1995
Year 2008
800
Year 2003
1000
2
6
x
4
Year 1991
Year 2008
10
8
Year 2003
10
Methodology
 Panel data model with fixed effects and clustered errors.
 336 annual observations, comprising the 24 municipalities in metropolitan area
of Buenos Aires between 1995 and 2008.
Dependent variable
 Crime rate against individuals excluding car accidents
Independent variables







Energy rate consumption
Number of students per inhabitant
Margin of victory
Effective number of parties
Cocaine seizure
Reelection
Intra party fragmentation
11
Variables are expressed in logs.
Panel Data Models
VARIABLES
Model 1
Model 2
Model 3
Model 4
Model 5
Model 6
Model 7
Model 8
0.182
0.0676
0.0505
0.0176
-0.0496
-0.0640
-0.190
-0.187
(0.156)
(0.202)
(0.192)
(0.203)
(0.200)
(0.206)
(0.216)
(0.221)
Number of students per
inhabitant
-0.0934
-0.271**
-0.181*
-0.223**
-0.307**
-0.351**
-0.228*
-0.273**
(0.109)
(0.114)
(0.0953)
(0.105)
(0.124)
(0.125)
(0.121)
(0.124)
Margin of victory
-0.00474
0.00786
-0.00362
0.00106
0.0127
0.0160
0.000684
0.00457
(0.0168)
(0.0168)
(0.0154)
(0.0164)
(0.0161)
(0.0176)
(0.0150)
(0.0165)
0.0811
0.0910*
0.151***
0.154***
(0.0525)
(0.0528)
(0.0522)
(0.0520)
0.130***
0.117***
0.140***
0.128***
(0.0318)
(0.0360)
(0.0341)
(0.0333)
Energy consumption
Effective number of parties
Cocaine Seizures
Reelection
0.124**
0.143**
0.141**
0.124**
(0.0522)
(0.0566)
(0.0558)
(0.0566)
Years in government
0.0592**
0.0532**
0.0691***
0.0590**
(0.0233)
(0.0251)
(0.0237)
(0.0263)
Intra-party fragmentation
Deterministic trend
Constant
Observations
R-squared
Number of muni
-0.0452
-0.0312
-0.0909
-0.0685
(0.0530)
(0.0541)
(0.0607)
(0.0630)
0.0203***
0.0215***
0.0246***
0.0254***
0.0165***
0.0175***
0.0243***
0.0244***
(0.00446)
(0.00489)
(0.00451)
(0.00479)
(0.00468)
(0.00473)
(0.00490)
(0.00505)
1.385
2.572**
2.343**
2.651**
3.527***
3.719***
3.970***
4.090***
(0.884)
(1.083)
(0.972)
(1.091)
(1.001)
(1.082)
(1.042)
(1.122)
336
336
336
336
336
336
336
336
0.567
0.570
0.568
0.565
0.531
0.538
0.520
24
24
24
24
24
24
24
0.525
12
24
Qualitative Analysis: Court Cases and Legislative Inquiries
involving PPC
• Candela case:
1. Kidnapping and murder of a 11 year old by a police/drug dealers
mixed gang. Legislative Inquiry
• Ephedrine case:
1. Triple homicide of Gral. Rodriguez.
2. Conviction of Martinez Espinosa (Maschwitz drug processing facility)
3. Involvement of high ranking state officials
4. Illegal financing of president Cristina Kirchner electoral campaign
2007
13
Mechanisms
Scenarios of bilateral monopoly between criminal organizations and
political actors which produced stable agreements over time are
broken.
Both politics and drug trafficking involve a territorial and multilayered
dimension.
Drug traffickers need specific territories in order to transport,
elaborate and sell drugs. In order to do so they require “safe” portions
of land which enable them to develop these activities with low risks of
being caught by authorities. Because of geography, transportation
difficulties, and communication costs criminal organizations act locally.
14
Mechanisms (II)
Drug market forces increased the number of criminal organizations on the
ground. As a consequence, the number of players on the market side
increased over the last years.
Process of fragmentation: party factions that compete against each other in
territorial disputes fighting for party power in a multilevel game.
Relationships between party factions and criminal groups at the local level
within the context of political competition influence increases in violence.
Re-election and fragmentation are simultaneously maintained through
electoral system design: “listas colectoras” and “listas espejo” (Mustapic,
2013).
The increasing number of political and drug trafficking groups competing for
territory within scenarios of either cooperation or competition between the
two distinct activities produce unstable agreements and tend to increase15
violence.
Mechanisms (III)
• Senate endorses judges and prosecutors appointments
• Governors appoint, remove and rotate in different settings police
agents.
• Police does not enforce internal oversight
• Mixed members gangs: police and drug dealers
16
Conclusions
The models presented in this paper show the significance of political variables to analyze crime.
Political variables on fragmentation and re-election of mayors show an impact over violence.
The longer actors stay on the ground, the greater their ability to develop ties of reciprocity, trust
and reputation with the police and criminal groups. Re-election for mayors without restrictions
seems to be a measure with negative effects over crime.
When scenario of stability for mayors (party predominance) and fragmentation of the political
system coexists with market pressures for new organizations into the market, violence increases.
17
THANK YOU
@afohrig
18
Theorizing the relationship between politics and crime
Non violence
Scenario
State sponsored protection racket
Indicator
High level of drug seizures, high level
of domestic consumption
Lack of drug related criminal activity
Low levels of seizures and low
domestic consumption
High domestic consumption and low
levels of seizures
Rise in the number of homicides
among gang members in territorial
disputes. Spatial concentration of
homicides.
Non-violent drug market
Broken state sponsored protection
racket due to the entrance of new
players into the market or new state
agencies intervening. Territorial
disputes or succession conflicts and
consequent fragmentation
Violence
Open conflict between the state and Increase in the number of criminal
criminal organizations
organizations disarticulated and their
members imprisoned. Rise in the
number of casualties.
Collusion with/diversification
other forms of organized crime
to Rise in crime rates against property
and individuals
19
Errors distribution
20
Distribution of Crime rate against individuals (2008)
502
750
570
367
518
485
482
559
538
264
676
476
587
927
635
500
614
427
474
(629,927]
(548.5,629]
(479,548.5]
[264,479]
No data
677
582
623
286
640
21
Effective number of parties (2008)
3
7
3
4
3
6
6
4
4
5
4
4
3
6
5
4
9
3
7
3
(5.5,9]
(4,5.5]
(3,4]
[3,3]
No data
5
3
5
3
22
Intra-party fragmentation (2008)
2
4
2
2
1
2
3
1
1
3
1
1
1
2
2
1
1
2
2
(2,4]
(1,2]
[1,1]
No data
2
4
2
2
1
23
Reelection (2008)
0
3
3
6
2
5
0
2
4
1
1
3
2
1
0
4
1
0
0
3
(3,6]
(1.5,3]
(0,1.5]
[0,0]
No data
0
1
0
3
24
Energy consumption rate (2008)
365641
285599
364942
243067
128795
184597
361316
298304
212289271985
144075
185186
273297
148537
204309
315148
249891
261239
208204
208571
233826
(279448,365641]
(223057.5,279448]
(184296.5,223057.5]
[128795,184296.5]
No data
183996
169818
140482
25