Upload
coleen-lloyd
View
214
Download
0
Tags:
Embed Size (px)
Citation preview
Crime, Development and Crime, Development and Welfare in Latin AmericaWelfare in Latin America
Conference “Confronting Crime and Violence in Latin Conference “Confronting Crime and Violence in Latin America: Crafting a Public Policy Agenda,” July 2007America: Crafting a Public Policy Agenda,” July 2007
Rodrigo R. SoaresRodrigo R. SoaresCatholic University of Rio de Janeiro (PUC-Rio), University of Maryland, NBER, IZACatholic University of Rio de Janeiro (PUC-Rio), University of Maryland, NBER, IZA
withwith
Joana NaritomiJoana NaritomiWorld BankWorld Bank
Overview1. Crime and Welfare
2. Latin Americai. Measurementii. Patterns
3. Candidate Explanationsi. Socioeconomic conditions and repressive policiesii. Quantitative relevance
4. Concluding Remarks
1. Crime and Welfare
• Latin America has been traditionally regarded as a particularly violent region of the world:
– Deaths to violence 200% higher than North America and Pacific, 450% higher than Western Europe, and 30% higher than Former Communist block (WHO);
– 44% of the population reported being victim of some type of crime in the previous year (ICVS);
– Crime and violence as the second most important public policy issue, ranking first for countries such as Argentina, El Salvador, and Venezuela (Latinobarómetro 2006).
1. Crime and WelfareViolent Deaths - Regions of the World - 1990s (WHO)
0
5
10
15
20
25
Latin America North America Western Europe Form. Communist Western Pacific
Vio
len
t D
eat
hs
(pe
r 10
0,0
00)
1. Crime and Welfare• Many potential welfare implications:
– Direct welfare loss due to increased mortality;
– Reduced investments in human and physical capital due to shorter planning horizon;
– Material costs, including both direct costs and expenditures on criminal justice and crime prevention;
– Loss of human capital and productivity of those deceased, incapacitated and incarcerated.
1. Crime and Welfare• Material Costs:
– Direct costs and expenditures on criminal justice and crime prevention: around 2.1% of the GDP per year for the United States, and 3.6% for Latin America (Bourguignon, 1999 and Londono and Guerrero, 1999);
– Considering monetary costs related to property crime, number rises to 2.6% for the US and 5.1% for Latin America (Bourguignon, 1999).
1. Crime and Welfare• Welfare loss from injuries and increased mortality:
– Recent estimation of the welfare value from gains in life expectancy are quantitatively very important.
– For violence, increased mortality has been shown to represent welfare loss of the same order of magnitude of material costs (Soares, 2006): 1 year of life expectancy associated with a yearly social cost of 3.8% of GDP.
– Colombia lost 2.2 expected years of life to violence: social loss analogous to permanent decline of 9.7% of yearly income, number for the US would be only 0.9% (Soares, 2006).
1. Crime and WelfarePresent Value of Social Cost of Violence from Reduced Life Expectancy (% of GDP), 1990s
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
CO
LOM
BIA
PH
ILIP
PIN
ES
VE
NE
ZU
ELA
CH
ILE
EL
SA
LVA
DO
RB
ELI
ZE
SU
RIN
AM
EM
EX
ICO
BR
AZ
ILR
US
SIA
N F
ED
ER
AT
ION
PU
ER
TO
RIC
OK
AZ
AK
ST
AN
BA
HA
MA
SC
RO
AT
IALA
TV
IAA
RG
EN
TIN
AS
AIN
T K
ITT
S A
ND
NE
VIS
ES
TO
NIA
UK
RA
INE
KY
RG
YZ
ST
AN
AZ
ER
BA
IJA
NT
AJI
KIS
TA
NU
NIT
ED
ST
AT
ES
BE
LAR
US
EC
UA
DO
RIS
RA
EL
CO
ST
A R
ICA
TR
INID
AD
AN
D T
OB
AG
OR
EP
UB
LIC
OF
MO
LDO
VA
LIT
HU
AN
IAT
UR
KM
EN
IST
AN
ALB
AN
IAB
AR
BA
DO
SP
OR
TU
GA
LC
UB
AK
UW
AIT
GR
EN
AD
AS
ING
AP
OR
EU
ZB
EK
IST
AN
AR
ME
NIA
GE
OR
GIA
FIN
LAN
DS
WE
DE
NP
OLA
ND
FR
AN
CE
UR
UG
UA
YU
NIT
ED
KIN
GD
OM
LUX
EM
BO
UR
GC
ZE
CH
RE
PU
BLI
CB
ELG
IUM
RE
PU
BLI
C O
F K
OR
EA
SLO
VA
K R
EP
UB
LIC
BU
LGA
RIA
AU
ST
RA
LIA
NE
W Z
EA
LAN
DH
ON
G K
ON
GC
AN
AD
AH
UN
GA
RY
GE
RM
AN
YIC
ELA
ND
RO
MA
NIA
MA
UR
ITIU
SM
ALT
AJA
PA
NS
LOV
EN
IAIT
ALY
MA
CE
DO
NIA
NO
RW
AY
NE
TH
ER
LAN
DS
AU
ST
RIA
IRE
LAN
DG
RE
EC
ES
PA
IN
> 250%
1. Crime and Welfare• Indirect consequences of reduced length of life expectancy:
– Changes in behavior due to shorter planning horizon: reduced incentives to take actions that generate long-term benefits and short-term costs;
– Decreased investments in human capital and health, reduced savings and investments in physical capital, and possibly reduced growth.
– Link from mortality to investment in human capital and growth through fertility (Lorentzen, McMillan, and Wacziarg, 2006, Kalemli-Ozcan, 2006).
– Connection leads to negative correlation between mortality and investment in human and physical capital, and can be a source of poverty traps.
1. Crime and Welfare• Intangible effects for the labor market and business climate:
– Deterioration of productivity, consumption, and labor force may constitute major part of Latin American’s cost: 7.1% of GDP according to Londono and Guerrero (1999).
– Crime has perverse effects on economic efficiency, reducing investment and employment in poor urban Colombian communities (Gaviria and Velez, 2002).
– In Brazil, 52% of managers rank crime as a major business constraint (World Bank’s Investment Climate Survey).
– But these dimensions are conceptually less clear and difficult to measure in a straightforward way.
1. Crime and Welfare
Social Cost of Violence (yearly cost as % GDP), Latin America and US, 1990s
mortality increase
public secur.
justice system
private prevent.
opportunity cost of
incarcerat.
monet. costs (medical,
etc.)growth
Sub-total
+ intangibles from Londono and Guerrero
(1999)
Latin America
2.0 1.1 0.5 1.4 0.1 0.6 0.1 5.8 12.9
US 0.9 0.5 1.3 0.6 0.6 0.2 0.0 4.1 -
2. Latin America• International comparisons of crime have to deal
with measurement error in crime rates.
• Underreporting in official data is related to institutional development (Soares, 2004) may bias conclusions from cross-country comparisons.
• Comparing victimization data and official records, on can estimate the reporting rate (fraction of crimes reported to authorities).
2. Latin AmericaIncome per Capita and Reporting Rate of Thefts, Cross-section of Coutnries, 1990s
R2 = 0.6523
-8
-7
-6
-5
-4
-3
-2
-1
0
6 6.5 7 7.5 8 8.5 9 9.5 10 10.5 11
ln(income per capita)
ln(r
ep
ort
ing
ra
te o
f th
eft
s)
2. Latin America• Ignoring this problem can lead to wrong conclusion
in terms of the correlation between various variables and development.
• But victimization surveys are available only for some countries and few periods of time.
• Use information from number of deaths due to violence as the best available alternative for analyzing the evolution of crime through time.
2. Latin AmericaCrime Rates (%) from Victimization Data (ICVS), 1990s
0.0
5.0
10.0
15.0
20.0
25.0
30.0
35.0
40.0
45.0
50.0
Latin America Africa Asia FormerCommunist
North America Oceania WesternEurope
Burglary Thefts Contact Crimes Any Crime
2. Latin AmericaCrime Rates (%) from Victimization Data (ICVS), LA Countries, 1990s
0.0
10.0
20.0
30.0
40.0
50.0
60.0
ARGENTINA BOLIVIA BRAZIL COLOMBIA COSTA RICA PARAGUAY
Burglary Thefts Contact Crimes Any Crime
2. Latin America• High crime rates in the region hide considerable
cross-country heterogeneity.
• Even more so when we look at evolution of death due to violence through time:
– One group with increasing trend;
– Another with stable or declining trend.
2. Latin AmericaMortality Rate by Violence, Selected Latin American Countries, Rising Trends
0
10
20
30
40
50
60
70
80
90
1979 1981 1983 1985 1987 1989 1991 1993 1995 1997 1999 2001 2003
de
ath
s to
vio
len
ce (
pe
r 1
00
,00
0)
ARGENTINA BRAZIL COLOMBIA VENEZUELA
2. Latin AmericaMortality Rate by Violence, Selected Latin American Countries, Declining and Stable Trends
0
5
10
15
20
25
30
35
40
45
50
1979 1981 1983 1985 1987 1989 1991 1993 1995 1997 1999 2001 2003
de
ath
s to
vio
len
ce (
pe
r 1
00
,00
0)
CHILE COSTA RICA MEXICOSources: Brazil: Datasus; Rest: WHO
2. Latin America• High crime rates in Latin America span various
different types of crime and are not artifact of the particular statistics used.
• What can explain this pattern?
• Why some countries have been successful at maintaining low violence and others at reducing it, while some have seen increasing violence?
3. Candidate Explanations• Hypotheses can be classified in two groups:
– Socioeconomic conditions conducive to an environment where crime is attractive to large fraction of the population;
– Government actions targeted at repression of criminal activities.
• From this interaction of forces – supply of potential criminals vs. repressive measures – an equilibrium level violence emerges.
3. Candidate Explanations• Concentrate discussion on 7 Latin American
countries and a group of comparison countries:
– Latin American countries: Argentina, Brazil, Chile, Colombia, Costa Rica, Mexico, and Venezuela ;
– Comparison group: Japan, US, South Korea, Spain, Sweden and Russia.
3. Candidate Explanations
• Socioeconomic conditions:
– Inequality;
– Growth;
– Age structure of the population.
3. Candidate Explanations• Repressive policies:
– Incarceration of offenders;
– Harsher penalties;
– Large police presence;
– Effective judicial systems;
– Respect to the law and a clean and efficient government apparatus.
3. Candidate Explanations
Socioeconomic Conditions and Repressive Policies
Violence Growth Gini Pop 15-29
Rule of Law
Police Judges Incarcer.
2000 1980-2000 2000 2000 2000 2000 2000 2001
Selected LA Countries 28.4 0.7% 54.2 27% -0.04 252 7.5 139
Comparison Group 12.4 2.2% 35.2 22% 1.0 398 15.2 282
3. Candidate ExplanationsGini Index, Selected Latin American Countries and Comparison Group
0
10
20
30
40
50
60
Argentina Brazil Chile Colombia CostaRica
Mexico Venezuela Japan Korea Russia Spain Sweden UnitedStates
Gin
i In
de
x
3. Candidate ExplanationsYoung Population, Selected Latin American Countries and Comparison Group
0.00
0.05
0.10
0.15
0.20
0.25
0.30
Argentina Brazil Chile Colombia CostaRica
Mexico Venezuela Japan Korea Russia Spain Sweden UnitedStates
% P
op
15
-29
3. Candidate ExplanationsPolice, Selected Latin American Countries and Comparison Group
0
100
200
300
400
500
600
Argentina Brazil Chile Colombia CostaRica
Mexico Venezuela Japan Korea Russia Spain Sweden UnitedStates
po
lice
pe
r 1
00
,00
0 in
ha
b
1,200
3. Candidate ExplanationsIncarceration, Selected Latin American Countries and Comparison Group
0
100
200
300
400
500
600
700
Argentina Brazil Chile Colombia CostaRica
Mexico Venezuela Japan Korea Russia Spain Sweden UnitedStates
inca
rece
ratio
n p
er
10
0,0
00
inh
ab
3. Candidate Explanations• High crime rates in Latin America do not seem that
surprising after all:
– Economic and demographic factors put a large fraction of the population at the margin of engaging in criminal activities;
– At the same time, policies toward repression of crime and violence are timid and likely ineffective.
• Is this enough to explain the observed differences?
3. Candidate Explanations• Empirical literature offers estimates of effects on crime:
– Incarceration: Levitt (1996);
– Police: Levitt (2002);
– Fraction of young population: Levitt (1999);
– Inequality: Fajnzylber, Lederman, and Loayza (2002b);
– Growth: Fajnzylber, Lederman, and Loayza (2002b).
• Taking these seriously, how much can we explain?
3. Candidate ExplanationsResponse of Violence in L.A. if Variables were set to Average of Comparison
Group
Variable Adjusted
Partial Effect
Cumulative EffectCounterfactual
Level in L.A.% reduction from 28.4
incarceration rate 17.3 39% 17.3
police 21.2 25% 12.9
fraction of young population 26.1 8% 11.9
inequality (gini) 20.3 29% 8.5
growth (income p.c.) 27.3 4% 8.4
3. Candidate Explanations• Violence in Latin America is not exceptionally high,
given socioeconomic conditions and repressive policies, and what is known about their effects.
• Russia is an outlier within the comparison group; if variables in Latin America were set to average of comparison group excluding Russia, violence would still fall to 14.1 (50% reduction).
3. Candidate Explanations• Quantitative roles of inequality, incarceration rates, and
police are the most important.
– Incarceration rates and number of policemen are policy variables directly under the control of the government.
– Inequality is an outcome variable that changes only very slowly through time (Deininger and Squire, 1996).
• Stronger policies in relation to incarceration and policing seem to be the most obvious immediate choice available.
4. Concluding Remarks• Quantitative exercise brings implicit idea that the effectiveness of
policies will be transported. This is obviously not the case.
• Effectiveness of any given intervention will depend on the way it is implemented and on the institutional context.
• Discussion also leaves out many relevant dimensions: police technology and training, effectiveness of judicial system, social norms related to violence, and interactions between citizen and State.
• Still, specific experiences show that it is possible to bring these together in an effective way: Bogotá, São Paulo, etc.