Transcript Lecture_Four-Deterre..

Part I
Strategies to Estimate Deterrence
Part II
Optimization of the Criminal
Justice System
Llad Phillips
1
Testing crime control
Llad Phillips
2
Schematic of the Criminal Justice System
Causes ?
Weak Link
Offense
Rate Per
Capita
Crime Generation
OF = f(CR, SV, CY, SE, MC
Expenditures
Expected
Cost of
Punishment
(detention,
deterrence)
Crime Control
Llad Phillips
CR = g(OF, X)
3
Suicide AS A Proxy For Causes of Homicide
Llad Phillips
4
FBI Vs. Vital Statistics
24
Murder Rate Per 100,000 from FBI Uniform Crime Reports;
Death by Cause, Suicide and Homicide, Vital Statistics from
US Statistical Abstract
20
16
12
8
4
0
1930
1940
1950
1960
1970
1980
1990
2000
2010
MURDER_RATE
VITALSTATS_HOMICIDE
VITALSTATS_SUICIDE
Llad Phillips
5
Homicide Rate & Suicide Rate
Murder Rate Correlated with the Suicide Rate, 1960-2007
11
MURDER_RATE_PER_100K
10
9
8
7
6
5
4
9
10
11
12
13
14
SUICIDE_RATE_PER_100K
Llad Phillips
6
Homicides per 100, 000 = a +
b*suicides per 100,000 + e, note
suicides is significant, explains~76%
OF= f( SE)
Llad Phillips
7
Schematic Model
Controls:
Imprisonment rate
Clearance ratio
Causes
Homicide
Llad Phillips
8
MURDER_RATE_PER_100K
12
10
The story over time
8
6
4
55
60
65
70
75
80
85
90
95
00
05
10
95
00
05
10
05
10
CR_HOMICIDE
100
90
80
70
60
55
60
65
70
75
80
85
90
USIMPRISON_RATE_PER_100K
600
500
400
300
200
100
0
55
Llad Phillips
60
65
70
75
80
85
90
95
00
9
A Control Story: US
_
Clearance ratio for homicide was falling
from 1960 0n
_
_
This could explain the rising homicide rate
from 1965-1975
Imprisonment rate was pretty stable until
1980 when it started rising
_
This could explain the falling homicide rate
from 1995-2009
Llad Phillips
10
US H om icide R ate Per 100,000 VS. C learance Ratio
12
MURDER_RATE
10
8
6
2010
1960
4
2
60
70
80
90
100
CR HOM
Llad Phillips
11
Homicides per 100,000 = a + b*CR + e
note: CR is significant, explains ~8.6%
OF = f(CR)
Llad Phillips
12
US Homicide Per 100,000 VS. Federal and State Prisoners Per 100,000
12
MURDER_RATE
10
8
6
2010
4
1960
2
0
100
200
300
400
500
600
FED_STATE_PRISON
Llad Phillips
13
Homicides per 100,000 =a + b*prisoners
per 100,000 + e (note: prisoners not
significant as the only explanatory variable)
Llad Phillips
14
Homicides per 100,000 = a +b*CR +c*
prisoners per 100,000 + e Note: CR and
Prisoners together are significant, explain
83%
Llad Phillips
15
OF = f(CR, SV)
homicide per 100,000 = a +b*CR +d*fed&state Prisoners per 100,000 + e
12
10
8
2
6
1
4
0
-1
-2
60
65
70
75
Residual
Llad Phillips
80
85
90
Actual
95
00
05
Fitted
16
Homicides per 100, 000 = a + b*Suicides
per 100,000 +c* CR + d* prisoners per
100,000 Note: all3 explanatory variables
are significant & explain 87%
Llad Phillips
17
Homicide Rate = a + b*CR
+c*Prisoners + d*suicide rate + e
12
The error e
Or residual has
a cycle. Indicates
An explanatory 1.5
variable is
1.0
missing. Best to 0.5
Find it. Next best,0.0
Model the error -0.5
10
8
6
4
-1.0
-1.5
60
65
70
75
Residual
Llad Phillips
80
85
Actual
90
95
00
05
Fitted
18
Homicide per 100,000 explained by
suicides, clearance ratio, & prisoners with
error modeled, explains 92.8 %
Llad Phillips
19
Error modeled or cleaned up cleaned up
12
10
8
1.0
6
0.5
4
0.0
-0.5
-1.0
70
75
80
Residual
Llad Phillips
85
90
Actual
95
00
05
Fitted
20
Outline
_
_
_
Human Capital & Other News
Studying for the Midterm
Deterrence:
_
_
Evidence pro
Evidence con
Llad Phillips
21
Human Capital news
Llad Phillips
22
About 60%
Of 9th graders
Get a diploma
somewhere
Llad Phillips
23
The high
Hurdle?
Algebra
Llad Phillips
24
Llad Phillips
25
Llad Phillips
26
Llad Phillips
27
Llad Phillips
28
Llad Phillips
29
Deterrence: conceptual issues
_
_
Controlling for causality
Simultaneity
Llad Phillips
30
Get
Expect
Source:
Llad
Phillips
Report to the Nation on Crime and Justice
31
Schematic of the Criminal Justice System
Causes ?
Control for Causality
Weak Link
Offense
Rate Per
Capita
Crime Generation
Expenditures
Expected
Cost of
Punishment
(detention,
deterrence)
Crime Control
Llad Phillips
32
Schematic of the Criminal Justice System
Causes ?
Weak Link
Offense
Rate Per
Capita
Recognize
Simultaneity
Crime Generation
Expenditures
Expected
Cost of
Punishment
(detention,
deterrence)
Crime Control
Llad Phillips
33
Greening the Earth
_
Greening UCSB
_
Rec-Cen
Llad Phillips
34
Human development Index and
Electricity Use
Llad Phillips
35
Production Function
UN Human Development Index & Electricity Consumption
1
0.9
0.8
0.7
Index
0.6
0.5
0.4
0.3
0.2
0.1
0
0
2000
4000
6000
8000
10000
12000
14000
16000
18000
Annual Kwhr Per capita
Llad Phillips
36
Llad Phillips
37
Llad Phillips
38
Llad Phillips
39
Policy Comment About Economic
Development
_
An Obama Keynesian strategy: invest in
infrastructure
_
Past investments in infrastructure
_
_
_
_
_
Llad Phillips
Canals
Railroads
Paved roads
Airways
?
40
Homicide and Non-negligent Manslaaaughter, Rates Per 100,000
16
California
14
12
USA
10
8
6
4
2
0
1900
1920
1940
HOMICIDECA
Llad Phillips
1960
1980
2000
HOMICIDEUSA
41
CA Homicide Rate Per 100,000 & Misery Rate in %
25
20
15
10
5
0
1900
1920
1940
1960
HOMICIDECA
Llad Phillips
1980
2000
MISERY
42
Causality?
Misery Index
Offense Rate
Mystery Force
Llad Phillips
43
Regress CAINDXPC = a + b*MISERY + e(t)
where e(t) = 0.96*e(t-1) + u(t)
0.04
0.03
0.004
0.02
0.002
0.01
0.000
0.00
-0.002
-0.004
55
60
65
70
75
Residual
Llad Phillips
80
85
Actual
90
95
00
05
Fitted
44
Schematic of the Criminal Justice System
Causes ?
Control for Causality
Weak Link
Offense
Rate Per
Capita
Crime Generation
Expenditures
Expected
Cost of
Punishment
(detention,
deterrence)
Crime Control
Llad Phillips
45
California Prisoners: 1851-1945
10000
8000
6000
4000
2000
0
60
70
80
90
00
10
20
30
40
CAPRISONERS
Llad Phillips
46
detrend = caprisoners - 19.656 - 48.569*time
6000
1930
5000
4000
3000
2000
1000
1900
0
-1000
1851
60
70
80
90
00
10
DETREND
Llad Phillips
20
30
40
1945
47
Regression of CAINDXPC on MISERY and CAPRPC
0.04
0.03
0.006
0.02
0.004
0.002
0.01
0.000
0.00
-0.002
-0.004
55
60
65
70
75
Residual
Llad Phillips
80
85
Actual
90
95
00
05
Fitted
48
Llad Phillips
49
Part I
Strategies to Estimate Deterrence
Llad Phillips
50
Questions About Crime
 Why
is it difficult to empirically
demonstrate the control effect of deterrence
on crime?
 What is the empirical evidence that raises
questions about deterrence?
 What is the empirical evidence that supports
deterrence?
Llad Phillips
51
Evidence Against the Death
Penalty Being a Deterrent
 Contiguous
States
Maine:
no death penalty
Vermont: death penalty
New Hampshire: death penalty
 Little
Variation in the Homicide Rate
Source:
Study by Thorsten Sellin in Hugo
Bedau, The Death Penalty in America
Llad Phillips
52
Isaac Ehrlich Study of the Death
Penalty: 1933-1969
 Homicide
Control
Rate Per Capita
Variables
 probability
of arrest
 probability of conviction given charged
 Probability of execution given conviction
Causal
Variables
 labor
force participation rate
 unemployment rate
 percent population aged 14-24 years
 permanent income
 trend
Llad Phillips
54
Ehrlich Results: Elasticities of
Homicide with respect to Controls
Control
Elasticity
Average Value
of Control
0.90
Prob. of Arrest
-1.6
Prob. of Conviction
Given Charged
Prob. of Execution
Given Convicted
-0.5
0.43
-0.04
0.026
Source: Isaac Ehrlich, “The Deterrent Effect of Capital Punishment
Critique of Ehrlich by Death
Penalty Opponents
 Time
period used: 1933-1968
period
of declining probability of execution
 Ehrlich
did not include probability of
imprisonment given conviction as a control
variable
 Causal variables included are unconvincing
as causes of homicide
Llad Phillips
56
U.S.
Llad Phillips
United States Bureau of Justice Statistics
http://www.ojp.usdoj.gov/bjs/
57
U.S.
United States Bureau of Justice Statistics
http://www.ojp.usdoj.gov/bjs/
Llad Phillips
58
What is the Empirical Evidence
that Supports Deterrence?
 Domestic
violence and police intervention
Experiments
 Traffic
Black Spots
Focused
Llad Phillips
with control groups
enforcement efforts
59
Traffic Black Spots
 Blood
Alley
Highway
 San
Marcos Pass
Highway
Llad Phillips
126
154
60
San Marcos Pass Experiment
 Increase
Highway Patrols
 Increase Arrests
Total
accidents decrease
Injury accidents decrease
Accidents involving drinking under the
influence decrease
Llad Phillips
61
Llad Phillips
62
Los Angeles Traffic Map
Domestic Violence & Police
Intervention
Llad Phillips
64
1993-2005
Llad Phillips
65
Female Victims of Violent Crime, 1973-2003
Llad Phillips
66
Homicides of Intimates, 1976-2005
Llad Phillips
67
Female Victims of Violent Crime
 In
1994
1
homicide for every 23,000 women (12 or
older)
 females
represented 23% of homicide victims in US
 9 out of 10 female victims were murdered by males
1
rape for every 270 women
1 robbery for every 240 women
1 assault for every 29 women
Llad Phillips
68
Victims of Lone Offenders*
Annual Average Numbers
Known
Female
Male
2,715,000
2,019,400
Intimate
1,008,000
143,400
Relative
304,500
122,000
Acquaintance
1,402,500
1,754,000
Stranger
802,300
1,933,100
United States Bureau of Justice Statistics
http://www.ojp.usdoj.gov/bjs/
Llad Phillips
70
Llad Phillips
71
Average Annual Rate of Violent
Victimizations Per 1000 Females
Family Income
Less than $10,000
$10,000 - $14,999
$15,000 - $19,999
$20,000 - $29,999
$30,000 - $49,999
$50,000 or more
Llad Phillips
Total
57
47
42
38
31
25
Intimate
20
13
11
10
5
5
72
Llad Phillips
73
United States Bureau of Justice Statistics
http://www.ojp.usdoj.gov/bjs/
Llad Phillips
74
Llad Phillips
75
Female victimization rates by relationship
Llad Phillips
76
Llad Phillips
77
Domestic Violence in California
http://caag.state.ca.us/
Llad Phillips
78
Domestic Violence Rates in California: 1988-1998
1988: 113.6 per 100.000
1998: 169.9 per 100,000
Llad Phillips
79
Domestic Violence in California
1988: 94% Male Arrests
1998: 83.5% Male Arrests
Llad Phillips
80
Police Intervention with
Experimental Controls
A
911 call from a family member
the
case is randomly assigned for “treatment”
A
police patrol responds and visits the
household
police
calm down the family members
based on the treatment randomly assigned, the
police carry out the sanctions
Llad Phillips
81
Why is Treatment Assigned
Randomly?
 To
control for unknown causal factors
assign
known numbers of cases, for example
equal numbers, to each treatment
 with
this procedure, there should be an even
distribution of difficult cases in each treatment group
Llad Phillips
82
911 call
(characteristics of household Participants unknown)
Random Assignment
code blue
code gold
patrol responds
patrol responds
settles the household
verbally warn the husband
Llad Phillips
settles the household
take the husband to jail
for the night
83
Part II
Optimization of the Criminal
Justice System
Llad Phillips
84
Questions About Statistical
Studies of Deterrence
_
Do we know enough about the factors that
cause crime?
_
_
Can we find variables that will control for
variation in crime generation?
We have better measures for the factors that
control crime than for the factors that cause
crime.
_
Unknown variation in crime generation may
mask the effects of crime control.
Llad Phillips
85
Schematic of the Criminal Justice System
Causes ?
Weak Link
Offense
Rate Per
Capita
Crime Generation
Expenditures
Expected
Cost of
Punishment
(detention,
deterrence)
Crime Control
Llad Phillips
86
Four-Way Diagram: Crime Generation & Crime Control
per capita expenditures on CJS
1
2
3
offense rate per capita
Source: Report to the Nation on Crime and Justice
Causal
factors
control
Source: Report to the Nation on Crime and Justice
Expenditures
per Capita
Crime Control Technology
South Dakota
North Dakota
$100
$0
0
2500 Index crimes
per 100,000 people
Offenses Per Capita
Optimization of the Criminal
Justice System (CJS)
 Minimize
damages to victims plus the costs
of control, subject to the crime control
technology
damages
to victims per capita = loss rate per
offense * offense rate per capita
Costs of control = per capita expenditures on
CJS
Total cost = damages + expenditures
Llad Phillips
91
Expenditures
per Capita
Total cost = expenditures per capita
Crime Control Technology
$200
South Dakota
North Dakota
$100
$0
0
2500 Index crimes
per 100,000 people
Offenses Per Capita
Expenditures
per Capita
Total cost = expenditures per capita
Crime Control Technology
$200
South Dakota
North Dakota
$100
Total cost = damages
to victims
$0
0
2500 Index crimes
per 100,000 people
Offenses Per Capita
5000 Index offenses per 100,000 people = 0.05 per capita
Expenditures
per Capita
Total cost = expenditures per capita
Crime Control Technology
$200
South Dakota
North Dakota
$100
Total cost = damages
to victims
$0
0
0.025
0.050 Offenses Per Capita
Index crimes
per capita
Total cost = $200 per capita = damages to victims = loss rate*0.05
so loss rate = $4,000 per Index Crime in South Dakota
Llad Phillips
94
Cost to Victims in US, 1993
Offense
Robbery
Loss Rate Reported
Offenses
$13,000
659,757
Damages,
Billions $
$8.6
Auto
Theft
Burglary
$4,000
1,561,047
$6.2
$1,500
2,834,808
$4.3
Larceny
$370
7,820,909
$2.4
Total
Source: National Institute of Justice,
$21.5
Victim Costs and Consequences(1996)
Llad Phillips
Source: Phillips: Lecture One
17
Expenditures per capita
Total cost = expenditures per capita
High
Family of Total Cost Curves
$100
Low
2500 Index crimes
per 100,000 people
Llad Phillips
Total cost = damages
to victims
Offenses Per Capita
96
Expenditures
per Capita
Total cost = expenditures per capita
Crime Control Technology
$100
South Dakota
North Dakota
Total cost = damages
to victims
2500 Index crimes
per 100,000 people
Llad Phillips
Offenses Per Capita
97
Application of the Economic
Paradigm
 Specify
the feasible options
the
states of the world: Crime control
technology
 Value
loss
the options
rate per offense
 Optimize
Pick
the lowest cost point on the crime control
technology
Llad Phillips
98
That’s all folks!
Crime Generation
1. variation of offense rate per capita with expected cost
of punishment
2. Shift in the relationship with a change in causal factors
Offense
rate per
capita
crime generation function
Expected cost(severity) of punishment
Crime Generation
1. variation of offense rate per capita with expected cost
of punishment
2. Shift in the relationship with a change in causal factors
Offense
rate per
capita
crime generation function
High causal conditions
Low causal conditions
Expected cost(severity) of punishment
Production Function for the Criminal Justice System (CJS)
1. Variation in expected costs of punishment with
criminal justice system expenditure per capita
Expected
costs of
punishment
production function
Criminal Justice System expenditures per capita
Four-Way Diagram: Crime Generation & Crime Control
per capita expenditures on CJS
offense rate per capita
Crime Generation
expected cost of punishment
Four-Way Diagram: Crime Generation & Crime Control
per capita expenditures on CJS
per capita
expenditures
on CJS
offense rate per capita
Production
Function
Crime Generation
expected cost of punishment
Four-Way Diagram: Crime Generation & Crime Control
per capita expenditures on CJS
square
per capita
expenditures
on CJS
Production
Function
450
offense rate per capita
Crime Generation
expected cost of punishment
Four-Way Diagram: Crime Generation & Crime Control
per capita expenditures on CJS
1
square
per capita
expenditures
on CJS
Production
Function
1
450
offense rate per capita
Crime Generation
expected cost of punishment
Four-Way Diagram: Crime Generation & Crime Control
per capita expenditures on CJS
1
square
per capita
expenditures
on CJS
Production
Function
1
450
offense rate per capita
Crime Generation
expected cost of punishment
Four-Way Diagram: Crime Generation & Crime Control
per capita expenditures on CJS
1
square
per capita
expenditures
on CJS
Production
Function
1
450
offense rate per capita
Crime Generation
expected cost of punishment
Four-Way Diagram: Crime Generation & Crime Control
per capita expenditures on CJS
1
square
per capita
expenditures
on CJS
Production
Function
1
450
offense rate per capita
Crime Generation
expected cost of punishment
Four-Way Diagram: Crime Generation & Crime Control
per capita expenditures on CJS
1
square
2
per capita
expenditures
on CJS
Production
Function
1
2 450
offense rate per capita
Crime Generation
expected cost of punishment
Four-Way Diagram: Crime Generation & Crime Control
per capita expenditures on CJS
1
square
2
per capita
expenditures
on CJS
Production
Function
3
1
2 450
offense rate per capita
Crime Generation
expected cost of punishment
Four-Way Diagram: Crime Generation & Crime Control
per capita expenditures on CJS
1
square
2
per capita
expenditures
on CJS
Production
Function
3
1
2 450
offense rate per capita
Crime Generation
expected cost of punishment
Female Victims of Violent Crime
Llad Phillips
113
Llad Phillips
114
Llad Phillips
115
Llad Phillips
116
Long Swings in the Homicide Rate in the US: 1900-1980
Source: Report to the Nation on Crime and Justice
Long Swings in
The Homicide Rate
United States Bureau of Justice Statistics
http://www.ojp.usdoj.gov/bjs/
Llad Phillips
118
California Homicide Rate Per 100,000: 1952-2003
16
14
12
Rate
10
8
6
4
2
0
1950
1960
1970
1980
1990
2000
2010
Year
Llad Phillips
119
Empirical Study of Certainty and Severity
Murder Rate Regressed on Clearance Ratio and Imprisoment Rate
12
10
8
2
6
1
4
0
-1
-2
1960
1965
1970
1975
1980
Residual
Llad Phillips
1985
1990
Actual
1995
2000
2005
Fitted
120