do human rights trials after transition make difference? hunjoon kim university of minnesota...

1
Do Human Rights Trials after Transition Make Difference? Hunjoon Kim University of Minnesota Abstract This research asks whether human rights trials have the capacity for improving human rights protection in transitional societies. The new dataset on human rights trials in the world (Kim; Sikkink and Walling) were used. Two-stage probit models were used to examine the possible endogenous relationship between domestic human rights trials and human rights practices. To model the spatial dependence across states, spatio- temporal lag variables were included. Contrary to the recent arguments made by realist scholars (Vinjamuri and Snyder 2003/4), human rights trials after transition lead to improvement in human rights protection. Moreover, both human rights trials and improvement turned out to have positive impact on other states within the same region, which is a possible sign of regional diffusion. Research Question with Political Transition and Human Rights Trials Background Since 1974, 100 states have experienced transition from nondemocratic regime to democracy or armed conflict to peace. By 2004, 40 states have used domestic transitional human rights trials to address past human rights violations. Research Question Do human rights trials improve human rights situation in countries that have gone through political transition? Methodology Two-Stage Probit Model 1. Estimate two reduced-form equations for (1) human rights trials and (2) human rights improvements. 2. Compute the linear predictors of two dependent variables and substitute them for the endogenous regressors in the structural equations. 3. Correct the bias in standard errors. Data Time-Series-Cross-Section Data with Binary Dependent Variables (BTSCS) Endogenous Dependent Variables DV1: Human Rights Trials (HRTi,t) Use of human rights trials in state i at year t (Source: Kim; Sikkink & Walling) DV2: Human Rights Improvement (HRIi,t) Improvement in the levels of human rights protection in state i at year t (Source: The Cingranelli-Richards Human Rights Database) Independent and Control Variables Table 1. Model Specification The symbol indicates that a variable is included in an equation; the symbol × indicates exclusion of a variable from an equation. Methodology (cont.) Complications and Solutions Findings and Result Findings and Conclusion Domestic transitional human rights trials improve human rights protection in transitional societies. However, improvement in human rights protection does not necessarily lead to the frequent use of trials. The probability of using trials increases if countries with a same religion in the continent used trials in the previous year. Also, the probability of experiencing improvement in human right protection increases if states within an immediate region experienced improvement in the previous year. (Diffusion effect) States are more likely to experience the improvement in human rights protection in 2~3 years after the last improvement. On the contrary, states are most likely to use transitional human rights trials immediately after their political transition or the previous use of trials. The chance of using trials decrease as years since transition or years since the last use elapse. C ountries w ith Transition C outries w ith Hum an R ights Trials 0 20 40 60 80 100 C um ulative N um ber 1974 1977 1980 1983 1986 1989 1992 1995 1998 2001 2004 Year Exogenous Variables EQ 1 HRT EQ 2 HRI Democracy Past level of human rights protection British legal tradition Economic standing Economic development Truth commission experience Regional HRT precedents at year t-1 × Commitment to international norms × Domestic human rights NGOs × Types of transition × Total population × Population growth × Involvement in international war × Involvement in civil war × Regional HRI precedents at year t-1 × Complications Solutions Inefficient and Incorrect Standard Errors Asymptotically correct covariance matrix (Maddala, 1983) Huber’s (1967) robust standard errors (clustering observations by states) Duration Dependence (Temporal correlation) Cubic polynomial – t, t 2 , t 3 (Carter & Signorino, 2006) Spatial Correlation Inclusion of the regional precedents variable at year t-1 as the spatio- temporal lag variable. Precedents(HRT) = W 1 · HRT (NT×1) (NT×NT) (NT×1) W1: spatial weights matrix1 Membership in the same continent and religion w 1ij =1, if state i at year t and state j at year t-1 share common continent and religion w 1ij =0, otherwise Precedents(HRI) = W 2 · HRI (NT×1) (NT×NT) (NT×1) W2: spatial weights matrix2 Membership in the same region (UN subregion) w 2ij =1, if state i at year t and state j at year t-1 share common region (UN subregion) w 2ij =0, otherwise Event Dependence (repeated spells) Inclusion of the number of previous events (sequence or count variable – n) as variables - n, n 2 , n 3 Heterogeneity Sample selection (including only country with political transition) Endogeneity Two-stage probit model 0 .1 .2 .3 .4 .5 .6 .7 .8 Probability (H R I= 1) 0 1 2 3 4 5 6 7 Yearsince lastH R I H um an R ights Im provem ent 0 .1 .2 .3 .4 .5 .6 .7 .8 P robability (H R T = 1) 0 2 4 6 8 10 12 14 Yearsince lastH R T 95% confidence interval H azard R ates H um an R ights Trials D uration D ependence

Upload: elinor-lloyd

Post on 29-Jan-2016

215 views

Category:

Documents


2 download

TRANSCRIPT

Page 1: Do Human Rights Trials after Transition Make Difference? Hunjoon Kim University of Minnesota Abstract This research asks whether human rights trials have

Do Human Rights Trials after Transition Make Difference?Hunjoon Kim

University of Minnesota

AbstractThis research asks whether human rights trials have the capacity for improving human rights protection in transitional societies. The new dataset on human rights trials in the world (Kim; Sikkink and Walling) were used. Two-stage probit models were used to examine the possible endogenous relationship between domestic human rights trials and human rights practices. To model the spatial dependence across states, spatio-temporal lag variables were included. Contrary to the recent arguments made by realist scholars (Vinjamuri and Snyder 2003/4), human rights trials after transition lead to improvement in human rights protection. Moreover, both human rights trials and improvement turned out to have positive impact on other states within the same region, which is a possible sign of regional diffusion.

Research Question

Figure 1. Cumulative Number of States with Political Transition and Human Rights Trials

Background

Since 1974, 100 states have experienced transition from nondemocratic regime to democracy or armed conflict to peace.

By 2004, 40 states have used domestic transitional human rights trials to address past human rights violations.

Research Question

Do human rights trials improve human rights situation in countries that have gone through political transition?

Methodology

Two-Stage Probit Model

1. Estimate two reduced-form equations for (1) human rights trials and (2) human rights improvements.

2. Compute the linear predictors of two dependent variables and substitute them for the endogenous regressors in the structural equations.

3. Correct the bias in standard errors.

DataTime-Series-Cross-Section Data with Binary Dependent Variables (BTSCS)

Endogenous Dependent Variables

DV1: Human Rights Trials (HRTi,t)

Use of human rights trials in state i at year t

(Source: Kim; Sikkink & Walling)

DV2: Human Rights Improvement (HRIi,t)

Improvement in the levels of human rights protection in state i at year t (Source: The Cingranelli-Richards Human Rights Database)

Independent and Control Variables

Table 1. Model Specification

The symbol ○ indicates that a variable is included in an equation; the symbol × indicates exclusion of a variable from an equation.

Methodology (cont.)Complications and Solutions

Findings and

Result

Findings and ConclusionDomestic transitional human rights trials improve human rights protection in transitional societies. However, improvement in human rights protection does not necessarily lead to the frequent use of trials.

The probability of using trials increases if countries with a same religion in the continent used trials in the previous year. Also, the probability of experiencing improvement in human right protection increases if states within an immediate region experienced improvement in the previous year. (Diffusion effect)

States are more likely to experience the improvement in

human rights protection in 2~3 years after the last

improvement. On the contrary, states are most likely to use

transitional human rights trials immediately after their

political transition or the previous use of trials. The chance

of using trials decrease as years since transition or years

since the last use elapse.

Countries with Transition

Coutries with Human Rights Trials

0

20

40

60

80

100

Cu

mu

lativ

e N

umbe

r

1974 1977 1980 1983 1986 1989 1992 1995 1998 2001 2004Year

Exogenous Variables

EQ 1

HRT

EQ 2

HRI

Democracy ○ ○

Past level of human rights protection ○ ○

British legal tradition ○ ○

Economic standing ○ ○

Economic development ○ ○

Truth commission experience ○ ○

Regional HRT precedents at year t-1 ○ ×

Commitment to international norms ○ ×

Domestic human rights NGOs ○ ×

Types of transition ○ ×

Total population × ○

Population growth × ○

Involvement in international war × ○

Involvement in civil war × ○

Regional HRI precedents at year t-1 × ○

Complications Solutions

Inefficient and Incorrect Standard Errors

Asymptotically correct covariance matrix (Maddala, 1983)

Huber’s (1967) robust standard errors (clustering observations by states)

Duration Dependence

(Temporal correlation)

Cubic polynomial – t, t2, t3

(Carter & Signorino, 2006)

Spatial Correlation Inclusion of the regional precedents variable at year t-1 as the spatio-temporal lag variable.

Precedents(HRT) = W1 · HRT (NT×1) (NT×NT) (NT×1)

W1: spatial weights matrix1

Membership in the same continent and religion

w1ij=1, if state i at year t and state j at year t-1 share common continent and religion w1ij=0, otherwise

Precedents(HRI) = W2 · HRI (NT×1) (NT×NT) (NT×1)

W2: spatial weights matrix2

Membership in the same region (UN subregion)

w2ij=1, if state i at year t and state j at year t-1 share common region (UN subregion) w2ij=0, otherwise

Event Dependence

(repeated spells)

Inclusion of the number of previous events (sequence or count variable – n) as variables - n, n2, n3

Heterogeneity Sample selection (including only country with political transition)

Endogeneity Two-stage probit model

0

.1

.2

.3

.4

.5

.6

.7

.8

Pro

bab

ility

(H

RI

= 1

)

0 1 2 3 4 5 6 7

Year since last HRI

Human Rights Improvement

0

.1

.2

.3

.4

.5

.6

.7

.8

Pro

bab

ility

(H

RT

= 1

)

0 2 4 6 8 10 12 14

Year since last HRT

95% confidence interval

Hazard Rates

Human Rights Trials

Duration Dependence