TUNESS Chart of the Week (TCW), Monday March 18, 2013
As we commemorate the 40th day since the assassination of the Tunisian opposition leader Chokri Belaid, we study this week the political assassinations of ruling or opposition leaders in the world which have been a recurring event throughout past decades and have very often changed the course of history or reshaped the political scene in the country or the region where it occurred. Overall, our findings as we explain below, confirm the empirical results of past contributions in the literature and reiterate the idea that given the current circumstances, Tunisia remains highly vulnerable to those kind of unlawful acts.
Jones and Olken (2007, 8) have recently noted that “ the annual rate of assassinations increased in the late 19th and early 20th century, decreased substantially during the 1940s and has been at a relatively high levels since 1950”. In this work we attempt to understand the underlying factors that increase the likelihood of those extrajudicial killing. Despite some theoretical and empirical attempts in the literature, scholars have not yet reached a firm conclusion on the set of universal variables that explain the occurrence of those events. The latter, Frey(2007, 3) ascertain: “ were undertaken for many different reasons ranging from political and ideological differences, to attract media attention, to the work of mentally deranged persons”.
The present short contribution is another attempt to shed some lights on this phenomenon. We estimate both Probit and Logit models to determine the role of certain political, institutional and economic factors in increasing the likelihood of political and/orextrajudicial killing eventssince 1980 in more than 170 countries, as defined in Cingranelli and Richards’(2010) Human Rights Database. Our dependent variable is coded 1 for a particular year if a given country experiences this event and 0 otherwise.
The Chart above illustrates the main finding of this study. The likelihood of extrajudicial killings is higher in countries where the justice system is subject to the control of certain branches of the local government. This result is intuitive and comes with no surprise since in countries where the rule of law are not enforced, certain violent groups might consider itself immune from the authority of justice. Although it appears not significant as shown in the regression table, the influence of the polity of the executive party (right or left) relative to the base (centre) seems to point toward a higherprobability of those events under right wing governments more than left wing governments. The finding of this research also speaks to previous conclusions in the literature where the control of local authorities over the freedom of speech and media were found to significantly increase the likelihood of those assassinations.
To investigate the effect of internal development on the likelihood of occurrence of those killings, we consider two universal measures of human developments namely, Education and GDP per capita. The results shown in table above indicate that while the years of education appears to have a significant negative impact across all specifications of the model over the probability of occurrence of those violent acts, the influence of GDP per capita is found to be insignificant. This result suggests that over the period under study those acts have taken place in wide array of societies including both developing and advances countries. Additionally, we find strong empirical support to our initial hypothesis that corrupt countries are more likely to experience extrajudicial killings relative to less corrupt countries. Conversely, we did not find strong statistical evidence about the effect of long standing political establishment on the likelihood of those assassinations which seems to ascertain that the latter have occurred under both young and long lasting ruling parties.
We further test the hypothesis commonly found in the literature that extrajudicial killings are more likely to take place during years of election where multiple political and social confrontations between certain groups of the population and the ruling governments are more frequent.We have thus included an additional control dummy variable equal to 1 for the years where a legislative or a presidential election took place and 0 otherwise. Despite the positive sign associated with this relationship suggesting that those events tend indeed to occur more frequently in election years, our results reveal that such a relationship is not significant at the 5 % level.
This week’s TCWhas examined the factors influencing the likelihood of extrajudicial killings. Using data from multiples databases we estimated several specifications of logistic regressions to identify some of those factors.
Study conducted by Bechir Bouzid, PhD , TUNESS Research Team
References:
Cingranelli, D. and Richards, D. (2010): “The Cingranelli and Richards (CIRI) Human Rights Data Project”, Human Rights Quarterly, Vol(32), No (2)
Frey, B. (2007):“Why Kill Politicians? A Rational Choice Analysis of Political Assassinations”, Working Papers Series No. 324, University of Zurich.
Jones, B. and Olken, B. (2007):“ Hit or Miss? The Effect of Assassinations on Institutions and War”, BREAD Working Paper, No. 150.