Loaiza Quintero, Osmar Leandro; Muñetón Santa, Guberney; Vanegas, Juan Gabriel; Bedoya Marulanda, John Fredy (2014). 'The relantionship between poverty and armed civil confltict: the case of Antioquia, Colombia' Paper presented at the annual conference of the HDCA, 2-5 September 2014, Athens, Greece.

Fighting poverty has become a main concern of development economics and of societies in general. However, now it is widely recognized that economic growth and re-distributional policies all help but not suffice to overcome poverty, as they put aside other elements of personal well-being. Poverty must be understood as a multidimensional issue, in which income is just an element among many others that help build a dignifying lifestyle. As Sen (1985, 1996, 2000) suggests, life quality can be seen through an approach in which it is judged through the capabilities a person has to achieve alternative functionings, that can be interpreted as a vector which lists all things that person regards as worth 'doing' or 'being'.


From this lens, poverty is seen as a problem of capability-deprivation. That is, poverty is the result of a lack of choice, which hinders functioning achievement due to a restricted capability set. It is plainly obvious that the capability approach pioneered by Sen calls for a multidimensional measurement of poverty. Nevertheless, the approach prioritizes some deprivations, based on some common values shared by society which determines the kind of needs it considers completely regrettable not to meet, i.e., to be well-nourished, avoid early death, to get education, etc. As deprivations restrict choice, and hence liberty, development is not regarded simply as a matter of material well-being, but as the process of widening liberties as a means to extinguish capability-deprivations. In this way any person could have the opportunity and ability to carry on a valuable life.

In that vein, violence is a phenomenon that highly restricts the ability of societies to overcome the deprivations that affect their members. Indeed, some voices have emerged that talk about a 'violence trap', instead of a 'poverty trap', as an obstacle that hampers the improvement of living conditions (The Economist, 2011). Usually it was said that individuals lacked the material basis to carry on a good life and the States lacked the resources to provide the services and investments required in a context conducive to economic development. Nevertheless, there´s clear cut evidence that portrays violence as the key obstacle, since countries free of armed conflicts are gradually finding a route out of poverty (World Bank, 2011), whereas poverty is concentrating in countries with civil conflicts, ethnic confrontations or organized crime (The Economist, 2011). This paper then seeks to identify the way armed conflict generates and reproduces poverty, putting aside a longstanding although controversial hypothesis which traces violence back to poverty, namely, which states poverty generates violence. The case of Antioquia, in Colombia, will provide the setting to study this relationship. Antioquia showcases areas of widespread poverty and also has a history of armed conflict that has put indelible marks on almost its entire population. Antioquia has topped the list of territories in Colombia with the highest numbers of violence victims and violence perpetrators. This paper then tries to study the effect of violence on poverty in Antioquia, while measuring this through a multidimensional poverty index that overcomes the limitations of traditional poverty measurements implemented in Colombia (like poverty lines).

This study uses information about armed actions by both illegal and legal groups in Antioquia. Poverty will be measured for each municipality of Antioquia using the multidimensional poverty index, following the methodology adopted by the National Planning Department (DNP) and pioneered by Sabina Alkire and James Foster under the Oxford Poverty and Human Development Initiative (OPHI). The source of information to implement this poverty measure will be a government database called SISBÉN (cross-section 2012), which is used in Colombia as an instrument to target social assistance programs. Then we will provide for each municipality in Antioquia a poverty measure with international standing and closer to the capability approach of Sen.

The spatial distribution of this poverty measurement will be analyzed through exploratory spatial data analysis, emphasizing the univariate and bivariate local Moran index. The bivariate Moran index allows to relate the poverty measure to an armed conflict variable. Moreover, the estimated multidimensional poverty index will be related to armed conflict data and a set of control variables in a linear regression framework. The control variables will comprise economic and institutional information about Antioquia's municipalities. Also, the multidimensional poverty index will be stratified according to the amount of deprivations found in a household. Each strata then will characterize a proportion of population according to poverty intensity, as proxied by the amount of unmet needs. Thus, the stratified version of the poverty measure leads to a multiple equations model (an equation for each strata) which will be estimated using the SUR estimator, which is a generalization of the generalized least squares estimation technique. Consquently, the linear regression framework will allow us to analyze the effect of violence on poverty incidence and intensity in Antioquia.