Samaca, Renata; Correa, Lorena (2014). 'Do people in extreme poverty in Colombia remains poor? Assessing the Colombian National Policy to overcome Extreme Poverty' Paper presented at the annual conference of the HDCA, 2-5 September 2014, Athens, Greece.


Over the last decade, the extreme Poverty in Colombia has decreased. In 2002 the percentage of the population living on less than one dollar a day was 33%. In 2012 this proportion declined to 22%. This reduction is due to a set of policies that the government has implemented in order to strengthen the National System of Social Protection, in particular, since 2007 it have been implement the Unidos Network as the National Strategy to overcome extreme poverty. In this way, the main objective of this proposal is to evaluate the social mobility of the households during the intervention of this strategy, and determine the main factors that allow to increase this mobility and to be more effective in the struggle against poverty.

Unidos is a network that gather public and private institutions in order to assure that the poorest households can access to social programs. To do this the network was designed with two main components: Family Support and Supply Management and preferential access. The first one identify the household's demands, and the second one target the social services to poorest population. The graduation model of this policy ensure that households are no longer poor, and in order to do this it is calculated and assessed the two main poverty measurements in Colombia: the monetary index (DANE – DNP, 2012) and the Multidimensional Poverty Index -  MPI (Alkire- Foster, 2011) , (Angulo et all, 2013).  Thereby only households that exceed the cutoff points of these two indicators graduate from the strategy. 

According to Behrman (2000) social mobility can be define as 'movements by specific entities between periods in socioeconomic status indicators', these entities can be the same individual in different periods of time, or across individual of different generations.  Although social mobility is related with equality of opportunities, is not always true that this equality implies higher social mobility (Galiani, 2006). In 2007, Galiani explains that 'analyzing the determinants of mobility involves exploring the channels through which offspring's income is correlated to its parents: inherited bequest, education, formal rules, skills, opportunities, working spirit, among others'. Nevertheless, there are others determinants of social mobility like equal opportunities, meritocracy, human capital accumulation, politico-economic considerations and beliefs.

Azevedo y Bouillon (2009) review evidence about social mobility in Latin America. The authors conclude many things. First, there seems to be high levels of immobility at the lower and upper tails of the income distribution, that is, there is social immobility in poorest people and richest people. Second conclusion is that education mobility and access to higher education are the main determinants of social mobility. Third, the labor market dynamics alter the level of social mobility as the returns of human capital vary with changes in the supply and demand for certain groups of workers. Finally, the authors found that the urbanization process and the increased opportunities for migration from poorer areas should promote higher mobility.

Latin-American countries are traditionally characterized by low social mobility, and in this context, Colombia is not the exception to this rule. Angulo, Azevedo, Gaviria and Páez research (2012) assess inter-generational social mobility in term of year of education, wealth index and household assets, and they found that it is quite low in the country.


Equally Ferreira and Meléndez (2012) found that, despite there have been some improvements, inequality in Colombia is very high in absolute terms. Their research found that inequality in income is determined by circumstances at birth and not by achievement or effort. Parental schooling is the variable with the largest correlation with individual income in the adult lives of their children. People born in small municipalities, rural areas and Pacific and Atlantic regions are also at disadvantage.


The data set and Methodology

The main information is collected in two surveys: the base line, gathering at the beginning of the intervention processes, and the graduation survey.  Despite these two surveys are designed in order to evaluate household performance, they are not perfectly comparable. Thus, a particular issue of this project is to design an income index and a multidimensional poverty index, which can be comparable in the two moments of time. For the first one (monetary) is necessary to adjust an imputation model (i.e Daza & Franco, 2009) that allow us to determine the income changes. For the second one we have to look for comparable variables across the surveys in order to assess the evolution of the household conditions in different dimensions.

Once the indicators are designed, household are classified according to relationship with the cutoffs points that determine extreme poverty and poverty in each indicator. Thus we are going to have nine groups that gather the different combinations between multidimensional and income poverty.

Since our main purpose is to identify if we have social mobility during the intervention, we  estimate the mainstream model in social mobility, which is determined by first–order Markov model. In this way, we assess the role of different dimensions (education, laboral conditions, houssing conditions and childhood conditions) in the social movility process, and therefore give some elements to improve the policiy design in extrem poverty.