| The purpose of this dissertation is to demonstrate the relevance for Colombia of analyzing the middle segment of the society -- as the core concept -- and socioeconomic polarization due to its linkage with the middle segment of the society, with a multidimensional framework for public policy purposes. Theoretically, I use Sen's multidimensional approach on capabilities and functionings, which has been applied for studying poor populations but not the middle segment of the society. I propose the multidimensional middle class (MMC) to empirically address the middle segment of the society with the relevant dimensions, rather than the middle income population (MIP) or the middle class (MC) as is currently done. Similarly, I propose an index of multidimensional polarization (IMP), rather than income polarization currently in use. The research questions that lead the dissertation are: i. What is the size -- in percentage -- of the Colombian MMC? Thus, is the MMC expanding or shrinking during the analysis period? ii. What is the trend of the Colombian IMP? iii. How does the size of the MMC fluctuate, compared to the IMP trend during the analysis period? Do trends suggest a linkage between both concepts after their multidimensional framed operationalization?. I implemented two methodologies to calculate the MMC and the IMP. To calculate the MMC, I implemented the Alkire-Foster (AF) methodology, and to calculate the IMP, I implemented the Gigliariano-Mosler (GM) methodology. I propose a novel application of these multidimensional methodologies with data available for years 2003, 2008, 2010, 2011, 2013, 2014 and 2015, based on the Colombian Living Standards Measurement Surveys (C-LSMS). The main findings are: i) the size of the MMC has a minimum value of 24.3% in 2003 and a maximum value of 36.5% in 2003. Therefore, excluding the year 2008, the MMC has an upward trend. ii) The IMP does not show a linear trend with income groups or with socio-occupational groups. The IMP with income groups increases from 2003 to 2008, then shows a decreasing trend. The IMP with socio-occupational groups increases from 2003 to 2008 as well; however, its trend from 2008 to 2015 is slightly upward but relatively stable. iii) The MMC and the IMP do not have a perfect inverse relationship; however, an increase in the MMC results in a reduction in the IMP, thus, the MMC and the IMP show time trends with similar patterns of unidimensional analysis. Therefore, I present the inverse relationship between the MMC and the IMP as an empirical regularity that supports the conclusion that societies with large middle segments have lower levels of polarization. |
