SUBJECTIVE POVERTY IN BOGOTÁ AND CUNDINAMARCA: A RESPONSE MODEL BINARY USING LASSO

Abstract

This research aims to identify the factors that influence the perception of poverty in Bogota and Cundinamarca. For this we used
data from the Multipurpose Survey in Colombia (2017) and estimated a binary response model using LASSO (least absolute
shrinkage and selection operator). The empirical testing showed that high levels of schooling, sufficient income to cover
minimum expenses, receiving rents, being a housewife, belonging to the special health regime or not receiving state aid or
remittances, decreased the probability of feeling poor. Meanwhile, people were more likely to consider themselves poor if they
presented at least one of the following conditions: they were unemployed, were disabled, belonged to the subsidized regime,
had low levels of schooling, received help from another person or some institution, or their minimum expenses were not covered
by their income. Therefore, to reduce the perception of poverty, public policies in Colombia should aim to create mechanisms
that promote higher education and job stability, instead of increasing subsidies.