The main objective is to estimate the determinants of the unemployment periods in Cali (Colombia) and its Metropolitan Area (MA) between 2012-2014, taking into account periods of 8, 16 and 48 weeks. The quantile regression technique has been used, after having corrected the selection bias. The results for Cali MA match Colombia's recent estimates with regards to the role of education, age, head of household, presence of children under 6 in the household, unpaid activities and the role of women involved. The results also suggest that determinants affect individuals differently for periods of 8, 16 and 48 weeks of unemployment.