This paper aims to estimate poverty lines for the department of Cauca in Colombia and compares them with the figures officially established through an innovative microeconometric methodology. It starts from a set of demand equations based on the Linear Expenditure Systems (LES) and then the poverty thresholds are estimated through a set of econometric models in two scenarios: knowing the minimum food expenditure and assuming the minimum expenditure savings. After resolving problems of identification, endogeneity and sample bias, the results show that indeed poverty thresholds for Cauca, regardless of the chosen modeling, are always undervalued in official data, which distorts the social and economic policy to help reduce the problem in this department.