Avaliação da inadimplência do IPTU na Cidade de Fortaleza

Abstract

This study aims to predict delinquency in the Urban Property Tax (IPTU) in the Municipality of Fortaleza. The methodology employed was an econometric binary logistic regression model, estimated using cadastral information from taxpayers and properties in Fortaleza for the fiscal years 2022 to 2024. The results indicate that variables such as taxpayer history, type of person, property category, single-installment payment, and income level are intrinsically related to the probability of IPTU delinquency. The model demonstrated strong predictive performance, with an area under the ROC curve of 83% and model specificity above 80%, confirming its ability to distinguish between compliant and delinquent taxpayers. These findings show that it is feasible to use rating models to guide municipal tax collection policies in a more targeted manner, helping to reduce delinquency and strengthen the Municipality’s financial autonomy in the context of the reform of the National Tax System.