Project CEMAPRE internal
| Title | Topics in Mode regression and the bootstrap - 2024 |
| Participants | Paulo Parente (Principal Investigator) |
| Summary | "In this project we are going to continue to work on the project started in 2021 on mode regression and the bootstrap.In 2023 we worked on the following tasks: 1- Paulo Parente showed that the method introduced in the paper by Parente and Smith (2018) holds validity in a broader context than the one originally considered. In their work, Parente and Smith (2018) reveal how the Kernel Block Bootstrap method can be applied to draw inferences about parameters in models defined by moment restrictions. The paper also introduced bootstrap procedures that utilize generalized empirical likelihood implied probabilities for drawing observations. In our project, we aim to demonstrate that the methods proposed by Parente and Smith (2018) allow for testing additional moment constraints and parametric restrictions expressed in a mixed form, as discussed by Gouriéroux and Monfort (1989) and Smith (2011) (Section 5, pp. 1209-1213). We are in the final stages of completing this article, which will soon be submitted to a research journal. 2- João Vieira extended the estimator for the parameters of a mode regression model for panel data with fixed effects to a model with interactive fixed effects (latent factors). Mode regression serves as an alternative to mean regression (least squares estimator) and quantile regression, enabling the estimation of the conditional mode of a random variable in a semi-parametric manner. Although this work remains unfinished, João Vieira intends to complete it by 2024." |