Ensaios sobre segregação no mercado de trabalho, avaliação de programas sociais e identificação não-paramétrica

• Autor(a) principal: Foguel, Miguel Nathan
• Data de Publicação: 2010
• Tipo de documento: Tese
• Idioma: por
• Título da fonte: Repositório Institucional da Universidade Federal Fluminense (RIUFF)
• Texto Completo: https://app.uff.br/riuff/handle/1/18748
• topic: Diferenciais de salários; Discriminação; Identificação; Modelos não-paramétricos; Avaliação de programas; Wage differentials; Discrimination; Identification, Non-parametric models; Programme evaluation; CNPQ::CIENCIAS SOCIAIS APLICADAS::ECONOMIA

The main body of this thesis contains three chapters. In chapter 1, we study the effect of gender segregation at the establishment level on the wages of males and females. While the available literature on this issue has used cross-section data, in that chapter we make use of a large panel of matched employer-employee data for Brazil ("Annual Social Information Report" - Relação Anual de Informações Sociais - RAIS). Estimates obtained with longitudinal methods show that the sign of the effect of interest seems to be negative. However, though this result is in line with the cross-section estimates, the longitudinal analysis reveals that the magnitude of the effect is smaller (that is, less negative). This difference in magnitude indicates that unobserved worker- and establishment-specific effects are (negatively) correla- ted with the proportion of females in establishments, implying that the tradicional cross-section estimates are biased. Chapters 2 and 3 of the thesis investigate the non-parametric identification of certain effects of interest when a subset of the covariates are functionally dependent of the other covariates in the model. In a non-parametric setting, this functional dependence is equivalent to the existence of perfect multi-collinearity in a linear regression model. We propose a set of suficient conditions upon which some aspects of the effects of the functionally dependent covariates are identified in a non-parametric fashion. In chapter 2, we study the identification of two effects that are relevant to the area of programme evaluation: (i) the time effect of the programme itself; and (ii) the effect associated with the influence of economic conditions that are external to the programme. Chapter 3 presents a generalisation of the identification analysis of chapter 2 to other areas of application.