Detalhes bibliográficos
Ano de defesa: |
2024 |
Autor(a) principal: |
Tecchio, Otávio Luiz |
Orientador(a): |
Ferman, Bruno,
Caetano, Carolina |
Banca de defesa: |
Não Informado pela instituição |
Tipo de documento: |
Dissertação
|
Tipo de acesso: |
Acesso aberto |
Idioma: |
eng |
Instituição de defesa: |
Não Informado pela instituição
|
Programa de Pós-Graduação: |
Não Informado pela instituição
|
Departamento: |
Não Informado pela instituição
|
País: |
Não Informado pela instituição
|
Palavras-chave em Português: |
|
Palavras-chave em Inglês: |
|
Link de acesso: |
https://hdl.handle.net/10438/35324
|
Resumo: |
This dissertation consists of two independent chapters. The first one is based on the paper Ferman and Tecchio (2023) and the second one is based on joint work with Carolina Caetano and Gregorio Caetano. The first chapter discusses the identification of dynamic effects of an irreversible treatment with a time-invariant binary instrumental variable (IV). For example, in evaluations of the dynamic effects of training programs, where a single lottery determines eligibility. A common approach in these settings is to report per-period IV estimates. Under a dynamic extension of standard IV assumptions, we show that such IV estimators identify a weighted sum of treatment effects for different latent groups and treatment exposures. However, there is the possibility of negative weights. We consider point and partial identification of dynamic treatment effects in this setting under different sets of assumptions. The second chapter considers the identification of causal effects when the treatment is endogenous and has bunching on an extreme of the support of its distribution. Our setting does not assume the existence of instrumental variables or special data structures such as panel. Instead, we show that the identification problem under endogeneity can be connected to the problem of identification of the conditional expectation of a latent variable in a nonparametric censoring model. Since the identification of the conditional expectation has been extensively studied in the literature on censoring models, many approaches can be adapted to our setting. In particular, we focus on the approach proposed in Chen et al. (2005), which enables the identification of the causal effect at the bunching point among the marginal observations with only nonparametric constraints on the relationship between observable and unobservable variables. |