MARC보기
LDR00000nam u2200205 4500
001000000436431
00520200228143643
008200131s2019 ||||||||||||||||| ||eng d
020 ▼a 9781088309131
035 ▼a (MiAaPQ)AAI13809075
040 ▼a MiAaPQ ▼c MiAaPQ ▼d 247004
0820 ▼a 330
1001 ▼a Sawada, Masayuki.
24510 ▼a Essays in Theoretical and Applied Econometrics.
260 ▼a [S.l.]: ▼b Yale University., ▼c 2019.
260 1 ▼a Ann Arbor: ▼b ProQuest Dissertations & Theses, ▼c 2019.
300 ▼a 83 p.
500 ▼a Source: Dissertations Abstracts International, Volume: 81-03, Section: A.
500 ▼a Advisor: Vytlacil, Edward.
5021 ▼a Thesis (Ph.D.)--Yale University, 2019.
506 ▼a This item must not be sold to any third party vendors.
506 ▼a This item must not be added to any third party search indexes.
520 ▼a In this dissertation, I study theoretical identification and inference problems on econometrics and apply econometric methodologies to data.In chapter 1, I identify the treatment effect heterogeneity across endogenous post-assignment decisions in the context of randomized experiments. Unlike existing approaches, I do not rely on any instruments or specific experimental design. Instead, I exploit a baseline survey to proxy for control outcomes. An essential requirement is that the proxy outcome is similar to the control outcome in the sense that the underlying rankings of the two outcomes across individuals are identically distributed conditional on the endogenous decisions. In an example of a microcredit experiment, randomly selected villages receive access and promotion to the treatment of borrowing from the microfinance. The random assignment of treatment access may not be an instrumental variable to identify the average treatment effect on treated (ATT) because the treatment access itself may have a direct impact on the outcome. I demonstrate that the degree of bias can be enormous when the take-up probability is relatively small, as the bias takes the form of the direct effect times the probability of no take-up divided by the probability of take-up. I show that the additional proxy variable allows identification of the ATT even when access to treatment has a direct effect on outcomes. I also show identification of the direct effect of the treatment assignment itself.In chapter 2, I propose an estimation procedure for the counterfactual subgroup effect introduced in chapter 1, exploiting the weak convergence results for semi-parametric distribution regression. I show that the weak convergence of the empirical process of the counterfactual distribution function to a tight zero-mean Gaussian process through the functional delta method by showing the counterfactual function is a known Hadamard differentiable mapping of observable conditional distribution functions. I propose an inference procedure for the subgroup mean and quantile effects based on an exchangeable bootstrap. I also show that the proposed procedure is robust to clustered sampling. I apply the proposed estimator to a microcredit experiment in Morocco. The treatment of microcredit access was randomly given to treatment villages, whereas access was wholly forbidden to control villages. Using my estimator, I find a strong positive ATT and a relatively small (and insignificant) positive direct effect. Although the small magnitude of the direct effect is consistent with the findings of the original authors, I find that the size of the IV estimate obtained under the assumption of no direct effect is 2.3 times larger than my estimate. Thus, the small magnitude of the direct effect results in substantial bias in the estimated ATT.
590 ▼a School code: 0265.
650 4 ▼a Economics.
690 ▼a 0501
71020 ▼a Yale University. ▼b Economics.
7730 ▼t Dissertations Abstracts International ▼g 81-03A.
773 ▼t Dissertation Abstract International
790 ▼a 0265
791 ▼a Ph.D.
792 ▼a 2019
793 ▼a English
85640 ▼u http://www.riss.kr/pdu/ddodLink.do?id=T15490563 ▼n KERIS ▼z 이 자료의 원문은 한국교육학술정보원에서 제공합니다.
980 ▼a 202002 ▼f 2020
990 ▼a ***1008102
991 ▼a E-BOOK