CEU Electronic Theses and Dissertations, 2021
| Author | Divényi, János Károly |
|---|---|
| Title | Three Essays on Causal Effects: Identification in a Novel Setting and Two Empirical Studies |
| Summary | The thesis consists of three chapters on causal effects: in each of them, I aim to deepen our understanding of causal relationships. In the first chapter, I investigate how we can identify causal effects from adaptively collected data which is becoming more and more widespread in the modern on-line world. Running Monte-Carlo simulations I illustrate how adaptivity biases the standard treatment effect estimators and I recommend a new strategy that balances between the welfare and estimation goals of the decision-maker. The remaining two chapters are empirical studies aimed at estimating causal effects in two policy-relevant settings. In the second chapter, I examine the effect of retirement on cognitive performance, taking a unifying approach by replicating previous studies and tracing back the differences in their results to various identification problems. I propose a new identification strategy that suggests that the effect is close to zero. In the third chapter (co-authored with S\'andor S\'ov\'ag\'o) we study the effect of elite schools on competency scores using non-parametric bound identification. Our results show that the effect is heterogeneous being more expressed for high-ability students; this result has important implications for the validity of previous studies that mainly estimate the effect around the admission cutoffs exploiting the discontinuity. |
| Supervisor | Lieli, Róbert; Kézdi, Gábor |
| Department | Economics PhD |
| Full text | https://www.etd.ceu.edu/2021/divenyi_janos.pdf |
Visit the CEU Library.
© 2007-2021, Central European University