Search Results

Title: Econometric Analysis of Functional Data
Resulting in 1 citation.
1. Seong, Dakyung
Econometric Analysis of Functional Data
Ph.D. Dissertation, Department of Economics, University of California, Davis, 2020
Cohort(s): Children of the NLSY79, NLSY79
Publisher: ProQuest Dissertations & Theses (PQDT)
Keyword(s): Behavior Problems Index (BPI); Handedness; Maternal Employment; Modeling, Instrumental Variables; Monte Carlo

Permission to reprint the abstract has not been received from the publisher.

Chapter 1 concerns an endogenous binary response model when available instrumental variables are possibly infinite-dimensional. Furthermore, the instrumental variables considered in this chapter could have too small signals in the first stage to obtain reliable results using existing estimators. The use of high-dimensional or infinite-dimensional instrumental variables under this environment has not received enough attention compared to the linear model in spite of its empirical importance. In this chapter, I propose two estimators based on ridge-regularization; the ridge-regularized conditional maximum likelihood estimator and the ridge-regularized nonlinear least square estimator. I derived asymptotic properties for both estimators. Compared to existing estimators that are only valid under strong identification, the proposed ridge-regularized estimators have good asymptotic properties even under weak identification. Monte Carlo simulations show that the proposed estimators have good finite sample properties regardless of the first-stage signal-to-noise ratio. Two empirical applications are also provided. In the first example, I investigate the effect of education on employment status for a full calendar year using variations in seasons of birth as instrumental variables. The second empirical example examines the effect of child behavioral developments on maternal employment status. To control for the potential endogeneity, I use children's handedness as instrumental variables in the second example.
Bibliography Citation
Seong, Dakyung. Econometric Analysis of Functional Data. Ph.D. Dissertation, Department of Economics, University of California, Davis, 2020.