Generalized Method of Moments with a large number of moment conditions with applications in financial economics

Han, Hyojin

Full Metadata

Overview

Year:: 2017
Contributor:: Han, Hyojin (creator); Renault, Eric (Advisor); McCloskey, Adam (Reader); Schennach, Susanne (Reader); Brown University. Department of Economics (sponsor)
Genre:: theses
Subject:: Econometrics--Asymptotic theory; Econometrics; Econometrics--Research
Extent:: xiii, 138 p.
DOI: https://doi.org/10.7301/Z07D2SMQ

Files

Description

Abstract:: This dissertation studies Generalized Method of Moments (GMM) models when the number of moments is allowed to diverge with the sample size and its application in financial economics. In the first chapter, my coauthor Eric Renault and I study how identification is affected in GMM estimation as the number of moments increases. We develop a general asymptotic theory extending the set up of of Chao & Swanson (2005) and Antoine & Renault (2009, 2012) to the case where moment conditions have heterogeneous identification strengths and the number of them may diverge to infinity with the sample size. The theory encompasses many cases including GMM models with many moments (Han & Phillips (2006)), partially linear models, and local GMM via kernel smoothing. We provide an understanding of the benefits of a large number of moments that compensate the weakness of individual moments by explicitly showing how an increasing number of moments improves the rate of convergence in GMM. In the second chapter, we develop an affine discrete-time option pricing model by using the general framework of discrete-time affine models by Darolles et al. (2006) for modeling a bivariate process of returns and stochastic volatility (SV). We exploit information in high-frequency data as summarized by realized variance (RV) which produces dynamics of RV as a SV-type extension of traditional high-frequency-based volatility (HEAVY) models by Shephard & Sheppard (2010) that are of the GARCH type. An empirical illustration is provided with the S&P500 index data. In the third chapter, we revisit a potential identification issue of affine models. Affine models provide conditional moment restrictions based on analytical characteristic functions (CF). The usual procedure for GMM is to apply minimum distance estimation using a set of induced unconditional moment restrictions. However, while the literature has focused on efficiency of GMM, the identification issue with a choice of instruments has been neglected. This chapter shows that identification failure could arise commonly in time-series models if moment conditions are not chosen carefully. Thanks to the affine structure, we provide analytical identification conditions for some choices of moments.
Notes:: Thesis (Ph. D.)--Brown University, 2017

Content

Access Conditions

Rights: In Copyright
Restrictions on Use: Collection is open for research.

Citation

Han, Hyojin, "Generalized Method of Moments with a large number of moment conditions with applications in financial economics" (2017). Economics Theses and Dissertations. Brown Digital Repository. Brown University Library. https://doi.org/10.7301/Z07D2SMQ

Relations

Collection:

Economics Theses and Dissertations

Theses and Dissertations for the Economics department.
...