Model selection and loss functions for structural time series: networks, spatial models, causality measures, and misspecified or redundant moment conditions

Full Metadata

Overview

Title
Model selection and loss functions for structural time series: networks, spatial models, causality measures, and misspecified or redundant moment conditions
Contributors
Scidá, Daniela (creator)
Renault, Eric (Director)
McCloskey, Adam (Reader)
Dungey, Mardi (Reader)
Brown University. Economics (sponsor)
Doi
10.7301/Z0TQ5ZXB
Copyright Date
2016
Abstract
This dissertation is comprised of the following three chapters: (1) "Causality and Markovianity: Information Theoretic Measures" (joint work with Eric Renault), my job market paper entitled (2) "Structural VAR and Financial Networks: A Minimum Distance Approach to Spatial Modeling," and my third year paper entitled (3) "GMM with Minimum Mean Squared Error." All three chapters are about econometric methodology for time series, with a particular focus on model selection and loss functions. They include both theoretical and empirical developments about Information Theory in Econometrics, Generalized Method of Moments, Networks, and Spatial Modeling. The common feature of all the econometric methodologies developed in this dissertation is the applicability to financial econometrics.
Keywords
Kullback causality measures
Granger causality
Markov property
non-nested hypotheses
financial networks
SVAR models
spatial models
GMM
misspecified moment conditions
AMSE reduction
optimal weighting matrix
redundancy
Financial institutions--Computer networks
Spatial systems--Mathematical models
Notes
Thesis (Ph.D. -- Brown University (2016)
Extent
xiv, 187 p.

Citation

Scidá, Daniela, "Model selection and loss functions for structural time series: networks, spatial models, causality measures, and misspecified or redundant moment conditions" (2016). Economics Theses and Dissertations. Brown Digital Repository. Brown University Library. https://doi.org/10.7301/Z0TQ5ZXB

Relations

Collection: