Three essays on time-varying parameters and time series networks

Research output: ThesisDoctoral ThesisScientific

251 Downloads (Pure)

Abstract

This thesis is composed of three essays on time-varying parameters and time series networks where each essay deals with specific aspects thereof. The thesis starts with proposing a 2SLS based test for a threshold in models with endogenous regressors in Chapter 2. Many economic models are formulated in this way, for example output growth or unemployment rates in different states of the economy. Therefore, it is necessary to have tools available which are capable of indicating whether such effects exist in the data or not. Chapter 3 proposes, to my best knowledge, the first estimator for the inverse of the long-run covariance matrix of a linear, potentially heteroskedastic stochastic process under unknown sparsity constraints. That is, the econometrician does not know which entries of the inverse are equal to zero and which not. Such situations naturally arise, for example, when modelling partial correlation networks based on time series data. Finally, in Chapter 4 this thesis empirically investigates how robust two commonly applied network measures, the From- and the To-degree, are to the exclusion of central nodes in financial volatility networks. This question is motivated by the current empirical literature which excludes certain nodes such as Lehman Brothers from their analysis.
Original languageEnglish
QualificationDoctor of Philosophy
Awarding Institution
  • Tilburg University
Supervisors/Advisors
  • Werker, Bas, Promotor
  • Boldea, Otilia, Promotor
Award date16 Mar 2018
Place of PublicationTilburg
Publisher
Print ISBNs978 90 5668 555 3
Publication statusPublished - 2018

Fingerprint

Time-varying parameters
Node
Endogenous regressors
Unemployment rate
Stochastic processes
Exclusion
Estimator
Output growth
Time series data
Covariance matrix
Modeling

Cite this

Rothfelder, M. (2018). Three essays on time-varying parameters and time series networks. Tilburg: CentER, Center for Economic Research.
Rothfelder, Mario. / Three essays on time-varying parameters and time series networks. Tilburg : CentER, Center for Economic Research, 2018. 142 p.
@phdthesis{fc7a10c07eee479aac22b36ce45f2037,
title = "Three essays on time-varying parameters and time series networks",
abstract = "This thesis is composed of three essays on time-varying parameters and time series networks where each essay deals with specific aspects thereof. The thesis starts with proposing a 2SLS based test for a threshold in models with endogenous regressors in Chapter 2. Many economic models are formulated in this way, for example output growth or unemployment rates in different states of the economy. Therefore, it is necessary to have tools available which are capable of indicating whether such effects exist in the data or not. Chapter 3 proposes, to my best knowledge, the first estimator for the inverse of the long-run covariance matrix of a linear, potentially heteroskedastic stochastic process under unknown sparsity constraints. That is, the econometrician does not know which entries of the inverse are equal to zero and which not. Such situations naturally arise, for example, when modelling partial correlation networks based on time series data. Finally, in Chapter 4 this thesis empirically investigates how robust two commonly applied network measures, the From- and the To-degree, are to the exclusion of central nodes in financial volatility networks. This question is motivated by the current empirical literature which excludes certain nodes such as Lehman Brothers from their analysis.",
author = "Mario Rothfelder",
note = "CentER Dissertation Series Volume: 554",
year = "2018",
language = "English",
isbn = "978 90 5668 555 3",
publisher = "CentER, Center for Economic Research",
school = "Tilburg University",

}

Rothfelder, M 2018, 'Three essays on time-varying parameters and time series networks', Doctor of Philosophy, Tilburg University, Tilburg.

Three essays on time-varying parameters and time series networks. / Rothfelder, Mario.

Tilburg : CentER, Center for Economic Research, 2018. 142 p.

Research output: ThesisDoctoral ThesisScientific

TY - THES

T1 - Three essays on time-varying parameters and time series networks

AU - Rothfelder, Mario

N1 - CentER Dissertation Series Volume: 554

PY - 2018

Y1 - 2018

N2 - This thesis is composed of three essays on time-varying parameters and time series networks where each essay deals with specific aspects thereof. The thesis starts with proposing a 2SLS based test for a threshold in models with endogenous regressors in Chapter 2. Many economic models are formulated in this way, for example output growth or unemployment rates in different states of the economy. Therefore, it is necessary to have tools available which are capable of indicating whether such effects exist in the data or not. Chapter 3 proposes, to my best knowledge, the first estimator for the inverse of the long-run covariance matrix of a linear, potentially heteroskedastic stochastic process under unknown sparsity constraints. That is, the econometrician does not know which entries of the inverse are equal to zero and which not. Such situations naturally arise, for example, when modelling partial correlation networks based on time series data. Finally, in Chapter 4 this thesis empirically investigates how robust two commonly applied network measures, the From- and the To-degree, are to the exclusion of central nodes in financial volatility networks. This question is motivated by the current empirical literature which excludes certain nodes such as Lehman Brothers from their analysis.

AB - This thesis is composed of three essays on time-varying parameters and time series networks where each essay deals with specific aspects thereof. The thesis starts with proposing a 2SLS based test for a threshold in models with endogenous regressors in Chapter 2. Many economic models are formulated in this way, for example output growth or unemployment rates in different states of the economy. Therefore, it is necessary to have tools available which are capable of indicating whether such effects exist in the data or not. Chapter 3 proposes, to my best knowledge, the first estimator for the inverse of the long-run covariance matrix of a linear, potentially heteroskedastic stochastic process under unknown sparsity constraints. That is, the econometrician does not know which entries of the inverse are equal to zero and which not. Such situations naturally arise, for example, when modelling partial correlation networks based on time series data. Finally, in Chapter 4 this thesis empirically investigates how robust two commonly applied network measures, the From- and the To-degree, are to the exclusion of central nodes in financial volatility networks. This question is motivated by the current empirical literature which excludes certain nodes such as Lehman Brothers from their analysis.

M3 - Doctoral Thesis

SN - 978 90 5668 555 3

PB - CentER, Center for Economic Research

CY - Tilburg

ER -

Rothfelder M. Three essays on time-varying parameters and time series networks. Tilburg: CentER, Center for Economic Research, 2018. 142 p. (CentER Dissertation Series).