Time series analysis and extreme value theory

Vasilakakis, Konstantinos; Βασιλακάκης, Κωνσταντίνος

dc.contributor.author	Vasilakakis, Konstantinos	en
dc.contributor.author	Βασιλακάκης, Κωνσταντίνος	el
dc.date.accessioned	2026-02-26T08:20:18Z
dc.date.available	2026-02-26T08:20:18Z
dc.identifier.uri	https://dspace.lib.ntua.gr/xmlui/handle/123456789/63603
dc.identifier.uri	http://dx.doi.org/10.26240/heal.ntua.31298
dc.description	Εθνικό Μετσόβιο Πολυτεχνείο--Μεταπτυχιακή Εργασία. Διεπιστημονικό-Διατμηματικό Πρόγραμμα Μεταπτυχιακών Σπουδών (Δ.Π.Μ.Σ.) “Μαθηματική Προτυποποίηση σε Σύγχρονες Τεχνολογίες και στα Χρηματοοικονομικά”	el
dc.rights	Default License
dc.subject	Financial Time Series	en
dc.subject	Extreme Value Theory	en
dc.subject	Financial Modeling	en
dc.subject	Financial Technology	en
dc.subject	Risk Management	en
dc.subject	Ανάλυση Χρονοσειρών	el
dc.subject	Θεωρία Ακραίων Τιμών	el
dc.subject	Οικονομετρία	el
dc.subject	Διαχείριση Κινδύνου	el
dc.subject	Χρηματοοικονομική Τεχνολογία	el
dc.title	Time series analysis and extreme value theory	en
heal.type	masterThesis
heal.classification	Econometrics	en
heal.classification	Applied Mathematics	en
heal.classification	Fintech	en
heal.language	en
heal.access	free
heal.recordProvider	ntua	el
heal.publicationDate	2025-07-04
heal.abstract	Periods of heightened financial stress are characterized by uncertainty, abnormally high market volatility, skewed returns and extreme breakdowns in historical asset correlations, which pose significant threats to investment performance and financial stability. This paper discusses the nature of major financial ETFs in stress periods, with a focus on risk, volatility, and tail event dynamics by combining time series econometric models and extreme value theory (EVT), key market sectors and instruments during severe financial crises, such as the 2020 COVID-19 pandemic, geopolitical and banking sector shocks in the early 2020s, and the ongoing inflationdriven tightening cycle. The time series analysis employs QVAR connectedness approach, ARMA and DCCGARCH copula models to model and predict leverage effects and volatility clustering, while EVT and R-Vine copulas are employed to model extreme losses, upper and lower tail dependences, and estimate Value at Risk during stressful market periods. We find that asset return distributions are strongly non-normal and fat-tailed in times of crisis, and are characterized by complex dependence structures across the world’s largest markets. Moreover, we show that asset volatility spillovers increase during periods of market stress, which diminishes diversification benefits and increases systemic risk. The research contributes to improving the understanding of financial contagion, tail risk behavior and inter-asset dependencies during periods of increased market tension.	en
heal.advisorName	Michaelides, Panayotis G.	en
heal.committeeMemberName	Παπαγεωργίου, Θεοφάνης	el
heal.committeeMemberName	Κωνσταντάκης, Κωνσταντίνος	el
heal.academicPublisher	Εθνικό Μετσόβιο Πολυτεχνείο. Σχολή Εφαρμοσμένων Μαθηματικών και Φυσικών Επιστημών	el
heal.academicPublisherID	ntua
heal.numberOfPages	67 σ.	el
heal.fullTextAvailability	false
heal.fullTextAvailability	false