1. Αποθετήριο Παραδοτέων Υποτρόφων Δράσεων ΕΣΠΑ Τριτοβάθμιας Εκπαίδευσης 2017-2023
2. Όλο το Αποθετήριο
3. Υποψήφιοι διδάκτορες

Τύπος:	Διδακτορική διατριβή
Τίτλος:	Measuring managerial skill in the mutual fund indrustry: a stochastic dominance based approach
Συγγραφέας:	[EL] Σαρρή, Δανάη[EN] Sarri, Danai
Επιβλέπων διατριβής:	[EL] Τσιώνας, Ευθύμιος[EN] Tsionas, Efthymios
Συμβουλευτική επιτροπή:	[EL] Αρβανίτης, Στυλιανός[EN] Arvanitis, Stelios [EL] Τοπάλογλου, Νικόλαος[EN] Topaloglou, Nikolaos
Ημερομηνία:	18/05/2020
Περίληψη:	In recent years, a significant increase of published research based on the Stochastic Dominance theory is observed. Computational advancements have been substantial for Stochastic Dominance tests to gain attractiveness in several areas, such as Economics and Finance. In this current study, we formulate Stochastic Dominance Efficiency (SDE) tests to construct portfolios as an extension of the classical Mean-Variance approach. We proceed with dimensionality reduction techniques to tackle the “curse of dimensionality” in the original dataset of 45,000 U.S. mutual funds, since the raw data dimensions make the aforementioned Stochastic Dominance formulations prohibitive. Towards that direction, a Principal Component Analysis is employed to reduce the number of the initial mutual funds assets by obtaining a small set of principal variables. The empirical findings reveal that through the construction of the optimal portfolios based on the Linear Programming formulations, the market efficiency is not rejected, yet it is inefficient in a Mean-Variance framework. The reduced set of principal variables can produce optimal portfolios, which are used to reassess and predict them, whereas to identify characteristics of individuals that manage them, such as managerial talent. Traditional forecasting models such as ARIMA and Holt Winters are compared with Long Short-Term Memory Networks, for their predictive power. Based on the assessment metric of Root Mean Squared Error, our validation shows that the LSTM based forecasting model outperforms the studied alternative approaches. Finally, extending the methodology of Berk and van Binsbergen (2015), we measure the managerial skill on the constructed portfolios based on the SDE formulations of second order using value added variable. Our findings indicate the presence of managerial skill that adds value of approximately $1.064 million annually.
Γλώσσα:	Αγγλικά
Τόπος δημοσίευσης:	Αθήνα, Ελλάδα
Σελίδες:	99
Θεματική κατηγορία:	[EL] Οικονομικά[EN] Economics [EL] Οικονομετρία[EN] Econometrics [EL] Χρηματοοικονομικά[EN] Finance
Λέξεις-κλειδιά:	alpha; Uncertainty; Stochastic Dominance; dimensionality reduction; PCA; Big Data; Prediction; Time Series; Skill; Neural Networks; LSTM; deep learning; Optimization
Κάτοχος πνευματικών δικαιωμάτων:	© Danai Sarri
Διατίθεται ανοιχτά στην τοποθεσία:	https://www.didaktorika.gr/eadd/handle/10442/47489
Σημειώσεις:	This research is co-financed by Greece and the European Union (European Social Fund-ESF) through the Operational Programme Human Resources Development, Education and Lifelong Learning in the context of the project “strengthening Human Resources Research Potential via Doctorate Research” (MIS-5000432), implemented by the State Scholarships Foundation (IKY)
Εμφανίζεται στις συλλογές:	Υποψήφιοι διδάκτορες