Πλοήγηση ανά Επιβλέπων "Kasimatis, Konstantinos"

A B C D E F G H I J K L M N O P Q R S T U V W X Y Z

Α Β Γ Δ Ε Ζ Η Θ Ι Κ Λ Μ Ν Ξ Ο Π Ρ Σ Τ Υ Φ Χ Ψ Ω

Τώρα δείχνει 1 - 2 από 2

Hedging an equity portfolio using volatility options
(09/27/2021) Simantirakis, Michail; Σημαντηράκης, Μιχαήλ; Athens University of Economics and Business, Department of Business Administration; Spyrou, Spyros; Agoraki, Maria-Eleni; Kasimatis, Konstantinos
The purpose of this dissertation is to demonstrate how classic portfolios can benefit from implied volatility. Information about the introduction of implied volatility indices and volatility derivative instruments are presented, as well as the desirable characteristics of the implied volatility, which lead investors to create portfolio strategies including the appropriate implied volatility derivatives. The strong negative correlation between implied volatility indices returns and the returns of equities combined with the asymmetric relation between their returns during negative and positive days of returns on equities are the main characteristics that enhance the interest for diversification or hedging through mixed portfolios of equities and implied volatility. A variety of academic research around volatility strategies is also presented. An empirical application of strategies is implemented using the core European Equities Index (STOXX 50) and the corresponding European Volatility Index (VSTOXX). The application investigates whether it is possible to provide diversification or hedging on an equity portfolio during tough periods in financial markets, like the COVID-19 market crash in March 2020. The implementation of the strategies uses threshold models of implied volatility which give signals for adding the appropriate volatility options in the equity portfolios when it is considered necessary. The strategies create different portfolios which are compared in between them and with the benchmarking equity index portfolio. Characteristics and performance metrics, such as Sharpe ratio, Loss deviation, and conditional-VaR are calculated for evaluating the performance and the risk of portfolios. The results of the empirical application despite the limitations are consistent with previous studies and show that implied volatility derivatives could provide the desired diversification and hedging during a market crash, but investors must be willing to pay the cost of the premiums when the market is stable or rising.
Predictability of bitcoin price using Twitter sentiment analysis
(09/27/2021) Papatheodorou, Konstantinos; Παπαθεοδώρου, Κωνσταντίνος; Athens University of Economics and Business, Department of Business Administration; Spyrou, Spyros; Agoraki, Maria-Eleni; Kasimatis, Konstantinos
The purpose of the specific research we conduct is to make accurate predictions of Bitcoin (BTC) price fluctuations using an extensive social Sentiment Analysis in two forms. Firstly, by taking the raw Sentiment data and combine them with BTC historical prices and secondly by applying the Autoregressive Moving Average model (ARIMA) in the raw Sentiment data and correlate the result with the movement of the historical prices.Social Sentiment Analysis have been used to classify the opinion of the under study tweets producing three possible results in each tweet classification: positive, negative or neutral. The tweets chosen for the research have been mined from Twitter API and belongs to a financial analyst with a crucial influence in public opinion about cryptos. The final result of the comparison between raw Sentiment data and BTC historical prices for the under study time period gave a non significant correlation of 3.9% (Pearson coefficient metric) and 0.2% (Spearman coefficient metric).Making use of the ARIMA model to smoothen the raw Sentiment data curve had a significantly better outcome in the predictions. That, happened by bisecting the data in two period to train and test the model. The final correlation of the predictions and BTC historical prices produced a correlation of 64.5% (Pearson coefficient metric) and 61.4% (Spearman coefficient metric) that indicates a significant correlation between predictions and historical data.