| dc.description.abstract | Over the last couple of years, volatility forecasting has gained great attention because of its significance in financial markets. In particular, the outbreak and rapid spread of the coronavirus (Covid-19) and the invasion of Ukraine have had striking impacts on global financial markets. Incidents like these have elevated the risk to unprecedented levels, causing investors to suffer severe losses quickly. As a result of strict quarantine policies, many countries' economic activities suffer. Some countries may experience long-term effects resulting in businesses failing and mass unemployment. The war in Ukraine has led to increased economic uncertainty among businesses, households, and financial markets.
This master´s thesis will investigate volatility forecasting on risky assets, cryptocurrencies, and major assets. We will do so by focusing on the implied volatility of Bitcoin and Ethereum, as it has not been researched to a great extent yet. However, it makes sense to compare this with traditional assets in multiple markets to prove the accuracy of our findings. Therefore, we consider the following assets in our thesis: 1) VIX with underlying S&P 500 (stock market); 2) GVZ with underlying GLD (commodity market); and 3) EVZ with underlying FXE (forex market). We investigate if implied volatility for Bitcoin and Ethereum can improve volatility forecasts by augmenting the HAR-RV model of Corsi (2009). We base our explanatory model using historical volatility, implied volatility, returns, and trading volume. Therefore, we make one of the first contributions using implied volatility of cryptocurrencies in volatility forecasting.
Our findings suggest implied volatility significantly impacts volatility forecasts in all major assets trading on regulated markets and Ethereum to an even greater extent than historical volatility. Our study, however, did not prove this for Bitcoin, which does not yield a significant relationship with implied volatility. Our findings for major assets are consistent with previous studies and extend the literature to include cryptocurrencies. | |