| Summary: | This thesis is mainly concerned with two broad topics: i) the impact of country ETF’s premiums and discounts over feedback trading; ii) modelling high-frequency realized volatility on liquid assets. Out of the first topic, it investigates whether feedback trading exists in US-listed country ETFs and whether it varies with their observed/forecast premiums and discounts by using a sample of twenty country ETFs for the 2000-2016 window, it shows that feedback trading is present in several of them, particularly those targeting Asia Pacific markets. For the second topic, it analyses the forecastability and tradability of realized volatility on financial markets, specifically, major stock indices and their tradable derivatives are used to help decision makers in taking better hedging or trading positions in the short term. This thesis also extends the study on energy commodities which do not have their own (implied) volatility futures to trade. A heterogenous autoregressive model including jumps is used to model realized volatility, additionally, recurrent neural networks and a hybrid model are also added to the toolbox. It has been noticed, that the linear heterogeneous autoregressive model produces on average the most stable results.
|
|---|
