Possible topics for Masters-Theses in WS 2015/2016 in the research area “Financial Econometrics”
Dear Master Candidates:
You are warmly welcome to write your theses at the Chair of Econometrics and Quantitative Methods in our research area Financial Econometrics and Quantitative Risk Management (QRM). Briefly speaking, all topics for Master-Theses in WS 2015/16 involve possible further development, comparison and application of different non- and semiparametric GARCH- or ACD-type models for modeling daily or high-frequency financial data and their application in QRM. An MA-thesis can be written in English as well as in German. Possible sub-areas in WS 2015/16 are (but not limited to):
Sub-area I: Comparison and application of Semi-GARCH and Spline-GARCH models
GARCH model and its well-known extensions such as APARCH, EGARCH and CGARCH (component GARCH) are very useful approaches for measuring financial market risk. However, financial returns in a long period are usually non-stationary due to e.g. s slowly changing unconditional variance trend (the scale function). Semiparametric generalizations of the above mentioned models are hence introduced into the literature, where it is proposed to estimate the scale function using kernel or local linear regression. Engle and Rangel (2008) proposed a similar Spline-GARCH model with another idea for estimating the scale function. The purpose of topics from this subarea is to extend the idea of Engle and Rangel (2008) to different Spline-GARCH models and to compare them with the Semi-GARCH-type models. Discussion on volatility forecasting and its application of those models in QRM is also wished.
Sub-area II: Comparison and application of Semi-ACD and Spline-ACD models
It is well-known that an ACD model is related to a squared GARCH. In the literature the idea of Semi-GARCH models is hence extended to Semi-ACD models. Those models are widely applicable for modelling different kinds of non-negative financial time series such as those of trading volumes and volatility indexes. If the log-transformation is suitable, such data sets can also be analyzed using an ARMA model with a slowly changing trend, which can be called a Semi-Log-ACD model. The trend in those models is usually estimated by nonparametric regression, which can however also be estimated using cubic splines (Engle and Russell, 1998). The purpose of topics from this subarea is to extend the idea of Engle and Russell (1998) to different Spline-ACD models, to apply them for modelling different kinds of financial time series and to compare them with existing Semi-ACD and Semi-Log-ACD models.
Sub-area III: Application of multiplicative component GARCH models
The multiplicative component GARCH model was introduced by Engle and Sokalska (2012) for estimating different factors in high-frequency returns. The purpose of the theses in this sub-area is to give a comprehend summary of the literature in this context, to discuss possible parametric and semiparametric extensions of this model, to investigate the properties and estimation of those new approaches and to apply them to own data. The main focuses are the analysis of the long-term dynamics, daily patterns and additional conditional fluctuations in high-frequency returns. Furthermore, you should discuss the application of your empirical results in financial trading, hedging, option pricing or QRM in detail.
Remark: Different topics from the same sub-area can be designed for a group of candidates. You are hence welcomed to write you these together your friends on those group topics.