Dependence in financial and high-dimensional time series

This thesis deals with the complex subtle and unexplored dependence structure in financial time series and high-dimensional macroeconomic time series. This thesis provides more flexible model specifications and simpler ways to account for autocorrelations in time series. First, this thesis proposes an autocorrelation-based factor model to estimate and forecast the yield curve, or the term structure of interest rates. This new model offers favorable in-sample fit and out-of-sample forecasting results. Second, it integrates more flexible distributions into the classic capital asset pricing model to better feature financial return characteristics. It develops the procedures for estimation and shows the economic value of this novel integration to institutional investors. Third, it sheds light on the dependence between the overnight return and the subsequent intraday return and examines trading opportunities and holding periods for day traders.
Hao Li graduated with B.A. in Finance (with honor in recognition of the thesis) from Nankai University, China. She obtained M.Sc. in Finance from the University of Groningen (Rijksuniversiteit Groningen), the Netherlands. While she was completing her M.Sc. in Economics at the University of Groningen, she started her Ph.D. in Financial Econometrics in the Department of Quantitative Economics at the University of Amsterdam, the Netherlands.