Econometric Modelling of Short Panels with Applications in Financial Econometrics
In recent years, there has been a significant increase in the availability of large panels of economic and financial data. The availability of such datasets is central to attaining a deeper understanding of complex interdependencies between economic and financial variables. Importantly, however, while large macro and financial panels are rich in valuable information, extracting this information requires econometric methods that are capable of dealing with computational and mathematical problems associated with large-dimensional data. Developing such methods, especially for use with financial data, is the main objective of this 4-year research project.
This research project is supported by the European Commission Research Executive Agency under the Marie Curie Career Integration Grant programme (FP7), for the period between August 2013 and July 2017.
- Modelling the time-varying correlation structure of a large number of financial assets. Although financial panels are generally characterised by a large time-series dimension, panels with a short time-series dimension (the so-called short panels) are also likely to arise, especially when availability of a long history of data is made impossible by a recent structural break, or when panels are naturally short due to data being observed at low frequency.
- Developing bias-correction methods for the incidental parameter bias, when the panels are characterised by cross-section dependence, as would be the case for financial panels. There is already a well-established literature on the incidental parameter issue, both in econometrics and statistics. However, this literature mainly focusses on the case of cross-sectionally independent panels. Hence, a second objective of this project is to extend this literature to cross-sectionally dependent panels. Such methods will also be useful in dealing with short panels.
The following papers have been submitted to academic journals during the first period of the project.
- Fitting Vast Dimensional Time-Varying Covariance Models (with Robert F Engle, Neil Shephard and Kevin Sheppard)