New project on modelling high-frequency data in financial economics

In the "Big Data" era, financial practitioners often have to tackle a large amount of high-frequency data collected for a large number of financial assets, in which setting, existing methods designed for estimating small volatility matrices perform poorly. This project will develop a novel nonparametric technique to accurately estimate the large spot volatility structure and subsequently construct optimal out-of-sample portfolios. It will not only make a profound methodological and theoretical contribution to the literature, but also provide state-of-art analytical tools for practitioners in economics and finance. This two-year project is supported by a British Academy/Leverhulme Small Research Grant.