Three essays in financial econometrics: fractional cointegration, nonlinearities and asynchronicities

Abstract

This thesis develops theoretical tools for the stylised facts of multivariate volatility processes and stock returns in financial markets. The first essay of this thesis contributes to the literature of fractionally cointegrated processes. Threshold adjustment is allowed in the error correction of bivariate fractionally cointegrated systems. Hypothesis testing for the presence of threshold and simulation evidence are provided to support the need of threshold specification in the adjustment dynamics of fractionally cointegrated processes. Empirical application considers the cointegrating relation and adjustment dynamics of S&P500 option implied volatility index spot and futures. Empirical finding shows that investors tend to hedge against volatility by using volatility-tracking products during market turbulence. The next two essays investigate some econometric issues that arise from the use of asynchronous data on modelling the joint dynamics of stock returns. The return correlation is inaccurate if asynchronicity is not taken into consideration. As a result, portfolio risk management can be highly distorted. Aiming to develop an accurate estimation on the return correlation dynamics, several econometric techniques are introduced to tackle this asynchronicity problem that allow financial practitioners to adequately adjust the asynchronous stock return series. This research also attempts to analyse asynchronicity problem as a measurement error problem, parameter estimates from the conventional vector autoregressive models are inconsistent if the vector of multivariate stock returns contains asynchronous returns. A good proxy of measurement error can effectively correct the asynchronous return vector and hence yield consistent parameter estimates.

More information

Identifiers

ORCID for Jose Olmo: orcid.org/0000-0002-0437-7812

Catalogue record

Export record