Rustam Ibragimov (Imperial College London)

Loading Map....

Date/Time
Date(s) - 17/01/2018
3:30 pm - 4:30 pm

Location
V101

Categories


Robustness, Heavy-Tailedness and Dependence in Economics and Finance: Market (In-)Efficiency, Volatility Clustering, Stock Return Predictability and Beyond

Abstract

Many key variables in finance, economics and risk management, including financial returns and foreign exchange rates, exhibit nonlinear dependence, heterogeneity and heavy-tailedness of some usually largely unknown type. Recent works in the literature have shown that heavy-tailedness – the property of financial and economic markets that governs large downfalls and large fluctuations in them – is of key importance for robustness of many key models and standard inference approaches in economics, finance, risk management and insurance. The works demonstrate that conclusions of many mainstream economic and financial models, including those on diversification optimality, are not robust to heavy tails — their implications may be reversed when heavy-tailedness is very pronounced. Motivated, in part, by these findings, a number of works in econometrics over the past years have focused on the development robust inference alternatives to widely used standard approaches, including those based on OLS regressions.

Among other research problems in empirical economics and finance, the presence of non-linear dependence (e.g., modelled using GARCH-type dynamics) and heavy-tailedness may problematic the analysis of (in-)efficiency, volatility clustering and predictive regressions in economic and financial markets using traditional approaches based on ACF’s of squared returns and asymptotic methods. Similar problems appear with commonly used predictive regressors. The talk will present several new approaches to deal with the above problems. The approaches are based on new methods of robust inference using conservativeness of t-statistics. In the methods, estimates of parameters of interest are computed for groups of data and the inference is based on t-statistics in resulting group estimates. This results in valid robust inference under a wide range of heterogeneity and dependence assumptions under the only conditions of asymptotic normality of group estimates. Numerical results and empirical applications confirm advantages of the new approaches over existing ones and their wide applicability in the study of market (in-)efficiency, volatility clustering, predictive regressions and other areas.

 

Comments are closed.