New frontiers in applied quantitative management and finance: nonlinear econometric models, fractional cointegration and economic

This project addresses the study of financial markets (and other) using some relatively new tools based on nonlinear time series and econophysics. Nonlinear models used in econometrics are usually constructed in order to address some specific issues of the events observed in real markets. One such type of model is the threshold class, useful to study asymmetries in the data. Other forms of assumptions concerning distributional aspects of the error terms have been also analysed by econometricians. Conditional heteroskedasticity fractional volatility are amongst these lines of research. Much less attention, however, has been paid to the nonlinear character of integrated data, at least in what concerns stock market analyses. Another line of research on financial markets, originally used by physics lies on the concept of entropy used in the context of econophysics.