Abstract
dc:descriptionMany econometric models that are commonly used in empirical financial and economic applications are linear. However, the existing literature provides several indications that nonlinear models may be more appropriate to describe relationships encountered in real phenomena. Excess kurtosis, volatility clustering, sensitivity of estimated parameters to alternative specifications of the estimation period are all points towards the presence of nonlinearities in financial and economic time series. Therefore, more complex models are needed in order to overcome the limitations of the linear regression models to deal with these stylized facts. This thesis examines some non-standard parametric time series models which aim at capturing several features of time series of interest, not accounted for by the usual models of the econometric literature. The first class of models considered here are autoregressive models (AR) whose parameters are autoregressive or moving average (MA) processes themselves. The models studied here, can be considered as natural alternatives to the standard linear autoregressive models with constant parameters in cases where the estimated parameters of the AR models exhibit persistent time variation.The second class of models analyzed here, aim at describing multivariate processes which exhibit unconditional heteroscedasticity (as opposed to the standard case of conditional heteroscedasticity modeled by GARCH models). These models are motivated by the empirical observations that the variance of the returns of certain financial assets exhibits time trends. As a result, we put forward a model in which the process generating asset returns exhibits trending second moments.The third category of models that this thesis is dealing with, are models which exhibit conditional heteroscedasticity. We focus specifically in models which exhibit very strong-persistent conditional heteroscedasticity which is manifested as a unit root in the conditional variance. These models are usually referred to as Integrated GARCH (IGARCH) models and describe an infinite-variance process. These models are analyzed in the context of an empirical puzzle that has been observed in the financial econometrics literature. In particular, it has been observed that asset returns observed at high frequencies exhibit IGARCH effects and fat-tailed distributions. However, as we move to lower frequencies, the distribution of returns looks more and more like the normal distribution. The puzzle lies in the fact that the observed Aggregational Gaussianity cannot be established theoretically, since the infinite variance in the original high-frequency series makes the application of the Central Limit Theorem prohibitive.Finally, in the fourth chapter we consider nonlinear models to assess the success of a procedure that evaluates mutual funds, namely that of Morningstar star rating system. For this purpose, we analyze multivariate regression models in which the regression coefficients are not constant, but functions of a set of observable variables. These models are applied in cases where the systematic risk of a managed portfolio, as measured by the beta coefficient, changes with time as a result of portfolio re-balancing by the active manager.
Degree
thesis:*- Grantor dc:publisher
- University of Piraeus (UNIPI)
- Year dc:date
- 2010
Author and committee
dc:creator, dc:contributor.*- Authors dc:creator
-
- Antypas, Antonios
- Αντύπας, Αντώνιος
Subjects
dc:subject × 25- Μη γραμμικά μοντέλα
- Αυτοπαλίνδρομα μοντέλα
- Μεταβλητοί συντελεστές
- Συσσωρευμένη κανονικότητα
- Μη πεπερασμένη διακύμανση
- Καταμερισμός αξιογράφων
- Ροπές με πολυωνυμική τάση
- Αξιολόγηση αμοιβαίων κεφαλαίων
- Non linear model
- Autoregressive models
- Time varying coefficients
- Aggregational gaussianity
- Infinite variance
- Asset allocation
- Trending second moments
- Mutual funds evaluation
- Morningstar
- Φυσικές Επιστήμες
- Μαθηματικά
- Κοινωνικές Επιστήμες
- Οικονομικά και Επιχειρήσεις
- Natural Sciences
- Mathematics
- Social Sciences
- Economics and Business
Rights
- Language dc:language
- eng
Identifiers
dc:identifier.*- Identifier
- 10.12681/eadd/37290