Dissertations / Theses on the topic 'Time-Varying Multivariate GARCH Models'

This thesis presents several non-parametric and parametric models for estimating dynamic dependence between financial time series and evaluates their ability to precisely estimate risk measures. Furthermore, the different dependence models are used to analyze the integration of emerging markets into the world economy. In order to analyze numerous dependence structures and to discover possible asymmetries, two distinct model classes are investigated: the multivariate GARCH and Copula models. On the theoretical side a new dynamic dependence structure for multivariate Archimedean Copulas is introduced which lifts the prevailing restriction to two dimensions and extends the multivariate dynamic Archimedean Copulas to more than two dimensions. On this basis a new mixture copula is presented using the newly invented multivariate dynamic dependence structure for the Archimedean Copulas and mixing it with multivariate elliptical copulas. Simultaneously a new process for modeling the time-varying weights of the mixture copula is introduced: this specification makes it possible to estimate various dependence structures within a single model. The empirical analysis of different portfolios shows that all equity portfolios and the bond portfolios of the emerging markets exhibit negative asymmetries, i.e. increasing dependence during market downturns. However, the portfolio consisting of the developed market bonds does not show any negative asymmetries. Overall, the analysis of the risk measures reveals that parametric models display portfolio risk more precisely than non-parametric models. However, no single parametric model dominates all other models for all portfolios and risk measures. The investigation of dependence between equity and bond portfolios of developed countries, proprietary, and secondary emerging markets reveals that secondary emerging markets are less integrated into the world economy than proprietary. Thus, secondary emerging markets are more suitable to diversify a portfolio consisting of developed equity or bond indices than proprietary

This thesis investigates the modelling and forecasting of multivariate volatility and dependence in financial time series. The first paper proposes a new model for forecasting changes in the term structure (TS) of interest rates. Using the level, slope and curvature factors of the dynamic Nelson-Siegel model, we build a time-varying copula model for the factor dynamics allowing for departure from the normality assumption typically adopted in TS models. To induce relative immunity to structural breaks, we model and forecast the factor changes and not the factor levels. Using US Treasury yields for the period 1986:3-2010:12, our in-sample analysis indicates model stability and we show statistically significant gains due to allowing for a time-varying dependence structure which permits joint extreme factor movements. Our out-of-sample analysis indicates the model's superior ability to forecast the conditional mean in terms of root mean square error reductions and directional forecast accuracy. The forecast gains are stronger during the recent financial crisis. We also conduct out-of-sample model evaluation based on conditional density forecasts. The second paper introduces a new class of multivariate volatility models that utilizes high-frequency data. We discuss the models' dynamics and highlight their differences from multivariate GARCH models. We also discuss their covariance targeting specification and provide closed-form formulas for multi-step forecasts. Estimation and inference strategies are outlined. Empirical results suggest that the HEAVY model outperforms the multivariate GARCH model out-of-sample, with the gains being particularly significant at short forecast horizons. Forecast gains are obtained for both forecast variances and correlations. The third paper introduces a new class of multivariate volatility models which is easy to estimate using covariance targeting. The key idea is to rotate the returns and then fit them using a BEKK model for the conditional covariance with the identity matrix as the covariance target. The extension to DCC type models is given, enriching this class. We focus primarily on diagonal BEKK and DCC models, and a related parameterisation which imposes common persistence on all elements of the conditional covariance matrix. Inference for these models is computationally attractive, and the asymptotics is standard. The techniques are illustrated using recent data on the S&P 500 ETF and some DJIA stocks, including comparisons to the related orthogonal GARCH models.

This thesis has treated the subject of DCC-GARCH model’s forecasting ability and Value-at- Risk applications on the Scandinavian foreign exchange market. The estimated models were based on daily opening foreign exchange spot rates in the period of 2004-2013, which captured the information in the financial crisis of 2008 and Eurozone crisis in the early 2010s. The forecasts were performed on a one-day rolling window in 2014. The results show that the DCC-GARCH model accurately predicted the fluctuation in the conditional correlation, although not with the correct magnitude. Furthermore, the DCC-GARCH model shows good Value-at-Risk forecasting performance for different portfolios containing the Scandinavian currencies.

The global foreign exchange market is the largest financial market with turnover in this market often outstripping the GDP of countries in which they are located. The dynamics in the foreign exchange market, especially price dynamics, have huge implications for financial asset values, financial returns and volatility in the international financial system. It is therefore an important area of study. Exchange rates have often departed significantly from the level implied by fundamentals and exhibit excessive volatility. This reality creates a role for central bank intervention in this market to keep the rate in line with economic fundamentals and the overall policy mix, to stabilize market expectations and to calm disorderly markets. Studies that attempt to measure the effectiveness of intervention in the foreign exchange market in terms of exchange rate trends and volatility have had mixed results. This, in many cases, reflects the unavailability of data and the weaknesses in the empirical frameworks used to measure effectiveness. This thesis utilises the most recent data available and some of the latest methodological advances to measure the effectiveness of central bank intervention in the foreign exchange markets of a variety of countries. It therefore makes a contribution in the area of applied empirical methodologies for the measurement of the dynamics of intervention in the foreign exchange market. It demonstrates that by using high frequency data and more robust and appropriate empirical methodologies central bank intervention in the foreign exchange market can be effective. Moreover, a framework that takes account of the interactions between different central bank policy instruments and price dynamics, the reaction function of the central bank, different states of the market, liquidity in the market and the profitability of the central bank can improve the effectiveness of measuring the impact of central bank policy in the foreign exchange market and provide useful information to policy makers.

Cette thèse tente d'évaluer l'intégration des marchés émergents dans une perspective régionale et intra-régionale. Elle contribue à la littérature existante en développant un modèle dynamique d’évaluation des actifs financiers à l’international (ICAPM) avec changement de régime. Spécifiquement, les rentabilités attendues peuvent passer du régime de segmentation parfaite au régime d’intégration parfaite ou inversement en fonction d’un certain nombre de facteurs nationaux, régionaux et internationaux qui sont susceptibles d’influencer le processus d’intégration financière. Le champ d’étude s’étend aux pays de l’Asie de Sud-est, d’Europe Sud-est, de l’Amérique Latine et du Moyen Orient sur la période 1996-2008. Nous développons le modèle de Bekaert et Harvey (1995) où la PPA n’est pas vérifiée, et les variances et covariances conditionnelles sont modélisées grâce à un processus GARCH multivarié. Cette approche permet de déterminer simultanément le niveau d’intégration au cours du temps de toutes les zones dans le marché mondial et le niveau d’intégration intra-régionale dans chaque région. Il permet aussi d’analyser la formation de la prime de risque totale. Nos résultats empiriques montrent que les marchés émergents restent encore très segmentés du marché mondial et des marchés régionaux. Ces résultats suggèrent que l’inclusion des actifs des marchés émergents continue à générer des gains de diversification substantiels, et que les règles d’évaluation devraient être conformes à un état d’intégration partielle
The purpose of this thesis is to study the dynamics of the global integration process of four emerging market regions into the world and the regional market, while taking into account the importance of exchange rate and local market risk. An international capital asset pricing model suitable for partially integrated markets and departure from purchasing power parity was developed in the spirit of Bekaert and Harvey (1995)’s regime-switching model in order to explain the time-variations in expected returns on regional emerging market indices. In its fully functional form, the model allows the market integration measure as well as the global and local risk premiums to vary through time. We mainly find that the integration degree in emerging market regions (Latin America, Asia, Southeastern Europe, and the Middle East) varied widely through time over the period 1996-2008 and is satisfactorily explained by global, regional and national factors. Even though it reaches fairly high values during several periods, and exhibit an upward trend towards the end of the estimation period, the emerging market regions under consideration still remain segmented from the world and regional market. These results thus suggest that diversification into emerging market assets continue to produce substantial profits and that the asset pricing rules should reflect a state of partial integration. Our investigation, which addresses the evolution and formation of total risk premiums, confirm this empirically

I have developed time-varying coefficient models for recurrent event data to evaluate the temporal profiles for recurrence rate and covariate effects. There are three major parts in this dissertation. The first two parts propose a mixed Poisson process model with gamma frailties for single type recurrent events. The third part proposes a Bayesian joint model based on multivariate log-normal frailties for multi-type recurrent events. In the first part, I propose an approach based on penalized B-splines to obtain smooth estimation for both time-varying coefficients and the log baseline intensity. An EM algorithm is developed for parameter estimation. One issue with this approach is that the estimating procedure is conditional on smoothing parameters, which have to be selected by cross-validation or optimizing certain performance criterion. The procedure can be computationally demanding with a large number of time-varying coefficients. To achieve objective estimation of smoothing parameters, I propose a mixed-model representation approach for penalized splines. Spline coefficients are treated as random effects and smoothing parameters are to be estimated as variance components. An EM algorithm embedded with penalized quasi-likelihood approximation is developed to estimate the model parameters. The third part proposes a Bayesian joint model with time-varying coefficients for multi-type recurrent events. Bayesian penalized splines are used to estimate time-varying coefficients and the log baseline intensity. One challenge in Bayesian penalized splines is that the smoothness of a spline fit is considerably sensitive to the subjective choice of hyperparameters. I establish a procedure to objectively determine the hyperparameters through a robust prior specification. A Markov chain Monte Carlo procedure based on Metropolis-adjusted Langevin algorithms is developed to sample from the high-dimensional distribution of spline coefficients. The procedure includes a joint sampling scheme to achieve better convergence and mixing properties. Simulation studies in the second and third part have confirmed satisfactory model performance in estimating time-varying coefficients under different curvature and event rate conditions. The models in the second and third part were applied to data from a commercial truck driver naturalistic driving study. The application results reveal that drivers with 7-hours-or-less sleep prior to a shift have a significantly higher intensity after 8 hours of on-duty driving and that their intensity remains higher after taking a break. In addition, the results also show drivers' self-selection on sleep time, total driving hours in a shift, and breaks. These applications provide crucial insight into the impact of sleep time on driving performance for commercial truck drivers and highlights the on-road safety implications of insufficient sleep and breaks while driving. This dissertation provides flexible and robust tools to evaluate the temporal profile of intensity for recurrent events.
PHD

Dans cette thèse, nous étudions des modèles de moyennes conditionnelles de séries temporelles à valeurs entières. Tout d’abord, nous proposons l’estimateur de quasi maximum de vraisemblance de Poisson (EQMVP) pour les paramètres de la moyenne conditionnelle. Nous montrons que, sous des conditions générales de régularité, cet estimateur est consistant et asymptotiquement normal pour une grande classe de modèles. Étant donné que les paramètres de la moyenne conditionnelle de certains modèles sont positivement contraints, comme par exemple dans les modèles INAR (INteger-valued AutoRegressive) et les modèles INGARCH (INteger-valued Generalized AutoRegressive Conditional Heteroscedastic), nous étudions la distribution asymptotique de l’EQMVP lorsque le paramètre est sur le bord de l’espace des paramètres. En tenant compte de cette dernière situation, nous déduisons deux versions modifiées du test de Wald pour la significativité des paramètres et pour la moyenne conditionnelle constante. Par la suite, nous accordons une attention particulière au problème de validation des modèles des séries temporelles à valeurs entières en proposant un test portmanteau pour l’adéquation de l’ajustement. Nous dérivons la distribution jointe de l’EQMVP et des autocovariances résiduelles empiriques. Puis, nous déduisons la distribution asymptotique des autocovariances résiduelles estimées, et aussi la statistique du test. Enfin, nous proposons l’EQMVP pour estimer équation-par-équation (EpE) les paramètres de la moyenne conditionnelle des séries temporelles multivariées à valeurs entières. Nous présentons les hypothèses de régularité sous lesquelles l’EQMVP-EpE est consistant et asymptotiquement normal, et appliquons les résultats obtenus à plusieurs modèles des séries temporelles multivariées à valeurs entières
The framework of this PhD dissertation is the conditional mean count time seriesmodels. We propose the Poisson quasi-maximum likelihood estimator (PQMLE) for the conditional mean parameters. We show that, under quite general regularityconditions, this estimator is consistent and asymptotically normal for a wide classeof count time series models. Since the conditional mean parameters of some modelsare positively constrained, as, for example, in the integer-valued autoregressive (INAR) and in the integer-valued generalized autoregressive conditional heteroscedasticity (INGARCH), we study the asymptotic distribution of this estimator when the parameter lies at the boundary of the parameter space. We deduce a Waldtype test for the significance of the parameters and another Wald-type test for the constance of the conditional mean. Subsequently, we propose a robust and general goodness-of-fit test for the count time series models. We derive the joint distribution of the PQMLE and of the empirical residual autocovariances. Then, we deduce the asymptotic distribution of the estimated residual autocovariances and also of a portmanteau test. Finally, we propose the PQMLE for estimating, equation-by-equation (EbE), the conditional mean parameters of a multivariate time series of counts. By using slightly different assumptions from those given for PQMLE, we show the consistency and the asymptotic normality of this estimator for a considerable variety of multivariate count time series models

This master thesis deals with extension of the univariate GARCH model to multivari- ate models. We present individual models and deal with methods of their estimation. Then we describe some statistical tests for diagnosting the models. We have programmed in the statistical software R one of them - the Ling-Li test. Afterwards we apply selected models to real data of stock market index S&P 500, stock market index Russell 2000 and stocks of crude oil. For the GO-GARCH model, we compare all available estimation methods and show their differences. Then we compare the results of all models with each other and also with univariate models in terms of estimates of conditional variances, estimates of conditional correlations and also in terms of computational complexity. 1

In this work we will describe methods for modeling multivariate financial time series. We will concentrate on both modeling expected value by multi- variate Box-Jenkins processes and primarily on modeling conditional corre- lations and volatility. Our main object will be DCC (Dynamic Conditional Correlation) model, estimation of its parameters and some other general- izations. Then we will programme DCC model in statistical software R and apply on real data. In applications we will concentrate on problem of high dimension of financial time series and on modeling conditional correlations data with outliers.

Research Doctorate - Doctor of Philosophy (PhD)
Non-linear time series data is often generated by complex systems. While linear models provide a good first approximation of a system, often a more sophisticated non-linear model is required to properly account for the features of such data. Correctly accounting for these features should lead to the fitting of a more appropriate model. Determining the features exhibited by a particular data set is a difficult task, particularly for inexperienced modellers. Therefore, it is important to move towards a modelling paradigm where little to no user input is required, in order to open statistical modelling to users less experienced in MCMC. This sort of modelling process requires a general class of models that is able to account for the features found in most linear and non-linear data sets. One such class is the STAR-GARCH class of models. These are reasonably general models that permit regime changes in the conditional mean and allow for changes in the conditional covariance. In this thesis, we develop original algorithms that combine the tasks of parameter estimation and model selection for univariate and multivariate STAR-GARCH models. The model order of the conditional mean and the model index of the conditional covariance equation are included as parameters for the model requiring estimation. Combining the tasks of parameter estimation and model selection is facilitated through the Reversible Jump MCMC methodology. Other MCMC algorithms employed for the posterior distribution simulators are the Gibbs sampler, Metropolis-Hastings, Multiple-Try Metropolis and Delayed Rejection Metropolis-Hastings algorithms. The posterior simulation algorithms are successfully implemented in the statistical software program R, and their performance is tested in both extensive simulation studies and practical applications to real world data. The current literature on multivariate extensions of STAR, GARCH, and STAR-GARCH models is quite limited from a Bayesian perspective. The implementation of a set of estimation algorithms that not only provide parameter estimates but is also able to automatically fit the model with highest posterior probability is a significant and original contribution. The impact of such a contribution will hopefully be a step forward on the path towards the automation of time series modelling.

Indiana University-Purdue University Indianapolis (IUPUI)
A popular non-invasive brain activity measurement method is based on the functional magnetic resonance imaging (fMRI). Such data are frequently used to study functional connectivity (FC) defined as statistical association among two or more anatomically distinct fMRI signals (Friston, 1994). FC has emerged in recent years as a valuable tool for providing a deeper understanding of neurodegenerative diseases and neuropsychiatric disorders, such as Alzheimer's disease and autism. Information about complex association structure in high-dimensional fMRI data is often discarded by a calculating an average across complex spatiotemporal processes without providing an uncertainty measure around it. First, we propose a non-parametric approach to estimate the uncertainty of dynamic FC (dFC) estimates. Our method is based on three components: an extension of a boot strapping method for multivariate time series, recently introduced by Jentsch and Politis (2015); sliding window correlation estimation; and kernel smoothing. Second, we propose a two-step approach to analyze and summarize dFC estimates from a task-based fMRI study of social-to-heavy alcohol drinkers during stimulation with avors. In the first step, we apply our method from the first paper to estimate dFC for each region subject combination. In the second step, we use semiparametric additive mixed models to account for complex correlation structure and model dFC on a population level following the study's experimental design. Third, we propose to utilize the estimated dFC to study the system's modularity defined as the mutually exclusive division of brain regions into blocks with intra-connectivity greater than the one obtained by chance. As a result, we obtain brain partition suggesting the existence of common functionally-based brain organization. The main contribution of our work stems from the combination of the methods from the fields of statistics, machine learning and network theory to provide statistical tools for studying brain connectivity from a holistic, multi-disciplinary perspective.