The growth of the financial risk management industry has been motivated by the increased volatility of financial markets combined with the rapid innovation of derivatives. Since the 1970s, several financial crises have occurred globally with devastating consequences for financial and non-financial institutions and for the real economy. The most recent US subprime crisis led to enormous losses for financial and non-financial institutions and to a recession in many countries including the US and UK. A common lesson from these crises is that advanced financial risk management systems are required. Financial risk management is a continuous process of identifying, modeling, forecasting and monitoring risk exposures arising from financial investments. The Value at Risk (VaR) methodology has served as one of the most important tools used in this process. This quantitative tool, which was first invented by JPMorgan in its Risk-Metrics system in 1995, has undergone a considerable revolution and development during the last 15 years. It has now become one of the most prominent tools employed by financial institutions, regulators, asset managers and nonfinancial corporations for risk measurement. My PhD research undertakes a comprehensive and practical study of market risk modeling in modern finance using the VaR methodology. Two newly developed risk models are proposed in this research, which are derived by integrating volatility modeling and the quantile regression technique. Compared to the existing risk models, these two new models place more emphasis on dynamic risk adjustment. The empirical results on both real and simulated data shows that under certain circumstances, the risk prediction generated from these models is more accurate and efficient in capturing time varying risk evolution than traditional risk measures. Academically, the aim of this research is to make some improvements and extensions of the existing market risk modeling techniques. In practice, the purpose of this research is to support risk managers developing a dynamic market risk measurement system, which will function well for different market states and asset categories. The system can be used by financial institutions and non-financial institutions for either passive risk measurement or active risk control.

In this paper, we reconsider the question how monetary policy influences exchange rate dynamics. To this end, a vector autoregressive (VAR) model is combined with a two-country dynamic stochastic general equilibrium (DSGE) model. Instead of focusing exclusively on how monetary policy shocks affect the level of exchange rates, we also analyze how they impact exchange rate volatility. Since exchange rate volatility is not observed, we estimate it alongside the remaining quantities in the model. Our findings can be summarized as follows. Contractionary monetary policy shocks lead to an appreciation of the home currency, with exchange rate responses in the short-run typically undershooting their long-run level of appreciation. They also lead to an increase in exchange rate volatility. Historical and forecast error variance decompositions indicate that monetary policy shocks explain an appreciable amount of exchange rate movements and the corresponding volatility. / Series: Department of Economics Working Paper Series

This thesis extends the application of the localized realized volatility model created by Ying Chen, Wolfgang Karl Härdle, and Uta Pigorsch to other futures markets, particularly the CAC 40 and the NI 225. The research attempted to replicate results though ultimately, those results were invalidated by procedural difficulties.

Following the Probabilistic Reduction(PR) Approach, this paper proposes the Student’s Autoregressive (St-AR) Model, Student’s t Vector Autoregressive (St-VAR) Model and their heterogeneous versions, as an alternative to the various ARCH type models, to capture univariate and multivariate volatility. The St-AR and St-VAR models differ from the latter volatility models because they give rise to internally consistent statistical models that do not rely on ad-hoc specification and parameter restrictions, but model the conditional mean and conditional variance jointly. The univariate modeling is illustrated using the Real Effect Exchange Rate(REER) indices of three mainstream currencies in Asia (RMB, Hong Kong Dollar and Taiwan Dollar), while the multivariate volatility modeling is applied to investigate the relationship between the REER indices and stock price indices in mainland China, as well as the relationship between the stock prices in mainland China and Hong Kong. Following the PR methodology, the information gained in Mis-Specification(M-S) testing leads to respecification strategies from the original Normal-(V)AR models to the St-(V)AR models. The results from formal Mis-Specification (M-S) tests and forecasting performance indicate that the St-(V)AR models provide a more appropriate way to model volatility for certain types of speculative price data. / Ph. D.

A modelagem da volatilidade desempenha um papel fundamental em Econometria. Nesta dissertação são estudados a generalização dos modelos autorregressivos condicionalmente heterocedásticos conhecidos como GARCH e sua principal generalização multivariada, os modelos DCC-GARCH (Dynamic Condicional Correlation GARCH). Para os erros desses modelos são consideradas distribuições de probabilidade possivelmente assimétricas e leptocúrticas, sendo essas parametrizadas em função da assimetria e do peso nas caudas, necessitando assim de estimar esses parâmetros adicionais aos modelos. A estimação dos parâmetros dos modelos é feita sob a abordagem Bayesiana e devido às complexidades destes modelos, métodos computacionais baseados em simulações de Monte Carlo via Cadeias de Markov (MCMC) são utilizados. Para obter maior eficiência computacional os algoritmos de simulação da distribuição a posteriori dos parâmetros são implementados em linguagem de baixo nível. Por fim, a proposta de modelagem e estimação é exemplificada com dois conjuntos de dados reais / The modeling of volatility plays a fundamental role in Econometrics. In this dissertation are studied the generalization of known autoregressive conditionally heteroscedastic (GARCH) models and its main principal multivariate generalization, the DCCGARCH (Dynamic Conditional Correlation GARCH) models. For the errors of these models are considered distribution of probability possibility asymmetric and leptokurtic, these being parameterized as a function of asymmetry and the weight on the tails, thus requiring estimate the models additional parameters. The estimation of parameters is made under the Bayesian approach and due to the complexities of these models, methods computer-based simulations Monte Carlo Markov Chain (MCMC) are used. For more computational efficiency of simulation algorithms of posterior distribution of the parameters are implemented in low-level language. Finally, the proposed modeling and estimation is illustrated with two real data sets

A modelagem da volatilidade desempenha um papel fundamental em Econometria. Nesta dissertação são estudados a generalização dos modelos autorregressivos condicionalmente heterocedásticos conhecidos como GARCH e sua principal generalização multivariada, os modelos DCC-GARCH (Dynamic Condicional Correlation GARCH). Para os erros desses modelos são consideradas distribuições de probabilidade possivelmente assimétricas e leptocúrticas, sendo essas parametrizadas em função da assimetria e do peso nas caudas, necessitando assim de estimar esses parâmetros adicionais aos modelos. A estimação dos parâmetros dos modelos é feita sob a abordagem Bayesiana e devido às complexidades destes modelos, métodos computacionais baseados em simulações de Monte Carlo via Cadeias de Markov (MCMC) são utilizados. Para obter maior eficiência computacional os algoritmos de simulação da distribuição a posteriori dos parâmetros são implementados em linguagem de baixo nível. Por fim, a proposta de modelagem e estimação é exemplificada com dois conjuntos de dados reais / The modeling of volatility plays a fundamental role in Econometrics. In this dissertation are studied the generalization of known autoregressive conditionally heteroscedastic (GARCH) models and its main principal multivariate generalization, the DCCGARCH (Dynamic Conditional Correlation GARCH) models. For the errors of these models are considered distribution of probability possibility asymmetric and leptokurtic, these being parameterized as a function of asymmetry and the weight on the tails, thus requiring estimate the models additional parameters. The estimation of parameters is made under the Bayesian approach and due to the complexities of these models, methods computer-based simulations Monte Carlo Markov Chain (MCMC) are used. For more computational efficiency of simulation algorithms of posterior distribution of the parameters are implemented in low-level language. Finally, the proposed modeling and estimation is illustrated with two real data sets

Dans cette thèse, nous considérons deux types d'application des effets de rétroaction en finance. Ces effets entrent en jeu quand des participants de marché exécutent des séquences de transactions ou prennent part à des réactions en chaîne, ce qui engendre des pics d'activité. La première partie présente un modèle d'exécution optimale dynamique en présence d'un flux stochastique et exogène d'ordres de marché. Nous partons du modèle de référence d'Obizheva et Wang, qui définit un cadre d'exécution optimale avec un impact de prix mixte. Nous y ajoutons un flux d'ordres modélisé à l'aide de processus de Hawkes, qui sont des processus à sauts présentant une propriété d'auto-excitation. A l'aide de la théorie du contrôle stochastique, nous déterminons la stratégie optimale de manière analytique. Puis nous déterminons les conditions d'existence de Stratégies de Manipulation de Prix, telles qu'introduites par Huberman et Stanzl. Ces stratégies peuvent être exclues si l'auto-excitation du flux d'ordres se compense exactement avec la résilience du prix. Dans un deuxième temps, nous proposons une méthode de calibration du modèle, que nous appliquons sur des données financières à haute fréquence issues de cours d'actions du CAC40. Sur ces données, nous trouvons que le modèle explique une partie non-négligeable de la variance des prix. Une évaluation de la stratégie optimale en backtest montre que celle-ci est profitable en moyenne, mais que des coûts de transaction réalistes suffisent à empêcher les manipulations de prix. Ensuite, dans la deuxième partie de la thèse, nous nous intéressons à la modélisation de la volatilité intra-journalière. Dans la littérature, la plupart des modèles de volatilité rétroactive se concentrent sur l'échelle de temps journalière, c'est-à-dire aux variations de prix d'un jour sur l'autre. L'objectif est ici d'étendre ce type d'approche à des échelles de temps plus courtes. Nous présentons d'abord un modèle de type ARCH ayant la particularité de prendre en compte séparément les contributions des rendements passés intra-journaliers et nocturnes. Une méthode de calibration de ce modèle est étudiée, ainsi qu'une interprétation qualitative des résultats sur des rendements d'actions américaines et européennes. Dans le chapitre suivant, nous réduisons encore l'échelle de temps considérée. Nous étudions un modèle de volatilité à haute fréquence, dont l'idée est de généraliser le cadre des processus Hawkes pour mieux reproduire certaines caractéristiques empiriques des marchés. Notamment, en introduisant des effets de rétroaction quadratiques inspirés du modèle à temps discret QARCH nous obtenons une distribution en loi puissance pour la volatilité ainsi que de l'asymétrie temporelle / In this thesis we study feedback effects in finance and we focus on two of their applications. These effects stem from the fact that traders split meta-orders sequentially, and also from feedback loops. Therefore, one can observe clusters of activity and periods of relative calm. The first part introduces an dynamic optimal execution framework with an exogenous stochastic flow of market orders. Our starting point is the well-known model of Obizheva and Wang which defines an execution framework with both permanent and transient price impacts. We modify the price model by adding an order flow based on Hawkes processes, which are self-exciting jump processes. The theory of stochastic control allows us to derive the optimal strategy as a closed formula. Also, we discuss the existence of Price Manipulations Strategies in the sense of Huberman and Stanzl which can be excluded from the model if the self-exciting property of the order flow exactly compensates the resilience of the price. The next chapter studies a calibration protocol for the model, which we apply to tick-by-tick data from CAC40 stocks. On this dataset, the model is found to explain a significant part of the variance of prices. We then evaluate the optimal strategy with a series of backtests, which show that it is profitable on average, although realistic transaction costs can prevent manipulation strategies. In the second part of the thesis, we turn to intra-day volatility modeling. Previous works from the volatility feedback literature mainly focus on the daily time scale, i.e. on close-to-close returns. Our goal is to use a similar approach on shorter time scales. We first present an ARCH-type model which accounts for the contributions of past intra-day and overnight returns separately. A calibration method for the model is considered, that we use on US and European stocks, and we provide some qualitative insights on the results. The last chapter of the thesis is dedicated to a high-frequency volatility model. We introduce a continuous-time analogue of the QARCH framework, which is also a generalization of Hawkes processes. This new model reproduces several important stylized facts, in particular it generates a time-asymmetric and fat-tailed volatility process

<p>We discuss the Normal inverse Gaussian (NIG) distribution in modeling volatility in the financial markets. Refining the work of Barndorff-Nielsen (1997) and Andersson (2001), we introduce a new parameterization of the NIG distribution to build the GARCH(p,q)-NIG model. This new parameterization allows the model to be a strong GARCH in the sense of Drost and Nijman (1993). It also allows us to standardized the observed returns to be i.i.d., so that we can use standard inference methods when we evaluate the fit of the model.</p><p>We use the realized volatility (RV), calculated from intraday data, to standardize the returns of the ECU/USD foreign exchange rate. We show that normality cannot be rejected for the RV-standardized returns, i.e., the Mixture-of-Distributions Hypothesis (MDH) of Clark (1973) holds. {We build a link between the conditional RV and the conditional variance. This link allows us to use the conditional RV as a proxy for the conditional variance. We give an empirical justification of the GARCH-NIG model using this approximation.</p><p>In addition, we introduce a new General GARCH(p,q)-NIG model. This model has as special cases the Threshold-GARCH(p,q)-NIG model to model the leverage effect, the Absolute Value GARCH(p,q)-NIG model, to model conditional standard deviation, and the Threshold Absolute Value GARCH(p,q)-NIG model to model asymmetry in the conditional standard deviation. The properties of the maximum likelihood estimates of the parameters of the models are investigated in a simulation study.</p>

We discuss the Normal inverse Gaussian (NIG) distribution in modeling volatility in the financial markets. Refining the work of Barndorff-Nielsen (1997) and Andersson (2001), we introduce a new parameterization of the NIG distribution to build the GARCH(p,q)-NIG model. This new parameterization allows the model to be a strong GARCH in the sense of Drost and Nijman (1993). It also allows us to standardized the observed returns to be i.i.d., so that we can use standard inference methods when we evaluate the fit of the model. We use the realized volatility (RV), calculated from intraday data, to standardize the returns of the ECU/USD foreign exchange rate. We show that normality cannot be rejected for the RV-standardized returns, i.e., the Mixture-of-Distributions Hypothesis (MDH) of Clark (1973) holds. {We build a link between the conditional RV and the conditional variance. This link allows us to use the conditional RV as a proxy for the conditional variance. We give an empirical justification of the GARCH-NIG model using this approximation. In addition, we introduce a new General GARCH(p,q)-NIG model. This model has as special cases the Threshold-GARCH(p,q)-NIG model to model the leverage effect, the Absolute Value GARCH(p,q)-NIG model, to model conditional standard deviation, and the Threshold Absolute Value GARCH(p,q)-NIG model to model asymmetry in the conditional standard deviation. The properties of the maximum likelihood estimates of the parameters of the models are investigated in a simulation study.

Dans cette thèse, nous considérons deux types d'application des effets de rétroaction en finance. Ces effets entrent en jeu quand des participants de marché exécutent des séquences de transactions ou prennent part à des réactions en chaîne, ce qui engendre des pics d'activité. La première partie présente un modèle d'exécution optimale dynamique en présence d'un flux stochastique et exogène d'ordres de marché. Nous partons du modèle de référence d'Obizheva et Wang, qui définit un cadre d'exécution optimale avec un impact de prix mixte. Nous y ajoutons un flux d'ordres modélisé à l'aide de processus de Hawkes, qui sont des processus à sauts présentant une propriété d'auto-excitation. A l'aide de la théorie du contrôle stochastique, nous déterminons la stratégie optimale de manière analytique. Puis nous déterminons les conditions d'existence de Stratégies de Manipulation de Prix, telles qu'introduites par Huberman et Stanzl. Ces stratégies peuvent être exclues si l'auto-excitation du flux d'ordres se compense exactement avec la résilience du prix. Dans un deuxième temps, nous proposons une méthode de calibration du modèle, que nous appliquons sur des données financières à haute fréquence issues de cours d'actions du CAC40. Sur ces données, nous trouvons que le modèle explique une partie non-négligeable de la variance des prix. Une évaluation de la stratégie optimale en backtest montre que celle-ci est profitable en moyenne, mais que des coûts de transaction réalistes suffisent à empêcher les manipulations de prix. Ensuite, dans la deuxième partie de la thèse, nous nous intéressons à la modélisation de la volatilité intra-journalière. Dans la littérature, la plupart des modèles de volatilité rétroactive se concentrent sur l'échelle de temps journalière, c'est-à-dire aux variations de prix d'un jour sur l'autre. L'objectif est ici d'étendre ce type d'approche à des échelles de temps plus courtes. Nous présentons d'abord un modèle de type ARCH ayant la particularité de prendre en compte séparément les contributions des rendements passés intra-journaliers et nocturnes. Une méthode de calibration de ce modèle est étudiée, ainsi qu'une interprétation qualitative des résultats sur des rendements d'actions américaines et européennes. Dans le chapitre suivant, nous réduisons encore l'échelle de temps considérée. Nous étudions un modèle de volatilité à haute fréquence, dont l'idée est de généraliser le cadre des processus Hawkes pour mieux reproduire certaines caractéristiques empiriques des marchés. Notamment, en introduisant des effets de rétroaction quadratiques inspirés du modèle à temps discret QARCH nous obtenons une distribution en loi puissance pour la volatilité ainsi que de l'asymétrie temporelle / In this thesis we study feedback effects in finance and we focus on two of their applications. These effects stem from the fact that traders split meta-orders sequentially, and also from feedback loops. Therefore, one can observe clusters of activity and periods of relative calm. The first part introduces an dynamic optimal execution framework with an exogenous stochastic flow of market orders. Our starting point is the well-known model of Obizheva and Wang which defines an execution framework with both permanent and transient price impacts. We modify the price model by adding an order flow based on Hawkes processes, which are self-exciting jump processes. The theory of stochastic control allows us to derive the optimal strategy as a closed formula. Also, we discuss the existence of Price Manipulations Strategies in the sense of Huberman and Stanzl which can be excluded from the model if the self-exciting property of the order flow exactly compensates the resilience of the price. The next chapter studies a calibration protocol for the model, which we apply to tick-by-tick data from CAC40 stocks. On this dataset, the model is found to explain a significant part of the variance of prices. We then evaluate the optimal strategy with a series of backtests, which show that it is profitable on average, although realistic transaction costs can prevent manipulation strategies. In the second part of the thesis, we turn to intra-day volatility modeling. Previous works from the volatility feedback literature mainly focus on the daily time scale, i.e. on close-to-close returns. Our goal is to use a similar approach on shorter time scales. We first present an ARCH-type model which accounts for the contributions of past intra-day and overnight returns separately. A calibration method for the model is considered, that we use on US and European stocks, and we provide some qualitative insights on the results. The last chapter of the thesis is dedicated to a high-frequency volatility model. We introduce a continuous-time analogue of the QARCH framework, which is also a generalization of Hawkes processes. This new model reproduces several important stylized facts, in particular it generates a time-asymmetric and fat-tailed volatility process