Submitted by Eduardo Alonso Marza dos Santos (eduardo.marza.santos@gmail.com) on 2015-06-21T10:30:55Z No. of bitstreams: 1 Eduardo_A_M_Santos.pdf: 646820 bytes, checksum: aaba122a576d7c75ad0e5803539c25d4 (MD5) / Approved for entry into archive by Suzinei Teles Garcia Garcia (suzinei.garcia@fgv.br) on 2015-06-22T11:46:18Z (GMT) No. of bitstreams: 1 Eduardo_A_M_Santos.pdf: 646820 bytes, checksum: aaba122a576d7c75ad0e5803539c25d4 (MD5) / Made available in DSpace on 2015-06-22T11:56:18Z (GMT). No. of bitstreams: 1 Eduardo_A_M_Santos.pdf: 646820 bytes, checksum: aaba122a576d7c75ad0e5803539c25d4 (MD5) Previous issue date: 2015-05-28 / The dissertation goal is to quantify the tail risk premium embedded into hedge funds' returns. Tail risk is the probability of extreme large losses. Although it is a rare event, asset pricing theory suggests that investors demand compensation for holding assets sensitive to extreme market downturns. By de nition, such events have a small likelihood to be represented in the sample, what poses a challenge to estimate the e ects of tail risk by means of traditional approaches such as VaR. The results show that it is not su cient to account for the tail risk stemming from equities markets. Active portfolio management employed by hedge funds demand a speci c measure to estimate and control tail risk. Our proposed factor lls that void inasmuch it presents explanatory power both over the time series as well as the cross-section of funds' returns. / O objetivo do trabalho é quanti car o prêmio de risco de cauda presente nos retornos de fundos de investimento americanos. Risco de cauda é o risco de perdas excepcionalmente elevadas. Apesar de ser um evento raro, a teoria de apreçamento de ativos sugere que os investidores exigem um prêmio de risco para reter ativos expostos a eventos negativos extremos (eventos de cauda). Por de nição, observações extremas têm baixa probabilidade de estarem presentes na amostra, o que di culta a estimação dos impactos de risco de cauda sobre os retornos e reduz o poder de técnicas tradicionais como VaR. Os resultados indicam que não é su ciente controlar somente para o risco de cauda do mercado de capitais. A gestão ativa de portfólio por parte dos gestores de fundos requer uma medida própria para estimação e o controle de risco de cauda. O fator de risco de cauda que propomos cumpre este papel ao apresentar poder explicativo tanto na série temporal dos retornos quanto no corte transversal.

Submitted by Edimilson Costa Lucas (costalucas@yahoo.com) on 2015-11-10T13:28:51Z No. of bitstreams: 1 EdimilsonCostaLucas_TESE.pdf: 2525096 bytes, checksum: 88b5fc4a39e14115350d9a7fddece121 (MD5) / Approved for entry into archive by Maria Tereza Fernandes Conselmo (maria.conselmo@fgv.br) on 2015-11-11T12:34:10Z (GMT) No. of bitstreams: 1 EdimilsonCostaLucas_TESE.pdf: 2525096 bytes, checksum: 88b5fc4a39e14115350d9a7fddece121 (MD5) / Made available in DSpace on 2015-11-11T12:38:48Z (GMT). No. of bitstreams: 1 EdimilsonCostaLucas_TESE.pdf: 2525096 bytes, checksum: 88b5fc4a39e14115350d9a7fddece121 (MD5) Previous issue date: 2015-10-19 / The occurrence of extreme weather events, such as increased temperature, hurricanes, floods and droughts has been increasingly common around of the world. The finance literature has documented efforts directed to the assessment of economic impacts from climate change that can bring significant consequences for the world economy. However, especially in Brazil, a key emerging market, little has been studied mainly with a view to assessing the impacts of climate events in the company level. Thus, this thesis analyzes, in an unprecedented manner, the impact of weather events on the value of companies belonging to two high national interest industries, in the form of two essays. First it analyzes the impact of extreme rainfall on the stock price of the Brazilian food sector. Therefore, it is conducted the research using daily data in share prices of six companies of this industry. From the location to the main area of activity of these companies, they are considered their daily data on extreme rainfall. With the use of hybrid methodology ARMA-GARCH-GPD, it was found that the evaluated companies, the extreme rainfall impacted significantly in more than half of the 198 days of extreme rainfall between 02/28/2005 and 12/30/2014, resulting in average losses daily around 1.97% on the day after the heavy rainfall. In terms of market value, this represents total average loss of around US$ 682.15 million in a single day. Second it evaluates the impact of climate variables and location on the value of companies in the energy sector in Brazil, from data on companies in the Brazilian electricity sector, as well as rainfall, temperature and geographical location of the companies. From the analysis of data in static panel and spatial panel, the results suggest that temperature and precipitation have significant effect on the value of these companies. This study can contribute in the process of structuring and creating a weather derivatives market in Brazil. / A ocorrência de eventos climáticos extremos, tais como aumento da temperatura, furacões, enchentes e secas, tem sido cada vez mais frequente ao redor do mundo. A literatura de finanças tem documentado esforços dirigidos à avaliação de impactos econômicos oriundos das variações climáticas, com consequências significantes na economia mundial. Entretanto, especialmente no Brasil, um dos principais mercados emergentes, pouco tem sido pesquisado, sobretudo com vistas à avaliação dos impactos de eventos climáticos no nível das empresas. Sendo assim, esta tese analisa, de forma inédita, o impacto de eventos climáticos sobre o valor de empresas pertencentes a duas indústrias de elevado interesse nacional, sob a forma de dois ensaios. Em primeiro lugar analisa-se o impacto de chuvas extremas sobre o preço de ações do setor de alimentos brasileiro. Para tanto, é conduzida a pesquisa empregando dados diários do preço de ações de seis empresas dessa indústria. A partir da localização da principal região de atuação dessas empresas, são considerados os respectivos dados diários referentes às chuvas extremas. Com o emprego da metodologia híbrida ARMA-GARCH-GPD, constatou-se que, nas empresas avaliadas, as chuvas extremas impactaram significantemente em mais da metade dos 198 dias de chuvas extremas ocorridos entre 28/02/2005 e 30/12/2014, acarretando perdas médias diárias ao redor de 1,97% no dia posterior a chuva extrema. Em termos de valor de mercado, isso representa perda média total ao redor de US$682,15 mi em um único dia. Em segundo lugar avalia-se o impacto de variáveis climáticas e localização sobre o valor das empresas do setor de energia do Brasil, a partir de dados referentes às empresas do setor elétrico brasileiro, bem como precipitação pluviométrica, temperatura e localização geográfica das empresas. A partir da análise de dados em painel estático e painel espacial, os resultados sugerem que temperatura e precipitação pluviométrica têm efeito significante sobre o valor dessas empresas. O presente estudo pode vir a contribuir no processo de estruturação e criação de um mercado de derivativos climáticos no Brasil.

Made available in DSpace on 2010-04-20T21:00:30Z (GMT). No. of bitstreams: 3 marciocecilioturma2004.pdf.jpg: 19602 bytes, checksum: 0772484d1cb46349dfbfb25620b5cdae (MD5) marciocecilioturma2004.pdf: 859203 bytes, checksum: 346a3e7d5751118ff894a182d7512b56 (MD5) marciocecilioturma2004.pdf.txt: 86793 bytes, checksum: e0c91b2715fc569bc6ec29bfce078e69 (MD5) Previous issue date: 2007-01-23T00:00:00Z / Há forte evidência que os retornos das séries financeiras apresentam caudas mais pesadas que as da distribuição normal, principalmente em mercados emergentes. No entanto, muitos modelos de risco utilizados pelas instituições financeiras baseiam-se em normalidade condicional ou não condicional, reduzindo a acurácia das estimativas. Os recentes avanços na Teoria de Valores Extremos permitem sua aplicação na modelagem de risco, como por exemplo, na estimação do Valor em Risco e do requerimento de capital. Este trabalho verifica a adequação de um procedimento proposto por McNeil e Frey [1999] para estimação do Valor em Risco e conseqüente requerimento de capital às principais séries financeiras de retornos do Brasil. Tal procedimento semi-paramétrico combina um modelo GARCH ajustado por pseudo máxima verossimilhança para estimação da volatilidade corrente com a Teoria de Valores Extremos para estimação das caudas da distribuição das inovações do modelo GARCH. O procedimento foi comparado através de backtestings com outros métodos mais comuns de estimação de VaR que desconsideram caudas pesadas das inovações ou a natureza estocástica da volatilidade. Concluiu-se que o procedimento proposto por McNeil e Frey [1999] mostrou melhores resultados, principalmente para eventos relacionados a movimentos negativos nos mercados . Futuros trabalhos consistirão no estudo de uma abordagem multivariada de grandes dimensões para estimação de VaR e requerimento de capital para carteiras de investimentos. / There is a strong evidence that financial return series are heavy-tailed, mostly in emerging markets. However, most of the risk models used by financial institutions are based in conditional or non-conditional normality, which reduces the accuracy of the estimates. The recent advances in Extreme Value Theory permit its application to risk measuring, such as Value at Risk and capital adequacy estimates. This work verifies the adequacy of a procedure proposed by McNeil and Frey [1999] to VaR and consequent capital requirement estimates for the main financial return series in Brazil. This semi parametric procedure combines a pseudo-maximumlikelihood fitting GARCH model to estimate the current volatility and the Extreme Value Theory (EVT) to estimate the tails of the innovations distribution of the GARCH model. Using backtestings the procedure was compared to other common methods of VaR estimation that disregard heavy tails of the innovations or the stochastic nature of the volatility. The procedure proposed by McNeil and Frey [1999] showed better results, mostly for negative events in the financial market2 . Further works will consist of studying a high dimensional multivariate approach to estimate VaR and capital requirements for portfolios of investment instruments.

Cette thèse s'inscrit dans le contexte de la statistique des valeurs extrêmes. Elle y apporte deux contributions principales. Dans la littérature récente en statistique des valeurs extrêmes, un modèle de queues de distributions a été introduit afin d'englober aussi bien les lois de type Pareto que les lois à queue de type Weibull. Les deux principaux types de décroissance de la fonction de survie sont ainsi modélisés. Un estimateur des quantiles extrêmes a été déduit de ce modèle mais il dépend de deux paramètres inconnus, le rendant inutile dans des situations pratiques. La première contribution de cette thèse est de proposer des estimateurs de ces paramètres. Insérer nos estimateurs dans l'estimateur des quantiles extrêmes précédent permet alors d'estimer des quantiles extrêmes pour des lois de type Pareto aussi bien que pour des lois à queue de type Weibull d'une façon unifiée. Les lois asymptotiques de nos trois nouveaux estimateurs sont établies et leur efficacité est illustrée sur des données simulées et sur un jeu de données réelles de débits de la rivière Nidd se situant dans le Yorkshire en Angleterre. La seconde contribution de cette thèse consiste à introduire et estimer une nouvelle mesure de risque appelé Conditional Tail Moment. Elle est définie comme le moment d'ordre a>0 de la loi des pertes au-delà du quantile d'ordre p appartenant à ]0,1[ de la fonction de survie. Estimer le Conditional Tail Moment permet d'estimer toutes les mesures de risque basées sur les moments conditionnels telles que la Value-at-Risk, la Conditional Tail Expectation, la Conditional Value-at-Risk, la Conditional Tail Variance ou la Conditional Tail Skewness. Ici, on s'intéresse à l'estimation de ces mesures de risque dans le cas de pertes extrêmes c'est-à-dire lorsque p tend vers 0 lorsque la taille de l'échantillon augmente. On suppose également que la loi des pertes est à queue lourde et qu'elle dépend d'une covariable. Les estimateurs proposés combinent des méthodes d'estimation non-paramétrique à noyau avec des méthodes issues de la statistique des valeurs extrêmes. Le comportement asymptotique de nos estimateurs est établi et illustré aussi bien sur des données simulées que sur des données réelles de pluviométrie provenant de la région Cévennes-Vivarais. / This thesis can be viewed within the context of extreme value statistics. It provides two main contributions to this subject area. In the recent literature on extreme value statistics, a model on tail distributions which encompasses Pareto-type distributions as well as Weibull tail-distributions has been introduced. The two main types of decreasing of the survival function are thus modeled. An estimator of extreme quantiles has been deduced from this model, but it depends on two unknown parameters, making it useless in practical situations. The first contribution of this thesis is to propose estimators of these parameters. Plugging our estimators in the previous extreme quantiles estimator allows us to estimate extreme quantiles from Pareto-type and Weibull tail-distributions in an unified way. The asymptotic distributions of our three new estimators are established and their efficiency is illustrated on a simulation study and on a real data set of exceedances of the Nidd river in the Yorkshire (England). The second contribution of this thesis is the introduction and the estimation of a new risk measure, the so-called Conditional Tail Moment. It is defined as the moment of order a>0 of the loss distribution above the quantile of order p in (0,1) of the survival function. Estimating the Conditional Tail Moment permits to estimate all risk measures based on conditional moments such as the Value-at-Risk, the Conditional Tail Expectation, the Conditional Value-at-Risk, the Conditional Tail Variance or the Conditional Tail Skewness. Here, we focus on the estimation of these risk measures in case of extreme losses i.e. when p converges to 0 when the size of the sample increases. It is moreover assumed that the loss distribution is heavy-tailed and depends on a covariate. The estimation method thus combines nonparametric kernel methods with extreme-value statistics. The asymptotic distribution of the estimators is established and their finite sample behavior is illustrated both on simulated data and on a real data set of daily rainfalls in the Cévennes-Vivarais region (France).

Neste trabalho propomos uma abordagem Bayesiana não-paramétrica para a modelagem de dados com comportamento extremo. Tratamos o parâmetro de locação μ da distribuição generalizada de valor extremo como uma função aleatória e assumimos um processo Gaussiano para tal função (Rasmussem & Williams 2006). Esta situação leva à intratabilidade analítica da distribuição a posteriori de alta dimensão. Para lidar com este problema fazemos uso do método Hamiltoniano de Monte Carlo em variedade Riemanniana que permite a simulação de valores da distribuição a posteriori com forma complexa e estrutura de correlação incomum (Calderhead & Girolami 2011). Além disso, propomos um modelo de série temporal autoregressivo de ordem p, assumindo a distribuição generalizada de valor extremo para o ruído e determinamos a respectiva matriz de informação de Fisher. No decorrer de todo o trabalho, estudamos a qualidade do algoritmo em suas variantes através de simulações computacionais e apresentamos vários exemplos com dados reais e simulados. / In this work we propose a Bayesian nonparametric approach for modeling extreme value data. We treat the location parameter μ of the generalized extreme value distribution as a random function following a Gaussian process model (Rasmussem & Williams 2006). This configuration leads to no closed-form expressions for the highdimensional posterior distribution. To tackle this problem we use the Riemannian Manifold Hamiltonian Monte Carlo algorithm which allows samples from the posterior distribution with complex form and non-usual correlation structure (Calderhead & Girolami 2011). Moreover, we propose an autoregressive time series model assuming the generalized extreme value distribution for the noise and obtained its Fisher information matrix. Throughout this work we employ some computational simulation studies to assess the performance of the algorithm in its variants and show many examples with simulated and real data-sets.

Afin d'estimer les forces statiques équivalentes du vent, qui produisent les réponses quasi-statiques et dynamiques extrêmes dans les structures soumises au champ de pression instationnaire induit par les effets du vent, une nouvelle méthode probabiliste est proposée. Cette méthode permet de calculer les forces statiques équivalentes du vent pour les structures avec des écoulements aérodynamiques complexes telles que les toitures de stade, pour lesquelles le champ de pression n'est pas gaussien et pour lesquelles la réponse dynamique de la structure ne peut être simplement décrite en utilisant uniquement les premiers modes élastiques (mais nécessitent une bonne représentation des réponses quasi-statiques). Généralement, les mesures en soufflerie du champ de pression instationnaire appliqué à une structure dont la géométrie est complexe ne suffisent pas pour construire une estimation statistiquement convergée des valeurs extrêmes des réponses dynamiques de la structure. Une telle convergence est nécessaire pour l'estimation des forces statiques équivalentes afin de reproduire les réponses dynamiques extrêmes induites par les effets du vent en tenant compte de la non-gaussianité du champ de pression aléatoire instationnaire. Dans ce travail, (1) un générateur de réalisation du champ de pression instationnaire non gaussien est construit en utilisant les réalisations qui sont mesurées dans la soufflerie à couche limite turbulente; ce générateur basé sur une représentation en chaos polynomiaux permet de construire un grand nombre de réalisations indépendantes afin d'obtenir la convergence des statistiques des valeurs extrêmes des réponses dynamiques, (2) un modèle d'ordre réduit avec des termes d'accélération quasi-statique est construit et permet d'accélérer la convergence des réponses dynamiques de la structure en n'utilisant qu'un petit nombre de modes élastiques, (3) une nouvelle méthode probabiliste est proposée pour estimer les forces statiques équivalentes induites par les effets du vent sur des structures complexes décrites par des modèles éléments finis, en préservant le caractère non gaussien et sans introduire le concept d'enveloppes des réponses. L'approche proposée est validée expérimentalement avec une application relativement simple et elle est ensuite appliquée à une structure de toiture de stade pour laquelle des mesures expérimentales de pressions instationnaires ont été effectuées dans la soufflerie à couche limite turbulente / In order to estimate the equivalent static wind loads, which produce the extreme quasi-static and dynamical responses of structures submitted to random unsteady pressure field induced by the wind effects, a new probabilistic method is proposed. This method allows for computing the equivalent static wind loads for structures with complex aerodynamic flows such as stadium roofs, for which the pressure field is non-Gaussian, and for which the dynamical response of the structure cannot simply be described by using only the first elastic modes (but require a good representation of the quasi-static responses). Usually, the wind tunnel measurements of the unsteady pressure field applied to a structure with complex geometry are not sufficient for constructing a statistically converged estimation of the extreme values of the dynamical responses. Such a convergence is necessary for the estimation of the equivalent static loads in order to reproduce the extreme dynamical responses induced by the wind effects taking into account the non-Gaussianity of the random unsteady pressure field. In this work, (1) a generator of realizations of the non-Gaussian unsteady pressure field is constructed by using the realizations that are measured in the boundary layer wind tunnel; this generator based on a polynomial chaos representation allows for generating a large number of independent realizations in order to obtain the convergence of the extreme value statistics of the dynamical responses, (2) a reduced-order model with quasi-static acceleration terms is constructed, which allows for accelerating the convergence of the structural dynamical responses by using only a small number of elastic modes of the structure, (3) a novel probabilistic method is proposed for estimating the equivalent static wind loads induced by the wind effects on complex structures that are described by finite element models, preserving the non-Gaussian property and without introducing the concept of responses envelopes. The proposed approach is experimentally validated with a relatively simple application and is then applied to a stadium roof structure for which experimental measurements of unsteady pressures have been performed in boundary layer wind tunnel

Nous appliquons la théorie asymptotique des expériences statistiques à des problèmes liés aux valeurs extrêmes. Quatre modèles semi-paramétriques sont envisagés. Tout d'abord le modèle d'échantillonnage de fonction de répartition de type Pareto. L'index de Pareto est le paramètre d'intérêt tandis que la fonction à variation lente, qui intervient dans la décomposition de la fonction de survie, joue le rôle de nuisance. Nous considérons ensuite des observations i.i.d. de fonction de répartition de type Weibull. Le troisième modèle étudié est un modèle de régression. On considère des couples d'observations $(Y_i,X_i)$ indépendants, les v.a. $X_i$ sont i.i.d. de loi connue et on suppose que la fonction de répartition de la loi de $Y$ conditionnellement à $X$ est de type Pareto, avec une fonction à variation lente et un index $gamma$ qui dépendent de $X$. On fait l'hypothèse que la fonction $gamma$ a une forme quelconque mais connue, qui dépend d'un paramètre $\ / Doctorat en sciences, Orientation statistique / info:eu-repo/semantics/nonPublished

Dans cette thèse, nous étudions la garantie tempête consacrée aux dommages causés par le vent et un développement de l'assurance comportementale à travers le risque automobile. Nous associons des informations extérieures comme la vitesse du vent aux données de l'assurance. Nous proposons la construction d'un indice tempête pour compléter et renforcer l'évaluation des dégâts causés par les tempêtes majeures. Nous définissons ensuite un partage du territoire français en 6 zones tempêtes, dépendant des corrélations extrêmes de vent, pour tester plusieurs scénarios. Ces différents tests et considérations nous permettent d'améliorer notre indice tempête. Nous nous appuyons sur les modèles de la théorie des valeurs extrêmes pour montrer l'impact de la variabilité sur le calcul des périodes de retour et besoins en fonds propres. Nous soulignons ainsi les difficultés rencontrées pour dégager des résultats robustes en lien avec les évènements extrêmes. Pour ce qui est de l'assurance automobile, nous testons différentes méthodes pour répondre aux évolutions techniques et réglementaires. Nous caractérisons la partition homme / femme en utilisant la procédure logistique, l'analyse des correspondances multiples ou les arbres de classification. Nous montrons qu'il est possible de compenser l'absence de la variable sexe par d'autres informations spécifiques à l'assuré ou à son véhicule et en particulier l'utilisation de relevés kilométriques. Enfin, nous nous intéressons à l'expérience acquise par les conducteurs novices. Nous étudions le comportement sur la route de l'assuré pour créer de nouvelles classes de risques / In this Ph.D. Dissertation we study the storm guarantee dedicated to the damage caused by the wind and a development of the behavioral insurance through the automobile risk. We associate external information like the wind speed to insurance data. We propose the construction of a storm index to complete and strengthen the evaluation of the damages caused by the major storms. Then we define a partition of the French territory in 6 zones storms, depending on extreme wind correlations to test several scenarios. These various tests and considerations allow us to improve our storm index. We lean on extreme value theory models to show the impact of the variability on the calculation of return periods and capital requirements. We underline the difficulties to obtain strong results in connection with the extreme events. Concerning car insurance, we test various methods to answer the technical and legal evolutions. We characterize the man/woman partition by using the logistic procedure, the multiple correspondence analysis or the classification trees. We show that it is possible to compensate for the absence of the sex variable with other information specific to the insurants or to their vehicle and in particular the use of kilometric data. Finally, we are interested in the acquired experience by young drivers. We study the behavior on the road of the insurants to create new classes of risks

L’objectif des travaux de la thèse est d’étudier les propriétés statistiques de correction des prévisionsde température et de les appliquer au système des prévisions d’ensemble (SPE) de MétéoFrance. Ce SPE est utilisé dans la gestion du système électrique, à EDF R&D, il contient 51membres (prévisions par pas de temps) et fournit des prévisions à 14 jours. La thèse comportetrois parties. Dans la première partie on présente les SPE, dont le principe est de faire tournerplusieurs scénarios du même modèle avec des données d’entrée légèrement différentes pour simulerl’incertitude. On propose après des méthodes statistiques (la méthode du meilleur membre etla méthode bayésienne) que l’on implémente pour améliorer la précision ou la fiabilité du SPEdont nous disposons et nous mettons en place des critères de comparaison des résultats. Dansla deuxième partie nous présentons la théorie des valeurs extrêmes et les modèles de mélange etnous proposons des modèles de mélange contenant le modèle présenté dans la première partieet des fonctions de distributions des extrêmes. Dans la troisième partie nous introduisons larégression quantile pour mieux estimer les queues de distribution. / The thesis has for objective to study new statistical methods to correct temperature predictionsthat may be implemented on the ensemble prediction system (EPS) of Meteo France so toimprove its use for the electric system management, at EDF France. The EPS of Meteo Francewe are working on contains 51 members (forecasts by time-step) and gives the temperaturepredictions for 14 days. The thesis contains three parts: in the first one we present the EPSand we implement two statistical methods improving the accuracy or the spread of the EPS andwe introduce criteria for comparing results. In the second part we introduce the extreme valuetheory and the mixture models we use to combine the model we build in the first part withmodels for fitting the distributions tails. In the third part we introduce the quantile regressionas another way of studying the tails of the distribution.

The thesis deals with extreme value distributions. The theoretical part is devoted to the basics of extreme value theory and to the characterization of extreme value distributions. There is the limit theorem for distributions of the maximum formulated and characteristics of the extreme value distributions deduced. There are parameter estimates for Weibull, lognormal and exponential distributions inferred using method of maximum likelihood and method of moments. There is also the theory of censored samples described. The practical part is devoted to statistical analysis of rainfall.