Orientador: Henrique Nunes de Oliveira / Banca: Raysildo Barbosa Lobo / Banca: Danísio Prado Munari / Resumo: O presente trabalho objetivou comparar, utilizando dados de um experimento de Microarray com um delineamento simples, os resultados de diferentes testes estatísticos a fim de verificar suas características na detecção de diferenças no nível de expressão dos genes. Os dados foram provenientes da South Dakota State University-EUA, do Department of Biology and Microbiology, Department of Animal Science, onde toda a parte experimental foi realizada. O material biológico envolveu quatro aves infectadas e quatro não infectadas com o vírus de bronquite infecciosa (IBV). O RNA utilizado foi extraído da camada epitelial da traquéia de animais controle e infectados com o vírus da IBV e, após a transcrição reversa foi marcado com os corantes fluorescentes (Cy3 e Cy5) e hibridizados com o microarray 13k cDNA de aves (FHCRC, Seattle, WA). A análise de dados dos resultados do experimento de microarray englobou dois estágios, sendo o primeiro denominado de Normalização, em que os dados foram pré-processados utilizando o procedimento Loess. A seguir foram realizadas as análises estatísticas propriamente ditas com testes de significância. Utilizou-se um modelo simples de ANOVA e aplicaram-se diferentes metodologias de análise. A análise das imagens revelou que dos 16192 spots em cada slide, apenas 10.926 puderam ser lidos sem defeitos no primeiro slide, 11.633 no segundo slide, 12577 no terceiro e 13.154 no quarto slide. A grande maioria dos spots em branco e controles negativos apresentou defeitos que determinaram sua eliminação. Um total de 13.597 spots foi lido no conjunto dos quatro slides, mas apenas 9.853 spots estavam representados em todos os slides. Concluiu-se que os experimentos de microarray, por tratarem de um conjunto muito grande de observações a serem analisados requerem análises estatísticas específicas. O método de Cui et al. (2005) reduziu... (Resumo completo, clicar acesso eletrônico abaixo) / Abstract: The aim of this research was to compare, using real data of an experiment of Microarray with a simple design, the results of different statistical tests in order to verify their characteristics in the detection of differences in the level of expression of the genes. The data were coming of South Dakota State University-EUA, of the Department of Biology and Microbiology, Department Animal of Science, where the whole experimental part was accomplished. The biological material involved four infected animals and four no infected with the virus of infectious bronchitis (IBV). Used RNA was extracted of the layer epitelial of the windpipe of animals control and infected with the virus of IBV and, after the reverse transcription it was marked with the fluorescent colors (Cy3 and Cy5) and hybridization with the microarray 13k cDNA of birds (FHCRC, Seattle, WA). The analysis of data of the results of the microarray experiment included two apprenticeships, being the first denominated of Normalization, in that the data were pre-processed using the procedure Loess. To follow the statistical analyses they were accomplished properly said through real data with significant tests. A simple model of ANOVA was used and different analysis methodologies were applied. The analysis of the images revealed that of the 16192 spots in each slide, only 10.926 could be read without defects in the first slide, 11.633 in the second slide, 12577 in the third slide and 13.154 in the fourth slide. The great majority of the spots in white and negative controls presented defects that determined it elimination. A total of 13.597 spots was read in the group of the four slides, but only 9.853 spots were represented in all of the slides. It was ended that the microarray experiments, for they treat of a very big group of observations to be analyzed request specific statistical analyses. The method of Cui et al. (2005) it reduced... (Complete abstract click electronic access below) / Mestre

Submitted by Leonardo Ishihara Monaka (l.monaka@terra.com.br) on 2018-10-25T20:08:35Z No. of bitstreams: 1 TCC_LMonaka_V finallissima_Rev2.pdf: 870927 bytes, checksum: 7bbead367066752977f09383d6cea552 (MD5) / Approved for entry into archive by Simone de Andrade Lopes Pires (simone.lopes@fgv.br) on 2018-10-26T00:31:57Z (GMT) No. of bitstreams: 1 TCC_LMonaka_V finallissima_Rev2.pdf: 870927 bytes, checksum: 7bbead367066752977f09383d6cea552 (MD5) / Approved for entry into archive by Suzane Guimarães (suzane.guimaraes@fgv.br) on 2018-10-26T13:37:50Z (GMT) No. of bitstreams: 1 TCC_LMonaka_V finallissima_Rev2.pdf: 870927 bytes, checksum: 7bbead367066752977f09383d6cea552 (MD5) / Made available in DSpace on 2018-10-26T13:37:50Z (GMT). No. of bitstreams: 1 TCC_LMonaka_V finallissima_Rev2.pdf: 870927 bytes, checksum: 7bbead367066752977f09383d6cea552 (MD5) Previous issue date: 2018-09-26 / Nos últimos anos, os operadores de Telecom no Brasil têm sido protagonistas para prover os investimentos da infraestrutura necessária para atender as demandas da nova era digital, e ao mesmo tempo, continuar gerando valor aos acionistas. A maior preocupação dos executivos em fazer cumprir o orçamento pactuado com a matriz, tem deixado para segundo plano a oportunidade de alavancar o debate de como a ferramenta de orçamento pode fornecer informações úteis para a empresa gerar vantagem competitiva. O objetivo deste trabalho de intervenção é realizar uma consultoria na empresa Telefónica | Vivo, a fim de avaliar se e como a aplicação da metodologia do orçamento flexível e análise de variação orçamentária na área de comercialização de terminais móveis da unidade de negócios B2B (Business-to-business) da operadora Vivo pode contribuir como instrumento de gestão para alavancar a competitividade da empresa. O estudo foi elaborado utilizando-se dados históricos reais da empresa, a partir dos quais foi efetuada uma comparação entre as informações geradas historicamente pela aplicação da metodologia orçamento estático (atualmente utilizada na empresa) e as geradas pela aplicação do orçamento flexível, objeto deste trabalho. Para validar essa comparação, foram realizadas entrevistas estruturadas com os executivos que possuem interação direta com o tema para avaliar a percepção desses usuários chave sobre a utilidade das informações geradas, utilizando um questionário padrão. O resultado do estudo revelou que a análise de variação utilizando o método do orçamento estático não provê maior detalhamento sobre quais foram as variáveis causais das variações. O resultado da análise utilizando a metodologia do orçamento flexível permitiu separar as variações por variável causal, antes sequer vistos de forma aplicada, e a maneira estruturada e transparente foi ponto crucial para a percepção positiva e unânime dos executivos. Por fim, espera-se que este trabalho aplicado possa contribuir para ampliar o conhecimento sobre o tema de aplicação do modelo de orçamento flexível e análise de variação nas empresas, dada a escassez de trabalhos sobre o tema no Brasil. / In recent years, Telecom companies in Brazil have been protagonists in providing the infrastructure investments needed to meet the demands of the new digital era and continuing to generate value for shareholders. The greater concern of the executives in fulfilling the budget agreed with the headquarter, deprives the debate of how the budget can provide useful information for the company to generate competitive advantage. This Applied Work consisted in the presentation of a work of intervention to carry out a business consulting in Telefónica | Vivo, in order to evaluate whether and how the application of the flexible budget methodology and budget´s variance analysis in the B2B Unit´s mobile handsets commercialization can contribute as a management tool to leveraging the company's competitiveness. The study was developed based on company´s real historical data, from which a comparison was made between the information generated historically by the application of the static budget methodology (currently used in the company) and those generated by the flexible budget application, object of this work. To validate this comparison, structured interviews using a standard questionnaire were conducted with executives who have direct interaction with the topic, in order to evaluate the key-users’ perception on the usefulness of the information generated. The result of the study revealed that the variance analysis using the static budget method does not provide detail on what were the causal variables of the deviations. The result of the analysis using the flexible budget methodology allowed separating the variations by causal variable, previously not even seen in an applied way, and the structured and transparent way was crucial for the positive and unanimous perception of the executives. Finally, it is expected that this applied work may contribute to increase the knowledge about the application of the flexible budget model and variance analysis in companies’ budgets, given the lack of academic work on the subject in Brazil.

Text in Afrikaans / Vrouemisdaad is 'n komplekse sosiale verskynsel. 'n Aspek van vrouemisdaad wat in die onderhawige proef skrif ondersoek word, is om aan die hand van bestaande teoriee te probeer verklaar waarom vroue soveel minder misdaad as mans pleeg. Ten einde antwoorde op die vraag te vind, is 'n kwantitatiewe ondersoek na die gemeenskap se houding rakende die vrouemisdadiger en 'n kwalitatiewe ontleding van die beskikbare literatuur gedoen. 'n Intensiewe verkennende studie van die beskikbare literatuur het getoon dat die gemeenskap, navorsers en akademici bepaalde persepsies huldig oor die vrouemisdadiger wat nie altyd met die werklike beeld strook nie. Alhoewel die getalle van vrouemisdadigers deur die jare toegeneem het, het die tipe misdade wat vroue pleeg weinig verander en die omvang van vrouemisdaad proporsioneel tot die bevolking dieselfde gebly. Geen enkele teorie of benadering op sigself bied 'n algemene verklaring vir die verskynsel van vrouemisdaad nie; dit kan hoogstens as gedeeltelike verklaring dien. 'n Algemene verklaringsmodel vir vrouemisdaad is egter opgestel wat terselfdertyd as samevatting van die geselekteerde teoriee dien. Met behulp van die argivale metode is 'n profiel van die gekommitteerde vrouemisdadiger saamgestel om 'n basiese kennissisteem van die vrouemisdadiger daar te stel. Statisties beduidende verbande is gevind tussen die karakteristieke van die blanke gekommitteerde vrouemisdadiger se huwelikstaat, aantal kinders en bedrog. Die kwantitatiewe ondersoek bestaan uit die ontleding van 516 respondente se persepsies oor die vrouemisdadiger. Uit die resultate van verskeie statistiese tegnieke blyk dit dat die ondersoekgroep, wat blanke respondente van vier voorstede uit struktuurstreekplansel 22 van Pretoria verteenwoordig, bepaalde persepsies en opvattings aangaande die vrouemisdadiger het. Die persepsiemeting het aan die lig gebring dat daar verskille in persepsies bestaan oor die vrouemisdadiger en vrouemisdaad in die blanke gemeenskap bestaan. Hierdie persepsies stem nie ooreen met die beeld wat blyk uit die onderhawige navorsing nie. Statisties beduidende verskille in persepsies tussen manlike en vroulike respondente van die ondersoekgroep is ook gevind. Daar bestaan verder statisties beduidende verskille tussen respondente van verskillende voorstede, onderwyspeile, taalgroepe, en ouderdomsgroepe. / It can be said that female crime is a complex phenomenon. An aspect of female crime that is investigated in this thesis is to explain, with reference to existing theories, why women commit fewer crimes than men. In an attempt to answer this question, a quantitative investigation regarding the attitude of society towards the female criminal as well as a qualitative study of the available literature, was conducted. An intensive exploratory study on female crime indicated that and academics have a certain of the available the community, perception of literature researchers the female criminal which does not always correspond with the facts. Although the number of female criminals has increased through the years the extent of female crime, proportionally to the population size, has remained constant, and the type of crime committed by women has remained relatively unchanged. It seems clear that no single theory or approach can explain female crime, it can at best give a partial explanation. An integrated explanation model for female crime has been compiled which simultaneously serves as a summary of selected theories. By means of the archival research method a profile of the female prisoner was compiled, to be used as a basis for the researcher's scientific knowledge of this phenomenon. Statistically significant relations were found to exist between the characteristics of the white female prisoner's marital status, number of children and fraud. The quantitative investigation consisted of an analysis of the responses of 516 respondents on an attitude scale. The results of various statistical techniques show that the research group, which represents respondents fr6m four suburbs from structure plan cell 22 of Pretoria, reveals certain attitudes and beliefs about the female criminal. This attitude measurement indicated that differences in perceptions regarding the female in the white community do criminal and female exist. Furthermore crime these perceptions do not correlate on the female criminal in with the information gathered this research. Statistically significant differences in attitude were found between male and female respondents, respondents from the different suburbs, with different qualifications, of different language groups, and of different age groups. / Sociology / D. Lit. et Phil. (Kriminologie)

Text in Afrikaans / Vrouemisdaad is 'n komplekse sosiale verskynsel. 'n Aspek van vrouemisdaad wat in die onderhawige proef skrif ondersoek word, is om aan die hand van bestaande teoriee te probeer verklaar waarom vroue soveel minder misdaad as mans pleeg. Ten einde antwoorde op die vraag te vind, is 'n kwantitatiewe ondersoek na die gemeenskap se houding rakende die vrouemisdadiger en 'n kwalitatiewe ontleding van die beskikbare literatuur gedoen. 'n Intensiewe verkennende studie van die beskikbare literatuur het getoon dat die gemeenskap, navorsers en akademici bepaalde persepsies huldig oor die vrouemisdadiger wat nie altyd met die werklike beeld strook nie. Alhoewel die getalle van vrouemisdadigers deur die jare toegeneem het, het die tipe misdade wat vroue pleeg weinig verander en die omvang van vrouemisdaad proporsioneel tot die bevolking dieselfde gebly. Geen enkele teorie of benadering op sigself bied 'n algemene verklaring vir die verskynsel van vrouemisdaad nie; dit kan hoogstens as gedeeltelike verklaring dien. 'n Algemene verklaringsmodel vir vrouemisdaad is egter opgestel wat terselfdertyd as samevatting van die geselekteerde teoriee dien. Met behulp van die argivale metode is 'n profiel van die gekommitteerde vrouemisdadiger saamgestel om 'n basiese kennissisteem van die vrouemisdadiger daar te stel. Statisties beduidende verbande is gevind tussen die karakteristieke van die blanke gekommitteerde vrouemisdadiger se huwelikstaat, aantal kinders en bedrog. Die kwantitatiewe ondersoek bestaan uit die ontleding van 516 respondente se persepsies oor die vrouemisdadiger. Uit die resultate van verskeie statistiese tegnieke blyk dit dat die ondersoekgroep, wat blanke respondente van vier voorstede uit struktuurstreekplansel 22 van Pretoria verteenwoordig, bepaalde persepsies en opvattings aangaande die vrouemisdadiger het. Die persepsiemeting het aan die lig gebring dat daar verskille in persepsies bestaan oor die vrouemisdadiger en vrouemisdaad in die blanke gemeenskap bestaan. Hierdie persepsies stem nie ooreen met die beeld wat blyk uit die onderhawige navorsing nie. Statisties beduidende verskille in persepsies tussen manlike en vroulike respondente van die ondersoekgroep is ook gevind. Daar bestaan verder statisties beduidende verskille tussen respondente van verskillende voorstede, onderwyspeile, taalgroepe, en ouderdomsgroepe. / It can be said that female crime is a complex phenomenon. An aspect of female crime that is investigated in this thesis is to explain, with reference to existing theories, why women commit fewer crimes than men. In an attempt to answer this question, a quantitative investigation regarding the attitude of society towards the female criminal as well as a qualitative study of the available literature, was conducted. An intensive exploratory study on female crime indicated that and academics have a certain of the available the community, perception of literature researchers the female criminal which does not always correspond with the facts. Although the number of female criminals has increased through the years the extent of female crime, proportionally to the population size, has remained constant, and the type of crime committed by women has remained relatively unchanged. It seems clear that no single theory or approach can explain female crime, it can at best give a partial explanation. An integrated explanation model for female crime has been compiled which simultaneously serves as a summary of selected theories. By means of the archival research method a profile of the female prisoner was compiled, to be used as a basis for the researcher's scientific knowledge of this phenomenon. Statistically significant relations were found to exist between the characteristics of the white female prisoner's marital status, number of children and fraud. The quantitative investigation consisted of an analysis of the responses of 516 respondents on an attitude scale. The results of various statistical techniques show that the research group, which represents respondents fr6m four suburbs from structure plan cell 22 of Pretoria, reveals certain attitudes and beliefs about the female criminal. This attitude measurement indicated that differences in perceptions regarding the female in the white community do criminal and female exist. Furthermore crime these perceptions do not correlate on the female criminal in with the information gathered this research. Statistically significant differences in attitude were found between male and female respondents, respondents from the different suburbs, with different qualifications, of different language groups, and of different age groups. / Sociology / D. Lit. et Phil. (Kriminologie)

The diploma thesis is focused on cost control management in a manufacturing company. The diploma thesis deals with the calculation of product. It describes existing system of calculations and on the basis of theoretical knowledge suggest effective solutions. This solution should help to find hidden costs.

The thesis is focused on managerial accounting, which serves as a tool for cost control in the construction business. The first part of my work includes introduction to basic concepts, which may be encountered in the use of management accounting. The second part focuses on the practical application of management accounting. With regard to the documents which I received, this section focuses on deviation control.

Cette thèse de doctorat consiste en trois chapitres qui traitent des sujets de choix de portefeuilles de grande taille, et de mesure de risque. Le premier chapitre traite du problème d’erreur d’estimation dans les portefeuilles de grande taille, et utilise le cadre d'analyse moyenne-variance. Le second chapitre explore l'importance du risque de devise pour les portefeuilles d'actifs domestiques, et étudie les liens entre la stabilité des poids de portefeuille de grande taille et le risque de devise. Pour finir, sous l'hypothèse que le preneur de décision est pessimiste, le troisième chapitre dérive la prime de risque, une mesure du pessimisme, et propose une méthodologie pour estimer les mesures dérivées. Le premier chapitre améliore le choix optimal de portefeuille dans le cadre du principe moyenne-variance de Markowitz (1952). Ceci est motivé par les résultats très décevants obtenus, lorsque la moyenne et la variance sont remplacées par leurs estimations empiriques. Ce problème est amplifié lorsque le nombre d’actifs est grand et que la matrice de covariance empirique est singulière ou presque singulière. Dans ce chapitre, nous examinons quatre techniques de régularisation pour stabiliser l’inverse de la matrice de covariance: le ridge, spectral cut-off, Landweber-Fridman et LARS Lasso. Ces méthodes font chacune intervenir un paramètre d’ajustement, qui doit être sélectionné. La contribution principale de cette partie, est de dériver une méthode basée uniquement sur les données pour sélectionner le paramètre de régularisation de manière optimale, i.e. pour minimiser la perte espérée d’utilité. Précisément, un critère de validation croisée qui prend une même forme pour les quatre méthodes de régularisation est dérivé. Les règles régularisées obtenues sont alors comparées à la règle utilisant directement les données et à la stratégie naïve 1/N, selon leur perte espérée d’utilité et leur ratio de Sharpe. Ces performances sont mesurée dans l’échantillon (in-sample) et hors-échantillon (out-of-sample) en considérant différentes tailles d’échantillon et nombre d’actifs. Des simulations et de l’illustration empirique menées, il ressort principalement que la régularisation de la matrice de covariance améliore de manière significative la règle de Markowitz basée sur les données, et donne de meilleurs résultats que le portefeuille naïf, surtout dans les cas le problème d’erreur d’estimation est très sévère. Dans le second chapitre, nous investiguons dans quelle mesure, les portefeuilles optimaux et stables d'actifs domestiques, peuvent réduire ou éliminer le risque de devise. Pour cela nous utilisons des rendements mensuelles de 48 industries américaines, au cours de la période 1976-2008. Pour résoudre les problèmes d'instabilité inhérents aux portefeuilles de grandes tailles, nous adoptons la méthode de régularisation spectral cut-off. Ceci aboutit à une famille de portefeuilles optimaux et stables, en permettant aux investisseurs de choisir différents pourcentages des composantes principales (ou dégrées de stabilité). Nos tests empiriques sont basés sur un modèle International d'évaluation d'actifs financiers (IAPM). Dans ce modèle, le risque de devise est décomposé en deux facteurs représentant les devises des pays industrialisés d'une part, et celles des pays émergents d'autres part. Nos résultats indiquent que le risque de devise est primé et varie à travers le temps pour les portefeuilles stables de risque minimum. De plus ces stratégies conduisent à une réduction significative de l'exposition au risque de change, tandis que la contribution de la prime risque de change reste en moyenne inchangée. Les poids de portefeuille optimaux sont une alternative aux poids de capitalisation boursière. Par conséquent ce chapitre complète la littérature selon laquelle la prime de risque est importante au niveau de l'industrie et au niveau national dans la plupart des pays. Dans le dernier chapitre, nous dérivons une mesure de la prime de risque pour des préférences dépendent du rang et proposons une mesure du degré de pessimisme, étant donné une fonction de distorsion. Les mesures introduites généralisent la mesure de prime de risque dérivée dans le cadre de la théorie de l'utilité espérée, qui est fréquemment violée aussi bien dans des situations expérimentales que dans des situations réelles. Dans la grande famille des préférences considérées, une attention particulière est accordée à la CVaR (valeur à risque conditionnelle). Cette dernière mesure de risque est de plus en plus utilisée pour la construction de portefeuilles et est préconisée pour compléter la VaR (valeur à risque) utilisée depuis 1996 par le comité de Bâle. De plus, nous fournissons le cadre statistique nécessaire pour faire de l’inférence sur les mesures proposées. Pour finir, les propriétés des estimateurs proposés sont évaluées à travers une étude Monte-Carlo, et une illustration empirique en utilisant les rendements journaliers du marché boursier américain sur de la période 2000-2011. / This thesis consists of three chapters on the topics of portfolio choice in a high-dimensional context, and risk measurement. The first chapter addresses the estimation error issue that arises when constructing large portfolios in the mean-variance framework. The second chapter investigates the relevance of currency risk for optimal domestic portfolios, evaluates their ability of to diversify away currency risk, and study the links between portfolio weights stability and currency risk. Finally, under the assumption that decision makers are pessimistic, the third chapter derives the risk premium, propose a measure of the degree of pessimism, and provide a statistical framework for their estimation. The first chapter improves the performance of the optimal portfolio weig-hts obtained under the mean-variance framework of Markowitz (1952). Indeed, these weights give unsatisfactory results, when the mean and variance are replaced by their sample counterparts (plug-in rules). This problem is amplified when the number of assets is large and the sample covariance is singular or nearly singular. The chapter investigates four regularization techniques to stabilizing the inverse of the covariance matrix: the ridge, spectral cut-off, Landweber-Fridman, and LARS Lasso. These four methods involve a tuning parameter that needs to be selected. The main contribution is to derive a data-based method for selecting the tuning parameter in an optimal way, i.e. in order to minimize the expected loss in utility of a mean-variance investor. The cross-validation type criterion derived is found to take a similar form for the four regularization methods. The resulting regularized rules are compared to the sample-based mean-variance portfolio and the naive 1/N strategy in terms of in-sample and out-of-sample Sharpe ratio and expected loss in utility. The main finding is that regularization to covariance matrix significantly improves the performance of the mean-variance problem and outperforms the naive portfolio, especially in ill-posed cases, as suggested by our simulations and empirical studies. In the second chapter, we investigate the extent to which optimal and stable portfolios of domestic assets can reduce or eliminate currency risk. This is done using monthly returns on 48 U.S. industries, from 1976 to 2008. To tackle the instabilities inherent to large portfolios, we use the spectral cut-off regularization described in Chapter 1. This gives rise to a family of stable global minimum portfolios that allows investors to select different percentages of principal components for portfolio construction. Our empirical tests are based on a conditional International Asset Pricing Model (IAPM), augmented with the size and book-to-market factors of Fama and French (1993). Using two trade-weighted currency indices of industrialized countries currencies and emerging markets currencies, we find that currency risk is priced and time-varying for global minimum portfolios. These strategies also lead to a significant reduction in the exposure to currency risk, while keeping the average premium contribution to total premium approximately the same. The global minimum weights considered are an alternative to market capitalization weights used in the U.S. market index. Therefore, our findings complement the well established results that currency risk is significantly priced and economically meaningful at the industry and country level in most countries. Finally, the third chapter derives a measure of the risk premium for rank-dependent preferences and proposes a measure of the degree of pessimism, given a distortion function. The introduced measures generalize the common risk measures derived in the expected utility theory framework, which is frequently violated in both experimental and real-life situations. These measures are derived in the neighborhood of a given random loss variable, using the notion of local utility function. A particular interest is devoted to the CVaR, which is now widely used for asset allocation and has been advocated to complement the Value-at-risk (VaR) proposed since 1996 by the Basel Committee on Banking Supervision. We provide the statistical framework needed to conduct inference on the derived measures. Finally, the proposed estimators

Cette thèse de doctorat consiste en trois chapitres qui traitent des sujets de choix de portefeuilles de grande taille, et de mesure de risque. Le premier chapitre traite du problème d’erreur d’estimation dans les portefeuilles de grande taille, et utilise le cadre d'analyse moyenne-variance. Le second chapitre explore l'importance du risque de devise pour les portefeuilles d'actifs domestiques, et étudie les liens entre la stabilité des poids de portefeuille de grande taille et le risque de devise. Pour finir, sous l'hypothèse que le preneur de décision est pessimiste, le troisième chapitre dérive la prime de risque, une mesure du pessimisme, et propose une méthodologie pour estimer les mesures dérivées. Le premier chapitre améliore le choix optimal de portefeuille dans le cadre du principe moyenne-variance de Markowitz (1952). Ceci est motivé par les résultats très décevants obtenus, lorsque la moyenne et la variance sont remplacées par leurs estimations empiriques. Ce problème est amplifié lorsque le nombre d’actifs est grand et que la matrice de covariance empirique est singulière ou presque singulière. Dans ce chapitre, nous examinons quatre techniques de régularisation pour stabiliser l’inverse de la matrice de covariance: le ridge, spectral cut-off, Landweber-Fridman et LARS Lasso. Ces méthodes font chacune intervenir un paramètre d’ajustement, qui doit être sélectionné. La contribution principale de cette partie, est de dériver une méthode basée uniquement sur les données pour sélectionner le paramètre de régularisation de manière optimale, i.e. pour minimiser la perte espérée d’utilité. Précisément, un critère de validation croisée qui prend une même forme pour les quatre méthodes de régularisation est dérivé. Les règles régularisées obtenues sont alors comparées à la règle utilisant directement les données et à la stratégie naïve 1/N, selon leur perte espérée d’utilité et leur ratio de Sharpe. Ces performances sont mesurée dans l’échantillon (in-sample) et hors-échantillon (out-of-sample) en considérant différentes tailles d’échantillon et nombre d’actifs. Des simulations et de l’illustration empirique menées, il ressort principalement que la régularisation de la matrice de covariance améliore de manière significative la règle de Markowitz basée sur les données, et donne de meilleurs résultats que le portefeuille naïf, surtout dans les cas le problème d’erreur d’estimation est très sévère. Dans le second chapitre, nous investiguons dans quelle mesure, les portefeuilles optimaux et stables d'actifs domestiques, peuvent réduire ou éliminer le risque de devise. Pour cela nous utilisons des rendements mensuelles de 48 industries américaines, au cours de la période 1976-2008. Pour résoudre les problèmes d'instabilité inhérents aux portefeuilles de grandes tailles, nous adoptons la méthode de régularisation spectral cut-off. Ceci aboutit à une famille de portefeuilles optimaux et stables, en permettant aux investisseurs de choisir différents pourcentages des composantes principales (ou dégrées de stabilité). Nos tests empiriques sont basés sur un modèle International d'évaluation d'actifs financiers (IAPM). Dans ce modèle, le risque de devise est décomposé en deux facteurs représentant les devises des pays industrialisés d'une part, et celles des pays émergents d'autres part. Nos résultats indiquent que le risque de devise est primé et varie à travers le temps pour les portefeuilles stables de risque minimum. De plus ces stratégies conduisent à une réduction significative de l'exposition au risque de change, tandis que la contribution de la prime risque de change reste en moyenne inchangée. Les poids de portefeuille optimaux sont une alternative aux poids de capitalisation boursière. Par conséquent ce chapitre complète la littérature selon laquelle la prime de risque est importante au niveau de l'industrie et au niveau national dans la plupart des pays. Dans le dernier chapitre, nous dérivons une mesure de la prime de risque pour des préférences dépendent du rang et proposons une mesure du degré de pessimisme, étant donné une fonction de distorsion. Les mesures introduites généralisent la mesure de prime de risque dérivée dans le cadre de la théorie de l'utilité espérée, qui est fréquemment violée aussi bien dans des situations expérimentales que dans des situations réelles. Dans la grande famille des préférences considérées, une attention particulière est accordée à la CVaR (valeur à risque conditionnelle). Cette dernière mesure de risque est de plus en plus utilisée pour la construction de portefeuilles et est préconisée pour compléter la VaR (valeur à risque) utilisée depuis 1996 par le comité de Bâle. De plus, nous fournissons le cadre statistique nécessaire pour faire de l’inférence sur les mesures proposées. Pour finir, les propriétés des estimateurs proposés sont évaluées à travers une étude Monte-Carlo, et une illustration empirique en utilisant les rendements journaliers du marché boursier américain sur de la période 2000-2011. / This thesis consists of three chapters on the topics of portfolio choice in a high-dimensional context, and risk measurement. The first chapter addresses the estimation error issue that arises when constructing large portfolios in the mean-variance framework. The second chapter investigates the relevance of currency risk for optimal domestic portfolios, evaluates their ability of to diversify away currency risk, and study the links between portfolio weights stability and currency risk. Finally, under the assumption that decision makers are pessimistic, the third chapter derives the risk premium, propose a measure of the degree of pessimism, and provide a statistical framework for their estimation. The first chapter improves the performance of the optimal portfolio weig-hts obtained under the mean-variance framework of Markowitz (1952). Indeed, these weights give unsatisfactory results, when the mean and variance are replaced by their sample counterparts (plug-in rules). This problem is amplified when the number of assets is large and the sample covariance is singular or nearly singular. The chapter investigates four regularization techniques to stabilizing the inverse of the covariance matrix: the ridge, spectral cut-off, Landweber-Fridman, and LARS Lasso. These four methods involve a tuning parameter that needs to be selected. The main contribution is to derive a data-based method for selecting the tuning parameter in an optimal way, i.e. in order to minimize the expected loss in utility of a mean-variance investor. The cross-validation type criterion derived is found to take a similar form for the four regularization methods. The resulting regularized rules are compared to the sample-based mean-variance portfolio and the naive 1/N strategy in terms of in-sample and out-of-sample Sharpe ratio and expected loss in utility. The main finding is that regularization to covariance matrix significantly improves the performance of the mean-variance problem and outperforms the naive portfolio, especially in ill-posed cases, as suggested by our simulations and empirical studies. In the second chapter, we investigate the extent to which optimal and stable portfolios of domestic assets can reduce or eliminate currency risk. This is done using monthly returns on 48 U.S. industries, from 1976 to 2008. To tackle the instabilities inherent to large portfolios, we use the spectral cut-off regularization described in Chapter 1. This gives rise to a family of stable global minimum portfolios that allows investors to select different percentages of principal components for portfolio construction. Our empirical tests are based on a conditional International Asset Pricing Model (IAPM), augmented with the size and book-to-market factors of Fama and French (1993). Using two trade-weighted currency indices of industrialized countries currencies and emerging markets currencies, we find that currency risk is priced and time-varying for global minimum portfolios. These strategies also lead to a significant reduction in the exposure to currency risk, while keeping the average premium contribution to total premium approximately the same. The global minimum weights considered are an alternative to market capitalization weights used in the U.S. market index. Therefore, our findings complement the well established results that currency risk is significantly priced and economically meaningful at the industry and country level in most countries. Finally, the third chapter derives a measure of the risk premium for rank-dependent preferences and proposes a measure of the degree of pessimism, given a distortion function. The introduced measures generalize the common risk measures derived in the expected utility theory framework, which is frequently violated in both experimental and real-life situations. These measures are derived in the neighborhood of a given random loss variable, using the notion of local utility function. A particular interest is devoted to the CVaR, which is now widely used for asset allocation and has been advocated to complement the Value-at-risk (VaR) proposed since 1996 by the Basel Committee on Banking Supervision. We provide the statistical framework needed to conduct inference on the derived measures. Finally, the proposed estimators

Portfolio selection has a long tradition in financial economics and plays an integral role in investment management. Portfolio selection provides the framework to determine optimal portfolio choice from a universe of available investments. However, the asset weightings from portfolio selection are optimal only if the empirical characteristics of asset returns do not violate the portfolio selection model assumptions. This thesis explores the empirical characteristics of traditional assets and hedge fund returns and examines their effects on the assumptions of linearity-in-the-mean testing and portfolio selection. The encompassing theme of this thesis is the empirical interplay between traditional assets and hedge fund returns. Despite the paucity of hedge fund research, pension funds continue to increase their portfolio allocations to global hedge funds in an effort to pursue higher risk-adjusted returns. This thesis presents three empirical studies which provide positive insights into the relationships between traditional assets and hedge fund returns. The first two empirical studies examine an emerging body of literature which suggests that the relationship between traditional assets and hedge fund returns is non-linear. For mean-variance investors, non-linear asset returns are problematic as they do not satisfy the assumption of linearity required for the covariance matrix in portfolio selection. To examine the linearity assumption as it relates to a mean-variance investor, a hypothesis test approach is employed which investigates the linearity-in-the-mean of traditional assets and hedge funds. The findings from the first two empirical studies reveal that conventional linearity-in-the-mean tests incorrectly conclude that asset returns are nonlinear. We demonstrate that the empirical characteristics of heteroscedasticity and autocorrelation in asset returns are the primary sources of test mis-specification in these linearity-in-the-mean hypothesis tests. To address this problem, an innovative approach is proposed to control heteroscedasticity and autocorrelation in the underlying tests and it is shown that traditional assets and hedge funds are indeed linear-in-the-mean. The third and final study of this thesis explores traditional assets and hedge funds in a portfolio selection framework. Following the theme of the previous two studies, the effects of heteroscedasticity and autocorrelation are examined in the portfolio selection context. The characteristics of serial correlation in bond and hedge fund returns are shown to cause a downward bias in the second sample moment. This thesis proposes two methods to control for this effect and it is shown that autocorrelation induces an overallocation to bonds and hedge funds. Whilst heteroscedasticity cannot be directly examined in portfolio selection, empirical evidence suggests that heteroscedastic events (such as those that occurred in August 1998) translate into the empirical feature known as tail-risk. The effects of tail-risk are examined by comparing the portfolio decisions of mean-variance analysis (MVA) versus mean-conditional value at risk (M-CVaR) investors. The findings reveal that the volatility of returns in a MVA portfolio decreases when hedge funds are included in the investment opportunity set. However, the reduction in the volatility of portfolio returns comes at a cost of undesirable third and fourth moments. Furthermore, it is shown that investors with M-CVaR preferences exhibit a decreasing demand for hedge funds as their aversion for tail-risk increases. The results of the thesis highlight the sensitivities of linearity tests and portfolio selection to the empirical features of heteroscedasticity, autocorrelation and tail-risk. This thesis contributes to the literature by providing refinements to these frameworks which allow improved inferences to be made when hedge funds are examined in linearity and portfolio selection settings.

Theory: - Controlling - its function and structure - Costs Analytical part: - Cost variance tracking and evaluation of its monthly prediction - Proposal of best practices for cost variance prediction done on monthly basis