Méthodes de Bootstrap pour les modèles à facteurs

Cette thèse développe des méthodes bootstrap pour les modèles à facteurs qui sont couram- ment utilisés pour générer des prévisions depuis l'article pionnier de Stock et Watson (2002) sur les indices de diffusion. Ces modèles tolèrent l'inclusion d'un grand nombre de variables macroéconomiques et financières comme prédicteurs, une caractéristique utile pour inclure di- verses informations disponibles aux agents économiques. Ma thèse propose donc des outils éco- nométriques qui améliorent l'inférence dans les modèles à facteurs utilisant des facteurs latents extraits d'un large panel de prédicteurs observés. Il est subdivisé en trois chapitres complémen- taires dont les deux premiers en collaboration avec Sílvia Gonçalves et Benoit Perron. Dans le premier article, nous étudions comment les méthodes bootstrap peuvent être utilisées pour faire de l'inférence dans les modèles de prévision pour un horizon de h périodes dans le futur. Pour ce faire, il examine l'inférence bootstrap dans un contexte de régression augmentée de facteurs où les erreurs pourraient être autocorrélées. Il généralise les résultats de Gonçalves et Perron (2014) et propose puis justifie deux approches basées sur les résidus : le block wild bootstrap et le dependent wild bootstrap. Nos simulations montrent une amélioration des taux de couverture des intervalles de confiance des coefficients estimés en utilisant ces approches comparativement à la théorie asymptotique et au wild bootstrap en présence de corrélation sérielle dans les erreurs de régression. Le deuxième chapitre propose des méthodes bootstrap pour la construction des intervalles de prévision permettant de relâcher l'hypothèse de normalité des innovations. Nous y propo- sons des intervalles de prédiction bootstrap pour une observation h périodes dans le futur et sa moyenne conditionnelle. Nous supposons que ces prévisions sont faites en utilisant un ensemble de facteurs extraits d'un large panel de variables. Parce que nous traitons ces facteurs comme latents, nos prévisions dépendent à la fois des facteurs estimés et les coefficients de régres- sion estimés. Sous des conditions de régularité, Bai et Ng (2006) ont proposé la construction d'intervalles asymptotiques sous l'hypothèse de Gaussianité des innovations. Le bootstrap nous permet de relâcher cette hypothèse et de construire des intervalles de prédiction valides sous des hypothèses plus générales. En outre, même en supposant la Gaussianité, le bootstrap conduit à des intervalles plus précis dans les cas où la dimension transversale est relativement faible car il prend en considération le biais de l'estimateur des moindres carrés ordinaires comme le montre une étude récente de Gonçalves et Perron (2014). Dans le troisième chapitre, nous suggérons des procédures de sélection convergentes pour les regressions augmentées de facteurs en échantillons finis. Nous démontrons premièrement que la méthode de validation croisée usuelle est non-convergente mais que sa généralisation, la validation croisée «leave-d-out» sélectionne le plus petit ensemble de facteurs estimés pour l'espace généré par les vraies facteurs. Le deuxième critère dont nous montrons également la validité généralise l'approximation bootstrap de Shao (1996) pour les regressions augmentées de facteurs. Les simulations montrent une amélioration de la probabilité de sélectionner par- cimonieusement les facteurs estimés comparativement aux méthodes de sélection disponibles. L'application empirique revisite la relation entre les facteurs macroéconomiques et financiers, et l'excès de rendement sur le marché boursier américain. Parmi les facteurs estimés à partir d'un large panel de données macroéconomiques et financières des États Unis, les facteurs fortement correlés aux écarts de taux d'intérêt et les facteurs de Fama-French ont un bon pouvoir prédictif pour les excès de rendement.

A Monthly Volatility Index for the US Economy

We estimate the monthly volatility of the US economy from 1968 to 2006 by extending the coincidentindex model of Stock and Watson (1991). Our volatility index, which we call VOLINX, hasfour applications. First, it sheds light on the Great Moderation. VOLINX captures the decrease in thevolatility in the mid-80s as well as the different episodes of stress over the sample period. In the 70sand early 80s the stagflation and the two oil crises marked the pace of the volatility whereas 09/11 is themost relevant shock after the moderation. Second, it helps to understand the economic indicators thatcause volatility. While the main determinant of the coincident index is industrial production, VOLINXis mainly affected by employment and income. Third, it adapts the confidence bands of the forecasts.In and out-of-sample evaluations show that the confidence bands may differ up to 50% with respect to amodel with constant variance. Last, the methodology we use permits us to estimate monthly GDP, whichhas conditional volatility that is partly explained by VOLINX. These applications can be used by policymakers for monitoring and surveillance of the stress of the economy.

Pricing derivatives with stochastic volatility

This Ph.D. thesis contains 4 essays in mathematical finance with a focus on pricing Asian option (Chapter 4), pricing futures and futures option (Chapter 5 and Chapter 6) and time dependent volatility in futures option (Chapter 7). In Chapter 4, the applicability of the Albrecher et al.(2005)'s comonotonicity approach was investigated in the context of various benchmark models for equities and com- modities. Instead of classical Levy models as in Albrecher et al.(2005), the focus is the Heston stochastic volatility model, the constant elasticity of variance (CEV) model and the Schwartz (1997) two-factor model. It is shown that the method delivers rather tight upper bounds for the prices of Asian Options in these models and as a by-product delivers super-hedging strategies which can be easily implemented. In Chapter 5, two types of three-factor models were studied to give the value of com- modities futures contracts, which allow volatility to be stochastic. Both these two models have closed-form solutions for futures contracts price. However, it is shown that Model 2 is better than Model 1 theoretically and also performs very well empiri- cally. Moreover, Model 2 can easily be implemented in practice. In comparison to the Schwartz (1997) two-factor model, it is shown that Model 2 has its unique advantages; hence, it is also a good choice to price the value of commodity futures contracts. Fur- thermore, if these two models are used at the same time, a more accurate price for commodity futures contracts can be obtained in most situations. In Chapter 6, the applicability of the asymptotic approach developed in Fouque et al.(2000b) was investigated for pricing commodity futures options in a Schwartz (1997) multi-factor model, featuring both stochastic convenience yield and stochastic volatility. It is shown that the zero-order term in the expansion coincides with the Schwartz (1997) two-factor term, with averaged volatility, and an explicit expression for the first-order correction term is provided. With empirical data from the natural gas futures market, it is also demonstrated that a significantly better calibration can be achieved by using the correction term as compared to the standard Schwartz (1997) two-factor expression, at virtually no extra effort. In Chapter 7, a new pricing formula is derived for futures options in the Schwartz (1997) two-factor model with time dependent spot volatility. The pricing formula can also be used to find the result of the time dependent spot volatility with futures options prices in the market. Furthermore, the limitations of the method that is used to find the time dependent spot volatility will be explained, and it is also shown how to make sure of its accuracy.

Méthodes de Bootstrap pour les modèles à facteurs

Cette thèse développe des méthodes bootstrap pour les modèles à facteurs qui sont couram- ment utilisés pour générer des prévisions depuis l'article pionnier de Stock et Watson (2002) sur les indices de diffusion. Ces modèles tolèrent l'inclusion d'un grand nombre de variables macroéconomiques et financières comme prédicteurs, une caractéristique utile pour inclure di- verses informations disponibles aux agents économiques. Ma thèse propose donc des outils éco- nométriques qui améliorent l'inférence dans les modèles à facteurs utilisant des facteurs latents extraits d'un large panel de prédicteurs observés. Il est subdivisé en trois chapitres complémen- taires dont les deux premiers en collaboration avec Sílvia Gonçalves et Benoit Perron. Dans le premier article, nous étudions comment les méthodes bootstrap peuvent être utilisées pour faire de l'inférence dans les modèles de prévision pour un horizon de h périodes dans le futur. Pour ce faire, il examine l'inférence bootstrap dans un contexte de régression augmentée de facteurs où les erreurs pourraient être autocorrélées. Il généralise les résultats de Gonçalves et Perron (2014) et propose puis justifie deux approches basées sur les résidus : le block wild bootstrap et le dependent wild bootstrap. Nos simulations montrent une amélioration des taux de couverture des intervalles de confiance des coefficients estimés en utilisant ces approches comparativement à la théorie asymptotique et au wild bootstrap en présence de corrélation sérielle dans les erreurs de régression. Le deuxième chapitre propose des méthodes bootstrap pour la construction des intervalles de prévision permettant de relâcher l'hypothèse de normalité des innovations. Nous y propo- sons des intervalles de prédiction bootstrap pour une observation h périodes dans le futur et sa moyenne conditionnelle. Nous supposons que ces prévisions sont faites en utilisant un ensemble de facteurs extraits d'un large panel de variables. Parce que nous traitons ces facteurs comme latents, nos prévisions dépendent à la fois des facteurs estimés et les coefficients de régres- sion estimés. Sous des conditions de régularité, Bai et Ng (2006) ont proposé la construction d'intervalles asymptotiques sous l'hypothèse de Gaussianité des innovations. Le bootstrap nous permet de relâcher cette hypothèse et de construire des intervalles de prédiction valides sous des hypothèses plus générales. En outre, même en supposant la Gaussianité, le bootstrap conduit à des intervalles plus précis dans les cas où la dimension transversale est relativement faible car il prend en considération le biais de l'estimateur des moindres carrés ordinaires comme le montre une étude récente de Gonçalves et Perron (2014). Dans le troisième chapitre, nous suggérons des procédures de sélection convergentes pour les regressions augmentées de facteurs en échantillons finis. Nous démontrons premièrement que la méthode de validation croisée usuelle est non-convergente mais que sa généralisation, la validation croisée «leave-d-out» sélectionne le plus petit ensemble de facteurs estimés pour l'espace généré par les vraies facteurs. Le deuxième critère dont nous montrons également la validité généralise l'approximation bootstrap de Shao (1996) pour les regressions augmentées de facteurs. Les simulations montrent une amélioration de la probabilité de sélectionner par- cimonieusement les facteurs estimés comparativement aux méthodes de sélection disponibles. L'application empirique revisite la relation entre les facteurs macroéconomiques et financiers, et l'excès de rendement sur le marché boursier américain. Parmi les facteurs estimés à partir d'un large panel de données macroéconomiques et financières des États Unis, les facteurs fortement correlés aux écarts de taux d'intérêt et les facteurs de Fama-French ont un bon pouvoir prédictif pour les excès de rendement.

School Anxiety Inventory: Reliability and Validity Evidence in a Sample of Slovenian Adolescents

The purpose of this study was to examine the reliability and validity of the School Anxiety Inventory (SAI) using a sample of 646 Slovenian adolescents (48% boys), ranging in age from 12 to 19 years. Single confirmatory factor analyses replicated the correlated four-factor structure of scores on the SAI for anxiety-provoking school situations (Anxiety about School Failure and Punishment, Anxiety about Aggression, Anxiety about Social Evaluation, and Anxiety about Academic Evaluation), and the three-factor structure of the anxiety response systems (Physiological Anxiety, Cognitive Anxiety, and Behavioral Anxiety). Equality of factor structures was compared using multigroup confirmatory factor analyses. Measurement invariance for the four- and three-factor models was obtained across gender and school-level samples. The scores of the instrument showed high internal reliability and adequate test–retest reliability. The concurrent validity of the SAI scores was also examined through its relationship with the Social Anxiety Scale for Adolescents (SASA) scores and the Questionnaire about Interpersonal Difficulties for Adolescents (QIDA) scores. Correlations of the SAI scores with scores on the SASA and the QIDA were of low to moderate effect sizes.

Recent advances in explaining hedge fund returns:implicit factors and exposures

We survey articles covering how hedge fund returns are explained, using largely non-linear multifactor models that examine the non-linear pay-offs and exposures of hedge funds. We provide an integrated view of the implicit factor and statistical factor models that are largely able to explain the hedge fund return-generating process. We present their evolution through time by discussing pioneering studies that made a significant contribution to knowledge, and also recent innovative studies that examine hedge fund exposures using advanced econometric methods. This is the first review that analyzes very recent studies that explain a large part of hedge fund variation. We conclude by presenting some gaps for future research.

Authentic Leadership Questionnaire: invariance between samples of Brazilian and Portuguese employees

The Authentic Leadership Questionnaire (ALQ) is used to assess authentic leadership (AL). Although ALQ is often used in empirical research, cross-cultural studies with this measure are scarce. Aiming to contribute to filling this gap, this study assesses the invariance of the ALQ measure between samples of Brazilian (N = 1019) and Portuguese (N = 842) employees. A multi-group confirmatory factor analysis was performed, and the results showed the invariance of the first- and second-order factor models between the Brazilian and Portuguese samples. The results are discussed considering their cultural setting, with the study’s limitations and future research directions being pointed out.

Assessing alternative models of individualism and collectivism: A confirmatory factor analysis

Six alternative structural models of individualism-collectivism are reviewed and empirically compared in a confirmatory factor analysis of questionnaire data from an Australian student sample (N=340). Central to the debate about the structure of this broad social attitude are the issues of (I) polarity (are individualism and collectivism bipolar opposites, or orthogonal factors?) and (2) dimensionality (are individualism and collectivism themselves higher-order constructs subsuming several more specific factors and, if so, what are they?). The data from this Australian sample support a model that represents individualism and collectivism as a higher-order bipolar factor hierarchically subsuming several bipolar reference-group-specific individualisms and collectivisms. Copyright (C) 2001 John Wiley & Sons, Ltd.

Are one factor logarithmic volatility models useful to fit the features of financial data? An application to microsoft data.

This paper provides empirical evidence that continuous time models with one factor of volatility, in some conditions, are able to fit the main characteristics of financial data. It also reports the importance of the feedback factor in capturing the strong volatility clustering of data, caused by a possible change in the pattern of volatility in the last part of the sample. We use the Efficient Method of Moments (EMM) by Gallant and Tauchen (1996) to estimate logarithmic models with one and two stochastic volatility factors (with and without feedback) and to select among them.

Dimensionality of the Rosenberg self-esteem scale and its relationships with the three- and five-factor personality models

The authors investigated the dimensionality of the French version of the Rosenberg Self-Esteem Scale (RSES; Rosenberg, 1965) using confirmatory factor analysis. We tested models of 1 or 2 factors. Results suggest the RSES is a 1-dimensional scale with 3 highly correlated items. Comparison with the Revised NEO-Personality Inventory (NEO-PI-R; Costa, McCrae, & Rolland, 1998) demonstrated that Neuroticism correlated strongly and Extraversion and Conscientiousness moderately with the RSES. Depression accounted for 47% of the variance of the RSES. Other NEO-PI-R facets were also moderately related with self-esteem.

Severity of chronic experimental Chagas' heart disease parallels tumour necrosis factor and nitric oxide levels in the serum: models of mild and severe disease

Heart tissue inflammation, progressive fibrosis and electrocardiographic alterations occur in approximately 30% of patients infected by Trypanosoma cruzi, 10-30 years after infection. Further, plasma levels of tumour necrosis factor (TNF) and nitric oxide (NO) are associated with the degree of heart dysfunction in chronic chagasic cardiomyopathy (CCC). Thus, our aim was to establish experimental models that mimic a range of parasitological, pathological and cardiac alterations described in patients with chronic Chagas’ heart disease and evaluate whether heart disease severity was associated with increased TNF and NO levels in the serum. Our results show that C3H/He mice chronically infected with the Colombian T. cruzi strain have more severe cardiac parasitism and inflammation than C57BL/6 mice. In addition, connexin 43 disorganisation and fibronectin deposition in the heart tissue, increased levels of creatine kinase cardiac MB isoenzyme activity in the serum and more severe electrical abnormalities were observed in T. cruzi-infected C3H/He mice compared to C57BL/6 mice. Therefore, T. cruzi-infected C3H/He and C57BL/6 mice represent severe and mild models of CCC, respectively. Moreover, the CCC severity paralleled the TNF and NO levels in the serum. Therefore, these models are appropriate for studying the pathophysiology and biomarkers of CCC progression, as well as for testing therapeutic agents for patients with Chagas’ heart disease.

Examining the virulence of Candida albicans transcription factor mutants using Galleria mellonella and mouse infection models.

The aim of the present study was to identify Candida albicans transcription factors (TFs) involved in virulence. Although mice are considered the gold-standard model to study fungal virulence, mini-host infection models have been increasingly used. Here, barcoded TF mutants were first screened in mice by pools of strains and fungal burdens (FBs) quantified in kidneys. Mutants of unannotated genes which generated a kidney FB significantly different from that of wild-type were selected and individually examined in Galleria mellonella. In addition, mutants that could not be detected in mice were also tested in G. mellonella. Only 25% of these mutants displayed matching phenotypes in both hosts, highlighting a significant discrepancy between the two models. To address the basis of this difference (pool or host effects), a set of 19 mutants tested in G. mellonella were also injected individually into mice. Matching FB phenotypes were observed in 50% of the cases, highlighting the bias due to host effects. In contrast, 33.4% concordance was observed between pool and single strain infections in mice, thereby highlighting the bias introduced by the "pool effect." After filtering the results obtained from the two infection models, mutants for MBF1 and ZCF6 were selected. Independent marker-free mutants were subsequently tested in both hosts to validate previous results. The MBF1 mutant showed impaired infection in both models, while the ZCF6 mutant was only significant in mice infections. The two mutants showed no obvious in vitro phenotypes compared with the wild-type, indicating that these genes might be specifically involved in in vivo adapt.

Differential associations of personal and general just-world beliefs with the Five-Factor and HEXACO models of personality.

Recent literature evidences differential associations of personal and general just-world beliefs with constructs in the interpersonal domain. In line with this research, we examine the respective relationships of each just-world belief with the Five-Factor and the HEXACO models of personality in one representative sample of the working population of Switzerland and one sample of the general US population, respectively. One suppressor effect was observed in both samples: Neuroticism and emotionality was positively associated with general just-world belief, but only after controlling for personal just-world belief. In addition, agreeableness was positively and honesty-humility negatively associated with general just-world belief but unrelated to personal just-world belief. Conscientiousness was consistently unrelated to any of the just-world belief and extraversion and openness to experience revealed unstable coefficients across studies. We discuss these points in light of just-world theory and their implications for future research taking both dimensions into account.

Continuous sky view factor maps from high resolution urban digital elevation models

This paper presents a new method to calculate sky view factors (SVFs) from high resolution urban digital elevation models using a shadow casting algorithm. By utilizing weighted annuli to derive SVF from hemispherical images, the distance light source positions can be predefined and uniformly spread over the whole hemisphere, whereas another method applies a random set of light source positions with a cosine-weighted distribution of sun altitude angles. The 2 methods have similar results based on a large number of SVF images. However, when comparing variations at pixel level between an image generated using the new method presented in this paper with the image from the random method, anisotropic patterns occur. The absolute mean difference between the 2 methods is 0.002 ranging up to 0.040. The maximum difference can be as much as 0.122. Since SVF is a geometrically derived parameter, the anisotropic errors created by the random method must be considered as significant.