La « Grande Récession » de 2008-2009 ou encore l'aggravation de la crise des dettes souveraines et de la dette publique dans la zone euro à l'été 2011, constituent de récents événements qui ont cristallisé les enjeux de l'analyse conjoncturelle, ceux relatifs notamment à la datation et à la prévision des inflexions cycliques de l'activité réelle. L'objet de cette thèse s'inscrit fondamentalement au sein de ces deux approches complémentaires du cycle économique.Le chapitre 1 dresse un portrait du cycle autour de trois conceptions distinctes de ses points de retournement : le cycle classique, le cycle de croissance et le cycle d'accélération. Nous discutons également de sa mesure eu égard aux diverses représentations possibles de l'activité agrégée d'un pays, ainsi qu'aux deux traditions existantes dans lesquelles s'inscrivent les modèles de datation. Nous mettons par ailleurs en lumière l'influence grandissante de l'environnement financier sur la dynamique cyclique des économies. Le chapitre 2 nous amène à développer deux algorithmes non-paramétriques dans le but de repérer les inflexions propres à chacun des cycles auparavant conceptualisés, mais aussipour en mesurer leurs principales caractéristiques. Le premier (resp. le second) algorithme repose sur une représentation univariée (resp. multivariée) de l'activité économique globale ; in fine, nous les appliquons aux données de la conjoncture française entre 1970 et 2010. Le chapitre 3 tire parti de nos résultats en matière de datation conjoncturelle afin de prévoir les récessions françaises depuis 1974. Au moyen de modèles probits, nous illustrons le rôle de variables financières et monétaires en tant qu'indicateurs avancés des fluctuations du cycle des affaires français. Nous montrons de plus que nos modèles prédictifs assurent uneparfaite détection des récessions pour un horizon égal à deux trimestres.Le chapitre 4 prolonge l'ensemble de l'analyse à plusieurs États membres de la zoneeuro, ces derniers étant observés depuis 1979. Nous construisons d'abord une chronologie de leurs cycles classiques respectifs puis, nous proposons un examen de leurs caractéristiques moyennes et de leur degré de synchronisation. Enfin, en s'appuyant sur des indicateurs financiers et monétaires dans le cadre d'un probit dynamique à effets fixes, nous parvenons à anticiper - jusqu'à un horizon de deux trimestres - les épisodes récessifs survenus dans les économies considérées. / The « Great Recession » of 2008-2009 and the sovereign and public debt crises which strengthened in the euro area in the summer of 2011 are recent events that have crystallized the challenges facing economic analysis, especially those related to dating and predicting cyclical inflections of real activity. The purpose of this thesis is to study these two complementary approaches to the economic cycle. Chapter 1 provides a portrait of the cycle using three distinct conceptions of its turning points: the classical cycle, the growth cycle and the acceleration cycle. We also discuss the measurement of the cycle with respect to various possible representations of aggregate activity of a country, as well as to two existing traditions which encompass dating models. Moreover, we highlight the growing influence of the financial environment over business cycle fluctuations.In chapter 2, we develop two non-parametric algorithms in order to identify theinflections that are particular to each of the previously conceptualized cycles, but also to measure their main characteristics. The first algorithm is based on a univariate representation of overall economic activity, the second on its ultivariate representation; ultimately, we apply the algorithms to the data of the French economy between 1970 and 2010. Chapter 3 builds on our results for cyclical dating to predict French recessions since 1974. Using probit models, we illustrate the role of monetary and financial variables as leading indicators of French business cycle fluctuations. In addition, we show that our models accurately detect recessions for a forecasting lag of two-quarters. Chapter 4 extends the entire analysis to several member states of the euro zone, with observations beginning in 1979. We first construct a chronology of their classical cycles, and then we propose an analysis of their main characteristics and their degree of synchronization.Finally, based on financial and monetary indicators in the context of a dynamic probit with fixed effects, we can anticipate the recessionary episodes which occurred in these economies with a horizon of two quarters.

CHAPTER 1:<p>The default g-priors predominant in Bayesian Model Averaging tend to over-concentrate posterior mass on a tiny set of models - a feature we denote as 'supermodel effect'. To address it, we propose a 'hyper-g' prior specification, whose data-dependent shrinkage adapts posterior model distributions to data quality. We demonstrate the asymptotic consistency of the hyper-g prior, and its interpretation as a goodness-of-fit indicator. Moreover, we highlight the similarities between hyper-g and 'Empirical Bayes' priors, and introduce closed-form expressions essential to computationally feasibility. The robustness of the hyper-g prior is demonstrated via simulation analysis, and by comparing four vintages of economic growth data.<p><p>CHAPTER 2:<p>Ciccone and Jarocinski (2010) show that inference in Bayesian Model Averaging (BMA) can be highly sensitive to small data perturbations. In particular they demonstrate that the importance attributed to potential growth determinants varies tremendously over different revisions of international income data. They conclude that 'agnostic' priors appear too sensitive for this strand of growth empirics. In response, we show that the found instability owes much to a specific BMA set-up: First, comparing the same countries over data revisions improves robustness. Second, much of the remaining variation can be reduced by applying an evenly 'agnostic', but flexible prior.<p><p>CHAPTER 3:<p>This chapter explores the link between the leverage of the US financial sector, of households and of non-financial businesses, and real activity. We document that leverage is negatively correlated with the future growth of real activity, and positively linked to the conditional volatility of future real activity and of equity returns. <p>The joint information in sectoral leverage series is more relevant for predicting future real activity than the information contained in any individual leverage series. Using in-sample regressions and out-of sample forecasts, we show that the predictive power of leverage is roughly comparable to that of macro and financial predictors commonly used by forecasters. <p>Leverage information would not have allowed to predict the 'Great Recession' of 2008-2009 any better than conventional macro/financial predictors. <p><p>CHAPTER 4:<p>Model averaging has proven popular for inference with many potential predictors in small samples. However, it is frequently criticized for a lack of robustness with respect to prediction and inference. This chapter explores the reasons for such robustness problems and proposes to address them by transforming the subset of potential 'control' predictors into principal components in suitable datasets. A simulation analysis shows that this approach yields robustness advantages vs. both standard model averaging and principal component-augmented regression. Moreover, we devise a prior framework that extends model averaging to uncertainty over the set of principal components and show that it offers considerable improvements with respect to the robustness of estimates and inference about the importance of covariates. Finally, we empirically benchmark our approach with popular model averaging and PC-based techniques in evaluating financial indicators as alternatives to established macroeconomic predictors of real economic activity. / Doctorat en Sciences économiques et de gestion / info:eu-repo/semantics/nonPublished