[pt] Modelos auto-regressivos (AR(p)) de séries temporais supõem que a dinâmica da série contém uma dependência linear nas observações passadas até uma defasagem p, e um erro aleatório independente e identicamente distribuído (i.i.d). Modelos de auto-regressão quantílica (QAR(p)) são uma generalização dos AR(p) em que os coeficientes auto- regressivos variam com o quantil da distribuição condicional, não sendo necessária, portanto, uma componente explícita de erro aleatório. Esta dissertação estuda a inferência estatística proposta para modelos QAR(p) por Koenker e Xiao (2004), com o auxílio de simulações de Monte Carlo. Enquanto a estimação mostra-se bem precisa, os resultados do teste de hipóteses, onde a hipótese nula supõe um modelo auto-regressivo (AR), não apresentam bons resultados, variando estes com o modelo gerador de dados. / [en] Autoregressive models (AR(p)) for time series assume that the series dynamics has a linear dependence on past observations up to a lag p, plus an independent and identically distributed (i.i.d.) random error. Quantile autoregressive models (QAR(p)) generalize the AR(p) by allowing different autoregressive coefficients for different quantiles of the conditional distribution and so there is no need for an explicit random error component. This dissertation studies the statistical inference proposed by Koenker e Xiao (2004) for QAR(p) models, by means of Monte Carlo simulations. While the estimation tools show themselves very accurate, the hypothesis test which considers an AR model as the null hypothesis yields poor results, and these vary with the data generating process

This doctoral thesis is a collection of four papers on the analysis of the term structure of interest rates with a focus at the intersection of macroeconomics and finance. "Risk in Macroeconomic Fundamentals and Bond Return Predictability" documents that factors related to risks underlying the macroeconomy such as expectations, uncertainty and downside (upside) macroeconomic risks are able to explain variation in bond risk premia. The information provided is found to be, to a large extent, unrelated to that contained in forward rates and current macroeconomic conditions. "Out-of-sample bond excess returns predictability" provides evidence that macroeconomic variables, risks in macroeconomic outcomes as well as the combination of these different sources of information are able to generate statistical as well as economic bond excess returns predictability in an out-of-sample setting. Results suggest that this finding is not driven by revisions in macroeconomic data. The term spread (yield curve slope) is largely used as an indicator of future economic activity. "Re-examining the predictive power of the yield curve with quantile regression" provides new evidence on the predictive ability of the term spread by studying the whole conditional distribution of GDP growth. "Modeling and forecasting the yield curve by extended Nelson-Siegel class of models: a quantile regression approach" deals with yield curve prediction. More flexible Nelson-Siegel models are found to provide better fitting to the data, even when penalizing for additional model complexity. For the forecasting exercise, quantile-based models are found to overcome all competitors. / <p>Diss. Stockholm :  Stockholm School of Economics, 2014. Introduction together with 4 papers.</p>

In dieser Arbeit untersuchen wir einige Möglichkeiten, financial und soziale Netzwerke zu analysieren, ein Thema, das in letzter Zeit in der ökonometrischen Literatur große Beachtung gefunden hat. Kapitel 2 untersucht den Risiko-Spillover-Effekt über das in White et al. (2015) eingeführte multivariate bedingtes autoregressives Value-at-Risk-Modell. Wir sind an der Anwendung auf nicht stationäre Zeitreihen interessiert und entwickeln einen sequentiellen statistischen Test, welcher das größte verfügbare Homogenitätsintervall auswählt. Unser Ansatz basiert auf der Changepoint-Teststatistik und wir verwenden einen neuartigen Multiplier Bootstrap Ansatz zur Bewertung der kritischen Werte. In Kapitel 3 konzentrieren wir uns auf soziale Netzwerke. Wir modellieren Interaktionen zwischen Benutzern durch ein Vektor-Autoregressivmodell, das Zhu et al. (2017) folgt. Um für die hohe Dimensionalität kontrollieren, betrachten wir ein Netzwerk, das einerseits von Influencers und Andererseits von Communities gesteuert wird, was uns hilft, den autoregressiven Operator selbst dann abzuschätzen, wenn die Anzahl der aktiven Parameter kleiner als die Stichprobegröße ist. Kapitel 4 befasst sich mit technischen Tools für die Schätzung des Kovarianzmatrix und Kreuzkovarianzmatrix. Wir entwickeln eine neue Version von der Hanson-Wright- Ungleichung für einen Zufallsvektor mit subgaußschen Komponenten. Ausgehend von unseren Ergebnissen zeigen wir eine Version der dimensionslosen Bernstein-Ungleichung, die für Zufallsmatrizen mit einer subexponentiellen Spektralnorm gilt. Wir wenden diese Ungleichung auf das Problem der Schätzung der Kovarianzmatrix mit fehlenden Beobachtungen an und beweisen eine verbesserte Version des früheren Ergebnisses von (Lounici 2014). / In this work we explore some ways of studying financial and social networks, a topic that has recently received tremendous amount of attention in the Econometric literature. Chapter 2 studies risk spillover effect via Multivariate Conditional Autoregressive Value at Risk model introduced in White et al. (2015). We are particularly interested in application to non-stationary time series and develop a sequential test procedure that chooses the largest available interval of homogeneity. Our approach is based on change point test statistics and we use a novel Multiplier Bootstrap approach for the evaluation of critical values. In Chapter 3 we aim at social networks. We model interactions between users through a vector autoregressive model, following Zhu et al. (2017). To cope with high dimensionality we consider a network that is driven by influencers on one side, and communities on the other, which helps us to estimate the autoregressive operator even when the number of active parameters is smaller than the sample size. Chapter 4 is devoted to technical tools related to covariance cross-covariance estimation. We derive uniform versions of the Hanson-Wright inequality for a random vector with independent subgaussian components. The core technique is based on the entropy method combined with truncations of both gradients of functions of interest and of the coordinates itself. We provide several applications of our techniques: we establish a version of the standard Hanson-Wright inequality, which is tighter in some regimes. Extending our results we show a version of the dimension-free matrix Bernstein inequality that holds for random matrices with a subexponential spectral norm. We apply the derived inequality to the problem of covariance estimation with missing observations and prove an improved high probability version of the recent result of Lounici (2014).