Vector autoregressive models have widely been applied in macroeconomics and macroeconometrics to estimate economic relationships and to empirically assess theoretical hypothesis. To achieve the latter, we propose a Bayesian inference approach to analyze the dynamic interactions among macroeconomics variables in a graphical vector autoregressive model. The method decomposes the structural model into multivariate autoregressive and contemporaneous networks that can be represented in the form of a directed acyclic graph. We then simulated the networks with an independent sampling scheme based on a single-move Markov Chain Monte Carlo (MCMC) approach. We evaluated the efficiency of our inference procedure with a synthetic data and an empirical assessment of the business cycles hypothesis.
We propose a new class of interacting Markov chain Monte Carlo (MCMC) algorithms designed for increasing the efficiency of a modified multiple-try Metropolis (MTM) algorithm. The extension with respect to the existing MCMC literature is twofold. The sampler proposed extends the basic MTM algorithm by allowing different proposal distributions in the multiple-try generation step. We exploit the structure of the MTM algorithm with different proposal distributions to naturally introduce an interacting MTM mechanism (IMTM) that expands the class of population Monte Carlo methods. We show the validity of the algorithm and discuss the choice of the selection weights and of the different proposals. We provide numerical studies which show that the new algorithm can perform better than the basic MTM algorithm and that the interaction mechanism allows the IMTM to efficiently explore the state space.
Multiple time series data may exhibit clustering over time and the clustering effect may change across different series. This paper is motivated by the Bayesian non-parametric modelling of the dependence between clustering effects in multiple time series analysis. We follow a Dirichlet process mixture approach and define a new class of multivariate dependent Pitman-Yor processes (DPY). The proposed DPY are represented in terms of a vector of stickbreaking processes which determines dependent clustering structures in the time series. We follow a hierarchical specification of the DPY base measure to accounts for various degrees of information pooling across the series. We discuss some theoretical properties of the DPY and use them to define Bayesian non-parametric repeated measurement and vector autoregressive models. We provide efficient Monte Carlo Markov Chain algorithms for posterior computation of the proposed models and illustrate the effectiveness of the method with a simulation study and an application to the United States and the European Union business cycles.
High dimensional vector autoregressive (VAR) models require a large number of parameters to be estimated and may suffer of inferential problems. We propose a new Bayesian nonparametric (BNP) Lasso prior (BNP-Lasso) for high-dimensional VAR models that can improve estimation efficiency and prediction accuracy. Our hierarchical prior overcomes overparametrization and overfitting issues by clustering the VAR coefficients into groups and by shrinking the coefficients of each group toward a common location. Clustering and shrinking effects induced by the BNP-Lasso prior are well suited for the extraction of causal networks from time series, since they account for some stylized facts in realworld networks, which are sparsity, communities structures and heterogeneity in the edges intensity. In order to fully capture the richness of the data and to achieve a better understanding of financial and macroeconomic risk, it is therefore crucial that the model used to extract network accounts for these stylized facts.
T he p ur p o se o f t h is p ap er i s t he co n str u cti o n o f a n e arl y wa r ni n g ind ic ato r fo r s ys t e mi c r i s k u s i n g e ntro p y me as u r es. T he a n al ys i s i s b a se d o n t he c r o s s-se ct io nal d is tr ib u tio n o f mar g i na l s ys t e mic r is k me a s ur es s uc h a s Mar gi n al E x p ected S ho r t fa ll, De lt a Co Va R a nd n et wo r k con n ect ed ne s s. T h e se m eas u re s ar e co n ce i ved a t a s i n gl e i n s ti t ut io n fo r th e f i na nc ia l i nd u s tr y i n t he E uro ar ea. W e e st i ma te e ntro p y o n t he s e me a s ur e s b y co n s id er i n g d i ffere n t d e fi ni tio n s (S ha n no n, T s al li s a n d Re n yi). F i na ll y, we t e st i f t he se e n tro p y i nd ica to rs s ho w fo r eca s ti n g ab il it ie s i n p r ed i cti n g b an k i n g cri se s. I n t hi s re gard , we u se t h e v ariab l e p res e nted i n B ab e c k ỳ e t al. (2 0 1 2) a nd Ale s si and De t ke n (2 0 1 1) fro m Eur o p ea n C e ntr al B a n k. E ntro p y i n d ic ato r s s ho w p ro mi s i n g fo r eca s t ab il it ie s to p r ed ict f i n an cia l a nd b a n ki n g cri si s. T he p ro p o sed ear l y wa r ni n g si g na l s r e v ea l to b e e ffe ct i ve i n fo r e cas ti n g fi n a nci al d i st re s s co nd it io n s.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.