Bayesian Graphical Models for Structural Vector Autoregressive Processes

Ahelegbey, Daniel Felix; Billio, Monica; Casarin, Roberto

doi:10.2139/ssrn.2198844

Cited by 60 publications

(110 citation statements)

References 60 publications

Supporting

Mentioning

110

Contrasting

Order By: Relevance

“…Methodologically, we build upon the Gaussian Graphical model for multi-variate systems proposed by Whittaker (1990), Dawid and Lauritzen (1993), Lauritzen (1996), , Wang and West (2009), Wang (2010), Rodriguez, Lenkoski, and Dobra (2011), Wang, Reeson, and Carvalho (2011) and Ahelegbey, Billio, and Casarin (2016). In particular Wang et al (2011) developed a dynamic matrixvariate graphical model which allows to capture conditional dependencies under a time-invariant graphs.…”

Section: Introductionmentioning

confidence: 99%

Modeling systemic risk with Markov Switching Graphical SUR models

Bianchi

Billio

Casarin

et al. 2019

Journal of Econometrics

Self Cite

View full text Add to dashboard Cite

We propose a Markov Switching Graphical Seemingly Unrelated Regression (MS-GSUR) model to investigate time-varying systemic risk based on a range of multi-factor asset pricing models. Methodologically, we develop a Markov Chain Monte Carlo (MCMC) scheme in which latent states are identified on the basis of a novel weighted eigenvector centrality measure. An empirical application to the S&P100 constituents shows that cross-firm connectivity significantly increased over the period 1999-2003 and the financial crisis of 2008-2009. Finally, we provide evidence that firm-level centrality does not correlate with market values and is instead positively linked to realized financial losses.

show abstract

Section: Introductionmentioning

confidence: 99%

Modeling systemic risk with Markov Switching Graphical SUR models

Bianchi

Billio

Casarin

et al. 2019

Journal of Econometrics

Self Cite

View full text Add to dashboard Cite

show abstract

“…In particular, to estimate 4 This identi cation strategy builds on the previous works by by Swanson and Granger (1997); Bessler and Lee (2002); Demiralp and Hoover (2003); Moneta (2008). Recently also Bayesian strategies have been used to tackle the same issue (see Ahelegbey et al, 2016). 5 For more details about Independent Component Analysis, its assumptions, algorithms and related concepts see Hyvarinen et al (2001) and references therein.…”

Section: Methodsmentioning

confidence: 99%

The Janus-Faced Nature of Debt: Results from a Data-Driven Cointegrated SVAR Approach

et al. 2017

View full text Add to dashboard Cite

Standard-Nutzungsbedingungen:Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Zwecken und zum Privatgebrauch gespeichert und kopiert werden.Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich machen, vertreiben oder anderweitig nutzen.Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen (insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, gelten abweichend von diesen Nutzungsbedingungen die in der dort genannten Lizenz gewährten Nutzungsrechte. Terms of use: Documents in AbstractIn this paper, we investigate the causal e ects of public and private debts on U.S. output dynamics. We estimate a battery of Cointegrated Structural Vector Autoregressive models, and we identify structural shocks by employing Independent Component Analysis, a data-driven technique which avoids ad-hoc identi cation choices. The econometric results suggest that the impact of debt on economic activity is Janus-faced. Public debt shocks have positive and persistent in uence on economic activity. In contrast, rising private debt has a milder positive impact on GDP, but it fades out over time. The analysis of the possible transmission mechanisms reveals that public debt crowds in private consumption and investment. In contrast, mortgage debt fuels consumption and output in the short-run, but shrinks them in the mediumrun.

show abstract

“…Clearly, the WSBM approach is independent by the chosen methodology in the network estimation and alternative techniques such as graph-based approaches (i.e. Ahelegbey et al, 2016a) and sparse models (i.e. Ahelegbey et al, 2016c;Hautsch et al, 2015) can be applied.…”

Section: Introductionmentioning

confidence: 99%

Financial Bridges and Network Communities

2017

Self Cite

View full text Add to dashboard Cite

Financial crises and systemic risks proved to play central roles as shock transmitters to the real economy, threatening the stability of the economic and financial systems. These phenomena have boosted a massive interest in the literature which deeply investigated systemic risk and contagion channels on the financial systems.Generally, systemic risk may arise as the interactions among financial institutions and markets which, consequently, lead to financial crises. A stylized fact that occurs often in real networks is the presence of a group of nodes which share common properties or play a similar role within the network. These community structures have been recognized also in finance with the presence of key nodes (community bridges) linked through short-cuts to otherwise separated communities. The case in point to better understand the role of the community structure in a network is provided by epidemiology. A parallelism with the financial stability indicates that the mitigation and prevention of the spread in infectious diseases (financial contagion) can be attained by seeking actively to immunize the super-spreaders. However, the presence of a community structure significantly affects the dynamic of the disease: immunization interventions focusing on nodes strongly linked with other communities (community bridges) are in this case more effective than the ones which aim to the highly connected nodes in the whole network. The reason is that community bridges are more relevant in spreading out contagion with respect to the nodes with fewer inter-community connections in the group: with the latter, contagion may stop before spreading out to the other communities. Hence, in a network with a community structure, classical connectedness measures can lead to the misidentification of a given SIFI at least in two cases: i) a financial institution shows a lower total degree with respect to the other nodes in the community, but a higher degree to the nodes belonging to other communities (false negative); ii) a financial institution shows a higher total degree with respect to the other nodes in the community, but a lower degree to the nodes belonging to other communities (false positive). Therefore, also in a financial network, the node immunization through the identification of highly connected nodes may not be effective in a network with community structure.On this ground, the aim of this paper is to investigate the topology of the financial networks focusing on the detection of financial communities and community bridges to overcome the weakness of classical connectedness measure. We denote these communities as the Systemically Important Financial Communities (SIFC) defined as a group of nodes that belong to the community with the highest inter-connectivity density of the network. In this regard, we propose measures of connectedness to describe the inter-and intra community connectivity in financial networks. In the empirical analysis, we investigate the European financial system from 1996 to 2013 including all the f...

show abstract

Bayesian Graphical Models for Structural Vector Autoregressive Processes

Cited by 60 publications

References 60 publications

Modeling systemic risk with Markov Switching Graphical SUR models

Modeling systemic risk with Markov Switching Graphical SUR models

The Janus-Faced Nature of Debt: Results from a Data-Driven Cointegrated SVAR Approach

Financial Bridges and Network Communities

Contact Info

Product

Resources

About