Roberto Casarin


University Ca' Foscari of Venice Department of EconomicsSan Giobbe 873/b, 30121 Venezia, Italy Office: Room A125 Phone: +39 041.234.91.49 +39 3297383331 E-mail: r.casarin@unive.itSkype: flubber618Linkedin

Since 2019 I am full professor of econometrics at University Ca' Foscari of Venice. I was previously associate and assistant professor at University Ca' Foscari, University of Brescia, research assistant at University of Padova, and research fellow at GRETA Associates. I received a Ph.D. in Mathematics, from University Paris Dauphine and CEREMADE, a Ph.D. in Economics from University Ca' Foscari of Venice, and a M.Sc. in Applied Mathematics from University Paris Dauphine and ENSAE. I have been visiting University Paris Sud, University of Bristol and University Paris Dauphine.

My research

It is mainly in Bayesian analysis: Bayesian nonparametric and semiparametric statistical inference, Monte Carlo simulation methods, time series models, text mining, graphical models, networks analysis. Here below a selection of the most recent research projects of mine.

A showcase of my projects

Here below a selection of some recent projects of mine published in international peer-reviewed journals.

I am director of the Venice Center for Risk Analytics VERA and of the International Master in Economics and Finance IMEF, board member of the European Society of Bayesian Econometrics ESOBE of the PhD program in Economics SSE, and of the Master in Data Analytics for Business and Societies DABS.

I am currently associate editor of Bayesian analysis and Econometrics.

I am research associate of the centers: ECLT, GRETA, member of the scientific societies: ISBA, IMS, ES, SIS, SIdE and of the projects: EABCN, ENBIS, SYRTO, Systemic Risk Hub, Performance Measures, SCSCF.

You can find me in the Math Genealogy Project with an Erdős number of 3. For further info on the impact of my research see: MathSciNet, ArXiv, GoogleScholar, SSRN, ORCID, Scopus, RePEc.

Hierarchical Species Sampling Models

We introduce a general class of hierarchical nonparametric distributions which includes new (e.g. hierarchical Gnedin), and well-known (e.g. Pitman-Yor and normalized measures) random measure. Our framework relis on generalized species sampling processes and provides a probabilistic foundation for hierarchical random measures. We show that hierarchical species sampling models have a Chinese Restaurants Franchise representation (see Figure) and can be used in Bayesian nonparametric inference.


Bassetti, F., Casarin, R., Rossini, L. (2020), Hierarchical Species Sampling Models, Bayesian Analysis, 15(3), 809-838.

Autoregressive Gamma Stochastic Volatility Models

We introduce a new family of discrete-time stochastic volatility models having two measurement equations relating both the observed returns and realized measures to the latent conditional variance. We provide an analytical filtering and smoothing recursions for the basic version of the model and an effective MCMC algorithm for heterogeneous volatility and leverage models. We present the first fully analytical pricing framework for discrete-time stochastic volatility models.


Bormetti, G., Casarin, R., Corsi, F. and Livieri, G. (2020), A stochastic volatility model with realized measures for option pricing, Journal of Business and Economic Statistics, 38(4), 856-871.

What Makes a Tweet be Retweeted?

We propose efficient inverse regression Bayesian method for analysis of tweet propagation of political messages. We find that politicians (see Figure) were able to identify the combination of sensitive words to enhance the probability of retweet of the message with an impact on political outcomes. Our work entail an examination of a neglected unit of analysis (trigram) in a language less studied (i.e. Spanish), and an innovative Bayesian approach to the predictive power that retweets have on electoral results . This new methodology can be used to adjust political messages as a means to increase voters engagement in political campaigns.


Casarin, R., German, M. and Ter Horst, E. (2019), What makes a tweet be retweeted? A Bayesian Trigram Analysis of Tweet Propagation during the 2015 Colombian Political Campaign, Journal of Information Science, forthcoming.

Multilayer Network Extraction

We propose a novel Bayesian graphical vector autoregressive model to extract multi-layer and multi-country networks. An example is provided (see Figure) illustrating the relationships for United States, Arabian Peninsula and key middle east producers (colored nodes), with a decomposition in production (red nodes) and rigs deployment (blue) layers. Production is less affected to changes in oil prices than rigs. Production is a lagged driver for prices, which indicates that the rigs-production linkasges may not be fully accounted for in the markets.


Casarin, R., Iacopini, M., German, M. and Ter Horst, E., Espinasa, R., Sucre, C. and Rigobon, R. (2020), Multilayer network analysis of oil linkages, Econometrics Journal, 23(2), 269–296.

Bayesian Nonparametric Sparse VAR

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. The BNP-Lasso prior is well suited for causal networks extraction, since they account for some features of real-world networks such as sparsity and communities structures.


Billio, M., Casarin, R., Rossini, L. (2019), Bayesian Nonparametric Sparse VAR Models, Journal of Econometrics, 212(1), 97-115.

A Bayesian Time-varying Spline Method for Multiple Equations

We expand the literature of risk neutral density estimation across maturities from implied volatility curves by developing a Bayesian time-varying spline method for multiple equations (see Figure). The dependence between the nonparametric risk-neutral densities allows for information borrowing across time and maturities.


Casarin, R., Molina, G., Ter Horst, E. (2019), A Bayesian Time-Varying Approach to Risk Neutral Density Estimation, Journal of the Royal Statistical Society, Series A, 182 (1), 165-195.

Markov Switching Graphical Models and Systemic Risk

We propose a new Markov Switching Graphical Seemingly Unrelated Regression model to investigate time-varying systemic risk based on a range of multi-factor asset pricing models. We show that regimes can be identified by using network statistics, and develop a novel weighted eigenvector centrality measure. Our empirical application to the S&P100 constituents shows that firm-level centrality does not correlate with market values and is instead positively linked to realized financial losses.


Bianchi, D., Billio, M., Casarin, R., Guidolin, M. (2019), Modeling Systemic Risk with Markov Switching Graphical SUR Models, Journal of Econometrics, 210(1), 58-74.

A Distributional Homogeneity Test

We propose a new Bayesian nonparametric homogeneity test for distributional changes. We provide an asymptotic approximation of the Bayes factor and show that it is related to the Shannon entropy. Our test is suitable for large high-dimensional datasets since it does not require time-consuming computations. On the FRED-QD dataset our test detected relevant structural changes in the US economy (see Figure).


Casarin, R., Costola, M. (2019), Structural changes in large economic datasets: A nonparametric homogeneity test, Economics Letters, 176, 55-59.

Opinion Dynamics and Disagreements on Networks

We propose new measures of disagreement based on connectedness. Building on the lifting approach in Hendrickx (2014), we consider general consensus dynamics and disagreement with antagonistic behaviour on signed networks. Synthetic and real-world financial networks of serial correlation are considered for illustrating the new measure (see Figure) and for studying opinion dynamics and convergence to consensus on prices.


Billio, M., Casarin, R., Costola, M., Frattarolo, L. (2019), Opinion Dynamics and Disagreements on Financial Networks, Advances in Decision Sciences, 23(4), 1-27