Dr Maria Kalli

Lecturer in Statistics


Prior to her academic career Maria worked in investment banking at Goldman Sachs in both New York and London at the Quantitative Strategies group of the Equities desk. 

She completed her BSc in Econometrics and Mathematical Economics at the London School of Economics, an MBA in Financial Engineering at New York University’s Stern Business School, an MSc in Mathematical Statistics at the University of Michigan and a Phd in Statistics here at Kent. She is a Fulbright Scholar, and a Senior Fellow of the UK Higher Education Academy.

Maria is the seminar organiser for the Statistics group and also organises the Financial Econometrics seminar series held jointly by SMSAS and KBS.

Research interests

Bayesian Statistics, especially Bayesian nonparametric methods, MCMC, Bayesian model selection and shrinkage, and Time Series analysis. The application areas she has an interest in are macroeconomics, financial econometrics, and micro econometrics. 


Current and past PhD students: 

  • Carl Gower, from 04/17 (Canterbury Christ Church University Business School) Bayesian time-varying model selection, with applications to macroeconomics and finance
  • Rachel Taylor, from 01/17 (COaST centre, Canterbury Christ Church University) Assessing the economic and social impact of cultural interventions in coastal areas - a Bayesian approach.
  • Michael Bedford, completed 05/15 (Canterbury Christ Church University Business School and SSPSSR) Bayesian predictive models and variable selection for Acute Kidney Injury.


Funding Committee Member:

  • National Institute for Health Research-Research for Patient Benefit-South East Coast (01/08-06/15)
  • National Institute for Health Research-Health Technology Assessment (06/12- 06/16) 

Grant reviewer for:

  • Engineering and Physical Sciences Research Council (from 01/13)
  • Economics and Social Research Council (from 01/12)
  • Medical Research Council (from 01/14) 
  • Statistics Editor for Manual Therapy Journal. (Jan 2011- Sept 2015)

  • Journal reviewer for Journal of Econometrics, Journal of the American Statistical Association, Journal of Business and Economic Statistics, Bayesian Analysis, Statistics and Computing, Journal of Applied Econometrics, Econometric Theory, Journal of the Royal Statistical Society-C, Journal of Financial Econometrics, Quantitative Finance, Review of Financial Studies, Computational Statistics and Data Analysis, Journal of Computational and Graphical Statistics, and Journal of Time Series Analysis. 

Member/Fellow of:

  • International Bayesian Society (ISBA) (from 01/08)
    Bayesian non-parametric section (from 01/10)
    Treasurer of the section (01/10-12/12)
    Economics, Finance and Business section (from 01/12)
    Computational section (from 01/12)
  • Royal Statistical Society (RSS) (from 01/09)
    Committee Member of East Kent RSS section (from 12/16)
  • Chartered Institute of Securities and Investments (from 06/12)



  • Kalli, M. and Griffin, J. (2018). Bayesian nonparametric vector autoregressive models. Journal of Econometrics [Online] 203:267-282. Available at: https://doi.org/10.1016/j.jeconom.2017.11.009.
    Vector autoregressive (VAR) models are the main work-horse model for macroeconomic forecasting, and provide a framework for the analysis of complex dynamics that are present between macroeconomic variables. Whether a classical or a Bayesian approach is adopted, most VAR models are linear with Gaussian innovations. This can limit the model’s ability to explain the relationships in macroeconomic series. We propose a nonparametric VAR model that allows for nonlinearity in the conditional mean, heteroscedasticity in the conditional variance, and non-Gaussian innovations. Our approach differs to that of previous studies by modelling the stationary and transition densities using Bayesian nonparametric methods. Our Bayesian nonparametric VAR (BayesNP-VAR) model is applied to US and UK macroeconomic time series, and compared to other Bayesian VAR models. We show that BayesNP-VAR is a flexible model that is able to account for nonlinear relationships as well as heteroscedas- ticity in the data. In terms of short-run out-of-sample forecasts, we show that BayesNP-VAR
    predictively outperforms competing models.
  • Griffin, J., Kalli, M. and Steel, M. (2017). Discussion of “Nonparametric Bayesian Inference in Applications”: Bayesian nonparametric methods in econometrics. Statistical Methods & Applications [Online]. Available at: https://doi.org/10.1007/s10260-017-0384-0.
    The use of Bayesian nonparametrics models has increased rapidly over the last few decades driven by increasing computational power and the development of efficient Markov chain Monte Carlo algorithms. We review some applications of these models in economic applications including: volatility modelling (using both stochastic volatility models and GARCH-type models) with Dirichlet process mixture models, uses in portfolio allocation problems, long memory models with flexible forms of time-dependence, flexible extension of the dynamic Nelson-Siegel model for interest rate yields and multivariate time series models used in macroeconometrics.
  • Kalli, M. and Griffin, J. (2015). Flexible Modelling of Dependence in Volatility Processes. Journal of Business and Economic Statistics [Online] 33:102-113. Available at: http://amstat.tandfonline.com/doi/abs/10.1080/07350015.2014.925457#.VOSycldJ54E.
    This article proposes a novel stochastic volatility (SV) model that draws from the existing literature on autoregressive SV models, aggregation of autoregressive processes, and Bayesian nonparametric modeling to create a SV model that can capture long-range dependence. The volatility process is assumed to be the aggregate of autoregressive processes, where the distribution of the autoregressive coefficients is modeled using a flexible Bayesian approach. The model provides insight into the dynamic properties of the volatility. An efficient algorithm is defined which uses recently proposed adaptive Monte Carlo methods. The proposed model is applied to the daily returns of stocks.
  • Kalli, M. and Griffin, J. (2014). Time-varying sparsity in dynamic regression models. Journal of Econometrics [Online] 178:779-793. Available at: http://dx.doi.org/10.1016/j.jeconom.2013.10.012.
    A novel Bayesian method for inference in dynamic regression models is proposed where both the values of the regression coefficients and the importance of the variables are allowed to change over time. We focus on forecasting and so the parsimony of the model is important for good performance. A prior is developed which allows the shrinkage of the regression coefficients to suitably change over time and an efficient Markov chain Monte Carlo method for posterior inference is described. The new method is applied to two forecasting problems in econometrics: equity premium prediction and inflation forecasting. The results show that this method outperforms current competing Bayesian methods.
  • Kalli, M., Walker, S. and Damien, P. (2013). Modelling the conditional distribution of daily stock index returns: an alternative Bayesian semiparametric model. Journal of Business and Economic Statistics [Online] 31:371-383. Available at: https://doi.org/10.1080/07350015.2013.794142.
    This paper introduces a new family of Bayesian semi-parametric models for the conditional distribution of daily stock index returns. The proposed models capture key stylised facts of such returns, namely heavy tails, asymmetry, volatility clustering, and the ‘leverage effect’. A Bayesian nonparametric prior is used to generate random density functions that are unimodal and asymmetric. Volatility is modelled parametrically. The new model is applied to the daily returns of the S&P 500, FTSE 100, and EUROSTOXX 50 indices and is compared to GARCH, Stochastic Volatility, and other Bayesian semi-parametric models.
  • Kalli, M., Griffin, J. and Walker, S. (2011). Slice Sampling Mixture Models. Statistics and Computing [Online] 21:93-105. Available at: http://dx.doi.org/10.1007/s11222-009-9150-y.
    We propose a more efficient version of the slice sampler for Dirichlet process mixture models described by Walker (Commun. Stat., Simul. Comput. 36:45–54, 2007). This new sampler allows for the fitting of infinite mixture models with a wide-range of prior specifications. To illustrate this flexibility we consider priors defined through infinite sequences of independent positive random variables. Two applications are considered: density estimation using mixture models and hazard function estimation. In each case we show how the slice efficient sampler can be applied to make inference in the models. In the mixture case, two submodels are studied in detail. The first one assumes that the positive random variables are Gamma distributed and the second assumes that they are inverse-Gaussian distributed. Both priors have two hyperparameters and we consider their effect on the prior distribution of the number of occupied clusters in a sample. Extensive computational comparisons with alternative “conditional” simulation techniques for mixture models using the standard Dirichlet process prior and our new priors are made. The properties of the new priors are illustrated on a density estimation problem.
Last updated