Keynote speakers

Harriet Richards is a Manager in the Climate Risk Team at the Bank of England. In the team, she assesses the impact of climate-related financial risks on the Bank's balance sheet, with a particular focus on residential mortgage collateral. Prior to this, Harriet worked at HM Treasury in a team that supported Mark Carney as Finance Advisor to the UK Prime Minister ahead of COP26. She has also worked as a credit risk analyst at the Bank of England. Harriet holds a MSc in Environmental Economics from the London School of Economics.

Title of keynote: Measuring and mitigating climate-related financial risks using scenario analysis - a case study on residential mortgages

The physical impacts of climate change and the transition to net zero pose financial risks that are relevant to a wide range of institutions across the financial system. However, they are challenging to quantify, which could limit financial institutions’ abilities to mitigate against these risks. Using residential mortgages as a case study, we explore how practitioners can use scenario analysis to quantify and mitigate climate-related financial risks. We focus on how practitioners can ‘extend’ macro-climate scenarios to undertake granular asset-level analysis of financial risks. We also discuss how the Bank of England has applied this approach in practice to mitigate climate risks to residential mortgage collateral posted with the Bank.

Emilio Carrizosa is a data scientist, Professor of Statistics and Operational Research in the University of Seville (Spain) and President of the Spanish Network for Mathematics in Industry math-in.

His research interests include modelling and optimization in Machine Learning processes, with main focus on interpretability and fairness issues. He has published more than 150 papers on optimization and data-drive decision making, receiving by his contributions, among others, the medal of the Spanish Statistics and Operational Research Society (SEIO), and in 2024 the prize SEIO-BBVA Foundation to the best paper in Statistics and Operational Research applied to Data Science and Big Data.

Title of keynote: Raiders of the Lost Interpretability

The adoption of so called Explainable AI, which is typically 'black box' machine learning models accompanied by post-hoc explainability tools, are becoming more common; however, the concern remains for high risk area: can we trust post-hoc explainers?

Stephen Miller PhD is a data science practitioner and researcher with nearly two decades of industry experience, including 10 years at Equifax where he has led innovation projects and teams as a technical leader and manager. As a Data Science leader within the Equifax Global DS Lab, he has authored 13 patents and applications and has presented multiple papers at previous conferences; including the CSCC 2023 Best Paper Award, for work on predictive modeling in affordability assessment.

Title of keynote: Machine Learning at the Credit Bureau: The Role of Predictive Modelling in a Regulated Landscape

Technological advancements are enabling the development and implementation of increasingly sophisticated predictive models, which are central to most credit decisions taken by lenders. At the same time, regulators are mandating that these decisions are fair and transparent for consumers. While sharing similar objectives, regulatory requirements vary across different countries. Changing macroeconomic conditions and lenders' own governance processes impose additional constraints on how models can be utilised in order to make such decisions.

Looking primarily through the lens of credit risk and affordability assessment, this talk will explore the ways the needs of regulators, consumers and lenders impact the development and use of predictive models and scores, together with the implications for lenders and credit bureaus.

Professor Billio has been Head of the Treviso Branch and Head of the Department of Economics of Ca’ Foscari University. She is currently coordinator of the Master Degree Programme in Economics, Finance and Sustainability and member of the Academic Senate of Ca’ Foscari University of Venice. She is Fellow of the Institut Louis Bachelier (Paris) and Research Fellow Leibniz Institute SAFE (Frankfurt).

Her main research interests include financial econometrics; risk management; business cycle analysis; systemic risk, financial stability and sustainable finance. She is participating in many research projects financed by the European Commission, World Bank, European Investment Bank, Eurostat and the Italian Ministry of Research (MIUR). She has been the scientific coordinator of the SYRTO EU-FP7 project devoted to systemic risk measurement, the coordinator of the H2020 project TrAnsparEEnS dedicated to the development of ESG ratings for small and medium enterprises, and the local coordinator of five European projects on Energy Efficiency (EeMAP, EeDaPP, EeMMIP, ENGAGE and DeliverEEM). Currently, she is responsible of the Sustainable Finance workpage of the Italian NextGenerationEU program GRINS (Growing Resilient INclusive and Sustainable) and coordinator of a DG Reform Technical Support Initiative (ESG Uptake - ESG risk management framework for the financial sector). The results of these and other research projects have appeared in peer-refereed journals including Journal of Econometrics, Journal of Financial Economics, Journal of Applied Econometrics, Journal of Business and Economics Statistics, Journal of the American Statistical Association.

Title of keynote: Network extraction and modelling

Multidimensional arrays (i.e. tensors) of data are becoming increasingly available and call for suitable econometric tools. Approaches are first revised for extraction of the network also discussing the importance of topology and structure of the data. A new dynamic linear regression model is then proposed for tensor-valued response variables and covariates that encompasses some well-known multivariate models such as SUR, VAR, VECM, panel VAR and matrix regression models as special cases. For dealing with the over-parametrization and over-fitting issues due to the curse of dimensionality, a suitable parametrization is exploited based on the parallel factor (PARAFAC) decomposition, which enables the achievement of both parameter parsimony and incorporates sparsity effects. The contribution is twofold: first, an extension of multivariate econometric models is provided to account for both tensor-variate response and covariates; second, the effectiveness of the proposed methodology is shown in defining an autoregressive process for time-varying real economic networks. Inference is carried out in the Bayesian framework combined with Monte Carlo Markov Chain (MCMC). Finally, the model is applied to analyse the temporal evolution of real economic networks.