Programme and keynote speakers

Karen Croxson is Deputy Chief Economist, Head of Research, and Social Data Science at the Financial Conduct Authority (FCA). She leads the FCA’s scientific research agenda, integrating economics, data science, and behavioural science to support evidence-based decision making across retail and wholesale markets, and improve FCA internal operations.

Before joining the FCA, she worked as a data science leader at QuantumBlack, a strategy consultant at McKinsey & Co, and an academic economist at Oxford University.

Keynote Topic: Credit Scoring — Evolution of the Market and Impact of AI

Karen will focus on the future evolution of the credit information market and some of the potential implications of big data and AI. Her presentation will outline the strategic objectives of the FCA and the importance of credit information for the healthy functioning of the retail lending sector and consumer outcomes.

She will discuss the FCA’s Credit Information Market Study, including scope and latest timings and next steps. Finally, she will discuss some recent thinking on potential scenarios for the evolution of the market and touch on some of the potential implications of AI and Machine Learning in the credit scoring context.

Andrew has spent the last 22 years leading Credit Risk Modelling teams in major UK banks.

After graduating, he joined Barclays and spent the first 11 years learning about general banking, doing various roles in:

• Corporate relationship management
• Risk analysis
• Development capital finance
• Project and people management

After a sabbatical to undertake an MBA he entered the world of credit risk modelling and found it so interesting that he decided not to leave!

Having lead the Commercial Banking Modelling team at Barclays for 14 years, Andrew moved to Lloyds Banking Group to undertake a similar role for 7 years. More recently he has joined the Prudential Regulation Authority (PRA) to see life from the other side of the fence, leading the Credit Risk Measurement team.

Keynote Topic: Managing Prudential Risk through Policy; Hybrid Mortgage PD Models

Dr Andrew Curtis is the Senior Manager at the PRA responsible for reviewing firms’ models under the Internal Ratings Based (IRB) approach, and in his keynote address to conference he will focus on regulatory policy initiatives in the area of retail mortgages.

The presentation will outline the strategic objectives of the PRA, and current concerns on unwarranted variability of risk weights produced by mortgage models across different firms. Andrew will discuss the PRA response to this issue, namely the requirement that all IRB firms move to a hybrid approach for modelling mortgage Probability of Default (PD).

He will build on previous statements from the PRA to outline key themes such as approaches to backcast long run default rates applicable where firms do not have sufficient internal data. Andrew will also discuss the issue of cyclicality, which affects how responsive the PD model is to changes in the underlying default rate of the mortgage portfolio.

Finally, he will outline some key considerations to successfully develop, implement, and embed these models in business and risk decisions, with a focus on the key role of effective model risk management when there is a significant change in rating philosophy.

Professor Hand is the former chair of the Statistics Section of Imperial College, London. He is a Fellow of the British Academy and a former President of the Royal Statistical Society. He has been awarded many honours, including the Guy Medal of the Royal Statistical Society, the Box Medal from the European Network for Business and Industrial Statistics, the International Research Medal of the IFCS, and was made Officer of the Order of the British Empire (OBE) for research and innovation in 2013.

Professor Hand has collaborated with most of the major players in the retail credit world, developing tools for credit scoring and fraud detection. Between 2010 and 2018 he took an extended sabbatical to work as Chief Scientific Advisor to Winton Capital Management. He has published over 300 scientific papers and 31 books, including Principles of Data Mining, Measurement Theory and Practice, The Improbability Principle, Statistics: A Very Short Introduction, The Wellbeing of Nations, and From GDP to Sustainable Wellbeing. His latest book, Dark Data: Why What You Don’t Know Matters, deals with the challenges for statistics, machine learning, and AI arising from incomplete and distorted data.

Keynote Topic: Consumer Credit, the Data Science, and AI Revolution

The media are full of stories about how data science, statistics, and AI are leading a revolution in our ability to gain understanding and make predictions about the world around us. But large data sets and sophisticated models have been the order of the day for decades in the world of credit scoring.

In David's talk he will examine recent developments and see what they have to offer the world of credit scoring, drawing attention to the challenges posed by and risks presented by modern developments.

Agus Sudjianto is an executive vice president, head of Model Risk and a member of Management Committee at Wells Fargo, where he is responsible for enterprise model risk management.

Prior to his current position, Agus was the modelling and analytics director and chief model risk officer at Lloyds Banking Group in the UK. Before joining Lloyds, he was an executive and head of Quantitative Risk at Bank of America. Prior to his career in banking, he was a product design manager in the Powertrain Division of Ford Motor Company.

Agus holds several US patents in both finance and engineering. He has published numerous technical papers and is a co-author of Design and Modelling for Computer Experiments. His technical expertise and interests include quantitative risk, particularly credit risk modelling, machine learning, and computational statistics.

He holds masters and doctorate degrees in engineering and management from Wayne State University and the Massachusetts Institute of Technology.

Keynote Topic: Machine Learning and Credit Underwriting

The financial industry has rapidly adopted machine learning for various applications. Large banks in the US are typically more cautious in adopting the methodology for high risk or regulated application such as credit underwriting.

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?

Alternatively, there are many recent developments on inherently interpretable, self-explanatory machine learning models without the problem of post-hoc explainers. The later offers many advantages beyond explainability such as more comprehensive model diagnostics and control to manage model risk which is particularly important for applications such as Credit Scoring.

This is the focus of Agus's talk where he will present examples of self-explanatory machine learning models including methods to incorporate constraints (for example, monotonicity or other shape constraints) and to generate appropriate adverse action reason code required by regulation in the US.