Professor Tarun Ramadorai

This course is available on the MSc in Finance (full-time), MSc in Finance (full-time) (Work Placement Pathway), MSc in Finance (part-time), MSc in Finance and Economics, MSc in Finance and Economics (Work Placement Pathway), MSc in Finance and Private Equity, MSc in Finance and Private Equity (Work Placement Pathway) and MSc in Finance and Risk. This course is not available as an outside option to students on other programmes. This course uses controlled access as part of the course selection process. For information on controlled access courses, including eligibility, application processes, deadlines, and departmental contact details, please refer to the Controlled Access Courses webpage.

All students on a programme listed under the Course Availability will be given a place. The course is not capped.

Please contact finance.teachingmanager@lse.ac.uk with any queries.

This course does not permit auditing students.

Co-requisites:

Students must complete FM422E or FM403 or FM436 or FM422 either before taking this course or in the same year as this course.

This course explores how large datasets, empirical methods including machine learning, and insights from behavioural finance and economics are transforming financial decision-making for households and other market participants. Our primary focus is two rapidly developing areas: credit analytics and asset management strategies.

The course begins by introducing key supervised machine learning tools used for classification and regression, and the range of particular choices involved in generating effective and accurate predictions. Although we use tree-based models as our main examples, the key concepts we highlight apply broadly across many machine learning approaches.

The first major application area is credit and mortgage analytics. Here, we examine how these methods can be employed to forecast default in both corporate credit markets and retail settings such as credit cards and peer-to-peer lending. We then focus on the mortgage market, one of the largest consumer credit markets, to discuss how machine learning is influencing mortgage selection, refinancing behaviour, and default prediction. Throughout, we will emphasise the underlying economic and financial factors driving rapid innovation in B2C lending.

Our attention then shifts to asset management, where new techniques and the use of large unstructured datasets are reshaping industry practice. After reviewing standard time-series portfolio construction methods, we consider how recent innovations and techniques can enhance these approaches, including those used in quantitative hedge fund strategies. We also explore how text and other unstructured data sources are increasingly used in tasks such as quantitative portfolio design and financial analysis.

Across all topics, the course maintains a strong focus on financial markets from supply, demand, and regulatory perspectives, making use of a variety of empirical models to illustrate key concepts.

33 hours of lectures and 11 hours of workshops in the Winter Term.

This course is taught in the interactive lecturing format. There is no distinction between lectures and classes/seminars; there are “sessions” only, and the pedagogical approach in each session is interactive.

In the absence of a suitable textbook in this new field, we will be reading straight from original research papers. In many cases these papers are less than a few years old. For fundamental machine learning methods, two useful textbooks are: “An Introduction to Statistical Learning” by Gareth James, Daniela Witten, Trevor Hastie and Robert Tibshirani; and “Elements of Statistical Learning” by Trevor Hastie, Robert Tibshirani, and Jerome Friedman.

Continuous assessment (100%).