CASE 6 - DATA-DRIVEN BANKRUPTCY PREDICTION

BACKGROUND

In the intricate world of investment banking, being able to assess the financial stability and potential risks associated with companies is of paramount importance. While traditional methods provide a myriad of metrics to analyze, synthesizing them into a single, digestible, and actionable score can significantly streamline the decision-making process for stakeholders. This concept is not foreign, as seen with ESG scores focusing on environmental, social, and governance factors. Drawing inspiration from such consolidated metrics, there's a growing demand for a similar, unified "Bankruptcy Score" to gauge a company's risk of insolvency.

PROBLEM DESCRIPTION

Your group is asked to conceptualize, design, and validate a comprehensive "Bankruptcy Score" – a single metric that accumulates a blend of financial, operational, and market-driven indicators to predict a company's risk of bankruptcy. This score should not only simplify the risk assessment process but also enhance its predictive accuracy.

GOALS

Develop your own Bankruptcy score method that presents key aspects such as * Consolidation of Metrics: Identify, evaluate, and consolidate a range of traditional and alternative data sources into a singular score. * Dynamic Adaptability: Ensure the score is adjustable based on industry, geography, and size of the company, allowing for more tailored risk assessments. * Transparency & Interpretability: Make the methodology behind the score transparent and easily interpretable for stakeholders. * Validation & Reliability: Backtest the score with historical bankruptcy cases to validate its predictive power and reliability.

SUGGESTED APPROACH

In order to achieve this the following activities is suggested * Technology exploration - Explore available data sources. - Review of the “US Company Bankruptcy Prediction Dataset” and references projects on how machine learning can be used to predict bankruptcy. * Project Specification - Define your own KPIs assessing various parameters of a company's health. * Method - System Architecture - Describe how data can be obtained, processed and aggregated, to keep your score up-to-date * Development of Minimum Viable Concept - How can this data be presented? (ex. Figma description) * Suggested limitations - The goal of this project could be proposing a method for predicting bankruptcy in a specific sector such as the financial sector such as larger financial institutions to regional banks in the US based on a knowledge transfer from this public available dataset.