Introduction

In this case study, we will delve into the project overview of Bank of America’s Global Corporate & Investment Banking (GCIB) division. The objective was to develop and prototype a portfolio management prediction tool that would assist the investment banking teams in modeling portfolio scenarios for their customers. This article will provide an in-depth analysis of the project, the product, challenges faced, and the ultimate solutions that were implemented.

Project Overview

The GCIB division of Bank of America comprises Corporate Banking, Capital Markets, Transaction Services, and Leasing. The project aimed to create a cutting-edge application for mining stock portfolios, leveraging multivariate dependence risk and portfolio optimization. The primary focus was on modeling changing dependence risk across four different period scenarios and optimizing portfolios with complex patterns of dependence.

Challenges Faced

One of the major hurdles encountered during this project was the clustering of financial time series in risky scenarios. It proved challenging to demonstrate to customers how different variables impact portfolio optimization over time. The existing tools utilized by the bank’s teams had significant performance issues and lacked clear visualizations to showcase variance effectively.

Objectives

The main objective was to develop an application capable of utilizing a customer’s current portfolio to run various scenarios based on risk tolerance. The tool would facilitate discussions about different possible scenarios with the customers.

My Role and Responsibilities

As a member of the prototype team, my responsibilities included selecting the prototype architecture and making decisions regarding the technology stack. The prototype was deployed on in-house servers, and ensuring its basic performance was also within my purview.

User Summary and Pain Points

The field service teams necessitated a tool that would assist them in refining a customer’s portfolio goals. The process involved swiftly selecting different multivariate scenarios and adjusting the customer’s goals. All of this needed to be accomplished using a tablet in the field. However, the available tools were not well-integrated, making configuration for smooth presentations cumbersome. Additionally, performance issues frequently arose, further delaying the customer’s preferred scenarios.

Proposed Solution

To address the pain points, the project team comprised a mix of front-end and back-end designers and developers. Each team member played a flexible role, with the project manager serving as the main point of contact. The technology stack for the prototype was chosen for its ability to facilitate rapid development. Microsoft-based technologies, including Progressive Web Application front-end, Tabular charting and visualizations, C#, .NET Core, and SQL Server, formed the core components of the solution.

Impact and Lessons Learned

The project team’s openness and collaborative spirit fostered innovative thinking throughout the development process. Despite being a relatively small team, we all gained valuable insights from one another. The most crucial lesson learned was the importance of maintaining focus on the Minimum Viable Product (MVP). Great ideas had to be thoroughly evaluated, communicated, and discussed by the team. The overall experience felt more akin to that of a startup rather than a major global bank.

Conclusion

Bank of America’s GCIB division successfully developed and prototyped a portfolio management prediction tool, enabling investment banking teams to model portfolio scenarios for their customers. By addressing the challenges faced and implementing the proposed solution, the bank enhanced its ability to demonstrate the impact of different variables on portfolio optimization. The project’s impact, combined with the lessons learned, transformed the experience into a dynamic and innovative endeavor.