Why Banks are Shifting to Open-Source Model Software for Financial Risk Management

Banks perform data analytics, statistical modelling, and automate financial processes using model software, making model software essential for financial risk management. 

Why banks are recently moving their model software to open-source¹: 

• Volume dependent costs: Open-source comes with flexible storage space and computation power and corresponding costs and capacity.  
• Popularity: Graduates are often trained in open-source software, whereas the pool of skilled professionals in some specific software is decreasing. 
• Customization: You can tailor it exactly to your own needs. Nothing more, nothing less. 
• Collaboration: Required level of version control and collaboration with other software available within the bank. 
• Performance: Customization and flexibility allow for enhanced performance. 

Often there is a need to reconsider the model software solution the Bank is using; open-source solutions are cost effective and can be set up in such a way that both internal and external (e.g. regulatory) requirements are satisfied. 

How Zanders can support in moving to open-source model software. 

Proper implementation of open-source model software is compliant with regulation.   

• Replication of historical situations is possible via version control, containerization and release notes within the open-source model software. Allowing third parties to replicate historical model outcomes and model development steps.   
• Governance setup via open-source ensures that everything is auditable and is released in a governed way; this can be done through releases pipelines and by setting roles and responsibilities for software users.  
• This ensures that the model goes from development to production only after governed review and approval by the correct stakeholders.  

Migration to open-source software is achievable. 

• The open-source model software is first implemented including governance (release processes and roles and responsibilities).  
• Then all current (data and) models are refactored from the current model software to the new one.  
• Only after an extensive period of successful (shadow) testing the migration to the new model software is completed.  

Zanders has a wealth of experience performing these exercises at several major banks, often in parallel to (or combined with) model change projects. Hence, if you have interest, we will be happy to share some further insights into our latest experiences/solutions in this specific area.  

Therefore, we welcome you to reach out to Ward Broeders (Senior Manager).