Model Risk: The perils of quantification

Estimation and hence quantification have been an integral part of scientific, economic, and social developments since humans learned to count. Precision and accuracy are at the heart of the motivations that drive such developments. However, quantification and modelling, like any other tools of civilizations, come with a false sense of accuracy and risk. Models, often originally developed to understand risk better, give rise to a risk specifically due to the usage of models, namely model risk. This risk, while always present, has only been in the forefront of discussion over the last couple of decades. In this discussion, we would like to understand model risk as a primary risk type, talk about some historical anecdotes, and look at approaches to manage it.
 
About the Speaker:
 
Anirban Dutta is a Director and the Head of Model Risk Governance and Framework for DWS, the asset management arm of Deutsche Bank. He is based in Pune, India. Before DWS, Anirban was a Director in the Model Risk Management function of Credit Suisse for over six years. He led the validation book of work of several modeling disciplines including economic capital, stress testing, liquidity, CFO, funding and valuations in resolution, and recovery and resolution planning. Most recently he was responsible for the model risk component of firmwide large regulatory stress testing programs like Loss Potential Analysis (LPA) and Internal Capital Adequacy Assessment Process (ICAAP) and overlay governance. 
 
Prior to joining Credit Suisse, he was responsible for the end-to-end product development and delivery of two financial technology start-ups in the areas of hedge fund investor services and smart data ingress and reporting engine. Anirban has a Ph.D. in mathematics from Western Michigan University, an M.A. in mathematics from Stony Brook University, an M.Stat. with specialization in probability from Indian Statistical Institute, and a B.Stat. (Hons.) from Indian Statistical institute.