The REGTECH Book

Merits and Demerits of a Shared Risk Engine

By Rohit Ghosh¹ and Harshwardhan Prasad²

¹AI Scientist and Founder, Skyra Capital

²VP, Quant Analytics, Morgan Stanley

The idea of a shared risk engine (SRE) as a concept is to build a cloud-based, secure risk engine as a shared service that simulates all the major risk-related functionalities at significantly lower cost.

An SRE would be a common platform shared across multiple organizations albeit in a safe and secure way without compromising on any data privacy. SREs can replicate all major risk engine functionalities like value at risk (VaR) calculation, risk reporting, stress testing, and so forth on a cloud instance that is shared across financial organizations without any of them needing to expose their internal data and model publicly. The main advantage of such a model is extreme cost reduction for any organization, for two main reasons. Firstly, major cost operations and functions of model building, data storage and quality, model validation, and even risk report building would be shared across all participants. Secondly, the cloud-based architecture would mean pay per usage, customized to model complexity and data requirements of individual banks. In this chapter, we demonstrate how we can achieve this target, and offer a brief overview of the architecture involved.

Computationally, in the first step, the central module (I₀) in SRE produces organization-specific risk application programming interfaces (ORAPIs) – a set of ...