BBVA takes new approach to simulating risk thanks to collaboration with fintech startup Simudyne
Institutions such as the Bank of England, the European Central Bank, and the US Treasury are constantly looking to avert another global financial crisis, and in doing so have increasingly turned to simulation as a means of anticipating future market activity – and, importantly, ensuring they are suitably equipped to survive and thrive in any scenario.
“Computer simulations have long been used to train humans, enabling them to formulate ready solutions to various situations. But simulations can benefit business too,” says Justin Lyon, CEO and founder of Simudyne. The UK-based startup was recently matched with BBVA’s Global Markets business unit through the Open Marketplace platform, completing a proof-of-concept project as part of the bank’s commitment to engaging innovators from the wider fintech ecosystem to help it develop new solutions to commercial challenges.
Lyon adds that his company “builds simulations that capture the real world. Banks are in possession of a lot of data. By exploring the thousands of permutations that data offers to see how a crisis could unfold, these simulations allow CEOs and risk management officers to explore all possible futures, offering them a playbook of decisions they can use.”
Innovative opportunity
A market leader in agent-based financial simulation – an intuitive computer model designed to capture the behaviour of individuals within a specific environment – Simudyne focuses on helping large global financial institutions better understand the risk they are exposed to.
BBVA was initially introduced to Simudyne through a fintech accelerator programme, and recognised agent-based simulation as an opportunity to optimise its own algorithmic strategies. The bank invited the startup to collaborate on a project to calibrate its Global Markets trading algorithms and improve their performance through testing in a realistic simulated exchange.
Explaining the scope of the project, Conrado García, Advanced Analytics & Algorithmic Trading specialist in the CIB department at BBVA, said: “From hedge funds on the buy side, to banks on the sell side, via traders and retail clients, there are a number of different agents involved in a real market. We used Simudyne’s modelling framework to simulate these agents and the interactions that occur between them.”
The benefits soon became clear. Training data-driven trading algorithms in a reactive environment with an unbounded – as opposed to a historical – dataset was seen to improve their performance, with the use of hypothetical stress scenarios particularly beneficial. What’s more, this approach ensured compliance with new regulations such as MiFiD 2, which emphasise the importance of testing algorithms that will, over time, interact independently with financial exchanges.
The project ran throughout Q1 2019, during which time BBVA was able to gain a greater understanding of where in the business Simudyne’s technology might be most productively applied. BBVA is now exploring if and how the tool could be more broadly used within the group.
Meanwhile – in a development that is characteristic of BBVA’s desire to create mutual benefit throughout its engagement and co-innovation with startups within the fintech ecosystem – Simudyne itself is applying data from the project and feedback from BBVA to improve its market simulation offering.
