Dynamically Setting the Price when Lending out Securities using Reinforcement Learning

What is it about?

The Securities Lending market has always been complex and ever-changing, hence, it's very hard to devise one type of price modelling method that fits all scenarios to maximize revenue. With our work on leveraging Reinforcement Learning (Contextual Bandit) to dynamically set the price depending on context of the market, we are able to strike a good balance in between a few available pricing methods and pick the one that gives the highest estimated reward (i.e., revenue).

Featured Image

Photo by Nicholas Cappello on Unsplash

Photo by Nicholas Cappello on Unsplash

Why is it important?

Combination of Reinforcement Learning (Contextual Bandit) and Securities Lending is relatively new. Hence, this piece of work sheds some lights on a new field of research and application. Business-wise, not only does this method gives an estimated higher revenue, it also gives some insights on which market supply-demand signals are more impactful when conducting the business.