Alessio Brini, Postdoctoral Associate at the Pratt School of Engineering of Duke University, was our guest at the D2 Seminar Series. He presented his talk on “Reinforcement Learning Policy Recommendation for Interbank Network Stability”
He presented his paper, a joint work with Gabriele Tedeschi and Daniele Tantari where they analyzed the effect of a policy recommendation on the performance of an artificial interbank market.
As stated in the abstract, financial institutions stipulate lending agreements following a public recommendation and their individual information. The former is modeled by a reinforcement learning optimal policy that maximizes the system’s fitness and gathers information on the economic environment. The policy recommendation directs economic actors to create credit relationships through the optimal choice between a low interest rate or a high liquidity supply. The latter, based on the agents’ balance sheet, allows to determine the liquidity supply and interest rate that the banks optimally offer their clients within the market. Thanks to the combination between the public and the private signal, financial institutions create or cut their credit connections over time via a preferential attachment evolving procedure able to generate a dynamic network. The results of their research shows that the emergence of a core-periphery interbank network, combined with a certain level of homogeneity in the size of lenders and borrowers, is essential to ensure the system’s resilience. Moreover, the optimal policy recommendation obtained through reinforcement learning is crucial in mitigating systemic risk.
What is a reinforcement learning policy and why is it important for the interbank market?
A reinforcement learning policy is an algorithm used in artificial intelligence that involves learning through trial and error. In the context of the interbank market, such a policy is used to model the public recommendation that directs economic actors to create credit relationships. This policy is important because it helps to maximize the system’s fitness and gather information on the economic environment, which can help to ensure the resilience of the interbank network. Additionally, reinforcement learning can help mitigate systemic risk by allowing financial institutions to make more informed and optimal choices when creating or cutting their credit connections.
What are the main questions you analyzed in your research?
The main research question analyzed in the introduction is the effect of an unconventional, environmentally dependent policy recommendation on the stability of the interbank system. The researchers use a reinforcement learning approach to model the policy recommendation and study how it impacts the spread of systemic risk in the interbank market. The originality of the research is the use of reinforcement learning to endogenize and identify the optimal strategy for banks in the interbank market and to model a policy recommendation that can help to tame systemic risk.
What is the major challenge in addressing such a kind of study?
The major challenge regarding this type of research I adapting the reinforcement learning framework to study an interbank network problem. Although reinforcement learning is common in many domains, it is not so in economics. The biggest challenge was to design a reward function reflective of the aim we wanted to pursue, such as the systemic stability of the system. Another challenging aspect we have faced is interpreting the learned policy regarding its input variables. This is generally considered an important aspect of the field of economics and finance because even if you have an algorithm that works well, the research needs to be able to explain the way it works.
Given the practical applications of this field, would you like to give us a glimps of future outcomes of your research?
Some future aspects revolve around using reinforcement learning for the agent-based model. We would like to extend the application of these algorithms to the more complex stylized model of the interbank market, which could include the role of the households and the central banks.
More information about his research and publications can be found on his personal page.
You can access the video recording of his seminar here (registration needed).