The National Bank of Georgia publishes the next paper in its NBG Working Paper Series. The working papers aim to present research on topics relevant to central banks and economics profession. The papers are circulated to elicit comments and encourage debate on ongoing research. The Working Paper Series are coordinated by the Macroeconomic Research Division and focus on topics relevant for central bank policy, including macroeconomic modeling, monetary and exchange rate policies, financial stability, etc.

 

The current working paper in the series titled "Application of Artificial Intelligence for Monetary Policy-Making" was prepared by Mariam Dundua, Leading Specialist in the Financial Technologies and Model Risk Division at NBG, and Otar Gorgodze, who worked on the project while he was head of Financial and Supervisory Technologies Department at the NBG, currently working as a Resident Advisor on Banking Supervision at the International Monetary Fund. The aim of this paper is to enrich the analytical toolkit of the monetary policy framework in Georgia using Artificial Intelligence. The study shows a new approach that searches for optimal monetary policy rules using Reinforcement Learning (RL). RL comes closest to human decision making and it seems to be particularly promising for application to complex economic problems. The analysis of AI generated monetary policy rules indicates that optimal policy rules exhibit significant nonlinearities. This could explain the practical shortcoming of traditional modelling tools for monetary policy. The generated transitions equations analysis also allows us to estimate the neutral policy rate, which came out to be 6.5 percent. This approach could be valuable for current monetary policy analysis in central banks. Future research and development of the approach could result in semi-autonomous policy conduct.   

 

The working papers can be accessed at the following page.