Estimating Policy Functions in Payment Systems using Reinforcement Learning

40 Pages Posted: 12 Feb 2021

See all articles by Pablo S Castro

Pablo S Castro

Google Inc.

Ajit Desai

Bank of Canada

Han Du

Government of Canada - Bank of Canada

Rodney Garratt

University of California, Santa Barbara (UCSB)

Francisco Rivadeneyra

Bank of Canada

Date Written: December 4, 2020

Abstract

This paper uses reinforcement learning (RL) to approximate the policy rules of banks participating in a high-value payments system. The objective of the agents is to learn a policy function for the choice of amount of liquidity provided to the system at the beginning of the day. Individual choices have complex strategic effects precluding a closed form solution of the optimal policy, except in simple cases. We show that in a simplified two-agent setting, agents using reinforcement learning do learn the optimal policy that minimizes the cost of processing their individual payments. We also show that in more complex settings, both agents learn to reduce their liquidity costs. Our results show the applicability of RL to estimate best-response functions in real-world strategic games.

Keywords: Artificial intelligence, Reinforcement learning, Payments systems

JEL Classification: A12, C7, D83, E42, E58

Suggested Citation

Castro, Pablo S and Desai, Ajit and Du, Han and Garratt, Rodney and Rivadeneyra, Francisco, Estimating Policy Functions in Payment Systems using Reinforcement Learning (December 4, 2020). Available at SSRN: https://ssrn.com/abstract=3743017 or http://dx.doi.org/10.2139/ssrn.3743017

Pablo S Castro

Google Inc. ( email )

1600 Amphitheatre Parkway
Second Floor
Mountain View, CA 94043
United States

Ajit Desai

Bank of Canada ( email )

234 Wellington Street
Ontario, Ottawa K1A 0G9
Canada

Han Du

Government of Canada - Bank of Canada ( email )

234 Wellington Street
Ontario, Ottawa K1A 0G9
Canada

Rodney Garratt

University of California, Santa Barbara (UCSB) ( email )

Francisco Rivadeneyra (Contact Author)

Bank of Canada ( email )

234 Wellington Street
Ontario, Ottawa K1A 0G9
Canada

HOME PAGE: http://sites.google.com/site/rivadeneyr/research

Do you have a job opening that you would like to promote on SSRN?

Paper statistics

Downloads
238
Abstract Views
1,035
rank
158,090
PlumX Metrics