We demonstrate the ability of reinforcement learning techniques to estimate the best-response functions of banks participating in high-value payments systems—a real-world strategic game of incomplete information.
We use retail payment data in conjunction with machine learning techniques to predict the effects of COVID-19 on the Canadian economy in near-real time. We find this approach yields a significant increase in forecasting precision over a linear benchmark model. This model can help policy-makers before official data are released.
