Research

PhD: Strategic optimisation of a commercial bank’s balance sheet under varying interest rate regimes Link to heading

Institution: University of Malta, Department of Communications and Computer Engineering Started: 2025 (M.Phil./Ph.D. track)

This research develops a comprehensive model for optimising commercial bank operations, integrating classical optimisation techniques with reinforcement learning (RL), and incorporating the effect of an exogenous bank rate on a bank’s strategic decisions. It builds a simulation environment from synthetic balance sheet data, trains both on-policy (e.g. PPO, A3C) and off-policy (e.g. DQN, DDPG) RL agents to optimise banking decisions, and benchmarks them against classical optimisation methods on metrics such as profitability, risk-adjusted returns, and regulatory compliance.

Research questions:

  1. How can reinforcement learning methods be effectively applied to model and optimise the operations of a commercial bank?
  2. Can these RL methods be made as sample-efficient as possible, reducing the number of interactions with the environment?
  3. Is a simulation environment derived from synthetic balance sheet datasets sufficient to train high-performing RL agents?
  4. What are the comparative advantages and limitations of RL-based optimisation versus classical optimisation approaches in banking?
  5. How does including an exogenous bank rate influence optimisation outcomes and decision-making within commercial banks?
  6. Will the trained RL model generalise across different types of commercial banks?

A full working paper / preprint will be linked here once available.

🔧 See the models I’m developing →

Background: Economics Link to heading

Over a decade of applied macroeconomic and econometric research across the Central Bank of Malta, the European Central Bank, and Bank of Valletta. See Experience for the full career history and Publications for research output.

Areas of expertise:

  • Macroeconomic forecasting and nowcasting (GDP, inflation, business conditions)
  • Labour market modelling (BVARs, labour market flows)
  • Housing and rental market economics, hedonic price indices
  • Applied econometrics and computational economics (Python, Stata, EViews, R, MATLAB)
  • Tourism and balance of payments analysis
  • Bank balance sheet optimisation, reinforcement learning for finance