Home/All Courses/Central Banking/Advanced Econometrics, Machine Learning & AI for Central Bank Research Training Workshop.
Central Banking · Physical, Online
Advanced Econometrics, Machine Learning & AI for Central Bank Research Training Workshop.
10 Days6 sessions available Certificate on completion
The Advanced Econometrics, Machine Learning & AI for Central Bank Research Training Workshop is an advanced professional programme designed to strengthen the analytical, quantitative research and forecasting capabilities of central bank economists, monetary policy researchers, financial stability analysts and macroeconomic modelling specialists.
The course integrates advanced econometric methods, time-series analysis, causal inference, machine learning, generative AI and emerging AI-based forecasting techniques to address practical central banking research challenges. Participants learn how to estimate and interpret macroeconomic relationships, forecast inflation and economic growth, evaluate monetary policy transmission, analyse exchange-rate movements, assess financial stability risks and develop reproducible empirical research for policy formulation.
Course objectives
Apply advanced econometric techniques to analyse macroeconomic relationships, monetary policy transmission, inflation dynamics, exchange-rate movements and financial stability.
Produce reproducible, statistically defensible and policy-relevant research outputs, including model documentation, forecast evaluations, empirical research papers and policy briefs.
Use AI and machine learning to improve predictive modelling, causal analysis, scenario assessment and the interpretation of economic and financial developments.
Integrate structured economic data, high-frequency indicators, alternative datasets and text-based information into empirical central bank research.
Develop, estimate, diagnose and validate econometric and machine-learning models for macroeconomic forecasting, nowcasting and policy analysis.
Who should attend
Central bank economists, monetary policy analysts, macroeconomic researchers, quantitative analysts, research department officers, economic modellers, inflation analysts, financial stability researchers, financial markets analysts, exchange-rate analysts, reserve management researchers, financial sector surveillance specialists, policy evaluation teams, central bank statisticians, econometricians, economic forecasting specialists, research consultants, postgraduate economists working on central banking applications.
Course content
Module 1: Advanced Econometric Foundations for Central Bank Research Review of probability and statistical inference, estimation theory and sampling distributions, maximum likelihood estimation, generalised method of moments (GMM), asymptotic properties
Module 2: Advanced Time-Series Econometrics and Macroeconomic Dynamics Stationarity and differencing, unit-root testing using ADF and Phillips–Perron procedures, KPSS testing, seasonal adjustment, trend-cycle decomposition, deterministic and stochastic
Module 3: VAR, SVAR, VECM and Monetary Policy Transmission Vector autoregression specification and lag selection, VAR stability and diagnostic testing, impulse response functions, forecast error variance decomposition, structural identification strategie
Module 4: Bayesian Econometrics, State-Space Models and Advanced Forecasting Bayesian inference and prior specification, posterior estimation, Bayesian VARs, shrinkage priors and high-dimensional macroeconomic systems, Gibbs sampling and Markov chain Mon
Module 5: Panel Econometrics and Cross-Country Macroeconomic Research Panel data structures and estimation, pooled OLS, fixed-effects and random-effects models, Hausman specification testing, dynamic panel models, instrumental variables, difference and s
Module 6: Causal Inference and Empirical Policy Evaluation Potential outcomes and treatment effects, causal diagrams and confounding, omitted-variable bias, instrumental variable estimation, difference-in-differences, event-study designs, regression disc
Module 7: Machine Learning Foundations for Economic and Financial Data Supervised and unsupervised learning, regression and classification, training, validation and test datasets, data leakage prevention, feature engineering, scaling and transformation,
Module 8: Machine Learning for Inflation, GDP and Macroeconomic Forecasting Macroeconomic forecasting design, benchmark model selection, direct and recursive forecasting, real-time and pseudo-out-of-sample evaluation, inflation and core inflation predict
Module 9: High-Frequency Data, Nowcasting and Alternative Data Analytics Mixed-frequency econometrics, MIDAS regression, bridge equations and dynamic factor nowcasting, ragged-edge datasets, data vintages and real-time databases, monthly and quarterly da
Module 10: Deep Learning, Natural Language Processing and AI for Economic Research Neural network foundations, multilayer perceptrons, recurrent neural networks, long short-term memory networks, temporal convolutional models, attention mechanisms, transf
Module 11: Financial Econometrics, Banking Data and Systemic Risk Analysis Financial return and volatility modelling, ARCH and GARCH extensions, stochastic volatility concepts, realised volatility, tail-risk measurement, Value at Risk and Expected Shortf
Module 12: AI-Assisted Research, Generative AI and Research Automation Generative AI applications in economic research, AI-assisted literature discovery and synthesis, research question refinement, hypothesis development, conceptual framework design, dra
Module 13: Explainable AI, Model Validation and Responsible Quantitative Research Model interpretability and explainability, feature importance, permutation importance, SHAP and partial dependence analysis, model calibration, residual and stability diagn
Module 14: Macroeconomic Scenarios, Policy Simulations and Decision Support Scenario design and counterfactual analysis, conditional forecasting, monetary policy reaction functions, Taylor-rule estimation, policy rate scenario analysis, inflation and out
Module 15: Capstone — Building an Integrated Econometric and AI Research Framework for a Central Bank Research question formulation, selection of a central bank policy problem, development of a theoretical and empirical framework, data acquisition and
Practical Exercises and Applied Research Projects Estimate an ARDL or VECM model to investigate the relationship between inflation, policy rates, exchange rates and money supply. Estimate a VAR or SVAR model and interpret impulse responses following a m
General notes
The programme combines established econometric techniques with modern predictive analytics, including ARIMA and SARIMA, vector autoregressions (VAR and SVAR), vector error-correction models (VECM), Bayesian VARs, panel econometrics, dynamic factor models, state-space models, machine learning, ensemble forecasting, neural networks and explainable AI. It also introduces AI-assisted analysis of economic reports, central bank communications, policy statements, research literature and alternative data.
Special attention is given to the challenges central banks face in emerging and developing economies, including short historical datasets, structural breaks, informal economic activity, data revisions, high-frequency data gaps, exchange-rate volatility, commodity price shocks and changing monetary transmission mechanisms.
The workshop emphasises model comparison, out-of-sample validation, economic interpretation, policy relevance and research reproducibility rather than treating AI as a replacement for economic theory or econometric judgement. This approach reflects ongoing research into machine learning for macroeconomic forecasting and central banking applications. <Link url="https://www.bis.org/publications/working-paper-930-big-data-and-machine-learning-central-banking" title="BIS research on big data and machine learning in central banking"/> and <Link url="https://www.imf.org/en/publications/wp/issues/2025/12/05/gdp-nowcasting-performance-of-traditional-econometric-models-vs-machine-learning-algorithms-572360" title="IMF research comparing econometric and machine-learning nowcasting models"/>.
Recommended Software, Tools and Research Platforms Econometric analysis: Stata, EViews, R, Python, MATLAB, SAS.
Advanced time-series and Bayesian modelling: R packages for time-series analysis and Bayesian estimation, Python statsmodels, PyMC, Stan, MATLAB econometrics tools, and suitable Stata or EViews procedures.
Machine learning and deep learning: Python, scikit-learn, XGBoost, LightGBM, TensorFlow, Keras, PyTorch.
Natural language processing and generative AI: Hugging Face Transformers, spaCy, approved enterprise AI assistants, document-processing tools, embedding models and retrieval-augmented generation frameworks.
Data engineering and reproducibility: Git, GitHub or approved internal Git repositories, MLflow, Jupyter, SQL, PostgreSQL, controlled research data repositories and automated analytical workflows
Visualisation and policy communication: Matplotlib, Seaborn, Plotly, Power BI, Tableau, LaTeX and reproducible report-generation tools.
The workshop can be adapted to a Stata- or EViews-centred workflow, an R-based research environment, or a Python-based econometrics and machine-learning workflow. Software choice should reflect institutional licensing, research standards, data governance and participant proficiency.
Expected Practical Outputs Participants will develop or practise producing an econometric research design, macroeconomic data dictionary, data quality assessment report, reproducible data-cleaning workflow, unit-root and cointegration test results, ARIMA or SARIMA forecast model, VAR or SVAR estimation report, impulse response analysis, VECM or ARDL specification, Bayesian forecasting model specification, panel econometric estimation, causal inference research design, inflation forecasting model, GDP nowcasting framework, machine-learning benchmark comparison, time-aware validation report, alternative-data indicator specification, central bank text sentiment indicator, financial stress or vulnerability model, explainable AI report, model validation checklist, forecast fan chart, policy scenario analysis, research tables and figures, empirical working paper, policy brief and reproducible analytical code or research notebook.
Econometrics Training for Central Bank Economists, Machine Learning for Macroeconomic Forecasting, AI for Central Bank Research, Advanced Time-Series Econometrics, Monetary Policy Modelling Training, Inflation Forecasting Training, GDP Nowcasting with Machine Learning, Bayesian Econometrics Training, VAR and SVAR Modelling, Causal Inference for Economic Policy, Financial Econometrics Training, AI for Economic Research, Macroeconomic Forecasting Workshop.
Frequently asked questions
The programme is designed for central bank economists, monetary policy researchers, quantitative analysts, financial stability specialists, macroeconomic modellers, economic statisticians and research professionals who need advanced empirical methods for monetary policy, forecasting and financial sector analysis.
The workshop covers advanced time-series econometrics, ARIMA and SARIMA, ARDL, VAR and SVAR, VECM, Bayesian VARs, dynamic factor models, state-space models, panel econometrics, GARCH models, local projections and causal inference techniques. The emphasis can be adjusted to the research priorities of the participating institution.
Econometrics provides structured approaches to estimating relationships, testing hypotheses and evaluating economic mechanisms. Machine learning can identify complex predictive patterns and improve forecasting in suitable data environments. Participants compare both approaches using appropriate benchmarks, out-of-sample validation and economic interpretation rather than assuming that one method always outperforms the other.
Yes. Participants explore natural language processing, central bank communication analysis, economic document extraction, AI-assisted literature review, research automation, explainable AI and the use of generative AI in research workflows. The course also addresses verification, confidentiality and responsible use of AI.
Some programming or statistical software experience is strongly recommended because the workshop involves model estimation, diagnostic testing, machine learning and applied research. Exercises can be tailored to Stata, EViews, R or Python, but the advanced modelling components require participants to engage directly with analytical software.
Yes. The programme can be tailored to monetary policy analysis, inflation forecasting, exchange-rate dynamics, macroeconomic nowcasting, financial stability, credit and liquidity risk, reserve management research, financial inclusion, climate-related financial risks or country-specific macroeconomic modelling.