# Econometrics & Forecasting

GER/ECF

## Who should attend?

• To be able to use econometric tools in order to determine correlations and adjustments between physical or economical series and to make forecasts and simulations.
Audience :
• Engineers, economists and financiers from all sectors.

Level : Proficiency

## Course Content

• ### STATISTIC BASIS

• Descriptive statistics (mean, median, standard-deviation, etc.), indices (Laspeyres, Paasche, Divisia).
• Statistical tests (normality, student, Fisher).
• Application: energy data set (quantities, prices).
• ### LINEAR REGRESSION MODELS & FORECASTING

• Simple and multiple linear regression models, ordinary least square estimator, R2.
• Application: energy demand model.
• Statistical tests validating econometric models: autocorrelation (Durbin-Watson, Lagrangian multiplier), heteroscedasticity (White and Breusch-Pagan), multicollinearity (BKW).
• Structural change on linear regression model (Chow test, Brown-Durbin & Evans test).
• Application: analysis of the substitution between oil, gas and electricity.
• Principle of forecasting with an econometric model (properties of the estimator, prediction interval).
• Application: forecasts on energy demand model.
• ### TIME SERIES ANALYSIS & FORECASTING

• Time series model.
• Smoothing techniques for short run forecasts: extrapolation techniques (moving average, time series decomposition with trend and seasonal pattern).
• Application: monthly energy demand series (with a seasonal pattern), forecast over 12 month.
• ARIMA models (AutoRegressive Integrated Moving Average), tests assessing the stochastic processes (number of autoregressive and moving average lags, stationarity).
• Application: ARIMA model simulations.
• ### TIME SERIES RELATIONSHIP: CO-INTEGRATION & CAUSALITY

• Introduction to co-integration techniques: unit root tests (Dickey-Fuller, Phillips-Perron, KPSS), Engle and Granger model, long-term equilibrium, Error Correction Model (ECM).
• Causality test.
• Application: co-integration techniques to Oil & Gas markets.
• Co-integration with multiple relationship: Johanson test (max. eigenvalue and Trace test) on a VAR (Vectorial AutoRegressive) model.
• Application: modeling the equilibrium between prices over several market places.
• Structural changes on co-integration model: long term and short term dynamic (Perron test, Gregory and Hansen test).
• ### CHANGES OF VOLATILITY ON ENERGY MARKET

• ARCH model (AutoRegressive Conditional Heteroscedastic) and generalization.
• Application: modeling volatility changes in the short term dynamic and on the equilibrium of Oil & Gas markets.

## Learning Objectives

• Upon completion of the course, the participants will be able to:
• use the main econometric techniques,
• perform an econometric estimation,
• develop models and make forecasts, in particular in the energy sector and on financial markets.

## Ways & Means

• Applications performed on computer (statistical tests, development of econometric models, forecasting, simulation, highlighting co-integration and causality relationship, etc.) using Excel and Eviews.