Belarusian Economic Research and Outreach Center

Home > News > The second module of BEROC Post Graduate School in Economics

The second module of BEROC Post Graduate School in Economics

The second module of Post Graduate School of Economics “Applied Time Series Econometrics for Non-Stationary Data” was held on January 24-28, 2012 at Belarusian State University, Minsk, K. Marks str. 31 (Economic Department)

Instructor: Igor Pelipas, Belarusian Economic Research and Outreach Center (BEROC), e-mail: pelipas@research.by

Course Overview:

The course dealt with applied econometric modelling, estimation, and testing of relationships for non-stationary time series. The following issues were discussed: introduction to stationarity and non-stationarity time series, deterministic and stochastic trends, integrated and cointegrated variables, unit roots, structural breaks, ARDL and VAR models, short-run and long-run relationships, equilibrium correction models, concept of exogeneity and its application for modelling, policy analysis and forecasting. The special attention was paid to the empirical model discovery using general to specific approach (Gets) and utilization of impulse indicator saturation technique (IIS) for handling outliers and multiply structural breaks during model selection.
The main objectives of the course were as follows:
- to provide students with an understanding of the main concepts of time series econometrics for non-stationary variables and the basic principles of applied time series analysis;
- to develop practical skills in applied times series analysis for non-stationary data, using relevant econometric software;
- to develop skills to understand and critically analyze the econometric results presented in the literature, and to interpret and adequately present the results of applied econometric research.
The course was accompanied by empirical modeling examples and exercises using econometric software OxMetrics.

Course Outline:

Day 1:

Introduction to time series econometrics: basic concepts. Stationary and non-stationary time series. Spurious regression. Cointegration. Long-run and short-run models. Introduction to OxMetrics: empirical example. General to specific approach and impulse indicator saturation.

Day 2:

Unit root testing. The Dickey-Fuller and Augmented Dickey-Fuller tests. Power and level of unit root tests. Structural breaks and unit root testing. Brief discussion of the various unit root tests.

Day 3:

Autoregressive distributed lag model (ARDL). Typology of liner dynamic models and testing of restrictions. Misspecification analysis. Cointegration and equilibrium correction models in a single equation framework. Two step Engle-Granger approach. Conditional equilibrium correction model. Concept of exogeneity. Brief discussion of the other approaches.

Day 4:

Cointegrated VAR. The Johansen-Juselius methodology. Formulating of the dynamic model. Deterministic components in the multivariate model. Testing for the reduced rank. Testing of weak exogeneity and exclusion of the variables. Testing for linear hypothesis on cointegration relations. Testing joint hypothesis on cointegration relations and feedback coefficients. Interpretation of the results.

Day 5:

Modelling the short-run multivariate system. Conditional and parsimonious VECM. Using VECM for forecasting. Structural breaks and forecasts failure. Robust forecasting devices. Concluding remarks.

Should you have any questions regarding the School do not hesitate contacting us at akelyeva@beroc.by

top