Пырлик Владимир Николаевич
- supposed to provide the students with a set of tools that are useful for both theoretical and empirical modeling of dynamic economic data coming in the form of both univariate and multivariate time series
- content covers (but not limited to) an overview of the crucial theoretical results of contemporary time series econometrics and of the approaches towards empirical application of these results to empirical data and tasks, including estimation of dynamic economic models and practical forecasting.
- students will be able to statistically describe and analyze various dynamic economic data coming in the form of time series
- construct and analyze models of the corresponding economic processes
- construct relevant predictions of the data
- Inroduction to the courseThe difference between time series and random samples. The difference between time series econometrics and cross-sectional econometrics. Opportunities and goals of time series econometrics. Examples of time series in economic life.
- StationarityStrong and weak stationarity. Examples of non-stationary time series. Typical types of non-stationarity in economic time series. Trends, types of trends. Structural shifts. Random walk. Non-stationarity in the variance. Formal tests for stationarity. Extended Dickey-Fuller criterion. The Kwiatkowski-Philips-Schmidt-Shin test. Stationary transformations. Difference transformations. Growth and growth rates. Log-differences. Interpretation of standard increments. The relationship between log differences and growth rates.
- Linear regression for stationary and ergodic time seriesErgodicity of the time series. The ergodic theorem. Centarl limit theolrema for dependent observations. Long-term desperation. Newey-West score. The method of least squares in linear regression for stationary time series. Model hypotheses and properties of OLS estimates. Popraka on autocoerrelation of residues. The problem of endogeneity. The distributed lag model and its variants. Testing complex hypotheses. Estimation of the commulative effect of a distributed lag model.
- Forecasting a single time series.Conditional mathematical expectation and the best in the mean-square sense of the forecast. Forecasting from linear models. Static and dynamic forecasts. Reliability of the forecast. Predictive quality metrics. Sarvnenie predictive strength of models. Validation and forecasting schemes. Alternative loss functions.
- Stationary linear regressionAutoregression processes. Moving average processes. Autoregression-moving average processes. Stationarity of processes. Random walk. Single roots. Characteristics of stationary linear processes. Wold's theorem. Box-Jenkins modeling and forecasting methodology.
- Structural breaks and stability
- Bell, W. R., Holan, S. H., & McElroy, T. (2012). Economic Time Series : Modeling and Seasonality. Boca Raton, FL: Chapman and Hall/CRC. Retrieved from http://search.ebscohost.com/login.aspx?direct=true&site=eds-live&db=edsebk&AN=445858
- Tsay, R. S. (2010). Analysis of Financial Time Series (Vol. 3rd ed). Hoboken, N.J.: Wiley. Retrieved from http://search.ebscohost.com/login.aspx?direct=true&site=eds-live&db=edsebk&AN=334288