UEH Standard Programs

Time Series Econometrics

Tiếng Việt

M00264

3

Môn học này cung cấp các kỹ thuật phân tích dữ liệu chuỗi thời gian hết sức cần thiết để thực hiện nghiên cứu khoa học và khóa luật tốt nghiệp. Sinh viên tiếp cận kiến thức của môn học không chỉ từ góc độ kiến thức nền tảng của các công thức toán học cần thiết nhất, từ các tình huống gần gũi trong thực tế, mà còn dưới dạng ‘ứng dụng’, ‘thực hành’ trên phần mềm Eviews (chính), và Stata (phụ) về hầu hết các mô hình kinh tế lượng chuỗi thời gian đơn biến và đa biến thông dụng hiện nay./ This course provides necessary techniques of analysing time-series data to doing research and thesis. Students reachs the knowledge not only from mathematical equations but also from the real case study in paractising in Eviews (mainly) and Stata (secondly) about most of current time series models.

This course is divided into two main components: (1) collection, management, and analysis of time series data; and (2) construction of econometric time series models for forecasting or testing economic hypotheses.

The collection and management of time series data has been briefly introduced in the Macroeconomics course (regarding macroeconomic indicators, balance of payments, monetary systems, and government budgets) and the Statistics in Economics and Business course (regarding the calculation of statistical indices such as CPI, PPI). Therefore, in this course, we will only introduce accessible data sources for research and guide students on how to manage time series data using Stata or R.

The analysis of time series data will focus on the following content: smoothing data, creating trend variables, creating seasonal dummy variables, and structural dummy variables; using time series graphs to detect movement trends, relationships between variables, and identifying structural breaks or thresholds in the data; understanding the significance of autocorrelation coefficients and correlograms; exploring the properties of stationary series, non-stationary series, random walks, differenced stationary series, trended stationary series, autoregressive (AR) series, and moving average (MA) series; conducting unit root tests and cointegration tests.

Regarding the construction of econometric time series models, we will divide them into univariate and multivariate models. The univariate model group is divided into two subgroups: simple and advanced. The simple subgroup will focus on moving average models, Holt exponential smoothing, Winter smoothing, Holt-Winter smoothing, and ARIMA models to forecast individual indices such as sales, inventory, commodity prices, and some macroeconomic indicators like GDP, CPI, interest rates, and unemployment. The advanced subgroup includes ARCH, ARCH-M, GARCH, GARCH-M, and TGARCH models to forecast highly volatile time series such as oil prices, gold prices, exchange rates, and stock prices. The multivariate model group mainly focuses on the VAR model, error correction mechanism (ECM) models or vector ECM (VECM) models, and Granger causality tests to forecast the relationships between economic indicators and test economic hypotheses. Recently, many applied studies (especially in the fields of macroeconomics, financial development, and energy) have used non-stationary panel data, so we will introduce some unit root tests and cointegration tests for this type of data. Additionally, threshold regression models have been and are being widely used, so we will also provide an overview to give students a basic foundation for self-study later on.