EC3304: Econometrics II
AY2014/2015, Semester 2, Lecturer: Eric Fesselmeyer
Course Coverage:
1. Time Series Regression
2. Estimation of Dynamic Causal Effect
3. Cointegration & Error Correction Model
4. Panel Data Regression
5. Binary Dependent Variable
6. Instrumental Variable
7. Experiments & Quasi-Experiments
This module is the second course in econometrics, and it builds heavily on the foundations laid by the first course. This course focus on the theoretical development and applications of various econometric tools that are widely used in economic research.
Times series regression is perhaps the most common tool used in macroeconomic research. This topic studies the regression of current value against past values of the same variable, and the results such regression produces. Some important concepts include autocovariances, autocorrelations, stationarity and ergodicity. The most emphasized property throughout the course is perhaps stationarity and the concept of unit roots.
Dynamic causal effects, on the other hand, regresses a dependent variable on exogenous regressors. If the mean-zero shocks are serially correlated, the usual OLS standard errors are wrong. This chapter introduces the heteroskedacity and autocorrelation consistent (HAC) standard errors, to overcome to issue. Cointegration studies the long-rin relationship between two integrated (non-stationary) time series and this relationship can be utilized to developed a refined dynamic model which can have a focus on long-run or short-run transitory effects (error-correction).
Panel data regressions focus on fixed effect regression. The two types of effects are state-fixed and time-fixed effects. The former is constant through time and the latter is constant across cross-sectional data. The next two sections examine binary dependent variables, usually a yes-no answer to the motivating question, and instrumental variables which are used in place of the original non-exogenous regressor.
The last topic on experiments and quasi-experiments were not covered due to time constraint. This chapter studies actual experiments and quasi-experiments, which are rare in economics but very influential because they avoid endogeneity problems of observational studies.
This is a rather difficult core module, and a lot of time and effort are required to understand the materials.
Workload: Moderate
Difficulty: Difficult
Grade: A+
Course Coverage:
1. Time Series Regression
2. Estimation of Dynamic Causal Effect
3. Cointegration & Error Correction Model
4. Panel Data Regression
5. Binary Dependent Variable
6. Instrumental Variable
7. Experiments & Quasi-Experiments
This module is the second course in econometrics, and it builds heavily on the foundations laid by the first course. This course focus on the theoretical development and applications of various econometric tools that are widely used in economic research.
Times series regression is perhaps the most common tool used in macroeconomic research. This topic studies the regression of current value against past values of the same variable, and the results such regression produces. Some important concepts include autocovariances, autocorrelations, stationarity and ergodicity. The most emphasized property throughout the course is perhaps stationarity and the concept of unit roots.
Dynamic causal effects, on the other hand, regresses a dependent variable on exogenous regressors. If the mean-zero shocks are serially correlated, the usual OLS standard errors are wrong. This chapter introduces the heteroskedacity and autocorrelation consistent (HAC) standard errors, to overcome to issue. Cointegration studies the long-rin relationship between two integrated (non-stationary) time series and this relationship can be utilized to developed a refined dynamic model which can have a focus on long-run or short-run transitory effects (error-correction).
Panel data regressions focus on fixed effect regression. The two types of effects are state-fixed and time-fixed effects. The former is constant through time and the latter is constant across cross-sectional data. The next two sections examine binary dependent variables, usually a yes-no answer to the motivating question, and instrumental variables which are used in place of the original non-exogenous regressor.
The last topic on experiments and quasi-experiments were not covered due to time constraint. This chapter studies actual experiments and quasi-experiments, which are rare in economics but very influential because they avoid endogeneity problems of observational studies.
This is a rather difficult core module, and a lot of time and effort are required to understand the materials.
Workload: Moderate
Difficulty: Difficult
Grade: A+