The Analysis of Non-Stationary Pooled Time-Series Cross-Section-Data. In: International Journal of Conflict and Violence (IJCV). Jg.8 H. 2. 28.5.2015, S. 222-242
Inhalt
- Abstract
- 1. Properties of OLS and Fixed-Effect Regressions with Non-Stationary Panel Data
- 2. Unit-Root Tests
- 2.1. Unit-Root Tests Requiring Cross-Section Independence
- 2.2. Panel Unit-Root Tests for PTCS-Data Exhibiting Cross-Sectional Correlation
- 2.4. Small Sample Properties
- 2.5. Issues in the Application of Panel Unit Root Tests
- 2.6. Remarks Regarding the Choice of a Test and the Interpretation of Unit-Root Tests
- 3. Testing for Cointegration
- 3.1. Testing for Cointegration in the Absence of Cross-Section Dependence and Breaks
- 3.2. Dealing with Cross-Section Dependence and Breaks
- 3.2.1. Allowing for Breaks in Absence of Cross-Section Dependence
- 3.2.2. Testing for Cointegration in the Presence of Breaks and Cross-Section Dependence
- 3.3. Small Sample Behavior
- 3.4. Practical Considerations
- 3.5. Remarks Regarding the Choice of Test and the Interpretation of Cointegration Tests
- 4. Estimating Cointegration Relationships in Non-Stationary PTCS Data
- 4.1. Fully Modified OLS and Dynamic OLS
- 4.2. Error-Correction Models
- 4.3. Estimation of Cointegration Parameters under Cross-Section Dependence
- 4.4. Small Sample Properties
- 4.5. Comments on the Estimation of Long-Run Relationships in Applied Research
- 5. An Example: Robbery Rates in the West German Federal States
- 6. Conclusion
- References