| 研究課題/領域番号 |
21K20121
|
|---|
| 研究機関 | 一橋大学 |
|---|
研究代表者 |
Jiang Peiyun 一橋大学, 社会科学高等研究院, 特任助教 (20906929)
|
|---|
| 研究期間 (年度) |
2021-08-30 – 2023-03-31
|
| キーワード | consistency / common break estimator / panel data / structural change / environmental analysis |
|---|
| 研究実績の概要 |
This research aims to achieve accurate inference and efficient estimation of structural changes in panel data models. We have two achievements in the FY2021.
(a) Project 1: The panel data frameworks often assume that the breaks occurred at a common location for all individuals. We proposed a test to evaluate the validity of common break assumption and applied it to analyze structural change phenomenon in the relationship between carbon dioxide emissions, energy consumption, and real output. We figured out that ten of seventy-two countries exhibited structural change at the same location and demonstrated the environmental Kuznets curve (EKC). More accurate inference has been confirmed compared to the existing applications' results, which is helpful for environmental policy analysis.
(b) Project 2: To take into account individual heterogeneity due to different institutional and cultural backgrounds, we relaxed the common break assumption and allowed that the break points are common in each group but vary across groups. We derived the limiting properties of the common break estimators when the common break assumption failed. And we demonstrated that the common break date estimator still achieves consistency and will converge in probability to one of the true break dates. This consistency property indicates an alternative approach to estimate heterogeneous breaks in panel data models, which is more convenient and reduces computation burdens.
|
| 現在までの達成度 (区分) |
現在までの達成度 (区分)
2: おおむね順調に進展している
理由
The two-year goals of this research are (1) to introduce a test to evaluate the validity of the common breaks assumption (2) to extend the theory in a general and realistic model by allowing for cross-sectional dependence in error terms and unobserved heterogeneity in breaks (3) to propose a new break point estimation procedure with fast and efficient computation (4) to combine theory with application to analyze economic issues. In the FY2021, we proposed a new test to detect common breaks in panel data models and applied it in environmental policy analysis, achieving goals (1) and (4). On the other hand, we investigated the asymptotic properties of common break estimators under misspecification to search for new estimation methodologies and completed 80% of (3). The final remaining goal is (2) for the FY2022.
|
| 今後の研究の推進方策 |
The evidence of distinct break points in the environmental Kuznets curve (EKC) indicates the significance of the validity of the common breaks assumption. If the common breaks assumption does not hold, the estimation and inferences will be misleading and inaccurate. We will continue the work on the common breaks assumption and further investigate more general and realistic models to achieve the goal (2) in the FY2022.
We consider a general panel model where the error terms are generally strongly correlated through a common factor structure. We have applied the common correlated effects estimator of Pesaran (2006) to estimate the factor structure and successfully derived the asymptotic distributions of the testing statistic under the null hypothesis. However, the Monte Carlo simulation results did not work as it is expected. Hence, we will deal with these contradictions. Or we plan to search for new ways to estimate the factor structures. We expect our method to be extended to a more general panel model and used to analyze real issues.
|
| 次年度使用額が生じた理由 |
In order to purchase computer and software for data analysis, attend international/domestic conferences, pay paper publication fees and so on, more expenses are expected in the FY2022.
|
| 備考 |
Jiang, Peiyun & Kurozumi, Eiji, 2021. "A new test for common breaks in heterogeneous panel data models," Discussion paper series, Hitotsubashi University. https://hias.hit-u.ac.jp/hias_dps/hias-e-107
|
