| 邓明.时变系数的空间误差合成模型[J].数量经济技术经济研究,2013,30(4):111-123 |
| 时变系数的空间误差合成模型 |
| Fixed Efficient Spatial Panel Data Model with Error Components |
| |
| DOI： |
| 中文关键词: 时变系数 空间误差合成模型 可行广义最小二乘 广义矩 |
| 英文关键词: Time-varying Coefficient Spatial Error Components Model FGLS GM |
| 基金项目: |
|
| 中文摘要: |
| 本文对时变系数的空间误差合成模型进行了研究，该模型的特征是利用扰动项中的空间个体成分将不同时期的方程联系起来；同时，允许自变量系数和误差项中的空间自回归系数随时间而变化，但每一个时期的自变量系数和误差项中的空间自回归系数是固定不变的。本文使用基于FGLS和GM方法的多阶段估计策略对模型参数进行了估计，证明了估计量的渐进异质性，并利用Monte Carlo方法模拟了其小样本性质。模拟结果表明，估计量的渐近性质随着样本容量的增加而改善。对中国省际知识生产及其空间溢出的实证案例也体现了该模型的应用价值。 |
| 英文摘要: |
| This paper researches the fixed efficient spatial panel data model with error components. This model combines the equations in different time by the individual factors of error disturbance, and allows the independent coefficients and the spatial autoregressive coefficients changing over time. However, the independent coefficients and the spatial autoregressive coefficients are fixed in each time. This paper estimates the coefficients by a multi-stage estimation based on FGLS and GM, proves the asymptotic consistence and simulates the small sample properties by Monte Carlo simulation. The simulation output reveals that the asymptotic property improved with the increase of sample size. The empirical research of China’s provicial knowledge product and spatial spillover also reflects the application value of the model. |
| 查看全文 |
