项目详情

• 项目名称：【估计动态面板+time effect】Stata   [已结项]
• 分类：
• 发布者： 水***莮
• 点击量：855
• 发布时间：2016-3-16 11:31
• 项目关键词：
• 项目描述：项目描述:类似复写原文的STATA，但是要加time effect
  要求：
  统计背景强，懂文中的7种估计方法，熟练STATA. 文中的7个方法基本都可以用STATA语句调用。
  
  OLS Ordinary Least Squares ignores the data's panel structure and generally produces an upward-biased coefficient estimate
  for the lagged dependent variable in the presence of unobserved heterogeneity (Bond (2002)). We use the Stata
  procedure “reg” to generate these coefficient estimates.
  FE Fixed Effects incorporates the data's panel structure but ignores the correlation between the lagged dependent variable and
  the regression error. FE yields a downward-biased coefficient estimate for the lagged dependent variable (Nickell (1981)).
  We use the Stata procedure “xtreg, fe” to generate these estimates.
  The following five “advanced” estimators use alternative techniques to avoid the biases associated with FE.
  AB Arellano and Bond's (1991) difference GMM first-differences the linear regression model and uses lagged dependent
  variable levels to instrument for the first difference of the lag. We employ the Stata procedure ‘xtabond’, which defaults
  to using one lag of the exogenous variables as the instrument set.
  BB Blundell and Bond's (1998) system GMM estimates a two-equation system of the regression in levels and in first
  differences. We use the Stata procedure “xtdpdsys” and limit the maximum number of lags to two.4 We specify the
  explanatory variables as predetermined,5 not fully exogenous.
  LD4 Four Period Long Differencing replicates the Huang and Ritter (2009) implementation of Hahn et al.'s (2007) estimator.
  They consider several differencing windows, butwe implement their 4-period variation in our simulations because it excludes
  the fewest firms (see Appendix A). Our estimates are computed with Stata code kindly provided by Rongbing Huang.
  LD Longest Differencing is an alternative adaptation of Hahn et al.'s (2007) balanced panel estimator, in this case allowing for
  unbalanced panels. No literature describes how to implement an LD estimator with unbalanced panels; our procedure is
  described in Appendix A.
  LSDVC Least Squares Dummy Variable Correction corrects the biased FE-estimated coefficients, using an estimate of the short-panel
  bias computed from each firm's data. It assumes that the independent variables are exogenous. Our estimates are computed
  with the user-written Stata procedure, ‘xtlsdvc’ (Bruno (2005)), which requires a vector of coefficient starting values. We
  experimentedwith using either the AB or the BB initial estimates, but found that the LSDVC estimates are robust to the initial
  matrix selection.We therefore report results derived from AB estimates of the initial coefficient matrix.
  
  完成时间：30天左右
  所需技术： STATA 统计 英语
  
  非中介，请看完附件确定会做，再联系我谢谢。其他的资料QQ联系详细谈
  
  
  
• 项目附件： 承接人可见
• 项目预算： ￥ 500-1000 元（ 已托管500元 ）
• 分享：

竞标详情