使用 plm 进行固定效应回归

Question

下面的代码使用 biglm 运行时间序列回归“excessr ~ mkt_rf”，因为函数 lm 不适用于我的真实数据集。

现在我想从 biglm 切换到 plm 以解决固定效果。不幸的是 plm 不起作用。

有谁知道我可以改变 plm 的工作吗？提前致谢！

library(biglm)
library(plm)
library(data.table)

union_with_factors = data.table(
  t = c(1,2,3,4,5,1,2,3,4,5,1,2,3,4,5),
  excessr  = c(10,23,13,53,43,76,34,12,45,13,42,31,4,53,64),
  FundId = c("x","x","x","x","x","y","y","y","y","y","z","z","z","z","z"),
  mkt_rf = c(1,1,2,1,3,1,1,2,1,3,1,1,2,1,3)
)

sp <- split(union_with_factors, union_with_factors$FundId)
beta <- sapply(sp, function(tmp){
  fit <- plm(excessr ~ mkt_rf, data = tmp)
  coef(fit)
})

Answer 1

假设个人的 ID 由 FundId 给出，时间 ID 由 t 给出，以下是如何应用固定效应回归：

library(biglm)
library(data.table)
library(plm)

union_with_factors = data.table(
  t = c(1,2,3,4,5,1,2,3,4,5,1,2,3,4,5),
  excessr  = c(10,23,13,53,43,76,34,12,45,13,42,31,4,53,64),
  FundId = c("x","x","x","x","x","y","y","y","y","y","z","z","z","z","z"),
  mkt_rf = c(1,1,2,1,3,1,1,2,1,3,1,1,2,1,3)
)
fit <- plm(excessr ~ mkt_rf, 
             data = union_with_factors, 
             index = c("FundId", "t"), 
             model = "within")
summary(fit)
fixef(fit)

有关详细信息，请参阅here 和plm 文档（控制台中的?plm）

编辑：在this post 和this article 之后，您似乎可以使用pmg（也在plm 包中）进行Fama-MacBeth 回归：

fama_macbeth <- pmg(excessr ~ mkt_rf, 
                   data = union_with_factors, 
                   index = c("FundId", "t"))
summary(fama_macbeth)