固定效应的plm回归

Question

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

现在，我想要从大到plm的帐户固定效果。不幸的是plm不起作用。

有人知道我能改变plm的作用吗？

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是由FundId提供的，而时间ID是由t提供的，下面是如何应用固定效果回归：

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，请参阅这里和plm文档(控制台中的?plm)

编辑:在这个职位和这篇文章之后的，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)