CODEWITHSUNDEEP
rdifference
banannanas
banannanas

Reputation: 56

Monthly/quaterly data when using did_imputation in R

I am using the did_imputation package in R for the Imputation Estimator from Borusyak, Jaravel, and Spiess (2021). I suspect to have some issues regarding the formatting of my variables, especially with regards to the date variables (of observation and time of treatment). Here is a small sample of my dataset:

data <- data.frame(
  coicop = c("CP01111", "CP01112", "CP01113", "CP01114", "CP01115"),
  date = as.Date(c("2010-02-01", "2010-02-01", "2010-02-01", "2010-02-01", "2010-02-01")),
  YoY_Inflation = c(-0.559, -6.039, 0.191, 0.537, -1.572),
  tname = c("2010-2", "2010-2", "2010-2", "2010-2", "2010-2"),
  gname = c(NA, "2012-1", NA, "2010-1", NA),
  stringsAsFactors = FALSE
)

I have tried the following command:

did_imputation(data = data, yname = "YoY_Inflation", tname = "tname", gname = "gname", idname = "coicop")

Unfortunately, this gives me an error message: "Error in .subset2(x, i, exact = exact): no such index at level 1"

The documentation does not specify in what format the date of treatment and date of observation have to specified. I have tried to use dates, which was not successfull either. I would very much like to preserve the monthly nature of the data. All examples I could find online are based on yearly data, which are formatted as numeric, integer, or double.

Does anybody have experience with this?

Upvotes: 2

Views: 35

Answers (0)

Related Questions