didcf

library(didcf)

Introduction

This vignette accompanies the package DiDCF. The algorithm comes from Gavrilova, Evelina, Audun Langørgen, and Floris Zoutman. “Difference-in-difference causal forests, with an application to payroll tax incidence in norway.” (2025).

How to install

Below I offer two options on how to install the package from GitHub. You can use devtools (which needs to be installed beforehand).

#install.packages(devtools)
library(devtools)
install_github("evelinagz/didcf")

estimate_didcf

The function estimate_DiDCF takes in your data and estimates difference-in-difference causal forests. Here is how you can use the function:

estimate_didcf(example_data$Y, 
               example_data$t_indicator[example_data$period==1], 
               model.matrix(~.,data=example_data[example_data$period==1,c("x_1","x_2")]), 
               1, 
               example_data$period, 
               example_data$unit_id)

The example can be found also in the help file related to the function. The inputs to the function are the following:

  1. Y - this is the outcome variable for all periods
  2. t_indicator - this is a cross-sectional indicator for treatment, it should be the same in every period, here it is taken from the base period 1
  3. X - these are the variables describing the margins of heterogeneity in the treatment effect of interest
  4. 1 - this is the value of the base period in the variable denoting time
  5. period - variable denoting time
  6. unit_id - variable denoting unit id

vip

The function vip is a handy tool to extract a table from the variable importance feature of the estimated causal forest. The table contains variable names and their respective importance values, sorted in descending order.

my_forest<-estimate_didcf(example_data$Y, 
               example_data$t_indicator[example_data$period==1], 
               model.matrix(~.,data=example_data[example_data$period==1,c("x_1","x_2")]), 
               1, 
               example_data$period, 
               example_data$unit_id) 
vip(my_forest[[1]])

Example Data

The example_data dataset included in this package contains a simulated dataset tailored to the payroll example from the paper. Here is how you can load and view the dataset:

# Load the example dataset
data(example_data, package = "didcf")

# View the dataset
head(example_data)