Skip to content

evanjflack/bacondecomp

Repository files navigation

Build Status Coverage status Example Jupyter Notebook

bacondecomp

bacondecomp is a package with tools to perform the Goodman-Bacon decomposition for differences-in-differences with variation in treatment timing. The decomposition can be done with and without time-varying covariates.

Installation

You can install bacondecomp 0.1.1 from CRAN:

install.packages("bacondecomp")

You can install the development version of bacondecomp from GitHub:

library(devtools)
install_github("evanjflack/bacondecomp")

Functions

  • bacon(): calculates all 2x2 differences-in-differences estimates and weights for the Bacon-Goodman decomposition.

Data

  • math_refom: Aggregated data from Goodman (2019, JOLE)
  • castle: Data from Cheng and Hoekstra (2013, JHR)
  • divorce: Data from Stevenson and Wolfers (2006, QJE)

Example

This is a basic example which shows you how to use the bacon() function to decompose the two-way fixed effects estimate of the effect of an education reform on future earnings following Goodman (2019, JOLE).

library(bacondecomp)
#> Loading required package: fixest
#> fixest 0.9.0, BREAKING changes! (Permanently remove this message with fixest_startup_msg(FALSE).) 
#> - In i():
#>     + the first two arguments have been swapped! Now it's i(factor_var, continuous_var) for interactions. 
#>     + argument 'drop' has been removed (put everything in 'ref' now).
#> - In feglm(): 
#>     + the default family becomes 'gaussian' to be in line with glm(). Hence, for Poisson estimations, please use fepois() instead.
df_bacon <- bacon(incearn_ln ~ reform_math,
                  data = bacondecomp::math_reform,
                  id_var = "state",
                  time_var = "class")
                  
# All 2x2 Comparisons
head(df_bacon)
#>    treated untreated    estimate      weight                     type
#> 2     1987      1985  0.04575585 0.031655309 Later vs Earlier Treated
#> 4     1986      1985  0.11390411 0.001978457 Later vs Earlier Treated
#> 5     1984      1985  0.12012908 0.001978457 Earlier vs Later Treated
#> 6     1985      1987  0.01021379 0.039569136 Earlier vs Later Treated
#> 9     1986      1987 -0.09374495 0.005440756 Earlier vs Later Treated
#> 10    1984      1987  0.09784857 0.013354583 Earlier vs Later Treated
# Summary of Early vs. Later, Later vs. Earlier, and Treated vs. Untreated
bacon_summary(df_bacon)
#>                       type  weight  avg_est
#> 1 Earlier vs Later Treated 0.06353  0.02868
#> 2 Later vs Earlier Treated 0.05265  0.03375
#> 3     Treated vs Untreated 0.88382 -0.00129
library(ggplot2)

ggplot(df_bacon) +
  aes(x = weight, y = estimate, shape = factor(type)) +
  geom_point() +
  geom_hline(yintercept = 0) + 
  theme_minimal() +
  labs(x = "Weight", y = "Estimate", shape = "Type")

References

Goodman-Bacon, Andrew. 2018. “Difference-in-Differences with Variation in Treatment Timing.” National Bureau of Economic Research Working Paper Series No. 25018. doi: 10.3386/w25018.

Paper Link