cic2

This is a multi-period implementation of the change-in-changes approach from Athey and Imbens (2006, Econometrica). This function is in a beta release and users should use caution when using this function in emprical work.

The function builds on the pte package and will return an overall treatment effect parameter as well as an event study. See, in particular, the argument ret_quantile below.

cic2(
  yname,
  gname,
  tname,
  idname,
  data,
  xformla = ~1,
  ret_quantile = NULL,
  gt_type = "att",
  anticipation = 0,
  cband = TRUE,
  alp = 0.05,
  boot_type = "empirical",
  biters = 100,
  cl = 1
)

Arguments

yname

Name of outcome in data

gname

Name of group in data

tname

Name of time period in data

idname

Name of id in data

data

balanced panel data

ret_quantile

This parameter determines which quantile will be reported by the cic2 function. By default ret_quantile=NULL; in this case, the function will return an estimate of the overall ATT and an event study for the ATT. Other choices should be between 0 and 1. For example, if the user specifies ret_quantile=0.9, then the function will return overall and event study parameters for the QTT(0.9). These ...would be better to return the overall distribution and then to average and invert in later steps...

alp

significance level; default is 0.05

boot_type

should be one of "multiplier" (the default) or "empirical". The multiplier bootstrap is generally much faster, but attgt_fun needs to provide an expression for the influence function (which could be challenging to figure out). If no influence function is provided, then the pte package will use the empirical bootstrap no matter what the value of this parameter.

biters

number of bootstrap iterations; default is 100

cl

number of clusters to be used when bootstrapping; default is 1

ret_dist

If set to be true, the function returns the observed distribution of outcomes and counterfactual distribution of outcomes for each (g,t) through the extra_gt_returns element of group_time_att object.