Cluster-robust inference for the doubly robust DiD estimator

The point estimate is the doubly robust moment of Sant'Anna and Zhao (2020), equation (2.6). The influence function is summed within cluster before it is squared, V = a (1/n^2) sum_c (sum_{i in c} psi_i)^2 with the finite cluster correction a = [G/(G-1)][(n-1)/(n-k)]. With one observation per cluster the sum collapses and V is the independent variance times a.

Usage

Drclt(y, D, X = NULL, cluster = NULL)

Arguments

y: Outcome change, one entry per unit.
D: Binary treatment indicator.
X: Optional baseline covariates.
cluster: Cluster label per unit; NULL gives one unit per cluster.

Value

List with estimate, se, se_iid, vif, n_clusters, dof_adj, k, n.

References

Bertrand, M., Duflo, E. and Mullainathan, S. (2004). Quarterly Journal of Economics 119(1), 249-275. Sant'Anna, P. H. C. and Zhao, J. (2020). Journal of Econometrics 219(1), 101-122.

Examples

V <- c(1, 2, 3, 4, 5, 6, 7, 8)
Drclt(V, V)
#> $estimate
#> [1] 5.666667
#> 
#> $se
#> [1] 1.774302
#> 
#> $se_iid
#> [1] 1.659708
#> 
#> $vif
#> [1] 1.142857
#> 
#> $n_clusters
#> [1] 8
#> 
#> $dof_adj
#> [1] 1.142857
#> 
#> $k
#> [1] 1
#> 
#> $n
#> [1] 8
#> 
#> $method
#> [1] "Cluster-robust DR-DiD"
#> 
#> attr(,"class")
#> [1] "morie_rich_result" "list"