| Type: | Package |
| Title: | Computes Powell's Generalized Synthetic Control Estimator |
| Version: | 1.0.0 |
| Date: | 2018-10-21 |
| Author: | Philip Barrett |
| Maintainer: | Philip Barrett <pobarrett@gmail.com> |
| Description: | Computes the generalized synthetic control estimator described in Powell (2017) <doi:10.7249/WR1142>. Provides both point estimates, and hypothesis testing. |
| License: | GPL-2 |
| Depends: | R (≥ 2.10), Rcpp (≥ 0.12.18), nloptr, reshape2 |
| LinkingTo: | Rcpp, RcppArmadillo |
| RoxygenNote: | 6.1.0 |
| Suggests: | knitr, rmarkdown, plm |
| VignetteBuilder: | knitr |
| URL: | https://github.com/philipbarrett/pgsc |
| Encoding: | UTF-8 |
| NeedsCompilation: | yes |
| Packaged: | 2018-10-21 20:37:31 UTC; philip |
| Repository: | CRAN |
| Date/Publication: | 2018-10-28 22:50:03 UTC |
Wrapper function for GSC estimation
Description
Wrapper function for GSC estimation
Usage
pgsc(dta, dep.var, indep.var, b.init, method, sol.it = NULL,
wt.init = NULL, print.level = 0, g.i = NULL, g.i.grad = NULL,
...)
Arguments
dta |
A data frame |
dep.var |
A string defining the dependent variable |
indep.var |
A vector of strings defining the independent (treatment) variables |
b.init |
An initial value for the treatment variable coefficients. Must have same length as 'indep.var' |
method |
The GSC iteration method to be used. Must be one of:
|
sol.it |
The first step solution used in the two-step methods. If omitted, a new one-step solution is computed. |
wt.init |
An initial value for the weighting matrix |
print.level |
The level of detail provided in the printed output |
g.i |
A function defining a restriction on the parameters. Used in hypothesis testing. |
g.i.grad |
The gradient of |
... |
Other arguments to be passed to the optimization |
Details
See the vignette "Using pgsc" for an extended example.
Value
Returns the point estimate of the model as a gsc object, a list with entries:
- b
The point estimate of the coefficients on the dependent variables
- diff
The difference between successive iterations
- err
The maximum error on the within-iteration optimization problems
- it
Number of iterations require to solve
- sig.i
The unit-specific MSEs from the solution
- W
The "full" weighting matrix for counterfactuals, containing own-unit weights (all zero) and unit-N weights
- wt
The "minimal" weighting matrix, omitting own-unit weights and weights on unit N (which can be computed as one-minus-rowsum)
Examples
data("pgsc.dta")
sol <- pgsc(pgsc.dta, dep.var = 'y', indep.var = c('D1','D2'),
b.init = c(0,0), method='onestep' )
summary(sol)
g.i <- function(b) b[1] ; g.i.grad <- function(b) c(1,0)
sol.r <- pgsc(pgsc.dta, dep.var = 'y', indep.var = c('D1','D2'),
b.init = sol$b, method='onestep', g.i=g.i, g.i.grad=g.i.grad )
summary(sol.r)
Synthetic data for PGSC testing.
Description
A dataset with an outcome given by a treatment and a set of factors.
Usage
pgsc.dta
Format
A data frame with 750 rows and 8 variables:
- n
The unit, here labeled as a US state
- t
The time period
- y
The outcome variable
- D1
The first treatment variable
- D2
The second treatment variable
- X1
The first observed confounding factor
- X2
The second observed confounding factor
- X3
The third observed confounding factor
Source
Generated by code in the package vignette "Using pgsc".
A wrapper for the wald test of a restricted solution
Description
A wrapper for the wald test of a restricted solution
Usage
pgsc.wald.test(dta, dep.var, indep.var, sol.rest, n.boot = 10000,
seed = 42)
Arguments
dta |
A dataframe |
dep.var |
A vector of strings of names of dependent variables. |
indep.var |
A vector of strings of names of independent (treatment) variables. |
sol.rest |
A restricted solution which is being tested |
n.boot |
The number of bootstrapped samples for the variance calculation. Default is 10000. |
seed |
Randomization seed. Default is 42. |
Details
See the vignette "Using pgsc" for an extended example.
Value
Returns the wald test as gsc.wald object, a list with entries:
- b
The point estimate of the coefficients on the dependent variables
- S
The Wald statistic
- s.boot
The bootstrapped Wald statistic
- p.value
The p-value for the Wald statistic.
Examples
data("pgsc.dta")
g.i <- function(b) b[1] ; g.i.grad <- function(b) c(1,0)
sol.r <- pgsc(pgsc.dta, dep.var = 'y', indep.var = c('D1','D2'),
b.init = c(0,1), method='onestep', g.i=g.i, g.i.grad=g.i.grad )
wald <- pgsc.wald.test( pgsc.dta, 'y', indep.var = c('D1','D2'), sol.r )
summary(wald)
plot(wald)