Synthetic Difference-in-Differences Estimation with Formula Interface

This function provides a formula-based interface to synthdid estimation, similar to lm(), plm(), and glm(). It automatically handles panel data conversion and returns a rich model object compatible with standard R modeling tools.

Usage

synthdid(
  formula,
  data,
  index = NULL,
  method = c("synthdid", "sc", "did"),
  se = FALSE,
  se_method = c("bootstrap", "jackknife", "placebo"),
  se_replications = SYNTHDID_SE_REPLICATIONS_DEFAULT,
  ...
)

Arguments

formula

A formula of the form outcome ~ treatment or outcome ~ treatment | covariates. The treatment variable should be a binary indicator (0/1).

data

A data.frame containing the panel data in long format.

index

A character vector of length 2 specifying the names of the unit and time variables, e.g., c("state", "year"). If NULL, assumes the first two columns are unit and time.

method

Estimation method: "synthdid" (default), "sc" (synthetic control), or "did" (difference-in-differences).

se

Logical. If TRUE, compute standard errors. Default is FALSE.

se_method

Standard error method: "bootstrap" (default), "jackknife", or "placebo".

se_replications

Number of replications for bootstrap/placebo standard errors.

...

Additional arguments passed to synthdid_estimate().

Value

An object of class c("synthdid", "synthdid_estimate") with components:

coefficients

Treatment effect estimate

call

The matched call

formula

The formula used

terms

The terms object from the formula

model

The model frame (if requested)

Y

The outcome matrix

N0

Number of control units

T0

Number of pre-treatment periods

weights

List with lambda, omega, and beta weights

setup

List describing the problem

estimator

Name of estimator used

index

Panel index variables

data_info

Information about the original data

See also

synthdid_estimate, panel.matrices

Examples

# \donttest{
data(california_prop99)

# Basic usage
result <- synthdid(PacksPerCapita ~ treated,
  data = california_prop99,
  index = c("State", "Year")
)

# With standard errors
result <- synthdid(PacksPerCapita ~ treated,
  data = california_prop99,
  index = c("State", "Year"),
  se = TRUE,
  se_method = "bootstrap"
)
#> Warning: bootstrap standard errors require more than one treated unit.

# Standard R methods work
print(result)
#> Synthetic Difference-in-Differences Estimate
#> 
#> Call:
#> synthdid(formula = PacksPerCapita ~ treated, data = california_prop99, 
#>     index = c("State", "Year"), se = TRUE, se_method = "bootstrap")
#> 
#> Treatment Effect:  -15.6 
#> 
#> Units:        38 control, 1 treated
#> Time Periods: 19 pre-treatment, 12 post-treatment
#> 
#> Convergence:  NOT CONVERGED - lambda (10000/10000 iters), omega (10000/10000 iters) 
#>               Consider increasing max.iter. Use summary() for details.
summary(result)
#> Call:
#> synthdid(formula = PacksPerCapita ~ treated, data = california_prop99, 
#>     index = c("State", "Year"), se = TRUE, se_method = "bootstrap")
#> 
#> Treatment Effect Estimate:
#>         Estimate Std. Error t value Pr(>|t|)
#> treated    -15.6         NA      NA       NA
#> 
#> Dimensions:
#>                          Value 
#>  Treated units:           1.000
#>  Control units:          38.000
#>  Effective controls:     16.388
#>  Post-treatment periods: 12.000
#>  Pre-treatment periods:  19.000
#>  Effective periods:       2.784
#> 
#> Top Control Units (omega weights):
#>               Weight
#> Nevada         0.124
#> New Hampshire  0.105
#> Connecticut    0.078
#> Delaware       0.070
#> Colorado       0.058
#> 
#> Top Time Periods (lambda weights):
#>      Weight
#> 1988  0.427
#> 1986  0.366
#> 1987  0.206
#> 
#> Convergence Status:
#>   Overall: NOT CONVERGED
#>   Lambda:  ✗ (10000/10000 iterations, 100.0% utilization)
#>   Omega:   ✗ (10000/10000 iterations, 100.0% utilization)
#> 
#>   Recommendation: Consider increasing max.iter or relaxing min.decrease.
#>   Use synthdid_convergence_info() for detailed diagnostics.
coef(result)
#>   treated 
#> -15.60379 
confint(result)
#> Warning: Standard error not available; cannot compute confidence interval
#>         Lower Upper
#> treated    NA    NA
plot(result)


# Compare methods
did_result <- synthdid(PacksPerCapita ~ treated,
  data = california_prop99,
  index = c("State", "Year"),
  method = "did"
)
# }