Two-stage least squares with a weak-instrument refusal gate

Routes to the native 2SLS engine (morie_iv_tsls), first-stage diagnostics (morie_iv_first_stage_diagnostics), Stock-Yogo critical values (morie_iv_stock_yogo), and the Anderson-Rubin confidence set (morie_iv_anderson_rubin_ci). Following Staiger-Stock, refuses to report the 2SLS point estimate when the first-stage F is below 10 and returns the identification-robust Anderson-Rubin set instead.

Usage

morie_iv_2sls(
  data,
  outcome,
  endogenous,
  instruments,
  exogenous = NULL,
  alpha = 0.05
)

Arguments

data: Data frame.
outcome: Outcome column.
endogenous: Endogenous regressor column (single).
instruments: Instrument column names.
exogenous: Optional exogenous covariate names.
alpha: CI tail. Default 0.05.

Value

Object of class "morie_iv": estimate (NA when the gate refuses), std.error, conf.int, first_stage_F, stock_yogo_10, effective_F and mop_critical_10 (the heteroskedasticity-robust Montiel Olea-Pflueger effective F and its TSLS critical value at tau = 10 percent), weak_instruments, ar_confidence_set, method, n, call.

References

Staiger & Stock (1997); Anderson & Rubin (1949); Stock & Yogo (2005).

Examples

set.seed(1)
n <- 200
z <- rnorm(n); u <- rnorm(n)
d <- z + 0.5 * u + rnorm(n)
y <- 2 * d + u + rnorm(n)
morie_iv_2sls(data.frame(y, d, z), "y", "d", "z")
#> Two-stage least squares -- 2SLS (native k-class, HC1) with Staiger-Stock gate 
#>   Estimate: 2.0538  (SE 0.1122)  95% CI [1.8326, 2.2750]
#>   First-stage F = 142.23 (Stock-Yogo 10% crit = 16.38)
#>   Effective F (robust) = 116.51 (Montiel Olea-Pflueger TSLS tau=10% crit = 23.11)