Compute Sandwich (Robust) Standard Errors — dcc11_sandwich

Heteroskedasticity-robust SEs using sandwich estimator: Var(theta) = I^-1 J I^-1

Usage

dcc11_sandwich_se(
  params,
  std_resid,
  weights,
  Qbar,
  distribution = "mvn",
  use_reparam = FALSE
)

Arguments

params: MLE parameter estimates
std_resid: T x k matrix of standardized residuals
weights: T-vector of observation weights
Qbar: k x k unconditional covariance matrix
distribution: "mvn" or "mvt"
use_reparam: Logical: parameters in (psi, phi) space?

Value

List with:

se: Standard errors (square root of diagonal of vcov)
vcov: Sandwich variance-covariance matrix: ($H^{-1} J H^{-1}$
bread: $H^{âˆ’1}$, inverse Hessian of the negative log-likelihood
meat: J, outer product of score vectors: $sum_t (grad_t)(grad_t)'$
params: Parameter values at which SEs were computed
param_names: Names of parameters
method: "sandwich"

Details

The sandwich (robust) variance estimator is: $$Var(\hat{\theta}) = H^{-1} J H^{-1}$$ where H is the Hessian and J is the outer product of score contributions. This is consistent under heteroskedasticity when the standard Hessian-based estimator may not be.