wald_varcov: Compute Wald-type asymptotic variance

In omkarakatta/ivqr: Instrumental Variables Quantile Regression

Compute Wald-type asymptotic variance

Description

See (3.11) from Chernozhukov and Hansen (2006)

Usage

wald_varcov(
  resid,
  alpha,
  tau,
  D,
  X,
  Z,
  Phi = linear_projection(D, X, Z),
  psi = 1
)

Arguments

resid	MILP residuals associated with point estimate
alpha	Alpha level of the test; defaults to 0.1; used to compute critical value and the Hall and Sheather bandwidth (for estimating J_Psi(tau) in CH (2006))
tau	Quantile (number between 0 and 1)
D	Matrix of endogeneous variables
X	Matrix of covariates (including intercept)
Z	Matrix of instruments (only relevant for obtaining Phi)
Phi	Transformed matrix of instruments; defaults to linear projection of D on X and Z
psi	Scalar coefficient in front of the variance-covariance matrix; useful for tuning MCMC subsampling (see mcmc_active_basis); defaults to 1

Value

Named list

1. varcov: variance-covariance matrix multiplied by psi

2. hs: Hall and Sheather Bandwidth

3. Other objects used to create varcov

omkarakatta/ivqr documentation built on Aug. 20, 2022, 11:04 p.m.