Inferential Statisticshard
0:00.0

Suppose a sequence of M-estimators θ^n\hat{\theta}_n satisfies i=1nψ(Xi,θ^n)=0\sum_{i=1}^n \psi(X_i, \hat{\theta}_n) = 0. In a misspecified model, the asymptotic variance of n(θ^nθ)\sqrt{n}(\hat{\theta}_n - \theta^*) is given by the sandwich estimator A1BA1A^{-1} B A^{-1}. What do the matrices AA and BB represent?