calculate_sigma2maxmodel {RCTS}

Calculates sigma2maxmodel

Description

Sigma2 is the sum of the squared errors, divided by NT. We need the sigma2 of the maxmodel to use (in term 2,3,4 of the PIC) instead of the configuration-dependent sigma2. (See paper AndoBai 2016). sigma2_max_model could actually be set to 1 as well, as it can be absorbed in parameter C of the PIC.

Usage

calculate_sigma2maxmodel(e, kg_max, S, S_cand, kg, k, k_cand)

Arguments

e

NxT-matrix containing the estimated error term

kg_max

scalar: maximum allowed number of estimated factors for any group

S

estimated number of groups

S_cand

vector with candidate values for the number of groups

kg

vector with the estimated number of group specific factors for each group

k

estimated number of common factors

k_cand

vector with candidate value for the number of common factors

Value

numeric