monte_rev2.m: simulation program except for HQ_c^{*} (by the Hallin and Liska's (2007) procedure) monte_revHL.m: simulation program for HQ_c^{*} genar.m: generate VAR(1) process generror.m: generate idiosyncratic error terms inco.m: for evaluating accuracy rates standard.m: for standardization Onatski.m: Onatski's (2010) ED estimator ERGR.m: Ahn and Horenstein's (2013) ER and GR estimators * Note that there are three options for demeaning. * demean option: 0==employing raw data, 1==demeaning cross-sectional averages, 2==demeaning cross-sectional and time averages * In usual factor analysis, "option 1" is taken. Note that eliminating cross-sectional averages eliminates the single common component in case of additive individual fixed effects. * In using Ahn and Horenstein's (2013) ER and GR estimators, their paper mentions "double differenced data". For this, choose "option 2". In case of the additive individual and time fixed effects, two common components will be eliminated. HQC: HQ_c^{*} estimator i.e., Hannan and Quinn criterion with Hallin and Liska's (2007) procedure subr_HQC.m: Hannan and Quinn's criterion with a subsample secint.m: finding second interval for Hallin and Liska's (2007) procedure * Note that Hallin and Liska's (2007) procedure needs the standardization, e.g., x=X-X_bar/std(X)