Tomohiro Ando and Jushan Bai, "Panel Data Models with Grouped Factor Structure under Unknown Group Membership", Journal of Applied Econometrics, Vol. 31, No. 1, 2016, pp. 163-191. All data, except as noted below, are from "Datastream", the property of Thomson Reuters. The data are available only to subscribers to their service. Here we provide detailed information for replicating the results. The file ab-data.zip contains seven files. They are all ASCII text files in DOS format. Unix/Linux users should use "unzip -a". 1. Analysis of U.S. mutual fund returns The following file contains the names of 536 U.S. mutual funds, collected from the Datastream database. Returns of these U.S. mutual funds are analyzed. -- USmutualfunds.txt The following 6 variables (Definitions are given in the paper) are obtained from the Fama and French website: http://mba.tuck.dartmouth.edu/pages/faculty/ken.french/data_library.html -- Mkt -- SMB -- HML -- LTR -- STR -- Mom -- F-F_Research_Data_Factors.txt contains Mkt, SMB, HML. -- F-F_LT_Reversal_Factor.txt contains LTR -- F-F_ST_Reversal_Factor.txt contains STR -- F-F_Momentum_Factor.txt contains Mom 2. Analysis of China's mainland stock markets The following file contains the names of China mainland firms, collected from the Datastream database. Returns of these stocks are analyzed. -- Chinastock.txt The following file contains the names of explanatory variables, collected from the Datastream database. All of these variables are transformed to returns as described in the paper. -- Chinapredictors.txt Tomohiro Ando, andoh [AT] kbs.keio.ac.jp Jushan Bai, jushan.bai [AT] columbia.edu