Eric M. Scheffel, "Accounting for the Political Uncertainty Factor", Journal of Applied Econometrics, Vol. 31, No. 6, 2016, pp. 1048-1064. Introduction The paper formulates, estimates, and identifies a structural dynamic factor model using a comprehensive set of time series data. The purpose of this Readme file is to accompany and describe the data employed in the above study. Further, this Readme file is bundled with a series of other files for replication purposes which are listed and described in the following: 1) data_joae_8188.csv (The complete set of untransformed or raw data employed in the publication in CSV format) 2) October_Updated_finaldata.csv (A subset of data originally downloaded from www.policyuncertainty.com -- no longer publicly available) 3) run.py (Original Python main program file used in producing key results in the publication) 4) ems-datadesc.pdf (A page extracted from the main publication describing the data and the transformations carried out on it prior to analysis) The first three files are ASCII files in DOS format. They are zipped in the file es-files.zip. Unix/Linux users should use "unzip -a". Additional comments The data file "data_joae_8188.csv" contains a total of 93 time series of monthly frequency which pertain to the U.S. economy. The total number of observations for all of the data is 320 and the date range is from 01-January-1985 to 01-August-2011. The majority of data included in "data_joae_8188.csv" has been sourced from the Federal Reserve Bank of St. Louis' research database FRED and the original short mnemonics for each times series have been preserved in the data set included with the publication. The total uncompressed size of the data file as stored on disk is 187.2 kBs. The only time series which was not originally obtained from FRED is Bloom, Baker & Davies' Policy Uncertainty Index which was downloaded from their website www.policyuncertainty.com in November of 2011 and then subsequently incorporated into the present structural dynamic factor model study. The original data obtained from www.policyuncertainty.com is stored in the file "October_Updated_finaldata.csv" but the same policy uncertainty index is also again incorporated into the final data set in "data_joae_8188.csv". All of the data accompanying this study are untransformed implying that any necessary de-trending and normalizing transformations are carried out as part of executing the Python code files "run.py". The types of transformations carried out on each time series is recorded in the file "datadesc.pdf", which also provides a more detailed description of each time series. The original Python code file "run.py" can only be run successfully if placed within a suitable Python "eco-system" of supporting libraries which can be readily identified by studying the series of "import" commands stated at the top part of the file "run.py". Also, since the results were only ever produced within a Linux system environment, no guarantee can be given that results are just as easily replicable within a Microsoft Windows environment. Finally, the program "run.py" is programmed in a way so as to prefer to download the relevant data in real-time directly from the St. Louis FRED database. An option to instead source the data from permanent storage is available but may require marginal alterations to the code. Some parts of the code were included experimentally but eventually never used for the final publication, such as for example Bai&Ng(2002) factor test statistic. Author Contact Eric Michael Scheffel Assistant Professor in Economics University of Nottingham Business School China eric.scheffel [AT] nottingham.edu.cn www.ericscheffel.com Ningbo, 19th January 2015