Joshua Abel, Robert Rich, Joseph Song, and Joseph Tracy, "The Measurement and Behavior of Uncertainty: Evidence from the ECB Survey of Professional Forecasters", Journal of Applied Econometrics, Vol. 31, No. 3, 2016, pp. 533-550. Program Code: The RichTracy_Data.zip file contains all of the necessary code to reproduce the results obtained in the paper. ReadMe.txt files are provided describing the programs and how to obtain the results. Input data are in .csv and .xlsx format, though the naming convention and content are matched to those of the text data included in the arst-data.zip file (see below). All intermediate and final datasets are created by running the Master.do file. Datasource 1: The first data source is from the Eueropean Central Bank, Survey of Professional Forecasters. This is a public data source administered by the ECB (http://www.ecb.europa.eu/stats/prices/indic/forecast/html/index.en.html). Data source questions can be directed to ecb-spf@ecb.europa.eu. All data files (which are in txt format) are zipped in the file ECB_data. Raw CSV files (matching the txt data files) are included in the RichTracy_Data.zip. A documentation file provided by the ECB can be found at (http://www.ecb.europa.eu/stats/prices/indic/forecast/shared/files/dataset_documentation_csv.pdf??18911e0be659e13bd3044a81b4a13dbd). A variable summary exerpted from this documentation file is below: The "individual rounds" files contain data for forecasts in the upper part of the spreadsheet and data for the assumptions in the lower part of the spreadsheet (except for early rounds when no assumptions were asked). The forecast data have the following column structure: 1) TARGET_PERIOD This column stands for the point (or period) in time to which the forecast refers to. Depending on the macroeconomic variable, the target period is given either as year (format "yyyy"), quarter (format "yyyyQq") or month (format "yyyymmm"), where the month is a mixed-case three-character tag of calendar months in English. See section 2 on forecast horizons for details. 2) FCT_SOURCE This column stands for forecast source or forecaster ID, is the code number assigned to an individual forecaster. This number remains the same for a specific forecaster over all forecast rounds (only the subset of forecasters that were actually responding in a particular round is reported). 3) POINT This column contains the point forecast of the forecaster for the macroeconomic indicator. 4) The following columns These columns contain the probability assigned to each of the intervals of the forecasted variable. The headers of the columns specify the intervals. The coding of the intervals is explained in Annex 4. Each file contains three tables -- one for each forecasted indicator. The tables are stored on the same worksheet one below the other, each separated by one blank row. The top table contains the inflation forecasts, the middle one the GDP forecasts and the third one reports the unemployment forecasts. Data in the tables are ordered first by target period and then by forecaster ID. Datasource 2: The second data source is from the Statistical Office of the European Communities, obtained via Haver Analytics DLXVG3. This is public use data containing quarterly HICP, GDP, and Unemployment rate data for the EU from 1995Q1-2014Q1. Each text file corresponds to either HICP, GDP, or the Unemployment rate series. The first and second columns of each text file correspond to the quarter, and the third column is the variable of interest, named by its corresponding DLX pneumonic. All data files (which are in txt format) are zipped in the file EuroArea_data.zip. A raw .xlsx spreadsheet (matching the txt data files) aggregates the individual series and is included in the file RichTracy_Data.zip. Please address any questions to: Dr. Robert Rich Research Dept. - Macroeconomic & Monetary Studies Macroeconomic & Monetary Studies Function Federal Reserve Bank of New York 33 Liberty Street New York, NY 10045-0001 USA