Matteo Pelagatti and Emilio Colombo, "On the Empirical Failure of Purchasing Power Parity Tests", Journal of Applied Econometrics, Vol. 30, No. 6, 2015, pp. 904-923. This short text explains how to use the data and code available in this archive to reproduce the simulation and application sections of the JAE article cited above. FILES RELATED TO SECTION 4, SIMULATION EXPERIMENT The simulations were implemented in the Ox language (www.doornik.com) because of its high efficiency and accurateness. The file containing the code is named "ppp_sim.ox" and its content is commented on a line-by-line basis. It should be relatively easy to understand the meaning of each line even for those not acquainted with the Ox language. The code uses three .xls files, which here are also made available as .csv files: "ItemWeights2005ox.xls" contains the weights of year 2005 used for the harmonized CPI; the columns contain the COICOP sector names and the row contains the countries. The names and order of the countries is COUNTRY NAME, ROW NUMBER Belgium, 1 Bulgaria, 2 Czech Republic, 3 Denmark, 4 Germany, 5 Estonia, 6 Ireland, 7 Greece, 8 Spain, 9 France, 10 Italy, 11 Cyprus, 12 Latvia, 13 Lithuania, 14 Luxembourg, 15 Hungary, 16 Malta, 17 Netherlands, 18 Austria, 19 Poland, 20 Portugal, 21 Romania, 22 Slovenia, 23 Slovakia, 24 Finland, 25 Sweden, 26 United Kingdom, 27 Iceland, 28 Norway, 29 Switzerland, 30 Croatia, 31 Turkey 32 "YearlyFR.xls" and "YearlyBE.xls" contain the harmonized CPI for each COICOP sector, for the countries France and Belgium respectively. The columns contain the COICOP sectors and the rows the years (from top to bottom: 2000-2010). FILES RELATED TO SECTION 5, EMPIRICAL EVIDENCE For the application we used R (www.r-project.org) which is not as efficient as Ox, but has the advantage of being open source and interactive. The code can be found in the file "application.r" and is commented and relatively self-explanatory to those who know R. It uses the packages "urca" for unit root testing and "XLConnect" for reading the .xslx data files. The data files used, which here are available also as .csv, are the following. "sectors.xlsx" contains only one column with the list of COICOP codes for the sectors used in CPI indexes. "be.xlsx", "de.xlsx", "fr.xlsx", "lu.xlsx", "nl.xlsx" : For the countries Belgium, Germany, France, Luxemburg, Netherlands, respectively, these files contain the harmonized CPI index time series for each COICOP sector. The first column contains the dates, and the next columns contain the COICOP sectors. The time series are monthly from 1996M1 to 2013M10. "wbe.xlsx", "wde.xlsx", "wfr.xlsx", "wlu.xlsx", "wnl.xlsx" : For the countries Belgium, Germany, France, Luxemburg, Netherlands, respectively, these files contain the weights of the harmonized CPI indexes for each COICOP sector. The first column contains the dates, and the next columns contain the COICOP sectors. The time series are yearly (the weights change on a yearly basis) from 1996 to 2013. All the .xls and .xlsx files are zipped in the file pc-data-xls.zip. All the .csv files and the single .r file are zipped in the file pc-files.zip. Unix/Linux users should use "unzip -a" with pc-files.zip but should not use the "-a" option with pc-data-xls.zip.