Konstantinos Drakos and Panagiotis Th. Konstantinou, "Investment Decisions in Manufacturing: Assessing the Effects of Real Oil Prices and their Uncertainty", Journal of Applied Econometrics, Vol. 28. No. 1, 2013, pp. 151-165. The plant-level data used in this study are the property of the Hellenic Statistical Authority (EL.STAT, http://www.statistics.gr). These are based on the Annual Industrial Surveys for Greece for the years between 1994 and 2005. As the data are confidential (we have signed a confidentiality agreement with the Hellenic Statistical Authority), they cannot be lodged here. The plant-level variable we have employed in our data set and permission for use may be obtained by contacting the Hellenic Statistical Authority. The plant-level variables we have employed are: (1) EMP := log of the total number of employees per plant. (2) EQ := the ratio of equity to industrial value added. (3) LO := the sum of bank loans and other loans, divided by industrial value added. (4) CF := the difference between total revenue and total cost, divided by industrial value added. (5) SL := the ratio of sales to industrial value added. For these, outliers corresponding to values above the 99th percentile have been excluded from the sample. Our investment indicator is based on: (6) GINV:= Total New Investment minus Total Disposals. This is constructed as described in page 1 of the online appendix. (7) DEP:= Petrol Expenditures divided by the sum of (Petrol Expenditures + Coal Expenditures + Lignite Expenditures + Natural Gas Expenditures) (8) profit rate := the difference between total revenue and total cost, divided by industrial value added. We drop observations that are below -.8948985934257507. ~~~~~ The macroeconomic variables we use and the programs used in constructing them are included in two zip files. - Most of the variables are included in the file "dk-macro-data.xls". The same data can be found in the following txt files: "Monthly-data.txt" Contains the raw monthly oil prices (nominal and real) on which all our uncertainty/risk measures are based. It also contains (on a monthly frequency) our baseline uncertainty measure, as well as two measures of uncertainty based on unconditional standard deviations ("naive" backward looking, the "perfect" forward looking). "Annual-OECD.txt" Contains annual data on manufacturing prices and the output gap. The former are used in creating our real oil price change measure employed in our analysis. The latter is employed in our robustness analysis. "Annual-Oil-Prices.txt" Puts together the (annual) real oil price change measure we use in our analysis. "REIR_{1}.txt" Contain (monthly) measures of Excess Increase Risk of Real Oil Prices (tail expectation), as well as the unpredictable component of the four-year real oil price change. These have been generated by file DK_REIR_oil_4y.m "REIR_{2}.txt" Contain (monthly) measures of Excess Increase Risk of Real Oil Prices (tail variances about the target). These have been generated by file DK_REIR_oil_4y.m "UR_1.txt" Contains (monthly) measures of Upside Risk of Real Oil Prices (tail expactation). These have been generated by file DK_oil_4y_level_boot.m "UR_2.txt" Contain (monthly) measures of Upside Risk of Real Oil Prices (tail variances about the target). These have been generated by file DK_oil_4y_level_boot.m "DataCalculations.txt" Contains yearly averages or risk/uncertainty measures, annual changes of real oil prices, and various macroeconomic variables used in our robustness results. - The baseline uncertainty measures are based on one-year-ahead predictions. These are estimated using Eviews 6.0, and can be found, along with other baseline estimations, in the file "jaedataset.wf1". The workfile consists of four "pages" which are described below: "Monthly.txt" Contains monthly nominal oil prices (in USD), the Euro/USD exchange and manufacturing prices. Based on these the 1-month and 48-month growth rates have been created. The same data are contained in file "Monthly Data.txt". These are the data used to produce Table A.4 of the appendix. The data can be accessed using "Monthly.txt". "Monthly-forecasts.txt" Contains the same variables as "Monthly", and also the predicted one-year-ahead conditional variances from the GARCH(1,1) model shown in Table A.4 of the appendix. Then, using the program jae_monthly_forecast_lump.prg, a new series is created (garch_real_1m_f) which is exactly the predicted conditional variance. By taking the square root of this variable we end up with the series "std_real_1m_f", which corresponds exactly to the variable "GARCH(1,1)-std(F)" contained in the file "Monthly Data.txt". The data can be accessed using "Monthly_forecasts.txt". "MonthlyREIR_UR.txt" Contains monthly observations of the risk/uncertainty series produced by the MATLAB files DK_REIR_oil_4y.m & DK_oil_4y_level_boot.m as well as the real oil price and its four year change. The data can be accessed using "MonthlyREIR_UR.txt". "Correlations.txt" Contains all the necessary series for creating Figures A.1 to A.3 in the appendix, and calculating the relevant correlations reported for each figure. The data can be accessed using "Correlations.txt". - The MATLAB file DK_REIR_oil_4y.m puts together measures of Excess Increase Risk of Real Oil Prices. The necessary background codes can be downloaded either from the webpage of Simone Manganeli (www.simonemanganelli.org/Simone/Research_files/DeflationCodes.zip) or from the data archive of Journal of Money, Credit and Banking (https://jmcb.osu.edu/volume-39-2007 ). The required routinres are in the sub-folder [Plots]. We also include the necessary files here, but the reader should clearly cite the paper by Kilian and Manganeli (2007) when using these. - The MATLAB file DK_oil_4y_level_boot.m puts together upside risk measures of Real Oil Prices. ~~~~~ There are two zip files. The file dk-binary-files.zip contains dk-macro-data.xls and jaedataset.wf1. The file dk-text-files.zip contains all other files, which are ASCII files in DOS format. Unix/Linux users should use "unzip -a" for dk-text-files.zip only. Panagiotis Th. Konstantinou Panagiotis.Konstantinou [AT] brunel.ac.uk