Marek Jarocinski, "Responses to Monetary Policy Shocks in the East and the
West of Europe: A Comparison", Journal of Applied Econometrics, Vol. 25, 
No. 5, 2010, pp. 833-868.

All data files are in the zip file mj-data.zip. All program files are in the
zip file mj-programs.zip. All files are ASCII files in DOS format.
Unix/Linux users should use "unzip -a".

Data:

The data are organised in three directories:

2008-cee - data on the central-eastern European countries
2008-euro - data on the euro area countries
2008-world - world commodity prices, US and German data

Data files are named after countries and are in the csv format. Data sources
are described in the Appendix to the paper, and data transformations are
described in subsection 3.1.

The series are organized in columns. The first column of each file is
the year and month in the yyyy-mm format.
Missing values are denoted by #N/A. The series names are
lneur - log exchange rate local currency per euro,
lndem - log exchange rate local currency per D-Mark,
r-mkt - money market short term interest rate,
lncpi - log CPI,
lnipsa - log Industrial Production, seasonally adjusted
lntr - log Total Reserves in foreign currencies, excluding gold

The series in the 'world1' data are
ffr - federal funds rate
r-mkt-ger - log German money market interest rate
lnipsa-ger - log German industrial production index
lneur-usd - log exchange rate German currency (D-Mark or euro) per US dollar
lnoil-eur - log oil prices in euros
lnnfd-eur - log non-fuel commodity prices in euros

The file 'dummies' contains dummy variables named 'd<year><month>' and
taking the value of 1 on the specified date and 0 otherwise.

Programs:

To replicate the results reported in the paper follow these steps:

1. Make sure that Ox (console version) is installed. Ox console
is free for academic use and is available from www.doornik.com

2. Download and install two Ox packages: utils and gnudraw. These packages,
written by Charles Bos, are available from www.tinbergen.com/~cbos where
installation instructions can also be found.

3. Execute the program 'run.ox' (for example, by typing at the command
prompt 'oxl run.ox'). 'run.ox' first sets options (such as the number of
Gibbs sampler draws, Burnout etc.) and the specifications to be loaded. Then
it calls all other routines needed in the estimation. Full simulations can
take several hours, so it is best to first set the number of Gibbs sampler
draws, burnout and the thinning coefficient to some small values (e.g. 10),
and check if the programs run correctly.

The files whose names start with 'spec-...' define specifications, i.e.
countries included, sample periods, control variables, priors etc. The
estimation is performed by the function 'estimate' stored in 'estimate.ox',
which produces results for a given specification.  Results produced are the
quantiles of the posterior of impulse response functions, draws of the
lambda parameter etc., their file names should be self-explanatory. All the
results are stored in a dedicated subdirectory for each specification, and
they are in the csv format (plus some eps graphic files). All other ox files
contain additional required functions. The directory 'packages' contains
other required packages (in addition to utils and gnudraw).

Note: When creating directories and manipulating files, the ox programs
assume that the operating system understands standard unix commands such as
rm,mv,mkdir If you are using windows, one way to get them is to download
unxutils from http://sourceforge.net/projects/unxutils and put
rm.exe, mv.exe, and mkdir.exe somewhere on the path.

Marek Jarocinski
marek.jarocinski [AT] ecb.int