Readme file for CM.exe

The zip-archive cm.zip contains the following files:

- CM.exe:
The program CM.exe computes solutions for the stochastic
growth model with flexible labor supply described in
Burkhard Heer and Alfred Maussner, Computation of Business Cycle Models:
A Comparison of Numerical Methods, Revised Version, January 2006.


- EDP.for, GaussNewton.for, HM.for, LA.for, Labor.for, MNR.for, PEA.for
  PJ.for, QN.for, Quickwin.for, Search1.for, Search2.for, SOA.for, Toolbox.for,
  and VI.for.

These files contain the Fortran source code of CM.exe. The source code
is compiled and linked using the Compaq Visual Fortran environment version 6.6C for
the Windows operation system. If you want to compile the source using another
Fortran compiler you probably will encounter some problems since we freely use
the quickwin library of this compiler to write program messages. In this
case you should outcomment the respective program statements.  

- Parameters.txt/Parameters_explained.txt

Interaction between the user and the program works via the ASCII-file
Parameters.txt. There, the user can input her values for the parameters of
the model and choose among many options that determine the solution method
and the output of the program. The parameters as well as the choice of
options are described in the file Parameters_explained.txt.

- Parameters_GE.txt/Parameters_US.txt

The parameter values for the German and US calibration are stored in these
two files.


- GSP1.txt/GSP2.txt

These two files store the settings for the search routines one and two, respectively.

The main output of the program will be directed to an ASCII file specified by
the user in Parameters.txt. Messages that the program produces during
run time are either printed on the sceen or directed to the file Logfile.txt.
If the program is instructed to compute Euler equation residuals the points
of the grid for the stock of capital are printed to a file named kt_??.txt,
where ?? is an abreviation of the method used:

 EP=extended path method
 GA=Galerkin projectin method
 LA=log-linear approximation
 PE=parameterized expectations
 SO=second order approximation
 VI=value function iteration

The points of the grid for the productivity shock are written to zt_??.txt and
the resiudals are written to emat_??.txt.

The value function iteration method outputs the policy function and the
grids for the stock of capital and the productivity shock in the binary files
vi_policy.bin, kgrid.bin, and zgrid.bin, respectively. These file can only be
read by the program.

The Galerkin projection method also produces a file named Impulse.txt containing
the impulse response to a one-time shock. The rows of this file show
 the number of the period (from 1 through 500), the
 level of the productivity shock, the value of the rhs of the Euler equation
 for capital, consumption, hours, output, and the capital stock.
If this file has less than 500 lines or if the any of the columns shows diverging rather than
converging numbers this indicates that the program has found a bad solution.

The program also writtes some other information on your hard disk. For instance, the
policy functions from the linear solution, which can be used to initialize the parameterized
expectations or the Galerkin projection algorithm, or the files that store the solution
for the coefficients of the polynomials used in the Galerkin projection.


