The documentation on the MC2 model can be downloaded by going to the
appropriate document and downloading from there.
Data can be downloaded
by going to the output phase when you are working with the MC2 model
and downloading from there.
If you want to download the MC2 model for use on your own machine,
this can be done by downloading the Fair-Parke (FP) program and the MC2 model
datasets that go with the FP program. You should have a computer with
at least 128MB of memory (although 64MB might work---I am not sure).
One of the datasets that needs to be downloaded is about 12 MB in size,
and if you don't want to download this dataset over the
internet, we can send you a zip disk with the dataset on it. You will
need a zip drive to read the disk. If you want the disk, email
ray.fair@yale.edu. Please list the size and type of computer that you
have.
The following are the downloadable datasets:
FPDST.ZIP 422 KB
Contains FPDST.EXE, the FP compiled program
MC8.ZIP 853 KB Contains the following:
- MC8.SP6 Main input dataset
- MC8US.INP US input dataset
- MC8US.DAT US data read by MC8US.INP
- MC8EST.INP Read by MC8.SP6
- MC8SHR.INP Read by MC8.SP6
- MC8SHR1.INP Read by MC8.SP6
- MC8SHR2.INP Read by MC8.SP6
- MC8XX90.INP Read by MC8.SP6
- MC8VAR.INP Read by MC8.SP6
MC8FINT.ZIP 12,358 KB Contains MC8FINT.DAT: all the data
(1960.1-2003.4)
FPDSTCOD.ZIP 222 KB
Contains FP code if you want to compile yourself.
The following discusses how to work with the MC2 model once you have
downloaded it.
To work with the MC2 model, first run at the DOS prompt:
FPDST > OUTUS
INPUT FILE=MC8US.INP;
This creates MC8US.COF, which is read by MC8.SP6. (This only needs
to be done once.)
Then enter commands at the end of MC8.SP6 and then at the DOS prompt enter:
FPDST > OUT
INPUT FILE=MC8.SP6;
When the job is done, examine OUT.
You need to know how to use the FP program in order to work with the
MC2 model. You may want to download the FP test examples from the website
and run these first in order to get familiar with the program.
(See The Fair-Parke Program.)
You may
also want to download an executable version of the FP program that uses
less memory than does FPDST.EXE. The program loads faster the less memory
it uses. FPDST.EXE is only needed when working with the MC2 model.
There are many comments in MC8.SP6 and in the .INP datasets that it
calls. You should read these carefully to make sure you know what is going
on. You should note the following:
- The data in MC8.SP6 begin in 1960.1, but the US model is estimated
beginning in 1954.1. This is why MC8US.INP is needed. The US model must
be estimated using this dataset and the estimates written to MC8US.COF,
which is then read by MC8.SP6.
- A number of "tricks" have to be used in MC8.SP6 to link the quarterly
and annual data. These are noted in the dataset. The annual data are
stored in the first quarter, with the remaining three quarters having
"missing" values. A lag of 4 for an annual country is a lag of one year,
not four years. All the estimation works for the annual countries
except when there is serial correlation of the residuals. For example,
RHO=8 means second order serial correlation for the annual countries,
but the estimation routine cannot handle this. The coefficient estimates
must be entered using the COEF; command, which is noted in MC8.SP6.
Solution, however, is fine.
- There are two sets of trade share variables (the alphas): A---- and
AA----. The AA----'s are the ones predicted by the trade share equations,
and the A----'s are the ones that are constrained to sum to 1. The
trade flow variables are denoted X----.
- A solution period must begin in the first quarter of a year and end in
the fourth quarter of a year. Otherwise the tricks of linking the
quarterly and annual data do not work.
- When working with the MC2 model (LA(344)=1), the FP program orders
the equations as the EQ commands are read in regardless of what equation
number is on the EQ command itself. For estimation, you must use the
number that the FP program has assigned to the equation. This is
also true of the MODEQ commands, which are used to specify the first
stage regressors. The PRINTMODEL command allows you to see the
numbering of the equations.
- Note the different treatment before and after 1999.1 regarding the EMU.
This is explained in comments in MC8.SP6.
- Once you learn how to work with the MC2 model in the FP program, all the
FP commands are at your disposal, and so much experimentation can be done.
You can add variables without worrying where they come in the list of
variables. However, if you add equations, you must know where the program
has put them (i.e., in the order in which they are read). If you put
an equation other than at the end, this will foul up the estimation and
MODEQ commands, which use the order that now exists. Unless you want to
change these commands, it is best that you add new equations at the end
of the existing equations. If you want to drop an equation without
affecting the order, use the command:
EXOGENOUS VARIABLE="name of left hand side variable" DROP;
This means that the equation will never be used, but it still remains in the
list and so the order is still the same.
|