The SIMPLE Model

J.F. Francois, Tinbergen Institute and CEPR
email: Francois@few.eur.nl internet: http://www.Dismal-Science.com/francois


Setting up and running the model
The experiments
Viewing results
Technical notes and documentation
Files in the Simple directory

© J.F. Francois, 1998-2002
Related material and updates are available from: www.Dismal-Science.com/handbook
Dismal-Science.com™


Running the SIMPLE Model with WinGEM and GEMPACK:

Many of the CGE exercises here involve a compiled and modified version of the GTAP model, called SIMPLE. The model is based on the GEMPACK version of the core GTAP model with significant modifications. It has been compiled so as to run independently (for this 2x3 application only). If you place all of the files from this archive into a single directory and follow the directions below, you should be able to run the model without additional software. To modify the theory, you will need to obtain a licensed version of GEMACK. Model files are found in the subdirectory "simple" on the CD-ROM in the "course" directory, and installed under the "course" directory on the C: drive if the course material has been installed from the CD-ROM.

WinGEM: Because the SIMPLE model is compiled, you do not need a full version of GEMPACK. However, you do need to install the Windwos front-end for GEMPACK, known as WinGEM. WinGEM is a free utility, and can be installed from the course CD. Installing WinGEM will place a GEMPACK program group under the Windows start menu, and a WinGEM icon on your front desk. Once you have loaded WinGEM, you will see a WinGEM progam menu bar across the top of your screen.

Setting the directories: Once you have started WinGEM, you will need to set up the working directory. Running the SIMPLE model involves using the "RUN TG" option. From the menu, choose "Simulation: Run TG".

Your next step involves changing both the working and default directory to the one where the SIMPLE model files are kept. This is usually c:/course/simple.

Selecting the experiment file: Your next step involve selecting the model that you will be running. From the "Simulation: Run TG" menu, you need to select a model file.

Finally, you are ready to run the model. This involves selecting an experiment file. The experiments are contained in command files, all of which have the name *.cmf (as in EXP3.cmf). *.CMF FILES

You must run EXP0.CMF first: The experiment EXP0 sets up the standard closure (i.e. theoretical structure) employed by all of the other model files. You need to run this experiment before running any other experiments.

Viewing social accounting data: Note that the model data (including a file in SAM format) can be viewed with the utility "Viewhar", reachable from the WinGEM menu under HA Files.

 

 

An Overview of the Experiments

The experiments are all defined through command files. When running this model, model feautures are manipulated through a mix of policy shock "switches", paramater file modifications, and closure "switches."

A full list of the experiments is provided below:

EXP1.CMF

Global free trade under non-nested GTAP model

EXP2.CMF

Same experiment, with IRTS and AC pricing in manufacturing

EXP3.CMF

Same experiment, with large group monopolistic competition in sector 2

EXP4.CMF

Same experiment, with cournot behaviour and CRTS in manufacturing

EXP5.CMF

Same experiment, with cournot and IRTS in manufacturing

NOTE!! You must run INSERT1.CMF and INSERT2.CMF before EXP4,5,6,11

EXP6.CMF

Same model as EXP5.CMF, but with change in competition policy. Experiment EXP6 involves a change in Japanese competition policy, involving a shift in the allowed degree of collusion in manufacturing from relatively collusive (halfway between perfect competition and monopoly pricing) to something very close to perfect competition. The model structure is identical to that employed in EXP5.CMF.

EXP7.CMF

Global free trade under the standard (nested) GTAP model

EXP8.CMF

Same as EXP1, but with steady-state capital stock adjustment under classical savings behaviour

EXP9.CMF

Same as EXP1, but with steady-state capital stock adjustment under Ramsey-type savings behaviour EXP10.CMF Same as EXP1, but with rigid real wages EXP11.CMF Same as EXP5, but with steady-state capital stock adjustment under classical savings behaviour EXP12.CMF Same as EXP9, but with global capital markets. This is a very unrealitisic experiment (even more so than most other CGE experiments!!!) but does demonstrate the long-run implications of capital mobility. Global trade liberalization tends to boost the productivity of capital, in ROW and also in Japan. At the same time, there is an induced substitution away from savings in Japan. The result is a sizeable inflow of capital into Japan from the rest of the world (even as there is capital accumulation globally and in both regions.) The result is an improvement in the steady-state Japanese trade balance of approximately $100 billion, as Japan must pay for this capital. Of course, this also implies that, in the medium-term, the trade balance deteriorates as Japan initially draws down and consumes it holdings of global capital. This experiment is benchmarked from an assumption that in initial data, shares of global capital stocks are equal, in value, to installed domestic capital, AND that capital is a perfect substitute between regions for all investors. Of course, this is not really true, so that this closure may greatly overstate the likely capital market effects.

EXP0.CMF

This is not really an experiment, but rather a solving of the model to set-up default model closure (i.e. what is endogenous and what is exogenous.)

Note: Experiments 1,2,4,5,6 involve national product differentiation (the Armington assumption), while experiment 3 involves firm level product differentiation. For these experiments, the Armington structure is non-nested, to accomodate the theory in the model on cournot pricing (which involves non-nested Armington).

 

 

Viewing Results

Once you have run the model, you can view the resuls using VIEWSOL. WinGEM will give you the option to choose VIEWSOL.

VIEWSOL is a graphic utility for examining model results.

 

 

Technical notes on the SIMPLE model

Beyond the GTAP model, the model code provided with this file (in the file SIMPLE.TAB) also allows for monopolistically competitive industries (i.e. firm-level product differentiation), cournot competition under constant and increasing returns, contestable markets with scale economies and average cost pricing, steady-state adjustment to capital stocks under classical and Ramsey-type savings behaviour, steady-state closures with a crude specification of global capital markets, and short-run sticky wage closures.

The aggregation of this model is based on a 2 region, 3 sector version of the GTAP dataset. Much larger aggregations are possible, and many additional aggregations are available from the GTAP consortium from its web site at:

http://www.agecon.purdue.edu/gtap.

The GTAP database is constantly being revised. More current versions of the SIMPLE model files, for later GTAP database versions, are available upon request to francois@few.eur.nl .

Supporting documentation for the theory underlying this model is available from the following sources:

[1] J.F. Francois, B.McDonald and H. Nordström (1996). 'Liberalization and Capital Accumulation in the GTAP Model.' GTAP consortium technical paper number 7, 1996.

[2] J.F. Francois (1998). 'Scale Economies and Imperfect Competition in the GTAP Model." GTAP consortium technical paper number 14, 1998.

[3] Hertel, T.W. and Tsigas, M. (1996) "The Structure of GTAP," in T.W. Hertel, ed., Global Trade Analysis, Cambridge University Press, Chapter 2.

Theory for the core model (perfect competition, static, constant-returns-to-scale) and the underlying notation is explained by Hertel and Tsigas. The other two papers ([1] and [2]) focus on imperfect competition and capital accumulation.

 

Model Regions and Sectors

regions

sectors

Japan

Rest of World

primary production

manufactures

services

 

Running the Model in DOS
As an alternative, you can also run the model in DOS. To do this, you need to read the readme.lst file contained in the SIMPLE directory.

The Simple directory contents

  DOCUMENTATION FILES

README.1st

Documentation for DOS-based work with the model

MODEL FILES

SIMPLE.AXS

TABLO generated file for SIMPLE.EXE

SIMPLE.AXS

TABLO generated file for SIMPLE.EXE

SIMPLE.EXE

executable version of the model SIMPLE

SIMPLE.TAB

The theory, in GEMPACK code, of SIMPLE

SIMPLE.STI

For GEMPACK users, the condensation file

SIMPLE.MIN

Fortran/GEMPACK files for the compiled model

SIMPLE.MOD

Fortran/GEMPACK files for the compiled model

SIMPLE.LIB

Fortran/GEMPACK files for the compiled model

SIMPLE.INF

Fortran/GEMPACK files for the compiled model

SIMPLE.F9M

Fortran/GEMPACK files for the compiled model

DATASET/PARAMETER FILES

GDAT.HAR

the original aggregated GTAP dataset

GSET.HAR

the GTAP set definitions

GDAT.EX4

Dataset produced by INSERT1.CMF and used by EXP4.CMF

GDAT.EX5

Dataset produced by INSERT2.CMF and used by EXP5.CMF and EXP6.CMF

SMALL.TXT

GTAP database aggregation definition file

GPAR%.DAT

Parameter file for EXP%.CMF

EXP%.DAT

Additional data/paremeters for experiment EXP%.CMF

MARK-NO.DAT

The MARK#.DAT files are for definition of Cournot

MARK-MJ.DAT

sectors through a switch that turns on markups.

SAM.HAR

The data in SAM format.

SHOCK FILES

TMSHAT.SHK

Import tariff shock file for free trade

TOHAT.SHK

Ouput subsidy shock file for free trade

TXSHAT.SHK

Export tax shock file for free trade

OTHER FILES

INSERT1.CMF

Recalibration for Cournot competition, CRTS

INSERT2.CMF

Recalibration for Cournot competition, IRTS

CLEANUP.BAT

A simple utility for cleaning up after GEMPACK simulations. This cleanus up manyf of the files left over after a GEMPACK simulation.