Navigation:  Setting Up a DSGE Model >

 

AiM Model File

 

Print this Topic Previous pageReturn to chapter overviewNext page

YADA mainly uses AiM as an engine for parsing and, optionally, solving DSGE models. The AiM software and algorithm is based on work by Anderson and Moore (1983, 1985). See also Adolfson, Laséen, Lindé, and Svensson (2008), Anderson (2010) and Zagaglia (2005). The QZ (generalized Schur) decomposition based model solving algorithms of Klein (2000) and Sims (2002) are also supported.

Alternatively, dynare can be used to parse a dynare model file, which is then translated into the AiM output files. Information about dynare model files are available from the dynare website and its manual. Note that the "aim_solver" option should be used when calling the "stoch_simul" function in dynare. YADA includes some examples of how a dynare model file can look like and be used by the dynare parser in YADA. The text below covers the native AiM case only, while dynare model files are described in detail in the dynare documentation.

The AiM model file is a text file that specifies the DSGE model in log-linearized form with all variable expressed in deviation from steady state. The syntax is simple with, for instance LEAD(X,h) reflecting expected value of a variable X h periods ahead and LAG(X,h) being the h-period lag of variable X. Currently, YADA allows for an arbitrary number of periods for the lead, but a maximum of 1 lag. The latter is not a restriction, but ensures that the dimension of the model is not unnecessarily enhanced.

Each equation is defined through 3 arguments: the name, type, and the formula. The example file "AnSchorfheideModel.aim" that comes with YADA includes a foreward-looking Phillips-curve through the commands:

EQUATION> EQ2Phillips

EQTYPE>   IMPOSED

EQ>       pihat = beta*LEAD(pihat,1) + kappa*yhat

                 - kappa*ghat

The variables in this case are pihat, yhat, and ghat, while beta and kappa are parameters. The latter parameter is estimated, while the former is a function of a parameter that is estimated. This means that beta is defined in the file with parameters to update, while kappa is defined in the prior distribution file.

The equations all have a name, specified to the right of the equation command. This name is used to identify via selection (see Set State Equations) which equations in the AiM model file are structural equations and which definitions of shocks and a constant.

The function Run AiM Parser that is located both on the toolbar and on the Actions menu runs the AiM parser software. If the AiM model file is correctly written the function will inform you about this and provide the necessary output files. Likewise it will report any errors it encounters.

The main advantage of using AiM is that it provides a very fast and efficient algorithm for solving a DSGE model. This statement remains true also for medium scale DSGE models, such as the New Area-Wide Model.

NOTE: AiM does not separate between variables and structural shocks. This is instead handled via YADA's functions for selecting state variables and state shocks.

 

Additional Information

A description about how the DSGE model is solved and written as a state equation with the three model solvers supported by YADA is found in Section 4 of the YADA Manual.
A more detailed description about how to setup the AiM model file can also be found in Section 18.1 of the YADA Manual.
An example file is located in the directory "example\AnSchorfheide" directly below the base directory for YADA. The file is simply called "AnSchorfheideModel.aim".
Further example files are located in the "LubikSchorfheide" and "SmetsWoutersAER" directories in the "example" directory.

 

 


Page url: http://www.texlips.net/yada/help/index.html?aim_model_file.htm