FRB/US model packages
The main FRB/US model package is a self-contained set of equations, data, programs and documentation that enables various types of simulations and provides information about the model's structure. The package contains the following files and subdirectories:
- A readme file with basic information about the organization of the package and how to get started;
- The subdirectory addins with EViews add-in commands that must be installed, following the instructions provided in the readme file;
- The subdirectory data containing the EViews database used in simulating the model as described in the readme file;
- The subdirectory mods containing model equations and parameter values;
- The subdirectory subs containing two libraries of subroutines, including the library mce_solve used for solving FRB/US under model-consistent (rational) expectations in EViews; the RE Solver Package is available for those wishing to download the mce_solve software and documentation by itself;
- The subdirectory programs containing some basic applications of the model including those described in the article "The FRB/US Model: A Tool for Macroeconomic Policy Analysis"
- The subdirectory documentation among whose documents are one with basic instructions for simulating FRB/US and another with a description of all of the model's equations.
FRB/US model package (ZIP) (Updated database: September 21, 2016)
The following zip file contains only the historical FRB/US database (csv format) and variable list.
FRB/US dataset and variable listing (ZIP) (Updated database: September 21, 2016)
The program frbus_supply.prg estimates a multivariate state-space model that forms the basis of the FRB/US specification of potential output and its components. The zip file below contains the code and the latest data set used in estimation.
The program frbus_supply.prg estimates a multivariate state-space model that forms the basis of the FRB/US specification of potential output and its components. The zip file below contains the code and the latest data set used in estimation. The zip file also includes a note (latent_note.pdf) that provides documentation of the state-space model. The model is closely related to one presented in Charles Fleischman and John Roberts, "From Many Series, One Cycle: Improved Estimates of the Business Cycle from a Multivariate Unobserved Components Model, " FEDS Working Paper 2011-46.
FRB/US supply-side package (ZIP) (Updated database: December 30, 2015)
The software library mce_solve provides code for the solution of linear and non-linear models under model-consistent (MC) or “rational” (RE) expectations in EViews. This code is more reliable and efficient than the RE algorithm built into EViews (Fair-Taylor) at solving FRB/US when any of its expectations are MC. The mce_solve library includes two RE algorithms: E-Newton and E-QNewton. These algorithms iterate to find a model's RE solution with a sequence of updates to either exogenous estimates of the model's future-dated endogenous variables or exogenous components of such estimates. For single simulations of linear RE models, E-Newton is likely to be faster for models of small-to-medium size and E-QNewton is likely to be faster for larger models. For nonlinear models, E-Newton tends to be penalized relative to E-QNewton. The E-Newton algorithm has a substantial advantage over E-QNewton on experiments that involve a large number of RE solutions, as long as the same expectations Jacobian can be used for each E-Newton solution.
The solution algorithms are described in detail in Flint Brayton,"Two Practical Algorithms for Solving Rational Expectations Models,” FEDS Working Paper 2011-44, and in the documentation included in the zip file. The zip file also contains some sample programs that illustrate how to use the algorithms.
NOTE: The programs for simulating the FRB/US model are written for use with the software EViews, available at www.EViews.com . The current version of FRB/US is compatible with EViews versions 7 and 8.