This page describes the changes made to RATS in Version 6.0 through 6.35, in reverse chronological order.

See RATS: Latest Release for details on the current release, RATS Updates for information about updating your RATS software or RATS Information and Features for general information on RATS.

Version 6.35

Version 6.35 began shipping in April, 2007 for Windows, and in May, 2007 for Macintosh, UNIX, and Linux platforms. The major improvements in 6.35 were as follows:

  • Significantly expanded support for State Space models, thanks to major additions to the DLM (dynamic linear modelling) instruction.
  • New SWEEP instruction: a multi-purpose tool for regressing one set of variables on another. Useful for "sweeping" out the effects of short-run variables in cointegration analysis, or dummies in other types of regressions. You can also allow a regression to vary across groups of data, building up the residuals or fitted values from the heterogeneous regressions.
  • Improved ODBC/SQL support (Professional Version only)
  • Graphing enhancements, including new options for controlling the frame around graphs, a new "spike" style, the ability to add a footer label to graphs, and the ability to nest SPGRAPH graph blocks.
  • A new format for the CALENDAR instruction that is easier to use and to read (the old formats still work).
  • Many other new options and more than a dozen new functions.

For a complete list of all the changes and improvements, see the April 2007 RATSLetter.

Version 6.3

See October 2006 RATSLetter for details.

The most significant news surrounding the release of RATS 6.3 relates to our Macintosh and UNIX/Linux products:

  • MacRATS 6.3 was the first release written specifically for the OS X operating system. It is a Universal application that runs on both Intel-based and PowerPC-based Macs.

  • UNIX RATS 6.3 brought our interactive RATS Editor Interface to the UNIX and Linux platforms for the first time ever. This interactive mode environment makes it much easier to do quick analyses using our menu-driven wizards, while also facilitating creating, modifying, and easily re-executing complete programs. It also makes working with graphics much easier. Batch mode is also still supported.

  • CATS 2.0 for Macintosh, UNIX, and Linux: The release of RATS 6.3 for Mac and UNIX/Linux also allows us to bring CATS 2.0 to those platforms for the first time. See CATS Version 2 for details.

Other Features

Other improvements appearing in 6.3 (on all platforms) included

  • New menu-driven Wizard for Panel Data Regressions

  • Pro versions now include ODBC/SQL support, and support for reading CRSP data.

  • DECLARE can now be used in a loop, procedure, or other compiled section. This makes putting existing code into a loop or a procedure. Also LOCAL commands can now be used anywhere in a compiled section (rather than just at the beginning), and can now include dimension fields for arrays.

  • New options on BOOT for doing block bootstrapping.

  • New KALMAN options for doing "dummy observation" priors.

  • FILTER now has an option for removing trends or seasons via regression.

Version 6.2

See March 2006 RATSLetter for details.

  • Support for Version 2.0 of CATS.
  • GARCH improvements: Support for IGARCH (univariate and multivariate); extends GARCH-M support to multivariate models.
  • New SHEET option on DATA for reading any worksheet in an Excel workbook (previously could only read the first sheet).
  • New menu-driven Wizards for VAR models and for accessing reserved variables.
  • Option for saving partial derivatives on NLSYSTEM.
  • Many new functions
  • Revised versions of the User's Guide and Reference Manual (provided in PDF format for updates from 6.0, hard copies also available).

Version 6.1

See July 2005 RATSLetter for details.

  • Five new menu-driven Wizards: for generating dummies; differencing series; filtering and smoothing series; Box-Jenkins ARIMA estimation; and exponential smoothing.
  • GARCH improvements: Support for additional regressions in multivariate models; support for asymmetry in multivariate models (previously only for univariate); option for selecting form of variances for diagonal, CC, and DCC models.
  • BOXJENK: new option for doing maximum-likelihood rather than conditional least squares, and support for estimation methods other than Gauss-Newton.
  • FORECAST: new option for saving standard errors and for doing static, rather than dynamic, forecasts (used to require a separate ERRORS command and using STEPS, respectively).
  • New filtering options on FILTER.
  • NLLS, NLSYSTEM: New option for providing a formula for the Jacobian, simplifying Full-Information Maximum Likelihood estimation
  • ESTIMATE (for estimating VARs) now offers a MODEL option, making it easier to work with multiple VAR specifications.
  • Thirteen new functions.

Version 6.02

See February 2005 RATSLetter for details.

  • Many new options on various instructions—see the newsletter for details.
  • Fifteen new built-in functions.
  • Two new menu-driven Wizards: for GARCH estimation; and for limited/discrete dependent variables estimation.

Version 6.01

See August 2004 RATSLetter for details.

  • New QZ instruction for doing QZ decomposition, used in calculating generalized eigenvalues.
  • LQPROG: Linear/quadratic programming now handles "greater than" constraints as well as "less than" constraints.
  • A menu-driven Wizards for the X11 seasonal adjustment process (Pro version only).
  • Several new functions.

Version 6.0

Version 6 was a major upgrade, offering dozens of improvements and new features. We list some of the highlights below, but for complete details, please see the March RATSLetter.

New GARCH Instruction

Built-in GARCH instruction makes estimating ARCH and GARCH models easier and faster than before. Options make it easy to select the distribution (Normal, t, or GED), specify an exponential model, or include asymmetry terms. In addition to the standard multivariate model, you can choose BEKK, Diagonal, Constant Correlation, or Dynamic Conditional Correlations.

More Menu-Driven Wizards

You can now read data, do transformations, estimate regressions, do hypothesis tests, display graphs, and more via a simple point-and-click interface.

New Estimators

New DDV and LDV instructions provide expanded support for discrete dependent variable and limited dependent variable models.

Interface Enhancements

In addition to the menu-driven Wizards mentioned above, RATS 6 includes several other interface enhancements, including more toolbar icons, full Undo/Redo capabilities, and the ability to view, edit, and graph any series in memory using the new "Series List" window and associated toolbar icons. A "most recently used files list" makes it easy to re-open recent files, and new preference settings make it easier to control the start-up directory and automatically load procedures. Finally, we have integrated most of the functionality of the stand-alone RATSDATA program directly into RATS.

Other Changes

Version 6 features many other improvements, including several other new instructions, more than 70 new functions, additions and improvements to nearly 30 existing instructions, and revised and expanded versions of the User's Guide and Reference Manual.

Still Using 5.0 or Older?

For those of you using Version 5.0, note that we introduced several other new instructions in Versions 5.10 and 5.11, including:

  • DBOX: Allows user-definable dialog boxes with controls such as check boxes, radio buttons and popup boxes.

  • FIXED: Sets tables of fixed values within a procedure.

  • FUNCTION: Allows user-definable functions which mimic the pre-defined RATS functions.

  • GSET: Does element wise operations on series whose elements are matrices.

  • NPREG: Performs non-parametric regressions using Nadaraya-Watson (kernel) techniques or Lowess.

  • REPORT: Generates flexible reports in table format.

  • RLS: Estimates by recursive least squares, generating recursive residuals and related statistics.

  • RREG: Estimates by robust regression.

See RATS Version 4 and 5 for more on the changes made in these older versions.