Below is a listing of errors found in the RATS Version 8 manuals. Note that these are fixed as they are found in the PDF manuals, so if you have a PDF shipped with a later 8.x, these may already be corrected.

For errors in earlier documentation, see Version 7 Errata, Version 6 Errata, or Version 5 Errata.

Reference Manual

  • CROSS Instruction, page 64:
    The example in the description of the RESULTS option should refer to the FROM and TO options, not "first" and "last".

  • DATA Instruction, page 78:
    In the list of formats supported for the FORMAT option, the text should read DTA rather than DATA. DTA is the format for Stata(r) files. Corrected in the PDF version of the manual that shipped with version 8.0.

  • ESTIMATE Instruction, page 161:
    The Variables Defined list should have %XX, rather than %X, as the variable containing the inv(X'X) matrix. Corrected in the PDF version of the manual that shipped with version 8.0.

  • PANEL Instruction, page 353:
    The ICOUNT and TCOUNT options are not documented properly. ICOUNT returns the number of valid observations in the current individual (that is, for y(it) the count of valid observations across all t for individual i). TCOUNT returns the number of valid observations for the current time period (that is, for y(it) the count of valid observations across all i at time period t). Note that this is in keeping with the defintions of the ISUM and TSUM operations--that is, both ISUM and ICOUNT operate across individuals, while TSUM and TCOUNT operate across time periods). Fixed after the release of 8.1.

  • STWISE Instruction, page 456:
    Lists the "general to specific" option as being a switch option: 
    GTOS/[NOGTOS]
    This is incorrect--GTOS is actually one of choices available for the METHOD option, not a stand-alone option, so GTOS is selected using the option: 
    METHOD=GTOS

  • X11 Instruction, page 513:
    The seasaonal factors produced by the FACTORS option are not forecasted out three years beyond the adjustment period. That was true in versions 7.1 and older, but is not part of the new X12-ARIMA methodology introduced with version 7.2. Instead, use the EXTENSION option with a forecasted input series to produce forecasted seasonal factors.

  • Functions
    The following functions were available in 8.0, but were not listed in Section 2 of the Reference Manual. See the Function Wizard or Help in RATS, or the revised PDF version of the Reference Manual for details on these:

    • %CHOICE, which enables specifying a "choice" option argument by supplying the choice as a text label. (Added in the 8.1 PDF)

    • %EQNRESID and %EQNRVALUE, for extracting a residual value from an equation. (Fixed in the 8.0 PDF).

    • %FILE returns an i/o unit from a supplied file name string. (Added in the 8.1 PDF)

    • %GRPARM was not listed in Section 2, but was documented in the GRPARM section. (Added in the 8.1 PDF)

    • The %LOGGEVDENSITY and %LOGGPDENSITY functions were documented in the detailed alphabetical listing, but were not listed in the summary list of Distribution/Probability functions at the beginning of Section 2. This was corrected in the version 8.o PDF.

User's Guide

  • Page 243, Cointegration and Error Correction Models
    The Pi symbol is missing in the second line of the second paragraph under equation (33). This should read:

    "...reduced rank assumption on the Pi term."

  • Page 291, McLeod-Li Test Reference
    There is a typo on the year reference for the McLeod-Li paper. It should read 1983, rather than 1993. The entry in the bibliography is correct. The incorrect 1993 date also appeared in the comments section of the mcleodli.src procedure file.

  • Page 297, BEK GARCH Model Formula
    In Equation 16 on page 297, the u and u' terms should reference period (t-1), not period (t). Corrected after the release of 8.1.

  • Page 311, Kalman Filtering, Equation 3
    There is a typo on the right-hand side of equation number 3 on page 311. Instead of:

    A(t)Sigma(t|t-1)A't

    it should read

    A(t)Sigma(t-1|t-1)A't

    (note "t-1" instead of "t"). Corrected after the release of 8.1.

  • Page 354, Threshold Autoregression Models
    On page UG-354, the first paragraph should say "large positive" not "large negative". Corrected in the PDF version of the manual that shipped with version 8.0.