RATS 8 Additional Topics

RATS VERSION 8
ADDITIONAL TOPICS
Additional Topics
This PDF supplement provides additional information on various aspects of using
rats, including more information on the rats interface, details on the supported
data file formats, coverage of some additional topics in econometrics, documentation
on several little-used instructions no longer included in the Reference Manual, and
more.
The RATS Interface

Data Formats
Functions
Topics in Econometrics
Deprecated Instructions
Error Messages
Errata
Table of Contents
1. The RATS Interface 5
1.1 Menu Operations...........................................................................................................6
1.2 Toolbar Icons...............................................................................................................17
1.3 Preference Settings: Customizing the RATS Editor....................................................19
1.4 Using RATS in Batch Mode.........................................................................................23
2. Data Formats 27
2.1 FORMAT=BINARY (Native binary format)..................................................................28
2.2 FORMAT=CDF (Comma delimited text formats).........................................................29
2.3 FORMAT=CITIBASE/FORMAT=DRI...........................................................................30
2.4 FORMAT=CRSP (CRSP Files)...................................................................................31
2.5 FORMAT=DBF (dBase Files)......................................................................................32
2.6 FORMAT=DIF (Data Interchange Format)..................................................................33
2.7 FORMAT=DTA (Stata Native Format).........................................................................34
2.8 FORMAT=FAME (FAME Databases)..........................................................................35
2.9 FORMAT=FORTRAN string........................................................................................37
2.10 FORMAT=FREE (Delimited text format, no names/dates)........................................41
2.11 FORMAT=HAVER (Haver Analytics native format)...................................................42
2.12 FORMAT=HTML (HTML table)..................................................................................43
2.13 FORMAT=MATLAB (MATLAB binary format)............................................................44
2.14 FORMAT=ODBC (ODBC/SQL databases)...............................................................46
2.15 FORMAT=PORTABLE (RATS Portable Format).......................................................48
2.16 FORMAT=PRN (Delimited text formats)...................................................................49
2.17 FORMAT=RATS (RATS data format)........................................................................50
2.18 FORMAT=TEX (TeX table)........................................................................................51
2.19 FORMAT=TSD (Tab separated text formats)............................................................53
2.20 FORMAT=WF1 (EViews workfile format)..................................................................54
2.21 FORMAT=WKS (Lotus WKS format)........................................................................55
2.22 FORMAT=XLS/XLSX (Excel workbook)....................................................................56
3. More on Functions 57
3.1 The Sweep Family of Functions..................................................................................58
3.2 Working With Polynomials..........................................................................................60
4. Statistical Techniques 61
4.1 Polynomial Distributed Lags........................................................................................62
Example 4.1 Polynomial Distributed Lags................................................................63
Additional Topics
4.2 Ridge and Mixed Estimators.......................................................................................64

Example 4.2 Shillers Smoothness Prior..................................................................66
4.3 Robust Estimation by Iterated Weight Least Squares.................................................67
Example 4.3 Robust Estimation...............................................................................68
4.4 Miscellaneous Regression Topics...............................................................................69
4.5 Band Spectrum Regression........................................................................................70
Example 4.4 Deseasonalization Using Spectral Techniques...................................71
5. Instructions 73
5.1 Deprecated Options and Parameters..........................................................................74
5.2 Deprecated Instructions..............................................................................................77
CNTRL Tools for Controlling Interactive Procedures...........................................78
DEFAULT Changing Default Options...................................................................79
DELETE Removing Series from a RATS Data File.............................................80
EDIT Screen Data Editing....................................................................................81
INCLUDE Adding Series to a RATS Data File.....................................................83
RENAME Renaming Data File Series.................................................................85
UPDATE Altering Data on a RATS Data File.......................................................86
6. RATS Error Messages 87

6.1 RATS Error Messages and Troubleshooting...............................................................88
7. Manual Errata 99
7.1 User's Guide..............................................................................................................100
7.2 Reference Manual.....................................................................................................100
4 RATS: Additional Topics

1. The RATS Interface
T his section includes information on the menu operations and toolbar icons avail-
able in rats, tells you how to use the FilePreferences operation to customize the
way rats behaves, and describes how to use rats in batch mode.
Menu Operations
Toolbar Icons
Preference Settings
Batch Mode Operation
Additional Topics
1.1 Menu Operations

The rats Editor provides eight menus: File, Edit, View, Data/Graphics, Statistics,
Time Series, Window, and Help. Most of these are very similar to those available in
many other Windows applications. We describe the menu operations briefly below.
The on-line help provides additional details on many of the menu operations.
The File Menu

The File menu provides standard file-handling and related operations.
New
opens a new text window, or a new rats format data file window.
Open...
opens an existing file. Using the file-type field in the dialog box, youc an open text
files (such as rats program and procedure files), rats graph files, or rats format
data files. For text files, the new window will only be flagged as the input window
if there are no other open text windows.
Recent Files
provides a list of the most recently used files. The list is preserved from one session
to the next, so you can easily reopen files you were using in a previous session.
Close
closes the active window. rats will give you a chance to save any changes before
closing the window. If you close the input window, rats does a Clear Memory op-
eration (see next page).
Merge...
inserts the contents of an existing file into the active window.
Save
saves the entire contents of the active window to disk using the current file name
(Save will prompt you for a file name if the file hasnt been saved previously).
Save As...
saves the contents of the active window to disk, using a filename you select. If you
have text selected in the window, Save As will ask if you want to save the selected
block, or the entire window.
Export...
exports the contents of the active window to a file. This operation is supported for
most windows other than text windows (where you just use FileSave As...). Export
allows you to choose from a variety of file formats.
Page Setup...
selects a printer, and allows you to set printerspecific options.

Additional Topics
Print
prints the contents of the active text or graph window. To print just a portion of a
text window, select the desired lines before doing Print.
Directory...
lets you change the default drive and directory for the current session.
Preferences...
customizes the behavior of the rats editor. See page page 19 for details. On Macintosh
versions, this operation is located on the rats menu.
Batch Mode
puts rats into batch mode. See page page 23.
Clear Memory
clears the current data and settings (all data series, other variables, CALENDAR and
ALLOCATE settings, etc.) from memory. Use Clear Memory (or the icon) when
you want to enter a new program, or re-execute a program from the beginning. This
does not delete any text, close any windows, or close open data files.
Exit
closes all windows and exits the rats program. rats will give you an opportunity
to save any changes before quitting. On Macintosh versions, this operation is called
Quit and is located on the rats menu.
RATS: Additional Topics 7

Additional Topics
The Edit Menu

The Edit menu provides cut and paste and other editing functions. See the on-line
help for further details.
Undo
undoes the recent changes. For example, if you have Cut a block of text, Undo will
replace the text. Undo also works with data editing windows. You can undo up
to the twenty most recent changes in each window. Note that Undo never reverses
calculationsit applies only to editing operations.
Redo
redoes the results of the most recent Undo.
Cut
for a text window, this deletes the selected text and copies it to the Clipboard. This
can also be used to delete series from a rats format file when applied to a Datafile
List window (the deleted series can be pasted to another file if desired).
Copy, Copy Special..., Copy as TeX

copies the selected information to the clipboard for later pasting, but does not
delete it from the window. You can use Copy to copy text, graphs, or the contents
of a spreadsheet window, or to copy-and-paste series from one rats format file to
another. For some types of windows, the Copy Special... operation will be available
allowing you to copy information to the clipboard in a particular format. Currently,
this is limited to the ability to copy the contents of Report Windows as a TeX tabular
item (the menu operation will show Copy as TeX to indicate this).
Paste
copies the contents of the clipboard into the active window. Most commonly used for
copying text, you can also copy-and-paste (or cut-and-paste) series from one rats
format file to another using Datafile List windows. Note that you cannot paste
graphics into rats.
Delete
deletes the selected text (does not copy it to the clipboard).
Select All
selects every line in the window. To execute every instruction in the input window,
you can do EditSelect All (or click on the Select all icon) and hit <Enter> (or
click the Run icon).
Find...
searches for a string of text. You can search from the top of the window, or forward
or backward from the current cursor position. Turn on the Case Sensitive check
box for a case-sensitive search.

Additional Topics
Find Next
repeats the most recent Find... operation.
Replace...
does a find-and-replace operation, locating a specified string in the active window,
and replacing it with another string.
To Lower Case
converts the selected lines to all lower-case.
Format comments
Takes the selected line(s) and turns them into a block of comments (lines prefixed
with an * that rats will ignore rather than execute as commands). The lines will
be reformatted to be of approximately uniform length. You can change this length
with the Comment Length box on the Preferences dialog box.
Comment-Out Lines
This turns the selected lines into comment lines, that will be ignored by rats
when executing instructions. This is done by adding the * symbol to the beginning
of each line.
Uncomment Lines
The opposite of the Comment-Out Lines operation, this removes the * symbols from
the beginning of the selected lines.
Indent Lines
Indents the selected lines. You can change the width of the indenting with the
Indenting box on the Preferences dialog box.
Unindent Lines
Removes indenting from the selected lines. The amount of indenting that will be
removed is the indenting amount set via the Preferences operation.
Show Last Error

moves the cursor to the line (instruction) that caused the most recent error.

Additional Topics
The View Menu

The View menu allows you to view lists of the variables and series currently stored
in memory and quickly generate simple graphs and statistical tables for series dis-
played in the Series Window or a rats data file window. The first three operations
are only available for certain types of windows.
Reset List
When a rats format file window is active, this allows you to change which series
are listed, and whether the series are sorted by name.
Show as B&W/Show as Color

When a Graph Window is active, you can use these operations to change the graph
from color mode to black and white mode or vice versa.
Change Layout
allows you to customize the appearance of the window. Only applies to some types
of windows.
Series Window
brings up a window showing all the series currently in memory. From this window,
you can use the other View menu operations or the toolbar icons to graph the series,
display basic statistics, export series, edit the series, and more.
Standard Functions
opens the Function Lookup dialog box, which you can use to browse through the built-
in functions, check the function syntax, and paste a function into the input window.
Standard Variables
displays a list all of the reserved variables defined by various rats instructions.
You can view an alphabetical list of all the variables, or list variables by category.
User Variables
displays a list of all the variables defined by the user in the current session.
All Symbols
displays a list of all global variable and function names, including reserved variables
and functions defined by rats as well as user-defined variables.
The following are available when the Series Window or a rats format file window
is active. These serve the same function as the corresponding toolbar icons (see page
page 17). The last two operations apply only for a list of series in memory.
Time Series Graph

produces a time series graph of the selected series.

Additional Topics
Histogram
produces a histogram plot of the selected series (can only be applied to one series
at a time).
Box Plot
produces box plots for the selected series.
Autocorrelations
produces a plot of the autocorrelations and partial autocorrelations for the series.
Statistics
produces basic descriptive statistics for the selected series
Covariance Matrix
produces a covariance\correlation matrix of the selected series
Data Table
displays the selected series as columns in a spreadsheet-style window, similar to
doing PRINT with the WINDOW option. You cannot edit series values using this data
viewdouble click on a series if you want to edit values in the series.

Additional Topics
The Data/Graphics Menu

Calendar
provides an easy way to enter the desired periodicity and starting date for the cur-
rent session. Note: If you will be using one of the Data Wizards, there is no need to
do the Calendar Wizard.
Data (RATS format)

The Data Wizards simplify the process of setting up your rats session, including
setting the CALENDAR and ALLOCATE commands, opening a data file, and reading in
data. If you will be working with a rats format file, first do FileOpen RATSData
to open the data file, select the series you want to read in from the file, and then do
the Data (RATS Format) Wizard to read in the data.
Data (Other formats)

If you will be working with any format other than rats format, just select the Data
(Other Formats) Wizard. From the dialog box, select the desired file format and open
the file you want to read.
Create Series (Data Editor)

opens a series editing window which you can use to create a new series in memory.
Trend/Seasonals/Dummies
generates trend series, seasonal and 0/1 dummies, or a series of draws from a ran-
dom normal distribution.
Transformations
provides a convenient way to do a variety of data transformations, from taking logs
or square roots to generating growth rates, and more.
Differencing
allows you to difference series using regular, seasonal, and/or fractional differences.
Filter/Smooth
implements several types of time-series filters.
Moving Window Statistics

extracts means, variances, fractiles, and extreme values from a moving window of
data.
X11 (Pro version)

implements the x11 seasonal adjustment process. Note: This feature is only sup-
ported in the Professional version of rats.
Graph Settings
This generates a GRPARM instruction, allowing you to set the font, style, and size
of labels used on graphs.

Additional Topics
Graph
generates time series graphs.
Scatter (X-Y) Graph

generates scatter plots.
Data Browsers: FRED Browser

accesses the online fred economic database maintained by the St. Louis Federal
Reserve (http://research.stlouisfed.org/fred2/), and displays a list of the main database
categories in a new window. You can double-click on categories and sub-categories
to drill down to a series list. You can download series to a rats format file (by drag-
ging and dropping onto a rats data file window), or bring the series into memory
by dragging and dropping it onto the Series Window.
Data Browsers: Haver DLX

This opens a Series List Window showing the contents of an already-open Haver
dlx database file. From this window, you can view/graph data, bring series into
memory, or export series to another file. Note: You must do OPEN HAVER data-
base, or have a database specified in the Data Sources section of the Preferences
dialog, prior to selecting this operation.
Data Browsers: FAME Database

This opens a dialog box you can use to open a fame database stored on your
computer or connect to a database stored on a remote server. After opening a
database, this displays a Series List Window showing the contents of the fame
database.

Additional Topics
The Statistics Menu

Univariate Statistics
generates descriptive statistics for a single series.
Covariance Matrix
computes a variance/covariance matrix for a set of series.
Linear Regressions
performs a variety of linear regressions, including ols, instrumental variables,
gmm, and ar1.
Limited/Discrete Dependent Variables

performs estimations involving limited and discrete
Dependent Variables
dependent variable models, such as probit and tobit.
Panel Data Regressions

performs panel data regressions, including fixed and random effects.
Recursive Least Squares

does recursive least squares regressions.
Regression Tests
does hypothesis testing on the most recently completed regression.
Nonparametric Regression
provides an interface to the NPREG instruction for non-parametric regressions.
(Kernel) Density Estimation

provides an interface to the DENSITY instruction for estimating the density func-
tion of a series.
Equation/FRML definition
defines an equation or formula. For formulas, you can define the list of nonlinear
parameters.

Additional Topics
The Time Series Menu

The Time Series menu offers operations for analyzing time series, including estima-
tion a variety of time series models, including arima and arch/garch models.
Autocorrelations
computes regular and partial autocorrelations of a series.
Cross Correlations
computes cross correlations and covariances for a pair of series.
Unit Root Tests

provides access to a number of unit root testing procedures.
ARCH/GARCH
estimates univariate and multivariate arch, garch, and related models.
Box-Jenkins (ARIMA) Models

estimates arima models.
Exponential Smoothing
implements exponential smoothing.
VAR (Setup/Estimate)
defines and estimates var models.
VAR (Forecast/Analyze)
generates forecasts, impulse responses, variance decompositions, and historical
decompositions for vars and other multi-equation models.
CATS Cointegration
If you have version 2 of the cats cointegration analysis procedure (sold separately),
you can use this operation to execute the procedure.
Single-Equation Forecasts
generates forecasts for a single equation.

Additional Topics
The Window Menu

The Window menu offers several operations for working with rats windows. Also, a
list of all open windows appears at the bottom of the Window menu. You can switch
to a window by selecting it from this list.
Tile Horizontal
tiles the open windows so they are all visible on the screen, using a horizontal
orientation (windows will be wider than they are tall).
Tile Vertical
tiles the open windows, using a vertical orientation (windows will generally be taller
than they are wide).
Cascade
stacks the open windows so the title bar of each window is visible.
Close All
closes all open windows. rats will give you a chance to save any changes to text
windows.
Keep Open on Close All

flags the top window so that it will remain open if you do WindowClose All.
Close All Graphs

closes all graph windows. You will not be asked if you want to save changes, so if
you want to save any of your graphs, be sure to do so before selecting this operation.
Close All Reports

closes all spreadsheet windows created using WINDOW options.
Use for Input

makes the active text window the input window.
Use for Output

makes the active text window the output window.
Paste AND Execute

This is a switch controlling how the other Wizards behave. When this is switch
is on, clicking OK in a Wizard dialog box automatically executes the instruction(s)
generated by the Wizard. If you would prefer that the Wizard only insert the in-
structions without executing them, select Paste AND Execute to turn off this setting.
Report Windows
This displays a list of the recently generated report windows. Selecting one of these
will display the report in a new Report Window. You can use this to re-display the
formatted output produced by instructions like REPORT, LINREG, TABLE, STATIS-
TICS, and many others. From the report windows, you can copy, paste, and export
information to spreadsheet programs and other applications.

Additional Topics
1.2 Toolbar Icons

The rats toolbars provide icons you can click on to accomplish a variety of tasks. The
toolbar appears at the top of the screen, just below the menu bar. Note that many of
the operations are also available via the View menu.
Text Editing Windows

The following icons appear in the toolbar when a text window is active. Some icons
are disabled (and appear dimmed) under certain conditions.
(Open) Shortcut for FileOpen. Opens a file.
(Save) Shortcut for FileSave. Saves the contents of the active window.
(Print) Shortcut for FilePrint. Prints the contents of the active window.
(Cut) Shortcut for EditCut. Removes the selected content from the
window and copies it to the clipboard.
(Copy) Shortcut for EditCopy. Copies the selected content to the clipboard.
(Paste) Shortcut for EditPaste. Pastes content from the clipboard into
the current window.
(Find) Shortcut for EditFind. Allows you to search for text.
(Functions) This opens the Function Wizard, which displays a list of all of the
built-in functions in rats. See Section 1.2 of the User's Guide.
(Input) This designates the active (front) window as the Input window.
It is equivalent to the WindowUse for Input operation. See page
Int-7 of the Introduction for more on the Input and Output icons.
(Output) This designates the active (front) window as the Output window.
It is equivalent to the WindowUse for Output operation.
, These set up split Input and Output windows, tiled horizontally
or vertically..
(Select A
ll) Shortcut for the EditSelect All operation. Selects all the text or
items in the active window.
(Run) Runs the selected instructions, or the instruction on the cursor
line, if any (equivalent to hitting <Enter>). Disabled if the active
window is not the input window. See page Int-7 of the Introduction
for more on the Run and Stop icons.
(Stop) While rats is processing instructions, the Run icon changes to
this Stop icon. You can click on this to halt the processing of the
instructions.

Additional Topics
(Ready/Local) Switches rats from Ready mode to Local mode. The icon will
change to Local/Ready ( ), indicating that rats is in Local
mode. Click the icon again to change back to Ready mode. You
can also hit <Ctrl>+L to switch modes. See page Int-7 of the
Introduction for details. This is omitted on Macintosh versions,
where you can just hit <Return> to insert a new line.
(Clear Mem.) This clears the memory. It is equivalent to FileClear Memory.
Graph Windows
When a graph window is active, rats will display the Save, Print, and Copy
icons described above, along with the additional icons for fixing and un-fixing the
graph proportions and for switching between color and grayscale mode. See page Int-
56 of the Introduction for details.
Series Window
See page Int-38 in the Introduction for details on the icons available when viewing
the Series Window (a list of the series stored in memory), which you can display by
selecting ViewSeries Window.
RATS Data File Windows

When you display a window listing the series stored in a rats format file, the same
icons available for the Series Window described on page Int-38 are available, except
for the Covariance Matrix and View Data buttons. The following additional but-
tons are also available (see Int-112 for additional information).
(Copy) Copies the selected series to the clipboard (can be pasted into
another rats format file).
, (Un/Redo) Undo and Redo operations.
(Rename) allows you to change the name of the selected series.

The stand-alone ratsdata utility program uses the same window format. You may
find it handy to use ratsdata for converting data to and from rats format files, or
for quickly creating or editing series or generating graphs (see the Help system in
ratsdata for details).
However, virtually all of the functionality of ratsdata has now been incorporated
in rats itself. So, in most cases you will probably find it easiest to just use the File
New and FileOpen operations in rats to do this kind of work.
Series Edit Windows

See page Int-14 in the Introduction for details on the icons available when editing a
series (either a series in memory or a series on a rats format file).

Additional Topics
1.3 Preference Settings: Customizing the RATS Editor

The Preferences... operation on the File menu (on the rats menu for Macintosh
users) lets you customize various aspects of the rats environment. Selecting the
Preferences operation brings up this dialog box:
You can use the Serial Number field at the top to store your rats serial num-
ber in the registry. This allows you to view the serial number at any time by doing
HelpAbout rats (ratsAbout rats on Macintosh). For convenience, you can also
store serial numbers for other rats products (such as the cats cointegration proce-
dures)just separate the serial numbers by semi-colon (;).
Note: in some cases, the installation program may be able to save the serial number
in the registry automatically.
The remaining fields are described on the following pages.
The Editor Tab

The first tab in the dialog box, labelled Editor, controls various settings affecting
the text editing features of the interactive mode interface.

Additional Topics
Open New Input Window at Start

This controls whether or not rats automatically opens up an empty text window,
designated as the input window, when you start the program.
Open New Output Window at Start
This controls whether or not rats automatically opens up an empty text window set
as the output window when you start the program. If Input Window is checked but
Output Window is not, one new window will open, set as both the input window and
the output window. If both boxes are checked, rats opens one input window and one
output window, tiled horizontally.
Open Series Window at Start
Turn this on if you want rats to automatically open the Series Window (displaying
the series in memory) when the program starts.
Buffer Text Output
Turn this off if you want rats to display each line of output as it is generated. Turn
it on if you want rats to save output in a buffer, and only update the screen peri-
odically, which allows it to execute much more quickly. You will want to leave this on
unless you need to see each line of output immediately as it is generated.
Editor Font Button
You can change the font and type size used in the rats editor by clicking on this but-
ton. This controls the appearance of text both on the screen and when printed. The
current font and size are shown in the dialog to the right of the button.
Indenting
This controls how far (in terms of number of characters) lines will be indented if you
use the Indent Lines operation on the Edit menu.
Comment Length
This sets the maximum line width (in number of characters) that will be used if you
do the Format Comments operation on the Edit menu.
The Directories Tab

The Directories tab offers the following fields:
Default Directory
The Default Directory field allows you to change the default startup directory. If you
select a directory here, and answer Yes when rats asks if you want to make the
changes permanent, that directory will be the default directory each time you start
the program. Note: use the FileDirectory operation if you just want to change the
default directory for the current session.
Procedure Directory
If you try to execute a procedure that rats has not yet compiled, it will automatically
search the current directory and the directory containing the rats executable file for
a .src file with a name matching that of the procedure. You can use this field to

Additional Topics
supply an additional search directory. So, if you have procedures you use regularly,
you may want to collect them in a specific directory, and enter the name of that direc-
tory in this field. See Section 15.2.1 of the Users Guide for details.
Procedure Library
The Procedure Library field allows you to specify a file of procedures that you want
rats to load automatically each time you start the program or clear the memory.
CATS Directory
If you have version 2.0 or later of our cats cointegration analysis procedure, enter
the directory containing your cats files here. This will allow you to use the CATS
Cointegration operation on the Statistics menu to launch the cats procedure.
Batch Output .ext
This field lets you specify the extension that is appended to output files when run-
ning in batch mode.
The Graphics Tab

The Graphics tab provides the following fields:
Fonts for Label Types
You can use this field to set the default font, size, and style for the various labels
available on graphs generated by rats. To change the settings for a label, select
(highlight) a label type and then click on the Change button. This displays a stan-
dard font selection dialog box which you can use to select the font.
Note that you can override these defaults using the GRPARM instruction or the Graph
Settings operation on the Wizards menu.
One Graph Window Only
If this is on, rats re-uses the same window each time you generate a new graph.
Otherwise, it displays each new graph in its own window.
The Data Sources Tab

This section contains settings relevant to some common sources of economic data.
FRED Key
This is a security key that grants you access (via the fred browser operation from
the Data/Graphics menu) to the online fred (federal reserve economic data)
database provided by the St. Louis Federal Reserve Bank. rats ships with a default
key that should work. If the Fed ever changes the key, you can use this field to enter
the new key. See research.stlouisfed.org/fred2/ for more information on the data-
base.
Haver Analytics Database
If you have files (such as the usecon database) from Haver in their dlx format, you
can set this field to point to the directory where the files are located. This allows you

Additional Topics
to use OPEN HAVER filename and DATA(FORMAT=HAVER) instructions to read the

data, without having to provide a path on the OPEN HAVER instruction.
CRSP Database
If you have crsp (Center for Research in Security Prices) data on your system, you
can set this field to the directory where those files are located. You can then read
data from the crsp database using DATA(FORMAT=CRSP). See www.crsp.com for
details.
Global Insight (Citibase/DRI Basics) Directory
If you have a copy of the Global Insight (formerly dri/citibase) Basics Econom-
ics database, you can use the Global Insight Directory field to tell rats where
these files are located so that you can read data from those files into rats using the
FORMAT=DRI option on the DATA instruction.

Additional Topics
1.4 Using RATS in Batch Mode

About Batch Mode
In batch mode, rats automatically reads instructions from a text file and stores
the resulting output in another (new) file. This is very convenient for running long
programs, because you can start rats and then leave the computer unattended while
it executes the program. You can create these programs using the rats editor or any
word-processing program that can save files as plain (unformatted) text. Similarly,
you can view the output with rats or any other word-processor.
Running Programs in Batch Mode

There are three ways to run winrats in batch mode (please see the ReadMe file
included with macrats and the unix/linux installation guide for details on running
in batch mode on those platforms).
One method is to start rats, put it into batch mode using the Batch Mode operation
on the File menu, and then drag and drop your program files from Windows Explorer
or a similar utility onto the active rats application. This is most useful if you have
several programs that you want to run.
Another method is to run from a command prompteither via the Run operation
on the Start menu, or by opening the Command Prompt shell. You simply type the
name of the rats application itself, followed by the name of the program you want to
execute and the command-line switch /RUN.
Finally, you can execute rats in batch mode from a short cut icon. This method is
most useful if you have a particular program that you need to run frequently.
In any case, rats will automatically run the specified program(s), and save the
output in a new file (or files) with the name filename.ext, where filename is the name
of the input file, and ext is the batch mode extension specified in the FilePrefer-
ences... operation in rats (the default extension is LIS). The file(s) will be saved in
the same directory as the program file(s).
Details on all three methods follow:
Drag and Drop Method

To run a program as a batch job using the drag and drop method, do the following:
1. Start rats.
2. Switch rats into batch mode by selecting the Batch Mode operation from the
File menu.
3. Open Windows Explorer and arrange the screen so that you can see both the
Winrats window and the Explorer window on the screen at the same time.
4. Select one or more program files from Explorer, drag the file icon(s) over to the
Winrats window, and release the mouse button.

Additional Topics
You can drag and drop the files one at a time, or you can drag and drop several files
simultaneously. Note that a check mark appears next to the Batch Mode operation on
the menu while rats is in batch mode. To switch back to interactive mode, just select
the Batch Mode operation again.
If you use batch mode frequently, you may want to create a copy of the Winrats icon
that will automatically start rats in batch mode. Heres how:
1. Make a copy of the Winrats program icon (see your Windows documentation or
help for details).
2. Right-click on the rats icon to display the properties dialog for the new icon.
3. Set the icon to Run minimized and click on OK.
Now, when you double click on this new icon, rats will automatically start up with
the Batch Mode switch turned on. You will need to click on the minimized icon to
open it up into a window. You can then drag and drop program files onto the window.
Command Prompt Method

If you are comfortable with using the Command Prompt shells in Windows, or are
used to using the older dos version in batch mode, you may find this method use-
ful. You simply need to open the Command Prompt window from within Windows,
and then type in the name of the rats executable file (ratsv8_std for Winrats,
ratsv8_pro or ratsv8_64 for 32-bit or 64-bit Winrats Professional), the name of
the input file, and the /RUN switch, separated by spaces. For example, the com-
mand line for Winrats might look like:
c:\winrats\ratsv8_std c:\test\myprog.rpf /run
This will run the job and save the output to MYPROG.LIS (unless you have used File
Preferences in Winrats to change the default output filename extension). You can
include commands like this in a dos batch (.bat) file if desired.
Short Cut Icon Method

There are several ways to create shortcut iconssee your Windows on-line help or
documentation for details. One method is to right-click on the desktop, and select
NewShortcut from the pop-up menu.
In the command line, enter the command to run the desired program, just as shown
under Command Prompt Method above.
Click on the Next button, enter a name for the icon, and click on the Finish but-
ton. This will create the batch job icon on your desktopyou can double-click on this
to run the job.

Additional Topics
Command Line Switches

rats offers several command line switches to control batch mode operationthese
can be used with both the Command Prompt and Shortcut Icon methods of run-
ning in batch mode:
/NOSHOWGRAPHS Suppresses display of graphs in batch mode.
/PRINTGRAPHS Prints graphs as they are generated.
/PLOT=filename Saves graphs to the specified plot file. Equivalent to putting an
OPEN PLOT instruction in your program.
/DATA=filename Opens the specified file as the DATA unit. Equivalent to an OPEN
DATA instruction.
/COPY=filename Opens the specified file as the COPY unit. Equivalent to an OPEN
COPY instruction.
/PROC=filename Designates filename as a Procedure Library filethe commands
in this file will automatically be processed before the specified
program file is executed.
To use a particular switch, just add it to the end of the command line (you can use
more than one switch if you wish). For example:
c:\winrats\ratsv8_std myprog.rpf /run /noshowgraphs /plot=graf.rgf
executes the instructions in MYPROG.RPF, saves the graphs in the file MYPLOT.RGF,
but does not display the graphs to the screen.

2. Data Formats
T his section describes all of the data file formats supported by rats, sorted by the
name of the FORMAT option used with the file format.
In each section, we provide a brief description of the file format itself, a list of the
rats instructions that can read or write that format, and information on interface
operations that can be used with the format. Where applicable, we provide additional
details and a short example.
For each format, the Interface Operations section dscribes which of the following
aspects of the rats interface can be used with the format in question:
Data Wizard
the Data Wizard operations on the Data/Graphics menu that you can use to read
data series into memory.
Series Window
a window showing all of the series in memory, which can also be used to read in
and export data. Use the ViewSeries Window operation to display this window.
RATS Data File Window

a window showing the contents of a rats format data file.
Report Window
a window display a rats Report, generated either by a REPORT instruction, a
WINDOW option on various other instructions, or by using the reloading a Report
Window using the Restore Report submenu on the Window menu.
Matrix Window
a window displaying the contents of an array. Opened using an MEDIT instruction
or by doing ViewAll Symbols and double-clicking on a matrix variable.
Additional Topics
2.1 FORMAT=BINARY (Native binary format)

Binary is the internal format used on a computer for representing numbers. It is
the fastest way of reading and writing data, but probably also the least useful, since
it isnt easily portable from machine to machine or program to program and has no
embedded description of the contents (10 columns with 100 rows and 1 column with
1000 look exactly the same). It is an Unlabeled format which can be used for input
and output of series and matrices.
RATS Instructions
data(format=binary) read series from binary into memory
store(convert=binary) copy series from binary to rats format
copy(format=binary) write series to binary file
prtdata(format=binary) write series from rats format file to binary
read(format=binary) read scalars and arrays from a binary file
write(format=binary) write scalars and arrays to binary file
ORG options are required on DATA, STORE, COPY and PRTDATA.
Interface Operations
None
Details
Binary format is the fastest way of reading and writing data, as it requires no trans-
lation from characters to numbers. It also preserves the internal representation of
all numbers. Speed, however, is the only real advantage it has over other formats. Its
disadvantages tend to outweigh this:
Because the data file is not in character form, it is very difficult to determine what
is on the file should you happen to forget.
It is impossible to change any number on the file without rewriting the entire file.
You cannot store dates or series labels on a binary file.
We dont recommend that you use BINARY format unless it's the only way available
to communicate data to another program (running on the same type of computer).
In reading binary data prepared using another program, you must be very careful to
read the data back in exactly the same way as it was written.
Real-valued data use a 64-bit representation. This is double precision, except on
machines with 64-bit single precision, like a dec Alpha.
With ORG=ROWS, rats reads (or writes) each series with a single call, so you
should use a separate fortran WRITE (READ) for each as well.
With ORG=COLS, each observation is read (written) with a single call; again you
should use a separate WRITE (READ) for each.

Additional Topics
2.2 FORMAT=CDF (Comma delimited text formats)

This is a text file with fields delimited with commas. Its a Labeled Table or Unla-
beled format (depending upon the file) for both input and output of series, reports
and matrices. For reading series, we recommend that you use the more general
FORMAT=PRN instead.
RATS Instructions
data(format=cdf) read series from cdf into memory
store(convert=cdf) copy series from cdf to rats format
copy(format=cdf) write series to cdf file
prtdata(format=cdf) write series from rats format file to cdf
read(format=cdf) read scalars or arrays from cdf into memory
write(format=cdf) write scalars or arrays from memory to cdf file
Data Wizard
You can read series using Data (Other Formats) on the Data/Graphics menu.
Series Window and RATS Data File Window
You can import series using FileImport... or the Import toolbar button. You can
export series using FileExport..., the Export toolbar button, or the Export operation
on the contextual menu.
Details
On DATA and STORE, you can, if required, use the NOLABELS option to skip automatic
label processing and use the LEFT and TOP options to isolate the data.
See Spreadsheet and Delimited Text Formats on page Int114 of the Introduction for more
details on the format.

Additional Topics
2.3 FORMAT=CITIBASE/FORMAT=DRI
This is a Time Series Database format available for series input, only on the Pro-
fessional versions of rats. It is the format used for the Basic Economics database
(formerly known as Citibase), which is maintained by Global Insight.
RATS Instructions
data(format=citibase) read from Basic Economics into memory
store(convert=citibase) copy series from Basic Econ. to rats format
None for reading or writing data. FilePreferences sets location of files.
Details
You must use the Preferences operation on the File menu to tell rats where the data
are located on your hard disk. With that done, you can read data with
data(format=citibase,other options) start end list of series
Because the series have dates, you can use this to select a sub-sample, or allow com-
pact or expand to a different frequency (see Changing Data Frequencies on page Int104 of
the Introduction).

Additional Topics
2.4 FORMAT=CRSP (CRSP Files)

This is available in the Pro version, but only under Windows and certain unix
platforms. You need a subscription to crsp (center for research in security prices)
data, available from the University of Chicagos Booth School of Business (www.crsp.
com). It is a Time Series Database format available for series input only.
RATS Instructions
data(format=crsp) read series from crsp into memory
None for reading or writing data.
You can use FilePreferences to set the default location of the crsp files.
Details
Your workstation must be configured with the proper set of environment variables
(CRSP_ROOT=, etc.) and you need to tell rats where the data are located. You can do
that by setting the crsp Directory setting in the FilePreferences... dialog box, or by
using an explicit OPEN CRSP instruction specifying the path in your program.
You can fetch the price, return, volume, bid and ask prices for stocks. These are done
by prefixing the crsp Permno for the stock with P for price, R for return, V for
volume, B for bid and A for ask. You can ask for any data frequency from daily to
annual, and youll get the aggregation for that frequency produced by crsp.
For example, the following will get the monthly returns for ibm (permno 12490) and
Microsoft (10107) over 1984:4 to 2004:12. Since Microsoft doesnt start trading until
1986, its returns will be nas until then.
open crsp c:\CRSP_Sample_Data\data\aiz\diz\
cal(m) 1984:4
all 2004:12
data(format=crsp) / r12490 r10107

Additional Topics
2.5 FORMAT=DBF (dBase Files)

This is a database format used by dBase and compatible programs. It's a Labeled
Table format for both input and output of series.
RATS Instructions
data(format=dbf) read series from dbf into memory
store(convert=dbf) copy series from dbf to rats format
copy(format=dbf) write series to dbf file
prtdata(format=dbf) write series from rats format file to dbf
Data Wizard
You can read series using Data (Other Formats) wizard on the Data/Graphics menu.
Details
The data are read from the file based upon the field (column) names. Each record is
treated as a separate observation. If the file includes a date field as the first field,
rats will be able to make use of the date information to select a sub-sample, recog-
nize skipped observations, or allow compacting or expanding to a different frequency
(Changing Data Frequencies on page Int104 of the Introduction). The dates can either use
the dbf date field format, which codes dates as year/month/day, or be entered as a
character field with the standard codings like year:period.
For writing data, each series that you write will be a field in the database. The dBase
specification allows for at most ten characters in a field name, so longer names will
be truncated. If you ask for dates, rats will either include a date field (for daily or
weekly data), or a character field with coded dates.
Example
This reads in several series of stock return data from a dbf file:
calendar(d) 1996:1:2
allocate 2009:12:31
open data stockdat.dbf
data(format=dbf)

Additional Topics
2.6 FORMAT=DIF (Data Interchange Format)

dif is a (rather bulky) text format used for transmitting the content of a simple
spreadsheet. Its a Labeled Table or Unlabeled format (depending upon the file) for
both input and output of series, reports and matrices, available in all versions of
rats.
RATS Instructions
data(format=dif) read series from dif into memory
store(convert=dif) copy series from dif to rats format
copy(format=dif) write series to dif file
prtdata(format=dif) write series from rats format file to dif
read(format=dif) read scalars or arrays from dif into memory
report(format=dif) (with ACTION=FORMAT) write a report to dif file
Data Wizard
Report Windows
You can write reports to a dif file using FileSave As, FileExport, EditCopy; the
Export and Copy toolbar items, and the Export and Copy contextual menu operations.
Matrix Windows
You can write the contents of a matrix window to a dif file using File-Save As, File-
Export; the Save and Export toolbars, and the Export contextual menu.
Details
For copy-and-pasting into dif, you will probably have to choose the Paste Special op-
eration in the target application. RATS copies the full precision of real numbers when
copying to the clipboard in dif format.

Additional Topics
2.7 FORMAT=DTA (Stata Native Format)

dta is the native format for Stata data files. Its a Labeled Table format available
for series input (only), on all rats versions.
RATS Instructions
data(format=dta) read series from Stata into memory
store(convert=dta) copy series from Stata into rats format
Data Wizard
You can import series using FileImport... or the Import toolbar button.
Details
Data in a dta file are stored as named series which will be read or imported using those
names. You can read all the series by leaving the series list blank on either DATA or
STORE, or you can request specific series by listing just the ones you want. If you import
directly into a data list, you will get all series.
rats will not recognize any dates on the file.
Example
open data pennxrate.dta
calendar(panelobs=34,a) 1970
data(format=dta) 1//1970:01 151//2003:01 year xrate ppp $
id capt realxrate lnrxrate oecd g7
reads 24 years of panel data for 151 countries.

Additional Topics
2.8 FORMAT=FAME (FAME Databases)

Fame is a time-series database management program provided by SunGard (see
www.sungard.com/fame). It is a Time Series Database format for input and output,
and is supported by the Professional versions of winrats and (on certain platforms)
unix. In addition to needing the proper version of rats, you also need a license for
the Fame software on the same platform.
RATS Instructions
data(format=fame) read series from fame into memory
store(convert=fame) copy series from fame to rats format
copy(format=fame) write series to fame file
On the Data/Graphics menu (for Pro versions), theres a Data BrowserFame Database
operation. This opens a Series Window showing the contents of the file. You can drag
and drop series from that into the Series Window or a RATS Data File window.
Details
Fame databases can be either local (accessed as standard operating system files) or
server. There are slight differences in how you use them.
Local databases are files with a .DB extension that you read like any other file:
cal(m) 1970:1
open data c:\fame\util\driecon.db
data(format=fame) 1970:1 2003:7 ipsb50001 cusa0
For server access, just use the base name of the database on the OPEN DATA, and in-
clude the additional option SERVER=server spec on the DATA or STORE instruction.
The server spec is a string with the port, followed by the server node, followed by
(if necessary) the user id and password. These are separated from each other by the |
character. For instance:
cal(q) 1980:1
open data driecon
data(format=fame,server="5555|booker") 1980:1 2010:2 gdp gicv
You can only add series to a local database. Use the APPEND option on the OPEN
COPY instruction to add to an existing file (rather than create a new one):
open(append) copy write.db
copy(format=fame) / x y z

Additional Topics
Fame Objects Supported by RATS

rats can read Fame objects with combinations of SERIES/FORMULA class, NUMERIC/
BOOLEAN/PRECISION type, and all frequencies except UNDEFINED. All series created
by rats are written into Fame files as objects with SERIES class and PRECISION
type. Missing values, which appear as NC, ND and NA in Fame data bases are trans-
lated into %NA in rats. Missing values in rats are converted to NA when exporting
to Fame data bases.
Symbolic Name Conversion Between RATS and FAME

rats and fame place different restrictions on symbolic variable names, so rats will
not accept some variable names which are legal in fame. To deal with this problem,
rats allows you to create a lookup tablea file containing a rats name and an as-
sociated fame name for one or more series.
The lookup table file must be an ASCII text file containing two columns of text
separated by at least one space. The first column on each line is a rats series name,
while the second column is the associated Fame name.
Using Lookup Tables

To use a lookup table, you must open the lookup file with an OPEN FAME instruction,
before executing your DATA or COPY instruction (FAME is a reserved I/O unit name).
On a DATA instruction, you must list only legal rats series names. Normally, rats
searches the Fame data base for the specified series. If you have opened a lookup
table, however, rats first looks for the series names in the lookup table. For names
that do appear in the table, rats searches the Fame data base using the associ-
ated Fame name, rather than the rats name. The series, if found, is then read into
memory using the rats name. For series not listed in the table, rats searches the
Fame data base using the rats name.
Similarly, when writing data using COPY, any series that appear in a lookup table
will be written using the Fame name, rather than the rats name.
A lookup table should be a text file with two columns. The rats name should be the
first item on each line, followed by the Fame name (separated by at least one space).
If the second column is omitted, the Fame name will be the same as the rats name
(equivalent to not including the name in the file). Anything after the second column
is ignored, so you can add comments after the Fame name if you like. The order of
series listed in a lookup table file is irrelevant. Both rats and Fame are case-in-
sensitive to series names, so mixed cases are legal in the lookup table. All lower case
characters are converted to upper case.

Additional Topics
2.9 FORMAT=FORTRAN string

This is an Unlabeled text format for input and output which uses fortran I/O
descriptors. Its available on all versions of rats. Its available in case other methods
for dealing with a poorly-formatted text file are inconvenient. However, we would
recommend you look first at importing the file into a spreadsheet program and using
its line parsing tools.
RATS Instructions
data(format='Fortran format string') read series using fortran format
copy(format='Fortran format string') write series using fortran format
read(format='Fortran format string') read data using fortran format
write(format='Fortran format string') write data using fortran format
ORG options are required on DATA and COPY.
None
Details
fortran format allows you to use the standard fortran I/O formatting codes to
indicate the format of the data file. You probably shouldnt try using it unless you
are fairly comfortable with fortran formats (described briefly below), and the other
methods above are cumbersome.
With fortran formats you can tell rats exactly how the data are organized when
using DATA or READ, or exactly how you want the data to be formatted when using
COPY or WRITE. For example, when used with COPY, FORMAT=FREE uses a very wide
field for each number so that it can reasonably handle very different magnitudes. The
result is that you can sometimes get a lot of extra digits. This takes up space (so you
see fewer numbers at a glance) and masses of digits can be hard to read.
If you are unfamiliar with fortran, a brief description is in order. The basic format
elements for real numbers take the forms Fw.d, Ew.d and Gw.d. rats (which only
simulates fortran I/O) also allows you to use the Iw format for input.
Fw.d For numbers with a fixed decimal place: a total of w positions
(total digits to the left and the right of the decimal) with d dig-
its to the right of the decimal. Very large and very small num-
bers will not fit an F format. For example, the number 1000000.
cannot be printed as F15.8.
Ew.d For numbers in scientific notation. When writing files, this for-
mat is useful because it can handle a number of any magnitude,
but it can be difficult to pick out large and small values at a

Additional Topics
single glance. For example, it takes a bit of work to see that the
first of 1.343E-02 and 8.547E-03 is the larger.
Gw.d This is a mixture of the F and E formats. If a number can rea-
sonably be printed with the F format, that is used. All others
(very large and very small) are displayed in the E format.
Iw For integer numbers. If there is a fractional part to the number,
it is ignored.
Aw This is used (by COPY) only for the date strings. It prints a char-
acter string, left justified, in a field of width w.
wX Indicates w spaces
/ Indicates a skip to the next line
In the description of a line, the fields are separated by commas. You can prefix a field
descriptor by a number of repetitions. For instance, F6.2,F6.2,F6.2 can be short-
ened to 3F6.2.
The FORMAT Option

The DATA, COPY, READ, and WRITE instructions all support the fortran format. The
FORMAT option is the same for all four instructions:
FORMAT="( format string )"
Enclose the format string in single quotes, for instance, FORMAT="(11X,4F15.7)".
The format string must fit on a single line. If you need an extremely long string, you
may want to put the string on its own line:
data(org=col,format= $
"(f8.5,2x,f8.4,2x,f8.5,2x,f8.3,2x,f8.5,2x,f8.5)" ) $
1947:1 1979:4 wage interest stocks mortgage charitable misc
With ORG=ROWS
Each entire series is read with the indicated format: below is a rats instruction and
its fortran language equivalent:
data(format="(11x,4f15.7)") 1 100 gnp m1
read(n,1000) (gnp(i),i=1,100)
read(n,1000) (m1(i) ,i=1,100)
1000 format(11x,4f15.7)
If you have blank lines separating the series, you may have to read the first series
with one format, then use a format such as (//(11X,4F15.7)) for the remaining
series: this particular format will skip two lines before reading a series.

Additional Topics
With ORG=COLS
The equivalent of a separate fortran read statement is used for each observation:
on the next page is a sample rats instruction and its fortran equivalent. (You
must delete any header lines at the top of the file).
data(org=col,format="(f9.4,2x,f10.3,2x,f9.4)") 1 100 gnp m1 ipd
do 100 i=1,100
read(n,1000) gnp(i),m1(i),ipd(i)
100 continue
1000 format(f9.4,2x,f10.3,2x,f9.4)
Mixed Character and Numeric Data

fortran format codes can also be useful if you have a text file containing both
numeric data and text labels. You cannot store character information in SERIES
variables, or use DATA to read in character/label information, but if the data are well
organized, it may be possible to read the numeric data into series and the text data
into LABEL variables. You will probably need to read the file twice: once to get the
numeric data and once to get the character data.
Suppose you have the following data file:
1.0450 ME 2.0 1.0
2.3210 MS 2.0 2.0
1.8930 MN 2.0 3.0
You could use the following program to read this data. First, a READ instruction using
a FORMAT option skip the numeric data and reads only the two-character label. The
data file is rewound (positioned to the top), and then the data are read using a DATA
instruction which skips over the character information to get the numbers. Finally, a
simple loop is used to display formatted output:
all 3
declare vect[labels] states(3)
open data mixed.dat
read(unit=data,format="(8x,a2)") states
rewind data
data(format="(f6.4,6x,2f6.1)",org=col) / x y z
do row=1,3
display states(row) @10 x(row) y(row) z(row)
end
Missing Data
You can use the codes (such as NA) for missing values used by other text files.
You can also leave a blank area where the missing values should be and use
BLANK=MISSING on the DATA instruction. This will interpret the blank area as a
missing value (as opposed to a standard practice of reading a zero).

Additional Topics
Using FORTRAN formats: An Example

The following data set was used in Section 2.9 of the Introduction. We show here how
to use fortran format to simplify the process of reading the data.
FIRM A
1996 11.3 11.6 10.9 12.3
1997 13.0 12.8 11.5 12.5
1998 12.9 13.0 13.2 13.6
FIRM B
1996 22.5 21.9 24.3 25.6
1997 21.9 21.8 22.6 23.5
1998 22.5 25.0 24.5 25.4
The data rows can be read with (12X,4F9.1). The problem is the other rows: there
is only one row preceding the first series (the row with FIRM A), but two (the blank
plus FIRM B) preceding the second. The simplest way to handle this is to add a
blank line at the beginning of the file and use
cal(q) 1996:1
all 1998:4
open data test.dat
data(format="(//(12x,4f9.1))",org=rows) / firma firmb
print / firma firmb
Note that we use PRINT to verify that the data has been read properly. How did we
come up with '(//(12X,4F9.1))'? First, the two slashes tell rats to skip two lines
before each block of data. Next, for a data block where the numbers are regularly
spaced (as they are here), just determine the position of the last digit in the first
data value on the line (column 21 in this case) and the field width of the data (here
9, which is the distance from the end of one number to the end of the next). The 12 in
12X is just the number of leading positions (219=12) to be skipped.
The 4F9.1 indicates that the fields are each nine characters wide (including lead-
ing blanks) with 1 digit after the decimal point, and that this format is repeated 4
times per row. After skipping two lines, rats will read data into FIRMA using the
12X,4F9.1 format until it has read the requested number of data points (12 in this
case). It then skips two more lines and again uses the 12X,4F9.1 format to read 12
data points into FIRMB.
An alternative if you either cant, or dont want to, alter the original file is to use
separate instructions so that you can apply different formats in succession:
data(format="(/(12x,4f9.1))",org=rows) / firma
data(format="(//(12x,4f9.1))",org=rows) / firmb

Additional Topics
2.10 FORMAT=FREE (Delimited text format, no names/dates)

This is a text file with fields delimited with spaces, tabs or commas that does not con-
tain series labels or date information (see FORMAT=PRN on page 49 for text files that do
include series names and, optionally, dates). Its an Unlabeled format for both input
and output of series, reports and matrices, available on all versions of rats.
RATS Instructions
data(format=free) read series from free into memory
store(convert=free) copy series from free to rats format
copy(format=free) write series to free file
prtdata(format=free) write series from rats format file to free
read(format=free) read scalars or arrays from free into memory
write(format=free) write scalars or arrays from memory to free file
Data Wizard
Details
See Text Files on page Int120 of the Introduction for details.

Additional Topics
2.11 FORMAT=HAVER (Haver Analytics native format)

This is a Time Series Database format for accessing data in the Haver Analytics na-
tive dlx format. It is available only in the Pro level, and only on Windows. It is for
series input. The Haver Analytics databases which are available for purchase from
Estima are provided in both rats and dlx formats.
RATS Instructions
data(format=haver) read series from Haver into memory
store(convert=haver) copy series from Haver to rats format
On the Data/Graphics menu (for WinRATS Pro), theres a Data BrowserHaver DLX
Database operation. This opens a window listing the contents of the file. You can drag
and drop series from that into the Series Window or a RATS Data File window.
Details
A dlx database consists of several files in the same directory with a common base name,
For instance, the usecon data base (a 12,000 series database with a broad range of
key U.S. macroeconomic series) comprises the files USECON.IDX and USECON.DAT. To
use DATA or STORE, you need to open a unit named HAVER with just that base name
(in this case USECON), with the full directory if required. In the FilePreference dialog,
in the Data Sources tab, you can set a permanent association of the HAVER unit with
a particular database, If youve done that, you dont need to include an OPEN HAVER
instruction to access that database.
Series are read according to their database names. Each series has its own range and
date scheme, so you can choose any subset of entries and can do compaction and expan-
sion as described in Changing Data Frequencies on page Int104 of the Introduction.
Example
The following reads six series (gdp, gdp deflator, gross private domestic investment,
unemployment rate, one year T-bill rate and M2) from the usecon data base. The
last three series are compacted from the monthly data on the database to quarterly
in the workspace.
open haver m:\data\haver\usecon
cal(q) 1960:1
data(format=haver) 1960:1 2010:2 gdpq dgdp fq lr fcm1 fm2

Additional Topics
2.12 FORMAT=HTML (HTML table)

html is a table-oriented text output (only) which renders the contents as an html
table for use in a web page. It is available on all versions of rats. You can apply it to
data series and reports.
RATS Instructions
copy(format=html) copy series from memory to an html table
prtdata(format=html) copy series from a rats format file to an html table
write(format=html) write scalars or arrays from memory to html file
report(format=html) used with the ACTION=SHOW and UNIT options, this
writes the report to an html table
Report Windows
You can write reports to html using FileSave As, FileExport, EditCopy; the Ex-
port and Copy toolbar items, and the Export and Copy contextual menu operations.
You can export series using FileExport..., the Export toolbar button, or the Export
operation on the contextual menu.
Matrix Windows
You can write the contents of a matrix window to an html file using File-Save As,
File-Export; the Save and Export toolbars, and the Export contextual menu.
Details
This produces just a part of a full html file, so you will have to insert it into a larger
document. From a report window, it copies real numbers as they are displayed on the
screen, so format them first to the precision that you want.

Additional Topics
2.13 FORMAT=MATLAB (MATLAB binary format)

matlab is the binary format used by the matlab software. This can be used in
various ways for input of series or matrices on any version of rats. Depending upon
the structure of the file, it can be either a Time Series Database format or an Unla-
beled format.
RATS Instructions
data(format=matlab) read series from matlab into memory
store(convert=matlab) copy series from matlab to rats format
read(format=matlab) read scalars or arrays from matlab into memory
Data Wizard
You can read series from a matlab (*.MAT) file using Data (Other Formats) on the
Data/Graphics menu.
Details
read(format=matlab)
Since READ is designed to read matrices, reading matlab data is straightforward: if
theres an NM matrix A on the file, READ A will dimension A as NM and read the
data into it. You can read multiple matrices off a single file. If youve saved scalars on
the file, you can read those as well.
data(format=matlab)
store(convert=matlab)
Data series will typically be stored in one of two ways: as separately named column
vectors, or as a single matrix with multiple columns. Either way, rats will not rec-
ognize any dates on the file. If the data are in separate columns, read them by name.
For the arrangement with multiple columns, use the option SHEET=matrix name,
then give series names to the columns.
Examples
The file CBES.MAT has one matrix named PSID. This has 3051 rows by 3 columns.
The columns represent (in order) age, education and income. The following will read
the data from that file and create three series:
open data cbes.bat
data(format=matlab,sheet="psid") 1 3051 age education income

Additional Topics
The file USMODEL_DATA.MAT has seven column matrices with names DC, DINVE, DY,
LABOBS, DW, PINFOBS and ROBS. Each column has 230 entries, representing quar-
terly data from 1950Q1 to 2007Q2. The following reads data from the file:
open data usmodel_data.mat
calendar(q) 1950:1
data(format=matlab) 1950:01 2007:02 dc dinve dy labobs dw $
pinfobs robs

Additional Topics
2.14 FORMAT=ODBC (ODBC/SQL databases)

This is a Labeled Table format for series input (only), available in the Professional
level of rats, to read data from any database format that supports Open Database
Connectivity (odbc) and sql.
RATS Instructions
data(format=odbc) read series from a database into memory
store(format=odbc) copies series from a database to a rats format file
None
Details
Before reading data from a database using odbc and sql, you must set up an odbc
Data Source for the desired database. The process for doing this varies somewhat
depending on the operating system in use. With recent versions of Windows, you can
set up a Data Source by opening the Windows Control Panel and double-clicking on
the Data Sources (odbc) control. This displays a dialog box you can use to define a
Data Source Name for your database.
Once you have the Data Source defined, you can open a connection to the database in
rats by using a command of the form:
open odbc dsn

where dsn is the Data Source Name for the database.
You can then read data from the database by executing a DATA command. On the
DATA instruction, you can either:
use the SQL option to provide a short (255 characters or less) sql string directly
on the DATA instruction, or
use the QUERY option to process a more complex sql queries, either from a sepa-
rate file or on the lines following the DATA instruction.
The DATA command creates a table (internally) which is then processed similar to the
spreadsheet files in Section 2.8 of the Introduction.
For example, the following creates a table with date and total daily sales (summing
the SUBTOTAL field by date to create the SALES field), then reads the data into rats,
creating the series SALES as monthly sums.
cal(m) 1995:1
open odbc "Sales"
data(format=odbc,compact=sum,
sql="select date,sum(subtot) as sales from invoice order by date") $
1995:1 2006:12

Additional Topics
The code below does the same thing using QUERY:

cal(m) 1995:1
open odbc "Sales"
open sqlfile "c:\rats\sqlquery.txt"
data(format=odbc,compact=sum,query=sqlfile) 1995:1 2006:12
where the file SQLQUERY.TXT contains the following lines:
select date,sum(subtot)
as sales from invoice
order by date
;

Additional Topics
2.15 FORMAT=PORTABLE (RATS Portable Format)

Portable format is the text version of the rats file format. It includes all of the infor-
mation stored on the rats file, but in text form. The most important use of is to cre-
ate archival humanreadable copies of rats format files. Its a Time Series Database
format which can be used for input and output in all versions of rats.
RATS Instructions
data(format=portable) read series from portable into memory
store(convert=portable) copy series from portable to rats format
copy(format=portable) write series to portable file
prtdata(format=portable) write series from rats format file to portable
Data Wizard
Examples
This writes a series out to a file in portable format, including one header (description)
line). You can add up to two of those.
copy(format=portable,header=1) / gnp
GROSS NATIONAL PRODUCT, CURRENT DOLLARS
This shows how a portable file is constructed. The series are displayed separately.
The key elements for each are the series name (on a separate line), the line contain-
ing the frequency and starting and ending dates, the header lines, and the row of ===
which ends the header.
GNP
Quarterly data from 1947:01 to 1990:04
GROSS NATIONAL PRODUCT, CURRENT DOLLARS
===============================================================================
1947:01 224.9000000 229.1000000 233.3000000 243.6000000
1948:01 249.6000000 257.1000000 264.0000000 265.5000000
etc.
dedit mydata.rat
open copy archive.txt
prtdata(unit=copy)
This writes all of the series stored in the rats data file MYDATA.RAT to a text file
called ARCHIVE.TXT.

Additional Topics
2.16 FORMAT=PRN (Delimited text formats)

This is a text file with fields delimited with spaces, tabs or commas. Its a Labeled
Table or Unlabeled format (depending upon the file) for both input and output of
series, reports and matrices, available on all versions of rats.
RATS Instructions
data(format=prn) read series from prn into memory
store(convert=prn) copy series from prn to rats format
copy(format=prn) write series to prn file
prtdata(format=prn) write series from rats format file to prn
read(format=prn) read scalars or arrays from prn into memory
write(format=prn) write scalars or arrays from memory to prn file
Data Wizard
Details

Additional Topics
2.17 FORMAT=RATS (RATS data format)

rats format is specially designed to deal with time series data. Its a Time Series
Database format which is available on all versions of rats. It supports all of the
time-series frequencies supported by rats, including panel and intra-day data. It
also allows you to store series with different frequencies in the same file.
RATS Instructions
data(format=rats) read series from rats format into memory
store(convert=rats) copy series from one rats format file to another
copy(format=rats) write series to rats file
prtdata(format=rats) write series from one rats format file to another
Data Wizard
You can read series by using FileOpen... to open a rats format file, selecting the
desired series from the rats Data File window, and using Data (RATS Formats) on
the Data/Graphics menu.
Series Window and RATS Data Window
You can use FileOpen and FileNew... to open or create a rats format file. Various
toolbar and contextual menu operation are available for generating graphs, editing
series, renaming series, and so on. You can bring series into memory from an open
rats file using the Data Wizard described above, or by dragging and dropping series
onto the Series Window.
You can write series from memory to the open file by dragging and dropping them
from the Series Window onto the rats Data File Window.
You can import series into the file from another using FileImport... or the Import
toolbar button. You can export series to another using FileExport..., the Export tool-
bar button, or the Export operation on the contextual menu.
Details
See RATS Format on page Int110 of the Introduction for more details on creating and using
rats format files.

Additional Topics
2.18 FORMAT=TEX (TeX table)

tex is a table-oriented text format for output (only) which renders the contents as a
tex tabular environment. You can apply it to data series and reports. Its available
on all versions of rats.
RATS Instructions
copy(format=tex) writes series to a tex file
prtdata(convert=tex) copy series from a rats file to a tex file
report(format=tex) used with the ACTION=SHOW and UNIT options, this
writes the report to a tex file
You can export series using FileExport..., the Export toolbar button, or the Export
operation on the contextual menu.
Report Windows
You can write reports to a tex file using FileSave As, FileExport, EditCopy as tex,
the Export and Copy toolbar items, and the Export and EditCopy as tex contextual
menu operations.
Matrix Windows
You can write the contents of a matrix window to a tex file using File-Save As, File-
Export, EditCopy as tex, the Save and Export toolbars, and the Export contextual
menu.
Details
To use the tex table in a document, you need to include the dcolumn package, which
allows for decimal alignment of columns. This is an example of what is produced:
\newcolumntype{.}{D{.}{.}{-1}}
\begin{tabular}{. . . }
\multicolumn{1}{c}{Test} & \multicolumn{1}{c}{Statistic} &
\multicolumn{1}{c}{P-Value}\\
\multicolumn{1}{l}{Joint} & \multicolumn{1}{r}{4.070} &
\multicolumn{1}{r}{0.000}\\
\multicolumn{1}{l}{Variance} & \multicolumn{1}{r}{0.121} &
\multicolumn{1}{l}{Constant} & \multicolumn{1}{r}{0.922} &
\multicolumn{1}{l}{X2} & \multicolumn{1}{r}{0.917} &
\multicolumn{1}{l}{X3} & \multicolumn{1}{r}{0.914} &
\end{tabular}

Additional Topics
Note that if you want to paste the table into a tex document, you need to use one of
the Copy as tex operations. The standard Copy operations also produce a straight tab-
delimited text copy of the data, which will generally be the one chosen by the Paste opera-
tion in the target application. Copy as tex puts only the tex version into the clipboard.
You will generally enclose the tabular environment produced by rats inside a table
environment which will handle the placement and captioning. If the tex code on the
previous page is saved in the file testtable.tex, the following would center it on
the page with the caption Example of tex Output.
\begin{table}[htb]
\centering
\include{testtable}
\caption{Example of TeX Output}
\end{table}
TeX formatting codes

In general, the export procedure will escape any character which would be interpret-
ed as a tex format code. For instance, the lag notation Y{1}, Y{2} and Y{3} out of a
standard regression output will appear as they would in rats output, as the { and }
get escaped to prevent the standard tex interpretation.
\newcolumntype{.}{D{.}{.}{-1}}
\begin{tabular}{. . . }
\multicolumn{2}{l}{Constant} & \multicolumn{1}{r}{0.0976}\\
\multicolumn{2}{l}{Y\{1\}} & \multicolumn{1}{r}{0.1356}\\
\end{tabular}
If you want to include special formatting codes, you can edit the information after
youve included in your final document. If you want to eliminate that step, you can
enclose a string (inside the quotes) with $ characters when you built a REPORT. For
instance, the following will use f1, f2 and f3 for the three lag coefficients.
report(action=define)
report(atrow=1,atcol=1) "1" %beta(1) %stderrs(1)
report(atrow=2,atcol=1) "$\phi _1$" %beta(2) %stderrs(2)
report(action=format,picture="*.####")
open texfile "regoutput.tex"
report(action=show,format=tex,unit=texfile)

Additional Topics
2.19 FORMAT=TSD (Tab separated text formats)

This is a text file with fields delimited with tabs. Its a Labeled Table or Unlabeled
format (depending upon the file) for both input and output of series, reports and ma-
trices, available in all versions of rats. For reading series, we recommend that you
use the more general FORMAT=PRN instead.
RATS Instructions
data(format=tsd) read series from tsd into memory
store(convert=tsd) copy series from tsd to rats format
copy(format=tsd) write series to tsd file
prtdata(format=tsd) write series from rats format file to tsd
read(format=tsd) read scalars or arrays from tsd into memory
write(format=tsd) write scalars or arrays from memory to tsd file
Data Wizard
Details

Additional Topics
2.20 FORMAT=WF1 (EViews workfile format)

wf1 is the EViewsTM workfile format. Its a Labeled Table format available for series
input (only), on all rats versions.
RATS Instructions
data(format=wf1) read series from wf1 format into memory
store(convert=wf1) copy series from wf1 format to rats format
Data Wizard
Details
Data in a wf1 file are stored as named series which will be read or imported using
those names. You can read all the series by leaving the series list blank on either
DATA or STORE, or you can request specific series by listing just the ones you want. If
you import directly into a data list, you will get all series.
Each file has a data frequency and start date, so you can read a subset of entries
or change the frequency and get the standard compaction or expansion handling
(Changing Data Frequencies on page Int104 of the Introduction).
Because these are often work files rather than simply data files, they will often
include extra series (like residuals) which are derived from the original data series.
If you use the Data Wizard, you may want to trim the list of series to include just the
data and do any further transformations yourself.
Example
open data yen1kp.wf1
calendar(m) 1973:1
data(format=wf1) 1973:01 1996:07 dm yen

Additional Topics
2.21 FORMAT=WKS (Lotus WKS format)

This is the native format used by the Lotus 1-2-3 spreadsheet and successor pro-
grams. While its no longer the main format used by any current software, there
are many legacy datasets which used it. Its a Labeled Table or Unlabeled format
(depending upon the file) for both input and output of series, reports and matrices,
available in all versions of rats.
RATS Instructions
data(format=wks) read series from wks into memory
store(convert=wks) copy series from wks to rats format
copy(format=wks) write series to wks file
prtdata(format=wks) write series from rats format file to wks
read(format=wks) read scalars or arrays from wks into memory
write(format=wks) write scalars or arrays from memory to wks file
report(format=wks) (with ACTION=FORMAT) write a report to wks file
Data Wizard
Report Windows
You can write reports to a wks file using FileSave As, FileExport, EditCopy; the
Export and Copy toolbar items, and the Export and Copy contextual menu operations.
Matrix Windows
You can write the contents of a matrix window to a wks file using File-Save As, File-
Export; the Save and Export toolbars, and the Export contextual menu.
The Copy operations will generally only allow pasting in wks format into a spreadsheet.
Details

Additional Topics
2.22 FORMAT=XLS/XLSX (Excel workbook)

These are the spreadsheet format for Microsoft Excel. Its a Labeled Table or Un-
labeled format (depending upon the file) for both input and output of series, reports
and matrices, available in all versions of rats. xls is the standard format used by
Excel through the 2003 version. xlsx is the standard format for Excel 2007 and
later.
RATS Instructions
For all of these, use FORMAT=XLS for Excel 2003 and older formats, and FORMAT=XLSX
for Excel 2007 format. Only XLS is permitted for output.
data(format=xlsx) read series from xlsx into memory
store(convert=xlsx) copy series from xlsx to rats format
copy(format=xls) write series to xls file
prtdata(format=xls) write series from rats format file to xls
read(format=xlsx) read scalars or arrays from xlsx into memory
write(format=xls) write scalars or arrays from memory to xls file
report(format=xlsx) (with ACTION=FORMAT) write a report to xlsx file
Data Wizard
You can read series from xls or xlsx files using Data (Other Formats) on the Data/
Graphics menu.
Report Windows
You can write reports using FileSave As, FileExport, EditCopy; the Export and
Copy toolbar items, and the Export and Copy contextual menu operations.
Matrix Windows
You can write the contents of a matrix window to a file using File-Save As, File-Ex-
port; the Save and Export toolbars, and the Export contextual menu.
The Copy operations will generally only allow pasting in xls format into a spreadsheet.
Details
See Spreadsheet and Delimited Text Formats on page Int114 of the Introduction for
more details on the format.

3. More on Functions
H ere we take a closer look at some useful rats functions.
Using %SWEEP
Polynomials
Additional Topics
3.1 The Sweep Family of Functions

The %SWEEP function, and related %SWEEPTOP and %SWEEPLIST functions, are very
useful, and probably dont get as much use as they deserve.
B=%SWEEP(A,k) is the following:
B (k, k ) = 1 A (k, k )
B (i, k) = A (i, k) A (k, k) for i k
B (k, j ) = A (k, j ) A (k, k ) for j k
B (i, j ) = A (i, j ) A (i, k) A (k, j ) A (k, k) otherwise
%SWEEP is reversible: %SWEEP(%SWEEP(A,k),k)=A.
The point of %SWEEP is the following: if A begins as a cross-product matrix of a set of

variables, successive sweeps will do a regression. If we fix 1 as the dependent vari-
able, after sweeping on 2 and 3, B(2,1) and B(3,1) will be the regression coefficients
and B(1,1) will be the residual sum of squares. Note that the sweep operator does not
produce a symmetric matrix.
The sweep functions are:

%SWEEP(A,k) Returns the result of sweeping A on pivot k.
%SWEEPTOP(A,k) Returns the result of sweeping A on each of the first k rows.
%SWEEPLIST(A,iv) Returns the result of sweeping A on each of the pivots listed
in the VECT[INTEGER] IV.
Consider the properties of the sweep function when applied to a cross product matrix:
1. The submatrix defined by the rows and columns of the pivots is ( X X )1 from a
regression on those variables.
2. The submatrix defined by the non-pivot columns with the pivot rows are the pro-
jection (regression) coefficients of the non-pivots on the pivots.
3. The submatrix defined by the pivot columns with the non-pivot rows are minus
the projection coefficients.
4. The submatrix defined by the non-pivots by non-pivots gives the sums of the
outer product of the projection residuals. (When divided by the number of obser-
vations, this will be the estimated covariance matrix of residuals).

Additional Topics
If we take a regression of trend on constant,

set trend 1 10 = t
cmom
# constant trend
disp %sweep(%cmom,1)
The output is:
0.10000 5.50000
1
( X X) matrix (regression coeff)
(regression coeff) -5.50000 82.50000 (Residual Sum of Squares)
One of the exercises in Hayashi (2000) is to do a Monte Carlo analysis of unit root
tests, doing one Dickey-Fuller test allowing for intercept and one with intercept and
trend. The sweep functions offer an efficient way to do this because the two regres-
sions can be done by sweeping first on the lag and intercept, then again on the trend.
cmom
# y y{1} constant trend
We sweep on 2 and 3 to regress on Y{1} and CONSTANT. The regression coefficients
are going to be in column 1 (non-pivot column). The sum of squared residuals is in
(1,1). The t-stat can be put together using these and the (2,2) element.
compute ss=%sweeplist(%cmom,||2,3||)
compute tmu=(ss(2,1)-1)/sqrt(ss(1,1)*ss(2,2)/(%nobs-2))
Now sweep on the trend as well and repeat
compute ss=%sweep(ss,4)
compute ttau=(ss(2,1)-1)/sqrt(ss(1,1)*ss(2,2)/(%nobs-3))

Additional Topics
3.2 Working With Polynomials

rats provides a suite of functions for manipulating polynomials. These are repre-
sented as a VECTOR of the v (1) , ,v (n + 1) , which denotes the polynomial:
v (1) + v (2) x + + v (n + 1) x n
%EQNLAGPOLY(EQN,S) Extracts a lag polynomial from an equation.

%POLYADD(P1,P2) Adds two polynomials.
%POLYCXROOTS(P) Returns a VECTOR[COMPLEX] with all roots of P, repre-
sented as complex numbers.
%POLYSUB(P1,P2) Subtracts two polynomials.
%POLYDIV(P1,P2,D) Divides P1 by P2, expanded out to degree d.
%POLYMULT(P1,P2) Multiplies two polynomials.
%POLYROOTS(P) Returns a VECTOR with all the real roots of P.
%POLYVALUE(P,X) Returns the value of P at x.
The most important of these is %EQNLAGPOLY(eqn,series), which pulls a lag

polynomial out of an equation or regression (the most recent regression if eqn=0).
If series is the dependent variable, it extracts the standard normalization for the
autoregression polynomial, that is, 1lag polynomial from the right side.
For example, this implements an ardl (autoregressive-distributed lag) model
linreg(define=ardl) logc
# logc{1 to 3} logy{0 to 3} seas{0 to -3}
compute arpoly = %eqnlagpoly(0,logc)

compute dlpoly = %eqnlagpoly(0,logy)
*
* Compute the long run effect
*
compute ardllt=%polyvalue(dlpoly,1)/%polyvalue(arpoly,1)
*
* Compute 8 lags of the expansion (degree 7 polynomial)
*
compute ardlest=%polydiv(dlpoly,arpoly,7)

4. Statistical Techniques
S tatistical practices change over time. rats has been around in some form for al-
most 30 years, beginning its life as a mainframe program on computers with less
power than a modern cell phone. Techniques which were commonly employed in the
1980s often arent even taught today, or have been replaced by modernized versions.
We include here a description of some regression procedures which are no longer
used enough to merit space in the main documentation, but might still be of some
interest to some users. We also include a description of several less-used options on
the rats estimation routines.
Polynomial Distributed Lags

Ridge and Mixed Estimators
Robust Estimation (M-Estimators)
Miscellaneous Regression Options
Band-Spectrum Regressions
Frequency Domain Deseasonalization
Additional Topics
4.1 Polynomial Distributed Lags

Polynomial distributed lags (or Almon lags or pdls) are the principal form of hard
restriction used for estimating distributed lags. See Chapter 2 of the Users Guide for
a more complete discussion of the topic of distributed lags.
For the equation:
L
(1) yt = b X
l = L0
l t l + ut
an unconstrained third degree PDL (lags 1 to L) takes the form:

L
(2) (a + bk + ck
k =1
2
+ dk3 ) X tk
where a, , c and d are the free parameters, that is, bl is constrained to follow a 3rd
order polynomial. If you expand this sum, a is the coefficient on the sum of X t-k , b
on kX t-k , etc. We use ENCODE to create these four series, which we then use in the
regression.
We impose an end constraint (polynomial is zero for lag L+1) by writing the
polynomial as (k (L + 1))(a + bk + ck2 ) .
We impose near (polynomial is zero for lag 1) and far constraints using
(k (L + 1))(k + 1)(a + bk).
The trick in handling the constrained PDLs is to get the EWISE instruction, which
creates the encoding matrix R, correct. Because lag 0 (k=0) corresponds to the first
row in R (J=1 in EWISE), the roots should be at J=0 (k=1) and J=L+2 (k=L+1).
As mentioned in the Users Guide section, you need to be very cautious in interpret-
ing the results from a pdl, since the t-statistics will usually be very high. This is
because they are testing whether a coefficient can be made zero, while maintaining
the hard shape constraint.
pdls have fallen out of favor mainly because the basic distributed lag like (1) usually
leaves very high serial correlation in the residuals. If you allow for richer dynam-
ics by adding lagged dependent variables (the ardl model), you will generally get
results that are easier to interpret.

Additional Topics
Example 4.1 Polynomial Distributed Lags

This example program is supplied with rats on the file PDL.RPF. We show both the
direct use of the ENCODE and UNRAVEL method for estimating a restricted regression,
and the use of the @PDL procedure for doing this type of analysis.
open data haversample.rat

calendar(m) 1947
data(format=rats) 1947:1 2007:4 fltg ftb3
set shortrate = ftb3
set longrate = fltg
*
* Run all estimates starting in 1951:1.
*
smpl 1951:1 *
declare rect r
*
* PDL with no end constraints
*
dim r(4,25)
ewise r(i,j)=j^(i-1)
encode(results=enc) r
# shortrate{0 to 24}
linreg(unravel) longrate
# constant enc
*
* Same estimation using the PDL procedure
*
@pdl(graph) longrate
# shortrate 0 24 3
*
* PDL with far constraint
*
dim r(3,25)
ewise r(i,j)=(j-26)*j^(i-1)
encode(results=enc) r
# shortrate{0 to 24}
linreg(unravel) longrate
# constant enc
*
* Same estimation using the PDL procedure
*
@pdl(constrain=far,graph) longrate
# shortrate 0 24 3

Additional Topics
4.2 Ridge and Mixed Estimators

Ridge regression (Hoerl and Kennard, 1970) and mixed estimation (Theil, 1971, pp.
347-352) are related procedures which are fairly easy to implement within rats.
These are Bayesian-like estimation techniques, in the sense that there are priors
which will generate them as their posterior modes. See Chapter 16 of the Users Guide
(Gibbs Sampling in particular) for more precise methods of handling Bayesian esti-
mation in rats.
A ridge estimator of b in y = X b + u is defined as
1
(3) b = ( X X + kI) X y
for some positive k. Least squares is the special case of k=0. Setting k>0 has the effect
of shrinking the estimate of the coefficients toward the origin, although individual
coeffficients may move away from zero.
In mixed estimation, non-sample information about b is put in the form
(4) r = Rb + v , Cov (v ) = V
The mixed estimator (assuming Cov (u) = s 2 I ) is

1
(5) = ( X X + 2 R V1R) ( X y + RV r)
2 1
The ridge estimator is a special case of this where R= I, r = 0, and V is scalar.
Computation Tools
1
Note that both techniques estimate b as ( X X + Q) ( X y + q). The tools for imple-
menting such estimators are the CMOMENT and LINREG(CMOMENT) instructions:
CMOMENT computes the matrices X X and X y using the actual data. In prac-
tice, these will actually be segments of the same matrix. We then replace these
matrices with X X + Q and X y + q.
LINREG(CMOMENT) computes the regressions using the altered versions of
X X and X y.
To simplify this process, on CMOMENT you should list the variables as regressors first,
followed by the dependent variable. That way, the matrix %CMOM, as computed by
CMOMENT will have the form
X X y X

X y y y

(actually just the lower triangle of this), so X X is the top corner, and X y is all but
the last element of the last row.

Additional Topics
Ridge Regression
Ridge regression is especially easy because all we need to do is add a constant to the
diagonal elements of . The simplest way to do this is with a DO loop:
cmoment
# constant x1 x2 x3 x4 y
linreg(cmoment) y Least squares
# constant x1 x2 x3 x4
do row=1,5
compute %cmom(row,row)=%cmom(row,row)+k Add to diagonals 1-5
end do row
linreg(cmoment) y Ridge regression
# constant x1 x2 x3 x4
Choosing an appropriate value for K is the real problem with ridge regression. See,
for example, the discussion in Judge, Griffiths, Hill and Lee (1980).
Mixed Estimation
We need to construct the matrices R, r and V in some fashion. r is often zero and V is
almost always diagonal, and is usually a scalar matrix (a constant times the iden-
tity), which simplify these considerably.
Since we do not know s2, in general, we must use an estimate s2 of it. We do a pre-
liminary LINREG to get the estimate s of s, then use matrix commands to replace
X X by X X + s2 R V1R , and
X y by X y + s2 R V1r .
Where V is scalar (say l 2 I ), this can be done quite simply (if the %CMOM array has
been arranged as described on the previous page) by stacking R on top of r into Rr
and then adding to %CMOM the array (s2 /l 2 )(Rr ) (Rr ) . This is what is done in Ex-
ample 4.2. If V is more general, its necessary to change the X X and X y matrices
separately. This is done most easily using the OVERLAY instruction.
Because the general case involves a bit of extra programming, we provide with rats
a procedure named @MIXED (on the file MIXED.SRC) which does the entire calcula-
tion. Its syntax is
@mixed( options ) depvar start end
# explanatory variables in regression format
The options are
capr=R matrix
lowr=r matrix
scalar=scalar for V being a scalar matrix
V=V matrix in general

Additional Topics
Example 4.2 Shillers Smoothness Prior

Shiller (1973) proposed a method for distributed lags which reduces the variability
in the lags seen in least squares estimates, while imposing a less stringent shape
restriction than the PDL. His smoothness prior (of degree d) takes the form:
d +1
(6) (1 L ) bi = vi
where b is the vector of lag coefficients. We do this using mixed estimation (Section
5.12). Since the dummy observations have a fairly simple form (in particular, the V
matrix is scalar), we will do the estimates directly.
FMATRIX creates these dummy observations. Note that the rows of R in Rb = v take
precisely the FMATRIX form with the option DIFF=d+1. We use several values for the
weight put on these.
This example uses a first degree prior (d=1) with lag zero of SHORTRATE left free.
Thus, there are 22 rows in the R matrix: the 24 affected lags minus (d+1). FMATRIX
starts at entry (1,3) because the CONSTANT and lag zero are the first two regres-
sors in the cross moment matrix. Since we need the original %CMOM array each time
through the DOFOR POWER loop, we store it in the array CMOMHOLD.
The example is supplied on the file SHILLER.RPF. See also SHILLERGIBBS.RPF,
which does the same model using Gibbs sampling.
open data haversample.rat
calendar(m) 1947
data(format=rats) 1947:1 2007:4 fltg ftb3
set shortrate = ftb3
set longrate = fltg
cmom
# constant shortrate{0 to 24} longrate
linreg(cmom) longrate
# constant shortrate{0 to 24}
declare rect dummy(22,27)

declare symm cmomhold(27,27) dummyxx(27,27)
fmatrix(diff=2) dummy 1 3
compute cmomhold=%cmom
compute dummyxx =tr(dummy)*dummy
compute scalefac=%seesq/.01
dofor [real] power = 1.0 2.0 4.0 8.0
compute kfactor=power*scalefac
compute %cmom =cmomhold+kfactor*dummyxx
compute title=$
"Shiller Smoothness Prior with k="+%strval(kfactor,"*.##")
linreg(cmom,title=title) longrate
# constant shortrate{0 to 24}
end dofor

Additional Topics
4.3 Robust Estimation by Iterated Weight Least Squares

Least absolute deviations (lad) is fairly complicated technically. An alternative way
of handling this is to use a criterion function which tail behavior similar to LAD, but
is more similar to least squares near zero, and thus is differentiable. For example,
the following uses a convex function, twice differentiable everywhere, which is as-
ymptotically u .
( )
12
2
(7) min c 2 + ut (b )
b
t
c is some constant. As c 0 , this approaches LAD. The first-order necessary condi-

tions (FONCs) for solution are
X tut
(8) 12
=0
t (c 2
+ ut2 )
12
which are the same as the FONCs for weighted least squares with SPREAD= (c 2 + ut2 )
We can thus estimate b by iterated weighted least squares.
Using the results from Section 2.2 in the Users Guide, we can compute the covari-
ance matrix as A-1BA-1, where
ut2
(9) B = X tX t
t (c 2
+ ut2 )
c2
(10) A = X tX t 32
(Note: we changed the sign on this)
t (c 2
+ ut2 )
This is included in ROBUST.RPF (Example 4.3) along with lad estimates.

Additional Topics
Example 4.3 Robust Estimation

open data zellner.prn
data(format=prn,org=columns) 1 25 valueadd capital labor nfirm
set logy = log(valueadd)

set logk = log(capital)
set logl = log(labor)
Compute linear regression, and the standardized residuals.
linreg logy / resids
# constant logk logl
prj(xvx=px)
set stdresids = resids/sqrt(%seesq*(1-px))
Rerun the regression, omitting the outliers (defined here as observations with stan-
dardized residuals greater than 2.5 in absolute value.
linreg(smpl=abs(stdresids)<=2.5) logy
Now do LAD estimator.
rreg logy
Iterated WLS M-estimator
compute c = sqrt(%seesq)
dec vector beta0
Start with least squares (spread = constant). Do 10 iterations or until convergence.
set spread = 1.0
do iters=1,10
compute beta0=%beta
set spread = sqrt(c^2+resids^2)
linreg(noprint,spread=spread) logy / resids
if %testdiff(beta0,%beta)<.0001
break
end do iters
disp "Iterations Taken" iters
Compute the sandwich estimator for the covariance matrix.
set f = resids/spread
set fprime = c^2/spread^3
mcov(matrix=b,lastreg) / f
mcov(matrix=a,lastreg,nosquare) / fprime
linreg(create,lastreg,form=chisquared,covmat=%mqform(b,inv(a)),$
title="Iterated Weighted Least Squares")

Additional Topics
4.4 Miscellaneous Regression Topics

The Option DFC
dfc=Degrees of Freedom Correction
The DFC option is used to enter the degrees of freedom lost in the dependent vari-
able in previous processing.
When you use regression residuals as the dependent variable, the standard errors of
the estimated coefficients of the regression will be understated, because the first re-
gression extracts some degrees of freedom. The Degrees of Freedom Correction
should be equal to the number of regressors in the first regression, minus the number
of those regressors which are repeated in the second regression.
linreg ip / detrend
# constant trend seasonal{ -10 to 0 } 13 regressors
linreg(dfc=12) detrend dfc=13 1
# constant intrate CONSTANT repeated
The ENTRIES Option

entries=number of supplementary card entries to process [all]
You can use the ENTRIES option on any instruction which has a supplementary card
in regression format: LINREG, CMOMENT, STWISE, PREGRESS, AR1, EQUATION (applies
to the supplementary card listing the MORE variables), LDV, DDV, PRB, LGT, MAKE,
ENCODE, MCOV, EXCLUDE and SUMMARIZE. You can also use it for the supplementary
card for TEST.
ENTRIES lets you put one of these instructions inside a loop and change the number
of regressors on the supplementary card to use. ENTRIES is largely obsolete because
you can use the ENTER instruction or one of the reglist functions to build supple-
mentary cards in a VECTOR[INTEGER].
An entry is any of the objects on the supplementary card, except for the control char-
acters # and $. You must count the { and } of a lag field. Thus M1{1 TO 10} counts
as 6 entries (M1, {, 1, TO, 10, and }). rats will use the number of entries specified by
the ENTRIES option, starting from the first entry on the card.
The code below is part of a procedure which computes something similar to a vec-
tor autoregression, but with varying number of lags among the variables. NUMVAR is
the number of variables in the system, VARS is a VECTOR[INTEGERS] with the list of
variables, and LAGS is a VECTOR[INTEGERS] with the number of lags for each. This
can handle anything up to a six variable system simply by changing NUMVAR.
do j=1,numvar
linreg(entries=6*numvar+1) vars(j)
# constant vars(1){0 to lags(1)} vars(2){0 to lags(2)} $
vars(3){0 to lags(3)} vars(4){0 to lags(4)} $
vars(5){0 to lags(5)} vars(6){0 to lags(6)}
end do j

Additional Topics
4.5 Band Spectrum Regression

Another type of frequency domain regression procedure is Engles(1974) band spec-
trum regression. In this procedure, the filtering procedure is to zero out certain
bands in the Fourier transforms of the dependent variable and regressors. This is
basically identical to Hannan efficient, except that a masking series is used instead
of the one computed there. For instance, to include just the lowest 3/4 of frequencies,
use
cset 1 = t<=nords*3/8.or.t>nords*5/8
in place of the FFT, CMULT, WINDOW and CSET instructions.

Additional Topics
Example 4.4 Deseasonalization Using Spectral Techniques

Sims (1974b) describes a procedure to deseasonalize data by removing all power
from frequencies in a band about the seasonals. This is a frequency domain filter-
ing procedure where the transfer function is a seasonal mask constructed using
CMASK. The width of the seasonal bands depends upon the width of the peaks in the
spectrum. The data in this example are Canadian retail sales. The example file is
FREQDEASON.RPF.
open data oecdsample.rat
calendar(m) 1960
Data are Canadian retail sales
data(format=rats) 1960:1 2007:3 canrett canretts
Transform to logs and take the trend out
set logret = log(canrett)
filter(type=hp) logret / removed
set prepped = logret-removed
compute nords=2*%freqsize(2007:3)
frequency 2 nords
rtoc
# prepped
# 1
fft 1
Create bands of width p/16. Leave the low frequencies in.
cmask 2 / (nords/12) (nords/32) (nords/12)+1
cmult 1 2
ift 1
Send the deseasonalized data back, add back the removed trend and exponentiate.
ctor 1960:1 %allocend()
# 1
# deseason
set deseason = exp(deseason+removed)

Graphs the deseasonalized series with the officially adjusted series.
graph 2
# canretts
# deseason

5. Instructions
T his section documents several instructions that are rarely used and thus are no
longer included in the Reference Manual. We also describe some options and pa-
rameters that are still supported by other instructions, but which have been super-
seded by newer parameters or (in most case) options.
Deprecated Instructions
Obsolete Options and Parameters
Additional Topics
5.1 Deprecated Options and Parameters

This section lists options and parameters which are still recognized by rats, but
have been superseded by newer features. In many cases, weve replaced parameters
with options, as it is usually easier to remember the appropriate option name than to
remember the proper sequence of parameters. In other cases, these have been obso-
leted by entirely different (and better) ways of doing things. Unless otherwise noted,
rats will still recognize the older parameters, but we recommend that you use the
new options.
ALLOCATE
Prior to Version 6 of rats, the ALLOCATE command was required in any program
that used data series. It is now optional. This is no longer needed, as you now set the
default range by supplying explicit start and end dates (or entry numbers) on your
first DATA, SET, or SMPL instruction.
AR1
A fifth coefficient used in versions prior to 7.0 for saving the coefficients has been
deprecated. Coefficients are now saved automatically in the vectors %BETA and %BE-
TASYS.
The DIFFERING option was renamed as HETEROGENOUS in Version 7. The name DIF-
FERING is still recognized, but not recommended.
ERRORS
The steps and VCV Matrix parameters used in older versions have been replaced
by the STEPS and CV options, respectively.
ESMOOTH
Two parameters (forecasts and steps) used in older versions have been replaced
by the FORECASTS and STEPS options described below.
ESTIMATE
Two additional parameters (residuals and coeffs) used in older versions have
been replaced by the RESIDUALS and COEFFS options described below.
IMPULSE
The steps, shock to and VCV Matrix parameters used in older versions have been
deprecated. These are replaced by the STEPS, COLUMN, and CV options, respectively.
INCLUDE
The comments parameter used in older versions has been replaced by the COMMENTS
option.

Additional Topics
KALMAN
The parameters residual, coeffs, and printflag used in older versions have been
replaced by options. The first two parameters, for storing the residuals and coef-
ficients, have been replaced by the RESIDS and COEFFS options (you can also use
COHISTORY to save coefficients). The functionality of the third parameter, used to
control printing of output based on a logical expression, has been incorporated into
the PRINT option .
MRESTRICT
Two additional parameters (resids and coeffs) used in older versions have been
deprecated. rats still recognizes these, but they are unnecessarythe residuals and
coefficients are now saved automatically in the %RESIDS series and %BETA vector,
respectively. You can also use the COEFF option to save the coefficients.
MVFRACTILE
The MVFRACTILE instruction functionality has been incorporated into the MVSTATS
instruction.
RENAME
The comments parameter (and accompanying text cards) used in older versions has
been replaced by the COMMENTS option described below.
RREG
A fifth parameter called coeffs, which saved coefficients into a series, has been dep-
recated. It will still function in Version 7, but is not recommended, as the coefficients
are automatically saved in the %BETA vector.
SEASONAL
Two additional parameters (seasonal and period) used in older versions have been
replaced by the SPAN and PERIOD options described below.
SIMULATE
Three additional parameters used in older versions have been replaced by options.
The steps and start parameters are replaced by the STEPS, FROM, and TO options.
The VCV Matrix parameter is replaced by the CV option.
STEPS
Two additional parameters (steps and start) used in older versions have been
replaced by the STEPS, FROM, and TO options described below.

Additional Topics
STWISE
A fifth parameter, coeffs, still works but is not recommended. The coefficients are
automatically saved in the vector %BETA.
SUR
The current options CV and CVOUT were called ISIGMA and OUTSIGMA, re-
spectively, in previous versions. The old names are still recognized in Version 8.
THEIL
Two additional parameters used in older versions, steps and data end, have been
replaced by the STEPS and TO options.
X11
Two additional parameters, adjusted and factors, used in earlier versions have
been replaced by options of the same names.

Additional Topics
5.2 Deprecated Instructions

This section lists (in alphabetical order) some obsolete or little-used instructions that
are still recognized by rats, but are no longer documented in the Reference Manual.
Descriptions begin on the next page.

Additional Topics
CNTRL Tools for Controlling Interactive Procedures

CNTRL is an old instruction used for controlling various aspects of program execution.
It is largely obsolete now that rats includes other instructions, such as USERMENU,
MESSAGEBOX, and SELECT for controlling interactive procedures.
cntrl( options ) pause message
Parameter
pause message This is a string of characters which is reproduced in the dia-
log box or prompt when you do CNTRL(PAUSE). Note that the
newer MESSAGEBOX and INFOBOX instructions are generally
more useful for interactive programs.
Options
page/[nopage]
CNTRL(PAGE) inserts a page break in the output file. You should only use this for
programs you intend to run in batch mode.
pause/[nopause]
Pauses and waits for a user response before continuing. This will be a dialog box,
or a simple message to hit a key, depending upon the environment. You can use
the pause message parameter to add a message to the user prompt. As noted
above, the MESSAGEBOX and INFOBOX instructions offer more flexible alterna-
tives.
user/[nouser]
Takes input from the keyboard until the user types a RETURN instruction.
interactive/[nointeractive]
Use this at the beginning of a program which is run as a batch program but uses
some interactive features.
Notes
If you are writing an interactive procedure that is to be executed in batch mode, you
will probably want to use the instruction ENVIRONMENT NOECHO to suppress the
echoing of lines as they are processed. Let the user see only what you want her to see.

Additional Topics
DEFAULT Changing Default Options

Most instruction options have a default setting, which rats uses unless you specify
otherwise. The DEFAULT instruction lets you change one or more of these default
values temporarily. The changes remain in effect until you quit rats, or change the
defaults again with another DEFAULT instruction. This can be very useful if a pro-
gram repeatedly uses a particular instruction with the same set of options.
default instruction name(options)
Parameters
The parameter field for DEFAULT looks like an instruction plus its option field, for
instance, GRAPH(NODATES). You can change more than one option for a given in-
struction with a single DEFAULT, but you need a separate DEFAULT for each separate
instruction whose defaults you wish to change.
DEFAULT only affects the options you specifyother default settings for the instruc-
tion remain unchanged.
Examples
default print(nodates)
default graph(nodates)
These two instructions set NODATES as the default option for PRINT and GRAPH. Sub-
sequent PRINT and GRAPH instructions will, by default, label output by entry number
rather than dates.
default linreg(noprint,smpl=regsampl)
This changes two of the default options for the LINREG instruction
The default print option becomes NOPRINT, so you would need to include a
PRINT option in a LINREG instruction to have it display the output.
The SMPL=REGSAMPL option means that subsequent LINREG instructions will
use REGSAMPL as the SMPL series.

Additional Topics
DELETE Removing Series from a RATS Data File

DELETE removes the specified series from the currently open rats data file (see the
instruction DEDIT). You must to use a SAVE instruction at some point to make these
changes permanent.
delete list of data file series, separated by blanks
Wizard
If you open the rats format file with Open on the File menu, the Delete and Cut
operations on the Edit menu will delete the selected series.
Parameters
list of series The list of series which you want to delete from the file.
Example
dedit pricedat.rat
delete turcpi grccpi prtcpi islcpi
save
removes the series TURCPI, GRCCPI, PRTCPI and ISLCPI from the rats data file
PRICEDAT.RAT.
Notes
When you do a DELETE instruction, it is likely that the length of your file will not
change. rats simply marks the old records as unavailable, but does not physically
change the size of the file.
See Also . . .
Intro. Section 2.7 rats format data files
DEDIT Opens (or creates) a rats format data file for editing.
STORE Adds data from one or more series to the rats file you are edit-
ing
INCLUDE Adds data from a single series to the rats file you are editing.

Additional Topics
EDIT Screen Data Editing

EDIT allows you to edit a series stored on a rats format file, or create a new series
on the file, using a spreadsheet-style editing screen. Alternatively, you can use the
instructions UPDATE and STORE: UPDATE is useful for minor changes, while STORE
replaces the entire data series.
The EDIT command is only available when running rats in interactive mode.
edit series to edit
Wizards
You can also edit series on a rats file by using the Open or New operations on the
File menu to open or create a rats file, and double-clicking on the series you want to
edit or clicking on the new series toolbar icon. You can edit series stored in memory
(rather than on a rats file) by doing ViewSeries Window and double-clicking on the
series you want to edit.
Parameters
series The list of series names you want to edit or create. If listing
multiple names, separate each by one or more blank spaces. If
you dont list any series, rats will display a scrolling list of all
the series on the file. You can select a series to edit from the
list, or enter a new series name to create a new series.
Option
modal/[nomodal]
With MODAL, an EDIT editing window functions like a modal dialog box. Execu-
tion of any pending instructions is suspended until the user closes the editing
window and new instructions can be executed. This can be useful when you want
the program to stop and wait for the end user to enter data before continuing.
Opening a RATS Format File

Before you can EDIT a series stored on a rats format file, you need to open or create
the file using the File-Open or File-NewRATSData Window operations, or the DE-
DIT or ENVIRONMENT RATSDATA= instructions.
Editing a Series
Once you have opened or created a rats data file, you can edit or create a series by
issuing a command like:
edit usargdps

Additional Topics
rats will display a series edit window which you can use to edit the values in the
series. See page Int-14 in the Introduction for details on using the series editing
window.
Creating a New Series

If you supply a new series name on the EDIT instruction, rats will ask if you want
to create the series. If you say Yes, the program will display a dialog box allowing
you to select the frequency and start date (Undated is one of the choices), and then
display the edit window for that series.
Saving Changes
When you are done editing, just close the editing window. You will see a dialog box
asking if you want to save the changes to the series. Answer Yes if you want to keep
the changes.
Important Note: You still need to save the changes to the file itself by issuing a
SAVE instruction if you opened the file using DEDIT, or by using FileSave if you
opened the file using FileOpen or FileNew. If you try to quit out of rats before sav-
ing changes to the data file, the program will ask if you want to save your changes.

Additional Topics
INCLUDE Adding Series to a RATS Data File

INCLUDE is one of three instructions that you can use to add data to a rats format
file (Users Guide, Section 2.5). It adds a single series to the file using data which you
have already brought into rats. The instruction EDIT allows you to edit or create a
series using a screen editor, while STORE includes many series or the contents of an
entire data file.
include fileseries series start end
Wizard
You can also store series on a rats format files using the mouse. First, open or cre-
ate a rats format file by doing FileOpen or FileNew. Then, use Series Window on
the View menu to display a list of the series in memory. Finally, drag and drop the
desired series from the series window onto the rats data file window.
Parameters
fileseries Name you want to assign to the series when it is saved on the
data file. There should not be a series with name file series
already on the file. If you wish to change an existing file series,
use STORE or EDIT, or DELETE the series before INCLUDEing.
series The series (in working memory) which supplies the values for
file series. Since you often will be using the same name on
the data file, the default for series is file series. However,
the names can be different.
start end Starting and ending entries of the range of data you want to
save. If you have not set a SMPL, this defaults to the defined
range of series.
Option
comments=STRING or VECT[STRINGS] containing comments [no comments]
Use this option if you want to add comments describing the series. You can add
up to two (80 character) comments to a series.

Additional Topics
Description
INCLUDE adds to the data file a series with name file series. rats takes the data
for this from series entries start to end. If you only specify the file series
parameter, INCLUDE takes data from that series and saves it on the file under the
same name. As with other editing instructions, you must SAVE the file when you are
finished with all your changes to make them permanent.
Supplying Comments
If you wish to add comments to a series, you can define the STRING or
VECT[STRINGS] ahead of time, or create it directly as the argument for the option,
as shown in the example below.
Example
cal(q) 1980:1
all 2006:4
* Read the AUSSALES series in from a text file:

open data aussales.dat
data(format=free) / aussales
* Open an existing RATS format file:

dedit austdata.rat
* Add the AUSSALES series to the RATS file, and include

* a comment describing the series:
include(comments="Australian Sales, Total") aussales sales

* Save the changes to the RATS file and close the file
save
quit

Additional Topics
RENAME Renaming Data File Series

Use RENAME to rename or change the comments for a series stored on the rats for-
mat data file that you are currently editing. It has no effect on the data in that series.
You must do a SAVE instruction later in order to make the changes permanent.
rename old name new name
Wizard
If you open a rats format file using FileOpen..., you can rename a series using the
Rename toolbar ion:
Parameters
old name Current name of the series.
new name New name for the series.
Option
comments=STRING or VECT[STRINGS] containing comments [no comments]
Use this option if you want to add comments describing the series. You can add
up to two (80 character) comments to a series.
Examples
dedit mydata.rat
rename fzm1 m1
rename(comments="GDP in 1992 dollars") gdp92 realgdp
save
This renames two series on the file, and attaches the comment GDP in 1992 Dollars
to the series REALGDP.
dedit prices.rat
open data prices.xlsx
store(convert=xlsx,org=cols)
rename 1234 a1234
save
converts an xlsx file to rats format, then fixes an illegal name (rats variable
names cant start with a digit).

Additional Topics
UPDATE Altering Data on a RATS Data File

UPDATE adds data to or replaces data on a rats data file series. It is a batch style
alternative to using the full-screen editor invoked by EDIT, or the FileOpen op-
eration. You must open the file for editing using DEDIT before using UPDATE. Any
changes you make only become permanent when you execute a SAVE.
update file series begin entry(expanded) end entry(expanded)

< data cards >
Parameters
file series Name of the series (on the rats data file) you want to update.
begin entry This is the date or entry number of the first entry you want to
change or add. You can use date fields, but note that these are
for the calendar setting of the series on the file. By default, it is
the current end of the file series + 1.
end entry The date or entry number of the end of the updated section.
If you omit this, UPDATE will add or alter one or more values
starting with begin entry.
Description
UPDATE reads data for the update from data cards following the instruction. Data
can be separated by blanks, tabs or commas. If you give explicit values for both the
begin entry and end entry parameters it will replace or add data for just those en-
tries. If you omit end entry or both, UPDATE will change or add as many entries as it
processes from a single line of data. With panel and intra-day data, you can only add
entries to the end of the series with UPDATE. Use EDIT, SET, or COMPUTE to modify
existing values of a panel data series.
You do not need to set up a full rats program (with CALENDAR and ALLOCATE) to use
UPDATE, as it takes all its information from the data file. You can update a file simply
with DEDIT, one or more UPDATEs, SAVE and END.
Example
dedit mydata.rat
update realgdp
1604.3 1611.4 1625.2 1644.7
update sales 2006:3
11234
save
adds four entries to the end of REALGDP and replaces the 2006:3 value of SALES.

6. RATS Error Messages
T his section lists the various error and warning messages that rats can display. In
many cases, we've included some comments that may help you resolve the issue.
When faced with an error or warning, be sure to take a moment to consider carefully
what the message says. While some causes are hard to identify properly, in many
cases the error message will tell you exactly what you need to know to resolve the
problem.
RATS Error Messages

Additional Topics
6.1 RATS Error Messages and Troubleshooting

This appendix lists all of the error and warning messages that rats can produce.
Many of these are largely self-explanatory. However, for some of the more common
errors, especially those that can have fairly obscure causes, we have tried to provide
suggestions to help you resolve the problem.
The messages are grouped by type. Each error has a unique error code consisting of
1 to 3 letters followed by a number, indicating the general type of error. For example,
basic syntax errors begin with the letters SX, while errors related to Vector Autore-
gressions begin with the VAR.
Many of the messages will include detailed information about the specific cause of
that error. For example, if you try to multiply a 3x3 matrix with a 4x4 matrix, rats
would produce the error:
## MAT2. Matrices with Dimensions 3 x 3 and 4 x 4 Involved in * Operation
In the listing below, information that varies depending on the context of the error
(such as the specific dimensions involved in the MAT2 error show above) are enclosed
in << and >>. Where possible, we indicate the type of information that rats will
display in an actual error message. For example, <<name>> means that rats will
display the name of a variable, instruction, option, etc., while <<data type>> indi-
cates that it will display a data type, such as SERIES or REAL.
Some of the error messages end with a question, such as Did you forget to
SOURCE?. Note that these only suggest a possible cause for the error.
Finally, note that a number of these error messages are obsolete, and thus will never
be generated by Version 7 or later of rats. We include them here only because they
are still present in the error message file used by rats (RATSERRS.MSG).
General Syntax Errors

SX1 Identifier <<name>> is Already in use as a(n) <<data type>>
Youve tried to re-declare an existing variable as a different type. Use a different
name for the variable, or clear the memory and re-start.
SX2 Expected << or >> to Terminate String
SX3 Expected a Label/Identifier
SX4 Expected Blank or Tab Here
SX5 Expected = Here
SX6 Expected ) Here
SX7 ( Must Follow Immediately After Function Name
Dont put a blank space between a function name and the (. For example, use
%ran(1.0), not %ran (1.0).
SX8 <<data type>> is not an Aggregate Data Type
SX9 <<data type>> is not a Basic Data Type
SX10 ] Expected After Type Declaration
SX11 Identifier <<name>> is Not Recognizable. Incorrect Option Field or
Parameter Order?

Additional Topics
Dont put a space between an instruction name and an option field. For example:
use print(nodates), not print (nodates). Check for misspelled names.
SX12 Expected a Variable Name Here
SX13 <<name>> Is Not a Valid RATS Variable Name
See Symbolic Names in Section 1.9 of the Users Guide for details on what rats
accepts as valid variable names.
SX14 Ill-formatted number
SX15 Trying to Store into Constant or Expression. Called Parameter by
Value?
If a PROCEDURE parameter is called by value, you cannot change the value stored in
that variable. If necessary, call the parameter by address instead.
SX16 Missing Operand or Adjacent Operators
Be sure to put at least one space before the equal sign in a SET instruction. For
example, use set x =t and not set x=t.
SX17 Missing Operator or Adjacent Operands
The SX17 can occur if you refer to N(entry) where N is a series number use
([series] N)(entry) instead. In SET or FRML expressions, you can use lag
notation: N{0}
SX18 Dates Must be a:b or a:b:c
Dont forget to separate the year and period (or the year, month, and day) in a date
expression with colons, such as: 1999:3:1.
SX19 This Character is Illegal Here
SX20 Expected <<character>> Here
SX21 A <<character>> Here is Unneeded or Unexpected
SX22 Expected Type <<data type>>, Got <<data type>> Instead
An expression may be returning a different data type than the declared type of
the variable you are setting, or you are using a variable of a type other than that
expected by the instruction (such as using a REAL variable as the series parameter
on SET).
SX23 Expected Variable/Element of Type <<type>>, Got Expression Instead
SX24 Expected Dimension Field Here
When using DIMENSION, dont put a space between a variable name and the left
parentheses. For example, use dim xmat(10,10), not dim xmat (10,10).
SX25 Logical Operators are .EQ./.NE./.LE./.LT./.GE./.GT./.AND./.OR./.N
OT.
SX26 Matrix Constructed with ||...|| has Too Complicated a Type
SX27 Illegal Combination of Data Types for Operation
SX28 <<{ or }>> For Lags Used In Improper Manner
SX29 Need a } to End the Lag Field
SX30 Characters After End of Expression
SX31 Expression to Left of , Has No Effect (year,period is obsolete no-
tation)
SX32 Expected */ To End Block Comment
SX33 Expression is Too Complicated. Break into Parts?

Additional Topics
Errors Related to Instruction Names or Usage

I1 Expected Instruction - <<string>> Is Not Recognizable As One
I2 Expected Instruction Here
The most common cause of these two errors are mis-spelled instruction names. If
you are doing an INPUT, READ, or DATA(UNIT=INPUT) instruction, you may also
have provided more data than the input instruction is expecting, causing rats to
try to interpret a line of data as a new instruction.
I3 Instruction Format Should Be <<instruction format>>
I4 <<instruction>> Requires RATS Pro Version
I5 RATS Linear Doesnt Include Instruction <<instruction>>
I6 <<instruction>> Required X Windows Version
Errors Involving Instruction Parameters

P1 / For Parameter Pairs Should Be Followed by Space
If using the / symbol for the default range, make sure you have a space before and
after the /.
P2 Second Half of Parameter Pair Is Absent
If you supply a value for a start parameter, you must also supply a value for the
end parameter (use * for either parameter to use the default value for that pa-
rameter).
P3 Instruction <<instruction>> Requires Parameter <<parameter>>
P4 <<parameter string>> is Invalid - Choose From <<list of choices>>
P5 <<parameter choice>> Cannot Be Used with <<parameter choice>>
Errors Related to Options

OP1 Looking For Option Field or White Space. This Character Is Unex-
pected.
OP2 Expected , to Separate Options or ) to End Option Field
Use a comma to separate each option in an option field, and close the option field
with a right parenthesis.
OP3 This Instruction Does Not Have An Option <<option>>
OP4 The Option <<option>> Is Already Set By This Instruction
You have probably listed the same option twice.
OP5 Form NO<<option>> is Not Permitted For Option <<option>>
OP6 Expected Option Here
OP7 <<argument>> Is Not One of the Choices for Option <<option>>
OP8 Option <<option>> Needs = Here
OP9 Options <<option>> and <<option>> May Not Be Used Together
OP10 Instruction <<instruction>> Requires Option <<option>>
If you have problems with an option field, carefully check the syntax and the list of
options for that instruction in the on-line help or the Reference Manual.
Errors Involving Supplementary Cards

SC1 Expected Supplementary Card (# ....) Here
SC2 Require Supplementary Card Field <<name>> (Position <<number>>)
SC3 Expected <<number>> Entries on Supplementary Card, Got Only <<num-
ber>>

Additional Topics
SC4 Cant Use / or * On This Type of Supplementary Card

SC5 Use of * Illegal At This Location
SC6 Need a LIST Instruction to Use CARDS
SC7 LIST Needs At Least <<number>> Entries, Had Only <<number>>
Errors Involving Data Series

SR1 ALLOCATE Instruction Needed Before Series or Equations Can Be Used
SR2 FREQUENCY Instruction Needed Before Complex Series Can Be Used
You must do an ALLOCATE instruction before defining or using real-valued series,
and you must use do a FREQUENCY instruction before defining or using complex-
valued series. If you are using procedures that work with series, you may need to do
your ALLOCATE instruction before SOURCEing in the procedure.
SR3 Tried to Use Series Number <<number>>, Only <<number>> Are Avail-
able
SR4 Tried To Use Series Number <<number>> (-series n1 n2 triples are
no longer legal)
SR5 <<writing/using>> Range <<entry>> to <<entry>> of Series <<se-
ries>> (Incorrect start,end or SMPL)
SR6 Missing a Necessary Parameter. Check Instruction Syntax
SR7 Block with <<number>> Elements is Too Big for This Version of RATS
SR8 Badly Formed TO triple
SR9 High TO Low Range: <<entry>> TO <<entry>> Is Illegal
SR10 Missing Values And/Or SMPL Options Leave No Usable Data Points
This can occur if you try to apply an operation to a series that contains no data,
or try to run an estimation where all observations are dropped because of missing
values in one or more series. Check all of the series involved using TABLE or PRINT.
Watch for transformations, such as taking logs or roots of negative numbers, that
can eliminate data. Make sure you are not trying to start an estimation too early in
the sample range. For non-linear estimation, check for valid initial conditions for all
parameters, make sure you arent taking logs or roots of negative numbers, etc.
SR11 Illegal range, start = <<entry>> , end = <<entry>>. Are your pa-
rameters correct?
SR13 Need to Include an Entry Range. No Previous Instruction Has Set
the Default.
If you dont use an ALLOCATE (which is now optional, rather than required), you will
need to specify an explicit ending period on the first instruction in your program
that uses series (usually a DATA instruction). Otherwise, you will get this error.
Regression/Estimation Problems
REG1 Cannot Execute Unless Preceded by a Completed Regression
REG2 LINREG(CMOM) First Requires CMOM/MCOV without CORR, CENTERED or
MATRIX option
REG3 <<instruction>> Fails - No Prior Regression or Regression Was Dif-
ferent Size
REG4 Variable <<name>> is Not in the Prior Regression
REG5 Variable <<name>> is Not in the Prior CMOMENT
REG6 Matrix <<name>> Is Singular. Check for Collinearity in First
<<number>> <<rows or columns>>

Additional Topics
REG7 Variance is Zero

REG8 No Instruments Available. Use an INSTRUMENT instruction first
REG9 Model is Underidentified - <<number>> Parameters To Estimate and
Only <<number>> Instruments
REG10 First <<number>> Instruments Are Linearly Dependent Over Regres-
sion Range
REG11 Regression Position <<number>> Should Be Between 1 and <<number>>
REG12 SIGMA Is Singular/Not PSD At Row <<number>>. Too Many Equations
for Data Set Size?
For sur estimation, you must have more time periods per series than equation.
REG13 Singular Regressions - Check for Collinearity among Rows 1 to
<<number>>.
REG14 Unsatisfactory Initial Estimates. Setting Last MA to 0.
REG15 Series <<name>> Has All <<ZERO or NONZERO>> Values Over Estimation
Range. Are You Using Zero-NonZero Coding?
REG16 Initial regression uses all the degrees of freedom
REG17 Either use series list parameters or SCRATCH option.
REG18 METHOD=CORC/MAXL Cannot Be Used With Gaps/Missing Values Switching
to HILU/SEARCH
This is just a warning that rats is switching to a different estimation method for
the AR1 instruction due to missing values in the sample range.
REG19 Need at least <<number>> Autocovariances. Adding 0s at end of CO-
VARIANCES series
REG20 ITERATE/BOXJENK Cannot Be Used with Gaps/Missing Values
arima estimation methods cannot be applied if there are any missing values in the
estimation range. If you didnt expect to have missing values, check the start and
end dates used for the regression, and use PRINT to check the data. You can use
SAMPLE to compress out the missing data if desired.
REG21 Cant Freely Estimate SW Matrix with <<number>> Conditions. Have
Only <<number>> Observations
Errors Involving Matrices

MAT1 Matrix <<name>> Has Not Been Dimensioned
You must dimension an array before you can set or reference individual elements.
MAT2 Matrices with Dimensions <<rows x columns>> and <<rows x columns>>
Involved in <<operator>> Operation
MAT3 Matrix with Dimensions <<rows x columns>> Involved in <<operator>>
Operation. Need <<rows x columns>>
MAT4 Must Use the Form matrix=<<function>>(...) for this Operation.
The function returns an array as a result, so an array name must be provided on
the left side of the equal sign.
MAT5 Needed Matrix with Dimensions <<rows x columns>>, Got <<rows x
columns>> Instead
MAT6 Trying to Store <<rows x columns>> Matrix Into VECTOR
MAT7 Function <<function name>> Requires a <<matrix type>> Matrix As
Argument
MAT8 Non-Positive Argument for function <<function name>>
MAT9 Base Address for OVERLAY is Invalid (Subscripts Out of Range)

Additional Topics
MAT10 You Cant OVERLAY with a Series, Only with an Array

MAT11 You Cant DIMENSION a Series
MAT12 Syntax: EWISE array(I,J)=expression or array(I)=expression
MAT13 Store into Out-of-Range Matrix or Series Element
This can occur if, for example, you try to set element 11 of a matrix that has only
been dimensioned to have 10 elements. If you are doing a non-linear estimation
using recursive FRMLs, as when estimating garch models with MAXIMIZE, where
series are being used to hold values of sub-formulasthose series must be initial-
ized over the entire estimation range. Set them to a value over the full range.
MAT14 Non-invertible Matrix. Using Generalized Inverse for SYMMETRIC.
rats tried to invert a matrix, but the matrix was non-invertible. If the matrix was
of type SYMMETRIC, rats has applied a generalized inverse routine.
MAT15 Subscripts Too Large or Non-Positive
Array elements are numbered from 1 to the dimensioned size. Dont use zero or a
negative number, or a number larger than the dimensioned size, in referring to an
array element. If this appears as a result of a MAXIMIZE command, you most likely
have specified the entry range incorrectly, or else one of your formulas is trying to
store a result into a series or array element that has not been defined yetmake
sure that any series used to hold residuals or variances have been explicitly defined
over the entire estimation range.
MAT16 Column Choice of %ld is Invalid
MAT17 Cant Use %s Range of %ld to %ld in %s operation
Errors Relating to Equations

EQ1 Cannot Use Equation Number <<number>> (Value should be Positive)
EQ2 Equation Includes More Than One Explanatory Variable
EQ3 Equation <<name or number>> Has Not Been Defined
EQ4 Equation <<name or number>> Has At Least One Undefined or NA Coef-
ficient
EQ5 Equation <<name or number>> Does Not Include Variable <<name>>
EQ6 Variable <<name>> Is Not the Dependent Variable of Equation
<<name>>
EQ7 Variable <<name>> Has Zero Coefficient in Equation <<name or num-
ber>>
EQ8 Cannot Replace the Dependent Variable with a Lag
EQ9 ALLOCATE Equation <<number>> is Increased in Size. %COEFF is No
Longer Valid
EQ10 Equation <<name or number>> Needs Residuals for This Operation
EQ11 Need To Use MODIFY Instruction to Start Equation Modification
Non-Linear Estimation Problems

NL1 FRML References More Than 20 Deep. Do You Have a Self-referencing
FRML?
NL2 NONLIN Instruction is Required Before Non-Linear Estimation Can Be
Done
NL3 FRML <<name>> Has Not Been Defined
NL4 FRMLs Cannot Have Moving Average Parameters
NL5 FRML <<name>> was Defined Without a Dependent Variable

Additional Topics
NL6 NONLIN Parameter <<name>> Has Not Been Initialized. Trying 0

This is just a warning that you have not provided an initial value for the specified
parameter. RATS will try to do the estimation using 0.0 as the initial value. If the
estimation fails to converge, use COMPUTE to supply better initial values before
doing the estimation. Also, note that 0.0 may not be a valid initial value for some
parameters (if the formula takes the log of the parameter, for example).
NL7 NONLIN Parameter Has Invalid Form (Common errors: using = > or < for
== >= or <=)
Errors Involving Vector Autoregressions

VAR1 A SYSTEM instruction is needed first
VAR2 Need SYSTEM and VARIABLES instructions first
VAR3 <<name>> requires either SYSTEM(KALMAN) or KFSET instruction
VAR4 SYSTEM definition is incomplete
VAR5 KFSET Lists <<number>> Arrays, TVARYING Lists <<number>>. They
should be equal
VAR6 <<name>> Instruction Lists <<number>> Arrays, Should Have <<num-
ber>>
VAR7 Need ESTIMATE or KALMAN(STARTUP=xxx) First
You must either ESTIMATE the model or do KALMAN(STARTUP=date) to initialize
the Kalman Filter before using KALMAN to do updating.
VAR8 SYSTEM Lists <<number>> Equations, VARIABLES Lists <<number>>
Variables. They should be equal
VAR9 You Cant ADD and DROP on the same KALMAN instruction
VAR10 Cant DROP Entry <<number>> - Filter Only Includes Entries <<num-
ber>> to <<number>>
VAR11 COHISTORY can only be used with a single equation model
Compiled Mode Problems

CP1 END(}) is Improperly Placed. Match with BEGIN({),DO,DOFOR,LOOP,PRO
CEDURE,MENU,FIND
CP2 This Instruction Must Precede All Executable Instructions
CP3 Instruction is Illegal Outside Compiled Section
CP4 ELSE Must Come Immediately After IF block
If you are using an IFELSE structure and want to execute several statements
when the IF command is true, you must enclose those statements in { and }.
CP5 CHOICE Must Come Immediately After MENU or Another CHOICE block
CP6 Need CHOICE instruction to follow MENU
CP7 No DO, DOFOR, LOOP, WHILE, UNTIL or FIND for BREAK/NEXT
CP8 Label <<name>> Is Undefined
CP9 Label <<name>> Is Already In Use
CP10 Statement Label is Illegal Here
CP11 Statement Label <<name>> is Unreachable From BRANCH
CP12 DO Loop Has Increment of 0
CP13 Switch and Choice Options Cannot be Done By Reference
CP14 Please Do Separate DIMENSION Instruction in Compiled Section
Inside a procedure or loop, you cannot include dimension fields on a DECLARE in-
structionuse a separate DIMENSION instruction in those cases.

Additional Topics
CP15 This Supplementary/Data/Text/CARDS card is not Used

CP16 Fatal Error - RATS Program Bug in Executing Compile
CP17 PROCEDURE Must be Initial Statement in a Compiled Section
CP18 <<name>> is not the Name of a PROCEDURE (Did you forget to
SOURCE?)
CP19 <<name>> is not a PROCEDURE Parameter
CP20 <<name>> may be used only in a PROCEDURE
CP21 SOURCE Instructions/Executes Nested Too Deeply. Recursive Proce-
dure?
CP22 <<name>> may not be used in a compiled section
CP23 Option abbrevation <<abbrev>> already in use in procedure <<name>>
Miscellaneous Errors Relating to Specific Instructions

X1 Syntax is CALENDAR(options) year month day with this frequen-
cy
X2 Dont Use CALENDAR 1 1 1 for Undated Data. See page 1-5 in the
Manual
X3 You Probably Mean <<number>>:1. Note that year:1 is needed with
Annual Data
Dont use just a year number, such as 1999, as a date reference when working
with annual data. Use 1999:1 instead.
X4 <<string>> is illegal picture. Pictures must look like *, ##, *.##
or ##.##
See DISPLAY for details on picture codes.
X5 ENTER(VARYING) Requires a Single VECTOR[...] type Variable
X6 You cannot ENTER an Array with more than one object on the list
X7 TREND=EXPONENTIAL/SEASONAL=MULTIPLICATIVE Cannot Be Used with Non-
Positive Data
X8 Unstable ESMOOTH Parameters. Try Different Values.
X9 You should specify an explicit seasonal with this frequency data
X10 Use the TABLE Instruction to Get Statistics on Several Series
X11 Series Lists Have <<number>> and <<number>> Variables. They Should
be Same Size
X12 Cannot Do COMPRESS with This Combination of Weights
X13 Redundant Restrictions. Using <<number>> Degrees, not <<number>>
See Section 6.2 of the Users Guide.
X14 Sorry, QUERY Cant Handle Arrays
X15 Expected <<number>> More Data Values
X16 MEDIT only accepts REAL Valued Arrays
X17 Problem and network do not match
The number of input and/or output series, hidden nodes, or DIRECT option setting
do not match those used to generate the memory vector for this neural network.
X18 Input pattern has different range than test data
The range of values in the input series exceeds the maximum or minimum values
specified when the network was first estimated.
X19 Unable to open X Window for display
X20 This Item Type Needs an INTEGER for output value

Additional Topics
X21 You Need a SERIES, LIST, STRINGS, or REGRESSORS Option to Define

the List
X22 DISTRIB=GED Cant Be Used with Multivariate Model. Switching to
Normal
X23 FILTER(TYPE=<<type>>) Needs <<value>> for WIDTH/SPAN option
Forecasting/Simulation Errors
FO1 Equations <<name or number>> and <<name or number>> both have de-
pendent variable <<name>>
Each equation in the system must have a unique dependent variable. See Section
11.2 in the Users Guide.
FO2 Need the equation parameter on each supplementary card
FO3 You need to do THEIL(SETUP) first
FO4 Identities must come last
FO5 SHOCKS array has <<number>> columns. Should have <<number>>
FO6 SIMULATE(SETUP) is no longer needed. See description of SIMULATE.
FO7 Gauss-Seidel solution is explosive. Check model. Try DAMP option.
FO8 Gauss-Seidel did not converge in <<number>> Iterations. Criterion
= <<value>>.
See Section 11.2 in the Users Guide for tips on getting your simultaneous equation
model to converge.
FO9 Simultaneous block is singular. Check your equations.
FO10 Either actual or forecast is NA. Skipping step(s).
FO11 IMPULSE, ERRORS and HISTORY cant handle model with FRMLs.
Memory/Internal Errors
M1 <<parameter or option>> = <<number>> is Illegal. Value should be
between <<number>> and <<number>>
M2 <<parameter or option>> = <<number>> is Illegal. Value should be
<<required value type>>
M3 RATS Internal Error!!! This should never happen
M4 A memory request for an additional <<number>> bytes cannot be
satisfied
M5 <<name>> has Value <<number>> , Must Be <<number>>
M6 User Abort
M7 Untrapped Floating Point Error: Overflow or NA
M8 Untrapped Miscellaneous Error
M9 RATS Internal Error!!! Save Your Work Now Using Save As
M10 Classroom RATS Limit of <<number>> Data Points Exceeded. You Need
<<number>>
M11 Integer divide by zero
M12 Polynomials cant have negative powers
M13 Table Must be 2xN RECT[INTEGER]
Dynamic Linear Modeling Errors

DLM1 Rank of Prediction Error Variance < Number of Observables
DLM2 No Observations Produce Valid Output. Check Data and Initial Val-
ues

Additional Topics
DLM3 The G Option Requires Both A and SW to Solve Out Initial Condi-
tions
DLM4 With TYPE=CONTROL, B Option is Required
DLM5 Probable Model Error. Diffuse prior was not reduced to zero rank
Linear/Quadratic Programming Errors

LPQ1 Contradictory constraints
LPQ2 Unbounded problem
LPQ3 General failure of LPQP
LPQ4 Simplex method does not converge
LPQ5 Active set method does not converge
LPQ6 All elements of b vector must be non-negative
Input/Output Errors
IO1 Unit <<I/O unit name>> is already open
IO2 Unit <<I/O unit name>> Is Already In Use For <<access type>>
IO3 Unable to open <<file name>>
Check to make sure the file isnt already open in another application.
IO4 Unit <<I/O unit name>> does not exist
IO5 Unit <<I/O unit name>> is permanent. You cannot <<OPEN or
CLOSE>> it
IO6 Units <<I/O unit name>> and <<I/O unit name>> are already
aliases
IO7 Unit <<I/O unit name>> is not open
IO8 An error occurred while writing to the disk, the disk may be full.
IO9 Invalid input <<string>> on line <<number>> while processing se-
ries <<name>> entry <<number>>
Usually occurs when a file being read with DATA includes text information. This is
only a warningrats will set the indicated entry to NA and continue reading data.
See the end of Section 2.8 in the Users Guide.
IO10 Unexpected end of file while processing line <<number>>. (series
<<name>> entry <<number>>)
IO11 Input/Output error on <<operation>> <<error code>>
IO12 A date is missing from series <<name>> after entry <<number>>
IO13 <<string>> is not a valid date on line <<number>>
IO14 Data are not in ascending date sequence
IO15 Using dates from the data file
IO16 Input file does not contain dates
IO17 Header names <<number>> series while there are <<number>> series
on file
The file should have one series name per column or row of data. This indicates that
there are too few or too many series names.
IO18 Unexpected end of line while processing line <<number>> (series
<<name>> entry <<number>>)
IO19 The file does not have a valid <<format type>> format
IO20 The file contains no valid data
IO21 RATS only process the first worksheet in a spreadsheet file
IO22 Format has nothing to read or write

Additional Topics
IO23 Format <<character>> and element <<type>> are of incompatible

types
IO24 FORTRAN Hollerith Field Length of <<number>> is Illegal/Clearly
Wrong
IO25 Illegal FORTRAN format <<format string>>
IO26 Unbalanced parenthesis in FORTRAN format
See Section 2.9 in the Users Guide on using FORTRAN formatting codes.
IO27 The file is not in RATS format.
IO28 File appears to be damaged
IO29 No file is currently being edited
IO30 There is no series <<name>> on the file
Check the list of series on the DATA instruction versus the list on the file. Or, omit
the list of series and add the VERBOSE option on DATAthis will read in all series
on the file and report their names.
IO31 Series <<number>> already exists on the file
IO32 At most two comments can be attached to a series
IO33 The file being edited is not version 4
IO34 Date Type is Not Supported by RATS Version 3 Format
IO35 Unable to rename file <<name>> to <<name>>
IO36 Unable to delete file <<name>>
IO37 Cannot use UNIT=INPUT with this format
IO38 Cannot convert between these Frequencies
IO39 Unexpected End of File. Only <<number>> Values Were Processed
IO40 Expected <<type>>. Got <<string>>.
IO41 <<name>> is not the Name of Series in Working Memory
IO42 Binary file contents do not match READ requests
IO43 Format <<format>> is illegal or not supported
IO44 Date scheme <<string>> is not currently supported on this file
type
IO45 Data can not be appended to unit <<I/O unit name>>
IO46 Separate year and period or year month day on UPDATE
IO47 <<format>> cannot process a <<string>>
IO48. File with <<number>> columns might not be readable by other pro-
grams
FAME Format Data Errors

FM1 FAME HLI error <<format>>
X11 Seasonal Adjust Errors (Professional Version Only)

X1101 X11 can take monthly or quarterly series only.
X1102 Series with missing values cannot be used in X11.
X1103 Series has non-positive values. Switching to additive adjustment.

7. Manual Errata
T his section lists errors found in the rats documentation since the manuals were
printed. In many cases, these have been corrected in the pdf versions of the
manuals provided with the software.
Manual Errata
Additional Topics
7.1 User's Guide

On page UG-354, the first paragraph should say "large positive" not "large negative".
Fixed in the pdf versions shipped with rats 8.0, but the error is present in hard
copy versions.
7.2 Reference Manual

DATA Instruction
In the list of formats supported for the FORMAT option, the text should read DTA
rather than DATA. DTA is the format for Stata files (fixed in the pdf version).
ESTIMATE Instruction
The Variables Defined list should say %XX, not %X (fixed in pdf version).
Functions
The following functions were not listed in Section 2.
%CHOICE option (still need to add to ref manpdf)
%EQNRESID and %EQNRVALUE, for extracting a residual value from an equation.
(Fixed in PDF).
%FILE option (still need in Ref man)
%GRPARM in Section 2 (is documented in references under instructions). need to
add to ref man sec 2
The following are documented in the main alphabetical listing, but were not listed in
the summary list of Distribution/Probability functions at the beginning of Section 2:
%LOGGEVDENSITY and %LOGGPDENSITY. (Fixed in the PDF)

Bibliography
Engle, R. (1974). Band Spectrum Regression. International Economic Review, Vol. 15, pp. 1-11.
Hayashi, F. (2000). Econometrics. Princeton, New Jersey: Princeton University Press.
Hoerl, A.E. and R.W. Kennard (1970). Ridge Regression: Biased Estimation for Non-Orthogonal
Problems. Technometrics, Vol. 12, pp. 55-82.
Judge, G.G., W.E. Griffiths, R.C. Hill and T-C. Lee (1980). The Theory and Practice of Economet-
rics. New York: Wiley.
Shiller, R.J. (1973). A Distributed Lag Estimator Derived from Smoothness Priors. Econometrica,
Vol. 41, pp. 775-788.
Sims, C.A. (1974). Seasonality in Regression. Journal of the American Statistical Association, Vol.
69, pp. 618-626.
Theil, H. (1971). Principles of Econometrics. New York: Wiley.

RATS 8 Additional Topics

Uploaded by

Copyright:

Available Formats

RATS 8 Additional Topics

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

RATS 8 Additional Topics

Uploaded by

Copyright:

Available Formats

RATS VERSION 8

The RATS Interface

4.2 Ridge and Mixed Estimators.......................................................................................64

6. RATS Error Messages 87

4 RATS: Additional Topics

1.1 Menu Operations

The File Menu

6 RATS: Additional Topics

RATS: Additional Topics 7

The Edit Menu

Copy, Copy Special..., Copy as TeX

8 RATS: Additional Topics

Show Last Error

RATS: Additional Topics 9

The View Menu

Show as B&W/Show as Color

Time Series Graph

10 RATS: Additional Topics

RATS: Additional Topics 11

The Data/Graphics Menu

Data (RATS format)

Data (Other formats)

Create Series (Data Editor)

Moving Window Statistics

X11 (Pro version)

12 RATS: Additional Topics

Scatter (X-Y) Graph

Data Browsers: FRED Browser

Data Browsers: Haver DLX

Data Browsers: FAME Database

RATS: Additional Topics 13

The Statistics Menu

Limited/Discrete Dependent Variables

Panel Data Regressions

Recursive Least Squares

(Kernel) Density Estimation

14 RATS: Additional Topics

The Time Series Menu

Unit Root Tests

Box-Jenkins (ARIMA) Models

RATS: Additional Topics 15

The Window Menu

Keep Open on Close All

Close All Graphs

Close All Reports

Use for Input

Use for Output

Paste AND Execute

16 RATS: Additional Topics

1.2 Toolbar Icons

Text Editing Windows

(Open) Shortcut for FileOpen. Opens a file.

RATS: Additional Topics 17

RATS Data File Windows

, (Un/Redo) Undo and Redo operations.

(Rename) allows you to change the name of the selected series.

Series Edit Windows

18 RATS: Additional Topics

1.3 Preference Settings: Customizing the RATS Editor

The Editor Tab