RATS

VERSION 8

INTRODUCTION

RATS
VERSION 8

INTRODUCTION

Estima

1560 Sherman Ave., Suite 510 Evanston, IL 60201 Orders, Sales Inquiries General Information Technical Support Fax: 8008228038 8478648772 8478641910 8478646221 Web: General E-mail: Sales: Technical Support: www.estima.com [email protected] [email protected] [email protected]

 2010 by Estima. All Rights Reserved. No part of this book may be reproduced or transmitted in any form or by any means without the prior written permission of the copyright holder. Estima 1560 Sherman Ave., Suite 510 Evanston, IL 60201 Published in the United States of America First Printing: November, 2010 PDF version revised September, 2011 for Version 8.1

Preface

Welcome to Version 8 of rats. This is the most extensive revision of the program and documentation in many years. When we set out to craft the new set of manuals, we started with a mission statement: that the job of software like rats is to help you process numerical information and get it into your final document as accurately as possible. With that in mind, we created this new book, intended to bring a user quickly up to speed, whether that person is an experienced user of rats or other statistical software, or an undergraduate new to statistical programming. Version 8 offers new and simpler ways to get information from rats into a final document, and that is emphasized in each of the short examples included here. In 2002, we began a project of implementing (all) the worked examples from a wide range of textbooks. We now number more than twenty texts in the collection, including everything from introductory econometrics and time series books to graduate level books covering almost the full range of modern econometrics. In addition, we have an ever-growing number of replication files for important papers. Our experience with this (over 1000 running examples) has led to great improvements in the program. Were grateful to Kit Baum, Tim Bollerslev, Chris Brooks, Kai Carstensen, Richard Davis, Steve De Lurgio, Frank Diebold, Jordi Gali, John Geweke, Bill Greene, Jim Hamilton, Fumio Hayashi, Andrew Mountford, Timo Tersvirta, Ruey Tsay, Harald Uhlig, Marno Verbeek, Mark Watson and Jeff Wooldridge for their help in providing data and programs, and checking results. A special tip of the hat goes to Bruce Hansen and Mark Watson, who not only post their code and data, but are very quick to act on questions about them. The new Users Guide has been trimmed down a bit by moving its first three chapters into this Introduction, but the important topics of State-Space Models and Switching Models and Structural Breaks have been given much more extensive coverage and now have their own chapters. ARCH/GARCH has also been elevated to chapter status. Some material from our recent e-courses on Bayesian Methods, State-Space Models and Vector Autoregressions has been worked into the updated documentation, and we benefitted from comments from the participants in those. Thanks to all those in the rats community who have contributed ideas, suggestions and procedures for others to use. Special thanks go to Walter Enders for writing the popular e-book, The RATS Programming Language, Rob Trevor for setting up and maintaining the rats discussion list, and to Paco Goerlich and Norman Morin for the numerous procedures theyve written. Jonathan Dennis from the University of Copenhagen (the principal programmer of the cats program) had a major role in designing and refining the REPORT and DBOX instructions. Chris Sims, as always, had a special role in influencing the direction of our coverage of time series. Tom Maycock handled the laborious task of converting the existing documentation over to a new layout program. You will note the very extensive cross referencing that this made possible. He is also responsible for the bulk of the technical support questions and maintains our extensive web site. To say that hes invaluable would be an understatement.

I would also like to thank my wife Robin and daughters Jessica and Sarah for their patience; Robin in particular for putting up with the 4:00 a.m. manual writing sessions. Thomas A. Doan October 2010

Table of Contents
Preface. ........................................................................................................................ Intiii .

1. Getting StartedA Tutorial

Int1

1.1. Using.the.RATS.Manuals........................................................................................Int2 1.2. Example.One:.The.RATS.Editor.............................................................................Int3 . 1.2.1. Learn.More:.Input.and.Output.Windows. ...................................................Int7 . 1.3. Example.Two:.Working.With.Data. .........................................................................Int8 . 1.3.1. Learn.More:.Series.Edit.Windows............................................................Int13 1.3.2. Learn.More:.Report.Windows. .................................................................Int15 . 1.4. Example.Three:.Transformations.and.Regressions..............................................Int17 1.4.1. Getting.the.Data.In...................................................................................Int18 1.4.2. Check.the.Data. .......................................................................................Int23 . 1.4.3. Data.Transformations.and.Creating.New.Series......................................Int25 1.4.4. Estimating.Regressions. ..........................................................................Int27 . 1.4.5. Learn.More:.Procedures. .........................................................................Int32 . 1.4.6. Learn.More:.Reading.Data.......................................................................Int34 1.4.7. Learn.More:.Annotated.Regression.Output..............................................Int35 1.4.8. Learn.More:.The.Series.Window..............................................................Int38 1.4.9. Learn.More:.Arithmetic.Expressions. .......................................................Int40 . 1.5. Example.Four:.Time.Series.Graphs.and.Analysis.................................................Int43 1.5.1. Filtering.and.Smoothing...........................................................................Int44 1.5.2. Graphing.the.Data....................................................................................Int46 1.5.3. Detrending,.Exponential.Smoothing,.Forecasting....................................Int48 1.5.4. Regression-based.Forecasting................................................................Int53 1.5.5. Learn.More:.Graph.Windows...................................................................Int56 1.5.6. Learn.More:.Long.and.Short.Program.Lines............................................Int57 1.5.7. Learn.More:.RATS.Program.Files.(RPF). ................................................Int58 . 1.5.8. Learn.More:.The.PRINT.Instruction. ........................................................Int59 . 1.5.9. Learn.More:.Data.Transformations. .........................................................Int60 . 1.5.10. Learn.More:.Forecasting........................................................................Int64 1.6. Example.Five:.Cross-Sectional.Data....................................................................Int66 1.6.1. Reading.the.Data.....................................................................................Int66 1.6.2. The.SMPL.Option:.Selecting.Sub-Samples. ............................................Int67 . 1.6.3. Testing.Regression.Restrictions...............................................................Int68 1.6.4. X-Y.Scatter.Plots:.The.SCATTER.Instruction...........................................Int71 1.6.5. Learn.More:.Comment.Lines.and.Comment.Blocks. ...............................Int73 . 1.6.6. Learn.More:.Linear.Regression.Instructions............................................Int75 1.6.7. Learn.More:.Error.Messages. ..................................................................Int79 . 1.6.8. Learn.More:.Entry.Ranges.......................................................................Int81 1.7. Example.Six:.Non-Linear.Estimation....................................................................Int84 . 1.7.1. Learn.More:.Non-Linear.Estimation.Instructions......................................Int89

Table of Contents
1.7.2. Learn.More:.Working.with.Matrices..........................................................Int90 1.8. The.RATS.Programming.Language......................................................................Int91 1.9. What.Next?...........................................................................................................Int93

2. Dealing With Data

Int95

2.1. The.Tools..............................................................................................................Int96 2.2. Data/Copy.Formats...............................................................................................Int98 2.3. Where.Are.Your.Data.Now?................................................................................Int101 2.4. The.Data.(Other.Formats).Wizard. .....................................................................Int103 . 2.5. Changing.Data.Frequencies...............................................................................Int104 2.6. Missing.Data.......................................................................................................Int107 2.7. RATS.Format......................................................................................................Int110 2.8. Spreadsheet.and.Delimited.Text.Formats...........................................................Int114 2.9. Text.Files.............................................................................................................Int120 2.10. File.Handling.Tips.and.Tricks............................................................................Int123

3. Graphics

Int125

3.1. Graphics..............................................................................................................Int126 3.2. Working.with.Graph.Windows.............................................................................Int127 3.3. Preparing.for.Publication.....................................................................................Int129 3.4. Graph.Styles.and.Style.Numbers........................................................................Int130 3.5. Labeling.Graphs..................................................................................................Int131 3.6. Keys.(Legends)...................................................................................................Int133 3.7. Overlay.(Two-Scale.or.Two-Style).Graphs..........................................................Int134 3.8. SPGRAPHMultiple.Graphs.on.a.Page............................................................Int136 3.9. GRTEXTAdding.Text.to.Graphs......................................................................Int139 . 3.10. Graphing.Functions...........................................................................................Int140 3.11. Highlighting.Entries...........................................................................................Int141 3.12. Fan.Charts........................................................................................................Int142 3.13. GCONTOURContour.Graphs........................................................................Int143 3.14. GBOXBox.Plots.............................................................................................Int144 3.15. Miscellaneous.Graph.Types..............................................................................Int145 3.16. Graph.Style.Sheets...........................................................................................Int149 3.17. Batch.Graph.Generation...................................................................................Int152 3.18. Choosing.Fonts.................................................................................................Int154

4. Resources

Int155

4.1. Installing.RATS. ..................................................................................................Int156 . 4.2. Additional.Documentation...................................................................................Int157 4.3. Examples.and.Procedures..................................................................................Int158 4.4. RATS.Forum.and.Online.Courses. .....................................................................Int159 . 4.5. Technical.Support. ..............................................................................................Int160 .

Bibliography Index Intvi Introduction to RATS

Int163 Int165

1.. Getting.StartedA.Tutorial T

he object of this chapter is to get you up and running as quickly as possible. The first thing to do is to get your program installed. If you havent done that, and need some help with it, see page Int156 in this book. Once youre ready to go, you should start working through this book. It introduces you to how rats gets information from you and supplies results to you. We provide several examples designed to give you a taste of what you can do with the program. To allow you to find your own pace, we have kept the examples fairly short; just long enough for each to introduce a few key concepts. The main discussion in each is intended to be followed in sequencewhats discussed in one is assumed in the next. However, each of the examples is followed by one or more Learn More subsections. These you can skip on first reading, particularly if youre never used a statistical package like rats before. These Learn More segments do include information that you will probably need eventually, so you probably will want to come back to them eventually. If you are very proficient at working with statistical languages, then you might want to go through these the first time, since many of the ideas are shared with other such programs.

Using the RATS Manuals The RATS Editor Working with Data Transformations and Regressions Time Series Graphs and Analysis Cross-Sectional Data Non-Linear Estimation Computations, Loops, and Procedures

Getting StartedA Tutorial 1.1 Using the RATS Manuals

There are three main manuals: this, the Users Guide and the Reference Manual. rats is a program with a very broad range of uses. The idea behind the Introduction to rats is to demonstrate the most important concepts of the program: the different window types and what you can do with them, the wizards, the data handling and graphics. If you are new to rats, we strongly recommend that you take the time to work through the examples in this chapter. Many of the examples are followed by Learn More sectionsthese are less critical on first read for someone new to the software, but you should eventually come back to them to become more proficient. The Users Guide describes how to use rats for various types of analysis (with extensive examples). In many cases, you can skip directly to the chapters in that which are of greatest interest. The Reference Manual provides a detailed description of each instruction and function in rats, as well as other elements of the rats language, and is designed to be used as the name describes: as a reference in case you need to know more about an instruction. The Users Guide also includes a combined Index covering the Introduction, Users Guide, and Reference Manual.

Conventions Used in the Manuals

Examples Instructions

We use the following font and style conventions in the rats manuals: Examples that can be typed in and executed as written are presented in bold Courier font. For example: linreg(define=req) rate 1960:1 1995:8 resids_rate # constant ip grm2 grppi{1} Instructions are the fundamental building blocks of the rats languagethey instruct rats to perform an action. In the example above, the instruction LINREG performs a linear regression. In body text, instruction names are presented in uppercase bold Courier font. For example BOXJENK, and DATA. Parameters are used to provide additional information to an instruction, such as the series names RATE and RESIDS_RATE in the example above. When describing instructions, the names of the available parameters are presented in lowercase italicized Courier. For example, start, end, series. Options are used to modify how instructions behave. In the sample above, DEFINE defines an equation. In body text, options are presented in uppercase Courier font (such as DATES and PRINT). Bold is used in option lists in the Reference Manual. Variables (series, matrices, scalars, and so on) are generally presented in uppercase Courier, such as RATE and IP. In some cases, mixed-case names are used for readability.

Parameters

Options

Variable Names

Int2

Introduction to RATS

Getting StartedA Tutorial 1.2 Example One: The RATS Editor

To introduce you to rats, we will take you step-by-step through several sample rats sessions. Well describe the key menu operations and rats instructions, and give you a feel for using the program. If you have not already installed rats on your computer, see page Int156 for installation instructions.

The RATS Editor

First, start the program in interactive mode. For Windows, open the WinRATS 8 folder on the Start menu or desktop and click on the winrats icon. For macrats, click on the macrats program icon. For unix, click on the ratsx file, or type ratsx at the command prompt. This loads the rats interactive interface, which we call the rats Editor. You will see the rats menu bar, the toolbar, and an empty worksheet window called NONAME00.TXT. (If you dont see this, select New...Editor/Text Window from the File menu or use the <Ctrl>+N keystroke). This environment allows you to type in and execute commands; use menu-driven wizards to perform tasks; save or print program files and output; export information to various types of files; display, save, and print graphs; and more. If you look at the File and Edit menus, youll see familiar operations, like Save and Print on File, and Cut, Copy and Paste on Edit; and they all work as you would expect (short-cut keys as well). What you wont find are operations for Bold, Italics, Centering, etc. The rats editor isnt intended to be used to produce a final document; its job is to help you process numerical information and get it into your final document as accurately as possible. While you do have some choice of font, its from the very limited number of monospaced fonts (Courier is by far the most common of these). In a monospaced or typewriter font, all characters are the same width, making it easy to line up different types of output. rats also produces output which is in specially formatted tables, but those are separate from the main editor window.

Input and Output Windows

rats allows you to have multiple windows open at one time, but you can only execute

instructions from the window that is designated as the Input Window. This window will have {i} appended to the window title, both on the windows title bar, and on the window list in the Window menu. Similarly, any output generated goes to the window designated as the Output Window, which will have {o} appended the name. This can be a separate window, or it can be the same as the Input Window. When you first start the program, the NONAME00.TXT window should be set as both input and output, shown by {io} in the window title bar.

Introduction to RATS

Int3

Getting StartedA Tutorial

Executing Instructions
In this tutorial, much of the work is done using point-and-click operations with menus and dialog boxes. However, we will also show you how to type in and execute instructions directly, which is often the fastest way to accomplish many tasks once you have some familiarity with the language. In general, you can execute an instruction by just typing it into the Input Window and hitting the <Enter> key. You can also execute a block of instructions by selecting (highlighting) the lines you want to execute using the mouse or the keyboard, and then hitting <Enter> or clicking on the Run icon . rats will execute all of the instructions, in order, from top to bottom.

What Should You Type In?

We put large pointers ( ) in the margin next to all the instructions (or groups of instructions) that we want you to enter and execute. We will also sometimes show you other sample uses of these instructions. Some of these will work with the sample data sets, but others will not. For now, we recommend that you only type in the instructions marked with a pointer. The instructions we discuss in this section are provided on a file called ExampleOne. rpf (rpf stands for rats program file). If you encounter any difficulties, you can open that file (using FileOpen) to see exactly how the instructions should look.

The DISPLAY Instruction

display "Hello, World"

To get started, type the line below into the blank input window and then hit the <Enter> key (which tells rats to execute the line you just typed):
rats should display the text Hello, World on the next line (or to your output win-

dow if using separate input and output windows). As you can see, DISPLAY does just what its name impliesit displays information. You can also use it to display the results of computations. For example, a popular econometrics textbook includes the following expression to compute by hand the least squares estimator for the intercept in a regression (Hill et al, 2008, page 22): (1) b1 = y b2 x = 283.5737 (10.2096)(19.6408) = 83.4160

display 283.5737-10.2096*19.6048

You can do this computation easily using DISPLAY. Try typing in the line below (with no spaces inside the expression), and hitting <Enter> to execute it:
rats will do the computation and display the result: 83.41653

Note that the * symbol is the multiplication operator. The addition and subtraction operators are the usual + and - . Other operators include / for division and ^ for

Int4

Introduction to RATS

Getting StartedA Tutorial

exponentiation. We describe the full set of operators in Section 1.4.9. Multiplication is never implied, the way it is in the algebraic expression above, so you must use the * to multiply two terms.

Rounding and Using Full Precision

The textbook actually reports a slightly different result (83.4160). Thats because the numbers presented in the book for equation (1) (and used as input in our computation) were themselves rounded versions of the numbers actually used to produce the result shown in the book. rats (and most other statistical software) does its calculations in double precision, which gives roughly fifteen significant digits. Now no practical data set records numbers to anything like that level of accuracy, so if the data going in are only good for four digits, its pointless to report results at fifteen digits. However, whenever possible you should keep all the intermediate calculations in full precision: in short, dont round intermediate values. The software is already rounding at fifteen, which sometimes isnt even enough for very poorly-behaved data. DISPLAY picks the format itself, and usually errs on the side of too many digits. If you want this number rounded to four decimal places, you can use picture codes, which provide a template for numerical output. Insert the picture code before the number or expression. For example, edit your original instruction by inserting ##.#### (spaces on either side) before the expression: Hit <Enter> to re-execute this line. The cursor can be anywhere in the lineit doesnt have to be at the end of the line (youre executing the line, not inserting a line break). As we said before, the job of the rats editor is to help you get information into your final document as accurately as possible. rats comes with over 1000 examples out of major textbooks. The most common type of error in those textbooks comes from taking a number computed by the software and manually re-typing it (incorrectly) into the manuscript. If you want a value, let rats do the rounding, and use copy and paste to get it into your document. Select the text you need, and do a copy operation. The text copy is so important, there are four ways to do it: the keyboard shortcut (<Ctrl>+C), the copy toolbar icon , the standard menu operation EditCopy and Copy on the contextual (right click) menu. DISPLAY can also show more than one calculation, and more than one field (which is why we wanted you to be careful about spaces). For instance, you can edit your line to add a description and hit <Enter> to execute:

display ##.#### 283.5737-10.2096*19.6048

display "intercept=" ##.#### 283.5737-10.2096*19.6048

Introduction to RATS

Int5

Getting StartedA Tutorial

The COMPUTE Instruction
COMPUTE is another important instruction. If you need to do an algebraic calculation, but dont need to see the result, you will probably do that using COMPUTE. An alternative to doing our computation with DISPLAY would be (execute both lines):

compute intercept = 283.5737-10.2096*19.6048 display "intercept=" ##.#### intercept

This does the calculation and saves the result into a new variable called INTERCEPT, and then displays that. Your screen should now look like this:

Its important to note a major difference between rats and math packages like Gauss and matlab. With either of those, the first display would be done with 283.5737-10.2096*19.6048; (display is implicit since there is no other place for result), and the COMPUTE with intercept=283.5737-10.2096*19.6048; With math packages, the basic unit of input is an expression. With rats, its an instruction. DISPLAY and COMPUTE are two examplesthere are roughly 200 others, many of which do very sophisticated calculations requiring (internally) thousands of separate sub-calculations. Many of those have point-and-click wizards on the Data/ Graphics, Statistics and Time Series menus to help you apply them to your data.

Int6

Introduction to RATS

Getting StartedA Tutorial 1.2.1 Learn More: Input and Output Windows
The Run/Stop Buttons
While rats is executing instructions, the Run icon changes to the Stop icon . Run returns as soon as the task is complete and rats is ready to accept more instructions. In many cases, you wont even notice this because it finishes so quickly. However, if you are doing something which does take a long time, and you decide that you dont really want it to continue, you can click on Stop to interrupt the execution.

Editing Without Executing (Ready vs Local Mode)

Hitting the <Enter> key in the Input Window normally executes the current line. If you need to insert a line break without executing the line (perhaps you want to input several lines before executing them), do the following: On Windows and unix/linux systems, you can edit lines without executing them by putting rats into local mode, either by clicking on the (Ready/ Local) button or by hitting <Ctrl>+L. Clicking on the Ready/Local button again (or hitting <Ctrl>+L again) puts rats back into ready mode. Ready means that rats is ready to accept instructions for execution. On a Macintosh, you can just use the <Return> key, rather than the <Enter> key, to insert a line break without executing a command.

Configuring Input and Output Windows

You can switch the input or output functions to a particular editor window using the menu items WindowUse for Input and WindowUse for Output or the corresponding and toolbar icons. We tend to use the editor windows in one of two ways: 1. For quick work, we simply start up rats and work with the NONAME window the way we just did in the example. If you arent really interested in developing a set of rats instructions to execute later and just want some quick answers, this is the simplest setup. 2. To develop a new (or existing) program, we designate one window as the Input Window, and a second window as the Output Window. With input and output in separate windows, it is much easier to keep a copy of the instructions that we decide we like. The and toolbar icons provide easy ways to get this splitwindow setupthey automatically open a second window, designate it as the output window, and tile the two windows horizontally or vertically, respectively. If you decide that you really like the second setup, you can set rats to default to that on the Editor tab of the Preferences (Section 1.3 of the Additional Topics pdf)set the Open New Output Window at Start box.

Introduction to RATS

Int7

Getting StartedA Tutorial 1.3 Example Two: Working With Data

While COMPUTE and DISPLAY are important, if youre getting started with rats, you probably have some time series data that you want to use. There are three main types of data collection: time series, cross section/survey, and panel/longitudinal. While rats can handle all three, its specialty is time series data, and the specialized techniques used in working with it. In this example, we introduce some basic tools for working with time series data. The CALENDAR instruction discussed here is provided on the file ExampleTwo.rpf (everything else in this section is done using menu operations). In time series data, the basic unit of data is a time series, which is a time sequence of data that (usually) has a specific starting place in time and a specific reporting frequency. What makes time series data very different from other forms is that the data you use may come from many sourcesin the U.S., the Census Bureau, the Federal Reserve and the Bureau of Labor Statistics (among many others) are each the original sources of key data series. Time series can start at different times; in some cases, they can end rather abruptly (such as exchange rates involving Euro zone countries).

The Calendar Wizard

To work with time series data, we first want to tell rats the frequency and starting date for our data set. One way to do that is to use the Calendar Wizard. Select the Calendar operation from the Data/Graphics menu. rats will display the Calendar Wizard dialog box shown below, which allows you to set frequency and date information. The most common frequencies are in the Standard Frequencies group. In this case, we want quarterly data, beginning in the first quarter of 1998, so click on the Quarterly button and change the Year field to read 1998, like this:

Int8

Introduction to RATS

Getting StartedA Tutorial

Now click on OK. rats will generate the corresponding CALENDAR instruction: calendar(q) 1998:1 in the input window and execute it automatically. The Q in parentheses stands for quarterly. This is called an option, and this option tells rats that we will be working with quarterly data. When you use options for an instruction, the left parentheses must appear immediately after the instruction name, with no spaces in between. The date 1998:1 tells rats that our data begin in the first quarter of 1998. This value is called a parameter. The parameters generally are information that is either necessary or very commonly used, while the options are just thatsomething that you might or might not use. Note that you generally do not need to use the Calendar Wizard if you intend to use one of the Data Wizards (page Int18) to read data in from a file, since they will set the date scheme themselves.

Creating a Data Series

If at all possible, you do not want to type in data. Most of your work will be done using data read in from files or database connections, but for now, well have you type in some numbers to introduce you to tools for creating, viewing, and editing data. First, open the Data/Graphics menu and click on the Create Series (Data Editor) operation. rats will display an empty Series Edit Window. As this is a new series, the first cell shows na, for Not Available. We use the na abbreviation both in rats output and in the manuals to signify a missing value. Here are the numbers to enter (data for 1998 in row one, for 1999 in row two, etc.): 98.2 100.8 102.2 100.8 99.0 101.6 102.7 101.5 100.5 103.0 103.5 101.5 Start by typing the first number (98.2) into the cell at the top. Then, hit the right arrow key. This saves the value you typed in and moves to the next cell. Type in the second number (100.8), hit the right arrow key again, and so on. Note that the right arrow still works when you get to the end of a line. Because this is a time series, the lines are just to help organize your view of the datawhat we really have is a continuous sequence of values. When you get to the final value, just hit <Enter> (rather than the right arrow) to save the value. If you only want to type in a few of the values, thats fineeverything we discuss below will still work, although the results will look different. If you are reading this as a pdf, you can also copy and paste the data lines above into the editor. If you need to edit a value, click on (or navigate to) the cell you want to change, edit the value, and hit <Enter>.

Introduction to RATS

Int9

Getting StartedA Tutorial

Once you are done, the window should look like this:

The View Menu

The operations on the View menu provide ways to quickly examine data series and other variables. These operations do not generate rats instructions the way the wizards do. However, they are very handy for verifying your data and doing some quick glance analysis before beginning work in earnest. The second most common error that weve found in replicating examples (and the one that is probably most significant in practice) is where a data set is simply wrongfor instance, the data are for a different set of years than the author thought. And these are for papers which have been circulated, presented in seminars, peer-reviewed and published. Do yourself a favor: make sure your data are what you think they are before doing any major work with them. The View menu can help with this. Try the ViewTime Series Graph operation (or click on toolbar item). Youll see a simple graph of the data. This is not intended to be a production graph; instead, its a quick line graph of the data that can help you spot any obvious errors like missing a decimal point. If you see that you have a value thats clearly wrong, you should be able to go back to the series editor window and use the Max or Min toolbar icons ( and ), which move to the maximum and minimum values in the series, respectively, to correct the problem. While there are several other quick-view operations (the ones at the bottom of the View menu), the other one that is very handy is Statistics. If you choose this (or the icon), you will see a new type of window called a report that looks like this:

Int10

Introduction to RATS

Getting StartedA Tutorial

Report Windows

A Report Window is like a read-only spreadsheet. The descriptive text and values are arranged in columns, with longer strings covering (spanning) several columns. Output from rats will either be in the form of Report Windows or text in the Output Window, and sometimes both. You can copy information out of a Report Window and paste it into a spreadsheet or word processing program. See page Int15 for more.

Test Results

The report for Statistics includes four different tests: the t-statistic, skewness, kurtosis, and Jarque-Bera. For each of these, the left column gives the test statistic and the right gives the marginal significance level (also known as the p-value). This is the probability that a random variable with the appropriate distribution will exceed the computed value (in absolute value for Normal and t). Wherever possible, rats will compute the significance level for tests, which allows you to avoid using lookup tables for critical values. You reject the null if the marginal significance level is smaller than the significance level you want. For instance, a marginal significance level of .03 would lead you to reject the null hypothesis if you are testing at the .05 level (.03<.05), but not if you are testing at the .01 level. Here, the only of the four that we would reject under standard procedures is mean=0, which has a test statistic so far out in the tails that the p-value shows as just being 0. (If you paste it into a spreadsheet, youll find that its actually roughly 10-21.) This would be a good time to warn you about over-interpreting test results. rats will often compute tests and summary statistics which might be useful. Its up to you to decide whether they actually make any sense in particular situation. The three test statistics here that pass are all tests for the Normal distribution (the excess in excess kurtosis is the difference between the sample value and what would be expected if the data were from a Normal population). Should we therefore conclude that the data are from a Normal population? Absolutely not. Those are all test statistics computed under the assumption that the observations are independent. This is our raw time series data; we certainly dont expect it to be independent, and from the

Introduction to RATS

Int11

Getting StartedA Tutorial

graph its clear that it has a bit of a trend and a rather pronounced seasonal. If we were looking at the statistics on the residuals from an estimated model, then those last three tests might be interesting; here, they arent. Some other software packages actually do many more of these automatic tests than rats, which has led people misusing them to conclude that a model that made no sense at all was fine because it passed all the tests.

Saving the Data

You may have noticed that we havent yet talked about saving the data. If you do File-Save, you will be prompted for a series name (and comments), or if you do FileClose, you will first be asked if you want to save the changes; if you answer yes, you will be asked the same question about the name and comments:

For our purposes, enter QUARTER as the name as shown above. Leave the comments blank. If you do the menu operation ViewSeries Window, you will now see a Series Window listing all the series in memory. Well talk more about this when we have more data. For now, this should just have one line for the series QUARTER. At this point, we have saved the data out of the Series Edit Window into a rats series. If you need a permanent copy, you need to take one more step: exporting the data to a file, which can be done with File-Export or File-Save As.

Series Names

Series (and other variable) names in rats can be from one to sixteen characters long. They must begin with a letter (or %, though you should generally avoid those, as theyre used for names reserved by rats), and can consist of letters, numbers, and the _ , $ and % symbols. Variable names arent case-sensitive. Generally, youll want to choose names which are just long enough so you will know what each represents. In most cases, your data files will include names for the series stored on the file, and rats will use those names when it reads in the data. Thus, the series names on the file should conform to the restrictions described above.

Int12

Introduction to RATS

Getting StartedA Tutorial 1.3.1 Learn More: Series Edit Windows

Change Layout
When using a Series Edit Window to edit a series, you can change the way the data appear on the screen using the menu operation ViewChange Layout or the toolbar icon . This brings up the dialog:

The main value of changing the Numbers Per Row is to match the view of the data to your original source if you need to double-check numbers that you have had to type in. The most common change in the layout is to adjust the decimal digits. By default, the data editor uses the shortest representation that can display all the data. That works well with original source data with a limited number of digits. However, with data that have been through some transformations (such as logs or averages), you might get ten (or even more) decimals. Change Layout lets you adjust that. Click on OK to close the Change Layout dialog box and apply your changes.

Copy and Paste

You can copy data out of a Series Edit Window by selecting a (consecutive) range of values and using the keystrokes <Ctrl>+C, the menu operation EditCopy, the toolbar icon or the Copy contextual menu item. When you paste into a target application, the values will have the same row and column arrangement as you have in the Series Edit Window in rats, that is, if you have 40 rows of 4 numbers, you will also get a 404 table in the target. If you paste as text, the data will be pasted as it appears in the window as well, with the number of digits shown. You can use Change Layout to adjust either the arrangement of the data or the number of digits. You can paste into a Series Edit Window by positioning the selection to the point where you want to insert values, then doing <Ctrl>+V, the menu operation Edit Paste, or the toolbar icon . The paste operation will take the values in order, even if the layout is different from what you have in the Series Edit Window. For instance, if you select and copy a column of data from a web site and paste it into the time series editor, the cells will be filled, in order, starting from the active cell, going first across and then down to the start of the next row, across that, etc. Note that you need the source data to be values from a single series in ascending time sequence.

Introduction to RATS

Int13

Getting StartedA Tutorial

Toolbar Icons
In addition to standard toolbar items like Select All, Copy, Paste, Print, the following specialized toolbar icons are available when using a Series Edit Window (Insert) Inserts a new cell at the current cursor position (cell value is set to NA). Removes the current cell and shifts the remaining data one position to the left to fill its place. Sets the current cell to the missing value code (NA, or Not Available). moves the cursor to the cell containing the largest (maximum) value in the series. moves the cursor to the cell containing the smallest (minimum) value in the series. searches for an (exact) value. displays a Report Window with basic statistics (mean, variance, skewness, etc.) for the series. displays a time series graph for the series. displays a histogram plot for the series.

(Remove) (NA) (Max. value) (Min. value) (Find) (Statistics) (Graph) (Histogram)

(Graph Xform) displays a time series graph with several power transformations of the series.

Int14

Introduction to RATS

Getting StartedA Tutorial 1.3.2 Learn More: Report Windows

Change Layout
A Report Window is like a read-only spreadsheet. You can adjust the appearance, but cant change the values themselves. Some (in this case most) of the values shown in a report are available at a higher precision than you see in the table. On the other hand, they will often show far more decimal places that you would likely use in any document that you would be preparing. To change the appearance of values, use the Change Layout menu operation (or the toolbar or the Reformat operation on the contextual menu). This will bring up the following dialog box:

If you set the Decimal Width box, the values will be formatted to that number of digits right of the decimal, with however many are needed to the left of it. If you set the Number Width box, a common format will be chosen that displays all the values within that number of characters. (If you set both, the decimal width choice determines the representation).

Copying To Other Applications

If you select information out of this and copy it (which you can do many ways: the keystroke <Ctrl>+C, EditCopy menu, Copy toolbar icon , and Copy contextual menu), and then paste into a word processor, you will get the information that is shown on the screen, with the number of digits shown. If your target application is a spreadsheet, you may need to choose Paste Special in order to control what comes through. (rats copies report information in many different formats). The dif, csv and xml spreadsheet formats will paste numbers at full (up to fifteen digit) precision. On the other hand, if you paste as (Unformatted) Text, you will get only the number of digits shown on the screen in rats, so if you adjust this using Change Layout, you will get what you see. Note that rats only copies content, not column widths and cell formats, so you can reformat the cells in your spreadsheet to show them the way that you want.

Introduction to RATS

Int15

Getting StartedA Tutorial

Copying to TeX
If you want to paste into a tex document (as a table), you need to use the EditCopy as TeX (or toolbar or Copy->TeX contextual menu operation). When you do that, rats puts into the clipboard a tex tabular environment for displaying the table, with the numbers as shown on the screen, so do the Change Layout first to get the representation you want. See the description of FORMAT=TEX in the Additional Topics pdf for more information on using tex with rats.

Exporting to a File

Select the cells that you want to save and use FileExport (or FileSave As, or the toolbar icon or the Export contextual menu). You have the choice of quite a few formats, most of which are also among the formats that rats copies when you do a Copy operation. The format here that is most likely to be useful is tex, since you can use an \include directive in your TeX document. By using that, rather than copy and paste, you can quickly replace the table if you need to re-generated the report.

Int16

Introduction to RATS

Getting StartedA Tutorial 1.4 Example Three: Transformations and Regressions

In this section, we will work with a sample data set from Pindyck and Rubinfelds Econometric Models & Economic Forecasts (1998) to demonstrate a variety of tasks, including data transformations and multiple regression models. All of the commands that we will describe here are also provided for you in a rats program file called ExampleThree.rpf. We recommend that you begin by entering the commands yourself as we discuss them. However, if you encounter difficulties, you can refer to (or execute the instructions on) ExampleThree.rpf to see exactly how things should look. The sample data set is provided on an xls (Excel) spreadsheet called ExampleThree.xls. The following data series are provided on the file: Rate IP M1 M2 PPI Three-month treasury bill rate. Federal Reserve index of industrial production, seasonally adjusted, index 1987=100. Money Stock M1, billions of US dollars, seasonally adjusted. Money Stock M2, billions of US dollars, seasonally adjusted. Producer Price Index, all commodities, index 1982=100, not seasonally adjusted.

Our primary goal will be to fit the following regression models to these data: (2)
Ratet = + 1 IPt + 2 ( M1t M1t3 ) + 3 PSUM t + ut
PPI t PPI t1 PPI t2 + + PPI t PPI t1 PPI t2

where PSUM t = (3)

Ratet = + 1 IPt + 2GRM 2t + 3GRPPIt1 + ut

where GRM2t =

( M2t M2t1 )
M2t1

, GRPPIt = 100

(PPIt PPIt1 )
PPI t1

Equation (2) is actually from Example 4.2 in the 3rd (1991) edition of Pindyck and Rubinfeld, while equation (3) is from Example 4.2 as it appears in the 4th (1998) edition. Our data were taken from Haver Analytics usecon database (available separately through Estima). The Haver series names for these are FTB3, IP, FM1, FM2, and PA, respectively. Weve converted them to the names shown above for our example. The series begin in January, 1959, and run through September, 1999. Some of the values are slightly different than the older data used in Pindyck and Rubinfeld.

Introduction to RATS

Int17

Getting StartedA Tutorial

Heres a portion of the data on the file:
Date IP M1 M2 PPI RATE 1959:01 36.000 138.900 286.700 31.700 2.84 1959:02 36.700 139.400 287.700 31.700 2.71 1959:03 37.200 139.700 289.200 31.700 2.85 1959:04 38.000 139.700 290.100 31.800 2.96 1959:05 38.600 140.700 292.200 31.800 2.85 1959:06 38.600 141.200 294.100 31.700 3.25 1959:07 37.700 141.700 295.200 31.700 3.24 1959:08 36.400 141.900 296.400 31.600 3.36 1959:09 36.400 141.000 296.700 31.700 4.00 1959:10 36.100 140.500 296.500 31.600 4.12 1959:11 36.300 140.400 297.100 31.500 4.21 1959:12 38.600 140.000 297.800 31.500 4.57

1.4.1 Getting the Data In

As in our previous example, we want to start by defining a date scheme and then reading data into memory. In this case, though, well be reading the data from a file, which is much more common. We could just type in the necessary instructions, but another option is to use the Data Wizard to do this. The Data Wizard approach is usually preferable for most file formats. However, for certain file formats, or in cases where you need to pull in data from multiple sources, you may need to type the instructions directly.

A Clean SlateThe Clear Memory Operation

If you still have rats running after completing the previous example, you will probably want to clear the memory of the settings and data entered earlier. You can do that by opening the File menu and selecting Clear Memory, or clicking on the Clear Memory toolbar icon: . You can also close the Input Window, then open a new one using the menu operation FileNew-Editor/Text Window. Any of these will clear any data series and other variables, as well as any settings defined by instructions like CALENDAR, from memory. This allows you to start fresh, as if you had never executed any instructions. Note that Clear Memory does not delete any text or close any windows.

The Data Wizard

rats actually offers two Data Wizardsone for use with rats format data files, and one for all other file types. Well discuss the wizard for rats format data files in

Section 1.7. For now, open the Data/Graphics menu and select the operation Data (Other Formats).

We will be working with an Excel xls format file called ExampleThree.xls, so select Excel 3.0-2003 Files (*.XLS) from the drop-down list of file types. The dialog box should now show files with that extension in the current directory.

Int18

Introduction to RATS

Getting StartedA Tutorial

The default start-up directory for rats should be the directory containing the main example programs, data files, and procedures that ship with rats, including ExampleThree.xls. So, you should see as one of the files listed in the dialog box. If not, you can use the dialog box to navigate to the appropriate directory. Select ExampleThree.xls and click the Open button. rats will display the Data Wizard dialog box. Click on the Show Preview button to view the contents of the file. Then, click on the Scan button, which scans the date information on the file to determine the appropriate CALENDAR setting. You should see something like this:

Data Organization The Data Organization section tells rats how the data are arranged on the file. Here, the data series run down the page in columns, so we want to use the Down Page setting (which is, by far, the most common and should be selected by default).

Introduction to RATS

Int19

Getting StartedA Tutorial

You only need to use the Header Rows field if you need to skip lines (other than a row of series names) at the top of file, such as lines of text describing the contents of the file. Our data file begins with the series labels (which we need) on row 1, so we leave Header Rows set to zero. Similarly, you can use Columns to Skip if you want rats to ignore columns at the left of the file. The Bottom Row and Right Column fields default to the last row and column numbers found in the file. You can reduce these values to skip rows at the bottom or columns at the right. For this example, leave these set to the default values. If you are reading a spreadsheet file containing data on multiple sheets, the Sheet field will let you pick which sheet you want to read. If you change sheets, the preview table will reload and the Bottom Row and Right Column fields will reset. Date/Calendar Handling The Format of Date Strings field shows what rats thinks is the format of the date information on the file (if any). If it thinks there is more than one possible interpretation, you can use this field to select from the choices it offers. Here, rats has correctly guessed that the dates are in the year:month form (abbreviated y:m). The File Dates field describes the frequency and starting date of the data as it appears on the file. When you click on Scan, rats examines the file to determine the frequency and starting date of the data set. In this case, it should report that the data is monthly, starting in January of 1959. The Target Dates field shows the starting date and frequency that will be used in your rats session. By default, this will match the File Dates setting. You can set or change it by clicking on the Set button. This is useful if you only want to read in a subset of the data, or if you want to work with the data at a different frequency. If you set a different target frequency, rats will automatically compact or expand the data to that frequency. The Compact by box allows you to select the compaction method that will be used when going to a lower frequency. That will only be active if the File Dates and Target Dates have a different frequency.

Read the Data

For now, accept the default settings and click on OK. rats will generate and execute the appropriate CALENDAR, OPEN DATA, and DATA commands. They should look something like this: OPEN DATA "C:\Users\Estima\Documents\WinRATS 8.0\ExampleThree.xls" CALENDAR(M) 1959:1 DATA(FORMAT=XLS,ORG=COLUMNS) 1959:01 1999:09 RATE IP M2 M1 PPI Lets take a closer look at each of these.

Int20

Introduction to RATS

Getting StartedA Tutorial

The OPEN Instruction
OPEN DATA "C:\Users\Estima\Documents\WinRATS 8.0\ExampleThree.xls" The OPEN DATA instruction tells rats the name and location of the data file you want to read. This instruction includes the full path to the file (which may be slightly different on your system). If you are typing in the instruction directly, you can omit the path if the file is located in the default directory (see page Int34).

The CALENDAR Instruction

CALENDAR(M) 1959:1 The CALENDAR instruction is similar to the one in our previous example, but here we have the option M for monthly data, rather than Q for quarterly, and the data start in January of 1959. Some of the Pindyck and Rubinfeld examples apply only to data from November, 1959. We will use date parameters to skip earlier observations when necessary in this example, but another alternative would be to set our CALENDAR to start in November. We could do that in the Data Wizard by clicking on the Set button under Target Dates and replacing the 1 in the Month/Period cell with 11 (for the 11th month), the resulting CALENDAR would be: calendar(m) 1959:11 With this setting, rats would skip the data for January through October, and start reading in data beginning with the November 1959 observation. If you have cross-sectional data, with no time series periodicity, omit the CALENDAR instruction.

Working with Dates

year:period

To refer to a date in rats, use the following formats: for annual, monthly, and quarterly data (or any other frequency specified in terms of periods per year). In this example, we have set a monthly CALENDAR, so 1996:2 translates to the 2nd month (February) of 1996. for weekly, daily, etc.

year:month:day

With annual data, period is always one, so any reference to a date in annual data must end with :1. The :1 after the year is very important, because without it, rats will assume you are specifying an entry number, not a date.

Introduction to RATS

Int21

Getting StartedA Tutorial

The DATA Instruction
DATA(FORMAT=XLS,ORG=COLUMNS) 1959:01 1999:09 RATE IP M2 M1 PPI This reads the data series RATE, M1, M2, IP, and PPI from the file. Here, we have two options, separated by commas: FORMAT and ORGANIZATION. You can abbreviate option names to three or more characters, which is what we did with the ORG option. Recall that the list of options must appear immediately after the instruction name, with no space between the instruction and the left parenthesis. For now, well go over these two options quickly. Because the data step is so important, we included a full chapter (Chapter 2) in this book to it; if you need more detail, you can check that out. The FORMAT Option FORMAT gives the format of the file you are reading. As noted above, ExampleThree. xls is an Excel spreadsheet file, and the option for that is FORMAT=XLS. rats supports about 20 formats, which are listed in Section 2.2. Note: rats does not try to determine the format of the file based on the file name or extensionthe FORMAT option must be set to match the format of the file being read. The ORGANIZATION Option The ORG option describes how the data are arranged on the file. The Down Page setting on the wizard corresponds to the setting ORG=COLUMNS while the Across Page setting corresponds to ORG=ROWS. You can also abbreviate the choices for options like this to three or more characters. For example: ORG=COL. The Entry Range The first two parameters (1959:01 and 1999:09) specify the starting and ending dates to be read in. You will generally omit explicit date ranges and just let rats figure out the appropriate range for a given operation. Here, though, the wizard includes the dates so that you know (and have a record of) the exact range being read. The Series Names For files that include series names (which rats requires for most formats), listing the series names on the DATA instruction is optionalif you omit the list, rats reads in all the series on the file. If you do provide a list, rats reads only those series. For text files without any series names (which can be handled with FORMAT=FREE), you must supply a list of series names if typing in the DATA instruction yourself (otherwise rats would have no way to identify the data). If you use the Data Wizard, rats will prompt you for names. The series list is also required when reading some database formats, to avoid accidentally reading a huge number of series.

Int22

Introduction to RATS

Getting StartedA Tutorial 1.4.2 Check the Data

Display the Series Window
First, we need to verify that the data have been read in properly. To begin, select the Series Window operation from the View menu. Youll see a window displaying a list of all the series:

This provides a quick check on the number of observations in each series, as well as the frequency and data range. Now, select (highlight) all of the series in the window and then select Statistics from the View menu, or click on the Basics Statistics toolbar icon: You should see the following table of summary statistics for each series. The most important items to check are the number of observations (Obs) and the minimum and maximum values. Be sure they are reasonable given what you know about the data.

You can explore the other View menu (and toolbar button) operations introduced on page Int10 with combinations of one or more seriesjust highlight the series you want to include before selecting an operation. Another way to generate the table of statistics is to use the instruction TABLE. Type in the following and hit <Enter>: The results should appear in the output window, and match those shown in the Statistics window above.

table

Introduction to RATS

Int23

Getting StartedA Tutorial

STATISTICS and the Univariate Statistics Wizard
statistics rate Heres resulting output:
Statistics on Series RATE Monthly Data From 1959:01 To 1996:02 Observations 446 Sample Mean 6.058587 Variance 7.701206 Standard Error 2.775105 of Sample Mean 0.131405 t-Statistic (Mean=0) 46.106212 Signif Level 0.000000 Skewness 1.186328 Signif Level (Sk=0) 0.000000 Kurtosis (excess) 1.587381 Signif Level (Ku=0) 0.000000 Jarque-Bera 151.440700 Signif Level (JB=0) 0.000000

For a more detailed set of sample statistics, you can use STATISTICS instruction with a single series:

The corresponding wizard is the Univariate Statistics operation on the Statistics menu. If you select that operation, rats will display the following dialog box:

The first step is to select the series you want to use from the Series drop-down listhere, weve selected the series RATE. Next, click on one or more of the Basic Statistics, Extreme Values, or Autocorrelations check boxes. Here, weve selected Basic Statistics, which generates a STATISTICS command as shown above.

Extreme Values generates an EXTREMUM instruction, which reports the maximum and minimum values of the series. Autocorrelations generates a CORRELATE instruction, which computes autocorrelations (and partial autocorrelations if you provide a series for the Partial Corrs field). You can check any combination of these three boxes. The other fields allow you to select the range used for the computations and to select various options for the autocorrelation computations.

Int24

Introduction to RATS

Getting StartedA Tutorial 1.4.3 Data Transformations and Creating New Series
The SET Instruction
In most of your work with rats, you will need to do at least a few data transformations, and you will often need to create new series from scratch. You can do that using the SET instruction, or one of several wizards. For our example, we need to define quite a few new series. Well start by generating a couple of differenced series using SET. Execute the following instructions:

set ppidiff = ppi - ppi{1} set m1diff = m1 - m1{3}

Be sure to put at least one blank space before the = sign. Lets examine the first transformation. The {1} notation (which we refer to as lag notation) tells rats to use the first lag of PPI in the transformation. This creates a new series called PPIDIFF, and sets it equal to the first difference of PPI: PPIDIFFt = (PPIt PPIt1) for each entry t in the default entry range. Note that PPIDIFF cannot be defined for 1959:1, because we do not have data for 1958:12, which would be the one period lag from 1959:1. rats recognizes this, and defines entry 1959:1 of PPIDIFF to be a missing value. Similarly, M1DIFF is defined as the three-lag difference of the series M1, so that M1DIFFt = M1t M1t3. M1DIFF will be defined starting in 1959:4. If you want to see the values of these series, click on or re-open the Series Window, highlight the series you want to view, and click on ViewData Table (or click on the icon). Next, well create some quarter to quarter growth rates, again using the {L} lag notation, and the / division operator: set grm2 = (m2 - m2{1})/m2{1} set grppi = (ppi - ppi{1})/ppi{1} We also need to create a three-period moving average term. First, we define PRATIO as the ratio of PPIDIFF (the first difference of PPI that we created above) to PPI. Then, we define PPISUM as the sum of the current and two lags of PRATIO. This could be done with a single SET, but is easier to read or modify this way:

set pratio = ppidiff/ppi set ppisum = pratio + pratio{1} + pratio{2}

Note that there are specialized instructions that can be used for some of these operations, such as DIFFERENCE and FILTER, but SET is the most important because it is the most flexible.

Introduction to RATS

Int25

Getting StartedA Tutorial

Data Transformation and Related Wizards
rats offers several wizards for doing transformations, creating dummy variables,

and other data-related operations. The Transformations operation on the Data/ Graphics menu is probably the most versatile.

For example, another way to create PPIDIFF as the first difference of PPI is to select Transformations, type in the name PPIDIFF in the Create field, select Difference from the By/As field, and select PPI in the From drop-down list.

You can use the Create field to type in or select the name of the series you want to create or redefine. The By/As field controls the type of transformation. Select GeneralInput Formula to enter your own formula for the transformation, or use one of the pre-defined transformations, including difference, log and square root. The Trend/Seasonals/Dummies, Differencing, and Filter/Smooth operations offer similar series creation and transformation capabilities.

Keep Data Transformations in Your Program!!

While you could save these transformed series to a data file and then rely on those saved versions in subsequent analysis, we strongly recommend that you continue to read in the original source data, and retain the instructions used to generate your transformations as part of your rats programs. This will help ensure that you can reproduce your results later on, and will avoid any confusion about exactly how the transformed series were derived. (Consider how many ways there are to compute a growth rate). Also, even someone who has never used rats should be able to tell exactly how the transformations were done simply by reading through the instructions.

Int26

Introduction to RATS

Getting StartedA Tutorial 1.4.4 Estimating Regressions

Our Equations
(4) (5) Recall that we want to estimate two regression equations: Ratet = + 1 IPt + 2 ( M1t M1t3 ) + 3 PSUM t + ut Ratet = + 1 IPt + 2GRM2t + 3GRPPIt1 + ut

Weve done the necessary transformations, and are ready to estimate the models.

Estimating a Linear Regression

Well start by using the Regression wizard to estimate equation (4). Select Regressions from the Statistics menu. In the dialog box, use the Dependent Variable dropdown list button to select RATE. Next, we need to select the explanatory variables. You can type the regressors directly into the Explanatory Variables field (using blank spaces to separate variables), or you can click on the button, which opens the dialog box shown below. You can add variables from the Available Series list to the regression by: doubleclicking on a series name in the Available list; selecting one or more series and clicking the button; or using the button to add lagged variables to the regression list For this regression, add CONSTANT, IP, M1DIFF, and PPISUM to this list of regressors (CONSTANT is a special built-in series name, used to include a series of ones in a regression):

Introduction to RATS

Int27

Getting StartedA Tutorial

Click OK to close the list. The main dialog should now look like this:

Click OK to run the regression. The Wizard will generate the instruction below. This will actually be in upper case to help you recognize which instructions are generated using wizardswe edit them to lower case using the menu operation EditTo Lower Case to keep with the style of the manual. linreg rate # constant ip m1diff ppisum

LINREG is the standard instruction for estimating linear regressions. Here, it regresses the dependent variable RATE on the independent variables IP, M1DIFF, and PPISUM, and includes a constant term (intercept) in the regression. The line beginning with the # symbol is called a supplementary card. Supplementary cards are used with many instructions to supply additional information to an instructionusually lists of series or equations. Supplementary cards always begin with the # character. The results will be somewhat different from those shown in the text book, because the data are not identical due to revisions made since Pindyck and Rubinfeld extracted their data. LINREG estimates using ordinary least squares (ols). The output produced by this command is shown on the following page:

Int28

Introduction to RATS

Getting StartedA Tutorial

Linear Regression - Estimation by Least Squares Dependent Variable RATE Monthly Data From 1959:04 To 1996:02 Usable Observations 443 Degrees of Freedom 439 Centered R^2 0.2526958 R-Bar^2 0.2475890 Uncentered R^2 0.8717009 Mean of Dependent Variable 6.0806546275 Std Error of Dependent Variable 2.7714419161 Standard Error of Estimate 2.4039938631 Sum of Squared Residuals 2537.0628709 Regression F(3,439) 49.4816 Significance Level of F 0.0000000 Log Likelihood -1015.1499 Durbin-Watson Statistic 0.0816 Variable Coeff Std Error T-Stat Signif ********************************************************************************** 1. Constant 2.11841571 0.42030566 5.04018 0.00000068 2. IP 0.06417324 0.00768853 8.34662 0.00000000 3. M1DIFF -0.04183328 0.01485687 -2.81575 0.00508547 4. PPISUM 58.26459284 8.01033322 7.27368 0.00000000

Loading into a Report Window

For many uses, particularly when youre examining several possible models, the text output shown above is most convenient, since the output from all the models will be together in a single file. However, when youve finally settled on a specification, this isnt what you want, since the format for the numbers is fixed the way you see them in the editor. You could round them yourself and type them into a document, but were trying to show you how to avoid that. Instead, you can reload the same information into a Report Window. To do that, you can open the Report Windows submenu on the Window menu and select the Linear RegressionLeast Squares report. The Report Windows list shows (up to) the last fifteen reports generated, with the most recent ones at the top of the list. Once youve reloaded the regression into a Report Window, you can use the operations described on page Int11 for reformatting numbers and copying the information for use in word processors. For example, you can select sections of the report and use ViewChange Layout to change the number of decimal places for the selected numbers, and then use EditCopy or FileExport to get the results into another program or file. In addition to reports generated by rats instructions like LINREG, you can also create your own reports with the precise information that you want. Thats a more advanced, but very useful, feature covered in Section 1.6.7 in the Users Guide.

Introduction to RATS

Int29

Getting StartedA Tutorial

Estimating the Second Equation
Now lets estimate equation (5). If you select the Regressions operation again, youll see the same dialog box as before, loaded with the settings from the last regression. To modify the equation, you can either edit the regressor list directly in the Explanatory Variables box, or you can click on the button to use the dialog box. In the dialog box, use to delete M1DIFF and PPISUM from the list. Add GRM2, and use to add lag 1 of GRPPI. The regressor list should look like this: constant ip grm2 grppi{1} Click OK in the main dialog box to execute the regression. Heres the command that is generated and the output. linreg rate # constant ip grm2 grppi{1}
Linear Regression - Estimation by Least Squares Dependent Variable RATE Monthly Data From 1959:03 To 1996:02 Usable Observations 444 Degrees of Freedom 440 Centered R^2 0.2215769 R-Bar^2 0.2162694 Uncentered R^2 0.8660034 Mean of Dependent Variable 6.0733783784 Std Error of Dependent Variable 2.7725545963 Standard Error of Estimate 2.4545026092 Sum of Squared Residuals 2650.8165458 Regression F(3,440) 41.7484 Significance Level of F 0.0000000 Log Likelihood -1026.6780 Durbin-Watson Statistic 0.1730 Variable Coeff Std Error T-Stat Signif ********************************************************************************** 1. Constant 1.20988109 0.52637190 2.29853 0.02199968 2. IP 0.06525852 0.00727000 8.97641 0.00000000 3. GRM2 136.19299812 34.77735583 3.91614 0.00010426 4. GRPPI{1} 101.94478474 17.17615119 5.93525 0.00000001

Range Parameters

Our first model used M1DIFF, which isnt defined for the first three periods, since it uses lag 3 of M1. This second model replaces that with GRM2, which only drops one point; the range for this regression is determined by the one-period lag of GRPPI which isnt defined until 1959:3. When you execute the LINREG, rats will scan the data, determine that 1959:4 is the earliest possible starting point for the first regression and 1959:3 is the earliest possible date for the second. Its ability to handle such entry range issues automatically is a very powerful feature. With time series models with lags, you need to be somewhat careful about doing comparisons of estimates based upon different ranges, as we have with these two.

Int30

Introduction to RATS

Getting StartedA Tutorial

The statistics which are sums (rather than averages) are especially affected by this: here the Sum of Squared Residuals and the Log Likelihood, which shouldnt be compared when computed over different ranges. Even the other statistics, which are based on averages, are only somewhat comparable. If we were seriously interested in choosing between these (we arent, since both have very low Durbin-Watson statistics, so neither is a serious model for the interest rate), we should re-estimate the second regression over the same range as the first. While you can redo the Regression Wizard, and reset the Sample Start box, its simpler to just edit the instruction and re-estimate it. We can change it to linreg rate 1959:4 * # constant ip grm2 grppi{1} Almost any instruction which operates across a set of data allows you to give an explicit range, or to let rats figure it out. Here, we need to override the standard handling, which for a regression uses the maximum range possible with the series involved. The estimation range has start and end parameters; were fine with the end being the automatic value, which is what the * means. What we need to override is the start, which we do by giving the start date of 1959:4. If we wanted to restrict the top end of the regression range (say to the end of 1992), we would use linreg rate 1959:4 1992:4 # constant ip grm2 grppi{1} If we wanted to restrict the end of the range, but use as much data as possible at the start, we would use: linreg rate * 1992:4 # constant ip grm2 grppi{1} If you want to use the default for both range parameters, you can just leave them out if nothing comes after them. If there are trailing parameters, you either have to use * *, or you can also use the shorthand / to cover both. For instance, LINREG has a fourth parameter (for the generated residuals), so if we wanted to estimate this over the full range and save the residuals into the series U, we could use linreg rate / u # constant ip grm2 grppi{1}

While not common with time series data, its also possible to skip entries out of the middle of the data set. Well look at that in Section 1.6, which looks at cross section data.

Introduction to RATS

Int31

Getting StartedA Tutorial 1.4.5 Learn More: Procedures

Procedures are collections of rats commands that can be executed with a single call to the procedure, using a syntax nearly identical to that used for built-in instructions. While you may eventually want to write your own proceduressee Section 15.2the main advantage for new users is that you have access to hundreds of existing procedures that greatly increase the scope of what you can do quickly and easily. We provide more than two hundred pre-written procedures with ratssee the procedures and Examples pdf file included with rats for the current list. As a simple example, type in and execute the following procedure call: @regactfit This executes a procedure called RegActFit which generates a plot showing the residuals and actual and fitted values from the most recent regression:

If you want to see what the procedure code looks like, youll find it on the file called REGACTFIT.SRC. It is a fairly simple procedure that provides a handy alternative to writing the necessary SET and GRAPH instructions yourself. You can find many more, including procedures written by other rats users from around the world, on our website. The website also includes a handy Procedure Browser for locating procedures of interest. You can find the browser in the Resources section of the website. Each procedure is usually stored on its own text file. Some of the procedures included with rats are described in the manual. For the others (including those available on our web site), see the comment lines at the top of the procedure file.

Int32

Introduction to RATS

Getting StartedA Tutorial

Executing a Procedure
The basic syntax for executing a procedure is: @procedure name( options ) parameters # < supplementary cards > (f needed) Everything is the same as a standard rats instruction except two things: The procedure name is preceded by @. You cant abbreviate the procedure name. The following uses the procedure @DFUNIT (DickeyFuller Unit root test) to do unit root tests on the RATE series from the data set in this section: @dfunit rate

Loading the Procedure

To use a procedure or function stored on a separate file, you need to have rats execute the instructions that define the procedure. In most cases, rats can do this automatically, by searching for a file with a .SRC extension whose name matches the name of the procedure. Otherwise, you can compile procedures stored on a file using the SOURCE instruction: source name of file with PROCEDURE or FUNCTION If rats cant load a procedure:

Check to make sure you did a full installation, including all of the files supplied with rats. See page Int156 (installing rats) for details. Do FilePreferences and check the Procedure Directory setting on the Directories tab. In order for rats to find procedure files automatically, this should point to the directory containing your procedures (Additional Topics, Section 1.3). If you know where the procedure file is installed on your computer, include a SOURCE instruction (with a complete path and filename) prior to the procedure call to source in the procedure. You may be running an out-of-date version of rats that did not ship with the procedure in questionyou can do HelpAbout RATS to check which version you are using. See www.estima.com, email [email protected], or call 800-822-8038 for information on updating. If you are using a network license at a company or university, your institution may already have the most recent version available check with your system support staff. If all else fails, you should be able to download the procedure using the Procedure Browser on the Resources section of our website.

Introduction to RATS

Int33

Getting StartedA Tutorial 1.4.6 Learn More: Reading Data

The Default Directory and the OPEN Instruction
When you use the Data Wizard to read a file, rats includes the full path to the file on the OPEN instruction. However, rats also maintains a default directory setting, which is where it will look for (or create) files if you dont specify the full path. You can use the menu operation FileDirectory if you want to change the default directory for the current session. If you want to make that change permanent, you can reset that as part of your Preferences (see Section 1.3 of Additional Topics). So, if you have the directory containing the rats example files set as the default directory (which it should be when you first install the program), you could type the OPEN instruction shown earlier like this: open data ExampleThree.xls This approach (omitting the path) often works better if you will be sharing programs and data with another user who has a different directory structure on their system. Your colleagues just need to make sure that their programs and data files are stored in the same directory, and that they set that as the default directory in rats.

Blank Spaces? Enclose in Quotes

If your path and filename includes any blank spaces, you need to enclose the entire string in single or double quotes, as in the code generated by the wizard. For example: open data "C:\My RATS Files\ExampleThree.xls" or open data 'C:\My RATS Files\ExampleThree.xls' For a path or filename that does not include any blank spaces, you dont need quote marks around the name, although there is no harm in including them.

Data from Multiple Files

You do not have to load all your data with a single instruction. You can use several OPEN DATA and DATA instructions to read from multiple data sources. While you could use the Data Wizard operation for each file, you would need to make sure to use the same Target Dates setting each time. Otherwise, the date mappings specified by the last CALENDAR setting will no longer match up properly with data read in previously. So, if you need to read from multiple files or sources, you may find it easier to just type in the instructions directly (or use the Wizard for the first file, and then type the instructions for the additional files directly).

Int34

Introduction to RATS

Getting StartedA Tutorial 1.4.7 Learn More: Annotated Regression Output

This is the output from the first linear regression with a description of what each element shows. Note that this set of output is specific to least squares regression; other forms of estimation will be similar, but may leave out some statistics that arent defined or have no usable interpretation.
(a) Linear Regression - Estimation by Least Squares (b) Dependent Variable RATE (c) Monthly Data From 1959:04 To 1996:02 (d) Usable Observations 443 (e) Degrees of Freedom 439 (f) Centered R^2 0.2526958 (g) R-Bar^2 0.2475890 (h) Uncentered R^2 0.8717009 (i) Mean of Dependent Variable 6.0806546275 (j) Std Error of Dependent Variable 2.7714419161 (k) Standard Error of Estimate 2.4039938631 (l) Sum of Squared Residuals 2537.0628709 (m) Regression F(3,439) 49.4816 (n) Significance Level of F 0.0000000 (o) Log Likelihood -1015.1499 (p) Durbin-Watson Statistic 0.0816 (q)Variable (r) Coeff (s) Std Error (t) T-Stat (u) Signif ********************************************************************************* 1. Constant 2.11841571 0.42030566 5.04018 0.00000068 2. IP 0.06417324 0.00768853 8.34662 0.00000000 3. M1DIFF -0.04183328 0.01485687 -2.81575 0.00508547 4. PPISUM 58.26459284 8.01033322 7.27368 0.00000000

Well use the following notation: y y y e T K The vector of values for the dependent variable The sample mean of the dependent variable over the estimation range The deviations from the mean of the dependent variable The vector of residuals The number of observations The number of regressors

(a) The type of model and estimation technique used. (b) The dependent variable (c) If you are using a CALENDAR, rats will list the frequency of the data and the beginning and ending of the estimation range. If you have date without a date scheme, this will be skipped. (d) The number of usable entries in the estimation range: T (e) The degrees of freedom: T-K (f) The centered R2 statistic: 1 ee yy

Introduction to RATS

Int35

Getting StartedA Tutorial

ee (g) The adjusted R2 statistic ( R 2 ):1 T K ( ) (h) The uncentered R2 statistic: 1 ee y y yy (T 1)

(i) The mean of the dependent variable: y (j) The standard error of the dependent variable (k) The Standard Error of Estimate: ee (T K ) yy (T 1)

(l) The Sum of Squared Residuals: e ' e (m) The regression F-statistic: y y ee ( K 1) ee T K) (

(n) The marginal significance level of the F, with K-1 and T-K degrees of freedom. (o) The log-likelihood: T e e log + 1 + log (2p ) 2 T

(p) The Durbin-Watson statistic:

(e
t =2

et1 )

e
t =1

2 t

(q) The names of the explanatory variables. Lags are shown as name{lag}. (r) The estimated coefficients. (s) The standard error of the coefficient estimate. (t) The t-statistic of the coefficient (coefficient/its standard error). (u) The marginal significance level for a (two-tailed) test for a zero coefficient.

Goodness of Fit Measures

You will notice there are three versions of the R2 statistic: the centered R2, the R 2 (R2 adjusted for degrees of freedom) and the uncentered R2. The centered and adjusted R2 are typically the only ones of interest.
rats also displays the mean and standard error of the dependent variable. These are

simply statistics on the dependent variable, and tell you nothing about the accuracy of the regression model. They are the same values you would get by doing a STATISTICS instruction on the RATE series over the same range.

Int36

Introduction to RATS

Getting StartedA Tutorial

Regression F and Significance Level
These are only included if the regression includes a CONSTANT (or its equivalent in some set of dummy variables). This is the result of an F-test for the hypothesis that all coefficients (excluding CONSTANT) are zero. The numbers in parentheses after the F are the degrees of freedom for the numerator and denominator, respectively.

Log Likelihood

For a linear regression, this is the log likelihood of the data assuming Normal residuals. Note that this includes the constants: the 1 + log( 2p ) terms only interact with T and not with the residuals and so could be dropped from any comparison of two linear regressions with the same number of observations. While some programs omit these, rats always includes them in all calculations of likelihoods or any density functions.

Durbin-Watson Statistic

The Durbin-Watson statistic tests for first-order serial correlation in the residuals. The ideal result is 2.0, indicating the absence of first-order serial correlation. Values lower than 2.0 (and particularly below 1.0) suggest that the residuals may be serially correlated.
rats always computes a Durbin-Watson statistic, even if you have cross-section data

where serial correlation isnt likely to be an issue. However, you should keep in mind that the tabled values for the Durbin-Watson statistic are known to be invalid in a variety of circumstances, such as the presence of lagged dependent variables.

Coefficient Table

rats attempts to come up with a common format for displaying the regression coeffi-

cients and standard errors. This makes the results easier to read than if the decimal points didnt align. The table lists all the variables in the regression, including the constant if you included it. The Coeff column lists the estimated coefficients for each variable. The t-statistic is the ratio of a coefficient to its standard error. The significance level (see page Int11) is for a two-tailed test for a zero coefficient. For models estimated by least squares, the t-statistic in the T-Stat column is treated as having a t distribution with T-K degrees of freedom. If the model is estimated by some other technique, the t-statistic is treated as having a Normal distribution.

Introduction to RATS

Int37

Getting StartedA Tutorial 1.4.8 Learn More: The Series Window

The Series Window is opened by selecting the ViewSeries Window menu operation. This is associated with your current rats session. If its open, it will get updated as you add other series. If you close the Input Window, or do a Clear Memory operation, the Series Window will be emptied, since it shows only the series is working memory. While you can use menu operations to add or edit series using the Series Window, those series will also be cleared when you finish a rats session, so you should only make changes like that if you are planning to export the data to a permanent file. You can double-click on a series to open up a Series Edit Window (Section 1.3.1). If you select one series, it will be the initially chosen series in the drop-down box on such things as the Linear Regressions Wizard (for the dependent variable). It has the following toolbar operations and associated menu operations:

(Select All) (Change Layout)

Shortcut for EditSelect All. Selects all the series in the list. Shortcut for ViewChange Layout. This is more useful in the possibly very large lists of series for data browsers, as it allows you to make a much shorter list by filtering out series which dont meet certain criteria. Shortcut for EditFind. Searches for a value across series. Shortcut for FileImport. Imports data from an external file. While this will work, its usually better to use the Data Wizard, which does much the same thing, but keeps a record of the instructions. Shortcut for FileExport. Exports the selected series in one of many formats. Shortcut for Data/GraphicsCreate Series (Data Editor). Opens a Series Edit Window for a new series. Shortcut for ViewTime Series Graph. Displays a time series graph for the series. Shortcut for ViewHistogram. Displays a histogram plot for the series. Shortcut for ViewBox Plot. Displays a box plot for the selected series.

(Find) (Import)

(Export) (New Series) (Series Graph) (Histogram) (Box Plot)

(Autocorrelations) Shortcut for ViewAutocorrelations. Computes and graphs the autocorrelations and partial autocorrelations of the series.

Int38

Introduction to RATS

Getting StartedA Tutorial

(Basic Statistics) Shortcut for ViewStatistics. Computes and displays descriptive statistics for the selected series, including the number of observations, mean, standard error, and maximum and minimum values of each series. Shortcut for ViewCovariance Matrix. Computes and displays a covariance and correlation matrix for the selected series. Shortcut for ViewData Table. Generates a (read-only) table of data for the selected series.

(Cov./Corr.)

(View Data)

Introduction to RATS

Int39

Getting StartedA Tutorial 1.4.9 Learn More: Arithmetic Expressions

The arithmetic expressions in rats form the basis of the general transformation instruction SET, as well as calculations involving scalar and array variables, which you do primarily with the instruction COMPUTE (see page Int6 and Users Guide Chapter 1). rats expressions are similar to those employed in many programming languages and applications. We give enough detail here to allow you to do almost any standard data transformation.

Constants

You can represent a numeric value using a variety of forms: with or without decimal points: 5 in scientific notation with a suffix of one of the forms: En, E+n, E-n, where n is a non-negative whole number: 2.3E5 (230000) -.4E-4 (-.00004) In addition, rats provides the following two constants: %PI The constant p %NA The missing value code 13.6 -.393

Arithmetic Operators
+ * / ^ or ** += = *= /=

rats supports the following arithmetic operators:

addition subtraction or negation multiplication division exponentiation Increment and assign (a += b is equivalent to a = a+b) Decrement and assign (a = b is equivalent to a = ab) Multiply and assign (a *= b is equivalent to a = a*b) Divide and assign (a /= b is equivalent to a = a/b)

You can use parentheses () to control the order of execution of the operations, and you can nest sets of parentheses if necessary. You cannot use brackets [] or braces {} as substitutes, because these have other meanings in rats. In the absence of parentheses, operations are done in the following order: 1. 2. 3. 4. 5. Negation ( - used to change sign) Exponentiation (exception: -a^b does a^b first) Multiply and divide Add and subtract Logical operators (see below)

If two operations have the same precedence, they are done from left to right, so A-B+C is equivalent to (A-B)+C. The one exception to this is ^: A^B^C is the same as A^(B^C).All of this is based upon natural order of operations in algebraic formulas. Just as a+b/c is interpreted as a+(b/c) in algebra, A+B/C is A+(B/C) in rats.

Int40

Introduction to RATS

Getting StartedA Tutorial

Functions
rats provides many useful functions (over 300). Functions accept zero or more argu-

ments and return a value or set of values. Function arguments must be enclosed within parentheses () and there should not be a space between the function name and the (. Some of the more important ones: LOG(x) SQRT(x) SIN(x) %IF(x,y,z) %VALID(x) natural log (loge) EXP(x) square root ABS(x) sine (of x in radians) COS(x) is y if x is non-zero and z if x is zero. is 0 if x is missing and 1 otherwise ex absolute value cosine (of x in radians)

See Section 2 in the Reference Manual for a complete list of available functions.

Examples
2

Expression b - 4ac 2-C log (1 + w2 )

RATS Code b^2 - 4*a*c 2^-c log(1 + w^2)

Expression 1 (1 + y a - bc d e
-u 2

RATS Code 1/(1 + y^2) a - b*c^d exp(-abs(u))

Logical and Relational Operators

Logical operators have many uses in rats, one of which is creation of dummy variables. They code true/false statements, returning a one for true and zero for false. These are shown below (A and B can be any expression or number.) Expression A==B (or A.EQ.B) A<>B (or A.NE.B) A>B (or A.GT.B) A>=B (or A.GE.B) A<B (or A.LT.B) A<=B (or A.LE.B) .NOT.A A.AND.B A.OR.B Translates to the Statement A is EQual to B A is Not Equal to B A is Greater Than B A is Greater than or Equal to B A is Less Than B A is Less than or Equal to B A is NOT non-zero (that is, A is false) Both A and B are non-zero (true) Either A or B or both are non-zero (true)

The logical operators are lowest in order of precedence, so A+B==C*D is done as (A+B)==(C*D).

The first six (the relational operators) take precedence over the last three, which are listed above in order of precedence: A.OR.B.AND.C.OR.D is done as A.OR.(B.AND.C).OR.D When testing for equality, be sure to use == (two equals), not = (just one)!!!

Introduction to RATS

Int41

Getting StartedA Tutorial

Examples
1 if Y > 3 Y > 3.0 = 0 if Y 3 1 if T 10 and T 20 T >= 10.and.T <= 20 = 0 Otherwise

Integer vs Real Numbers

rats distinguishes between integer and real (floating point) numbers. Integers are

whole numbers which serve primarily as entry numbers, loop indices and subscripts. For instance, the date 1970:1 is processed as an integer, and subscripts such as T-1 are integer expressions. A constant typed without a decimal point is considered to be an integer: 100.0 is real, 100, 1, 2 and 3 are integer. When an operation mixes an integer with a real, the integer is converted to its equivalent as a real number before the operation is completed. For instance, if any one of A, B or C is real, (A+B+C)/3 is the same as (A+B+C)/3.0. Division of two integers results in the integer quotient with no fraction, so 10/3 is 3 and 1/2 is 0. This may or may not be what you intend, so you need to be careful in such situations. If you need to force rats to convert an integer to a real in a situation where it would not be done automatically, use the function FLOAT, for instance FLOAT(T). Similarly, you can convert reals to integers with the function FIX. FIX truncates any remainder. If you want to round first and then convert to integer, do something like FIX(%ROUND(X,-1)) which rounds to the nearest 10 and converts to an integer.

Missing Values

Operations involving a missing value produce a missing value as a result, except for equality and inequality (== and <>). Because missing values propagate automatically, you dont have to worry about making special tests to insure that an expression only operates on valid numbers. An invalid operation of any form will also produce the missing value code as a value. Examples are SQRT(-1.0) and 1/0. rats does not distinguish between illegal operations (such as SQRT(-1.0)) and operations which could be assigned a value of infinity, such as 1./0. Within an expression, you can represent a missing value only as %NA; not as NA or . or any of the alternatives that can be used on data files.

Referencing Series Elements

You can access a particular entry of a series by listing the entry or element in parentheses immediately after the variable name. For series, you would use an expression of the form seriesname(entry), such as: FOODPROD(5) or DISPINC(1939:1). As weve already seen, in a SET instruction you reference the current value of a series (the value at the entry being computed) as seriesname alone, and its lag as seriesname{lag}.

Int42

Introduction to RATS

Getting StartedA Tutorial 1.5 Example Four: Time Series Graphs and Analysis
Now well look at an example to demonstrate some time series filtering and analysis techniques, and explore the graphing capabilities of rats. It uses monthly data on housing starts in the United States, and is also from Pindyck and Rubinfeld (1998), page 480. The instructions are provided on the file ExampleFour.rpf. Before starting, close the Input Window and start a new one (with FileNew: Editor/Text Window). This time, we will use separate input and output windows. Once the editor window is open, hit the toolbar. The data file we want is called EX152.XLS. Use the Data Wizard (Other Formats) operation to open the file. Click on Show Preview to view the contents, and then click on Scan to process the date information. The dialog box should look like this:

The dates are in the form year:month:day, but the file only contains one observation per month. rats is able to recognize this, and correctly identifies this as monthly data. Click on OK. rats will generate and execute the following instructions: open data ex152.xls calendar(m) 1986:1 data(format=xls,org=columns) 1986:1 1995:10 hs6fr

Introduction to RATS

Int43

Getting StartedA Tutorial

Examine the Data
As always, we recommend that you examine the data before proceeding. For example, select Series Window from the View menu, click on the HS6FR series, and do View Time Series Graph. You should see the following:

As you would expect, the series exhibits a very strong seasonal behavior. There also appears to be a decreasing trend over the first few years, followed by an increasing trend over the remaining years.

1.5.1 Filtering and Smoothing

We will examine some techniques for smoothing the data. Start by selecting Filter/ Smooth from the Data/Graphics menu, which will bring up this dialog box:

Int44

Introduction to RATS

Getting StartedA Tutorial

Because we only have one series in memory (HS6FR), rats automatically selects it as the Input Series. The default filter type is a centered moving average filter, which is what we want (centered means that for a window width of 2N + 1 periods, the filtered value at time T is computed using observations T-N through T+N). Type in FLAT7 as the name for the Output Series, and enter 7 as the Width:

Click on OK. rats will generate a FILTER instruction with the appropriate options: filter(type=centered,width=7) hs6fr / flat7 This seven-period moving average should smooth the series considerably, but will probably retain some of the seasonal behavior. To reveal the underlying trend behavior of this series, we can try the Hodrick-Prescott (hp) filter. It is designed to separate the trend behavior of a series from its cyclical component. Select Filter/Smooth again. This time, youll have to select HS6FR as the input series (rats doesnt guess, now that there are more than one series in memory). Enter HPFILTER as the Output Series and select Hodrick-Prescott as the Filter Type. You can leave the Tuning parameter field blank to use the default value. The dialog should look like this:

Introduction to RATS

Int45

Getting StartedA Tutorial 1.5.2 Graphing the Data

Lets examine the results by generating a graph that includes all three series. This time we will use the Graph operation on the Data/Graphics menu. Unlike the View Time Series Graph operation, this actually generates a rats instruction, making it easy to reproduce the graph later. It also gives us more control over the appearance of the graph. Select the Graph operation, which displays the dialog box shown below. The Series field in the middle lists all the series available in memory. We want to graph all three, so click on HS6FR and then <Shift>+click on HPFILTER to highlight the whole list. Click on the button to add the three series to the Base Series list on the left (as opposed to the Overlay Series list on the right, which is used to graph some series using a second vertical scale). Thats all you need to do to produce a graph, but lets add some features. First, type in Moving Average versus HP Filtering in the Footer field at the bottom. The dialog should now look like this:

Next, click on the Key tab, and select Upper Right (Inside) as the position for the key. Then click OK to generate the graph. Heres the resulting Graph Window:

Int46

Introduction to RATS

Getting StartedA Tutorial

As expected, FLAT7 is considerably smoother than the original series, but still shows a significant amount of seasonality. The HPFILTER series, however, seems to fit the overall trend behavior of the original series quite well, with nearly all of the cyclical behavior and short-term fluctuations removed. Here is the GRAPH instruction generated by the wizard. Due to space constraints, we use the line continuation symbol $ here, which tells rats that a command is continued on the next line: graph(style=line,footer="Moving Average versus HP Filtering",$ key=upright) 3 # hs6fr # flat7 # hpfilter The main instruction includes several options: STYLE, which controls the way data is represented; FOOTER, which adds a text label below the graph; and KEY, which adds a key. Both STYLE and KEY select from a specific list of choicessee GRAPH in the Reference Manual for a list of the possible choices. The number 3 at the end of the (continued) GRAPH line tells rats that you are graphing three series. When you are done viewing the graph, close the graph window or click on the input window to bring it back to the front.

Introduction to RATS

Int47

Getting StartedA Tutorial 1.5.3 Detrending, Exponential Smoothing, Forecasting

Pindyck and Rubinfeld also examine the use of exponential smoothing techniques. First, they remove the trend component from the series by regressing it on a linear trend series. Then they apply exponential smoothing to get a smoothed version of the detrended series. They then add back the trend component removed in the first step to get a smoothed version of the original series. This can actually be done in a single step in rats. Well demonstrate that in a moment, First, here is the step-by-step version. The Filter/Smooth operation offers one way of removing a trend using a regression. Select Filter/Smooth, and choose HS6FR as the input series. Type in DETREND as the output series, select OLS Trend/Seasonal Removal as the filter type, and turn on the Trend checkbox:

Click OK to generate the DETREND series. The instruction looks like this: filter(remove=trend) hs6fr / detrend Next, we need to compute and save the trend component that was removed, by subtracting the detrended series from the original series. You could use the Transformations wizard (on Data/Graphics) to do this, but you may find it easier just to type in the SET instruction yourself: Now, we can do our exponential smoothing on the detrended series. Because it is time-series specific, the Exponential Smoothing wizard is located on the Time Series menu. Go ahead and select Exponential Smoothing. Select DETREND as the input series, None as the Form for Trend, and type in the name ESMOOTH for the output series. The textbook uses a value of 0.2 for the level smoothing parameter a, so enter that value in the Level field. The dialog box should look like this:

set removed = hs6fr - detrend

Int48

Introduction to RATS

Getting StartedA Tutorial

Click on OK to generate the ESMOOTH instruction:

esmooth(alpha=.2,smoothed=esmooth) detrend Now we need to add back in the trend component removed earlier. We can replace the existing values of ESMOOTH with the sum of those values and REMOVED: Now, select the Graph operation again, and choose HS6FR, HPFILTER, and ESMOOTH as the series to be graphed. To label the graph, enter Exponential Smoothing vs HP Filter as the header. On the Key tab, choose Above (Outside) and turn off the Box Around Key?:

set esmooth = esmooth+removed

Introduction to RATS

Int49

Getting StartedA Tutorial

This generates the instructions and graph shown below: graph(style=line,header="Exponential Smoothing vs HP Filter",$ key=above,nokbox) 3 # hs6fr # hpfilter # esmooth

Modelling the Trend and Seasonal Behavior

As noted earlier, there is an easier way to do this, which is to include a component for the trend behavior in the exponential smoothing model. Select Time SeriesExponential Smoothing again. This time, select the original series HS6FR as the series to model. Then, select the Linear choice for the trend model in the Form for Trend box, and turn on the Estimate switch in the Smoothing Parameters box (which tells rats to estimate the smoothing parameters used in the model). Thats all you need to reproduce the previous analysis, but lets go a couple of steps further, by also including a seasonal term in the model and using the resulting model to produce forecasts. In the Form for Seasonal box, select Multiplicative. Provide a name for the smoothed series (we used ESMOOTH again here). Finally, enter a name in the Forecasts to box to hold the forecasts (we use ESMOOTH_FORE), and enter 8 as the number of forecast steps to compute:

Int50

Introduction to RATS

Getting StartedA Tutorial

When you are ready, click OK to do the smoothing. The resulting instruction is shown below, along with a GRAPH instruction that plots both the original series and the forecasts. esmooth(trend=linear,seasonal=multiplicative,estimate,$ smoothed=esmooth,forecast=esmooth_fore,steps=8) hs6fr graph(style=line,header="Exponential Smoothing Forecast", $ key=above,nokbox) 2 # hs6fr # esmooth_fore

Introduction to RATS

Int51

Getting StartedA Tutorial

Seasonally Adjusting Data
If you just want to seasonally adjust (de-seasonalize) the data, you would include a seasonal component in the ESMOOTH model, but omit the trend term. The ESMOOTH below does just that, and saves both the smoothed (seasonally adjusted) values and the fitted values. This also uses the ESTIMATE option to estimate the values of the model coefficients.

esmooth(seasonal=multiplicative,estimate,$ smoothed=smoothed,fitted=fitted) hs6fr

graph(header="Seasonally Adjusted Data",key=above,nokbox) 3 # hs6fr # smoothed # fitted

As you can see from the graph, the seasonal component has been removed from SMOOTHED, leaving the underlying trend and short-term fluctuations. Exponential smoothing offers a relatively simple form of seasonal adjustment. You can also treat seasonality using state-space models (Users Guide, Chapter 10), or, if you have a Professional level of rats, you can use the more sophisticated Census X11/X12 adjustment procedure.

Int52

Introduction to RATS

Getting StartedA Tutorial 1.5.4 Regression-based Forecasting

The graph shows that a general upward trend from 1992 on in the seasonally adjusted series. Suppose we want to generate an out-of-sample forecast of that trend through 1996. One way to do that is to do a simple regression on a constant and a trend series. First, we need to define the trend series. Because well be forecasting out-of-sample, we need to define that trend series beyond the end of the actual data to include the forecasting range (the special built-in CONSTANT series is automatically available at all time periods). Select Trend/Seasonals/Dummies from the Data/Graphics menu. Enter TREND as the series name, and 1996:12 as the ending date:

This generates the instruction: set trend * 1996:12 = t which uses the (normally omitted) start and end parameters on SET. The * symbol used for the start parameter tells rats to use the earliest date in the default range (entry one in this case). Specifying 1996:12 as the end date defines the trend through the end of our forecast range. The T on the right side of the equal sign is a reserved variable in rats. It returns the current entry being set or evaluated. So, this creates a series called TREND that contains the number 1 in entry 1, the number 2 in entry 2, and so on. Next, select Linear Regressions from the Statistics menu, and enter SMOOTHED (the seasonally adjusted series) as the dependent variable, 1992:1 as the starting date, CONSTANT and TREND as the explanatory variables, and FORECAST_EQ as the equation to define. The dialog box should look like this:

Introduction to RATS

Int53

Getting StartedA Tutorial

This generates the following LINREG:

linreg(define=forecast_eq) smoothed 1992:1 * # constant trend The DEFINE option creates an equation called FORECAST_EQ containing the structure and estimated coefficients from the regression. Well use that to produce forecasts. Now select Single-Equation Forecasts from the Time Series menu. The Equation field tells rats which equation you want to forecast. The default choice (Last Regression) would actually work, because the last regression performed is the one we want to forecast, but go ahead and select FORECAST_EQ. Then, enter 1995:11 and 1996:12 as the starting and ending date for the forecasts. Finally, enter HSFORE in the Forecasts To field to save the computed forecasts into a series with this name:

Int54

Introduction to RATS

Getting StartedA Tutorial

Click OK to compute the forecast. This generates the following UFORECAST instruction (the U stands for univariate, or single-equation): uforecast(from=1995:11,to=1996:12,equation=forecast_eq) hsfore To see the results, execute the followingnote the use of the date ranges to limit the graph to our estimation and forecasting range (we dont need a range on HSFORE because it is only defined over the range we want to see): graph 3 # hs6fr 1992:1 * # smoothed 1992:1 * # hsfore

Introduction to RATS

Int55

Getting StartedA Tutorial 1.5.5 Learn More: Graph Windows

Graph Windows are, not surprisingly, windows used to display graphs. Each graph generated gets its own window. You may note that Graph Windows have relatively few toolbar icons. With rats, you create the appearance that you want using the graphics commands, rather that adding them (manually) afterwards. This gives you graphs that are perfectly reproducible. Chapter 3 is devoted to the tools for making high-quality graphs. rats allows all types of special effects, such as shading regions, adding labels and text, and configuring colors, line and fill patterns. Programs can sometimes generate a large number of graphs. To reduce clutter, you might want to minimize graphs that you dont need to look at right now (but want to keep available)you can always reload them from the Window menu. Theres also a Close All Graphs on the Windows menu which will close (permanently) all the current graph windows.

Resizing Graphs

You can resize a graph window by clicking and dragging a corner of the graph. rats will automatically adjust the elements of the graph (font size, etc.) to fit it within the new window. If you squeeze the graph down to be very narrow in one dimension or the other, the fonts may become hard to read. However, as long as you keep the proportions within normal ranges, it should remain readable. When you print a Graph Window or export it for use with a word processor, rats normally uses proportions close to a golden ratio, with the width being about 1.5 times the height. This will not necessarily match the proportions that you are seeing on the screen. If youve changed the shape of the graph and want to use those proportions, click on the Fix button ( ). To cancel that, click on the ( ).

Color or Black and White

By default, graphs are displayed in color, with different colors used for each series. If you print a graph to a black and white printer, rats will automatically convert to using different patterns to distinguish series. You can preview what the graph will look like in black and white mode by clicking on the toolbar button. The precise set of patterns (or colors, for that matter), can be configured, if the standard ones dont work well or your publication needs a standard style. See page Int149.

Saving and Copying Graphs

You can use FileSave As... or FileExport... to save a graph to disk. If you want to be able to re-open the graph in rats, save the graph in our rgf (rats Graph Format). If you want to import the graph into another document, use one of the other formats, with PostScript usually being the best option. You can also include instructions in your program to save graphs automatically (page Int152). You can also use EditCopy, <Ctrl>+C, or Copy on the contextual menu to copy a graph to the clipboard for pasting into another application. The formats available depend on the platform you are using (page Int129).

Int56

Introduction to RATS

Getting StartedA Tutorial 1.5.6 Learn More: Long and Short Program Lines
Continuation Lines, $ Symbol
rats will accept instructions of any physical length. Sometimes, though, a line

becomes so long that it is difficult for you to handle. To split a line into more manageable segments, put a $ character at the end of the unfinished line. You just have to be careful where you break a line. Dont break up tokens, things like numbers, variable names and quoted strings. The best locations are before or after commas or arithmetic operatorswith after being the easiest to read, since its clear that the expression isnt done. An example is set gnp = consumption + investment + govtexpend + $ export - import There is no limit on the number of continuation lines allowed for a single instruction. Some statistical and programming languages require a symbol, often a semi-colon (;), at the end of each command to signal to the program that a command is complete. That is not the case with ratsit automatically assumes a command is complete when it hits the end of the physical line of text. Thus the only time you have to include a symbol (the $) at the end of a physical line is if the command isnt done, and continues to the next line of text.

Multiple Statements on a Line, the ; Symbol

You actually can use the semi-colon (;) in rats to signal the end of a command, but this is only necessary if you want to include multiple commands on a single line. This is usually only a good idea for a set of short instructions which you see almost as one, or for attaching short comments to instructions. Examples: set gdp = log(gdp) ; set m1 = log(m1) smpl 1960:1 1985:3 ;* Set Sample for Remaining Instructions

Introduction to RATS

Int57

Getting StartedA Tutorial 1.5.7 Learn More: RATS Program Files (RPF)
At this point, you should have a complete Editor Window with the instructions for the analysis that weve done with this housing data series. The advantage of using the separate I/O windows is that we have the input commands by themselves. We will now save this as a rats program file which we can run later. We use the extension .RPF for program files. These are simple text files, so you can open them with any word processing program, but the .RPF extension allows us to associate the files with the rats software.

Saving a Program

Select FileSave (or FileSave As... or the toolbar) and assign the file a name. For example, you might save this as MyProgram.rpf. You can also save the file when you close the windowyoull be first asked if you want to save the changes, and if you answer Yes, youll be prompted for the file name.

Executing a Saved Program

Now, close all the open windows (using WindowsClose All). Because you are closing the input window (with your instructions), this will clear the rats memory as well. Do FileOpen... and re-open the file you just saved. Again, do the to get a separate output window. To re-execute the entire program, do EditSelect All (or click on the button) and hit <Enter> (or click on the run icon: ).
rats will execute all of the instructions in the file in order, generating text output in

the output window, along with the various graphs and report windows. Thats how easy it is to reproduce a set of results! You can do the same thing with any of our example programs.

.PRG Extension

Historically, we used the extension .PRG for program files, but have phased that out because it was not unique to rats, so any of a number of programs might actually end up being associated with it, and because it caused problems with some e-mail filters since one possible use of .PRG is for executable batch files. However, since there are a large number of existing rats program files which were given .PRG extensions, we still include that as a possible extension in the file dialog.

Int58

Introduction to RATS

Getting StartedA Tutorial 1.5.8 Learn More: The PRINT Instruction

Another useful instruction for examining your data is PRINT. This displays series to the output window or (with a WINDOW option) to a spreadsheet-style window. If you want rats to print all the data for all the series in memory, just type:

print To view data over a limited range of entries, you can add start and end parameters with the desired dates (or entry numbers): print 1990:1 1996:6 You can also print only certain series by listing them after the dates: print 1990:1 1996:6 hs6fr esmooth esmooth_fore If you want all the data but only for certain series, you can replace the start and end parameters with a forward slash (/). This tells rats to use the default range for the series: print / hs6fr esmooth esmooth_fore To display the information to a spreadsheet style window, add a WINDOW option and a title for the window: print(window="ESMOOTH Forecasts") / hs6fr esmooth_fore You can copy and paste the contents of the resulting window into another program, or export the window contents to a spreadsheet or other file type using FileExport.... PRINT is also very useful when you are troubleshooting a program. When you get unexpected results, such as an error message complaining about missing values, try printing out the series involved. You may notice an error that had slipped through earlier checks. You can use the PICTURE option to supply picture codes (page Int5) to control the formatting of the output. In this case, we use an * on the left of the decimal, which tells rats to use as many digits as necessary, while the .# tells rats to round the output to one decimal (see DISPLAY in the Reference Manual for more on picture codes):
print(picture="*.#") / hs6fr ENTRY HS6FR 1986:01 105.4 1986:02 95.4 1986:03 145.0 1986:04 175.8

Introduction to RATS

Int59

Getting StartedA Tutorial 1.5.9 Learn More: Data Transformations

SET (introduced on page Int25) is the general data transformation instruction. rats also has several special-purpose instructions for doing transformations, such as SEASONAL for creating seasonal dummies and DIFFERENCE and FILTER (page Int44) for difference and quasi-difference transformations. However, most of the time you will be using SET. Below, we describe SET in more detail, and present a kit of standard data transformations. Most of these examples are not based upon the data sets we have been using, so you wont necessarily be able to execute them as part of the tutorial. It should be easy to adapt them to your own needs, however. We also describe the various Wizards available for doing many of these types of operations.

The Instruction SET

The general form of SET is set( options ) series start end = function(T) In the function part of the SET instruction, you can use constants, scalar variables, other series, matrices, calls to built-in or user-defined functions, and any of the arithmetic and logical operators available in rats. In its most basic form, SET defines one series as a simple transformation of another series. For example: set loggdp = log(gdp) sets each entry of the series LOGGDP to the log of the corresponding entry of GDP, using the built-in function LOG() (see page Int41 for more on functions). The start and end parameters are optional, and you can usually skip them when you are doing your initial transformations after DATA. To set a series as a function of other variables, you use standard arithmetic notation: set totalcon = durables + nondur + services set scaled = resids/sigmasq

Trend Series

You use the variable T to create trend series and dummy variables based upon time. T is equal to the number of the entry being set, where the entry you specified on CALENDAR is given number 1. For instance, the following creates (in order) linear, quadratic and exponential (5% growth) trend series: set trend = t set trendsq = t^2 set exptrend = exp(.05*t)

Int60

Introduction to RATS

Getting StartedA Tutorial

Seasonal Dummies
rats provides a special instruction SEASONAL for creating seasonal dummies. You

use it in one of two forms:

seasonal seasons seasonal(period=1948:2) february 1948:1 2009:12 The first creates SEASONS as a dummy for the last period of the year (4th quarter or December). By using SEASONS and its leads in a regression, you can get a full set of dummies without having to define a separate one for each period. With monthly data, the lag field SEASONS{-10 TO 0} covers dummies for February (lag 10, which is a 10-period lead) to December (lag 0). The second creates FEBRUARY as a February dummy defined from 1948 to 2009. The PERIOD option specifies the first period to receive a value of 1 (February, 1948).

Other Dummies

Dummy variables are easy to construct using logical and relational operators (see page Int41), since these operators return the values zero or one. set dummy = t>=1972:3.and.t<=1974:3 set large = pop>5000 set female = .not.male The first example creates a dummy series called DUMMY using a logical expression. It stores the value 1 in entries where the logical expression is true, and a 0 in the other entries. In this case, the expression is true between 1972:3 and 1974:3 inclusive, so DUMMY is 1 for those entries and 0 elsewhere. The second sets LARGE to 1 when the corresponding entries of POP are greater than 5000, and 0 elsewhere. In the third example, entries of FEMALE are 1 when MALE is 0, and are 0 elsewhere.

Trend/Seasonals/Dummies Wizard

You can create many of the trend, seasonal, and dummy variables described above using Trend/Seasonals/Dummies on the Data/Graphics menu.

Lag and Lead Transformations

Transformations involving lags or leads of a series can be written using the T subscript, lag notation, or a combination of both. set pavge = ( price + price{1} ) / 2.0 set pavge = ( price(t) + price(t-1) ) / 2.0 set inflation = 400.*log( deflator/deflator{1} ) The first two are identical. They create PAVGE as the average of the current and first lag values of PRICE. The first uses lag notation, the second uses T explicitly. Which style you use is a matter of taste. The third example computes annualized growth rates of DEFLATOR (in percent) using the log difference approximation.

Introduction to RATS

Int61

Getting StartedA Tutorial

All of these transformations involve a lagged value, so the first entry will be set to missing: if PRICE is defined over 1922:1 to 1941:1, PAVGE will be defined only over 1923:1 to 1941:1.

Differencing

Simple differencing is, of course, easy to handle with SET: set diffgdp = gdp - gdp{1}

rats also offers the instruction DIFFERENCE for regular, seasonal, or fractional dif-

ferencing (see page Int31 on the use of the /):

difference gdp / diffgdp difference(sdiffs=1,diffs=1) gdp / ddsgdp As noted earlier, you can also use Transformations or Differencing on the Data/ Graphics menu to create differenced series.

Growth Rates

There are several ways to compute growth rates or approximations to them. The first two SET instructions below compute (for quarterly data) annualized growth rates, the third and fourth compute period over period rates and the last computes the year over year growth. set growx = 4.0*log(x/x{1}) set growx = ((x/x{1}) ^ 4 - 1.0) set growx = log(x/x{1}) set growx = x/x{1} - 1.0 set growx = x/x{4} - 1.0 Transformations on the Data/Graphics menu can also create simple growth rate series.

Benchmarking and Normalizing

These generally require one of the two options for SET; either FIRST or SCRATCH: set(first=100.) capital 1920:1 1941:1 = .90*capital{1}+invest computes CAPITAL from an investment series beginning with a value of 100.0 for the first period (1920). set(scratch) gdp 1950:1 1998:4 = 1000.0*gdp/gdp(1975:3) renormalizes GDP to take the value 1000 in 1975:3. You need the SCRATCH option because it is replacing the values of the GDP series that are needed in the calculation, and is using two entries of it (current entry and 1975:3) . See SET in the Reference Manual for more on these options.

Int62

Introduction to RATS

Getting StartedA Tutorial

The %IF FunctionConditional Expressions
The logical function %IF(x,y,z) returns the value y if x is non-zero, and returns z if x is zero. This is a very useful function. For example: set w = %if(test<0, %na, w) set testseries = %if(t<=1990:12, seriesone, seriestwo)

The first makes all entries of W for which the series TEST is negative equal to missing (%NA is how you represent the missing value in expressions). The other entries are not affected. The second stores the values of SERIESONE in TESTSERIES for all entries through 1990:12. Entries from 1991:1 on are taken from SERIESTWO. Note that the %IF function only evaluates a singlevalued expression each time it is called. It is the SET instruction itself that loops over the entries in the sample range, calling %IF once for each entry in that range.

Missing Values

SET propagates missing values through the formula. With only two exceptions (the %VALID(x) function and the %IF(x,y,z) function), any operation which involves a missing value returns a missing value. SET also sets to missing any observation which involves an illegal operation, such as divide by zero and square root of a negative number. set stockprice = %if(%valid(stockprice),stockprice,-999.0) This replaces all the missing values of STOCKPRICE with the value 999.00. You might use this in preparation for exporting this series to a file format (such as ASCII) which doesnt support special codes for missing values.

Introduction to RATS

Int63

Getting StartedA Tutorial 1.5.10 Learn More: Forecasting

Whats Available?
rats can forecast using a wide range of modelling techniques:

Simple regression models (Section 1.5.4 and Users Guide Section 5.2) Simultaneous equations (Users Guide, Chapter 8) Vector autoregressions (vars) (Users Guide, Chapter 7) Exponential smoothing (Section 1.5.3 and Users Guide Sections 6.2 and 6.3) Box-Jenkins (arima) models (Users Guide, Sections 6.2 and 6.4) Spectral methods (Users Guide, Sections 6.2 and 6.5) Chapter 5 of the Users Guide provides a general overview of forecasting, while specific kinds of forecasting models are addressed in the sections noted above.

Wizards and Instructions

You can use the PRJ instruction (page Int71) to compute out-of-sample fitted values for the most recently executed regression, as long as the necessary right-hand-side variables are defined for the required time periods. We introduced the UFORECAST instruction and the Single-Equation Forecast Wizard on page Int54. The other, more general, forecasting instruction is FORECAST. It can forecast one equation or many, and can also be used to forecast non-linear models. You can use VAR (Forecast/Analyze) on the Time Series menu to generate a FORECAST instruction once youve defined a proper model. By default, both UFORECAST and FORECAST do dynamic forecasting. That is, when computing multi-step forecasts, they use the forecasted values from initial steps as the lagged dependent variable values in computing the later forecast steps. Both can also do static forecasting if you use the STATIC option. Static forecasting means that actual values are used for any lagged dependent variable terms in the model. These instructions are discussed in detail in Chapter 5 and 6 of the Users Guide and in the Reference Manual.

Equations and Formulas

Section 1.5.4 introduced the idea of defining equations and using UFORECAST to compute forecasts using equations. FORECAST can also produce forecasts for a formula, or a set of formulas. Note the distinction between equations and formulas:

Int64

Introduction to RATS

Getting StartedA Tutorial

Equations are descriptions of linear relationships. You can define them directly using the instruction EQUATION, but usually you create them using the DEFINE options of estimation instructions, as shown in Section 1.5.4. With UFORECAST, you use an EQUATION option to supply the equation you want to forecast. With FORECAST, you either list the equations you want to forecast on supplementary cards (one card per equation), or you use the MODEL option to specify a MODEL (group of equations and/or formulas) that you want to forecast. MODELS are usually constructed using the GROUP instruction. (also called FRMLs) are descriptions of (possibly) non-linear relationships. These store a SET style function together with the dependent variable, and are normally defined using the FRML command. Before you can forecast a formula or set of formulas, you must group them into a model using the GROUP instruction. You then use the MODEL option on FORECAST to forecast the model. See page Int84 and Chapter 8 of the Users Guide for more on formulas.

Formulas

Simulations

You can compute random simulations of a model using UFORECAST, or with the SIMULATE command, whose syntax is virtually identical to that of FORECAST. You can also use other tools to supply your own shocks to a systemsee Chapter 16 of the Users Guide.

Introduction to RATS

Int65

Getting StartedA Tutorial 1.6 Example Five: Cross-Sectional Data

For our next example, we will look at a cross sectional data set. This data is taken from Chapter 2 of Verbeek (2008) and is provided on a text file called WAGES1.DAT. The file contains survey data for 3294 young workers, and includes four series: WAGE (hourly wage in dollars), MALE (a 1/0 dummy variable, equal to 1 for male respondents and 0 for female respondents), EXPER (years of experience at job), and SCHOOL (years of schooling). This example is provided on the file ExampleFive.rpf. To get started, close the windows from the previous session and do FileNew-Editor: Text Window to create a new input window. When you are ready, select the Data Wizard (Other Formats) operation and open the WAGES1.DAT. (you may need to select Text Files (*.*) from the list of formats in order to see the file in the dialog box). Click on Show Preview to examine the data on the file:

1.6.1 Reading the Data

The data run down the page in columns, so the default settings should be correct. This file includes series names, so rats wont need to ask you to provide names for each series.

Int66

Introduction to RATS

Getting StartedA Tutorial 1.6.2 The SMPL Option: Selecting Sub-Samples

Our first task is to examine the average wages for both males and females. As noted earlier, the series MALE is equal to 1 for males, and 0 for females. Recall from Section 1.4.2 that we can use the STATISTICS instruction to compute averages. But how to compute an average of WAGE using only certain observations (for males only, and for females only)? The answer is to use the SMPL option (short for sample), which allows you to specify a sub-sample using a logical and/or numeric expression. The syntax is simply smpl=expression(t) Observations for which expression(t) returns a logical true or a non-zero numeric value are included in the operation. Observations where expression(t) returns a logical false or the number 0 are omitted from the operation. For example: stats(smpl=male) wage computes summary statistics for WAGE, using only observations where MALE is 1
Statistics on Series WAGE Observations 1725 Skipped/Missing 1569 Sample Mean 6.313021 Variance 12.242031 Standard Error 3.498861 SE of Sample Mean 0.084243 t-Statistic (Mean=0) 74.938512 Signif Level (Mean=0) 0.000000 Skewness 1.921402 Signif Level (Sk=0) 0.000000 Kurtosis (excess) 8.845542 Signif Level (Ku=0) 0.000000 Jarque-Bera 6685.148147 Signif Level (JB=0) 0.000000

Using the Univariate Statistics operation, the dialog box would look like this:

Most instructions (and associated wizards) that deal with series have the SMPL option.

Introduction to RATS

Int67

Getting StartedA Tutorial

To get the averages for female workers, you just apply the logical operator .NOT. to the MALE series: stats(smpl=.not.male) wage This computes results using only observations where MALE is not equal to 1. This could also be written as: stats(smpl=male<>1) wage You could also create a series (using SET or the Transformations wizard) containing the reversed dummy variable and use that for the option. While probably not important here, defining the separate series can make it easier to see whats happening. If you create a FEMALE series using: set female = .not.male you can then use stats(smpl=female) wage If you are familiar with certain other statistics program, you might have expected that you would need an IF statement to accomplish this. While there is an IF instruction in rats, it is used to evaluate a single condition, not as an implied loop over a range of observations.

1.6.3 Testing Regression Restrictions

linreg wage # constant male

Another way to examine the respective wages for males and females is to regress the WAGE series against a constant and the MALE dummy variable:

The coefficient on the CONSTANT will be the average female wage, while the coefficient on the MALE variable will be the difference between male wages and female wages (that is, it reflects the impact of being male on wage). The results show an average female wage of $5.15, with males making an additional $1.17 per hour, for an average of $6.32, with a standard error of $0.11 on the differential. To determine whether gender truly affects wages earned, Verbeek considers whether other factors, such as years of schooling and years of experience on the job, may explain some or all of the discrepancy. One way to do that is to compare the restricted regression used above (restricted because the schooling and experience variables are omitted) with the unrestricted regression which does include the additional variables to determine whether the additional components are significant.
rats offers an instruction specifically for testing these types of exclusion restrictions.

Lets look at the regression with some added factors: linreg wage # constant male school exper

Int68

Introduction to RATS

Getting StartedA Tutorial

You can simply write exclude # school exper This produces the following:
Null Hypothesis : The Following Coefficients Are Zero SCHOOL EXPER F(2,3290)= 191.24105 with Significance Level 0.00000000

The null hypothesis that the added coefficients on schooling and experience are zero (and thus can be excluded from the model) is tested with an F statistic with 2 and 3290 degrees of freedom. The result is significant beyond any doubt, so these variable do seem to help explain wages. However, the regression coefficient on MALE actually goes up in the expanded regression, so the use of the additional factors cant explain the difference between males and females.

The Regression Test Wizard

You can also generate this test using the Regression Tests operation on the Statistics menu, which displays this dialog box:

Here, we want to select the Exclusion Restrictions option, which tests that some set of coefficients all have zero values. Thats the simplest of the three options. The Other Constant Restrictions is used when you have a test on one or more coefficients with a null hypothesis that they take specific values, not all of which are zero. (A common case, for instance, is that the intercept is zero and a slope coefficient is one). General Linear Restrictions is used when you have restrictions that involve a linear combination of several coefficients (examples are that coefficients summing to one, or coefficients being equal, so their difference is zero).

Introduction to RATS

Int69

Getting StartedA Tutorial

When you click OK with the Exclusion Restrictions on, you get a second dialog box, listing the regressors from that last regression. Select (highlight) the SCHOOL and EXPER series:

Then click OK to run the test. This generates a TEST instruction with a ZEROS option. With TEST, you use numbers rather than variable names to indicate which terms are being tested. In this case, we are testing that the third and fourth regressors in the LINREG are (jointly) zero, so the instruction generated by the wizard is: test(zeros) # 3 4

More on Hypothesis Testing

Chapter 3 of the Users Guide provides full details on hypothesis testing, including detailed information on the instructions described above, as well as the specialized instructions SUMMARIZE and RATIO. You will also find out how to implement serial correlation and heteroscedasticity tests, Chow tests, Hausman tests, unit root tests, and many more.

Int70

Introduction to RATS

Getting StartedA Tutorial 1.6.4 X-Y Scatter Plots: The SCATTER Instruction
Weve looked at creating time series (data vs. time) graphs with GRAPH. You can also graph one series against another to produce a scatter or XY plot. To demonstrate, lets look at the relationship between wages earned and years of schooling. First, we will regress WAGE against a constant and SCHOOL, and compute the fitted values from this regression. You can do that using the Linear Regressions wizard, adding a series name in the Fitted Values to field to save the fitted values, or you can type the instructions directly, using the instruction PRJ (short for project), which computes the fitted values from the most recent regression: linreg wage # constant school prj wagefit To draw the graph, select Scatter (XY) Graph from the Data/Graphics menu to bring up the Scatter Graph wizard. Select (highlight) both WAGE and SCHOOL from the list of series and click on to add the pair to the list of series to be graphed. When prompted, choose the combination with SCHOOL as the X-axis variable and click OK. Then, select Symbols as the Style. Now, select WAGEFIT and SCHOOL and click on the button to add this pair to Overlay Series list. SCHOOL should again be the X-axis variable. Select Line as the Style, and turn on the Same Scale as Base? switch (we want both pairs of series to be graphed using the same vertical scale). In the dialog box below, weve also added labels for the horizontal and vertical axis:

Introduction to RATS

Int71

Getting StartedA Tutorial

Clicking OK generates the following SCATTER instruction, and the graph shown below: scatter(style=symbols,overlay=lines,ovsame,$ vlabel="Hourly Wages",hlabel="Years of School") 2 # school wage # school wagefit

As you can see, the OVERLAY option allows us to combine (overlay) two different styles (possibly with two different scales) on a single graph.

Int72

Introduction to RATS

Getting StartedA Tutorial 1.6.5 Learn More: Comment Lines and Comment Blocks
If you open up ExampleFive.RPF, you will see, in addition to the instructions that we just generated, lines like the following: * Adds school and exper to the regression and test the joint * significance of the two additional variables. These are comments that we added to explain whats going on to someone reading the example file, but not following it in this section. Its generally a good idea to add comments to any program file that you save.

Comment Lines

A line with * as the first non-blank character is treated as a comment and is ignored. We use comments rather extensively in the examples of this manual to help you understand the programs. However, note that we also often use marginal comments which are set off from the rest of the line by being in italic type rather than bold face. This is not legal in an actual rats program.

Comment Blocks

You can also mark an entire block of lines as a comment by using the symbol /* to mark the beginning of the block, and */ to mark the end of the block. (Note: the /* must be the first (non-blank) characters on the lines). While this provides a convenient way to enter several lines of comments, it is probably most useful commenting out an existing block of instructions. For example: * This is a comment line /* This is a comment block linreg y # constant x1 */ linreg y # constant x1 x2 This skips the first regression, but does the second.

Tips on Using Comments

Your first goal is always to get the calculations correct. However, if you have a program (or part of one) that you expect that you might need again or that others might want to read or use, its a good idea to add comments describing what each section of code does. Also, if you have a part of your program which youre not sure is correct, commenting it can often help you spot errors. (If you cant explain why youre doing something, that might be a good sign that youre doing it wrong).

Introduction to RATS

Int73

Getting StartedA Tutorial

Now its possible to overdo the comments. You may have seen some programs with a very elaborate comment block like: ***************************************************** * This is a program that analyzes the population * * of kangaroos in New South Wales. * *****************************************************
rats provides several ways to help manage larger blocks of comments. The first are

the comment block symbols. So

/***************************************************** This is a program that analyzes the population of kangaroos in New South Wales. ******************************************************/ gives a similar appearance to the block above, but you can more easily edit the comments as you dont have to worry about having the * symbols at the start of each line. The Edit menu provides operations for commenting or un-commenting sections of your code. If something not quite as pretty will do, you can use the Format Comments operation on the Edit menu. This takes a selected block of text and creates a set of one or more comment lines of roughly equal length (except for the final line). If some of the lines are already comments, the *s and leading blanks are stripped out before the lines get formatted. As an example, we pasted in the first two sentences from above as a single line and did Format Comments. The result is: * If something not quite as pretty will do, you can * use the Format Comments operation on the Edit * menu. This takes a selected block of text and * creates a set of one or more comment lines of * roughly equal length (except for the final line). The desired line length can be set on the Editor tab in the Preferences dialog (Section 1.3 of the Additional Topics pdf). The comments above used a line length of 50, to fit the page, but this is usually set to something longer than that (70-80 tend to work best). You may want wider comments in your actual program files. Note that the comment length isnt a hard limit. In order to roughly balance the lines, it might go a bit above that. There are also a couple of other comment operations on the Edit menu. CommentOut Lines takes the selected text and inserts * at the beginning of each; converting each into a comment. The reverse is Uncomment Lines, which strips the lead * off any line which has one. These allow you to take a block of lines out of the execution and put them back in.

Int74

Introduction to RATS

Getting StartedA Tutorial 1.6.6 Learn More: Linear Regression Instructions

Robust (HAC) Standard Errors
Youll notice that weve used only a few features off the Linear Regressions Wizard (page Int28). Well talk a bit here about what else is available.

This is a checkbox which you can use to change the way that the regression standard errors are computed. It wont change the coefficients; just the standard errors, the covariance matrix, and any regression tests that you do afterwards. If you apply this to our last linear regression in this example, you will get the following instruction: linreg(robust) wage # constant school so you could also simply have inserted the ROBUST option into the existing instruction. If you compare the output, youll see a few differences. First, it now says:
Linear Regression - Estimation by Least Squares With Heteroscedasticity-Consistent (Eicker-White) Standard Errors

so while it still estimates by least squares, the standard errors are now computed using the Eicker-White methods, which are robust to certain types of heteroscedasticity (hence the name of the option). If you arent familiar with this, see Section 2.2 in the Users Guide. You also might notice that the regression F isnt shown: thats a test which is computed under the assumption of homoscedastic errors, and youve said (by using the ROBUST option) that that isnt true. The other summary statistics are the same, as are the coefficients, but the standard errors are (slightly) different, so the T-Stat and Signif columns change as well. When you click on the Robust (HAC) Standard Errors check box, you will also notice that two additional fields below it become active: the Lags/Bandwidth and Lag Window Type. These arent of any special interest in this example, since this is cross section data, but these allow you to compute standard errors which are also robust to autocorrelation (the hac is short for Heteroscedastic Auto Correlated). If we applied that here (not that we should), the new instruction would be something like: linreg(robust,lags=3,lwindow=newey) wage # constant school

Residuals

Further down in the wizard, you will notice a drop down box labeled Residuals To. Whenever, you run a regression, rats creates a series named %RESIDS and fills it with the residuals. If you want to graph the residuals, or compute its statistics, you can just use %RESIDS like any other series. However, if you run a new regression, %RESIDS will now get the residuals from that. If you want to keep a copy of the residuals from one specific regression, you can give a (valid) series name in this box, which will generate the new instruction (see page Int31 for a description of the /):

Introduction to RATS

Int75

Getting StartedA Tutorial

linreg wage / u # constant school

Show Standard Output

This is a checkbox which you will almost always want to be on. If you click it off, you will generate an instruction like this, adding the NOPRINT option: linreg(noprint) wage # constant male

Almost all instructions which ordinarily would produce some output will have the option of NOPRINT. Why would you want to do that? Perhaps all you really want from the regression are the residuals. Or the defined equation for forecasting. When you run a regression, almost anything computed by the LINREG is available for further calculations. For instance, the coefficients are in a vector called %BETA; the sum of squared residuals are in %RSS. Perhaps you only need those. If you plan ahead and use NOPRINT, you dont have to read past the unwanted regression output.

Show VCV of Coefficients

This is a checkbox that you will usually want to be off. If it is on, the estimation instruction will include the VCV option (using the larger regression from the example): linreg(vcv) wage # constant male school exper

This places the following below the standard regression output (and puts a similar report into the WindowReport Windows list):
Covariance\Correlation Matrix of Coefficients Constant MALE SCHOOL EXPER Constant 0.216203148 -0.15560300 -0.90642459 -0.54623329 MALE -0.007790538 0.011594097 0.09663502 -0.10932520 SCHOOL -0.013822338 0.000341249 0.001075567 0.18093999 EXPER -0.006035400 -0.000279728 0.000141010 0.000564669

This is a square table with covariances on and below the diagonal and correlations above the diagonal. For instance, the estimated variance of the coefficient on MALE (the number at the intersection of the MALE row and MALE column) is roughly .0116. The estimated correlation between the coefficients on MALE and EXPER is where the MALE row and the EXPER column meet (above the diagonal), roughly -0.109. The covariance of the same two coefficients is at the EXPER row and MALE column (-.00028). Note that the header says Covariance\Correlation. We use the \ to help you remember the structure of the matrix: covariance below and correlation above. We recommend that you not use this on a regular basis simply because there is little useful information in it. The covariances are used in doing joint hypothesis tests as in Section 1.6.3, but you will be using instructions like TEST and EXCLUDE that do the computations for you, rather than using numbers out of this table.

Int76

Introduction to RATS

Getting StartedA Tutorial

The Method Popup Box
Everything we have done so far has used the (default) choice of Least Squares in the Method box at the top left. There are several other choices available, some of which generate LINREG with a different set of options; others generate different instructions with a similar overall look to LINREG. The next two choices are Two Stage Least Squares and GMM (IV with Optimal Weights). If you choose one of those, youll notice that the Instruments box and button become enabled. These are (closely related) instrumental variables estimation methods which require both a list of regressors and a list of instruments. (You can read more about instrumental variables in Section 2.5 of the Users Guide.) You add variables to the instruments list just as you do for the regressors: you can either type them directly into the box, or you can use the button to pop up a dialog box to manage the list. If we choose Two Stage Least Squares, and use CONSTANT and SCHOOL as the Explanatory Variables and CONSTANT, EXPER and MALE as the Instruments, we generate the instructions: instruments constant exper male linreg(inst) wage # constant school

The instruments are listed on the separate INSTRUMENTS instruction line. That allows you to do several LINREGs off the same list without having to re-enter the instrument information. Two-stage least squares is done by adding the INST (short for INSTRUMENT) option to LINREG. The next two choices for the Method box are AR(1) Regression and AR(1)-Instrumental Variables. These will generate a different instruction: AR1, which is for linear regressions with first-order autocorrelated errors. See Section 2.4 of the Users Guide for more on handling autocorrelated errors. We wont look further at this here, but would like to point out that parts of the dialog change when you choose this (and all the remaining methods). In particular, the Robust (HAC) Standard Errors is replaced by a different popup box for the AR1 Method. The next choice down on the Method box is Robust Regression, which generates an RREG (Robust REGression) instruction. This is still a linear regression, but uses a different estimation technique which is more robust to outliers. See Section 2.12 in the Users Guide if youre interested. The final choice is Stepwise Regression, which generates an STWISE instruction. Stepwise regression is nowhere near as popular as it was perhaps thirty years ago, but still has its place. If youre interested in this, see the description of STWISE in the Reference Manual.

Introduction to RATS

Int77

Getting StartedA Tutorial

Other Related Instructions
There are several other instructions which have similarities in form to LINREG, in the sense that the instruction looks like: instruction(options) dependent variable # list of explanatory variables start end

but have option sets that are so different that they cant easily be fit into the same wizard. These are the next several choices on the Statistics menu: The Limited/Discrete Dependent Variables wizard handles the DDV (Discrete Dependent Variables) and LDV (Limited Dependent Variables) instructions. DDV is used for techniques applied to data where the dependent variable is discrete (usually just values 0 and 1) such as logit and probit. LDV is used for techniques such as tobit which are applied when the dependent variable is continuous, but the observed range is somehow limited. These are covered in Chapter 12 of the Users Guide. The Panel Data Regressions wizard generates a PREG instruction, which includes special techniques for estimating the linear regression with panel data (often also known as longitudinal or cross section-time series data); techniques such as fixed effects and random effects. See Section 12.5 in the Users Guide for more. The Recursive Least Squares wizard generates an RLS instruction. When applied over the same range as a LINREG, it will give exactly the same outputthe point is that it does a whole set of calculations for reduced ranges that can be used to check for stability in the overall linear model. See Section 2.6 of the Users Guide. There are quite a few other estimation instructions, but they have quite a different form. The instructions included here have models which are based upon a linear combination of a set of explanatory variables. If we need a non-linear function, we will need a different way to enter that than just a list of variables. Well see one of these other instructions (NLLS, for non-linear least squares) in Section 1.7. There are several other instructions which apply to systems of linear equations rather than just single equations: SUR (Section 2.7 of the Users Guide), which does fairly general systems, and ESTIMATE, which estimates the equations of a Vector AutoRegression (var)the subject of Chapter 7 in the Users Guide.

Regression Format

Regression format is the manner which rats uses to list regressors, either on the supplementary card for LINREG and related instructions, or on a list of instruments. It consists of a list of series and lag fields. You create a lag field by appending to the series name the list of desired lags in braces ({}). You can use the notation {n1 TO n2} to represent all lags between n1 and n2. Leads are represented as negative lags. Examples of lag fields (using data from an earlier example with time series data) are: M1{0 TO 4 8} RATES{1} SEASON{-1} M1 lags 0,1,2,3,4, and 8 RATES lagged once One period lead of SEASON

Int78

Introduction to RATS

Getting StartedA Tutorial 1.6.7 Learn More: Error Messages

In a perfect world, we wouldnt need to talk about errors or error messages. Unfortunately, you will undoubtedly encounter an occasional error message in your work with rats. In this section, we discuss how to interpret error messages and fix errors.

RATS Errors and Error Messages

Whenever rats detects an error, it displays an error message describing the problem. Where possible, these messages provide you with specific information about the error. Each error message begins with an error code: a one-, two-, or three-letter code indicating the general type of error, plus a numeric code identifying the specific error. For example, general syntax errors begin with the code SX, syntax errors involving options have the code OP (for OPtion), and errors involving linear regressions begin with the code REG. See Chapter 6 in the Additional Topics PDF for a complete listing of all the rats error messages, along with some suggested causes and remedies.

Syntax Errors

Syntax errors probably appear most often. These occur while rats is reading a line (as opposed to errors which occur as rats is trying to execute a line it has read successfully). Syntax errors generally mean that rats couldnt make sense of the line it was processing. Missing parentheses, misspelled instruction or option names, illegal characters, and incorrectly entered numbers are some common causes of syntax errors.
rats displays two lines of information: the error message, and then the portion of the line which caused the error, enclosed in >>>> and <<<< symbols. rats only

prints the line up to the point at which the error occurred. It will display a maximum of 20 characters, counting back from the point of the error. For example, suppose you tried the following: set trendsq 1 50=t^2 This should have at least one blank space between the end parameter (50) and the =. Because that space is missing, rats will produce the following error message: Error SX15. Trying to Store Into Constant or Expression >>>>set trendsq 1 50=t<<<< As you can see, rats stopped processing the line immediately after it detected the illegal assignment operation.
rats often needs to look one character ahead to identify particular operators. For

instance, if it sees an =, it must check the next character to see if the operator is ==. Thus, the true culprit may be one character before the one shown.

Introduction to RATS

Int79

Getting StartedA Tutorial

If the reason for a syntax error isnt immediately obvious, check the entire line, not just the point at which rats stopped reading the line. rats can often interpret the actual mistake in a way that allows it to go on reading a few more characters from the line. For instance, data (unit=input) / gnpdata produces the error Error SX11. Identifier INPUT Is Not Recognizable >>>>data (unit=input)<<<<
rats reads all the way to the right parenthesis before it generates the error, but the

problem is actually the (illegal) blank space between DATA and the left parenthesis that starts the option field. This blank causes rats to interpret (UNIT=INPUT) as the start parameter, rather than as the option field. It is legal to use an expression like this for start. However, INPUT is not a recognizable variable name, so rats produces the error message.

Other Types of Errors

Once rats has successfully interpreted the line, it does not echo back the characters on the line. All you will see is the error message.

Show Last Error

If you are just typing in instructions one at a time, it will be clear which instruction caused the problem. Find it, fix it and re-execute the corrected line (hit <Enter>). However, if you select a block of lines and execute, the problem could be with any of the selected lines. If its a syntax error, you might be able to figure out where the problem is by matching up the echoed characters, but that wont help if its an execution error. In that case, try the Show Last Error operation on the Edit menu, which will position the cursor on the instruction line that generated the error.

Int80

Introduction to RATS

Getting StartedA Tutorial 1.6.8 Learn More: Entry Ranges

The SMPL Instruction
The SMPL instruction lets you control the default range. It is an important instruction in tsplike programs, but is less so in rats because: You can set explicit entry ranges on individual instructions where necessary. You can use default ranges on most transformation and regression instructions. SMPL is useful in situations where you want to run a sequence of regressions, forecasts, or other operations over a common fixed interval (other than the default range). For instance, suppose you have data from 1922:1 through 1941:1, but you want to run two regressions over the period 1923:1 to 1935:1. smpl 1923:1 1935:1 linreg foodcons # constant dispinc trend linreg foodprod # constant avgprices

Once you set a SMPL, rats uses the SMPL range as the default. To clear a SMPL, just issue a SMPL instruction with no parameters. We recommend that you do any preliminary transformations before you set a SMPL. If you need to skip entries in the middle of a data set, use the SMPL option (pages Int67 and Int82).

Using Entry Numbers Instead of Dates

calendar(m) 1959:1 the instruction linreg rate 1 24 # constant ip m1diff ppisum is equivalent to: linreg rate 1959:1 1960:12 # constant ip m1diff ppisum

You can use hard-coded entry numbers rather than dates. For example, given a CALENDAR of

because 1960:12 is the twenty-fourth entry given the CALENDAR setting. With time series data, you will usually use the dates, but using entry numbers is sometimes easier.

Introduction to RATS

Int81

Getting StartedA Tutorial

Dates are actually handled as integer-valued variables, which means that you can combine dates and integer entry numbers in an expression. For example, another way to do the regression above is: compute start = 1959:1 compute end = start+23 linreg rate start end # constant ip m1diff ppisum This means that you need to be careful about using the proper format for dates when using them. Suppose accidentally you left off the :1, and wrote: linreg rate 1959 1960 # constant ip m1diff ppisum

rats would try to run the regressions using entries one thousand nine hundred and

fifty nine and one thousand nine hundred and sixty. Clearly not what you intended, so be sure to include the :period anytime you are referring to a date.

Exceeding the Default End Period

The default end period is not a binding constraintyou can define series beyond this default limit as needed by using explicit date/entry ranges.

Selecting Subsamples: The SMPL Option

rats provides two related methods for dropping observations out of the middle of a

sample. We introduced the SMPL option available on many instructions on page Int67. It allows you to include and omit selected observations within the start to end range of the instruction. The formal description is: smpl=SMPL series or formula The SMPL series is a series or formula with non-zero (or logical true) values at the entries (between start and end) you want to include in the estimation, and zero values (false) at the entries to omit. It can be an existing series, or a formula like that used in a SET instruction. Its usually a dummy variable series of some form. It may be a dummy in the data set, or one constructed for this particular purpose. You can also set a SMPL series which applies to all instructions that support the SMPL option. You do this with smpl(series=SMPL series) We prefer to use the SMPL options in most situations. If you use the SMPL instruction, you must remember to reset it when you are done with the analysis which uses the reduced sample.

Int82

Introduction to RATS

Getting StartedA Tutorial

Missing Values
rats will automatically omit observations which have a missing value in the depen-

dent variable or any of the regressors, so you dont need a SMPL option for them. If you have a regression with lags, you could end up losing several data points since any observation which needs an unavailable lag of a variable will be dropped.

Skipping a Time Period

Suppose you want to leave the years 1942 to 1946 out of a regression over 1919 to 2009. You could construct a SMPL series for this with: set notwar 1919:1 2009:1 = 1.0 Set all entries to one set notwar 1942:1 1946:1 = 0.0 Set entries for war years to zero linreg(smpl=notwar) depvar # regressors The second SET only applies to the restricted range from 1942:1 to 1946:1, and so wont affect the values at other time periods. You could do the same thing with linreg(smpl=t<1942:1.or.t>1946:1) depvar # regressors but the first way is easier to understand at a glance.

Subsample Based Upon Value

To run a regression for just those entries where a series exceeds some value or meets some similar criterion, use relational operators. This generates dummies for three subsamples based upon the series POP, then runs the regressions over those. set small = pop<2000 set medium = pop>=2000.and.pop<=6000 set large = pop>6000 linreg(smpl=small) depvar # regressors linreg(smpl=medium) depvar # regressors linreg(smpl=large) depvar # regressors linreg(smpl=pop<2000) depvar # regressors

Again, we could collapse these by using the SMPL option into three instructions like

again, however, at a loss in readability. A few extra keystrokes is generally worth it if is produces a clearer program.

Introduction to RATS

Int83

Getting StartedA Tutorial 1.7 Example Six: Non-Linear Estimation

(6) Ct = + Yt + t Our final example (file ExampleSix.RPF) is taken from Greene (2008), Section 11.3.1. This estimates a model of consumption of the form:

where C is (real) consumption, Y is (real) disposable personal income and a, b and g are unknown parameters. The data file is ExampleSix.RAT, which is a rats data format file. If you open this (choose File-Open, pick RATS Data Files in the file type box, and select the file), you will see the following RATS Data File Window:

If you think that it looks a lot like a Series Window, youre right. They are formatted the same way, and share most of the same operations on the View menu and toolbars. Of the five series on the file, we need only two: REALCONS and REALDPI. Select those two (click on one of the series, then <Ctrl>+click or <Command>+click on the second), and choose the Data/GraphicsData (RATS Format) menu operation. This will pop up the following dialog box:

Youll notice that this is much simpler than the Data Wizard for the other formats. What this wizard will do is to look at the series that you selected, and guess that you want the coarsest frequency and the maximum common range. For instance, if we had both monthly and quarterly data, with some series starting in 1947 and some

Int84

Introduction to RATS

Getting StartedA Tutorial

in 1954, the guess would be quarterly data starting in 1954. If the common range is what you want, you can just ok the dialog. In this case, all the data are quarterly, and start and end at the same dates, so the only question is whether we want the whole range. If we wanted just 1960:1 to 1995:4, we could change the start and end dates in the dialog to request that range. We do want the full range, so we get open data examplesix.rat calendar(q) 1950:1 data(format=rats) 1950:01 2000:04 realcons realdpi If g were known, we could create a transformed variable and estimate a and b using LINREG; for instance, if it were 1.2, that would be done with: set ypower = realdpi^1.2 linreg realcons # constant ypower

Since its unknown, we cant, and have to estimate this using the instruction NLLS (Non-Linear Least Squares). Estimating linear models with least squares and related techniques is a fairly simple process, involving straightforward matrix computations. Non-linear estimation, however, is an inherently more general and complex taskthe variety of different models you can estimate is virtually limitless, and fitting these models requires the application of complex iterative optimization algorithms, rather than simple computations. The estimation process can demand a fair bit of expertise and judgment on the part of the user, because the algorithms can be sensitive to initial conditions and are prone to converging to local, rather than global, optima. Thus, it should be no surprise that fitting non-linear models in rats requires more effort and attention than fitting simple ols models. There are several steps we must do before we can even estimate the model. First off, we need to decide what we will call the free parameters; we cant use actual Greek characters, but we have to choose instead some legal rats variable names. Some other programs force you to use a specific way of writing this (perhaps c(1)+c(2)*realdpi^c(3)). rats allows you to use symbolic names, which is much simpler, particularly if you ever decide to change the formula and add or remove some of the parameters. alpha, beta and gamma are an obvious choice here, but well choose the shorter a, b and g. If we were doing a SET instruction to calculate the right-hand side formula, we could write that as a+b*realdpi^g, which is, in fact, the way that we will write this. We need two instructions to tell rats what the free parameters will be called and to define the right-hand side formula. These, respectively, are NONLIN and FRML. You might find it easiest to put these in manually, but you can also apply the Equation/ FRML Definition Wizard from the Statistics menu:

Introduction to RATS

Int85

Getting StartedA Tutorial

Equation/FRML Definition Wizard

Choose FRML (General) from the Create popup menu. Choose the name that you want to assign to this formula in the Equation/FRML Name box; well use CFRML. Choose the dependent variable (REALCONS) in the Dependent Variable popup. Put in the names A B G in the Free Parameters box and the formula A+B*REALDPI^G in the Formula box. As you can see, since you have to type in both the formula and the parameter names, there isnt all that much that the wizard really does other than help you make sure you get all the information in. For instance, if you just put in the formula definition without providing the A B G names, you will get an error message that A is unrecognizable when you try to ok the dialog. If you put everything in correctly, the wizard will produce: nonlin a b g frml cfrml realcons = a+b*realdpi^g

The NONLIN instruction tells rats that on non-linear estimation instructions that follow, the free parameters will be A, B and G. The FRML instruction (FRML short for FoRMuLa) defines the expression that well be usingCFRML will be a special data type also known as a FRML. You dont need to include the dependent variable when defining a FRML; if you dont have one, just leave that part of the instruction out.

Guess Values

However, were still not quite ready. (As we said, non-linear estimation is quite a bit harder than linear). Non-linear least squares is an iterative process; Greene, in fact, shows how to do (roughly) what rats will do internally as a whole sequence of linear regressions. That sequence has to start somewhere, at what are known as the guess values or initial values. The default for rats is all zeros, which sometimes works fine. Zeros wont work here, because if B is zero, the value of G doesnt matter: G is said to be unidentified (more than one value for G, here all values, give the same result).

Int86

Introduction to RATS

Getting StartedA Tutorial

An obvious choice to start are least squares results when G is one, which would just be a LINREG of REALCONS on CONSTANT and REALDPI. We can set this up with linreg realcons # constant realdpi compute a=%beta(1),b=%beta(2),g=1.0

%BETA is a vector defined by rats which has the coefficients from the last estimation instruction. %BETA(1) is the first coefficient and %BETA(2) is the second. We use those in a COMPUTE instruction to put our guess values into the three parameters.

The NLLS Instruction

nlls(frml=cfrml)

Finally, were ready. The model is estimated with

Because of all the work required to prepare everything, the instruction itself is quite simple. NLLS has quite a few options, but most of them are to control the estimation process itself if the model proves hard to fit. The output is shown here:
Nonlinear Least Squares - Estimation by Gauss-Newton Convergence in 65 Iterations. Final criterion was 0.0000003 <= 0.0000100 Dependent Variable REALCONS Quarterly Data From 1950:01 To 2000:04 Usable Observations 204 Degrees of Freedom 201 Centered R^2 0.9988339 R-Bar^2 0.9988223 Uncentered R^2 0.9997776 Mean of Dependent Variable 2999.4357843 Std Error of Dependent Variable 1459.7066917 Standard Error of Estimate 50.0945979 Sum of Squared Residuals 504403.21571 Regression F(2,201) 86081.2782 Significance Level of F 0.0000000 Log Likelihood -1086.3906 Durbin-Watson Statistic 0.2960 Variable Coeff Std Error T-Stat Signif ******************************************************************************** 1. A 458.79905447 22.50138682 20.38981 0.00000000 2. B 0.10085209 0.01091037 9.24369 0.00000000 3. G 1.24482749 0.01205479 103.26415 0.00000000

Its not that much different from LINREG output. The only addition is the line describing the number of iterations. You would like this to indicate that the estimation has converged. If it doesnt, you will see in its place something like:
NO CONVERGENCE IN 30 ITERATIONS LAST CRITERION WAS 0.0273883

Introduction to RATS

Int87

Getting StartedA Tutorial

When you get this message, or something like it, pay attention; its telling you that something seems to be wrong. We get quite a few technical support questions where people are trying to interpret results (the bottom part of the output) while ignoring the message that the estimates might be wrong. This turns out to be a simple one to correct; we generated the warning message above by adding to NLLS the option ITERS=30 (which is less than the default of 100 iterations). Just taking off the restricted number of iterations (by omitting the ITERS option) gives us the proper result. In many cases, though, you have quite a bit more work to do to correct the reported problem. If you do non-linear estimation, you need to read (carefully) through the first few sections in Chapter 4 of the Users Guide.

Fitted Values

We introduced the PRJ instruction for computing fitted values on page Int71. PRJ also works for many other types of estimations. However, it doesnt work after NLLS, or other instructions that arent based upon a linear specification. Instead, you can apply your defined FRML in a SET instruction: set fitted = cfrml A FRML can be used within the SET expression (or, for that matter, in another FRML definition) almost exactly like a series, that is, CFRML by itself means the value at the current entry, CFRML{1} means the value the previous entry, etc. However, these types of expressions are the only places where you can use a FRML like a series. You cant use them for the dependent variable in a regression, or in the regressor list, or many other places. If you want to see the value produced by a FRML, do a SET as shown, and then examine the series created.

Int88

Introduction to RATS

Getting StartedA Tutorial 1.7.1 Learn More: Non-Linear Estimation Instructions

Many of the most important (and complicated) instructions in rats do non-linear estimation. See Chapter 4 of the Users Guide on non-linear estimation techniques, which describes the various optimization algorithms, methods of creating and maintaining non-linear parameter sets and formulas, and includes descriptions of several of the basic instructions. The following are similar to NLLS in that they require a NONLIN instruction and one (or more) FRMLs to be defined before you can use them. Both are covered in Chapter 4 of the Users Guide. MAXIMIZE does maximum-likelihood estimation, applicable to a very large variety of models. NLSYSTEM estimates systems of non-linear equations, rather than just the one equation that NLLS handles. The remaining instructions have either entire chapters or at least sizable sections in the Users Guide (all references are to that book) devoted to their use. GARCH (Chapter 9) estimates univariate and multivariate arch and garch models. DLM (Chapter 10 ) estimates and analyzes state-space models. NPREG (Section 13.1) estimates non-parametric regressions. NNLEARN and NNTEST (Section 13.3) fit neural networks. CVMODEL (Section 7.5) estimates covariance matrix models for structural vars. LQPROG (Section 13.2) solves linear and quadratic programming problems. FIND (Section 4.13) can be used for just about any other type of optimization problem. If the function to be optimized can be written out using rats instructions, FIND can be used to optimize it.

Introduction to RATS

Int89

Getting StartedA Tutorial 1.7.2 Learn More: Working with Matrices

Matrices (or arrayswe use the terms interchangeably) are very useful in rats. For example, on page Int87 we showed how to use the %BETA vector to access estimated coefficients. %BETA is just one of many arrays defined by various rats instructions. For example, LINREG and many others also store the estimated standard errors and t-statistics in the vectors %STDERRS and %TSTATS, respectively. Arrays like these can be useful in reporting results, and for doing further computations, such as the Hausman test (Users Guide, page UG99).

DISPLAY and COMPUTE

By now you should be familiar with DISPLAY and COMPUTE, and wont be surprised to learn they work with arrays as well as scalars. For example, after doing the NLLS instruction in our example, you could (re) display the estimated coefficients and standard errors (with fewer decimals) by doing: display *.### %beta display *.### %stderrs

COMPUTE is the primary instruction for doing matrix computations. rats supports the standard arithmetic operators as well as several special matrix-specific operators. The program also provides dozens of built-in functions (such as inverse, transpose, and so on) for doing matrix calculations. See Section 1.7 of the Users Guide for more.

Defining Arrays

You can use DECLARE to create your own arrays, although you can often skip the declaration step. COMPUTE, READ, or INPUT can store values into arrays. If you need to construct an array from a set of series, you can do that using the MAKE instruction. You can define arrays of labels, strings, series, equations, even arrays of arrays, and so on. These structures can be very useful in writing more sophisticated code, particularly for implementing complex, repetitive tasks. In Chapter 3 we make extensive use of arrays of strings for labeling graphs, while instructions like FORECAST use arrays of series to store results.

Other Types of Matrices

In-Line Matrix Notation

You can use in-line matrixnotation (Users Guide, Section 1.7) to provide the values of an array in an expression. Use two vertical lines (||) to mark the start and end of an array, and a single bar (|) for a row break. This is particularly useful for defining a matrix in the middle of an expression, such as in a non-linear formula. You can also use in-line matrix notation for options that accept an array for the arguments. For example, this estimates a Box-Jenkins model with ar terms on lags 1, 4, and 12: boxjenk(ar=||1,4,12||) y

Int90

Introduction to RATS

Getting StartedA Tutorial 1.8 The RATS Programming Language

You can do quite a bit with the combination of the built-in instructions that weve introduced here, like LINREG and NLLS, plus the many others like BOXJENK, GARCH and MAXIMIZE that weve only mentioned in passing. But the real power of rats comes from combining the sophisticated and carefully tested calculations of instructions like those with an extensive programming language. Thats what makes possible everything from the hundreds of procedures that are included with rats (Section 1.4.5), to the highly interactive cointegration program cats, which is entirely written in the rats language. While you may never get to the point of writing your own procedures, you will almost certainly need to use some of the more basic programming features. The DO and DOFOR loops are relatively simple, and make many tasks much easier. Most of the early chapters in the Users Guide assume that you can use these, so well introduce them here. See Chapter 15 of the Users Guide for more on the programming features.

Loops
DO

rats offers five different looping instructions:

is a standard loop instruction which loops over an integer index. It is like a fortran do or a basic for. is less standard but often very useful. It loops over a list of items: series, matrices, numbers, etc. loops as long as an expression is true. loops until an expression is true. It always executes the enclosed instructions at least once (WHILE may not). loops unconditionally. You need to use a BREAK instruction to get out of the loop.

DOFOR WHILE UNTIL LOOP

A DO loop has the following structure:

do index = startvalue,endvalue,increment instructions executed for each value taken by index end do The index is a simple variable name, by convention, short names like I, J and K are often used. The increment is 1 by default, and the ,1 can be left out if it is. A DOFOR loop repeatedly executes all of the instructions down to the matching END, with index taking a new value from the list of values on each pass through the loop. This continues until the list is exhausted. While the index of a DO loop must be an integer variable, a DOFOR index can be almost any type of variable. dofor index = list of values instructions executed for each value taken by index end dofor

Introduction to RATS

Int91

Getting StartedA Tutorial

Examples
do time=1990:1,1992:12 linreg y time time+47 # constant y{1} set c1 time time = %beta(1) set c2 time time = %beta(2) end do This does a rolling regression, using 48 data points for each regression. The first regression starts in 1990:1, the second in 1990:2, and so on. Dates in rats are handled as integer entry numbers, so you can use dates in integer computations and in DO loops. On each trip through the loop, the SET instructions copy estimated coefficient values from %BETA (set by LINREG) into entry TIME of the series C1 and C2. When the loop is complete, these series will contain all 48 pairs of coefficient estimates. dofor i = realgnp to realbfi set(scratch) i = log(i{0}/i{1}) end dofor This does a percent change calculation for series from REALGNP to REALBFI. We use I{0} rather than just I so that I is treated as a series, not an integer value.

Conditional Execution

The IF instruction allows you to control the flow of program execution based on the condition of a logical expression. For example, in the following, Condition is true is displayed only if the variable TEST contains the value 1. if test==1 display "Condition is true" else display "Condition is false"

Important note: The IF instruction only evaluates a single conditional expressionit does not include any sort of implicit looping over series or matrix entries. If you want to set values of series or an array based on a condition, you probably want to use the %IF() function, rather than using IF inside a loop. See page Int63 for details. IF-ELSE blocks are usually part of a larger programming structure, like a DO loop or a PROCEDURE. On the rare occasion that you need a stand-alone IF or IF-ELSE, enclose the whole set of lines in { }; for instance, { if test==1 display "Condition is true" else display "Condition is false" }

This lets rats know the limits of the programming code (what we call a compiled section) that needs to be treated as a unit.

Int92

Introduction to RATS

Getting StartedA Tutorial 1.9 What Next?

There is almost no limit to what you can do with rats. It is only a question of learning how. You may need more information on the instructions weve already covered. Or, you may be interested in other types of analysis. With that in mind, we would like to offer some suggestions on how to find out what you need to know. Yes, there are a lot of instructions, and the manuals are large. Dont worry! rats is a very powerful and general program, and you may only need to use a fraction of its capabilities. Some suggestions for learning about rats: First, look through the rest of this Introduction. Youll learn more about data handling and graphics, two things you will likely be using almost any time you work with the program. Take a look at the Reference Manual to get a feel for how it is organized and how details on each instruction are presented. Look up some instructions youve already seen, like LINREG. When you see something new in an example, use the Reference to find out exactly what is going on. Check out the Users Guide for in-depth information on various types of analysis. You can often jump straight to the topic of interest; if any of the previous chapters are needed to understand whats happening, well let you know that right away. Use the Tables of Contents and the Index (theres a combined index at the end of the Users Guide)! They are probably the quickest way to find the information you want. See the Procedures and Examples pdf file included with the program for a complete list of the many examples and procedures included with rats. You can find even more examples and procedures on our website. Appendix C of the Reference Manual lists all the rats instructions, grouped by their general function This is a quick way to find instructions you may need for a particular task. You can use the help facility as a quick reference for the interface and all of the instructions.Most important, dont be afraid to try something.

Introduction to RATS

Int93

2.. Dealing.With.Data

lmost every time that you use rats, you will need to bring data into the program. Since collecting and organizing your data set may be the single most timeconsuming part of your analysis, you would obviously like rats to accept your data with as little fuss as possible. We have provided you with a wide range of options for handling this important step. You should note that rats is not a data set-oriented program. Simply selecting a data file does not make the contents immediately available for analysis: you must use the DATA instruction or a Data Wizard to bring the data into working memory. This may seem to be an unnecessary extra step, but it offers numerous advantages. For example, you have the option of reading in only certain series and certain entries from the data set. Converting data from one frequency to another can be handled automatically as you read in the data.
rats supports reading and writing data from a wide range of sources and in many

different formats. The most popular are RATS format and Excel spreadsheets. The Professional version of rats supports sql commands and odbc connections for reading many types of databases. Estima provides (for purchase) two sets of macroeconomic data. The oecd Main Economic Indicators has key data series for a large set of major countriesfor more information see http://www.estima.com/datainfo_oecd.shtml. These are provided in a set of rats format data files with one for each country. The Haver usecon data (http://www.estima.com/datainfo_haver.shtml) is an extensive set of us data. These are provided in both rats format and in dlx format. The RATS format version can be used on any platform, while the dlx (which we call FORMAT=HAVER) can be used on Windows only.

DATA, COPY and Related Instructions Converting Frequencies Missing Values RATS Format Excel and Other Spreadsheet Formats Database Formats Free and FORTRAN Formats Tips and Tricks

Data 2.1 The Tools

Below, we discuss the instructions you can use to read data from a file into a rats session and to write data from rats out to a file. If you are new to rats, please see the tutorial in Chapter 1 for additional information.

Getting Data In: The DATA Instruction and Data Wizards

Most of your projects in rats will require reading data in from a file or files. The critical instruction for this is DATA, which reads data from a file into one or more series variables stored in memory. You can write out the appropriate OPEN DATA and DATA instructions directly, or you can use one of the Data Wizards (on the Data/Graphics menu) to generate the necessary instructions. The FORMAT option on DATA tells rats what type of file you will be reading. Depending on the type of file, you may also need to use the ORGANIZATION option to specify whether the data series on the file are arranged in rows or columns. Note that the default format option for reading data is simple free-format text. rats does not attempt to determine the format of the data file itself based on the filename extensionyou need to set the FORMAT option yourself.

Getting Data Out: The COPY Instruction and FileExport...

The companions instructions to OPEN DATA and DATA are OPEN COPY and COPY, which write data series to a file. COPY can produce data files in a wide range of formats, including spreadsheets, rats format files, text files, and binary files. For rats format files, COPY can only create a new file. To write data to an existing rats file, use the instructions DEDIT, STORE, and SAVE (page Int110). If you prefer to use menu operations to write the data, you can: 1) Use ViewSeries Window to list the series available in memory. 2) Select (highlight) the series you want to write to a file from that list. 3) Do FileExport..., select the desired file format from the drop-down list, enter a name for the file, and click OK. You can also open or create a rats format file (using FileNew or FileOpen) and then drag and drop series from the Series Window onto the rats file window.

Supported Formats

We list the formats in Section 2.2, divided into three broad categories. You can find detailed information on each format (indicating where it can be used and any special details) in the Additional Topics pdf.

Int96

Introduction to RATS

Data
Other Instructions for Reading and Writing Data
DATA and COPY can only read and write SERIES type variables. For other types of variables (such as scalar reals, integers, arrays, labels, strings), you can use the INPUT or READ instructions to read data into rats, and the WRITE or DISPLAY instructions to write data to a file. The REPORT instruction can be useful for creating specialized reports of data and statistics, similar to those created by rats instructions like LINREG or STATISTICS.
rats allows you to write data in a wide range of forms using COPY. You can also open or create rats format files using FileNew or FileOpen, and move data to and from memory one rats format file to another (or from the working data series in memory to the rats file) by dragging and dropping series from one window to another.

Moving Data Around

In fact, with rats format used as an intermediate stop, you can easily translate data from, for instance, a poorly formatted text file to a wellorganized spreadsheet. The rats file format is directly portable across different platforms (Mac, Windows, unix) with no translation required. You simply need to be able to transfer binary data. XLS, XLSX, WKS, PRN, DBF, DTA, DIF and the rats Portable format are some other machine-independent formats which rats supports.

Double-Check Your Data!

It is very important to be sure that data are being read in properly, particularly when reading from a data file for the first time. The tutorial in Chapter 1 includes a number of suggestions for doing this, particularly using quick views of the sample statistics. In a rats program file, you can use the TABLE instruction immediately after the DATA instruction(s) to get the same type of quick statistics table.

Introduction to RATS

Int97

Data 2.2 Data/Copy Formats

rats can read or write data in around twenty different formats. Some of these are

strictly for input, some strictly for output, and some can be used for either. The details on each of these is provided in the Additional Topics PDF. The different formats generally fall into one of three broad categories.

Time Series Databases

Here, we look at file formats specifically designed to handle time series data. Data are organized by named series. Each of these series has its own calendar scheme and range of data. These are usually fairly complicated proprietary formats, since they have to be able to code up many types of date schemes, and must allow for very large amounts of data. The most important of these is rats format (FORMAT=RATS). The only other one which is supported in the standard release of rats is rats Portable Format, which is a text version of rats format, useful mainly for archiving data in a human-readable form. We use FORMAT=PORTABLE for describing this format. The other formats listed below are available only in the Professional version of RATS, and some arent included on certain platforms. These are all proprietary formats, so our ability to support them depends upon the type of support offered by the owner of the format. You will also need a subscription to a database or service provided by the owner in order to use them. These other formats are: Citibase/DRI/Global Insight native format. We call this FORMAT=CITIBASE. This is available on all platforms.
crsp (center for research in security prices) data, available from the University of

Chicagos Booth School of Business (www.crsp.com). We call this FORMAT=CRSP. Its available for Windows and some unix systems.

Fame is the native format for the fame database management program provided by SunGard. We call this FORMAT=FAME. This is available on Windows and certain unix systems. Haver dlx is the native format for Haver Analytics (www.haver.com). The Haver usecon data which can be purchased from Estima are provided in both rats format and in dlx format. We call this FORMAT=HAVER. Its available on Windows only.
fred (federal reserve economic data) is a (free) on-line database provided by the

St. Louis Federal Reserve Bank (research.stlouisfed.org/fred2). This is available on all platforms, but requires an active internet connection. You can read series in directly with FORMAT=FRED, but you also have the option of browsing the database interactivelyon the Data/Graphics menu, select Data Browsersfred(online) browser. This displays a list of the main database categories in a new window. Double-click on a category (or sub-category) to see a Data File Window with the series available in that category.

Int98

Introduction to RATS

Data
Labeled Tables and Spreadsheets
These form a rectangular table of data with each column representing a series and each row representing an observation. Each column is labeled with the series name. Each series has at least nominally the same range, and same date scheme (if any) though there might be missing values in some of the series. The most commonly used of these is one of the Excel formats, though they also can be in the third, less structured category. For formats through Excel 2003, use the option FORMAT=XLS. For Excel 2007, use FORMAT=XLSX. You can pull data off more than one worksheet within an Excel workbook (using the SHEET option on DATA), but can only access one worksheet per DATA instruction; use additional DATA instructions to get data from more than one worksheet.
rats accepts several delimited text formats. FORMAT=PRN will accept data fields

which are separated by commas or white space (spaces or tabs). For writing data, you can also use FORMAT=CDF (comma delimited format) or FORMAT=TSD (tab separated data) to get specific field separators. EViews workfiles fall into this category. rats can process data series (but not other objects). This uses FORMAT=WF1.
dta is the native format for Stata data files. Use FORMAT=DTA for that.

If you have the Professional level of rats, you can also read data using sql queries on databases which support odbc (such as Oracle and Access). This is a bit more complicated than other formats since you have to provide the sql query that sets up the table of interest. This is FORMAT=ODBC.
dif is a (rather bulky) text format for transmitting the content of a simple spread-

sheet in a text, rather than binary form. Its rarely used now, except as a copy-paste format for data to and from spreadsheets. It uses FORMAT=DIF.
wks is an older spreadsheet format used by the Lotus spreadsheets. Its perfectly

adequate as a storage format for data, and quite a few legacy data sets are saved in it. As with Excel, it can also be in the third category. This uses FORMAT=WKS.
dbf is the database format used by dBase and compatible programs. It uses

FORMAT=DBF.

Unlabeled Text and Spreadsheets

These either dont have labels for individual series, or have labels which have illegal characters for rats series names (such as spaces or parentheses). The most important of these is free-format, which is just a text file with numbers, delimited with commas or white space (spaces, tabs, line breaks). While we would never recommend saving your own data in this, there are a large number of existing data files done in such an unstructured way. Also, if you ever have to scan data out of a book, you will likely end up with this. This uses FORMAT=FREE.

Introduction to RATS

Int99

Data
Another unlabeled text format is the collection of fortran formats. This is likely only of value if you have a very old text data file which, in order to keep the size down, squeezed data into undelimited fields. (For instance, first five positions are the first number, second five are the next number).
matlab is the native data format for the matlab programming language. This

could also fall into the Time Series Databases category, if each series is a separate one-column matrix. If, instead, you have a single matrix with multiple columns forming your dataset, it falls into this group. This uses FORMAT=MATLAB. The least-recommended of all the formats is native binary. Never save data in this format; theres no identifying information on the file, you cant read the data without the proper program and it isnt necessarily portable from one type of computer system to another. Its FORMAT=BINARY. Any of the spreadsheet formats (XLS, XLSX, WKS, DIF, PRN) can also be treated this way if you have extra comment lines or missing or unusable labels.

Series Labels on the Files

The advantage of the Time Series Databases and the Labeled Tables is that they have series labels already on the file, so you can pick and choose which series you want. However, the name you use for a series in your program must be the same as the name for the series on the data file. That is, if M1 is called FMS on the data file, you will have to use the name FMS in your program also. The commercial databases need naming conventions to distinguish thousands of series, so their names can often be a bit cryptic. While the data must initially be read using the names from the file, you can (and probably should) always copy the information into more meaningful names, such as: set m1 = fms

Formats for Input Only

CITIBASE, CRSP, DTA, FRED, MATLAB, ODBC, WF1, XLSX

rats supports two types of specialized output formats for generating tables. These

Formats for Output Only

are FORMAT=TEX, which generates a TeX tabular environment, for inclusion in a TeX document and FORMAT=HTML, which generates an html table for inclusion on a web page.

Int100

Introduction to RATS

Data 2.3 Where Are Your Data Now?

The suggestions below are designed to help you figure out the best approach to getting your data read into rats so you can make use of them. Details on the various file formats supported by rats follow later in this chapter and in Additional Topics.

In a Time Series Database

Except for fred, these can all be read using the proper sequence of CALENDAR, OPEN DATA and DATA instructions. See Section 2.7 for more on the use of RATS format; otherwise, check the description of the format in the Additional Topics PDF. Again, note that most of these are available only for the Professional level of rats.
rats, fred, Fame and Haver dlx all have browser modes, which are Data File

Windows. You can use that to navigate through the contents of the file, using the ViewReset List menu item or Change Layout toolbar icon ( ) to restrict the list based upon frequency, start or end year, comments or names (or any combination). You can drag and drop series you select onto the Series Window, or onto a RATS Data File Window, or export them directly to some other type of file (such as a spreadsheet). You open the browser for a rats data file by opening it with the menu File-Open; the other browsers are submenus for Data/GraphicsData Browsers. If you are working with the browser for a rats file, you can also select the series you want and use the Data (RATS Format) Wizard. That will look at the ranges and frequencies of the series that you selected, and come up with a common data range.

In a Labeled Table

These are all included in the Data (Other Formats) Wizard, which we would recommend you use the first time you try to work with a particular file. The wizard will always read all the series from the file, which might not be what you want. However, you can always edit the DATA instruction produced by the wizard to eliminate the unwanted series. Save the edited program file, and use that as the basis for further work. See Section 2.8 and the next paragraph for more on the spreadsheet formats (XLS, XLSX, WKS, DIF, PRN). Because there are many ways to pull data out of these, when you use them as a Labeled Table, you need to include the option ORG=COLUMNS.

In a (Less Structured) Spreadsheet

Spreadsheets can often include additional information besides the data and labels (and possibly dates) that are required for treatment as a Labeled Table. If you have a well-structured labeled table within the spreadsheet, you probably can read it as a Labeled Table by using the LEFT and TOP options to restrict the attention of the DATA instruction to the area that you want. For instance, if you have ten lines of descriptive headers, you would add the option TOP=11. The Data (Other Formats) Wizard can help you set those options.

Introduction to RATS

Int101

Data
If you have a properly labeled table, except that the series are arranged by rows, rather than columns, you can use the option ORG=ROWS. You can combine that with the TOP and LEFT options if needed to isolate the data table from other information. There are several other options for reading data out of less-structured spreadsheets which are covered in Section 2.8. Note, however, that in many cases, the simplest approach is simply to pull the data into a spreadsheet program and edit out the unwanted parts of it.

In a Text File

What you should do depends a lot on the format that you have inherited and how big the data set is. If the data is a well-structured Labeled Table, you can just read it as described above. If the file is fairly well-organized in columns and rows, but without series names, you can probably still process the file using the Data (Other Formats) operation. In most cases, you can just select Text files (*.*) as the file type. Youll just have to provide labels for the series. If the data file is fairly stable, we recommend that you then write the data from rats into a rats format or spreadsheet format file for later use. See Section 2.9 for more tips on reading text files.

In a Database File

If the data are in a database you can access using sql commands, you should be able to read it with the Professional version of rats, using DATA with FORMAT=ODBC. See the description of the FORMAT=ODBC in the Additional Topics pdf. If you have quite a bit of data in printed form, youre probably best off scanning it to create a text file, and continuing as described under In a Text File. Since character recognition software is still imperfect, youll have to be extra careful in checking that the data came in properly. If you need to key in your data, you can:

On Paper

type the data into a spreadsheet, a database application, or a text file (using rats or any other text editor) and read the data using the steps described in the relevant section above. Use the DataCreate Series operation and type the data directly into rats memory, and then use FileExport.... Open or create a rats format file (using FileOpen or FileNew), and then use DataCreate Series to type the data directly into a rats format data file. The rats data editor used by the Create Series operation is designed to handle timeseries data sets easily, where you deal with one series at a time. If your data set is more case-oriented, a spreadsheet or data base editor is likely to be more convenient, since you can edit across series. Whatever method you use, be sure to check the data after reading it into rats.

Int102

Introduction to RATS

Data 2.4 The Data (Other Formats) Wizard

For most users, the bulk of your work will probably involve reading data from Excel spreadsheets or text files, and as demonstrated in Chapter 1, the Data Wizards (Other Formats) operation is usually the easiest way to go about this. Because youll use it often, lets take another look at a sample Data Wizard dialog:

To get to this point, weve done the following: Selected the FileData Wizard (Other Formats) operation, selected Excel 2007 Files (*.xlsx) from the list of file types, and opened the file HAVERSAMPLE.XLSX. Clicked on Show Preview to view the file, and used the Sheet button to preview the two sheets on this file. Here, weve selected the Monthly sheet. Clicked on Scan to have rats determine the frequency and dates of the data. Used the Set button to tell rats to read the data at a quarterly frequency, rather than at its native monthly frequency. Because weve set our Target Dates field to a lower frequency than the source data, the Compact by button is active. You can use this to choose the compaction method rats will use. Here, weve selected the default choicetaking simple averages of the subperiods (page Int104). If we didnt want the Wizard to generate a CALENDAR instruction at all (for example, if we had already set a quarterly CAL earlier in the session), we could just turn off the Reset Workspace Dates switch.

Introduction to RATS

Int103

Data 2.5 Changing Data Frequencies

The DATA instruction can automatically convert data from one frequency to another. For this to work:

the source file must contain dates that rats can process. The program cannot determine the frequency of the source data if the file does not contain dates. you must set the CALENDAR instruction to the desired (new) frequency. You can do this by typing in CALENDAR directly, by using the Data/GraphicsCalendar operation, or by setting the Target Dates field in the Data Wizard (page Int103). Given a mismatch between the frequency of the data on the file and the CALENDAR seasonal, rats will automatically compact or expand the data to match the CALENDAR setting. This works for any of the file formats for which rats can process dates (any of the Time Series Database formats and most of the Labeled Table formats).

Compacting Data

To compact from a higher frequency to a lower frequency, just follow these steps: 1) As noted above, make sure the source data file contains valid date information at the original (higher) frequency. See page Int106 for tips if you dont already have date information on a file. 2) If youre using a Data Wizard, just make sure that the File Dates field accurately reflects the higher frequency of the source data,. Then, set the Target Dates fields to the desired lower frequency. Click on ok to read the data. If you are typing in commands directly, set the CALENDAR to the target (lower) frequency. For example, if you are compacting monthly data to a quarterly frequency, specify a quarterly CALENDAR. Then use OPEN DATA and DATA to read the file.
rats will automatically compact the data to match the CALENDAR frequency using

the method specified by the COMPACT or SELECT option, or the Compact by field on the Data Wizard. The default is COMPACT=AVERAGE.

Compact and Select

The COMPACT and SELECT options allow you to select from several compaction methods. Note that the two options are mutually exclusive. If you try to use both, rats will honor the SELECT choice. The choices are: compact=[average]/sum/geometric/first/last/maximum/minimum AVERAGE SUM GEOMETRIC FIRST/LAST MAX/MIN simple average of the subperiods sum of the subperiods geometric average of the subperiods first or last entry of each subperiod, respectively maximum value or minimum value from each subperiod

Int104

Introduction to RATS

Data
select=subperiod to select compacts data by selecting a single subperiod from within each period. Suppose you have a quarterly CALENDAR and you want to read a monthly data series. DATA with SELECT=3 will read in the third month from each quarter (that is, the March, June, September, and December observations from each year). With the default option of COMPACT=AVERAGE, each quarterly value will be the average of the three months which make up the quarter. It is more complicated when you move between dissimilar frequencies: weekly to monthly, for instance. If you use SELECT=2, it will select the second full week within the month. Suppose you have the weekly observations shown below. Since weekly data are classified in rats according to the end of the week, March 13, and not March 6, will end a full week for this data set. If you compact data using the other methods, rats will give 6/7 of the March 6 value to March, and 1/7 to February. If the input data are: Week ending: March 6 15.0 March 13 11.0 March 20 13.0 March 27 18.0 April 3 20.0

the value for March (with COMPACT=AVERAGE) is: (6/7x15.0 + 11.0 + 13.0 + 18.0 + 4/7x20.0)/(6/7 + 3 + 4/7) = 14.97

Expanding Data to Higher Frequencies

The steps are the same as those described under Compacting Data on page Int104. If rats detects that the data on the file being read are at a lower frequency than the current CALENDAR setting or Data Wizard target frequency, it will automatically expand the data to the higher frequency by setting each subperiod equal to the value for the full period. For example, if you set a quarterly CALENDAR and read in annual data, rats will set the value of each quarter in a given year to the annual value for that year. For dissimilar frequencies, the value of a period which crosses two periods of the lower frequency data is a weighted average of the two values. For instance, in moving from monthly to weekly, the value for the week ending March 6 will be: 1/7 x February + 6/7 x March For a more complex interpolation, you would first read the data as described above to expand to the higher frequency, and then apply the desired interpolation routine to the expanded data. rats ships with several interpolation procedures you can use, but we primarily recommend the @DISAGGREGATE procedure (supplied on the file DISAGGREGATE.SRC), as it provides options for choosing from several models and techniques. The following example reads in quarterly gdp data at a monthly frequency. It uses DISAGGREGATE to produce a monthly series from the step function that DATA gener-

Introduction to RATS

Int105

Data
ates. Since gdp is ordinarily quoted in annual rates, we want to maintain the average at the higher frequency; if we had a series that quoted in totals (sales figures, for instance), we would use MAINTAIN=SUM. calendar(m) 1947 open data haversample.rat data(format=rats) 1947:1 2006:4 gdp set trend = t @disaggregate(factor=3,model=loglin,maintain=average) gdp / gdpm # trend

No Dates on Source File?

If you have a data set that you want to compact or expand, but your data file doesnt include date information, you can either

add dates to the original file directly (see page Int119 for suggestions if using a spreadsheet), or read the data into rats under its actual frequency, write the data out to a new file that does include dates, and then read the data back in from the new file, letting rats compact or expand as needed. For example, suppose the quarterly gdp data from the previous example was provided in a plain text file. You know that the data is quarterly, and starts in 1947Q1, but the text file doesnt contain any date information, so you cant expand it directly. Instead, you would first read the data into rats, using the appropriate quarterly CAL setting: calendar(q) 1947 open data gdpnodates.txt data(format=free,org=columns) 1947:1 2006:4 gdp Next, write the data out to a new file, using the DATES option to include dates (based on the CAL information) in the file. (You could also export the data from the Series Window). Here, well create a rats format file: open copy gdpwithdates.rat copy(for=rats,dates) close copy You can then process it as above, but reading from GDPWITHDATES.RAT: calendar(m) 1947 open data gdpwithdates.rat data(format=rats) 1947:1 2006:4 gdp set trend = t @disaggregate(factor=3,model=loglin,maintain=average) gdp / gdpm # trend

Int106

Introduction to RATS

Data 2.6 Missing Data

Coding
Internally, rats represents missing data with a special value. On most machines this is the value +infinity. On output, a missing value is denoted as NA for Not Available. In a rats expression, this value is available as %NA. Thus, set fixx = %if(x<0.0,%na,x) will set FIXX to be missing for entries where X<0 and to X otherwise. You can test whether a value is missing using the %VALID function: %VALID(x) is 0.0 if X is missing and 1.0 otherwise. You can also use X==%NA.
rats has special ways of handling missing values for each of the formats.

Spreadsheet Files

The +infinity coding is the same as is used by the major spreadsheet programs for N/As. Thus, if a cell contains the explicit missing value function (NA() in Excel), rats will read a missing value at that cell. When you export data to the spreadsheet, rats passes the code value (as a number, not a function), so it will display as NA or #N/A on the spreadsheet.
rats will also interpret blank cells within the spreadsheet as missing values.

Spreadsheet-Style Text (PRN) Files

Missing values are a problem if you have to save the data in a space or tab separated text format (PRN or TSD). You cant simply leave the cell blank in the text file, as you can with an actual spreadsheet, because a blank cell is indistinguishable from the blanks used to separate data items. Instead, youll have to use one of the codings described in the next paragraph.

Text Files

rats will accept the following as codes for missing values in text-format data files:

The characters NA or #N/A (upper or lower case) A decimal point, followed by zero, one, or two non-numeric characters (for example, . or .NA) Note, however, that you cannot use these within a rats expression: use %NA for that.
rats will interpret any block of non-numeric characters as a missing observation. However, if the characters dont fit the description above, rats will issue a warning

message that it has encountered invalid input.

Introduction to RATS

Int107

Data
Numeric Codes
If you have data which use a specific numeric value (999 or something similar) to indicate a missing value, use the option MISSING=missing value code on DATA. Whenever possible, use integers for missing value codes. The realities of numerical precision in the computing world mean that rats may not be able to match exactly a decimal-valued code such as 999.99. Suppose your data set uses 9999 for missing values. You could use something like the following: data(format=prn,org=columns,missing=-9999) / sales revenue This would read the series SALES and REVENUE, and convert any values of 9999.0 to missing values. If you have several missing value codes (such as 999 = no response, 998 = invalid response), you have two options: Edit the file and replace the two codes by a single code. Read the data and alter it within rats. This is an example of the latter: data(format=free,org=col) / income charitable mortgage set income = %if(income==-999.or.income==-998, %na, income) The %IF function tests whether the current entry of INCOME is equal to either of our two missing value numeric codes. If so, the function returns the %NA missing value code, which is then stored in INCOME. Otherwise, the existing value of INCOME is retained. To apply this procedure to multiple series, we can use a DOFOR loop: dofor i = income charitable mortgage set i = %if(i{0}==-999.or.i{0}==-998, %na, i{0}) end dofor

Note that the index variable I will actually be an integer variable containing the series number associated with the current series. As a result, we need to use the lag notation {0} (specifying the zero lag) on the right hand side of the SET instruction. This tells rats to treat I as a series, not as an integer number.

Skipped Dates

Suppose that you have a (daily) data set in which holidays and other non-trading days are skipped. If you have no date information on the file, there is little you can do: you will have to treat it as an irregular time-seriesomit the CALENDAR entirely, or use CAL(IRREGULAR).

Int108

Introduction to RATS

Data
If you do have dates on the file, DATA (and STORE with CONVERT) will code entries corresponding to the skipped dates as missing values. For example, consider the following portion of an XLS file, which skips the (U.S.) Thanksgiving holiday on November 23, 2000: SALES_DATA 2000:11:20 3590.50 2000:11:21 4256.05 2000:11:22 2987.23 2000:11:24 6799.87

You might read the data with the following instructions: calendar(d) 2000:1:2 open data sales.xls data(format=xls,org=col) 2000:1:2 2000:12:31 sales_data The data in the SALES_DATA series for the week of Thanksgiving would be: 2000:11:20 2000:11:21 2000:11:22 2000:11:23 2000:11:24 3590.50 4256.05 2987.23 NA 6799.87

If you really want to skip the holidays without having gaps in your data (in the sense that you want to treat the data for November 22 and November 24 as adjoining entries), you cannot treat the data set as Daily because rats insists that daily data be five days a week. If you use CALENDAR(IRREGULAR), rats will ignore the dates on the data file and read the data from the file into consecutive entries. There is little difference between the two ways of handling the holidays, unless you do some type of time-series analysis that uses lags or leads. Another option for these situations is to go ahead and read the data using a DAILY or SEVENDAY CALENDAR, as in the Thanksgiving example above, then use the SAMPLE instruction to create compressed versions of the data without the missing-values: sample(smpl=%valid(sales_data)) sales_data / sales_nomissing You can use the uncompressed data for printing or generating graphs (where you want to include date information), and use the compressed series for estimating Box Jenkins, garch, and other similar models that dont accept missing values. If you want to patch over a missing value like this with the previous value (which would make no sense with the sales data, but might if you have, for instance, price data), you could do that with something like: set patched = lastprice=%if(%valid(price),price,lastprice) This will use the value in the series PRICE if PRICE isnt missing, and will use the last non-missing value if it is.

Introduction to RATS

Int109

Data 2.7 RATS Format

rats format is designed for the special problems of time series data. It has many

advantages over other formats:

Data access is more flexible: you can set up a single file with many series and select only the ones you need for a particular application. You can also select a specific set of entriesyou dont have to read in entire series. The data series do not need to have identical structures. For example, you can mix monthly, quarterly and daily series on one file and can have series running over different intervals. rats can automatically compact or expand data to match the current CALENDAR frequency. See Section 2.4. There are many ways to add data to a file or edit series already on a file. Data retrieval is extremely fast.
rats also allows a certain amount of flexibility in reading the Labeled Table files

(page Int99): you can select specific series off a file and restrict yourself to a subset of entries, and, under certain circumstances, you can compact and expand.

Stored Information

For each series on the file, rats retains the following information: the series name the frequency of the data (annual, quarterly, monthly, panel, etc.) the dates (or entries) for which the series is available up to two lines of comments the data itself

Reading Data into RATS

To work with data from a rats format file, you can:

Open the file using OPEN DATA, and read the data using DATA(FORMAT=RATS), or Open the file using FileOpen, select (highlight) the series you want, and use the Data (RATS Format) wizard from the Data/Graphics menu to read in the data. Open the file using FileOpen and drag and drop series from the rats file onto the Series Window (displayed using ViewSeries Window). With rats format, you can read any combination of series from the file. You can also read specific sets of observations from the full data set by using the start and end parameters on the wizard or the DATA instruction. This is one of the most important advantages of rats format. If you omit the list of series when using DATA, rats will read all the series on the file. Be cautious when working with large files, or you might read in many unneeded series.

Int110

Introduction to RATS

Data
To change the name of the series as stored on the file, you can either open the file using File-Open and use the Rename toolbar button, or else open the file with DEDIT and use the RENAME instruction.

Instructions for Creating and Editing RATS Files

rats provides special instructions for creating and editing rats format data files. Never try to make changes to a rats data file using a text editor. The instructions for working with rats format files are:

COPY When used with the option FORMAT=RATS, this writes a set of series to a new rats format file. This is the easiest way to write data to a rats file, but cannot be used to append data to an existing fileuse DEDIT, STORE, and SAVE for that. DEDIT initiates the editing of a new or existing file. Once you have opened a file using DEDIT, you can use the following instructions. Several of these are rarely used now, since its easier to make changes on-screen using the RATS Data File Window. STORE SAVE CATALOG PRTDATA QUIT adds specified series to the file, or converts the contents of a data file of another format. saves changes to the file. produces a directory of series on the file. prints one or more of the data series stored on the file. ends an editing session without saving the changes (do SAVE then QUIT if you want to save changes and then close the file).

The following are no longer documented in the Reference Manual, but are included in the Additional Topics pdf. EDIT performs on-screen editing of an existing or new series. adds a single series to the file. makes changes to a small set of entries of an existing series. deletes a series from the file. renames an existing series.

INCLUDE UPDATE DELETE RENAME

Introduction to RATS

Int111

Data
RATS Data File Windows
When you open an existing rats file using FileOpen, or create a new file using File NewRATSData Window, rats displays the contents of the file in a window. From this window, you can:

Read data into memory by selecting the series you want from the data window and then doing Data (RATS Format) from the Data/Graphics menu. Read data into memory by dragging and dropping series from the rats file window to the Series Window (displayed using ViewSeries Window). Write data to the rats file by dragging series from the Series Window and dropping them on the rats Data File Window. Use FileImport to bring data from other files into the rats format file. Use FileExport to export data from the rats format file. Double-click on a series to view or edit the data in that series. Select Create Series from Data/Graphics (or click on the icon series on the file. Rename a series by selecting the series and clicking on the ) to create a new icon.

Use the ViewReset List menu item or toolbar icon to restrict the list based upon frequency, start or end year, comments or names (or any combination) Use the View menu operations or corresponding toolbar icons to display time series graphs, histograms, and box plots, generate a table of basic statistics, or compute a set of autocorrelations for the selected series. You can also right-click (or <Command>+Click on the Mac) on a series (or list of series) and select operations like Cut, Copy, and Export... from the pop-up menu.

Examples of DATA

Suppose the data file MYDATA.RAT has REAL_GDP, quarterly from 1947:1 to 2010:1, and TBILLS, monthly from 1946:1 to 2010:5. cal(q) 1947:1 Sets a quarterly CALENDAR open data mydata.rat data(format=rats) 1947:1 2010:1 real_gdp tbills

will read REAL_GNP and quarterly averages (the default compaction option) of TBILLS, over the range 1947:1 through 2010:1. cal(m) 1954:1 allocate 1999:12 open data mydata.rat data(format=rats) / tbills Sets a monthly CALENDAR

will read the (monthly) data for 1954 through 1999 for TBILLS.

Int112

Introduction to RATS

Data
Examples of Editing
dedit company.rat cal(q) 1950:1 open data ascii.dat data(format=free) 1950:1 2003:4 order invent shipment backlog store order invent shipment backlog save reads four series from a freeformat file and saves them on the rats file COMPANY. RAT. cal(d) 1991:1:2 dedit(new) options.rat open data options.wks store(convert=wks,org=col,adddates) save

converts all data from the (undated) wks file OPTIONS.WKS, creating OPTIONS.RAT. The data will be stored as daily data, beginning with entry 1991:1:2.

The RATSData Utility Program

rats ships with a stand-alone, menu-driven utility program called ratsdata. Most of the functionality of this program has now been incorporated directly into rats, us-

ing the New, Open, Import and Export operations on the File menu, and the various operations (and associated toolbar icons) on the View menu.

You may still find ratsdata useful for some tasks. Windows users will find shortcut icons for ratsdata in the Winrats folder on the Start menu and desktop. Mac users will have a ratsdata icon in their Macrats folder. Linux and unix users can run the ratsdata executable file. See the ratsdata help for details on using the program.

Using Older RATS Format Files

rats Version 8 uses the same rats file format as Versions 4 through 7, so files can

be interchanged seamlessly among these versions. It can also read from files created with Versions 2 and 3 of rats, which should cover any rats format file created since around 1987.

Introduction to RATS

Int113

Data 2.8 Spreadsheet and Delimited Text Formats

Supported Formats
Spreadsheets and comma, tab, or space delimited text files are popular tools for data storage, and many commercial database vendors allow users to download data in these formats. rats supports several formats associated with such programs, with these choices for the FORMAT option on DATA and COPY: FORMAT Option XLSX XLS WKS PRN CDF TSD DIF File Type Microsoft Excel 2007 xlsx spreadsheets Microsoft Excel (2003 and earlier) xls spreadsheets Lotus 123 worksheets (wk3, wks, wk1, wrk and wr1 files) Print Files (text files with a spreadsheetstyle layout) Commadelimited text files Tab and space delimited text Data Interchange Format

xls/xlsx are typically the best choices from among these. rats supports these formats in several ways:

You can read data from a file using the Data Wizard (Other Formats) operation, the instruction DATA, or by importing into a Series Window. You can write data to a file using the instruction COPY or by exporting from the Series Window.

You can convert data from a spreadsheet directly to a rats format file using the CONVERT option of STORE or by importing into a RATS Data File Window. You can read and write matrices using READ and WRITE. You can export data from Report Windows using the export operations.

As Labeled Tables

All of these can be treated either as Labeled Tables or as Unlabeled Data, depending upon the form that the information takes. To be handled as a Labeled Table, its necessary for the data to look something like the sample at the top of the next page. We refer to this as organization by column (series arranged in columns, one observation per row). The series can also run across the rows: dates (if present) in the first row and each row of data beginning with the series label in column A. This is organization by row or variable. In either case, the block of data should be in the upper left hand corner of the worksheet (although options like TOP and LEFT can be used to ignore data above and to the left of the data block).

Int114

Introduction to RATS

Data

Dates

The column (or row) of dates is optional. Dates permit more flexible use of the data. If your file includes dates, the date information must appear in the first column or row of the data block. You can enter the dates as strings or as functions or values. If you do enter them as functions or values, be sure to format the cells with a date format. If you use strings, the dates must be have some type of delimiter between the year and period, or year, month and day; for instance, 1999:1, 1999Q1 or 1999-3-31 are acceptable, 199901 isnt. If the dates arent yearDperiod or yearDmonthDday (D being some non-numeric delimiter), you need to use the DATEFORM option on DATA to provide the form being used. For instance, if youre using the notation qq-yyyy (011999, 02-1999, etc.), include the option DATEFORM="qq-yyyy". The Data Wizard will generally be able to properly detect any such non-standard coding for the dates. With annual data, you can use just the year number in the data file (2010 rather than 2010:1), but rats will only recognize this as a date if it is entered as a string rather than a numeric value. Remember that in your rats programs themselves, you must include the :1 for annual date references.

Labels

The labels at the beginning of each series are not optional if you want to treat the file as a Labeled Table. rats uses the series labels to determine which series to read from the file. Labels should normally appear in row 1 (if organized by columns) or column A (if organized by rows). If you have additional rows of header information before the row of series labels (on data arranged by column), you can use Header Rows before Dates/Labels field in the Data Wizard or the TOP option directly on DATA to skip those. Variable names must start with a letter, consist of letters, digits, _ , $ or %, and be sixteen characters or less in length (they can actually be longer, but only the first sixteen are significant). If you dont have the option of changing the names (for in-

Introduction to RATS

Int115

Data
stance), your best strategy is to skip the automatic label processing. This is done using the option NOLABELS on DATA, combined with the TOP option to skip the line with the (unusable) labels. For instance, suppose the sample file had labels of Investment (GE), Value (GE) and Capital (GE). Neither the spaces nor the parentheses are permitted in a rats variable name. Instead, we could read this with: open data sample.xls cal(q) 1999:1 data(format=xls,org=col,nolabels,top=2) 1999:1 2001:1 ige fge cge Since we skipped the label processing, we have to provide our own labels for the series, and we have to read all the series from the file.

Reading Data

You can read data from a spreadsheet or delimited text file using the Data (Other Formats) Wizard or by typing in a DATA instruction with the appropriate FORMAT and ORG options. For multi-sheet workbooks, you can use the SHEET option to tell rats which worksheet you want to read. By default, it will read data from the first worksheet on the workbooks list of sheets. To use the Wizard, select Data (Other Formats) from the Data/Graphics menu, specify the format of the file you want to read (such as Excel 2007 Files (*.XLSX)) from the drop-down list in the dialog box, select the file you want to read, and then use the Wizard dialog box to read the file. If the source file has more than one worksheet, this will have a popup box to allow you to select the worksheet you want. If you are typing the instructions manually, use CALENDAR or the Calendar operation to set the frequency and start date (if using dated data), then use OPEN DATA to specify the file to be read, followed by a DATA instruction with the appropriate options. You can list the specific series you want to read on the DATA instruction, or omit the list and let rats read all the series on the file. Whether you can select specific observations out of the file depends in part on whether or not the file includes a date column:

If there is no date column or row on the file, rats assumes the first observation on the file corresponds to the first observation you request on DATA. You can ask only for entries at the beginning of the file, never a block in the middle or end. If you need to be able to read in a subset of the data, you could first convert the file to rats format. If there are dates on the file, and the frequency on the file matches the current CALENDAR seasonal, DATA will locate the entries requested and fill in any skipped dates with missing values. If the frequency of the current CALENDAR does not match the frequency of the data on the file, rats will compact or expand the data as it reads it in. See Section 2.4. If you dont use a CALENDAR instruction, rats will read observations from the beginning of the file. Any dates on the file are ignored.

Int116

Introduction to RATS

Data
Examples
Make sure that you use the proper extension (XLSX, XLS, or PRN for example) on the file name. rats will not add it automatically. cal(q) 1999:1 open data sample.xls data(format=xls,org=col) 1999:1 1999:4 cge fge reads series CGE and FGE, for the dates 1999:1 through 1999:4, from the sample worksheet earlier in this section. cal(q) 1999:1 open data sample.xls data(format=xls,org=col) 1999:1 2000:3 This reads all the series on the sample file. Because there are no data for 2000:2 and 2000:3, those entries are filled with NA. cal(m) 2002:1 open data daysales.xlsx data(format=xlsx,org=col,compact=sum) 2002:1 2009:12 This reads data from an Excel 2007 xlsx file, summing the daily data on the file to create monthly data in working memory. cal(q) 1950:1 open data sales.xls data(format=xls,org=col,sheet="quarterly",verbose) 1950:1 2006:4 This reads data from an Excel workbook containing several spreadsheet pages. Here, we are reading from the sheet entitled Quarterly. The VERBOSE option produces information about the data file, including the frequency and starting date. It will also indicate if any frequency conversion is being performed to match the CALENDAR frequency.

Introduction to RATS

Int117

Data
ORG=MULTIROWS
Consider the following. It has three years of quarterly data for two firms. This could be in any of the spreadsheets formats. FIRM A 2007 11.3 11.6 10.9 12.3 2008 13.0 12.8 11.5 12.5 2009 12.9 13.0 13.2 13.6 FIRM B 2007 22.5 21.9 24.3 25.6 2008 21.9 21.8 22.6 23.5 2009 22.5 25.0 24.5 25.4 We cant easily convert this to a Labeled Table because the data for a single series span multiple rows. One approach (probably the simplest in this case) is to eliminate the FIRM A and FIRM B lines and delete the 2007, 2008, and 2009. In this is in a spreadsheet, you could just delete the first column, and save whats left as text, and read it as free-format (see Section 2.9) with cal(q) 2007 open data firms.txt data(format=free,org=rows) 2007:1 2009:4 firma firmb

An alternative is to use a set of options on DATA to isolate just the information that we need. All the spreadsheet formats have a third choice for the ORG option, which is ORG=MULTIROWS. This is for precisely this type of situation, where a data series covers several lines. In order to use this, however, we need to know (in advance) exactly how many data points there are in each series, since the limit on a series isnt defined by the end of the data file. The data for the first series starts at column 2 of row 2 and has twelve data points while the second starts at column 2 of row 7. We have to read each series separately with something like: open data firms.xls cal(q) 2007 all 2009:4 data(format=xls,left=2,top=2,nolabels,org=multi) / firma data(format=xls,left=2,top=7,nolabels,org=multi) / firmb Both the date scheme (quarterly 2007:1) and end date (2009:4 or twelve entries) need to come from us since those wont be recognized from the file.

Int118

Introduction to RATS

Data
Writing Data
There are several ways to write data to a spreadsheet file. The most common is to use the COPY command to write data series. Begin by opening a file for output with OPEN COPY. Then use COPY to write the data to the file. We recommend that you use the ORG=COL option. You can include dates on the file with the DATES option. Do a CLOSE COPY command if you want to be able to open the file in another application right away. For spreadsheets, this creates a new file (overwriting any existing file of the same name), so you must write the entire file using a single COPY instructionyou cannot use multiple COPY commands to append data to existing spreadsheets. You can also write series to a spreadsheet by opening the Series List Window using ViewSeries Window, selecting the series you want to export, and doing FileExport. open copy sample.xls copy(format=xls,org=col,dates) 1999:1 2000:1 ige fge cge produces the sample worksheet from earlier in the section.

Adding Dates to a Spreadsheet

One way to add dates to a spreadsheet file which doesnt have them is to read the data into rats, set the desired CALENDAR, and write the data back to a new file. If you do that with COPY, use the DATES option; if you export it, make sure you check the Label Observations with Dates box. If you prefer to add dates directly to a spreadsheet, heres an easy way to add monthly or quarterly date labels to a spreadsheet file. First, create a blank column or row and format it as one of the built-in date formats. Next, enter the date number corresponding to your first date in cell A2 (or B1 for ORG=ROWS) using the appropriate date function. In Excel, for example, you could enter the date June 1, 1999 with the formula: =date(1999,6,1) To fill the rest of the date column or row with monthly dates, enter the following formula in cell A3 (or C1): =date(year(a2+31),month(a2+31),1) If ORG=ROWS, use +b1+31 in the formulas, rather than +a2+31. Finally, copy this formula to the rest of the cells in the date column or row. For quarterly dates, just use +92 in place of +31.

Introduction to RATS

Int119

Data 2.9 Text Files

Free Format
We use the term free format to refer to text files that contain only numbers, with no series names, date labels, or other alphanumeric labels. Numbers on the file can be separated by blanks, tabs, commas, or new lines. You can use the Data (Other Formats) operation to read data from free format files. Just select Text Files or Comma Delimited from the file-type list before selecting the file you want to read. When you complete the Data Wizard dialog box, rats will prompt you for a name for each series being read in. If typing in the DATA instruction directly, use FORMAT=FREE and the appropriate ORG option (ORG=COLS or ORG=ROWS), and include a list of the names you want to assign to the series. See below for more on reading free format files.

PRN and Other Delimited Formats

prn format refers to a text file structured like a spreadsheet. The name dates back to

the days of Lotus 123, where it described a printable (text) version of a spreadsheet. To be read as a Labeled Table, prn files need to have data structured as described on page Int114. Series should be arranged in columns or rows, and the file must contain names for each series. prn files can also contain dates, in the form of date-format strings such as yyyy/mm/dd. Note that you cant use any type of numerical coding for the dates. Numbers on the file can be separated by blanks, commas, tabs, and/or carriage-returns. The columns dont need to be nicely aligned since its the delimiters that determine where one column ends and the next starts. You can read the file using the Data (Other Formats) operation, or by typing in a DATA instruction with FORMAT=PRN and the appropriate ORG option.
rats also offers two related choices: FORMAT=CDF for comma-delimited files and

FORMAT=TSD for tab-delimited files.

For reading data, CDF, TSD, and PRN are interchangeable. Regardless of which you use, rats will accept commas, tabs, or spaces as separators. For creating a file with COPY, the option will determine the type of separator used: commas for CDF, tabs for TSD, and spaces for PRN.

Free Versus PRN

Free format files can be very convenient if you want to type in a short data set, because you can create them with any editor that can save files as plain text. They are also the most natural format for data scanned in from printed material. However, because these files do not contain series labels or date information, they are not good for long-term use, and rats cannot read them with much flexibility. Usually, you will want to convert free-format files to rats format files, or to one of the spreadsheet formats.

Int120

Introduction to RATS

Data
prn format is a generally more useful and reliable. The presence of series names

reduces the chance of mistakes identifying data, and also give you the option of only reading in selected series. The ability to include date labels is clearly an advantage for dated time series data. If you have a text file that does have variable labels, or other non-numeric characters you could: read the file using FORMAT=PRN. This allows rats to process the series labels and dates (if any). import the data into a spreadsheet and read that file into rats (see Section 2.8). Spreadsheet programs have very sophisticated parsing dialogs for taking text and dividing it into spreadsheet cells. If you have an otherwise well-formatted text file which does not include series labels, you could: edit out all the non-numeric characters and read the file as FORMAT=FREE. edit the file to add series labels and read it as FORMAT=PRN.

Reading Free Format Files

When reading a free format file, rats reads series line by line, according to the following:

If your data are organized by row, each variable must begin on a new line. The data for a given series can, however, extend over more than one linejust make sure that the data for the next series begins on a new line. If DATA does not find enough entries on a line to fill a series, it automatically moves on to the next line and continues reading numbers. Extra rows of spaces are fine (for instance, separating data for two series). FORMAT=FREE will just keep scanning until it hits the next set of numbers. If your data are organized by column, the data for each new observation must begin on a new line. As above, data for a particular observation can extend over multiple lines. If DATA does not find enough entries on a line to fill all the series for an observation, it automatically moves on to the next line and continues reading numbers. When it has read data for all series for that observation, it drops to the next line to start reading in the next observation. rats interprets the characters NA or #N/A (upper or lower case), or a period followed by zero, one, or two non-numeric characters, as a missing value. If it encounters any other non-numeric characters, rats will interpret them as a missing observation and display a warning message.

Introduction to RATS

Int121

Data
Troubleshooting
Free format allows data for a single observation (or a single variable) to cover several lines. The disadvantage of this is that it becomes more difficult to pinpoint errors. If you forget one number early in the file, DATA will automatically pull in the next line to fill the omitted value, and throw the data off. If you get the error Unexpected end of file... when reading with FORMAT=FREE, it means that rats reached the end of the data file before it filled all of the data series. To determine what happened, do the following: Check your ALLOCATE setting or the start and end parameters on DATA to make sure they are set correctly. Note that if, for example, you have ALLOCATE 2003:5:14, you can do DISPLAY 2003:5:14 to see how many data points are implied by the ALLOCATE range. Make sure that you have the proper ORGANIZATION option. Check that the data file has the number of observations and variables that you expect, at least at first inspection. If all seems to be correct and its a small enough file, you can do a quick check for typographical errors. If all this fails to locate the problem, you will have to let DATA help you find the problem. For illustration, suppose you have a file that looks like this: 1,2,3 10.20,30 100,200,300 1000,2000,3000 This is supposed to have 4 observations on each of three series, but the second line has a decimal point between the 10 and the 20, where it should have a comma, so the line only contains two values (10.20 and 30) rather than three. If we read this with data(org=col) 1 4 first second third print we will get the error message,
Unexpected end of file while processing line 5. (series FIRST entry 4).

We can tell the following from this: 1. DATA thinks it needs a fifth line to read the requested data. Our data set is supposed to have four linesthis tells us that, in fact, the file has as many lines as we think. 2. rats was trying to read the fourth observation from a fifth line, so for some reason we are precisely one line off at the end of the file. We can then examine the values of the series with PRINT, looking for values which have ended up in the wrong series, or locating where a series gets off sequence.

Int122

Introduction to RATS

Data 2.10 File Handling Tips and Tricks

Cant Open a File?
If you have created a data file in rats and are having trouble opening the file in another application, you need to make sure that rats doesnt still have the file open. For example, if you are working in interactive mode and do: open copy test.dat copy(format=prn,org=columns) / x y z and then immediately try to open TEST.DAT in another application while rats is still running, you will probably get an error message. And, if you view a directory listing of TEST.DAT using Windows Explorer or a similar utility, it will appear as a zero-byte file. Thats because rats still has the file open. In order to access the file in another application, just close the file by issuing the command: close copy or else by quitting the rats application. Similarly, rats may not be able to read data from a file that is currently open in another application.

File Units

In rats, you open a file by associating a file name with an input/output unit. For example, the statement: open data myfile.txt simply associates the filename myfile.txt with the DATA unit, which is one of the reserved unit names in rats. Every rats instruction that can read or write data has a UNIT option which allows you to specify the source or destination unit for that operation. Fortunately, this business of file unit names is generally transparent to the user, because all of the relevant instructions have default settings for the UNIT option. For example, the DATA instruction applies to the DATA unit by default. So, you can read data from a file by simply doing something like: open data sample.rat data(format=rats)

Because DATA is the default unit for the DATA instruction, it will automatically read the data from the SAMPLE.RAT file, which has been associated with the DATA unit by the OPEN command. If no file has been associated with the DATA unit, rats will prompt you for a filename. In some cases, however, you may find it helpful to make use of the UNIT option, particularly when you want to have several data files or output files open simultaneously. This is made easy by the fact that rats lets you define your own unit names.

Introduction to RATS

Int123

Data
For example, suppose you want to read data from two different files in your program. You can either: Use an OPEN DATA command to associate the first file with the DATA unit, and then read the data in using DATA. Then, repeat this procedure for the second file: open data first.rat data(format=rats) / x open data second.rat data(format=rats) / y Or, you can define your own custom unit names, and associate one file with each name. You can then read from the files in any order by specifying the appropriate unit name. For example: open data1 first.rat open data2 second.rat data(format=rats,unit=data2) / y data(format=rats,unit=data1) / x Because you are using different unit names for each file, both files remain open, so you can go back and read from either one without having to repeat the OPEN command.

Int124

Introduction to RATS

3.. Graphics

SCATTER for x-y graphs, GCONTOUR for creating contour graphs, and GBOX for creating box plots. All of these instructions offer a very large collection of options, allowing you to create publication-quality graphs easily. In addition to a variety of basic graph types (such as line graphs and bar graphs), rats offers a number of special effects, such as overlay graphs (displaying data in two different forms or with two different scales in a single graph) and arrays of smaller graphs. You can also define graphics style sheets, which allow you to customize various attributes of your graphs, including colors, patterns, line thicknesses, and more. Finally, graphs created can be exported in several well-known formats (such as PostScript) for placement in documents.

ats offers four main graphing instructions: GRAPH for time-series graphs,

Displaying Graphs Saving and Printing Graphs Graph Style Sheets Exporting Graphs for Publication Graph Labels Special Graphs and Examples

Graphics 3.1 Graphics

The section provides general information on working with graphs in rats. We will cover the instructions used to generate graphs, as well as the parts of the rats interface which deal with graphs, including exporting graphs for use in other applications.

Overview of the Graphics Instructions and Wizards

rats has four main instructions for creating graphs:

GRAPH creates time series plots, or, more generally, plots of a series which is arranged as a sequence. The Graph wizard on the Data/Graphics menu generates a GRAPH instruction. You can produce line graphs, filled line graphs, several types of bar graphs, and high-low-close graphs. SCATTER creates scatter (x vs. y) plots. In addition to series vs. series, it does graphs where the x axis is a grid on the real line, either equally or unequally spaced. The Scatter(X-Y) wizard on the Data/Graphics menu generates a SCATTER instruction. GCONTOUR (Section 3.13) produces contour plots. GBOX (Section 3.14) produces box (or box and whisker) plots. These are very flexible instructions which allow you to create a wide variety of graphs. The auxiliary instructions SPGRAPH, GRTEXT, and GRPARM further enhance the abilities. With rats, you can: use color, dashed line patterns, or shades of gray to distinguish series. rats can automatically translate color into patterns or grayscale for output to black and white printers. generate a graph with series having different scales (which we refer to as overlay graphs) using a single GRAPH or SCATTER instruction (Section 3.7). The overlay options also allow you to mix different styles on a single graph. For example, you could plot one series as a line and another using a bar-graph style. create pages of equally sized postage-stamp graphs. This is done using SPGRAPH (Section 3.8). add explanatory text within the graph, done with GRTEXT (Section 3.9). use graph style sheets to set the color and thickness of lines, choose from a range of fill patterns for bar graphs, and more (Section 3.16). change fonts and font sizes, done with GRPARM (Section 3.18) The GRAPH and SCATTER instructions were introduced in the tutorial in Chapter 1. For complete details on these and the other instructions listed above, see the remainder of this chapter and the relevant sections in the Reference Manual.

Int126

Introduction to RATS

Graphics 3.2 Working with Graph Windows

Resizing Graph Windows
We already briefly discussed Graph Windows in Section 1.5.5. As mentioned there, when you resize a Graph Window, the contents resize with it. This is the standard behavior that you would get when you insert a graph into a document. However, the windows that you see on the screen are not (in general) in the proportions used if you export or copy the graph to insert into a publication. Instead, rats does that in a roughly golden ratio with the width being 1.5 times the height. If you paste that into a document in standard orientation, the graph will generally cover about 1/3 of the page. You can force rats to use different proportions by first resizing the window to get the appearance that you want, then using the toolbar icon. If you resize the window further, youll see that the graphs shape doesnt change, just the size. That shape will be maintained if you copy, print, or export the graph.

Color or Black and White

rats can render a graph either in color, or in black and white. Color is almost always

the easiest to view on the screen, so thats the standard way that graphs are shown, but most (print) publications still require graphs in black and white. If you need a grayscale graph, click on the toolbar icon: . If you copy or export the Graph Window, whether that is done in color or black and white will depend upon how you are showing the graph at the time. If you want to change back to color, just click on the same toolbar icon, which will now be in color. Grayscale graphs can be hard to read if you have lines that cross many times. If you have such a graph, you may need to either change the style numbers (page Int130), or define your own style sheets (page Int149).

Saving and Opening Graphs

If you want to save a graph so you can reload it later, you need to save it in rgf format (rats Graphics Format). To open a saved rats graph, select the Open... operation from the File menu, choose rats graphics from the list of file types, and then open the desired graph file. This will show the graph in a new Graph Window, exactly as if it has just been generated by the program.
rgf is designed for temporary storage of graphs. Its specification changes as we add

features to our graphics system, so you should only open files with the same version of rats that created them. It is, however, portable across systems, so a file can be shared by Windows, Mac and unix users as long as they are running the same version number of rats.

Introduction to RATS

Int127

Graphics
Printing
To print a Graph Window, do the following: 1. Make sure the desired graph window is active. 2. Select Print from the File menu, or click on the toolbar icon.

If you have a black and white printer, your graph will print with colors replaced by the black and white patterns, whether or not that is showing in your window. If its a color-capable printer, then the graph will print in color if youre showing the graph in color, or in black and white if thats what youre showing. Printers will allow you to choose between portrait and landscape mode. The results of the two choices are shown below. When youre printing a graph, youre more likely to use landscape, since it uses the full page. Portrait generally will only give a partial page since the graph itself is wider than it is tall, but the paper, in that orientation, is the reverse.

Portrait:
Portrait Mode
140
RATE IP PPI

Landscape:
100 120 140 20 40 60 80 0 1959 1962 1965 1968 1971 1974 1977 1980 1983 1986 1989 1992 1995

120

RATE IP PPI

100

80

60

Landscape Mode

40

20

0 1959 1962 1965 1968 1971 1974 1977 1980 1983 1986 1989 1992 1995

Int128

Introduction to RATS

Graphics 3.3 Preparing for Publication

You have three ways to include rats graphs in a publication:

Export the graph to one of the standard graphics file formats and then import the graph into your word-processor or page-layout program. Use EditCopy to copy a graph to the system clipboard, and then paste it into your word-processing application. Print a copy of your graph directly from rats and include this in your publication. For example, if you have Adobe Acrobat or similar software, you can print directly to a pdf format file. The first two are obviously more flexible, since most word processors will allow you to size and shape the graph to fit your document. Make sure that you have the desired settings for the proportions and color as discussed on page Int127. Those will be applied when exporting or copying the graph.

Exporting/Saving

To export a graph for placement into another application: 1. Make sure the desired graph window is active. 2. Choose the Export or Save As operation from the File menu or the icon. Select the desired format. toolbar

PostScript is the preferred format for this because it gives the most accurate translation of the graph, and works very well if you will be producing your output as a pdf file, or on a PostScript printer. pdf format should also work very well for any applications that support importing pdf files. wmf (Windows Metafile) can work well for color graphs, but isnt as good with patterns. Note that PostScript comes in two orientations, Portrait and Landscape. When you place the graph file on a page, these will have the same orientations shown on the previous page for printing. Portrait is the standard here, and is what we use in all the graphs in this manual other than samples of landscape.

Copying and Pasting

To copy a graph to the clipboard: 1. Make sure the desired graph window is active. 2. Choose the Copy operation from the Edit menu or the toolbar icon. Windows versions copy both bitmap and Windows metafile versions of the graph to the clipboard, while Macintosh versions copy in pdf format. To paste the graph, switch to the other application and select the Paste or Paste Special... operation from the Edit menu (Paste Special... allows you to select the format that is pasted).

Introduction to RATS

Int129

Graphics 3.4 Graph Styles and Style Numbers

Selecting Colors, Lines, and Patterns
rats graphs data using lines, fill patterns, or symbols, depending upon the settings

you choose for the STYLE and OVERLAY options: for instance, the painted styles like bar, stacked bar, and polygonal all use fill patterns. For each category (line, fill, symbol) rats supports up to thirty different styles for color mode, and a corresponding thirty for black and white mode. Styles are identified and selected by number, from 1 to 30, plus style 0 which is used by the SHADE option (Section 3.11). The default styles used by rats have been chosen to be fairly easily distinguishable roughly for style numbers 1 through 10. You can use Graph Style Sheets (Section 3.16) to create your own definitions for the style numbers. By default, rats uses style number one for the first series being graphed (the series listed on the first supplementary card), style number two for the second series, and so on. You can use the stylenum parameter on the supplementary cards of GRAPH and SCATTER to select different styles. For example: graph 2 # x1 / 4 # x2 / 2

selects style number four for the X1 series and style number two for the X2 series.

Color and Black and White Styles

rats normally displays graphs in color, using different colors to distinguish series.

In color mode, with the default style definitions, lines are drawn as solid lines in different colors, fill patterns as solid fills in different colors, and symbols using the same symbol in different colors. These usually are easy to distinguish.
rats switches to the corresponding black and white (grayscale) styles if you: print to

a black and white printer; use the ( ) toolbar button; or use the PATTERNS option on the graphing instruction. With the default styles, lines are drawn in black using different dash patterns to distinguish series. Fills are drawn using different hatch patterns, and different symbols (in black) are used.

If you have a graph with several intertwined lines, it can be hard to see the detail of the lines when you cant use color. The solid black (style 1) isnt the problem; its the dashed lines which lose detail and may seem to disappear at times. If you have one series which is fairly smooth and one which is more volatile, list the volatile one first so it gets style 1. If that wont work, you may need to use thicker solid grayscale linesstyles 8, 9, 10 and 11 are defined that way (this may look a bit odd in color). However, it may not be possible to make the graph work if its too busy, and you may need to think of a different way to show the data.

Int130

Introduction to RATS

Graphics 3.5 Labeling Graphs

To help you produce publication-quality graphs, rats gives you a great deal of flexibility in labeling graphs. We look at these in the paragraphs below.

Major Labels

The four main graphing instructions, as well as SPGRAPH, all offer HEADER, FOOTER, SUBHEADER, HLABEL, and VLABEL options. Their functions are shown in the graph below. In addition, in its title, the Graph Window will assume the name of the HEADER or the FOOTER (HEADER if you use both).

HEADER is the long label at the top of the graph

100

SUBHEADER is the second label at the top, immediately below the HEADER

VLABEL labels the vertical scale

80

60

40

20

0 10 20 30 40 50 60 70 80 90 100

FOOTER adds a label at the bottom left

HLABEL labels the horizontal scale

Other Labels

When the main labels are used on SPGRAPH (Section 3.8), they are placed at these locations relative to the full graph page, and not to any single graph on it. You can use the XLABELS and YLABELS options on SPGRAPH to label the rows and columns of the graph matrix, or you can add the standard label types to the individual graphics instructions, which will add (smaller) labeling in the zone assigned to that graph. The GRTEXT instruction (Section 3.9) is a very flexible tool that allows you to add strings of text anywhere inside or in the margins of a graph. You can select the exact location, font, size, and style of the text string. SCATTER and GCONTOUR (Section 3.13) offer an XLABELS option to supply your own list of labels for the X-axis tick marks, while GBOX (Section 3.14) provides a LABELS option for labeling each of the series in the graph.

Introduction to RATS

Int131

Graphics
Supplying Labels
You can supply either a literal string of text enclosed in quotes, or a LABEL or STRING variable defined ahead of time (some options require a VECTOR of STRINGS). For example, the following GRAPH commands produce the same result: graph(header="US Real GDP") # rgdp

compute hlab = "US Real GDP" graph(header=hlab) # rgdp

Constructing Strings

Quoted strings are fine for single graphs, but if you need to generate many standardized graphs, it can be rather tedious to duplicate code and hand edit it. Fortunately, all these options will permit you to use a STRING type variable instead of a quoted string. These can be constructed, or input (see below). compute header="Coherence of "+%l(x1)+" and "+%l(x2) graph(header=header,noticks) # coherenc compute header="Autocorrelations "+diffs+" x "+sdiffs graph(header=header,max=1.0,min=-1.0,style=bargraph) # corrs Note: %L(series) is a special function which returns the label of a series.

Reading Strings From a File

If you have a standardized program, but the headers or labels cant be constructed easily as they have no obvious pattern, make up a separate file with a list of headers (one per line) and read them in using the instruction READ: open data labels.rgf declare vector[string] header(230) read header do i=1,230 ... graph(header=header(i),...) # ... end do i

Line Breaks in Strings

You can use two backward slash characters (\\) to insert line breaks in strings used in graphs. See the GRTEXT examples in Section 3.9.

Int132

Introduction to RATS

Graphics 3.6 Keys (Legends)

The KEY option on GRAPH and SCATTER allows you to add a key (legend) at a choice of several locations. Four of these are outside the graph box (KEY=BELOW, ABOVE, LEFT or RIGHT). Four of these are inside the graph box (UPLEFT, LOLEFT, UPRIGHT, LORIGHT). This second group doesnt take space away from the graph itself, but you have to be careful to choose a location (if it exists) which doesnt interfere with the graph. There is also KEY=ATTACHED, which puts the graph labels on the graph itself at locations where the labeling will be as clear as possible. This is available only for a few types of graphs (line, step, and symbols time series graphs). By default, rats uses the series names for the key labels. If these are cryptic database names, you can provide your own labels on the graph using the KLABELS option. KBOX controls whether a box is drawn around the key, while KHEIGHT and KWIDTH allow you to control the size and shape of the key. If you only want to see text in the key box (perhaps you dont need to show a sample line because you are graphing only one series), use NOKSAMPLE. This example uses descriptive titles rather than the original series names for the key. The key will be placed below the graph. graph(key=below,klabels=||"30 Year Bonds","3 Month Bills"||) 2 # ftb30y 1998:1 1999:12 # ftbs3 1998:1 1999:12 This uses KEY=ATTACHED, and also KLABELS, using in-line matrix notation (page Int90) to provide the VECTOR of STRINGS. graph(style=line,key=attached,header="Figure 3.15",$ klabels=||"AGRICULTURE","RETAIL","SERVICES","MANUFACTURING"||) 4 # ga8gff # ga8gr # ga8gs # ga8gm
Figure 3.15
1000

750

MANUFACTURING

500

SERVICES

RETAIL

250
AGRICULTURE

0 1960 1962 1964 1966 1968 1970 1972 1974 1976 1978 1980 1982 1984 1986 1988

Introduction to RATS

Int133

Graphics 3.7 Overlay (Two-Scale or Two-Style) Graphs

The OVERLAY option makes it easy to do a GRAPH or SCATTER plot with two different vertical scales. It also allows you to combine two different styles. For example, you can graph some series using a line style and others using a bar graph style. By default, the scale for the first set of series appears on the left side of the graph, while the scale for the second set of series appears on the right side of the graph. You can use the OVSAMESCALE option (along with OVERLAY) if you want to use the same scale for both axes, but with different styles for different series.

Creating Overlay Graphs

Here is the basic procedure for creating an overlay graph:

Use the OVERLAY option on your GRAPH or SCATTER instruction to tell rats to do an overlay graph. OVERLAY offers the same choices as STYLE, but it sets the style used for the overlaying (right-scale) series. The STYLE option itself applies only to the left-scale series. If you are only graphing two series (or two pairs of series for SCATTER), you dont need to do anything else. Just list the series on separate supplementary cards, as you would for an ordinary graph. The second series (or pair) are graphed using the OVERLAY style and scale. If graphing more than two series, use the OVCOUNT option to tell rats how many series should be graphed using the right-side (overlay) scale. For OVCOUNT=n, rats graphs the last n series (or pairs) listed using the right-side scale.

Notes

When you use the OVERLAY option, rats ignores the AXIS, EXTEND, and SCALE options on GRAPH and the AXIS, EXTEND and VSCALE options on SCATTER. This is because running an axis line or extending grid lines across the graph is likely to be very distracting, as they will apply to only one of the two scales. All other options, such as HEADER, HLABEL, and GRID, work normally. You can use the OVLABEL option to label the right-side scale. Use the standard VLABEL option if you want to label the left-side scale. The OMAX and OMIN options set the maximum and minimum values for the right-side scale. They function just like the MAX and MIN options, which will control only the left-side scale in a two-scale graph. If you want to use one of the painted styles, such as POLYGONAL or BARGRAPH, you should probably use that as the STYLE (left-scale), as these are drawn first. If you use a paint style for the overlaying series, they may cover up the first block of series. You can use the OVRANGE option to offset the two scales so they arent directly across from each other. OVRANGE=fraction gives each scale the given fraction of the verti-

Int134

Introduction to RATS

Graphics
cal size (fraction should be in the range 0.5 to 1.0). OVRANGE=.6 for instance will have them overlap only slightly (the left scale getting the bottom 60% and the right getting the top 60%).

Example

This simple example (from GRAPHOVERLAY.RPF) graphs two series. The series HSF (housing starts) is graphed using the POLYGONAL style, and its scale appears on the left side of the graph. The series FCME (mortgage rates) is graphed as a LINE, and uses the right-side scale. You will probably need to experiment with the stylenum parameters to get overlay graphs to look good. Here, we have used fill pattern 2 (medium gray) for the HST series, and line type 1 (solid black) for the FCME series. Without these, HST would be graphed using a solid black pattern (pattern 1), and FCME, which would be a dashed line, would not be as visible. We also used KLABEL to supply custom key labels. open data haversample.rat calendar(m) 1972 data(format=rats) 1972:1 1988:12 fcme hst graph(key=below,header="Housing Starts vs Mortgage Rates",$ klabel=||"Housing Starts","Mortgage Rate"||,$ style=polygonal,min=0.0,overlay=line) 2 # hst / 2 # fcme / 1

Housing Starts vs Mortgage Rates

2500 14 13 2000 12 11 10 1000 9 8 7 0 1983 1985 1987 1989 1991 1993 1995 1997 1999 2001
Housing Starts Mortgage Rate

1500

500

Introduction to RATS

Int135

Graphics 3.8 SPGRAPHMultiple Graphs on a Page

spgraph(vfields=2,hfields=2) four graph, scatter, or gcontour instructions spgraph(done) You can put two or more individual graphs on a single picture using SPGRAPH with the HFIELDS (horizontal fields) and VFIELDS (vertical fields) options. For example:

This SPGRAPH divides the available space in half in each direction, creating four zones of equal size. The graphs (by default) fill the zones by rows beginning at the top left. Note, however, that the characters on the small graphs are sized to be proportional to what they would be on a full page; that is, each quarter graph will look like a reduced version of the same graph done by itself. The labeling (particularly on the scales) can sometimes get too small to be readable, particularly if you get to four or more graphs in either direction. Depending upon the situation, you may want to increase the character sizes using GRPARM, or you might want to suppress the axis labeling altogether. To do this, use NOTICKS and SCALE=NONE on GRAPH and VSCALE=NONE and HSCALE=NONE on SCATTER or GCONTOUR. You also need to be careful that you do not leave false impressions by allowing each graph to find its own range. This could inflate the appearance of what may be, in reality, some very small effects. Most of the work in doing a matrix of graphs is, in fact, just preparation. The following (from example SPGRAPH.RPF) graphs the exchange rate of the dollar versus four foreign currencies. These are first given a common reference by converting them to percentage appreciation of dollars per foreign unit since the start of the sample. As all but the uk series (GBRUSXSR) are originally stated in foreign currency per dollar, they have to be flipped. We then use TABLE to get the maximum and minimum values attained across all the countries, and use these as the MAX and MIN on each GRAPH. This ensures that the very modest movements relative to the Canadian dollar arent exaggerated relative to the much larger movement versus the yen. The LABELS instruction is used to give a more readable label to the series, naming them by their country instead of the database coding. Their new labels are used in the GRAPH instructions within the loop on the HEADER option. cal(m) 1996:1 open data oecdsample.rat data(format=rats) 1996:1 1998:12 canusxsr frausxsr $ jpnusxsr gbrusxsr Flip to dollars per foreign unit and assign descriptive labels set canusxsr = 1.0/canusxsr set frausxsr = 1.0/frausxsr set jpnusxsr = 1.0/jpnusxsr labels canusxsr frausxsr jpnusxsr gbrusxsr # "Canada" "France" "Japan" "UK"

Int136

Introduction to RATS

Graphics
Convert to percent appreciation dofor i = canusxsr frausxsr jpnusxsr gbrusxsr compute base=([series]i)(1996:1) set i = 100*(i{0}/base - 1.0) end dofor i Use TABLE to get the maximum and minimum across all series table(noprint) / canusxsr frausxsr jpnusxsr gbrusxsr Set up the SPGRAPH for a 22 matrix spgraph(vfields=2,hfields=2,$ header="U.S. Dollar vs Major Currencies",$ subheader="Percent appreciation over the sample") dofor i = canusxsr frausxsr jpnusxsr gbrusxsr graph(max=%maximum,min=%minimum,header=%l(i)) 1 # i end dofor spgraph(done)

U.S. Dollar vs Major Currencies

Percent appreciation over the sample
Canada
35 30 25 20 15 10 5 0 -5 1996 1997 1998 35 30 25 20 15 10 5 0 -5 1996 1997 1998

Japan

France
35 30 25 20 15 10 5 0 -5 1996 1997 1998 35 30 25 20 15 10 5 0 -5 1996

UK

1997

1998

Introduction to RATS

Int137

Graphics
Here is another example using SPGRAPH. This is taken from an example in Diebold (2004). The example program file (DIEBP235.RPF) is included with rats. spgraph(hfields=2,vfields=2,$ header="Figure 9.17 Recursive Analysis: Breaking Parameter Model") scatter # x y3 rls(cohist=cohist,sehist=sehist,sighist=sighist,csum=cusum) $ y3 / resids # x set upperco = cohist(1)+sehist(1)*2.0 set lowerco = cohist(1)-sehist(1)*2.0 graph(header="Recursive Estimates") 3 # cohist(1) # upperco 10 * # lowerco 10 * set upperres = 2.0*sighist set lowerres = -2.0*sighist graph(header="Recursive Residuals") 3 # resids # upperres # lowerres set cusum = cusum/sqrt(%seesq) set upper5 startr end = .948*sqrt(%ndf)*(1+2.0*(t-startr)/%ndf) set lower5 startr end = -upper5 graph(header="CUSUM test") 3 # cusum # upper5 # lower5 spgraph(done)
Figure 9.17 Recursive Analysis: Breaking Parameter Model
Data Series
640 560 480 400 320 240 160 80 0 -80 0 50 100 150 200 -2 -4 -6 20 40 60 80 100 120 140 160 180 200 4 2 0 8 6

Recursive Estimates

CUSUM test
175 150 125 100 75 50 25 0 -25 -50 20 40 60 80 100 120 140 160 180 200 250 200 150 100 50 0 -50 -100 -150 -200 20 40 60

Recursive Residuals

80

100

120

140

160

180

200

Int138

Introduction to RATS

Graphics 3.9 GRTEXTAdding Text to Graphs

The GRTEXT command allows you to add text anywhere inside a graph image or in the margins of the graph, with options allowing you to specify the position of the text, the alignment, and the font style. To use GRTEXT, you need to enclose the graphing and GRTEXT command(s) inside an SPGRAPH block. The code excerpt below is from the HISTOGRAM.SRC procedure included with rats, which generates histogram plots. We use GRTEXT to include computed statistics on the graph. Note the use of \\ symbols to put line breaks into a single stringthis is often easier than using multiple GRTEXT commands to achieve the same effect. spgraph display(store=s) "Mean" %mean "\\Std Error" sqrt(%variance) $ "\\Skewness" %skewness "\\Exc Kurtosis" %kurtosis if distrib==2 { set nx = 1.0/sqrt(%variance)*$ %density((fx-%mean)/sqrt(%variance)) scatter(style=bargraph,overlay=line,ovsamescale) 2 # fx dx # fx nx } else { scatter(style=bargraph) 1 # fx dx } grtext(position=upright) s spgraph(done)
0.45 0.40 0.35 0.30 0.25 0.20 0.15 0.10 0.05 0.00 -2 -1 0 1 2 3 4

Mean Std Error Skewness Exc Kurtosis

1.02374 1.06368 -0.10658 0.09784

Introduction to RATS

Int139

Graphics 3.10 Graphing Functions

If you want to graph a function y = f (x ), use the instruction SCATTER. Create a grid series with x values and a corresponding series of y values. SCATTER with the option STYLE=LINE will create the graph. You can also use other style choices, such as BAR and POLYGONAL. In general, you should not need more than about 100 grid points on the x-axis for a smooth function. Note that the LINE, BAR and POLYGONAL styles on SCATTER only work correctly if the x series is sorted into increasing order.

The following computes a series of AR1 regressions for a grid of values for r between .5 and 1.0 and graphs the residual sum of squares against r. This is from example file GRAPHFUNCTION.RPF. @GridSeries(from=.5,to=1.0,size=.005,pts=gpts) rhos set rss 1 gpts = 0.0 do i=1,gpts ar1(noprint,rho=rhos(i)) invest 1950:1 1985:4 # constant ydiff{1} gnp rate{4} compute rss(i)=%rss end do i scatter(style=lines,vlabel="Residual Sum of Squares",$ hlabel="Value of rho",header="Multiple Mode in AR1") # rhos rss
Multiple Mode in AR1
42000 39000

Residual Sum of Squares

36000 33000 30000 27000 24000 21000 18000 0.5 0.6 0.7 0.8 0.9 1.0

Value of rho

Int140

Introduction to RATS

Graphics 3.11 Highlighting Entries

The options GRID and SHADING permit you to draw attention to particular entries. GRID does this by drawing a vertical line from top to bottom of the graph at specific entries. SHADING paints a shaded box over any set of consecutive non-zero entries. Shading or grid lines are done before any of the series. If the graph uses one of the painted styles (POLYGONAL, BAR, STACKED or OVERLAP), it may cover the highlighting, so be careful if you want to use those. SCATTER and GCONTOUR have separate options for grids or shading in each direction. Note that, unlike GRAPH, where the highlighting information is provided in data series, for SCATTER the grid lines are provided using a VECTOR in the HGRID and VGRID options and the shading zones are in a N2 RECTANGULAR for HSHADE and VSHADE. set raterise = t>=1973:8.and.t<=1974:12.or.$ t>=1978:3.and.t<=1981:11 graph(shading=raterise,header="U.S. Private Housing Starts",$ subheader="With Rising Mortgage Rate Intervals") 1 # hsf 1972:1 1987:12

is from GRAPHOVERLAY.RPF. It puts shading over 1973:8 to 1974:12 and 1978:3 to 1981:11. You might find that you need to darken the shading for publication. To do that, redefine FILL_BW_0 (see page Int149). The default value is .90; if thats too light, you might try .85.

U.S. Private Housing Starts

2500 2250 2000 1750 1500 1250 1000 750 1972 1974 1976 1978 1980 1982 1984 1986

With Rising Mortgage Rate Intervals

Introduction to RATS

Int141

Graphics 3.12 Fan Charts

Fan charts provide a useful way to display confidence bands, particularly with multiple levels. The confidence bands are displayed in shades of the same color, which are darkest nearest the center and growth fainter towards the outer bands. With the GRAPH instruction, its usually best to do these as an overlay, with the point forecast being done as a line. Make sure that you use OVSAME with OVERLAY=FAN so the bands are located in the correct location. This is from Makradakis, et. al (1998), part of textbook example MWHP366.RPF. It computes point forecasts and upper and lower 80% and 95% confidence bands. The pre-forecast data and the point forecasts are done as line graphs; the confidence bands as an overlay. The order of listing of the series covered by the fan isnt important. At each point, the values are ordered from bottom to top. With four series, there are three bands. The center is done in a darker gray (for black and white) and the outer ones in a lighter shade. uforecast(equation=weq,stderrs=stderrs) wfore 1973:1 1974:12 set lower95 1973:1 1974:12 = wfore+%invnormal(.025)*stderrs set upper95 1973:1 1974:12 = wfore+%invnormal(.975)*stderrs set lower80 1973:1 1974:12 = wfore+%invnormal(.1)*stderrs set upper80 1973:1 1974:12 = wfore+%invnormal(.9)*stderrs graph(footer="Figure 7-26 Forecasts and Prediction Intervals",$ ovcount=4,overlay=fan,ovsame) 6 # writing 1969:1 1972:12 # wfore # lower95 # lower80 # upper80 # upper95
Figure 7-26 Forecasts and Prediction Intervals
1200 1100 1000 900 800 700 600 500 400 300 1969 1970 1971 1972 1973 1974

Int142

Introduction to RATS

Graphics 3.13 GCONTOURContour Graphs

The GCONTOUR instruction generates contour plots. Its design is quite different from GRAPH and SCATTER, because it needs the x and y values to be in the form of a grid, and the function values must be a matrix with dimensions size of x grid by size of y grid. This matrix is usually created using the instruction EWISE, which fills an array based on a function of the row and column numbers. See EWISE in the Reference Manual and Section 1.7 of the Users Guide for details. The following is taken from the file GCONTOUR.RPF. This analyzes the log likelihood of a model with a break point in the mean as a function of the location of the break point (shown on the y axis) against the mean of the post break process (x axis). Set up the grids for BREAKS (1,...,100) and MUS (.2,.4,...,20) compute breaks=%seqa(1.0,1.0,100) compute mus =%seqa( .2, .2,100) Generate the log likelihood for all the combinations of MUS and BREAKS. dec rect f(100,100) ewise f(i,j) = -50.0*log((x2-2*mus(i)*over(fix(breaks(j)))+$ (100-breaks(j))*mus(i)^2)/100.0) Do the contour graph with a grid line across the actual break (50) gcontour(x=mus,y=breaks,f=f,vgrid=||50||)

100

80

60

40

20

0 0 5 10 15 20

Introduction to RATS

Int143

Graphics 3.14 GBOXBox Plots

Box plots (also known as box and whisker plots) provide a quick way to examine some basic statistical properties of one or more data series, and they can give you a visual indication of how the observations in the data are distributed. This simple example (from GRAPHBOXPLOT.RPF) draws box plots for two data series which list taxes as a share of total output for 22 countries in two different years. The first series contains the data for 1965, while the second series contains data for 1983: open data CountryTaxData.xlsx calendar(panelobs=22) data(format=xlsx,org=columns) 1//1 1//22 Country Tax1965 Tax1983 gbox(header="Taxes as Share of Output for 22 Countries",$ subheader="1965 vs 1983",labels=||"1965","1983"||,$ frame=half,extend) 2 # tax1965 # tax1983

Here, weve used the LABELS option to tell GBOX to use 1965 and 1983 as the labels for the two series, rather than the series names. In this case, we use in-line matrix notation to provide the VECTOR of STRINGS. This could also be done by storing the labels into a variable of type VECTOR[STRING] ahead of time: compute [vector[string]] gboxlab = ||"1965","1983"|| gbox(header="Taxes as Share of Output for 22 Countries",$ subheader="1965 vs 1983",labels=gboxlab,$ frame=half,extend) 2 etc.

Taxes as Share of Output for 22 Countries

55 50 45 40 35 30 25 20 15 10 1965 1983

1965 vs 1983

Int144

Introduction to RATS

Graphics 3.15 Miscellaneous Graph Types

Autocorrelations
Autocorrelations and partial autocorrelations are bounded between -1.0 and 1.0. Graphing them without fixing the plot limits can produce confusion, because the plot ranges for different sets of correlations will probably be quite different. We have found the following set of options to be very helpful: graph(number=0,style=bargraph,max=1.0,min=-1.0) 1 # corrs NUMBER=0 causes the time axis to be labeled with 0,1,2,etc. STYLE=POLYGONAL also works well for correlations, and STYLE=SPIKE can be useful if you have to graph a large number of correlations. See Section 6.4.1 of the Users Guide for more details on computing and graphing correlations. This shows a very basic graph of correlations. If you use any rats procedures which compute and graph autocorrelations (such as the BJIDENT procedure included with rats), you are likely to run into some graphs that supply additional information. For instance, its possible to highlight the correlations which are statistically significant using the SHADING option, which we demonstrate in another context on page Int141.

1.00 0.75 0.50 0.25 0.00 -0.25 -0.50 -0.75 -1.00 0 5 10 15 20

Introduction to RATS

Int145

Graphics
High-Low-Close Graphs
You can use the option STYLE=VERTICAL for high-low-close graphs and the like. At each entry, it draws a line connecting the highest and lowest values and puts hash marks at the locations of the series. If there are three or more series, it puts a filled circle on the first series so it can be picked out easily. The example GRAPHHIGHLOW.RPF pulls monthly values of the Dow Jones Industrial Average into annual series. It uses three DATA instructions, one to get the maximum, one the minimum and one the final value during the year. Since each DATA instruction will reset SPDJI, the data are copied to a new series name after each. The close series should be listed on the first supplementary card so it gets tagged with a special symbol. For this graph, we look at results from 1982 through 2006. open data haversample.rat calendar(a) 1982:1 all 2006:1 data(format=rats,compact=max) / spdji set djmax = spdji data(format=rats,compact=min) / spdji set djmin = spdji data(format=rats,compact=last) / spdji set djlast = spdji graph(style=vertical,header="DJIA: 1982-2006") 3 # djlast # djmin # djmax
DJIA: 1982-2006
12500

10000

7500

5000

2500

0 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006

Int146

Introduction to RATS

Graphics
Forecasts
If you graph forecasts with the last few observations of historical data (as separate series), you will find there will be a break in the graph between the end of the historical data and the start of the forecasts. One way to improve the appearance is to add the final historical value to the beginning of the forecast series. That way, its line will connect smoothly with the historical series. The following (from GRAPHFORECAST.RPF) uses ESMOOTH (exponential smoothing) to forecast from 2007:1 to 2008:12. The final actual data value from 2006:12 is added at that entry to the forecast series (JPNFORE). A grid line is added to the graph at 2006:12. cal(m) 1960:1 open data oecdsample.rat data(format=rats) 1960:1 2006:12 jpniptotrs

esmooth(trend=select,seasonal=select,$ forecast=jpnfore,steps=24) jpniptotrs set jpnfore 2006:12 2006:12 = jpniptotrs graph(header="Forecasts of Japanese IP",grid=t==2006:12) 2 # jpniptotrs 2004:1 2006:12 # jpnfore 2006:12 2008:12
Forecasts of Japanese IP
130 125 120 115 110 105 100 95 2004 2005 2006 2007 2008

Introduction to RATS

Int147

Graphics
This is another example, taken from pages 349-360 of Diebold (2004). It includes upper and lower confidence bands, and shades the forecast area. The full example program (DIEBP348.RPF) and associated procedure files are included with rats. Set dummy variable FOREZONE to 1 for 1995:1 and later:
set forezone * 2010:12 = t>=1995:1

Estimate the model through 1994:12

boxjenk(const,diffs=1,ar=1,define=diffeq) logyen * 1994:12 resids

Use UFORECAST to generate forecasts and standard errors:

uforecast(equation=diffeq,stderrs=sefore) fyen 1995:1 1996:7

Generate upper and lower bound series from results:

set upper 1995:1 1996:7 = fyen+2.0*sefore set lower 1995:1 1996:7 = fyen-2.0*sefore

Graph the actual values, forecasts, and upper and lower bounds. Shading is applied where FOREZONE is non-zero. The stylenum parameter is used to specify line style 3 for both the upper and lower series:
graph(header=$ "Figure 12.17. Log Yen/Dollar Rate: History, Forecast and Realization",$ shading=forezone) 4 # logyen 1990:1 1996:12 # fyen 1995:1 1996:7 # upper 1995:1 1996:7 3 # lower 1995:1 1996:7 3

Figure 12.17. Log Yen/Dollar Rate: History, Forecast and Realization

5.1 5.0 4.9 4.8 4.7 4.6 4.5 4.4 4.3 4.2 1990 1991 1992 1993 1994 1995 1996

Int148

Introduction to RATS

Graphics 3.16 Graph Style Sheets

The default styles are designed to work well in most circumstances, but you can also define your own custom styles to suit your preferences or organizational standards. You do this using Graph Style Sheets. Style sheets are external (text) files that allow you to define up to 30 different choices each for color lines, grayscale lines, color patterns, grayscale patterns, color symbols, and grayscale symbols. You can also define fill pattern style 0, which is used by the SHADE option. You have control over the following attributes:

For lines in color graphs: the line color, the pattern (solid line or one of seven dashed patterns), and thickness. For lines in grayscale graphs: the grayscale level, pattern, and thickness. For fill patterns (as used in bar graphs and similar styles) in color graphs: the color and the pattern, with seven patterns from which to choose (solid and six types of hatch patterns). For fill patterns in grayscale: the grayscale level and the pattern. For symbols in color graphs: the shape (twelve choices), the color, and whether or not the symbol is filled or not filled. For symbols in grayscale graphs: the shape, the grayscale level, and whether or not the symbol is filled or not filled.

Using Graph Style Sheets

To define your own styles, you need to create a text file containing the definitions (see below for details), and then read those definitions into rats using OPEN and GRPARM. You can create the text file with rats or any other text editor or word processor. Each style definition should appear on a separate line. You can define as many or as few of the styles as you want; if you dont redefine a style, it will just keep the previous settings. Youre most likely to want to redefine the black and white styles, since those will be used most often in publications. A line in the text file will take one of the following forms: LINE_COLOR_NN=pattern,color,thickness LINE_BW_NN=pattern,gray,thickness FILL_COLOR_NN=pattern,color FILL_BW_NN=pattern,gray SYMBOL_COLOR_NN=pattern,color,filled SYMBOL_BW_NN=pattern,gray,filled The first part of the definition specifies the type of representation you are defining (LINE, FILL, or SYMBOL). The second part (COLOR or BW) tells whether it is a color or black and white style.

Introduction to RATS

Int149

Graphics
The third part (NN) is the style number that youre defining. It should be between 0 and 30. Styles 1 through 30 can be selected by the user. Style 0 is reserved for shading performed using the SHADE option, so if you want to adjust the pattern or gray level used for shading, define the FILL_COLOR_0 and/or FILL_BW_0 styles. The arguments are as follows: pattern a number indicating the pattern choicesolid or dashed for lines, hatch pattern for fills, and symbol shape for symbols. See the next page for the available choices. is represented as a 24 bit (six digit) hexadecimal number. The first two hexadecimal digits are the level of red (00=no red to FF=red fully on), the next two are the level of green and the final two the level of blue. is a real number between 0 and 1 representing the degree of grayness (fraction of white). 0 means black, 1 means white. Note that its much easier to distinguish the lighter end of this (near 1) than the darker end: 0 and .25 look very similar, .90 and .95 look quite different. The default values for the first four black and white fills are solid black, solid .90 gray, solid .50 gray and solid .80 gray.

color

gray

thickness is a real-valued scale factor where 1.0 represents a standard line thickness. To make a line three times the standard thickness, use 3.0. filled is 0 for not filled (outline only) and 1 for filled.

Examples

The line graphs shown in Chapter 1 used a style set with these definitions: LINE_BW_01=0,0.00,1.0 LINE_BW_02=0,0.60,1.0 LINE_BW_03=0,0.80,1.0 Here, we are redefining the styles for the first three black and white line styles. The first value after the = selects line pattern 0 (a solid linesee next page) for all three. The second parameter sets the gray scale value. Here, we use solid black, 60% gray, and 80% gray. (Note again that in gray scale, a higher number is lighter). The third parameter sets the line thicknesswere using the default size of 1.0. Were only changing the black and white styles, so color versions of the graphs will use the default styles (solid lines, with black, blue, and green as the colors for the first three styles). But in black and white mode, the graphs will use solid lines in black, 60% gray, and 80% gray, respectively, rather than default of a solid line and two dashed line styles (all in black). We read these into rats from an external text file prior to generating the graphs using instructions like this: open styles graphstyles_00_60_80.txt grparm(import=styles)

Int150

Introduction to RATS

Graphics
If we wanted thicker lines, we could replace the 1.0 values above with 2.0 or 3.0 for lines that are twice or three times as thick.

Background Color

The default background color for the main graph box is white. You can redefine this color or pattern to be used by using the instruction GRPARM(BACKGROUND=stylenum)

Pattern Definitions

The available line, fill, and symbol choices are shown below. To select a line, fill, or symbol for a given style, use the number in the left-hand column as the pattern parameter. For example, SYMBOL_COLOR_2=1,FF0000,1 defines color symbol style number two as a red, filled square (1 being the pattern code for a symbol). Code Line Pattern Fill Pattern Symbol

0 1 2 3 4 5 6 7 8 9 10 11 12

Introduction to RATS

Int151

Graphics 3.17 Batch Graph Generation

Generating and saving graphs one at a time can be slow and error-prone if you are producing many of them, or if you have a standard program that you need to run on a regular basis. In such cases, you may prefer to have your program save your graphs automatically. There are two ways to do this, as described below. Because you arent processing each one manually, you cant switch them to over to black and white mode, if thats what you need. Instead, use the instruction GRPARM(PATTERNS) before you do any of the graphs. (You only need to do this once). This changes the default appearance of all types of graphs from color to black and white (patterns).

OPEN PLOT filename

If you issue an OPEN PLOT instruction, any subsequent graphs you generate will be saved to the specified rgf format file, until you either do a CLOSE PLOT instruction, or use another OPEN PLOT instruction to open a different graph file. This works in both interactive and batch modes, and offers the option of saving multiple graphs to a single file. There is no problem with putting several pages of graphics on a single file if you just intend to print them. However, if you are going to reopen the graphs later for viewing or for translation to another format, you should put only one graph on a file. If your program generates several graphs, you can either use a separate OPEN PLOT instruction for each graph as shown below, or use the ENVIRONMENT GSAVE instruction. open plot gdp.rgf graph(header="US Real GDP") # gdp90 open plot consump.rgf graph(header="Consumption of Durable Goods") # gcd

ENVIRONMENT GSAVE=template GFORMAT=format

The ENVIRONMENT instruction with the GSAVE=template parameter is your best choice for automatically saving multiple graphs into separate files (one graph per file), and for automatically saving files in formats other than rgf, such as PostScript. You just supply a filename template, and rats saves any subsequent graphs using that template plus a sequence number for the filename. The template will usually be a filename string that includes an asterisk (*) symbol. For each graph you generate, rats will construct a filename by replacing the asterisk with a sequence number, and save the graph using the constructed filename. Omit the asterisk if you want to save a single graph under a specific name.

Int152

Introduction to RATS

Graphics
By default, this saves graphs in rgf format. You can use the GFORMAT=format parameter to select other formats. Choices for the format parameter include: RGF, PORTRAIT, LANDSCAPE, WMF, and PDF, where PORTRAIT and LANDSCAPE are PostScript format, saved in portrait and landscape orientations, respectively (see page Int129). The availability of some formats is platform-dependent. The code below plots a series of impulse response graphs and saves them in a set of PostScript format files named ImpulseResp1.EPS, ImpulseResp2.EPS, and so on. environment gsave="ImpulseResp*.eps" gformat=portrait list ieqn = 1 to neqn do i=1,neqn graph(header=header(i),key=loleft,number=0) neqn cards impblk(i,ieqn) end do i

Printing Graphs Automatically

ENVIRONMENT PRINTGRAPHS

If you want rats to print graphs automatically, include the instruction

in your program (prior to the graphing instructions). This will spool the graphs to your default printer as they are generated.

Introduction to RATS

Int153

Graphics 3.18 Choosing Fonts

Font Handling
The FONT option on GRPARM (and GRTEXT) allows you to choose the font for the specified label(s). With the exceptions noted below, the fonts you select will be used when displaying, printing, and saving the graph. You can select from any font installed on your system (use the Fonts folder on the Windows Control Panel or Macintosh System folder, or a font-handling utility, to see a list of the installed fonts.) You must type the font name exactly (although case does not matter). On a Windows system for example, you might use FONT="Times Roman" or FONT="Arial". If you export a graph to PostScript from Windows, rats will automatically insert dashes between words in multi-word font names, as PostScript does not accept spaces in font names. It will also substitute standard PostScript fonts for their Windows counterparts (Helvetica for Arial, for example). If you want to change or add font substitutions, and are familiar with PostScript, you can edit the PROLOG.PST file (found in your rats directory). You can also supply exact PostScript names. rats may not be able to use those fonts when displaying the graph, but will use them if you export the graph to a PostScript file. Be sure to use the full PostScript name. For example, the regular version of Adobe Garamond would be AGaramondRegular rather than AGaramond. On the Macintosh, font names are generally the same for both display and PostScript devices. This means you can use the same font name for displaying output to the screen, printing on a PostScript printer, or exporting to a PostScript file. unix systems are generally similar to Macintosh systemsuse exact PostScript names for both display and printing purposes, as well as exporting to a file.

Examples

grparm(font="Symbol") axislabels 14 compute [vect[strings]] flabel=$ ||"0","p/6","p/3","p/2","2p/3","5p/6","p"|| scatter(style=line,vlabel="Coherence",vmin=0.0,vmax=1.0,$ xlabels=flabel) # freq coher The axis labels will be done in 14 point Symbol font. Although this choice affects both the horizontal and vertical axes, the numbers on the vertical axis will still be shown as numbers, since the Symbol font doesnt remap the numerical characters. However, the ps in the XLABELS option strings will be shown as ps. The seven character strings in FLABEL will be spaced equally (on center) from one end of the x-axis to the other.

Int154

Introduction to RATS

4.. Resources I

n addition to the information and examples presented here in the rats manuals, there are numerous other resources available that can help you get the most out of the program. We describe some of those here in this chapter.

Installing RATS Additional Documentation Examples and Procedures RATS Forum and Online Courses Technical Support

Resources 4.1 Installing RATS

For UNIX/Linux Users For Macintosh Users
Please see the unix rats installation guide included with your software.

To install rats on your system, just drag the macrats folder from the cd to your hard drive. If you do not want to install all of the files included with rats, first create a new folder for rats on your hard drive, then open the rats folder on the cd and copy over only the components you wish to install.

For Windows Users

To install rats on your system: 1. Insert the cd in your cd rom drive. 2. In most cases, the installation program still start automatically. If the installation program does not start automatically, you can run it manually by selecting the Run operation from the Start menu, typing in d:setup (where d is the drive containing the cd), and clicking on ok. File Locations By default, the core winrats application files will be installed in the subdirectory: C:\Program Files\Estima\WinraTS 8\ while all of the user files, including the example programs, procedures, and pdf documentation files will be installed in the rats user directory. By default, this will be a folder called WinRATS 8 in your Documents folder (My Documents on older versions of Windows). For example, under Windows 7, this would usually be: C:\Users\(username)\Documents\WinraTS 8\ which you can access using the Documents or My Documents shortcuts available in Windows. If you want to install the files in a different location, use the Custom button displayed on the Setup Types screen in the installer. You can control the locations of both the application files and the user files. For example, you may wish to install all of the user files in an easy-to-find location, such as C:\RATS The Custom installation button also allows you to choose which components you wish to install.

Int156

Introduction to RATS

Resources 4.2 Additional Documentation

PDF Files
rats ships with several documents in Adobe pdf format. You will find these files in the Manuals subdirectory of your rats user directory (see page Int156). Windows users will have shortcuts to these in the Winrats folder on your Start menu.

In addition to pdf versions of this Introduction and the Users Guide and Reference Manual, you will find the following files providing additional information: An Additional Topics manual, covering some additional statistical topics and features not included in the Introduction, Users Guide and Reference Manual Short files (Procedures and Examples, Paper Replication Programs, and Textbook Examples) listing the examples and procedures included with rats. You may also find a Supplement pdf describing features and changes made to the program since these manuals were written.

The Help System

In the Windows version, you can access the rats help system via the Help menu. Youll find quick references for all of the rats instructions, details on the various menus and dialog boxes, and other information on using rats. For Macintosh and unix/linux users, the help system is provided as a collection of html files. Double click on the rats Help shortcut icon in your rats folder to open the main page. From there, you can navigate to other help topics.

Introduction to RATS

Int157

Resources 4.3 Examples and Procedures

rats ships with nearly a thousand example programs. Referring to these can help

you learn the program, and they can also provide frameworks you can modify to suit your own needs. We also include hundreds of rats procedures that greatly extend the capabilities of the program, allowing you to implement complex tasks with a single procedure call. See page Int32 for details. You will find the main set of examples and procedures located in your rats user directory (page Int156). Programs and data sets for replicating the examples from many popular econometric textbooks are provided in the TextbookExamples subdirectory. Similarly, the PaperReplicationExamples subdirectory offers program, data, and procedure files for replicating the results from many important papers. You may want to consult the pdf files listing the available examples and procedures to help locate those that may be of interest

Using the Website

All of the examples and procedures shipped with rats are also available on our website, along with dozens of others written by rats users around the world. You may find that the Procedure Browser and Search tools on the web site are the easiest ways to locate files of interest. You can find links to the examples and procedures by going to: www.estima.com and clicking on the Resources tab, then on Procedures/Examples. The Textbook Examples link takes you to lists of the example programs for the various textbooks, grouped by book. The Procedure Browser link takes you to a very handy tool for searching through the available files (excluding the textbook examples). You can also get to the Procedure Browser directly using this address: www.estima.com/cgi-bin/procbrowser.cgi The Browser allows you to search for files related to a particular subject area, or that reference a particular authors work or techniques. You can also search for recently added files, or limit the search to files that work with specific (older) versions of rats. Alternatively, you can use the Search tab on the site to do a general keyword search. This is particularly handy for locating relevant textbook examples (as these are not covered by the Procedure Browser) and applicable posts on our web forum.

Int158

Introduction to RATS

Resources 4.4 RATS Forum and Online Courses

The RATS Software Forum
www.estima.com/forum Here, registered rats users can exchange ideas, ask statistical and programming questions, share example code and procedures, and get answers to technical questions from Estima staff. We also frequently post new procedures and examples on the forum. Please note that only users with legal, licensed copies of rats are allowed to register as users on the forum. You dont have to register to view topics and posts, but registration is required to post any messages on the forum. If you ask a technical support question on the forum, please do not email the same question to Estimas support email address, as this will lead to an unnecessary duplication of effort on our part. We are happy to provide support through either medium, but please choose one or the other. We host a discussion forum on our website, at:

Online Courses and Course Materials

www.estima.com/courses.shtml for details on any upcoming courses.

Estima also offers online courses delivered via private sections of the user forum on our website. Please see the

We also offer for sale the packages of course materials from previous courses. These packages include the pdf handbooks providing the instruction materials along with example programs and procedures developed for the course. Again, see the website for details on course materials.

Introduction to RATS

Int159

Resources 4.5 Technical Support

Support is not cheap, but we think it is an important part of what you have paid for. The next three pages describe the procedures for obtaining support and the level of support which you can expect from us.

Your Serial Number

Your serial number will be printed on a label on the cd jacket. (Note that classroom rats doesnt include this). This serial number is your key to updates and technical help. You can use the Preferences dialog (on the File menu in Windows or unix, or the RATS menu on the Mac) to save your serial number in the registry for later reference. You may also want to write the serial number in your manual, or somewhere else handy where you can find it. If you are a new purchaser, please fill out the registration/address change form available on our website. This is especially important if you bought the program through a reseller or through your purchasing department as we will be unable to send you newsletters and other notices without an address. When contacting technical support, please supply your serial number so we can verify which version of rats you are using.

Contacting Estima Technical Support

If you have access to e-mail, this is probably the best way for you to obtain technical support. You can provide us with a copy of your program and data set, and other very specific information about the nature of your question. This makes it much easier for us to resolve the problem or question, and to provide detailed answers. We try to respond to questions within one business day. Please be sure to use a descriptive subject line on your email. A message with a subject like "Please help" is all too likely to end up in a spam-filter folder. You can also contact us by calling the technical support number listed below. Technical support is available from Monday through Friday, 9 am to 5 pm US Central time. When you contact technical support please be able to provide the following: 1. Your serial number (found on the label on the cd envelope). We do spot checks to make sure that only those who have paid for the product get support. 2. The product name and version number (also from the original disk). 3. As much detail about your question as possible. In particular, if you are getting an error message you dont understand, be sure to include that error message in your e-mail, or have it handy to read to us over the phone. rats produces fairly specific and detailed error messages, so providing the specifics of any errors you encounter will make it much easier for us to diagnose the problem. As noted earlier, please dont submit same technical questions via both email and the forumchoose one medium or the other.

Int160

Introduction to RATS

Resources
You can contact technical support at: Voice: Email: Web site: FAX: Mail: (847) 864-1910 [email protected] www.estima.com (847) 864-6221 Estima 1560 Sherman Ave, Suite 510 Evanston, IL 60201

If you write or FAX us with a question, please remember to include your name and a phone or FAX number.

Can RATS Do ...?

Many of these questions can be answered by checking the index and the table of contents in the manuals carefully, or by visiting our web site (www.estima.com), where your question may be addressed by the list of Frequently Asked Questions or by one of the many procedures and example programs available for downloading. If you cant find the answer there, you can e-mail, fax, or call our technical support department with your question. The answers to these sorts of questions generally fall into four categories: 1. Yes (and you should have known it by looking at the index or table of contents). Please be sure to check these first. 2. Yes, you can use the instruction ... 3. Yes, but it takes a little work (a short sequence of instructions). 4. No, or at least it would be very difficult and require extensive programming. Many of the suggestions made by users have helped to improve our product, so we try to be as helpful as we can. With types 2 and 3, we will usually tell you exactly how to do it. However, with type 4, we can only give you a general idea of what you have to do if you decide to press on.

Statistical Questions

rats has many capabilities which may be unfamiliar to some of you. If you decide

to explore some new territory, we will be happy to steer you to some good references or to explain how rats does particular computations. However, while we can help to clear up basic misunderstandings about the use of the rats instructions, we cannot give involved statistics lessons over the phone or email. If you are interested in discovering new techniques, watch the newsletter for information regarding upcoming courses.

Introduction to RATS

Int161

Resources
Bugs and Potential Bugs
In a program as complex as rats, there are undoubtedly some bugs remaining. In addition, because rats has many features of a programming language, it is quite possible for you to experience problems due to errors in your own code. The more complex your program, the more likely it is that the latter is true. If your program is not running correctly, you should do the following:

Check carefully that you are using the proper syntax on the instruction(s) causing the problem. See the Reference Manual in particular. If you are doing extensive operations with loops and COMPUTE instructions, put in some debugging statements (DISPLAY and PRINT are the most useful for this) to see where things go awry. Check the list of frequently asked questions, and the list of known bugs, available on our web site (www.estima.com). If, after all this, you have a strong suspicion that you have located a bug, contact Technical Support. If you have done a thorough job on the preceding steps, you can often ask a direct question such as Is there a known bug in ....?, and we may be able to give you a quick answer. If you have not been able to isolate the problem, we will almost certainly have to ask you to send us the input file and data and as much other information as you can supply.

The RATSletter

The RATSLetter is a newsletter for registered users. We distribute it (approximately) twice a year. It includes new product announcements, answers to common questions, bug reports, tips on the use of the program, lists of contributed procedures, among others. We would appreciate questions of general interest and suggestions. If you have a program or part of one which you are proud of, send it in and let the rest of the rats community see it.

Updates

We typically have a major update every two to three years, where we change the main version number and redo the documentation. Between these we have several minor updates, where we add features to the program, revise procedures, add new textbooks and paper replications, etc. Many of the most important features get added first in these minor updates since thats when we do most of the programming. If you want to stay up-to-date automatically, we have update subscription programs where you pay in advance to receive the updates at one fixed rate. See www.estima.com/updatesubs.shtml

www.estima.com

Our web site offers news and information on rats and other Estima products, answers to frequently asked questions, and many examples and procedures you can download.

Int162

Introduction to RATS

Bibliography
Diebold, F.X. (2004). Elements of Forecasting, 3rd Edition. Cincinnati: South-Western Greene, W.H. (2008). Econometric Analysis, 6th Edition. New Jersey: Prentice Hall. Hill, R.C., W.E. Griffiths, and G.C. Lim (2008). Principles of Econometrics, 3rd Edition. New York: Wiley. Makridakis, S., S.C. Wheelwright, and R.J. Hyndman (1998). Forecasting, Methods and Applications, 3rd Edition. Hoboken: Wiley. Pindyck, R. and D. Rubinfeld (1998). Econometric Models and Economic Forecasts, 4th Edition. New York: McGraw-Hill. Verbeek, M. (2008). A Guide to Modern Econometrics, 3rd Edition. New York: Wiley.

Introduction to RATS

Int163

Index
Symbols
operator, Int40. = operator, Int40. ; multiple statements per line, Int57. .AND. operator, Int41. .EQ. operator, Int41. .GE. operator, Int41. .GT. operator, Int41. .LE. operator, Int41. .LT. operator, Int41. .NE. operator, Int41. .NOT. operator, Int41. .OR. operator, Int41. {..} for lags in expressions, Int25. in regressor lists, Int78. @procedure, Int33. * for comments, Int73. for default entry, Int31. operator, Int40. ** operator, Int40. *= operator, Int40. / for default entries, Int31. operator, Int40. for integers, Int42. /= operator, Int40. \\ for line breaks, Int132. + operator, Int40. += operator, Int40. < operator, Int41. <= operator, Int41. <> operator, Int41. == operator, Int41. > operator, Int41. >= operator, Int41. || for in-line matrices, Int90. $ line continuation, Int57, Int75. %IF function, Int41, Int62, Int107. used with SET, Int63. %L function, Int132. %NA missing value, Int40. %PI constant, Int40. %RESIDS series, Int75. %ROUND function, Int42. %VALID function, Int41, Int63.

AR1 instruction, Int77. ARIMA models, Int64. Arithmetic expressions, Int40. Arrays, Int90. Autocorrelations graphing, Int145.

Batch mode graphs in, Int127. Benchmarking, Int62. Binary data, Int123.

CALENDAR instruction, Int21. Calendar wizard, Int8. CATALOG instruction, Int111. Comments, Int73. COMPUTE instruction, Int6. for matrices, Int90. Concatenation of strings, Int132. Conditional blocks, Int91. Constants, Int40. CONSTANT series, Int27. Continuation lines, Int57, Int75. COPY instruction, Int96, Int111. with spreadsheets, Int119. Cross sectional data, Int66. CRSP format, Int98. CUSUM test example, Int138. CVMODEL instruction, Int89.

Filter/Smooth, Int44. Graph, Int46. Scatter (X-Y) Graph, Int71. Transformations, Int26. Trend/Seasonals/Dummies, Int53. DATA instruction, Int22, Int96. changing data frequencies, Int 103. with holidays omitted, Int108. spreadsheet files, Int116. Data wizards, Int18. (Other Formats), Int19, Int43, Int103. (RATS Format), Int84, Int110. Dates range of, Int30, Int81. referring to, Int21. on spreadsheet files, Int114, Int119. DDV instruction, Int78. DEDIT instruction, Int111. DEFINE option, Int64. DELETE instruction, Int111. DIEBP228.RPF example, Int138. DIEBP235.RPF example, Int138. Differencing, Int62. Differencing wizard, Int26. DIF files, Int114. @DISAGGREGATE procedure, Int105. DISPLAY instruction, Int4. for matrices, Int90. DLM instruction, Int89. DOFOR instruction, Int91. DO instruction, Int91. Dummy variables, Int60, Int61. Durbin-Watson statistic, Int37. Dynamic Linear Models, Int89.

Data changing frequency of, Int104. cross-section, Int66. detrending, Int48. reading from a file, Int18, Int96. subset of, Int67. transformations, Int25. writing to a file, Int96. Data/Graphics menu Calendar, Int8. Create Series (Data Editor), Int9. Data Browsers, Int98. Data (Other Formats), Int18. Data (RATS Format), Int84.

EDIT instruction, Int111. Edit menu Comment-Out Lines, Int74. Copy as TeX, Int16. Format Comments, Int74. Show Last Error, Int80. To Lower Case, Int28. Uncomment Lines, Int74. ELSE instruction, Int92. Entry ranges, Int30, Int81. Equation/FRML wizard, Int86. Equations, Int64. Error messages, Int79. ESTIMATE instruction, Int78.

Introduction to RATS

Int165

Index
EViews work files, Int99. EXAMPLEFIVE.RPF example, Int66. EXAMPLEFOUR.RPF example, Int43. Example programs, Int158. EXAMPLESIX.RPF example, Int 84. EXAMPLETHREE.RPF example, Int17. Excel files, Int99, Int114. Exponential smoothing, Int48. Exponential Smoothing wizard, Int48. Expressions, Int40. logical, Int41. ODBC, Int99. PORTABLE, Int98. PRN, Int114. RATS, Int110. TSD, Int114. WF1, Int99. WKS, Int114. XLS, Int114. XLSX, Int114. Formulas, Int64. creating from regression, Int 64. FRED database, Int98. Frequency of data series changing, Int103. differing, Int34. FRML instruction, Int85. Functions using, Int41. window, Int56. wizard, Int46. Graph window, Int56, Int127. Graph wizard, Int46. Growth rates, Int62. GRPARM instruction, Int126. GRTEXT instruction, Int126.

Haver Analytics data, Int95. Holidays on data file, Int108. Hypothesis tests results of, Int11.

Fame format, Int98. Fan chart, Int149. File menu Clear Memory, Int18. Directory, Int34. New Editor/Text Window, Int 3, Int18. Files binary, Int123. formats, Int22. organization of data on, Int22. reading strings from, Int132. spaces in names, Int34. spreadsheet, Int114. Filtering, Int44. Filter/Smooth wizard, Int26, Int48. FIND instruction, Int89. Fitted values, Int64. for non-linear regression, Int 88. FIX function, Int42. FLOAT function, Int42. Fonts for graphs, Int133. Forecasts, Int53, Int64. graphing, Int147. static, Int64. FORMAT option, Int22. CDF, Int114. CITIBASE, Int98. CRSP, Int98. DIF, Int114. DTA, Int99. FAME, Int98. FREE, Int120. HAVER, Int98. MATLAB, Int100.

GARCH instruction, Int89. GCONTOUR.RPF example, Int 143. Goodness of fit, Int30. GRAPHBOXPLOT.RPF example, Int144. GRAPHFORECAST.RPF example, Int147. GRAPHFUNCTION.RPF example, Int140. GRAPHHIGHLOW.RPF example, Int146. GRAPH instruction, Int126. GRAPHOVERLAY.RPF example, Int135. Graphs autocorrelations, Int145. background, Int151. copying and pasting, Int129. displaying, Int46, Int127. exporting, Int129. fan charts, Int149. fonts on, Int154. highlighting entries, Int141. high-low-close, Int146. instructions for, Int126. keys/legends, Int132. labeling, Int131. orientation of, Int130. overlay/two-scale, Int134. patterns, Int151. preparing for publication, Int 129. scatter (X-Y) plots, Int71. shading, Int141.

IF instruction, Int92. INCLUDE instruction, Int111. Input window, Int7. Instructions long (splitting up), Int57, Int75. multiple per line, Int57. INSTRUMENTS instruction, Int 77. Integer numbers, Int42. Interpolating data, Int105. Interrupt program, Int7. I/O units tips on, Int123.

J K L
Labels on data files, Int115. for graphs, Int131. Landscape mode, Int130. LDV instruction, Int78. Least squares, Int27. Limited/Discrete Dependent Variables wizard, Int78. LIML. See Limited information maximum likelihood. Linear regressions, Int27. annotated output, Int35. Linear Regressions wizard, Int 35. LINREG instruction, Int28.

Int166

Introduction to RATS

Index
Logical operators, Int41. creating dummies using, Int61. LOOP instruction, Int91. LQPROG instruction, Int89.

Q R

QUIT instruction, Int111.

Marginal significance levels, Int 11. MATLAB data files, Int99. Missing values, Int62, Int63. on data files, Int107. in expressions, Int42, Int63. %NA constant, Int40. Moving average, Int25. MWHP366.RPF example, Int142.

Neural networks, Int89. NLLS instruction, Int87. NNLEARN instruction, Int89. NNTEST instruction, Int89. Non-linear estimation, Int84. initial values, Int86. NONLIN instruction, Int85. Normalizing series, Int62. NPREG instruction, Int89.

OECD MEI data, Int95. OPEN instruction, Int34, Int96. Operators precedence, Int40. ORGANIZATION option, Int22. with free-format, Int121. Output window, Int7. Overlay graphs, Int134.

RATS Data File window, Int112. RATSDATA program, Int113. RATS format files instructions for, Int111. older versions of, Int113. RATS forum, Int159. Ready/Local button, Int7. Recursive Least Squares wizard, Int78. Recursive residuals graphing example, Int138. @REGACTFIT procedure, Int32. Regression format, Int78. Regressions output annotated, Int35. Regression Tests wizard, Int69. Relational operators, Int41. RENAME instruction, Int111. Report window, Int11, Int15, Int29. Residuals, Int75. Robust regression, Int77. ROBUST option, Int75. Rounding, Int5. RPF file type, Int58. RREG instruction, Int77. R-squared statistics, Int30. Run button, Int7.

Panel Data Regressions wizard, Int78. Picture codes, Int5. Portrait mode, Int130. Precedence of operators, Int40. PRG file type, Int58. PRINT instruction, Int59. PRJ instruction, Int64. forecasting with, Int64. PRN files, Int114. Procedures, Int32. PRTDATA instruction, Int111.

SAVE instruction, Int111. SCATTER instruction, Int71, Int126. Scatter plots, Int71. Scatter (X-Y) Graph wizard, Int 71. Scientific notation, Int40. Seasonal adjustment, Int52. dummies, Int61. Serial correlation tests for, Int37. Serial number, Int160. Series, Int9. compacting frequency, Int104. creating, Int9, Int25. dummy variables, Int61. editing, Int13. expanding frequency, Int105.

frequency of, Int103. missing values, Int62, Int107. names, Int22. print, Int59. reading from a file, Int96. transformations, Int25. Series Edit window, Int9, Int13. Series window, Int23, Int38. SET instruction, Int25. Simulations, Int65. Single-Equation Forecasts wizard, Int54. Smoothing, Int44. exponential, Int48. SMPL instruction, Int81. SMPL option, Int67. SPGRAPH instruction, Int126. SPGRAPH.RPF example, Int136. Spreadsheets adding dates to files, Int119. reading data from, Int114. Stata data files, Int99. STATISTICS instruction, Int24. Statistics menu Limited/Discrete Dependent Variables, Int78. Panel Data Regressions, Int 78. Recursive Least Squares, Int 78. Regressions, Int27. Regression Tests, Int69. Univariate Statistics, Int24. Stepwise regression, Int77. Stop button, Int7. STORE instruction, Int111. Strings concatenating, Int132. line breaks in, Int132. reading from files, Int132. STWISE instruction, Int77. Style Sheets graph, Int149. Sum of coefficients, Int77. Supplementary cards, Int28. SUR instruction, Int78. Syntax errors, Int79.

TABLE instruction, Int23. Technical support, Int160. TeX copy to, Int16. Textbook examples, Int158.

Introduction to RATS

Int167

Index
Time Series menu Exponential Smoothing, Int48. Single-Equation Forecasts, Int 54. Toolbar icons series edit window, Int14. series window, Int38. Transformations wizard, Int26. Trend/Seasonals/Dummies wizard, Int26, Int53, Int61. Trend series, Int53, Int60. Linear Regressions, Int35. Panel Data Regressions, Int 78. Recursive Least Squares, Int 78. Regression Tests, Int69. Scatter (X-Y) Graph, Int71. Single-Equation Forecasts, Int 54. Transformations, Int26. Trend/Seasonals/Dummies, Int26, Int53. Univariate Statistics, Int24. WKS files, Int114.

UFORECAST instruction, Int55. UNTIL instruction, Int91. UPDATE instruction, Int111.

X Y Z

View menu, Int10. Change Layout Report Window, Int15. Series Edit Windows, Int13. Data Table, Int25. Series Window, Int12, Int38. Statistics, Int10, Int23. Time Series Graph, Int10.

WHILE instruction, Int91. Window, Int13. change layout, Int13. Graph, Int56, Int127. Input, Int7. Output, Int7. RATS Data File, Int112. Report, Int11. Series, Int23. toolbar, Int38. Series Edit, Int9. Window menu Close All Graphs, Int56. Report Windows, Int29. Use for Input, Int7. Use for Output, Int7. Wizards Data (Other Formats), Int18, Int103. Data (RATS Format), Int110. Differencing, Int26. Equation/FRML, Int86. Exponential Smoothing, Int48. Filter/Smooth, Int26, Int48. Graph, Int46. Limited/Discrete Dependent Variables, Int78.

Int168

Introduction to RATS