INTRODUCTION TO EVIEWS

You can find EVIEWS 4.1 on the Manchester Computer Clusters as follows All Programs Programs Faculty FSSL Econometric & Statistical This document will give a general introduction into the workings of EVIEWS. It does not work through a particular example, but you can use the data which are provided in the EVIEWS Tutorial data.xls spreadsheet to practice the basic techniques. There is an additional document eviews tutorial.doc which gives you a few problems to work through with these data.

Notation and miscellaneous

EVIEWS commands or menu commands are printed in the courier type. LMC - left mouse click RMC - right mouse click DMC - double click Workfiles are the workhorses of EVIEWS. They store your data and results of your analysis. Each workfile can contain a big number of data sets. When you create a new file (FILE - NEW) you only have to open a new workfile. You will not need databases, programs or text files. As you open a new workfile you will be asked about the nature of the data. If you have data observed at regular intervals you can select the appropriate frequency and the start and end dates. Often, however, it will be convenient to choose Unstructured / Undated as your Workfile structure type, especially if you have cross sectional data. You then merely have to indicate how many observations you have (in the End Date field).

/opt/scribd/conversion/tmp/scratch679/68953151.doc

Page 1 of 5

While each workfile will contain several data series, each data series is stored as its own object. Each workfile comes with a default data series "resid" and a coefficient vector c, which, at any time, will store the residuals and coefficient estimates of the regression last run (the or in some versions , are an indication from EVIEWS that this object is a parameter vector, whereas the small stylised line plot in the symbol to resid indicates a series of data) . When you create or import new data series, they will appear as new data series, just as "resid" does. Double-clicking on a data series opens a window where you van view (spreadsheet, statistics and graphics) the data series. If you want to perform an operation, which involves more than one data series, you got to highlight the data series in question (using the shift or control key) and then DMC. Choose open group if you want to display the time series or open equation if you want to run a regression involving the time series. When the operation you want to perform with the chosen set of variables is finished, close the group again. Note, that this does not delete the individual variables. Well get back to this once we imported a few more data series. The white area beneath the command menu is called the command line. Some actions can be performed by typing a short command in this line, which can be quicker than using the menus. The more you familiarise yourself with EVIEWS the more often you will use these commands.

Importing Data
Two alternative strategies exist to import data. Both strategies will be described in turn. From Excel Have the Excel file ready with data in columns. Memorise the first data cell (eg A2), the name of the data sheet (if multiple sheets) and the number of variables to be imported. In EVIEWS: PROCS - IMPORT - READ TEXT LOTUS EXCEL Then choose file (which has to be closed in Excel) Confirm available information In the "Names of series " field either enter the requested names of the data series or just the number of data series to be imported. In the latter case EVIEWs will assign names automatically. OK Copy and Paste Copy data into clipboard from spreadsheet or text editor In EVIEWS: QUICK - EMPTY GROUP (EDIT SERIES) LMC on upper left cell RMC and PASTE Instead of pasting data into the spreadsheet you can also just type data into the spreadsheet. On the top of the window you can see an Edit+/- button. This locks or unlocks the spreadsheet for editing. EVIEWS will give the new series names like SER01 and unfortunately you wont be able to change the names right here. The following procedure will enable you to rename the series to more sensible names.

/opt/scribd/conversion/tmp/scratch679/68953151.doc

Page 2 of 5

Close the group by clicking on the X in the top right corner of the Group window and confirm that you want to delete this group (recall that this will not delete the data themselves!). You will then see the new data series icons in your worksheet file. RMC on the data series for which you want to change the name and choose the Rename option. Give the data series a sensible name. If you have plenty of data series, you should put some thought into how to name the series such that you remember what is what. OK

Creating new data series

Often you will have to create new data series. Imagine you have a time series of share prices "p", which you want to transform into a time series of log returns. QUICK - GENERATE SERIES Enter the requested equation. To calculate the returns of "p": rp = log(p) - log(p(-1)) OK "rp" is now the time series of returns. Note that "p" = pt whereas "p(-1)" = pt-1. To perform this operation you can also use the command window. As you improve your EVIEWS skills you will appreciate that the command line often offers a quick way to achieve what you want. Type into the command line genr rp = log(p) - log(p(-1)) or series rp = log(p) - log(p(-1)) ENTER Genr and series is the command which makes EVIEWS generate a new data series. After you imported your first data series, you should save the entire workfile, which is accomplished in the usual Microsoft manner by choosing the appropriate option in the File dropdown menu. Workfiles have the standard file extension wf1.

/opt/scribd/conversion/tmp/scratch679/68953151.doc

Page 3 of 5

Displaying and editing data series

You can view the data series as spreadsheets (and edit them if required) or as a variety of graphics. Select the data series (one or several) you want to view. DMC and open group. You will see the spreadsheet with your data. On the top of the window with your data series you can see several buttons. The most important ones are the View, Name and the Edit+/- button. Should you want to edit the data you have to press the Edit+/- button before editing them. Renaming your data series, you use the Name button. When clicking on the View button a list of options appears. Take a view minutes to explore them. The most useful are the Line Graph, Descriptive Statistics and Correlogram options.

Regression
There are several ways to run a regression. Assume you have two variables "y" and "x" in your workfile and you want to regress "y" on "x", ie. you want to estimate the parameters and in yt = + xt + t. Select "y" and "x". DMC and Open equation In the equation specification window you will see "y x c". "y" and "x" are the variables you selected. EVIEWS automatically chucks in the constant "c". The convention is, that the first variable is the dependent variable and all following variables are the independent variables. If you want to estimate without a constant term you just delete the "c". If you want to add further variables you type them into the window. If you want to add xt-1 you simply have to add "x(-1)" without the need to first create this time series. OK You will receive a standard regression output, which you should be able to interpret. In any doubt, consult your Data Analysis notes or a standard statistics textbook.. Again you can perform this operation from the command line ls y c x enter and you receive the same output. The View button on the top left of the Equation window now gives you a list of options in order to visualise your regression results and to perform a variety of tests. If you want to re-estimate the equation with a slightly different specification, choose the Estimate button. Sometimes you will need the regression residuals for further manipulation and you will want to store them. Recall that the "resid" time series always contains the residuals of the last regression. If you want to store the residuals, use the generate data option. Eg. genr res1 = resid The residuals will be stored in "res1". While the resid series will change its values as you run a new regression, the res1 series will now remain unchanged. Also note that the object in your workfile will always store the parameters of the last regression you run, much like the resid data series object. As with the residuals, if you want to store the parameter values you need to create a new coefficient vector by typing the following command in the command line: coef c_save = c, which saves a new coefficient object named c_save. You might want to save a particular specification and its results. You can do this by using the Name button. You are asked for a name of the equation you want to store. Once you did this your equation will appear as an object in the workfile. /opt/scribd/conversion/tmp/scratch679/68953151.doc Page 4 of 5

When you close a regression window you will be asked whether you want to delete the regression. Recall that this will merely delete the regression window but not the data series themselves. Hence you can safely answer yes to this, unless you want to save the entire regression specification (see previous paragraph).

Useful EVIEWS commands

log(x) x(-1) x(-2) d(x) scalar a = 21.3 scalar b =3^3 genr a = b*b genr dy = y - y(-1) calculates the natural logarithm of x refers to x lagged by one period refers to x lagged by two periods calculates the difference, x x(-1) sets a to a particular value sets b to 3*3*3 = 27 generates a series, here the square of b sets dy to the differenced series of y

/opt/scribd/conversion/tmp/scratch679/68953151.doc

Page 5 of 5