EAST CAROLINA UNIVERSITY Department of Economics Economics 6390 Research Some Eviews Notes for Assignment #3 Let’s see

what needs to be done in order to produce the empirical results you will need. 1: Create a Workfile It would be useful to look up ‘Workfile Basics’ in the HELP menu of EViews. An EViews ‘Workfile’ is a file which stores all of the information (data, graphs, estimation results, etc.) that EViews needs/produces for your particular project. To get started you must first create a ‘Workfile’. To do this: 1. 2. Click the ‘File’ button in the top-left corner of the EViews toolbar and click on ‘New’ and then ‘Workfile’. At this point you’ll need to supply EViews with some information. In particular, you need to indicate: (a) the workfile ‘frequency’ (for the “housing starts” data, the ‘frequency’ is ‘monthly’). (b) the ‘start date’, i.e., the date of the initial observation. (for the “housing starts” data, the ‘start date’ is ‘1946.01'). (c) the ‘end date’, i.e., the date of the last observation (for the “housing starts” data, the ‘end date’ is ‘1994.11'). -Once you’ve supplied (a), (b), and (c), click ‘OK’. Now your workfile is created. 2: Save your Workfile Since things can go wrong at any time while working on your computer, it is crucial to remember to ‘save’ your work at various points. I tend to be very risk averse in this matter and often ‘save’ my files (in EViews, my word processor, etc.). It is very frustrating to work through a long session with the program and end up losing all of your results because of a failure to ‘save’ your file. So, once you’ve created your workfile, I suggest that you immediately save it. Since, in the process of doing the work for this course, you will end up with a large number of EViews files, I strongly suggest that create a special directory/folder to store all of your EViews files (the raw data files, the workfiles, and anything else). For the purposes of this demonstration, let’s assume that you have the EViews program files stored in the following directory: c:\eviews Then, I suggest that you create a folder/subdirectory on ‘c:\eviews’ this directory to store everything. Let’s assume you call this subdirectory ‘data’. In this case the ‘path’ for the subdirectory would be: c:\eviews\data To make this your ‘default’ directory (to which all ‘saved’ items will be, bey default, saved), issue the following command: cd c:\eviews\data Then, assuming this subdirectory has been created and you have made it your default directory, you can now save your workfile into it by clicking ‘File’ and ‘Save’ and indicating the file type (the default will be ‘workfile’). Dr. Philip Rothman

and 0 otherwise.txt’ filename extensions as ascii/text files. A window with the title ‘Open’ will then appear. in this window you should enter the number ‘1'. This will open up a ‘spreadsheet view’ of the data.g. iii. Lotus. from Figure 6. an Excel file. and then ‘Text-Lotus-Excel’. Actually. There are three steps: 1.3: Import the Data (Not Necessary if Data Have Been Supplied to You in an EViews Workfile) After you have created and saved your workfile. January) observations. to the right of this window click on ‘Reset sample to: Workfile range’ button. 4: Estimate a Model with Seasonal Dummies The key thing you need to do is estimate the required OLS regressions. and then click ‘OK’. ‘1946:01 1994:11’. to . Let’s assume that you’ve stored a copy of this file in the following directory: c:\eviews\data 2. In the ‘Sample to Import’ window enter. This tells EViews that you want to read in data from either a text/ascii.dat” there is just one series and the first row of the file contains the name of the series. To see such a time series plot simply click on the variable called ‘hstarts’. For example.e. since the data are ordered ‘in Columns’. You are now ready to start analyzing the data. You can download an ascii/text file with the “housing starts” data from the web page for this assignment.. Then click ‘OK’. iv. Within EViews. (b) EViews automatically recognizes with both ‘. Note: to have this done automatically for you. so that the entire time series is read in.. Create the ‘seasonal dummy variables’. ii. e.data’ for this example). Similarly. Note that the name of this file is “hstarts.dat”. So. or Excel file. In the top-left window the ‘Names of series or Number of series .’ In “hstarts. So. to create a dummy variable called ‘d1’that takes on the value 1 for all first month (i. Next. you need to: 1. To see a time series plot. click ‘View’ and then ‘line graph’. You now have to indicate: i.. But before you do this you might want to see what the “housing starts” data look like. let’s see what’s necessary in order to obtain the regression results reported in Table 6. 3. at this point an ‘Ascii Text Import’ window opens up. click on the file. This can be done with the EViews command @SEAS. Leave the ‘Data Order’ button unchanged. which simply means that each row of the file contains an additional observation. which presents the data in separate cells as in. There are two things you need to do at this point: (a) Indicate where the file is located (‘c:\eviews.4 of the text you already know what they look like. issue the following command in the EViews command window: genr d1 = @seas(1) Note: it’s necessary to hit the ‘Enter’ key to have EViews run this and any other command. then ‘Import’.1. There are several other options on the ‘View’ menu that you might want to check out. I suggest that you once again ‘save’ your workfile. click on ‘File’.dat’ and ‘. But before you move onto this step. Obtain the file with the data you need..

e. 106. if we were to include a ‘constant’ term. To run a regression that would be satisfy the requirements of Problem #5 on p. . (b) From the toolbar at the top of the main EViews window. The first variable listed after ‘ls’ is the dependent variable in the regression. issue the following command: ls hstarts c d1 d2 d3 d4 d5 d6 d7 d8 d9 d10 d11 In this example a ‘constant’ term has been added to the model and the ‘d12’ seasonal dummy variable has been dropped. on p. This is necessary since the workfile sample range runs from 1946:01 to 1994:11.. Two ways you can do this are: (a) From the EViews command window issue the following command: ls hstarts d1 d2 d3 d4 d5 d6 d7 d8 d9 d10 d11 d12 The ‘ls’ command instructs EViews to run an OLS regression. 109. Set the current sample range to 1946:01 to 1993:12.1 and generate all of the regression results needed to analyze the all of the homework problems dealing with the “housing starts” data.. 106 in the text. To do this enter the following on the EViews command window: smpl 1946:01 1993:12 3. if you want to see what happens if all twelve seasonal dummies are included and a constant term is added to the model. and your computer will scream at you if you run such a regression’”. The system variable EViews uses for a regression ‘constant’ term is called ‘c’. February) observations and 0 otherwise.. This will open up an ‘Equation Specification’ window in which you can list the variables to be used in the regression (the first variable name in the list must be the dependent variable) and indicate the sample range to be used. This is what the author means when he writes. So. issue the following commands in the EViews command window: genr d2 = @seas(2) Using the @SEAS command you should be able to create all of the dummy variables you need for this homework assignment. Recalling the discussion on the bottom paragraph of p. All of the other variables listed are the independent variables in the regression. issue the following command: ls hstarts c d1 d2 d3 d4 d5 d6 d7 d8 d9 d10 d11 d12 You should get the ‘Near singular matrix’ error message. we would have to exclude one of the seasonal dummy variables. click ‘Quick’ and then ‘Estimate Equation’. 2. Note: in this regression we have excluded a ‘constant’ term.create a dummy variable called ‘d2’ that takes on the value 1 for all second month (i. With the information given above you should be able to replicate the results reported in Table 6. “. Estimate the model by OLS. which indicates that the specified model can not be estimated.