Questions

Notice : if you work with cross section data instead of time series, choose the undated frequency and give a number to each country (if we are talking about countries) or industriy or whatever it is.

Afteropening your workfile with File Open Workfile name, use the Genr command by writing an expression in the command line just below the main menu. For instance, you will type Genr NewSerie=(Serie1*Serie2)/Serie3. To deflate a nominal GDP serie with a price index, you could write : Genr RealGDP=NomGDP/Price. The operators of expressions are the classical +, -, *, / and a few more described in the help of the software. The new serie will appear next to the others on the workfile.

Open your workfile by selecting File Open Workfile Name and type Plot in the command line. You have to specify the name of the serie to plot. For instance, Plot RealGDP. The graph comes to screen in a reduced window. Select the window on the whole screen. Click on the Options button of the graph menu and choose the graph options with the new dialog box. After closing the graph window you can save your graph by choosing a name in the dialog box that appears at screen. Your graph will appear as a graphic object next to the others series objects in your workfile.

After opening your workfile, type Smpl first period last period in the command line just below the main menu. For instance, if you have a sample between 1950 and 1995, you could type Smpl 1970 1990 to change the sample range.

After opening your workfile, type in the command line just below the main menu :
LS Dependent Variable Name C Independent Variables Names separated by a space between the names. For instance, if we regress consumption on yield after creating the consumption and yield series, we have to write the following instruction : LS Consumption C Yield. The following table is supplied to you by the software :

LS // Dependent Variable is CONSUMPTION                Sample: 1948 1994  Included observations: 47  Excluded observations: 0 after adjusting endpoints

Variable   C YIELD

Coefficient 6086.105 0.861840

Std. Error 1045.405 0.010610

T-Statistic 5.821767 81.22654

Prob.T-Stat 0.0000 0.0000

R-squared Adjusted R-squared S.E. of regression Sum squared resid Log likelihood Durbin-Watson stat

0.993226 0.993075 2518.732 2.85E+08 -433.749 0.211522

Mean dependent var S.D. dependent var Akaike info criterion Schwartz criterion F-statistic Prob(F-statistic)

85584.15 30267.53 15.70464 15.78337 6597.751 0.000000

        
As you can see, this table give you the main statistical values you need to analyse your regression. The menu of the equation window enable you to get others interesting views of your regression. For instance, if you click on the Resids button then you get a residuals graph.
 

 

you are doing a pooling : Eviews directly apply the GLS method;

you can convert the GLS case into an OLS case. For instance, if you have a serial correlation problem you can apply a Cochrane-Orcutt procedure by writing the following command : LS CONSUMPTION C YIELD AR(1), where AR(1) is used to apply this procedure;

you proceed in the same way as you estimate an equation with the OLS method, but you click on the estimate button of the equation window to choose the right estimation methods and their options. Please be cautious and choose the right estimation methods and parameters after checking them in the econometric litterature. You are not just doing press-buttons econometrics.

After computing your regression, the t, F and Durbin-Watson test are directly available in the equation window. If you want to compute other statistical tests, click on the view button of the equation window and select either Coefficient Tests or Residuals Tests or Stability Tests. To come back to regression results click on the stats button.

Type seas or smooth in the command line placed just below the main menu. Indicate the name of the serie into the dialog box. Another dialog box comes to screen and you choose the right methods and parameters to smooth the time serie or to adjust it for seasonal variations.

After getting your regression results, click on the forecast button of the menu in the equation window. Apply the best method to your case. So you have to think to that very well... Look at our econometric bibliography if you have some doubt.
 

After opening your workfile, click on Object New System and give it a name. A new window comes to screen. Write in it your system using its structural form. The system parameters as the constants and coefficients are writtenn C(1) C(2) C(3) ... C(n). For instance, you could type the following system with five parameters to estimate :

CONSUMPTION = C(1) + C(2)*YIELD
INVESTMENT = C(3) + C(4)*YIELD + C(5)*YIELD(-1)
YIELD = CONSUMPTION + INVESTMENT + GOVERNMENTEXPENSES
 
Notice : YIELD(-1) means the yield at the former period.

After writting your model, you can estimate it. Only click on the estimate button of the system window. A dialog box will ask you to choose the right method of estimation. Apply the method that suits well to your case. For further tips, type system in the search engine of the software help and choose system estimation. Advice : have a look to our econometric bibliography too.

Sometimes cross-section and time-series data are merged or pooled. The result could be interpreted as a cross section of time series or a time series of cross sections. After opening your workfile, select Objects New Object and choose Pool giving it a name. Identify the elements which differentiate the countries or years in the new window you can view on the screen. For instance, if you got variables for 24 countries and for the years 95 and 96, type 95 and 96 as identifiers. But be careful ! You should have created the series before doing your pooling. The names of the series have to contain the identifier. In our case, you should have written the yield series in the following way : YIELD95 and YIELD96. After that click on the sheet button of the pool window and type the series names replacing the identifier by a question mark ?. For instance, YIELD? PRICE? DEMAND?. Another window comes to screen with the data. Click on the estimate button and write the variables names with a ? instead of the idenfier part. For instance, DEMAND? in Dependent Variable and YIELD? PRICE? in Common coefficients. Choose the right methods and parameters. Perhaps you need to check in the econometric litterature or in the software help.