21-12-2012, 06:50 PM
Applied Econometrics
1Applied Econometrics.ppt (Size: 190.5 KB / Downloads: 22)
Defining econometrics and its methodology
Amalgam of economic theory, mathematics and statistics
Methodology
Statement of theory and hypothesis
Specification of the mathematical model of the theory
Specification of the statistical, or econometric model
Obtaining the data
Estimation of parameters of the econometric model
Hypothesis testing
Forecasting or prediction
Using the model for control or policy purposes
Data Transformation.xls
Introduction to E.Views
Creating Workfile (Go to FILE---NEW---WORKFILE)
A window will appear asking for the type of data (Time series, Cross Section and Panel) you want to enter. There are three options
Unstructured/Undated – Cross Section Data
Dated-regular frequency – Time Series Data
Balanced Panel – Panel Data
Along with the type of data window will also ask for the START DATE and END DATE of data for time series and panel data and number of observations for cross section data. You will provide the information accordingly.
Features of Workfile
Main Menu Bar
Workfile Command Bar
Status Bar
White window
Originally two folders will appear in white window. One with parameter sign “beta” is for constant (intercept) and the second one with name “resid” is for error term series. After regression error term series will appear automatically in this folder.
Saving Workfile
Be very careful in saving Workfile. Safe File with extension .wf1. Sometimes while saving E.Views Workfile “commandlog” will appear in SAVE tab which is not the correct file to save.
Data Entry
Time series data or Dated-regular frequency data
Feeding Data in E.Views directly
Creating Object
Select OBJECT from workfile command bar. Select NEW OBJECT from dropdown menu. A window will appear with two tabs.
Left Tab---Type of Object
Right Tab---Name for Object
Select SERIES from left tab and give that series a NAME in right tab. Try to give brief name, for example, inf for inflation, y for income, exc for exchange rate, etc.
Once the above mentioned task is done you will notice a folder appearing in white window with the name specified by you. Open this folder. Click the command button EDIT+/-. The window is active now. You can enter the data or copy the series from excel. After feeding the data clicking EDIT+/- again is recommended.
Some Precautions
If you are copying data from excel file some things should be observed before copying the data.
Merged cells should be separated first.
There should be no space in the names of series. If there is space E.Views will treat it as two different series. For instance, if you have given exchange rate a name as exc rate then E.Views will treat exc and rate as two different series.
Graphs
Suppose we want to analyze the trend of inflation (inf) over time. One of the effective ways to do it is to draw line graph which can be drawn as follows.
Open the folder “inf”. Go to VIEW. Select GRAPH from dropdown menu and then click LINE.
You will observe that spreadsheet has been changed into graph. To save the graph as separate folder click on FREEZE. A window with same graph will appear. Give it a name by clicking on NAME and close the window.
In your workfile a folder will appear with bars on it. This is frozen window and cannot be used for further manipulation. However, you can change the formatting of graph and copy the graph to MS.
Interpretation of Results
As the model is run in log-log form, coefficients will be explained in percentage form. The interpretation for (k/elf) is :
With one percent increase in capital labor ratio gdp as a ratio of elf will increase by 0.27 percent. The coefficient of k/elf is also significant at one percent.
The next step is to explain the goodness of fit model which can be seen through R-square and adjusted R square. R square is 0.97 which means that 97 percent variations in dependent variable are being explained by independent variables included in the model
another task is to explain F-test for the joint significance of model which is not relevant for Bivariate regression model. We will do it for multivariate regression model.