Panel Cointegration

In this section, you will learn to carry out the Pedroni (Engle-Granger based) Panel Cointegration in EViews 8. We will explain in detail about them in the following steps:

Let us assume, the name of the MS-Excel datasheet is "Panel-Data-Representation.xlsx". You have to upload this dataset in EViews. Follow the step-by-step guide to upload this dataset in Eiews 8:

Go to File -> Import -> Import from file...



Upon selecting the data, the following window will appear:



Once you click Next > , the following window will appear:



Once you click Next > , the following window will appear:



Once you click Finish, the following window will appear:



Now in this window, you can see that all the five variables (N, Y, EC, PT, and POP) are uploaded. All the variable names are in lower-case letters. Apart from the variables, you will also find the CODE, YEAR, dateid, c, and resid. You will not require these components for this test. This confirms that the data has been uploaded properly.

Once the data is uploaded in EViews 8, we need to carry out logarithmic transformation on the data. Follow the step-by-step guide to perform logarithmic transformation of the dataset in Eiews 8:

Go to Object -> Generate Series...



Upon clicking, the following window will appear. Here you will have to type the equation lec=log(ec) for logarithmic transformation...



Once you click OK, a new variable LEC will be generated. In this way, you will have to log-transform the remaining four variables using the following four equations:

ly=log(y)
ln=log(n)
lpop=log(pop)
lpt=log(pt)

At the end, you will have five log-transformed variables, i.e. LN, LY, LEC, LPT, and LPOP. The EViews 8 window will look like the following:

Select the newly generated log-transformed variables (as shown in the following window), and go to Quick -> Group Statistics -> Johansen Cointegration Test



Upon clicking, the following window will appear:



Once you click OK, the following window will appear:



At the beginning, the Individual intercept option will be chosen by default. The Lag length should be chosen as Automatic Selection, with Max lag: as *. Once you click OK, the following window will appear:



There are total eleven statistics reported in the result. In order to ascertain the cointegrating association between the variables, at least 6 statistics should be significant. In the result, both the values of Panel PP-Statistic and Panel ADF-Statistic, Group PP-Statistic, and Group ADF-Statistic are significant at 1% level. As 6 out of 11 statistics are significant, cointegrating association is ascertained using Individual intercept.

Now, again go to Quick -> Group Statistics -> Johansen Cointegration Test



Upon clicking, the following window will appear:



Once you click OK, the following window will appear:



Now, we have to select Individual intercept and individual trend option. The Lag length should be chosen as Automatic Selection, with Max lag: as *. The window will now look like the following:



Once you click OK, the following window will appear:




Like before, there are total eleven statistics reported in the result. In order to ascertain the cointegrating association between the variables, at least 6 statistics should be significant. In the result, both the values of Panel PP-Statistic and Panel ADF-Statistic, Group PP-Statistic, and Group ADF-Statistic are significant at 1% level. As 6 out of 11 statistics are significant, cointegrating association is ascertained using Individual intercept and individual trend.

Now, again go to Quick -> Group Statistics -> Johansen Cointegration Test



Upon clicking, the following window will appear:



Once you click OK, the following window will appear:



Now, we have to select No intercept or trend option. The Lag length should be chosen as Automatic Selection, with Max lag: as *. The window will now look like the following:



Once you click OK, the following window will appear:




Like before, there are total eleven statistics reported in the result. In order to ascertain the cointegrating association between the variables, at least 6 statistics should be significant. In the result, both the values of Panel PP-Statistic and Panel ADF-Statistic, Group PP-Statistic, and Group ADF-Statistic are significant at 1% level. As 6 out of 11 statistics are significant, cointegrating association is ascertained using No intercept or trend.

As a whole, the cointegrating association is ascertained using three options, i.e. Individual intercept,Individual intercept and individual trend, and No intercept or trend. Therefore, we can conclude that the cointegrating association exists among the panel variables.