搜档网
当前位置:搜档网 › eviews_16

eviews_16

eviews_16
eviews_16

Panel unit root tests require that you construct a data set and use the POOL feature in Eviews to set up the panel. All the basics are contained in a separate guide “panel estimation”.

For the illustration below I will use the Big Mac example. The data can also be downloaded from the course website. As seen below the POOL setting requires that you write the “identifiers”, that is, the country names in this example.

Next, click on VIEW and select “unit root test” as shown below.

Once you click on “unit root test” the following window will appear.

The top left hand side window asks you to place the name of the series. Recall that in a panel setting the “?” is a wild card that replaces all series that start with “ls” (here the nominal exchange rate) with lsFRANCE and so on. You also need to click on test type and choose “Im, Pesaran, Shin” in the test type window. While several tests exist the one considered in Enders is due to these authors. Once you’ve decided what type of lag augmentation is appropriate (e.g., Schwarz vs some other selection technique – see the top right hand side window), whether you want to choose a maximum lag length, as opposed to letting Eviews decide for you, as well as whether a trend belongs in the test

equation, then click OK and the following will appear.

For the purposes of this example, I chose a subset of 6 countries (these are shown at the top). Next, the test statistic described in Enders page 225 is shown next. Finally, the individual ADF test statistics are shown as a re the components of the test equation (i.e.,

E(t), and E(Var), as well as the lag augmentation chosen for each series in the panel.

相关主题