ardl eview

3
Package: ARDL Bound approach Author: Yashar Tarverdi Mamaghani Author’s email: yashar [email protected] or [email protected] Date: 06/11/2014 Add-in type: Time Series Default proc name: ardlbound Version: 2.0 Description ARDL bound finds the best lag structure (based on user selection) for an ARDL model. It also runs the Wald test and finds the F-value required for ARDL bound test. This add-in and ARDL bound test is based on Pesaran et al(2001). Please cite this add-ins as : ”Tarverdi M Yashar, (2014). ardlbound: An Eviews Add-in to perform ARDL bound test based on Pesaran et al (2001).” Changes log: Number of variables that can be used in the add-in increased to 10. Complexity removed from add-in code to make it more expandable for non- programmers. Criterion ”No autocorrelation in residual” is omitted. Option to show the detailed table is omitted as all the results are shown in the table. Manual After running the program, in the first window user can define the pa- rameters for this test. Parameters such as : name of variables,criterion,type of unit root test and many others. In picture below the options and their function can be seen. 1 name of the variables in correct order (up to 10 variables) 2 Significant level (This applies for all tests) 3 Here you can choose which Unit Root test is going to be used to stationary test.There are four choices ADF, KPSS, DFGLS, PP. 4 Criterion that is going to be used for finding the best structure 5 Maximum number of lag you want to consider in your model. 6 Whether you want to consider ”Intercept” in estimation. By default it is selected. 1

Upload: sarwar41ss

Post on 06-Nov-2015

7 views

Category:

Documents


0 download

DESCRIPTION

ardl in eviews

TRANSCRIPT

  • Package: ARDL Bound approach

    Author: Yashar Tarverdi Mamaghani

    Authors email:yashar [email protected] or [email protected]

    Date: 06/11/2014

    Add-in type: Time Series

    Default proc name: ardlbound

    Version: 2.0

    Description ARDL bound finds the best lag structure (based on user selection)for an ARDL model. It also runs the Wald test and finds the F-value requiredfor ARDL bound test. This add-in and ARDL bound test is based on Pesaran etal(2001).Please cite this add-ins as :Tarverdi M Yashar, (2014).ardlbound: An Eviews Add-in to perform ARDL bound test based on Pesaran et al(2001).

    Changes log:

    Number of variables that can be used in the add-in increased to 10. Complexity removed from add-in code to make it more expandable for non-

    programmers.

    Criterion No autocorrelation in residual is omitted. Option to show the detailed table is omitted as all the results are shown in

    the table.

    Manual After running the program, in the first window user can define the pa-rameters for this test. Parameters such as : name of variables,criterion,type of unitroot test and many others.In picture below the options and their function can be seen.

    1 name of the variables in correct order (up to 10 variables)

    2 Significant level (This applies for all tests)

    3 Here you can choose which Unit Root test is going to be used to stationarytest.There are four choices ADF, KPSS, DFGLS, PP.

    4 Criterion that is going to be used for finding the best structure

    5 Maximum number of lag you want to consider in your model.

    6 Whether you want to consider Intercept in estimation. By default it isselected.

    1

  • Figure 1:

    How it works The program goes through following steps:

    1. Check the variables to be time series.

    2. Check the series to be stationary(This method(bound test) is applicable totime series which become stationary after taking maximum two differences.)

    3. Based on the selected max lag, this add-in estimate an ECM with all lagcombination possible.The equation is yt = 0+0yt1+1xt1+

    iytj +

    ixtj

    4. Based on selected criterion, it highlights the best lag combination alongwith the F-value from Wald test(This F-value is required for comparison withtables in Pesaran et al(2001)).

    5. At the end a table is presented. And the line that delivers satisfactory criterionis colored in Red.(Figure 2). Also, the other criterion that might be satisfactorybecomes orange(Figure 3).

    Figure 2:

    2

  • Figure 3:

    Contact and Support I would appreciate any feedback and comment about theadd-in.

    Yashar Tarverdi Mamaghani

    3