## ardl

For making suggestions and/or requests for new features you'd like added to EViews.

Moderators: EViews Gareth, EViews Moderator

tvonbrasch
Posts: 331
Joined: Fri Apr 15, 2011 5:35 am

### ardl

Hi

Just installed eviews9. Really like the ardl function.

it would be nice if one could impose restrictions on the long run solution and the let Eviews search across models (for a given long run specification with restrictions imposed).

Can you add this functionality to Eviews 9?
T

EViews Gareth
Fe ddaethom, fe welon, fe amcangyfrifon
Posts: 11674
Joined: Tue Sep 16, 2008 5:38 pm

### Re: ardl

Nice idea, but too late for ev9.
EV10 perhaps

tvonbrasch
Posts: 331
Joined: Fri Apr 15, 2011 5:35 am

### Re: ardl

Hi

Thomas

smecointelligency
Posts: 2
Joined: Tue Dec 26, 2017 4:45 am

### Re: ardl

Good Morning, Hello,
Please, I estimated an ARDL model by
Eviews9 and another time by Eviews10, and after, in the
bounds test (ARDL Bounds Test) I find a difference in the
results and in the display too:

In Eviews9: F-statistic =
6.299721 / k = 5
In Eviews10: F-statistic =
16.37981 / k = 5

what does
it mean?

Cordially

***********************************************************************************************************************************************************************

In Eviews9: F-statistic =
6.299721 / k = 5

ARDL
Bounds Test
Date: 12/26/17 Time: 14:15
Sample: 1989 2013
Included
observations: 25
Null Hypothesis: No
long-run relationships exist

Test Statistic Value k

F-statistic 6.299721 5

Critical
Value Bounds

Significance
I0 Bound I1 Bound

10% 2.26 3.35
5%
2.62 3.79
2.5% 2.96
4.18
1% 3.41 4.68

Test
Equation:
Dependent Variable: D(CEDI)
Method: Least Squares
Date:
12/26/17 Time: 14:15
Sample: 1989 2013
Included observations: 25

Variable Coefficient Std.
Error t-Statistic Prob.

D(IRD) -10.52031 12.48692
-0.842506 0.4127
D(MMDI) 0.059484
0.040760 1.459352 0.1651
D(MMDI(-1)) -0.101842 0.050627
-2.011622 0.0626
C
3068.236 1686.466 1.819329
0.0889
DIDI(-1) 0.038992
0.110981 0.351340 0.7302
IN(-1) -4.033732 10.04597
-0.401528 0.6937
IRD(-1)
-30.32519 20.32496 -1.492017
0.1564
MMDI(-1) 0.157618
0.040771 3.865971 0.0015
UE(-1) 29.26800 16.73319
1.749098 0.1007
CEDI(-1)
-0.594558 0.288519 -2.060723
0.0571

R-squared
0.787975 Mean dependent var
256.5458
0.660761 S.D. dependent var
328.5518
S.E. of regression
191.3625 Akaike info criterion
13.63539
Sum squared resid
549294.3 Schwarz criterion
14.12294
Log likelihood -160.4424
Hannan-Quinn criter.
13.77062
F-statistic 6.194057
Durbin-Watson stat 2.093658
Prob(F-statistic) 0.001054

***********************************************************************************************************************************************************************

In Eviews10: F-statistic =
16.37981 / k = 5

ARDL Long
Run Form and Bounds Test
Dependent Variable:
D(CEDI)
Selected Model: ARDL(1, 0, 0, 1, 2,
0)
Case 2: Restricted Constant and No
Trend
Date: 12/26/17 Time: 14:16
Sample: 1987 2013
Included
observations: 25

Conditional Error Correction Regression

Variable
Coefficient Std. Error t-Statistic
Prob.

C 5538.655
824.9302 6.714089 0.0000
CEDI(-1)* -1.017627
0.156009 -6.522872 0.0000
DIDI** 0.268324 0.067106
3.998530 0.0012
IN**
-11.25555 5.365446 -2.097785
0.0533
IRD(-1) -27.29649 13.55232
-2.014156 0.0623
MMDI(-1) 0.103375
0.036360 2.843070 0.0123
UE** 8.653475 13.98972
0.618560 0.5455
D(IRD)
-11.67631 8.963649 -1.302629
0.2123
D(MMDI) -0.022249 0.037384
-0.595151 0.5606
D(MMDI(-1)) -0.076970 0.035610
-2.161491 0.0472

* p-value incompatible with t-Bounds
distribution.
** Variable interpreted as Z =
Z(-1) + D(Z).

Levels Equation
Case 2:
Restricted Constant and No Trend

Variable Coefficient Std.
Error t-Statistic Prob.

DIDI 0.263676 0.046233
5.703220 0.0000
IN
-11.06059 5.793404 -1.909169
0.0756
IRD -26.82367
13.29405 -2.017720 0.0619
MMDI 0.101584 0.034788
2.920112 0.0106
UE
8.503584 13.56173 0.627028
0.5401
C 5442.717
386.9041 14.06736 0.0000

EC = CEDI - (0.2637*DIDI
-11.0606*IN -26.8237*IRD + 0.1016*MMDI +
8.5036*UE + 5442.7169 )

F-Bounds
Test Null Hypothesis: No levels
relationship

Test
Statistic Value Signif. I(0) I(1)

Asymptotic: n=1000
F-statistic
16.37981 10% 2.08 3
k
5 5% 2.39 3.38
2.5% 2.7
3.73
1%
3.06 4.15

Actual
Sample Size 25 Finite Sample:
n=30
10%
2.407 3.517
5%
2.91 4.193

1% 4.134 5.761

---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------

EViews Gareth
Fe ddaethom, fe welon, fe amcangyfrifon
Posts: 11674
Joined: Tue Sep 16, 2008 5:38 pm

### Re: ardl

Was your EViews 9 up to date? Changes were made to the calculations.

smecointelligency
Posts: 2
Joined: Tue Dec 26, 2017 4:45 am

### Re: ardl

I have eviews 9 and eviews 10,
I confirmed in eviews 9 the bounds test with the WALD test and I found the same result (F-statistic = 6.299721), but on Eviews 10 I found (F-statistic = 16.37981), please, Where is the problem?

sam SAM
Posts: 28
Joined: Tue Dec 26, 2017 4:31 am

### "THE DIFFERENCE BETWEEN EVIEWS 9 AND EVIEWS 10 IN THE BOUNDS TEST"

Good Evening,
I have eviews 9 and eviews 10, I confirmed in Eviews 9 the test of the limits with the test WALD and I found the same result (F-statistic = 1.345929); no cointegration. And I confirmed in Eviews 10 this test with the WALD test and I found the same result (F-statistic = 10.20527), but; there is cointegration.

So, according to software eviews 9 there is not a cointegration, and according to software eviews 10 it exists!

Best wishes

EViews Gareth
Fe ddaethom, fe welon, fe amcangyfrifon
Posts: 11674
Joined: Tue Sep 16, 2008 5:38 pm

### Re: ardl

Is your copy of EViews 9 up to date?

sam SAM
Posts: 28
Joined: Tue Dec 26, 2017 4:31 am

### Re: ardl

May 2015

EViews Gareth
Fe ddaethom, fe welon, fe amcangyfrifon
Posts: 11674
Joined: Tue Sep 16, 2008 5:38 pm

### Re: ardl

Update it.

sam SAM
Posts: 28
Joined: Tue Dec 26, 2017 4:31 am

### Re: ardl

Hi, Good Morning,
ie the problem is at Eviews 9 level?

EViews Gareth
Fe ddaethom, fe welon, fe amcangyfrifon
Posts: 11674
Joined: Tue Sep 16, 2008 5:38 pm

### Re: ardl

We changed the way some ARDL calculations were made in later versions of 9.

sam SAM
Posts: 28
Joined: Tue Dec 26, 2017 4:31 am

### Re: ardl

Sorry, The problem is that I confirmed the results of Eviews 9 with the Wald test, and the same on Eviews 10. in the end I found a contradiction !
So, according to software eviews 9 there is not a cointegration, and according to software eviews 10 it exists!
where is the problem?
Last edited by sam SAM on Thu Jan 11, 2018 10:25 am, edited 2 times in total.

EViews Gareth
Fe ddaethom, fe welon, fe amcangyfrifon
Posts: 11674
Joined: Tue Sep 16, 2008 5:38 pm

### Re: ardl

The blog posts outline what EViews 10 is doing. You can decide for yourself whether it matches your model etc...

sam SAM
Posts: 28
Joined: Tue Dec 26, 2017 4:31 am

thank you