[Louison Cahen-Fourot]

Dear Daniel,

Thanks a lot for this quick and neat reply. My bad for not having kept in mind the point on the non-standard distributions from the Pesaran et al. paper.

As a side thought, if I may make a suggestion for further development of the ardl command: It would be nice to include an option to run nonlinear asymetric ardl models such as proposed by Shin et al. (2014). I am aware of the command nardl by Marco Sunder, I have used it quite a lot and it works fairly well, but as far as I know it doesn't allow for non-zero threshold in partial sum decompositions. That would be a nice improvement, and it will would be great to have it all in the same place within the ardl command. In the meantime, your command works well with hand-made partial sum decompositions.

Many thanks again for your command and your time,

Best,

Louison

[Daniel Schneider]

Dear Louison,

-ardl- and -nardl- are separate projects and I do not see any merging of functionality happening in the future. Merging functionality / code would require a substantial work effort and it is probably better to keep them as separate, tested entities. But many thanks for pointing towards the -nardl- command and for your suggestion.

Best,
Daniel

[John heric]

Dear Sebastian:
I tried run "net install ardl, from(http://www.kripfganz.de/stata/)" in my STATA-14 but i always get the error messages as followed:

net install ardl, from(http://www.kripfganz.de/stata/)
connection timed out -- see help r(2) for troubleshooting
http://www.kripfganz.de/stata/ either
1) is not a valid URL, or
2) could not be contacted, or
3) is not a Stata download site (has no stata.toc file).


Most was due to some restrictions to Internet access. what should i do?

Much appreciated.

[Sebastian Kripfganz]

Dear Sebastian:
I tried run "net install ardl, from(http://www.kripfganz.de/stata/)" in my STATA-14 but i always get the error messages as followed:

net install ardl, from(http://www.kripfganz.de/stata/)
connection timed out -- see help r(2) for troubleshooting
http://www.kripfganz.de/stata/ either
1) is not a valid URL, or
2) could not be contacted, or
3) is not a Stata download site (has no stata.toc file).


Most was due to some restrictions to Internet access. what should i do?

Much appreciated.

I have sent you a private message here on Statalist.

[Mahana Noorma]

Dear Sebastian,
Dear all,

I applied the bounds testing approach (developed by Pesaran et al. (2001)) to a specific case and I have a question regarding the long-run coefficients.
According to Giles’s blog (2013: http://davegiles.blogspot.ch/2013/06...nds-tests.html), and further papers on the bounds testing approach, there is a correspondence between the coefficients of the long-run equation and those of the "unconstrained error correction model":

Long-run equation
y_t = α₀ + α₁x_1t + α₂x_2t + v_t

Unconstrained ECM:
Δy_t = β₀ + Σ β_iΔy_t-i + Σγ_jΔx_1t-j + Σδ_kΔx_2t-k + θ₀y_t-1 + θ₁x_1t-1 + θ₂ x_2t-1 + e_t

Source: http://davegiles.blogspot.ch/2013/06/ardl-models-part-ii-bounds-tests.html

As explained by Giles, the long-run coefficients can be extracted from the unconstrained ECM:
the long-run coefficients for x₁ and x₂ are -(θ₁/ θ₀)= α₁ and -(θ₂/ θ₀)= α2 respectively.

However, according to my results, this “coefficients correspondence” is not realized and I am unable to understand why.
I already posted this question a few months ago but since then, I read the above mentioned blog of Giles and a few papers in which the bounds testing procedure was applied. Most of the authors mention this “coefficients correspondence”. Thus, I am confused and would need some further help.

Here are my results:

Unconstrained ECM

ARDL(1, 0, 0, 0, 0, 0, 0, 0)
Dependent Variable: D(Y)
Sample (adjusted): 1973 2013
Number of obs = 41 after adjustments
Variable	Coef.	Std. Err.	t	P>|t|
D(Y(-1))	-0.149489	0.118913	-1.257125	0.2208
D(X1)	-0.06227	0.075834	-0.821134	0.4197
D(X2)	-1.641474	0.493317	-3.32742	0.0028
D(X3)	0.135946	0.273259	0.4975	0.6234
D(X4)	7.858689	2.3929	3.28417	0.0031
D(X5)	-0.468855	0.108366	-4.326576	0.0002
D(X6)	-0.011257	0.01183	-0.951626	0.3508
D(X7)	-101.6557	142.0518	-0.715624	0.4811
Y(-1)	-0.945555	0.261508	-3.615784	0.0014
X1(-1)	-0.036331	0.012209	-2.975747	0.0066
X2(-1)	-0.632525	0.293756	-2.15323	0.0416
X3(-1)	-0.149936	0.139885	-1.071849	0.2944
X4(-1)	-2.337494	1.939226	-1.205375	0.2398
X5(-1)	-0.291691	0.10937	-2.667013	0.0135
X6(-1)	-0.012298	0.015918	-0.772534	0.4473
X7(-1)	-66.49316	154.2351	-0.431116	0.6702
C	28.02979	7.87529	3.559208	0.0016
R-squared	0.952677	Mean dependent var		-0.130624
Adjusted R-squared	0.921128	S.D. dependent var		2.622106
S.E. of regression	0.736398	Akaike info criterion		2.519657
Sum squared resid	13.01475	Schwarz criterion		3.230162
Log likelihood	-34.65296	Hannan-Quinn criter.		2.778384
F-statistic	30.19684	Durbin-Watson stat		2.025592
Prob(F-statistic)	0

Long-run equation

Long-run equation
Dependent Variable: Y
Sample (adjusted): 1971 2013
Number of obs = 43 after adjustments
Variable	Coef.	Std. Err.	t	P>|t|
C	32.37835	8.3102	3.896217	0.0004
X1	-0.032359	0.012787	-2.530517	0.016
X2	-1.013083	0.25197	-4.020655	0.0003
X3	-0.136791	0.156776	-0.872522	0.3889
X4	2.914955	2.137927	1.363449	0.1814
X5	-0.389945	0.165139	-2.36131	0.0239
X6	-0.030134	0.010307	-2.923763	0.006
X7	-294.3322	158.4416	-1.85767	0.0716
R-squared	0.571425	Mean dependent var		1.861392
Adjusted R-squared	0.48571	S.D. dependent var		1.955403
S.E. of regression	1.402299	Akaike info criterion		3.680344
Sum squared resid	68.82546	Schwarz criterion		4.008009
Log likelihood	-71.12739	Hannan-Quinn criter.		3.801176
F-statistic	6.666562	Durbin-Watson stat		2.387055
Prob(F-statistic)	0.000049	Wald F-statistic		4.324163
Prob(Wald F-statistic)	0.001537

For example, if I want to get the long-run coefficient of the variable X1: - (-0.036331)/( -0.945555)= -0.03842 is not equal to -0.032359
I can not find what I am doing wrong…and would appreciate some help.

Thank you very much.
Kind regards

[Sebastian Kripfganz]

Mahana Noorma:
Please understand that we usually cannot give advice on EViews.

As far as I know, the long-run equation is estimated separately in EViews and not obtained from the unrestricted ECM. When the short-run dynamics are ignored, it is clear that the estimates will differ.

[Mahana Noorma]

Dear Sebastian,

Thank you so much for your reply.

Kind regards

[bui quang hien]

Dear All,

Currently, I am investigating ardl using http://www.stata.com/meeting/chicago..._kripfganz.pdf
(slide 9)

*CASE1
. webuse lutkepohl2
. ardl ln_inv ln_inc ln_consump , maxlag(8) maxcomb(1500) ec dots fast
------------------------------------------------------------------------------
D.ln_inv | Coef. Std. Err. t P>|t| [95% Conf. Interval]
-------------+----------------------------------------------------------------
ADJ |
ln_inv |
L1. | -.1882187 .063747 -2.95 0.004 -.3152092 -.0612281
-------------+----------------------------------------------------------------
LR |
ln_inc | -1.381501 1.827062 -0.76 0.452 -5.021196 2.258194
ln_consump | 2.272314 1.875768 1.21 0.230 -1.464407 6.009035
-------------+----------------------------------------------------------------
SR |
ln_inv |
LD. | -.1358362 .1031304 -1.32 0.192 -.3412826 .0696101
L2D. | -.0336715 .1069373 -0.31 0.754 -.2467014 .1793585
L3D. | .2087659 .1018203 2.05 0.044 .0059294 .4116025
L4D. | .4012027 .097397 4.12 0.000 .207178 .5952275
|
ln_consump |
D1. | 1.36573 .4398616 3.10 0.003 .4894806 2.241979
|
_cons | -.0030844 .0657358 -0.05 0.963 -.1340368 .127868
------------------------------------------------------------------------------

So, LR and SR formula will be rewrite as below.
How to calculate constant parameter for LR?
Does LR and SR formula rewrite correct?
*LR:
ln_inv = -1.381501 ln_inc + 2.272314 ln_consump + constant

*SR:
D.ln_inv = -.1882187 ECT(t-1) -.1358362 D.ln_inv (t -1) -.0336715 D.ln_inv (t -2) + .2087659 D.ln_inv (t -3)+ .4012027 D.ln_inv (t -3)
+ 1.36573 D.ln_consump (t) -.0030844

*CASE2:
. ardl ln_inv ln_inc ln_consump , maxlag(8) maxcomb(1500) ec1 dots fast
------------------------------------------------------------------------------
D.ln_inv | Coef. Std. Err. t P>|t| [95% Conf. Interval]
-------------+----------------------------------------------------------------
ADJ |
ln_inv |
L1. | -.1882187 .063747 -2.95 0.004 -.3152092 -.0612281
-------------+----------------------------------------------------------------
LR |
ln_inc |
L1. | -1.381501 1.827062 -0.76 0.452 -5.021196 2.258194
|
ln_consump |
L1. | 2.272314 1.875768 1.21 0.230 -1.464407 6.009035
-------------+----------------------------------------------------------------
SR |
ln_inv |
LD. | -.1358362 .1031304 -1.32 0.192 -.3412826 .0696101
L2D. | -.0336715 .1069373 -0.31 0.754 -.2467014 .1793585
L3D. | .2087659 .1018203 2.05 0.044 .0059294 .4116025
L4D. | .4012027 .097397 4.12 0.000 .207178 .5952275
|
ln_inc |
D1. | -.2600243 .3042968 -0.85 0.396 -.8662145 .346166
|
ln_consump |
D1. | 1.793422 .5861078 3.06 0.003 .6258353 2.961008
|
_cons | -.0030844 .0657358 -0.05 0.963 -.1340368 .127868
------------------------------------------------------------------------------
How to calculate constant parameter for LR?
Does LR formula rewrite same as CASE1 ? I still confused "ln_inc L1."
*LR:
ln_inv = -1.381501 ln_inc + 2.272314 ln_consump + constant
-------------+----------------------------------------------------------------
LR |
ln_inc |
L1. | -1.381501 1.827062 -0.76 0.452 -5.021196 2.258194
|
ln_consump |
L1. | 2.272314 1.875768 1.21 0.230 -1.464407 6.009035
-------------+----------------------------------------------------------------

Thank you very much!

[Daniel Schneider]

Hi Bui,

I am not sure whether I understood your question fully, but here are a couple of statements that may be helpful:

• The long-run relationship will be the same for options ec and ec1. Most short-run coefficients will be unaffected too.
• In order to pull the constant into the long-run relationship, use option restricted.
• I have double-checked the calculation of the constant in the long-run relationship and could not find any mistake.

Please also see the help file for ardl. In particular, sections "Deterministic components", "Long-run coefficients expressed in time t or t-1", and "The error-correction term" should help you. If anything remains unclear, please let me know.

[bui quang hien]

Dear Mr.Schneider,
I can executed long-run including coefficient parameter. However, when use option "restricted.", coefficient of SR term is hidden. Did you know why it occurred?

Thank you!

[Sebastian Kripfganz]

The constant term can either be included in the short-run part or in the long-run relationship but not in both at the same time.

[bui quang hien]

Dear Mr.Kripfganz,
-------------+----------------------------------------------------------------
SR |
golds |
D1. | .1662788 .0740476 2.25 0.027 .0196315 .3129261
|
reer |
D1. | -.4335587 .1802798 -2.40 0.018 -.7905934 -.076524
|
interest|
D1. | .0078001 .0026632 2.93 0.004 .0025257 .0130744
LD. | -.0067299 .0024835 -2.71 0.008 -.0116484 -.0018115
|
_cons | -19.7153 12.30821 -1.60 0.112 -44.09106 4.660461
------------------------------------------------------------------------------

. matrix list e(lags)

e(lags)[1,8]
lrm2s golds oil stock reer interest ex cpi
r1 1 1 0 0 1 2 0 0

May I ask why some coefficient were hidden in ADRL (1, 1, 0, 0, 1, 2, 0, 0)? Ex: oil, stock, ex, cpi
I try to compare with https://www.nrb.org.np/ecorev/pdffiles/vol25-1_art2.pdf. The result give all coefficient of all variables.

The constant term can either be included in the short-run part or in the long-run relationship but not in both at the same time.

[Sebastian Kripfganz]

The lag classification that the author uses in that linked NRB paper is not conventional. He counts the number of lags in the error-correction representation of the model, while the usual way would be to count the lags in the levels representation. This is what ardl stores in the matrix e(lags). Compare with the ardl output when you do not use the ec or ec1 option. Please also see the remarks sections in the ardl help file on "Terminology" and "Long-run coefficients expressed in time t or t-1", as well as slides 8, 9, and 14 to 17 in my Stata Conference presentation:
http://www.stata.com/meeting/chicago..._kripfganz.pdf

[Johanna Melinder]

Dear Sebastian,
Dear all,

I am using your program for my Master thesis and it has helped me a lot. So first of all, a big thank you!

I would like to test for structural break(s) in my model at a known period in time, with a particular interest on if the coefficients for some variables have changed. Using a normal OLS regression I would have included interaction terms for the variables of interest. (eg Y = c+ x1+x2 + D + D*x1 where D=1 after the break point). But since I have a mixture of I(0) and I(1) variables where the I(1) variables seems to be cointegrated I want to use ardl.

My question is therefore: can I use a similar approach in the ardl frame work? What would the model look like then?
If not, are you aware of any other approach I can use to test for a structural break in my model?

Thank you very much in advance!

Best,
Johanna

[Sebastian Kripfganz]

Currently, the ardl module does not accept interaction terms with the factor-variables notation. However, you can generate your interaction terms first as new variables and then run the ardl command more or less as usual, e.g.

gen x1_dummy = x1 * dummy
ardl y x1 x2 x1_dummy, exog(dummy)

Note that I have specified the break-point dummy with the exog() option because I suppose that you do now want this dummy to be lagged.

In addition, you maybe want to further restrict the number of lags for the interaction term to be the same as for the x1 variable itself. There is no option to directly achieve this restriction, but you can proceed in two steps and use the lags() option at the second step:

ardl y x1 x2 x1_dummy, exog(dummy) lags(. 3 . 3)

assuming in this example that the optimal lag order for x1 obtained in the first step was 3.

Last but not least, you should be aware that the critical values of the Pesaran et al. (2001) bounds test (estat btest) are no longer valid when you have such structural breaks.

Finally, if you want to use any standard postestimation commands usually available after the regress command, you can store the underlying regress estimation results from ardl with the regstore() option, e.g.

ardl y x1 x2 x1_dummy, exog(dummy) regstore(ardl_results)
estimates restore ardl_results
// followed by any regress postestimation command of your choice