[Samaneh Gh]

Hello everyone, great post.
I have a question, I'm trying to use the commands xi: and i. like this:

xi: qregpd l_igini l_egdp00 i.reg, id (id) fix (year) quantile (0.25)

where all the variables are numbers in that command line.

and Stata throws me the next mistake.

"factor variables and time-series operators not allowed"

Please someone to help me understand what I'm doing wrong or how can I solve the problem?
Thanks in advance.

When you use "xi:" you don't have to use i anymore. Try the command with reg instead of i.reg.

[Juan Miranda]

Yes thanks.
This's was the problem.

[Mohammad Azeem Khan]

buHello Mr. Matthew J. Baker

"Problem with optimization technique"

Actually, I am a new user of quantile regression technique. In my panel data model, I have 36 cross-sectional IDs and the Time period is 18 years. In my model, some of the covariates are of macro nature which means they only change with time but not among the cross-sectional units. I applied the qregpd command with MCMC optimization technique. Where I used 1000 draws and 500 burns and arate was 0.5. But my results are not consistent in the sense that when I repeated the same command at the same quantile, my coefficients are changing in magnitudes. Some of them are even changing their signs. Could you please tell me that why is that so? I also used other optimization techniques but the problem is still there.

Sincere Thanks in advance.

[Barbora Sedova]

Please, can someone advice: https://www.statalist.org/forums/for...on-explanation

[Anna Bennet]

Dear Mr. Matthew,
Thank you very much for your new packages. I would like to ask are there any tests which allow to check validity and weakness of instruments within these packages and could I use internal instrumental variables? Thank you in advance.

[Sabyasachi Tripathi]

Dear Dr. Matthew J. Baker

I want to calculate heteroskedasticity after qregpd regression.
Could you please help me how can I perform the test.
Are there any other post tests require after qregpd regression.
Please help.

With Kind Regards
Sabyasachi

[Zara Ruzim]

1

[setareh amd]

Dear Matthew J. Baker
I have problem with qregpd estimation. Each times I estimate, I've received different result. I attached my estimation here.
I would be grateful if you could help me with quantile regression.
Thank you in advance.
Best regards

[Atrayee Choudhury]

Hello Dear Matthew,

I have used qregpd command using MCMC method for Powell (2015). Very helpful indeed.

I have one quick query. Can we use interaction variables as one of the explanatory variables in Powell's method? Have came across few papers which have used interaction variables as explanatory variables in.. panel quantile regression but they have used Koenkar's (2005) method. Haven't seen any literature yet which have used an interaction explanatory term in Powell's method. I have tried using interaction term as one of explanatory variable in qregpd (MCMC method) and the results are quite desirable. But before proceeding I wanted to ask you if Powell's method allows the use of interaction terms as explanatory variable

[Atrayee Choudhury]

Hello Dear Matthew J. Baker

I have used qregpd command using MCMC method for Powell (2015). Very helpful indeed.

I have one quick query. Can we use interaction variables as one of the explanatory variables in Powell's method? Have came across few papers which have used interaction variables as explanatory variables in.. panel quantile regression but they have used Koenkar's (2005) method. Haven't seen any literature yet which have used an interaction explanatory term in Powell's method. I have tried using interaction term as one of explanatory variable in qregpd (MCMC method) and the results are quite desirable. But before proceeding I wanted to ask you if Powell's method allows the use of interaction terms as explanatory variable

[Joao Santos Silva]

Dear Atrayee Choudhury,

I realize that you directed your question to Matthew, but while you wait for his reply, and for what it is worth, here is my view on it: There is nothing in Powell's method that prevents you from using interactions; these are treated as any other explanatory variable. I would add, however, my usual warning: Powell's method is extremely elegant but it estimates a model that is very different from what many users have in mind. so, make sure the estimator is doing what you think it is doing and if not consider using alternative estimators that are also available in Stata.

Best wishes,

Joao

[Atrayee Choudhury]

Hello Joao Santos Silva I completely get your point. Thanks a lot!

[Fadi Ansar]

Hi All,
can anyone help me why iam getting below results while using qregpd commands .fmy data is consist of T=14 and N=70
[CODE]qregpd Zscore d_MP HHI_Asset_Aggregate CapitalRatio NIM CBT lasset lgdp linf ,id(bankname) fix(Y
> ear) optimize(mcmc) noisy draws(1000) burn(100) arate(.5)
Adaptive MCMC optimization
................................................. 50: f(x) = -146.271289
................................................. 100: f(x) = -147.176643
................................................. 150: f(x) = -146.271289
................................................. 200: f(x) = -144.099618
................................................. 250: f(x) = -144.099618
................................................. 300: f(x) = -146.271289
................................................. 350: f(x) = -146.271289
................................................. 400: f(x) = -146.271289
................................................. 450: f(x) = -146.271289
................................................. 500: f(x) = -146.271289
................................................. 550: f(x) = -146.271289
................................................. 600: f(x) = -146.271289
................................................. 650: f(x) = -146.271289
................................................. 700: f(x) = -146.271289
................................................. 750: f(x) = -146.271289
................................................. 800: f(x) = -146.271289
................................................. 850: f(x) = -144.099618
................................................. 900: f(x) = -144.099618
................................................. 950: f(x) = -146.271289
................................................. 1000: f(x) = -144.099618


Quantile Regression for Panel Data (QRPD)
Number of obs: 632
Number of groups: 85
Min obs per group: 1
Max obs per group: 9

Zscore Coef. Std. Err. z P>z [95% Conf. Interval]

d_MP 2.66e-16 7.53e-18 35.28 0.000 2.51e-16 2.80e-16
HHI_Asset_Aggregate -173.4132 . . . . .
CapitalRatio 1.63e-15 1.40e-16 11.65 0.000 1.35e-15 1.90e-15
NIM 1.31e-16 9.56e-18 13.65 0.000 1.12e-16 1.49e-16
CBT -.3285899 . . . . .
lasset 1.00e-16 1.65e-18 60.73 0.000 9.70e-17 1.03e-16
lgdp 2.65e-16 9.55e-17 2.78 0.005 7.81e-17 4.52e-16
linf -1.482125 . . . . .

No excluded instruments - standard QRPD estimation.


MCMC diagonstics:
Mean acceptance rate: 0.438
Total draws: 1000
Burn-in draws: 100
Draws retained: 900
Value of objective function:
Mean: -145.7849
Min: -151.5055
Max: -144.0996
MCMC notes:
*Point estimates correspond to mean of draws.
*Standard errors are derived from variance of draws.


[CODE]

[Chandan Sharma]

Dear Prof. Baker;
Thanks for your post. i have three queries:
Post estimation qregpd, how to plot quantile wise estimates?

How to check the validity of instruments?
post estimation using qregpd, are any diagnostics tests available?
Best
CS

[Shobhana Sikhawal]

Hello Dear Matthew J. Baker

I am using qregpd command using MCMC method for Powell (2015).

As Joao Santos Silva has mentioned in their reply that we can use the interaction term in panel quantile regression, my query is that can we calculate the marginal effects in this set-up? I have been searching for the command to calculate marginal effects in the presence of the interaction term in panel quantile regression, however haven't found any solution.