TH-cam recently changed the way my content will be monetised. My channel now needs 1,000 subscribers. So it would be amazing if you show your support by both watching my videos and subscribing to my channel if you haven’t done so already. Monetising my videos allows me to invest back into the channel with some new equipment so this small gesture from you will be extremely huge for me. Many thanks for your support….CrunchEconometrix loves to teach, help me stay online.
Thank you for your help!! For those using STATA 15 , the eststo command doesn't work (at least it doesn't for me, I think STATA updated the command) in steps 4 and 5. The work around is to run the model one at a time then use "estimates store fixed" (after running the fixed effects) and "estimates store random" (after running the random effects) commands immediately after. Then proceed to step 6. So for me step 4 command was: . xtreg depvar expvar1 etc, fe . estimates store fixed Step 5 was the same using re and random in place of fe and fixed. Hope this helps.
@@dr.markevers8331 Awesome! I'll appreciate it if you can share the link to my TH-cam Channel with your friends and academic community in the 🇺🇸. They will find the content helpful too 😊.
Hello Dr. thank you for your videos, they are very very useful. I'm watching you from Costa Rica and I will certainly share them with other economics students.
Good morning Crunch Econometrix. I am located in Ghana. I hope to attend one of your workshops in data analysis in the future. I am also successfully in estimating the model with your help. Please what are the possible diagnostic test I could run for the random and fixed effect model. Thank you.
@@CrunchEconometrix I am revising for my final coursework exams. Am sure even during practice and very soon at my PhD level I will continue consulting your content. Stay blessed.
Dr Adeleye, thank you very much for the videos. Please can you explain why you are using this code in addition "cons urbp yr2-y13" Please what is urbp? why you didnt write yr1-yr13? Thank you
Hi Daniel, thanks for the encouraging feedback. "cons" and "urbp" are the regressors. Even if I code yr1-yr13 Stata will drop yr1 due to collinearity. You can always try it to see what you get.
Hello Bosede Ngozi Adeleye, First, I want to say thanks for your time, I have one question but not about this lecture. In order to determine the model for balanced panel data, conducted necessary tests: F test for the fixed-effect model and Breusch-Pagan Lagrange Multiplier (LM) test. Then the Hausman test is needed if found both fixed and random effects. The Hausman test recommended the fixed-effect model. Now, which test should I run? One-way fixed effect (“within” estimation) or Two-way fixed (“within” estimation)/Between” estimation
Hi Sohail, one-way (either time or cs dummies) and two-way (both time and cs dummies) fixed is applicable to LSDV (which is a pooled OLS fixed effects procedure). For panel fixed effects, only time dummies are applicable as the cs dummies are eliminated because they are time-invariant. Hope these hints are helpful.
@@CrunchEconometrix Thanks, your suggestions are very helpful and still need little help. (1)what do you mean by cs dummies? (2)The commands should be for panel fixed effects xtreg DV IDV i.year,fe (3)How to remove serial correlation, heteroskedasticity and collinearity using Stata in the balanced panel data? any link or your lecture, please recommend me. Thanks again for your kind consideration.
Thank you for the great explanation! Sorry how do we interpret the R square in fixed effect model as we have three R square ( between, within and overall)? Or does it mean we dont interpret the r square in Fixed and random effect due to absence of adjusted r square?
@@abrahamcharles8895 Awesome! I'll appreciate it if you can share the link to my TH-cam Channel with your students and academic community in Tanzania 🇹🇿...May God bless you as you do, amen 🙏
Hello Madam, I am a Ghanaian studying in Germany. Many thanks,your videos have been very helpful. My question is I am using IV 2SLS using the xtivreg2 function, I wanted to know if I could follow your steps in this video to use the Hausman test and whether to use it on the first stage or on the second stage. Secondly, could you please direct me on how to create a 3 level fixed effects model with each variable having subscripts (itr) standing for countries, years and regions of the world. I already have my data countries, years and regions. I would be super grateful. Thanks in advance
Hi Akuafo, thanks for the encouraging feedback. Deeply appreciated. I use the Hausman test for the FE/RE techniques and not IV-2LS. You may need to check other online resources on the possibility. I also have no idea about 3 level fixed effects, at the moment. Kind regards.
Hello prof am thanks for the video really helpful, my question is I have model that has 3 parts some are less than 5% and other more so can I use both random and fixed effects at same time . Thanks
Hi! Your content is really helping me with my thesis. What variables do you include in the equations for the hausman test? Is it just the dependent and independent variable? I have an interaction in my dataset, do I include that as well?
Hi I want to thank you for the explanation. I have been following your videos from Mauritius for quite a while now and they have been really helpful in the course of my dissertation. My first question is that I have run the Hausman Test and obtained a p-value of 0.9719. Does this mean that I should not reject the null hypothesis and use a random effects model due to correlation being present? And secondly, is there any difference between a random effects model and a random effects regression on stata? I would greatly appreciate your answer. Thanks.
Hi! I have a question and was hoping that after 5 years maybe you would still respond hahah. In the video I saw that one of your variables was ommitted. I have the same when doing the regression. In my Random effects regression, one of the industry dummies and one of the year dummies get ommitted. In the Fixed effects, all industry dummies get ommitted. What do you suggest to do? The hausman test suggests using the random effects. But I don't know how to interpret the ommitted variables
Thank you for your help Professor. I wanted to ask if this is the test that is also used to test for endogeneity in panel data? If so, how must one approach the test for endogeneity?
Hi Advaith, thanks for the positive feedback. Hausman test does not address endogeneity. You can check online resources for more information on the various endogeneity tests. Please may I know from where (location) you are reaching me?
I have run my model as per your video .hausman Fixed.My chi result is Prob>chi2 = 0.2990; it is more than 5% so is it mean that Null Hypothesis (Random Model is appropriate) is correct? and I should go for Random Effect? or my model fails?
CrunchEconometrix. Thanks for your support through the mail I sent you. I have since downloaded the and they are helpful. However, am using Stata 14 and am having difficulties in location the estimate store on the interface of the software. Any command to use for it please Thank you once again
Hi Faith, it implies that the syntax is yet to be downloaded. Use either "ssc install eststo" OR "findit eststo" to install it on your computer. May I know from where (location) you are reaching me?
@@CrunchEconometrix I am in Ghana. Hope to one attend one of your workshops. I have a lot of friends in Lagos and hoping to visit Lagos before the year ends. I was able to estimate the random and fixed effect models. I researched diagonostic test for from your videos but could not find any. Any recommendation please. Thank you.
Thanks for the video Professor. Please, I understand it is not ideal to use too many regressors (independent variables). In my case, I classified exchange rate regime(explanatory variable) into 3 dummies already and capital control is my 2nd explanatory variable. I have reserve accumulation as dependent variable.Trade Openness, Inflation and Short term debt are my control variables. #Can I use only Trade openness and drop the rest of the control variables? #Do I use the dummies for the exchange rate regime when running the Hausman test or I have to use the exchange rate regime variable? Thank you. With much appreciation for your guidance.
Hi Emma, b4 performing HT, run correlation analysis to ensure that highly collinear variables are not included together in the same model. After that, run the HT with reserve, exchange rate, capital and trade. Run other simulations by interchanging the control variables. Also, perform HT using the dummies of the exchange rate variable. Choose the model that gives you the most significant results.
@@wordin2mins Bravo Emma!!! I am so very excited (clapping). Please tell your friends and colleagues about my TH-cam Channel. Congratulations, once again. Proud of you!
Good day Ma. I'm doing a panel data study on ecowas. The results of my f-test and breusch-pagan lm tests were significant, but the p-value for the hausman test was inconclusive. It read that "model fitted on these data fails to meet the asymptomatic assumptions of the Hausman test". However, when I switched the order of the command on the hausman test, the p-value was 0.0137. Can I go ahead to reject the null hypothesis based on the second p-value or do I proceed with another panel data estimation technique?
Hello, I just wondered if you knew what the error "(V_b-V_B is not positive definite)" meant? I am getting this error when running the hausman test in Stata.
Hi Guiliana, you may want to read up on the Hausman test for detailed understanding of the Vb matrix...but I will suggest you focus on the pvalue of the test and take your decision.
Hi Ma'am, just a quick question. I am reading a paper and i'm unsure of something, how do we get country-specific effects? Is it a variable that the researcher has to create by him(her)self? Thank you for your videos, they have helped a lot
Hi Martin, thanks for the observation. The link to the Google drive is deactivated. The action was taken in August 2019 due to unethical and sinister actions of some persons who persistently attempted hacking my Google drive. I have since placed a detailed notice about this on my TH-cam Community Page. Dofiles and datasets are no longer free (though, some datasets are FREE so you need to check which ones). Availability is on my website cruncheconometrix.com.ng/shop/ upon payment of a token after which you are allowed a one-time download. Thanks for your understanding and patronage. CrunchEconometrix Team.
Hi, is there a way to export table result in a report as outreg2 command does for regressions? i'm looking for around the web but i didn't get any answers. Thank you for your useful videos.
Hi, you muddled this up: "is there a way to export table result in a report as outreg2 command does for regressions?" Kindly rephrase to understand what you need?
Hi, ma'am. I am reaching you from Indonesia. For my data set, the Hausman test fails to reject the null and so I understand it is better to use RE. However, I want to ask whether it is still okay to use FE? Because FE is consistent under the null and hypothesis? I am just afraid to use RE because there may be some omitted variables (i.e. any individual characteristics that may or may not influence the other explanatory variables). I hope you understand my question. Thank you.
@@raaksharammohun6279 Yes it is time series but you must reduce the variables to 4 or 5 to avoid loss of degrees of freedom and also increase the years from 10 to at least 30 to avoid small sample size bias.
@@CrunchEconometrix Thank you for your help. For the problem of small size, can i make use of log to solve the problem? Or do i have to do for a period of 30 years?
Hi Sohail, kindly follow the steps as shown in this video for the Hausman test...and if in doubt please read the Hausman paper (you can search online for it). Thanks.
Hello Feyi, To perform Chow test in EViews: 1) Plot the graph of the series to identify the breakpoint 2) run the first regression w/o the dummy variables and their interaction terms 3) perform the Chow test: go to View, Stability diagnostics, Chow test, impute the breakpoint date, click ok 4) if the F-stat is significant, then the null hypothesis of no break is rejected 5) take the plot of the CUSUMSQ and you'll see that the plot deviates out of the 5% significance level boundary To correct this: 1)create a dummy variable, (d) which takes 0 for years w/o break, 1 for breakpoint years 2) generate new series with the dummy variable for all the regressors in the model. That is, generate d*X1, d*X2, d*X3 etc 4) run the second regression with all the regressors and the newly generated series of the the dummy variables and their interaction terms 5) take the plot of the CUSUMSQ and you'll see that the plot stays within the 5% significance level boundary. ...try this and let me know how it goes. I intend to do a video on this within the week. So send me your data to cruncheconometrix@gmail.com, I may decide to use it for the online tutorial on structural break.
Hi Denisa, please watch my video on "reshape wide to longitudinal data". You will find it very helpful. May I know from where (location) you are reaching me? Thanks.
Concern yourself with the pvalue. If significant use the FE estimator and RE estimator if otherwise. To interpret that sentence is to get involved in some econometric jargons which I'm not prepared. But if you are so interested, then seek other online resources for the meaning.
Hello Madam, Is it correct if I use fixed effect model and random effect model on level data when some of my variables are stationary at first difference
Thanks madam, Can you advice me what is the method I should use to analyse relationship between dependent and independent variables of a panel which comprises 9 cross sections for 20 years
TH-cam recently changed the way my content will be monetised. My channel now needs 1,000 subscribers. So it would be amazing if you show your support by both watching my videos and subscribing to my channel if you haven’t done so already. Monetising my videos allows me to invest back into the channel with some new equipment so this small gesture from you will be extremely huge for me. Many thanks for your support….CrunchEconometrix loves to teach, help me stay online.
CrunchEconometrix 🤝
This was very helpful, you summed this up better than any of my lecturers have so far. Thank you.
Thanks for the encouragement, Ahmed. Deeply appreciated!
That was really concise. Loved it. Thank you so much.
You are very welcome, Sam😊
Thank you for your help!!
For those using STATA 15 , the eststo command doesn't work (at least it doesn't for me, I think STATA updated the command) in steps 4 and 5. The work around is to run the model one at a time then use "estimates store fixed" (after running the fixed effects) and "estimates store random" (after running the random effects) commands immediately after. Then proceed to step 6.
So for me step 4 command was:
. xtreg depvar expvar1 etc, fe
. estimates store fixed
Step 5 was the same using re and random in place of fe and fixed.
Hope this helps.
Wow! Thanks Mark, I'll keep this in mind and advise those with related issues. Please may I know from where (location) you are reaching me?
CrunchEconometrix I’m in Lubbock, Texas at Texas Tech.
@@dr.markevers8331 Awesome! I'll appreciate it if you can share the link to my TH-cam Channel with your friends and academic community in the 🇺🇸. They will find the content helpful too 😊.
Hello Dr., thanks for the videos, they are very helpful
Thanks for your encouraging feedback...deeply appreciated!🥰🙏
Hello Dr. thank you for your videos, they are very very useful. I'm watching you from Costa Rica and I will certainly share them with other economics students.
I'm encouraged by your feedback, Felipe. May God bless you as you share my videos. Much love from Nigeria 🇳🇬! 💞
your explanation is great.... many thanks...
Thanks Abderahman, for the positive feedback and remarks on my video. Deeply appreciated! May I know from where (location) you are reaching me?
Thank you for your explanation. That's Great !!!
Thanks for the encouraging feedback, Memel. Deeply appreciated!!!
Good morning Crunch Econometrix. I am located in Ghana. I hope to attend one of your workshops in data analysis in the future. I am also successfully in estimating the model with your help. Please what are the possible diagnostic test I could run for the random and fixed effect model. Thank you.
Hi Faith, read more about cross-sectional dependence test, BP LM test, hetero and serial correlation test for panel data analysis.
thank u sooooooo much. it was really helpful for me .
Glad it helped, Nada!
This is helpful. Thanks.
Thanks Obadiah, for the encouraging feedback. Deeply appreciated!
@@CrunchEconometrix I am revising for my final coursework exams. Am sure even during practice and very soon at my PhD level I will continue consulting your content. Stay blessed.
Dr Adeleye, thank you very much for the videos. Please can you explain why you are using this code in addition "cons urbp yr2-y13" Please what is urbp? why you didnt write yr1-yr13?
Thank you
Hi Daniel, thanks for the encouraging feedback. "cons" and "urbp" are the regressors. Even if I code yr1-yr13 Stata will drop yr1 due to collinearity. You can always try it to see what you get.
Hi thanks for the useful video. I have one question please how to check if there is an endogeneity problem in this model
Hi Abdelhadi, you may need to check out other online resources for that....thanks for the encouraging feedback, deeply appreciated! 🙏
@@CrunchEconometrix Already did, thanks for your efforts
Hello
Bosede Ngozi Adeleye,
First, I want to say thanks for your time, I have one question but not about this lecture. In order to determine the model for balanced panel data, conducted necessary tests: F test for the fixed-effect model and
Breusch-Pagan Lagrange Multiplier (LM) test. Then the Hausman test is needed if found both fixed and random effects. The Hausman test recommended the fixed-effect model.
Now, which test should I run? One-way fixed effect (“within” estimation) or Two-way fixed (“within” estimation)/Between” estimation
Hi Sohail, one-way (either time or cs dummies) and two-way (both time and cs dummies) fixed is applicable to LSDV (which is a pooled OLS fixed effects procedure). For panel fixed effects, only time dummies are applicable as the cs dummies are eliminated because they are time-invariant. Hope these hints are helpful.
@@CrunchEconometrix Thanks, your suggestions are very helpful and still need little help.
(1)what do you mean by cs dummies? (2)The commands should be for panel fixed effects xtreg DV IDV i.year,fe (3)How to remove serial correlation, heteroskedasticity and collinearity using Stata in the balanced panel data? any link or your lecture, please recommend me. Thanks again for your kind consideration.
Thank you for the great explanation! Sorry how do we interpret the R square in fixed effect model as we have three R square ( between, within and overall)? Or does it mean we dont interpret the r square in Fixed and random effect due to absence of adjusted r square?
Hi Abraham, use the "within" R-sq. Please may I know from where (location) you are reaching me?
CrunchEconometrix Thank you so much Madam! I appreciate your assistance! Im from Tanzania!
@@abrahamcharles8895 Awesome! I'll appreciate it if you can share the link to my TH-cam Channel with your students and academic community in Tanzania 🇹🇿...May God bless you as you do, amen 🙏
Hello Madam, I am a Ghanaian studying in Germany. Many thanks,your videos have been very helpful. My question is I am using IV 2SLS using the xtivreg2 function, I wanted to know if I could follow your steps in this video to use the Hausman test and whether to use it on the first stage or on the second stage.
Secondly, could you please direct me on how to create a 3 level fixed effects model with each variable having subscripts (itr) standing for countries, years and regions of the world. I already have my data countries, years and regions. I would be super grateful. Thanks in advance
Hi Akuafo, thanks for the encouraging feedback. Deeply appreciated. I use the Hausman test for the FE/RE techniques and not IV-2LS. You may need to check other online resources on the possibility. I also have no idea about 3 level fixed effects, at the moment. Kind regards.
Hello prof am thanks for the video really helpful, my question is I have model that has 3 parts some are less than 5% and other more so can I use both random and fixed effects at same time . Thanks
Hi Abdirahman, kindly follow the guidelines for choosing between FE and RE models. Thanks
Hi! Your content is really helping me with my thesis. What variables do you include in the equations for the hausman test? Is it just the dependent and independent variable? I have an interaction in my dataset, do I include that as well?
Hi Jari, I will advise that you exclude the interaction term. Thanks.
Hi I want to thank you for the explanation. I have been following your videos from Mauritius for quite a while now and they have been really helpful in the course of my dissertation. My first question is that I have run the Hausman Test and obtained a p-value of 0.9719. Does this mean that I should not reject the null hypothesis and use a random effects model due to correlation being present? And secondly, is there any difference between a random effects model and a random effects regression on stata? I would greatly appreciate your answer. Thanks.
Hi Zahraa, thanks for the positive feedback and kind remarks. Deeply appreciated. YES to the first and NO to the second question. Kind regards.
Thank you.
You are welcome, Sir🙏🥰
Hi! I have a question and was hoping that after 5 years maybe you would still respond hahah. In the video I saw that one of your variables was ommitted. I have the same when doing the regression. In my Random effects regression, one of the industry dummies and one of the year dummies get ommitted. In the Fixed effects, all industry dummies get ommitted. What do you suggest to do? The hausman test suggests using the random effects. But I don't know how to interpret the ommitted variables
Hi there, Stata drops them due to collinearity. See the information at the top of your Stata output. Nothing much you can do about it.
Thank you for your help Professor. I wanted to ask if this is the test that is also used to test for endogeneity in panel data? If so, how must one approach the test for endogeneity?
Hi Advaith, thanks for the positive feedback. Hausman test does not address endogeneity. You can check online resources for more information on the various endogeneity tests. Please may I know from where (location) you are reaching me?
@@CrunchEconometrix Thank you Professor. I'm reaching you from India
Thank you so much
U're welcome, Chiedza. Please may I know from where (location) you are reaching me?
@@CrunchEconometrix I am from Zimbabwe
Thank you so much. How do you create dummies (EAR ECA ...etc ) is there any command for that?
Hi Abdur, kindly watch my video on "Building Panel Data" for indepth information on what to do. Thanks.
maam how to deal with omitted variables such as firm age in fixed effects model
Hi Anuradha, not sure what you mean by this.
I have run my model as per your video .hausman Fixed.My chi result is Prob>chi2 = 0.2990; it is more than 5% so is it mean that Null Hypothesis (Random Model is appropriate) is correct? and I should go for Random Effect?
or my model fails?
Use RE estimator.
CrunchEconometrix. Thanks for your support through the mail I sent you. I have since downloaded the and they are helpful. However, am using Stata 14 and am having difficulties in location the estimate store on the interface of the software. Any command to use for it please
Thank you once again
Hi Faith, it implies that the syntax is yet to be downloaded. Use either "ssc install eststo" OR "findit eststo" to install it on your computer. May I know from where (location) you are reaching me?
@@CrunchEconometrix
I am in Ghana. Hope to one attend one of your workshops. I have a lot of friends in Lagos and hoping to visit Lagos before the year ends. I was able to estimate the random and fixed effect models. I researched diagonostic test for from your videos but could not find any. Any recommendation please. Thank you.
Thanks for the video Professor. Please, I understand it is not ideal to use too many regressors (independent variables). In my case, I classified exchange rate regime(explanatory variable) into 3 dummies already and capital control is my 2nd explanatory variable. I have reserve accumulation as dependent variable.Trade Openness, Inflation and Short term debt are my control variables.
#Can I use only Trade openness and drop the rest of the control variables?
#Do I use the dummies for the exchange rate regime when running the Hausman test or I have to use the exchange rate regime variable? Thank you.
With much appreciation for your guidance.
Hi Emma, b4 performing HT, run correlation analysis to ensure that highly collinear variables are not included together in the same model. After that, run the HT with reserve, exchange rate, capital and trade. Run other simulations by interchanging the control variables. Also, perform HT using the dummies of the exchange rate variable. Choose the model that gives you the most significant results.
@@CrunchEconometrix Thanks Professor for your guidance. I succeeded in my project. I am much grateful!
@@wordin2mins Bravo Emma!!! I am so very excited (clapping). Please tell your friends and colleagues about my TH-cam Channel. Congratulations, once again. Proud of you!
@@CrunchEconometrix I'm glad you share in my joy Professor. Sure, I will keep informing them.
Please can you tell me how to rum Hausman test for reg3 (i am using simultaneous equations with 3sls estimation)
And sargan test pleeeeeeeeese
Hi Kabel, you may need to check other online sources. I have no videos on 3sls. Thanks.
Good day Ma. I'm doing a panel data study on ecowas. The results of my f-test and breusch-pagan lm tests were significant, but the p-value for the hausman test was inconclusive. It read that "model fitted on these data fails to meet the asymptomatic assumptions of the Hausman test". However, when I switched the order of the command on the hausman test, the p-value was 0.0137.
Can I go ahead to reject the null hypothesis based on the second p-value or do I proceed with another panel data estimation technique?
Hi Joy, given the outcome of the Hausman test, proceed to estimate the FE model.
@@CrunchEconometrix Thank you ma
Hello, I just wondered if you knew what the error "(V_b-V_B is not positive definite)" meant? I am getting this error when running the hausman test in Stata.
Hi Guiliana, you may want to read up on the Hausman test for detailed understanding of the Vb matrix...but I will suggest you focus on the pvalue of the test and take your decision.
Hi Ma'am, just a quick question. I am reading a paper and i'm unsure of something, how do we get country-specific effects? Is it a variable that the researcher has to create by him(her)self? Thank you for your videos, they have helped a lot
Hi Renae, watch my videos on "Error Component Models" and "LSDV". Follow the guidelines.
Thank you 🙏🏾
Hi. The the link to the engee2.xlsx dataset leads to trash folder in your drive. You may need to restore the content for your viewers to access
Hi Martin, thanks for the observation. The link to the Google drive is deactivated. The action was taken in August 2019 due to unethical and sinister actions of some persons who persistently attempted hacking my Google drive. I have since placed a detailed notice about this on my TH-cam Community Page. Dofiles and datasets are no longer free (though, some datasets are FREE so you need to check which ones). Availability is on my website cruncheconometrix.com.ng/shop/ upon payment of a token after which you are allowed a one-time download. Thanks for your understanding and patronage.
CrunchEconometrix Team.
Hi, Could you please explain the reason for using i(c_id) after fe? Thank you
Hi Nelee, it is to cluster analysis on the group.
@@CrunchEconometrix Thank you for the reply :)
Hi, is there a way to export table result in a report as outreg2 command does for regressions? i'm looking for around the web but i didn't get any answers. Thank you for your useful videos.
Hi, you muddled this up: "is there a way to export table result in a report as outreg2 command does for regressions?" Kindly rephrase to understand what you need?
@@CrunchEconometrix I was asking a command to export outcome of hausman test without copy and paste the table.
@@alessandrocremaschini414 No command for that. What you require is the statistic and its pvalue which can be easily written out.
Hi, ma'am. I am reaching you from Indonesia. For my data set, the Hausman test fails to reject the null and so I understand it is better to use RE. However, I want to ask whether it is still okay to use FE? Because FE is consistent under the null and hypothesis? I am just afraid to use RE because there may be some omitted variables (i.e. any individual characteristics that may or
may not influence the other explanatory variables). I hope you understand my question. Thank you.
Hi Nadhira, nothing to be afraid of. Follow the test results and use the FE for robustness checks.
Hello, can you explain me how to do hausman test for time series data? Thank you
Hi Raaksha, the Hausman test is not applicable to time series analysis only panel data.
@@CrunchEconometrix okaay, i'm doing a regression analysis on 1 country with 14 variables for a 10 year period. Is that time series data or panel?
@@raaksharammohun6279 Yes it is time series but you must reduce the variables to 4 or 5 to avoid loss of degrees of freedom and also increase the years from 10 to at least 30 to avoid small sample size bias.
@@CrunchEconometrix Thank you for your help. For the problem of small size, can i make use of log to solve the problem? Or do i have to do for a period of 30 years?
@@raaksharammohun6279 Solution: Increase years to at LEAST 30.
Hello Madam,how i can select the (Ho)null H to run the hausman test,because someone selected Ho as FE is appropriate and H1 as RE is appropriate .
Hi Sohail, kindly follow the steps as shown in this video for the Hausman test...and if in doubt please read the Hausman paper (you can search online for it). Thanks.
please can you help me to test the weak form of market efficiency on stata?
i have to fit this : r(it) = ai + E(it) an do wald test on it
Hi Abiboulaye, no idea.
Thank you ma. Please ma can you help with chow test? Thank you
FEYISAYO AYODELE ok Feyi, what what application are you using?
E views ma
FEYISAYO AYODELE Ok, I'm on transit right now, I'll get back to you in an hour..
Hello Feyi,
To perform Chow test in EViews:
1) Plot the graph of the series to identify the breakpoint
2) run the first regression w/o the dummy variables and their interaction terms
3) perform the Chow test: go to View, Stability diagnostics, Chow test, impute the breakpoint date, click ok
4) if the F-stat is significant, then the null hypothesis of no break is rejected
5) take the plot of the CUSUMSQ and you'll see that the plot deviates out of the 5% significance level boundary
To correct this:
1)create a dummy variable, (d) which takes 0 for years w/o break, 1 for breakpoint years
2) generate new series with the dummy variable for all the regressors in the model. That is, generate d*X1, d*X2, d*X3 etc
4) run the second regression with all the regressors and the newly generated series of the the dummy variables and their interaction terms
5) take the plot of the CUSUMSQ and you'll see that the plot stays within the 5% significance level boundary.
...try this and let me know how it goes. I intend to do a video on this within the week. So send me your data to cruncheconometrix@gmail.com, I may decide to use it for the online tutorial on structural break.
Feyi,
May I know your institution and program of study?
Hi how do you decide between Pool OLS or fixed effect?
Do a Google search on "poolability test".
Bosede Ngozi Adeleye thank. Can we see F test in fixed effect model.
@@hllp2497 That will depend on the software you are using which you can verify from the output.
Hi, I cannot manage to import the data, error message is: ... not Stata Format. What I can do? I am very new to Stata. Thanks in advance
Hi Denisa, please watch my video on "reshape wide to longitudinal data". You will find it very helpful. May I know from where (location) you are reaching me? Thanks.
when I used Hausman fe test /
what is mean this result? I'd like interpret this sentence: (V_b-V_B is not positive definite)
Concern yourself with the pvalue. If significant use the FE estimator and RE estimator if otherwise. To interpret that sentence is to get involved in some econometric jargons which I'm not prepared. But if you are so interested, then seek other online resources for the meaning.
Hello madam, what if my Hausman test is significant at = 58.78
Prob>chi2 = 0.0979?
That's 10% significance. You may decide to reject the null hypo and use FE estimator.
Hello Madam, Is it correct if I use fixed effect model and random effect model on level data when some of my variables are stationary at first difference
Hi Amitha, FE and RE models are performed on short-panels (when N>T). But with long panels when N
Thanks madam, Can you advice me what is the method I should use to analyse relationship between dependent and independent variables of a panel which comprises 9 cross sections for 20 years
Why we create year dummies in stata
To control for variation of the depvar across time.
how to perform hausman test on multiple variables without creating dummies
Kindly watch my videos on Hausman test. Well explained.
why your test is different with this vedio?: th-cam.com/video/MXk8ry7H0sU/w-d-xo.html
That doesnt say about fixed/random
Read the Hansen paper to fully comprehend the test.