วีดีโอ

😊

Glad you liked it!

Yes, I did wonder as well. I check his content - but it seems to be all removed. Let's hope he is fine

What exactly is confusing about it?

This is the way. Watch to the end!

In panel data, the standard approach is to use the cluster option in Stata, which also accounts for serial correlation.

You can use time and firm dummies. However, there is a risk of overfitting. You might get great in-sample model fit - but once you try to predict out-of-sample the quality drops.

Thanks, much appreciated!

It should be here: github.com/GerhardKling/DataScienceStata/tree/main/Lecture4

@@YUNIKARN thank you

If you are interested in pursuing a PhD at the University of Aberdeen, we can have a chat. Otherwise, I cannot provide any individual consultation due to regulations. All the best!

@@YUNIKARN its ok, let us talk for advice and make decision together , thank you

@@YUNIKARN I am still waiting for your support , let us talk and see if i can pursue it

I responded. If you want to study at the University of Aberdeen, I am happy to help. However, I cannot provide individual consultations

Glad it was helpful! May the Force be with you!

Well, the truth hilarious. I am working on a panel data course. You find two Stata courses on Udemy (www.udemy.com/user/gerhard-kling-2/ ). May the Force be with you!

I am glad the videos are useful. I am working on a panel data course (GMM, difference-in-difference etc.). Put video editing takes a very long time. Enjoy your research project!

Two years is short. Most PhD programmes will require three years. So you will need to think about self-funding one year or finding alternative arrangements (write-up periods etc.). I worked at Utrecht University. The Dutch system is quite different from the UK as most PhD students have a teaching position. Do you want to try the Dutch market or UK/US? You need a good fit with a potential supervisor. Good luck!

@@YUNIKARN It's entirely up to me, by the way; money is not a problem, nor is getting employment. I am prepared to provide for my own needs in order to maintain my research's flexibility. This is the reason I asked you to share your thoughts on the matter. I want to work closely with the university of my choice and be totally independent. That's the reason I'm funding myself. But I believe that two years is reasonable. I wish not to spend more time. I am at Erasmus University and TU Delft

@@alimtanta2934 I had PhD students who completed within 2 years (both with industry experience). With a good proposal, you should be able to get admission. I suggest contacting potential supervisors before making an application. Good luck!

​@@YUNIKARNThank you very much for your response. Your channel is very educational. I appreciated

Yes, I take it as a compliment

Glad it helped!

There is an implementation in Stata called xtbreak. My Udemy course on Applied Time Series explores structural breaks in a lot more detail (testing for knowns and unkown breaks)

@@YUNIKARN thanks for a good reply , now I have used the test and found two breaks in the data and wanted to use Lagrange multiplier Lee and Strazicich (2003) but could not find the command in stata I searched online could you please help with the command if it available in stata or even are thanks

@@bolisolwak1390 Only available in R or RATS as far as I know. Here is a link to the GitHub. So either run it in R/RATS or translate into Stata github.com/hannes101/LeeStrazicichUnitRoot

@@YUNIKARN Thanks once again, I am not familiar with that software if I could get code for R it would sound good to me do you have the link for R ?. Thanks.

@@bolisolwak1390 I suggest running R through Anaconda (www.anaconda.com/). Free download for R and R Studio. If you have not used R before, it will take a while to get used to

Thanks for watching! I am looking into the job market. It is not easy to get reliable data. On Monday, I will present some data on foreign direct investment. India or China - who is ahead?

A pleasure. Thanks for watching!

I am glad the video was useful. May the Force be with you!

Thanks, they did an awesome job!

Thanks, much appreciated. May the Force be with you!

Yes, N refers to time. Everything is discrete

You're most welcome

I have done more on panel data in my Udemy course - see www.udemy.com/course/getting-started-with-stata/ I am also working on new material but this will take quite a bit of time. May the Force be with you!

I did a video on "Why does Stata drop my variables?", which will be published on Monday 1st April 2024 - April Fool's Day Special. Link: th-cam.com/video/CgHueBDaEnU/w-d-xo.html If you have four dependent variables, it suggests that you will have four separate models to explain them. Of course, you can have FE or RE as these models are separate. If you do not know which one is actually dependent, then you need to explore causality (or causal order if you have a time dimension). May the Force be with you!

Video sources can be cited using urls. By the way, the book "Getting started with Stata" (based on the Udemy course) should be published in 2025. So you could also cite the forthcoming book. Thanks for wacthing!

See my previous response

All codes are on GitHub. I did not use any code to calibrate the parameters. This is a separate issue

Thank you, if possible can you share the code of how the calibration of parameters was done through email?

As I said, I did not cover calibration as this is a separate issue and requires a bit more finance background. I might do it in future

No space. Say variable names is VAR. Then d.VAR without space

@@YUNIKARN it still doesn't work

Have you tsset your data. Stata needs to know the time dimension

You can do your on VIF. Just regress all explanatory variable one by one on the set of all other explanatory variables. Obtain the R squared. The VIF is then 1/(1-R squared). Or add a constant to your regression model. The implementation of the VIF command is a bit odd in Stata. Second problem: the fixed effects dummies seem to explain other dummies in the model. Stata drops these to avoid collinearity (the inverse of the data matrix does not exist as the matrix does not have full rank). Check the definitions of your dummies

Thanks for the interesting question - not an easy decision. I wish you all the best! Best wishes, Gerhard

Really good question. There are many things to consider (e.g., pension). I actually recorded a video. It should be ready on Monday 11th Macrh at 5pm UK time. May the Force be with you!

@@YUNIKARN Thanks, Professor. I am waiting for the video.

No, see gets home made chicken. I am also not vegan - just had a vegan day

​@@YUNIKARNdon't have too many vegan days , that food looks like plasticene food

Hahaha ... Will do a meat day soon

@@YUNIKARNI am so happy that you and your dog aren’t vegan, I already wanted to comment about it saying how bad it is but I see that you, nor the dog are vegan👍

Glad it was helpful!

I have a playlist focused on Excel for data analysis (th-cam.com/play/PLtKGNX49fOCbdq_XS4ciWCW8OfZgNqJXc.html) - but no material on Power BI at the moment. I am working on a new Udemy course on Company Valuation, which will use Excel ... more to come.

Thanks, much appreciated! May the Force be with you!

I am glad you find this material useful. The fully revised course with all codes and datasets is available on Udemy (see link below and on my channel pages) www.udemy.com/user/gerhard-kling-2/

@@YUNIKARN So now we need to pay to use the data? ☹ Thank you anyway!

Most datasets can be obtained from public sources for free

I would keep them in separate sections. Usually, regressions use abnormal returns (derived from event studies) as dependent variable. You need to be careful with outliers. Event studies are known to produce outliers. I did a series on event studies. This is the first of four videos th-cam.com/video/ZJRaON2kHJ4/w-d-xo.html

Finance tends to be a better choice as it is mostly an empirical discipline. So your quantitative skills will be in demand. Moving from engineering into asset pricing is very common as well. However, most economists tend to prefer applicants with a BSc/MSc in Economics. This is also an issue if you want to enter the acadenic job market. All the best!

@@YUNIKARN Thanks for the reply. I am also looking at what options should I look into for research. Like you suggested asset pricing and its a new term for me. Are there any more terms I should know? Also what are my options after a PhD in Finance? I saw video which said most of them go into academia.

@@akkshheyagarwaal7629 Academia is not always the best option after a PhD. You must have a plan B. I did a few videos focused on the academic job market, which might be useful. I also covered some aspects of topic choice (focused on dissertations) in other videos. Drop me an email if you have any further questions (email on channel pages). May the Force be with you!

Hello Dr. Kling, I could not find your email on your channel. I have sent you a request on LinkedIn to connect.

@@akkshheyagarwaal7629 LinkedIn rules!

Sure, but it is less coding and hence less fun 😉

Glad it was helpful! Happy New Year!

Thik hai - have a great 2024!

I have done a video focused on PhDs in Finance th-cam.com/video/fpHrMxsTeik/w-d-xo.html - but many issues also apply to your case. BTW, I completed a BSc/MSc Maths and started a second PhD in Maths, which I did not complete. In general, if your aim is to get into industry, do not do a PhD - except in very high-demand areas (AI/ML or Quantitative Finance if your topic is applied). It is better to focus on getting industry experience. Do you want to do a PhD in Maths or in a different area? May the Force be with you!

@@YUNIKARN Yeah, the problem is I don't know yet. I did my dissertation in "weather derivatives" which we used a mean-reverting Ornstein-Uhlenbeck process to price a financial product; that nearly killed me! My BSc was in engineering and I think I don't feel confident enough for a Math PhD. I was thinking maybe PhD in computational finance which is more inclined towards ML/computer science would be a better choice. I know one thing though, I love this sector of Quant Finance (commodity derivatives/CAT bonds,...) so much that I'm willing to put the work in. I've found an advisor in University of Glasgow, but I'm not sure if UK is the place I wanna do my PhD anymore. US has more opportunities/more funding. Anyway, sorry for my looong comment, I'm just lost. God bless.

@@6Ligma Glasgow is a great university. I suggest applying for several PhD programmes and scholarships in the US. Just try and see whether you get good offers. Ideally, stay applied whether it is more CS or Finance is a matter of choice. All the best!