You can find the spreadsheets for this video and some additional materials here: drive.google.com/drive/folders/1sP40IW0p0w5IETCgo464uhDFfdyR6rh7 Please consider supporting NEDL on Patreon: www.patreon.com/NEDLeducation
Fantastic video! I have been looking for a good way to deal with correlated variables for a long time. I had heard of Cholesky decomposition but didn't know it could be used for this.
Hi, and thanks so much for the kind words! I have got a separate video on Cholesky here that covers the concept in a lot more detail: ru-vid.com/video/%D0%B2%D0%B8%D0%B4%D0%B5%D0%BE-B0wj9xYqw5Y.html
Further to my last comment - is that what forces the correlations to be maintained and therefore the portfolio means and standard deviations to be similar
Thanks… just submitted one of my assignments :) … with JetFuel Prices correlated with Exchange rate against US$… both moving and yet have some correlation..
Awesome video. How would you model correlated electricity and fuel prices in this way? Which distribution would you assume? Also how do you accurately estimate the correlation of time series data? Thanks! I’ve already learnt a ton, you’re a star
Loved the video! Was wondering about the literature behind the method applied here for generating the correlated simulated returns. I would like to reference it in my paper! Thanks!!
Hi Labros, and glad you liked the video! This is one of the statistical methods that everybody knows but few know the original source it came from :) I believe you can use it without referencing (found many papers that do so), but my best bet at who proposed it originally is Solow (1985): link.springer.com/article/10.1007/BF01031616
Hello sir nice video but am wondering whether Cholensky Decomposition can be used if Distribution assumed is log-normal instead of normal? If not could you make a video where you show alternative method on how to do this if we assume Returns Distribution is log-normal?
great video - as usual - just wondering why standard deviation (variance) is often multiplied by square of time period (specially 252 ) ? i noticed that you haven't done that in the volatility calculation in this simulation . - thanks for the lessons once again
Hi Amit, and glad you liked the video! This is to annualise daily data - as there are 252 trading days in a year and volatility scales as a square root (variance scales linearly) with time. Similarly, you can use SQRT(12) to annualise monthly data or SQRT(52) for weekly data.
very helpful. can you explain how to do this with a portfolio of commodity futures including longs and shorts? for example, i may be spreading gold, so im long june and short dec. Thanks
Great video, as always. Out of curiosity (I can't find it amongst your videos) do you have a video on the calculations of the five factor model, and your thoughts about it? Otherwise that is a topic I would love to suggest! Best W
@@NEDLeducation it's perhaps out of scope for you videos, but it would be interesting to hear your thoughts on some of the models being presented. The good, the bad, the ugly. Perhaps that's for another playlist, but I would listen to it :)
@NEDL Awesome video! Made this very difficult concept very easy to digest and understand, well done! I don't know if it's possible, but was wondering if I could reach out to you and ask a few questions about something related to this topic, but using it in what might be an unusual fashion? Let me know and if not, no hard feelings! Either way, thanks for the great video!
Hi Terry, and thanks for the comment! Yes, it is possible to reach out to me with questions and even schedule a one-to-one chat - I provide this as my Patreon benefits :)
Thank you so much for the video. I implemented it in a totally different field and it works perfectly. My only question is how to implement the same method for different types of distributions (Beta, Gamma, Lognormal)? I would highly appreciate if you can just give me a hint. @NEDleducation
