This is our first video, in what we hope will be a series of videos explaining many of the concepts of modern finance. Let us know what you thought, and if you have any suggestions as to future videos, please let us know!
You did something that my professor failed to do, I appreciate the amount of effort you put in for this video and surprised to see the subscriber count, am interested in knowing what tools you have used to make the animations.Thanks a lot by the way i really appreciate it
I’m really glad you enjoyed it! I absolutely love explaining finance and financial concepts. These videos take a while for me to create and I’m working on one now explaining Black Scholes in detail. As for the software, all the puppets and backgrounds are made by me, by hand, in Adobe Illustrator, with the exception of some backdrops and props which are edited pictures in Illustrator. The video is compiled and animated using Adobe After Effects. Hope that helps!
The content is really well presented and very useful. One comment: would you have the courage to drop the “warnings” about math in the discussions? How can it be conceivably acceptable for anyone dealing or wanting to deal with finance not to be willing to deal with the corresponding math???
It's not a matter of courage, it is a matter of ensuring the "right" audience is viewing the video. I can see in the metrics a major drop after the warning is given - this is good, as it prevents viewers who don't want to watch the video seeing it and then giving it a "dislike", which harms my content. As for not wanting to deal with the math - you'd probably be surprised by the number of financial professionals who find mathematics "boring" and "tedious".
Thanks for great tutorial. I have two questions. 1. For portfolio volatility why not simply calculate volatility of each asset's return (via standard dev) and sum them? 2. I've calculated portforlio volatility of 1 years worth of closing price of three stocks (Air canada, AMZN, and Suncor Energy), volatility is very low 5.9*10^-11. Is it supposed to be very low?
Hi Haneul, and thanks for the questions. 1. A portfolio's volatility cannot be found by summing standard deviation, or even using a dot product (sumproduct) for the values. The reason for this is that there are covariations which exist between these assets. Let's say that you have two equally volatile positions A and B. However, A and B have inverse volatilities (when one is up the other is down). If you only take their respective standard deviations and apply to their respective sizes, your volatility will be the standard deviation of A and B. However, because when A is up and B is down, the portfolio actually has lower risk, likely near 0 (hypothetically) because the two cancel each other out (perfectly in our scenario which is not realistic but useful as an example). This is why diversification among uncorrelated assets is important! It can actually reduce risk. 2. This is probably incorrect. One issue that you might be facing is that you have not annualized the data. You have taken daily returns, and therefore you must multiply the standard deviation value by the SQRT(252)
@haneulkim4902 The other issue might be that you are taking just the closing prices. You need to take the percentage change between days. This is also likely your issue. Take the (current day - previous day)/previous day for each day to get the daily percent change.
@@financeexplainedgraphics Thank you! Why do we need to annualize the standard deviation? and I did use return (= price_t -1/ price_t), not closing price. Lastly please post more videos~ your videos are very well structured and informative :)
@@haneulkim4902 Glad I could help! The reason you want to annualize is so that it "looks" right. You don't have to annualize, but it is best practice. Most times when professionals quote volatility they do so on an annualized basis. When you say a security has an volatility of 20% annualized, it is more familiar to investors than saying your volatility is 1.26% - so when you said your vol was extremely low, I thought it might not have been unannualized (but 5.9*10^-11 is WAY to low 😂 even for daily vol). As for the videos, I am trying. They take a very long time to make and I am a bit of a perfectionist. I also have a day job which takes up a lot of the energy I would otherwise put into making videos. However, I am so glad you liked it! Still working on Black Scholes, which is a beast of an equation to explain in ~15 minutes.
@@financeexplainedgraphicsI see, the reason it was so low was because for calculating return I was using (Return = price at t / price at t-1) instead of (Return = price_at -t - price_at_t-1) / price at t-1. Wow already excited for the next video, thanks again for an amazing video and answering my questions!
Optimizers can do it for you, but as a beginner I strongly believe it is important to learn the fundamentals of the theory (risk vs. reward and volatility as risk). The math is secondary and for someone without knowledge in linear algebra it could be more challenging. If you use an optimizer it can be convenient, but it does little to help you understand what is going on and why it is determining what it has calculated. One thing investment apps don't do is calculate the expected return, which is necessary for computing the optimization. So developing your own theories and investment strategies is still key.
Figure out/eQasian Q..uestion works , it is mathematic B and A level. You go in a mathematic web site and when you see there your skills, than you need to start reading from year 5, so you make sure that you need to connect the past knowledge with new formulas. Ask more Q..uestions , dont bother to ask somethink clever.
You can calculate in Excel or any software you prefer. The linked Excel doc shows how this is done. Standard calculators do not perform these operations, and doing it by hand is very impractical.
At minute 2:46 What do i put into the covariances in the matrix?, The mean of the historical return, or the vector or what. Im still in highschool and wana learn this so please help me out
Impressive that you want to work on this while still in high school. Looking at the 2:46 time stamp, think of it this way: You have assets (stocks, bonds, whatever) and each are called A, B, and C. The vectors (vectors are essentially single column "matrices") are the historical returns of the asset. For instance, you might have a stock A that has a return vector of [+0.2%, -1.1%, +1.3%, +0.7%, -0.8%] over a given 5 day period. Now you do this for each of your assets and find what their returns were each day (or month or whatever you like, just be consistent and make sure all assets have the same number of observations). Then, you are going to create a matrix. A matrix has rows and columns, and a covariance matrix, like the one in our video, is ALWAYS square (number of rows = number of columns). HOWEVER, the size of the matrix will depend on the number of assets. We have three assets, so the matrix will be 3x3. If we have 10 assets, the matrix would be 10x10. So, assuming we have 3 assets, we will have a 3x3 matrix, which means our matrix will have 9 data points. Each data point is the covariance of two of our vectors, which equals a single number. A covariance is a statistical term and you can read up on in a number of places (Wikipedia is great for this). If you are doing this in Excel (which I recommend as a great first step to financial modeling), you can use a function covariance.s( ). Inside the parenthesis you will put the range (your vectors) of returns for the assets separated by a comma. Make sure you have the same number of observations in each range (vector). You can see this in the Excel document in the video description. Now, you have your matrix, which will shown (in this case) 9 covariances. However, there will be something interesting about this matrix. You will see that it is special, because it is a symetric matrix. If your matrix is not symetric, you have messed up. Why is this a symetric matrix? Because cov(A,B) is the same as cov(B,A), etc. Here is something else that is interesting, if you found the variance of any single vector (A, B, or C) that variance would be the same as the cov(A,A) for variance A, and cov(B,B) for variance B, and so on. So what you have in your matrix is the variances of A, B, and C running along the diagonal from the top left to the bottom right, and in the fields you have a mirror of values which are the covariances of A,B B,A A,C C,A B,C and C,B. In other words, our 3x3 matrix looks something like this in variable terms: | varA X Y | | X varB Z | | Y Z varC | That's probably more than you asked for. I highly recommend you play around with the excel file. This math is well beyond what most people are taught in high school. If you are interested in learning more about this type of math, you should explore classes on Linear Algebra, which is the specific mathematics applied to matrix and vector relationships (just like the ones in the video). I highly recommend the MIT Classes that are free on TH-cam for Linear Algebra (Professor Strang is a legend in this field and he recently retired but all his classes are available on their channel). You don't need to understand calculus at all to understand Linear Algebra, but I would highly recommend you familiarize yourself with the concept of Vectors and Matrices if this type of math and financial analysis interests you. Let me know if you have any more questions.
I wont dislike this but let it be clear I am complaining. This video is useful only to someone that has no idea about those. I was looking for a shortcut instead of having to read the book chapters but I guess there is no free lunch.
This is our first video, in what we hope will be a series of videos explaining many of the concepts of modern finance. Let us know what you thought, and if you have any suggestions as to future videos, please let us know!
I hope your channel grows!
Great and super content
خیلی عالی بود ممنونم
Great video! Recommend putting "Markowitz Model" in the title (or at least description) for improved SEO on other students looking to review for exams
Great suggestion!
@@financeexplainedgraphics I'm here to confirm that I looked this up and found it because the name was in the title 😆
The best explanation on TH-cam
I'm honored! Thank you!
great explanation! now things look more easier than in class. Thanks a lot!
You did something that my professor failed to do, I appreciate the amount of effort you put in for this video and surprised to see the subscriber count, am interested in knowing what tools you have used to make the animations.Thanks a lot by the way i really appreciate it
I’m really glad you enjoyed it! I absolutely love explaining finance and financial concepts. These videos take a while for me to create and I’m working on one now explaining Black Scholes in detail. As for the software, all the puppets and backgrounds are made by me, by hand, in Adobe Illustrator, with the exception of some backdrops and props which are edited pictures in Illustrator. The video is compiled and animated using Adobe After Effects. Hope that helps!
I have exactly the same experience. Thank you for this video❤
Excellent explanation. simple and easy to understand.. Thankyou so much!
You explain the concepts in finance really well! Please, make more of these. Now waiting for the video about beta 😊
It's on its way!
This was a great video and made this so much easier to understand than others I've seen
Good presentation and explanation. Thank you for teaching
Awesome work, this video was quite insightful
Glad it could be of help!
Beautiful video
excellent! thank you very much
Thank you for the video. A new subscriber.
OOOH my goodness, everything clicks now!
Awesome video - Thank you 🙏🏾
Ask your finance doc if MPT is right for you!
Pretty excelent explanation of Markowitz. The xls applies for fixed income too?
It does! Any asset can be modelled this way and added to the model
The content is really well presented and very useful. One comment: would you have the courage to drop the “warnings” about math in the discussions? How can it be conceivably acceptable for anyone dealing or wanting to deal with finance not to be willing to deal with the corresponding math???
It's not a matter of courage, it is a matter of ensuring the "right" audience is viewing the video. I can see in the metrics a major drop after the warning is given - this is good, as it prevents viewers who don't want to watch the video seeing it and then giving it a "dislike", which harms my content.
As for not wanting to deal with the math - you'd probably be surprised by the number of financial professionals who find mathematics "boring" and "tedious".
thank you!
Thankyou
Thanks for great tutorial. I have two questions.
1. For portfolio volatility why not simply calculate volatility of each asset's return (via standard dev) and sum them?
2. I've calculated portforlio volatility of 1 years worth of closing price of three stocks (Air canada, AMZN, and Suncor Energy), volatility is very low 5.9*10^-11. Is it supposed to be very low?
Hi Haneul, and thanks for the questions.
1. A portfolio's volatility cannot be found by summing standard deviation, or even using a dot product (sumproduct) for the values. The reason for this is that there are covariations which exist between these assets. Let's say that you have two equally volatile positions A and B. However, A and B have inverse volatilities (when one is up the other is down). If you only take their respective standard deviations and apply to their respective sizes, your volatility will be the standard deviation of A and B. However, because when A is up and B is down, the portfolio actually has lower risk, likely near 0 (hypothetically) because the two cancel each other out (perfectly in our scenario which is not realistic but useful as an example). This is why diversification among uncorrelated assets is important! It can actually reduce risk.
2. This is probably incorrect. One issue that you might be facing is that you have not annualized the data. You have taken daily returns, and therefore you must multiply the standard deviation value by the SQRT(252)
@haneulkim4902
The other issue might be that you are taking just the closing prices. You need to take the percentage change between days. This is also likely your issue. Take the (current day - previous day)/previous day for each day to get the daily percent change.
@@financeexplainedgraphics Thank you! Why do we need to annualize the standard deviation? and I did use return (= price_t -1/ price_t), not closing price. Lastly please post more videos~ your videos are very well structured and informative :)
@@haneulkim4902 Glad I could help! The reason you want to annualize is so that it "looks" right. You don't have to annualize, but it is best practice. Most times when professionals quote volatility they do so on an annualized basis. When you say a security has an volatility of 20% annualized, it is more familiar to investors than saying your volatility is 1.26% - so when you said your vol was extremely low, I thought it might not have been unannualized (but 5.9*10^-11 is WAY to low 😂 even for daily vol).
As for the videos, I am trying. They take a very long time to make and I am a bit of a perfectionist. I also have a day job which takes up a lot of the energy I would otherwise put into making videos. However, I am so glad you liked it! Still working on Black Scholes, which is a beast of an equation to explain in ~15 minutes.
@@financeexplainedgraphicsI see, the reason it was so low was because for calculating return I was using (Return = price at t / price at t-1) instead of (Return = price_at
-t - price_at_t-1) / price at t-1.
Wow already excited for the next video, thanks again for an amazing video and answering my questions!
As a beginner, is it worth it to figure out how this equation works? or do investment apps kind do that for you already?
Optimizers can do it for you, but as a beginner I strongly believe it is important to learn the fundamentals of the theory (risk vs. reward and volatility as risk). The math is secondary and for someone without knowledge in linear algebra it could be more challenging. If you use an optimizer it can be convenient, but it does little to help you understand what is going on and why it is determining what it has calculated.
One thing investment apps don't do is calculate the expected return, which is necessary for computing the optimization. So developing your own theories and investment strategies is still key.
@@financeexplainedgraphics thanks
Figure out/eQasian Q..uestion works , it is mathematic B and A level. You go in a mathematic web site and when you see there your skills, than you need to start reading from year 5, so you make sure that you need to connect the past knowledge with new formulas. Ask more Q..uestions , dont bother to ask somethink clever.
In 4:40 you forgot the indexes of r
Haha, wow! You are right, I did. I guess we can just pretend the indexes of w match r 😉
Great catch.
Wow
Is there a calculator to us for this matrix?
You can calculate in Excel or any software you prefer. The linked Excel doc shows how this is done. Standard calculators do not perform these operations, and doing it by hand is very impractical.
At minute 2:46 What do i put into the covariances in the matrix?, The mean of the historical return, or the vector or what. Im still in highschool and wana learn this so please help me out
Impressive that you want to work on this while still in high school.
Looking at the 2:46 time stamp, think of it this way:
You have assets (stocks, bonds, whatever) and each are called A, B, and C. The vectors (vectors are essentially single column "matrices") are the historical returns of the asset. For instance, you might have a stock A that has a return vector of [+0.2%, -1.1%, +1.3%, +0.7%, -0.8%] over a given 5 day period. Now you do this for each of your assets and find what their returns were each day (or month or whatever you like, just be consistent and make sure all assets have the same number of observations).
Then, you are going to create a matrix. A matrix has rows and columns, and a covariance matrix, like the one in our video, is ALWAYS square (number of rows = number of columns). HOWEVER, the size of the matrix will depend on the number of assets. We have three assets, so the matrix will be 3x3. If we have 10 assets, the matrix would be 10x10.
So, assuming we have 3 assets, we will have a 3x3 matrix, which means our matrix will have 9 data points. Each data point is the covariance of two of our vectors, which equals a single number.
A covariance is a statistical term and you can read up on in a number of places (Wikipedia is great for this). If you are doing this in Excel (which I recommend as a great first step to financial modeling), you can use a function covariance.s( ). Inside the parenthesis you will put the range (your vectors) of returns for the assets separated by a comma. Make sure you have the same number of observations in each range (vector). You can see this in the Excel document in the video description.
Now, you have your matrix, which will shown (in this case) 9 covariances. However, there will be something interesting about this matrix. You will see that it is special, because it is a symetric matrix. If your matrix is not symetric, you have messed up. Why is this a symetric matrix? Because cov(A,B) is the same as cov(B,A), etc.
Here is something else that is interesting, if you found the variance of any single vector (A, B, or C) that variance would be the same as the cov(A,A) for variance A, and cov(B,B) for variance B, and so on. So what you have in your matrix is the variances of A, B, and C running along the diagonal from the top left to the bottom right, and in the fields you have a mirror of values which are the covariances of A,B B,A A,C C,A B,C and C,B.
In other words, our 3x3 matrix looks something like this in variable terms:
| varA X Y |
| X varB Z |
| Y Z varC |
That's probably more than you asked for. I highly recommend you play around with the excel file. This math is well beyond what most people are taught in high school. If you are interested in learning more about this type of math, you should explore classes on Linear Algebra, which is the specific mathematics applied to matrix and vector relationships (just like the ones in the video). I highly recommend the MIT Classes that are free on TH-cam for Linear Algebra (Professor Strang is a legend in this field and he recently retired but all his classes are available on their channel). You don't need to understand calculus at all to understand Linear Algebra, but I would highly recommend you familiarize yourself with the concept of Vectors and Matrices if this type of math and financial analysis interests you.
Let me know if you have any more questions.
I wont dislike this but let it be clear I am complaining. This video is useful only to someone that has no idea about those. I was looking for a shortcut instead of having to read the book chapters but I guess there is no free lunch.
Sadly we cannot offer any free lunches 😔
cracks