New Practice Questions
1. How many durations? Just one, really! @Branislav posted here (https://trtl.bz/2FEklWn) an impressive summary of what's he's learned about duration in the FRM. I added my agreement that, in fact, there is only one duration concept but it has three faces: Macaulay, modified and effective. We've been discussing bond duration and convexity for over a decade in the forum, and I've learned so much from candidates and practitioners. When I started teaching the FRM, for example, I didn't realize the units for all of them are properly years, until a member posted the proof! Yea, I still remember that day when I was so ... um, wrong. You might think I know my stuff, but hosting the forum means that I often get "owned," mathematically speaking
.
2. How many correlations? Many, actually! This week saw several threads about correlation, including more interpretation of Malz's implied default correlation (https://trtl.bz/2FAcvgo) and whether Malz comports with Meissner (https://trtl.bz/2FSxyv9). Also, when does an increase in correlation imply less risk? https://trtl.bz/2FL4uWl
3. Square-root-rule (SRR): Here is a smart question which does not take for granted our commonly used square-root-rule https://trtl.bz/2FJ1rxz We routinely scale a daily volatility or VaR over ten days by multiplying by sqrt(10). But what are the conditions for doing that, including are we technically making an assumption about whether the returns are arithmetic or geometric? (here's a hint: the SRR does not require normality).
External
1. Free rstudio conference materials: In January, I attended the rstudio::conf (https://www.rstudio.com/conference/) in Austin. I also went to last year's in San Diego and hope to go next year when it's held in San Francisco. This is probably the most important conference for the R community and people routinely say it's the best conference they've ever attended. I've long believed that developing some data science fluency (if not some code practice) is a good career plan for those of us in technical (non-sales) fields such as financial or risk analysis, even if we are not developers. Hence my multi-year personal investment in learning data science where I happen to prefer R, but of course Python is very popular. Typical of the utterly generous community around R, the talks and materials for that conference have been made freely available at the resource page https://resources.rstudio.com/rstudio-conf-2019 (and materials here on github at https://github.com/rstudio/rstudio-conf/tree/master/2019)
2. Better board practices? In the last WIR, I pointed to a resource on the Theranos debacle. Although I admit I'm interested in narratives related to individual psychology and startup culture, I am also mostly fascinated by the utter failure in governance. John Carreyrou's book did not really solve this aspect of the puzzle for me: clearly Elizabeth Holmes populated her board with "cabinet members, congressmen and military officials" to create prestige and distract from actual science, but how exactly is an entire well-compensation board allowed to completely fail to perform any of its duties? In my previous life I was a management consultant to boards, and while most were impressive, it led me to believe that traditional boards (especially marquee hires) are one of capitalism's Achilles' heel. The performance standards for board members at many public companies should be much higher. So I enjoyed Board 3.0: An Introduction https://corpgov.law.harvard.edu/2019/03/26/board-3-0-an-introduction/ published at the Harvard Law Forum, which is a great resource on governance practices. They also just published: Crisis Resilience and the Board—Taking Risk Oversight to the Next Level (https://corpgov.law.harvard.edu/201...oard-taking-risk-oversight-to-the-next-level/)
3. Many (mental) models: As an investor and student of risk, I'm attracted to the idea of mental models as a means of coping with complexity . Shane Parrish says "a mental model is simply a representation of how something works. We cannot keep all of the details of the world in our brains, so we use models to simplify the complex into understandable and organizable chunks." (https://fs.blog/mental-models/). I'm almost finished reading The Model Thinker by Scott Page (https://amzn.to/2FOaHle). This is a strong reference-like catalog of about 25 models--only barely quantitative in exploration--including many that I had never heard of before. How many of them do we study in the FRM? Let's see, I count at least seven: normal distributions, power-law distributions, linear models, concavity and convexity, random walks, path dependence, and Markov models. Not bad, GARP! What's next on my reading list? Oh geez, where to start. I am looking forward to Infinite Powers: How Calculus Reveals the Secrets of the Universe, excerpted here https://www.sciencefriday.com/articles/the-language-of-calculus/
- P1.T4.912. Key rate exposure technique in multi-factor hedging applications (Tuckman Ch.5) https://trtl.bz/2U4wsGf
- P2.T6.903. The International Swaps and Derivatives Association (ISDA) Master Agreement (Gregory Ch.4) https://trtl.bz/2U17iIN
- Fixed Income: Bond's full/flat price on settlement date (FRM T4-22) https://trtl.bz/2V0Jxgg
- TI BA II+: How to compute bond price on realistic (between coupons) settlement date (TIBA-02) https://trtl.bz/2TI25AB
- R Programming Introduction: Matrices (R intro-05) https://trtl.bz/2uD1buZ
1. How many durations? Just one, really! @Branislav posted here (https://trtl.bz/2FEklWn) an impressive summary of what's he's learned about duration in the FRM. I added my agreement that, in fact, there is only one duration concept but it has three faces: Macaulay, modified and effective. We've been discussing bond duration and convexity for over a decade in the forum, and I've learned so much from candidates and practitioners. When I started teaching the FRM, for example, I didn't realize the units for all of them are properly years, until a member posted the proof! Yea, I still remember that day when I was so ... um, wrong. You might think I know my stuff, but hosting the forum means that I often get "owned," mathematically speaking
.2. How many correlations? Many, actually! This week saw several threads about correlation, including more interpretation of Malz's implied default correlation (https://trtl.bz/2FAcvgo) and whether Malz comports with Meissner (https://trtl.bz/2FSxyv9). Also, when does an increase in correlation imply less risk? https://trtl.bz/2FL4uWl
3. Square-root-rule (SRR): Here is a smart question which does not take for granted our commonly used square-root-rule https://trtl.bz/2FJ1rxz We routinely scale a daily volatility or VaR over ten days by multiplying by sqrt(10). But what are the conditions for doing that, including are we technically making an assumption about whether the returns are arithmetic or geometric? (here's a hint: the SRR does not require normality).
External
1. Free rstudio conference materials: In January, I attended the rstudio::conf (https://www.rstudio.com/conference/) in Austin. I also went to last year's in San Diego and hope to go next year when it's held in San Francisco. This is probably the most important conference for the R community and people routinely say it's the best conference they've ever attended. I've long believed that developing some data science fluency (if not some code practice) is a good career plan for those of us in technical (non-sales) fields such as financial or risk analysis, even if we are not developers. Hence my multi-year personal investment in learning data science where I happen to prefer R, but of course Python is very popular. Typical of the utterly generous community around R, the talks and materials for that conference have been made freely available at the resource page https://resources.rstudio.com/rstudio-conf-2019 (and materials here on github at https://github.com/rstudio/rstudio-conf/tree/master/2019)
2. Better board practices? In the last WIR, I pointed to a resource on the Theranos debacle. Although I admit I'm interested in narratives related to individual psychology and startup culture, I am also mostly fascinated by the utter failure in governance. John Carreyrou's book did not really solve this aspect of the puzzle for me: clearly Elizabeth Holmes populated her board with "cabinet members, congressmen and military officials" to create prestige and distract from actual science, but how exactly is an entire well-compensation board allowed to completely fail to perform any of its duties? In my previous life I was a management consultant to boards, and while most were impressive, it led me to believe that traditional boards (especially marquee hires) are one of capitalism's Achilles' heel. The performance standards for board members at many public companies should be much higher. So I enjoyed Board 3.0: An Introduction https://corpgov.law.harvard.edu/2019/03/26/board-3-0-an-introduction/ published at the Harvard Law Forum, which is a great resource on governance practices. They also just published: Crisis Resilience and the Board—Taking Risk Oversight to the Next Level (https://corpgov.law.harvard.edu/201...oard-taking-risk-oversight-to-the-next-level/)
3. Many (mental) models: As an investor and student of risk, I'm attracted to the idea of mental models as a means of coping with complexity . Shane Parrish says "a mental model is simply a representation of how something works. We cannot keep all of the details of the world in our brains, so we use models to simplify the complex into understandable and organizable chunks." (https://fs.blog/mental-models/). I'm almost finished reading The Model Thinker by Scott Page (https://amzn.to/2FOaHle). This is a strong reference-like catalog of about 25 models--only barely quantitative in exploration--including many that I had never heard of before. How many of them do we study in the FRM? Let's see, I count at least seven: normal distributions, power-law distributions, linear models, concavity and convexity, random walks, path dependence, and Markov models. Not bad, GARP! What's next on my reading list? Oh geez, where to start. I am looking forward to Infinite Powers: How Calculus Reveals the Secrets of the Universe, excerpted here https://www.sciencefriday.com/articles/the-language-of-calculus/
Last edited:
