ES is a complement to value at risk (VaR). ES is the average loss in the tail; i.e., the expected loss conditional on the loss exceeding the VaR quantile. For more financial risk videos, visit our website! www.bionicturtl...
David, your video illustrations of risk and finance topics have always helped me crystalize concepts in Risk. So big thank you! Keep up the great work.
@tiburonski: interesting distinction...as they are both weighted quantiles(spectral measures), I'd *think* it's okay to view ES as a conditional quantile, but since it's an average not a median, I see your point...i think the metaphor can be "convulted" (like Oprisk) with freq distribution (crash or no?) and severity. VaR = Prob no crash (0 deaths); e.g., "I won't die next year with 99.99x% confidence. Then ES = prob of death if crash; e.g., if i do crash, my ES = 0.9x probability of death.
Good idea, I honestly like the Rachev ratio, too (alas, it doesn't appear on any of our exams, yet, but it should be on the FRM imo). Thanks for the suggestion!
Definition of alpha in table is different with the formula. Given that (1 - alpha) is confidence level (defined in table), the formula is averaging the cumulative loss that exceeds quantile over the non-red area instead of the red one. It is important note that finance people sometimes defined the alpha different with statisticians. [ alpha = 95% is defined as 95% confidence level such that q(0.95) = 1.645.] Therefore, the formula is correct when alpha is defined as confidence level.
please correct me if I am wrong, but I get $40 for the Tail-VAR (average loss size above the VARalpha = 200/5) and the Expected Shortfall of $2 where expected loss given loss above VARalpha is (1-alpha)*TailVAR. Look forward to your reply. thanks
They should have used cdf. this is confusing. Also the formula is incorrect, by way of convention. alpha is the type 1 error or 1-confidence level. The integration is from 0 to alpha and the averaging (conditional expectation), i.e. the denominator is 'alpha' not 1-alpha. I am so glad I didn't take Bionic Turtle's prep.
Given that you, Bionic Turtle, have already talked about ES/VaR, why not talk about the generalized Rachev ratio, a significantly better optimization tool than the Sharpe ratio for non-normal risk.
Hi, thanks for this. I'm not finding the first example as intuitive as the second example. Could you include the formula in ccolumn F, especially cell F6 as this would help my understanding. Thanks in advance.
+Johnny Suryoadji he's representing losses, i.e. a positive number means it's a loss. the negative area will be a negative loss, i.e. a profit (hope this makes sense!)
