Thomas Obitz
Member
Hello,
I am trying to make sense of Dowd's chapter 3, example 1.6 - and I simply do not get it. I am wondering whether there a mistake in the example?
1) In the second paragraph, he calculates p as the probability mass left of the lower boundary q - h/2. That looks arbitrary. Hull (in Risk Management and Financial Institutions, chapter 14.2) simply uses 1 - confidence = 5%.
2) When he calculates f(q), he takes the probability mass between +/- h/2 and plugs that into the formula. However, I understand that f(q) is actually the PDF at the quantile. So he would have to divide by h, which reduces the standard error considerably. That would also remove the dependency on the choice of h, which he complains about in the text, but which does not make any sense to me, as h is simply a variable which has to be infinitely small. Again, Hull simply uses the PDF.
Any advice?
Regards
Thomas
I am trying to make sense of Dowd's chapter 3, example 1.6 - and I simply do not get it. I am wondering whether there a mistake in the example?
1) In the second paragraph, he calculates p as the probability mass left of the lower boundary q - h/2. That looks arbitrary. Hull (in Risk Management and Financial Institutions, chapter 14.2) simply uses 1 - confidence = 5%.
2) When he calculates f(q), he takes the probability mass between +/- h/2 and plugs that into the formula. However, I understand that f(q) is actually the PDF at the quantile. So he would have to divide by h, which reduces the standard error considerably. That would also remove the dependency on the choice of h, which he complains about in the text, but which does not make any sense to me, as h is simply a variable which has to be infinitely small. Again, Hull simply uses the PDF.
Any advice?
Regards
Thomas
