One again, I agree with sle that normalizing by ATM IV is a good idea since it breaks down the positive link b/w level and skew, and make econometric/statistical modeling easier since you can model two processes separately without worry about joint dynamics.
Think like this: it is normal for skew to be 6 vol points b/w 25 delta and 75 delta when 50 delta IV is 30. but when 50 delta vol is 80, 6 skew points are nothing, while 16 points, which is still 20% of ATM vol is the more likely scenario.
sle: is my interpretation of your statement "if vol level and skew are correlated, you are not normalizing them correctly" now correct?
njrookie
Think like this: it is normal for skew to be 6 vol points b/w 25 delta and 75 delta when 50 delta IV is 30. but when 50 delta vol is 80, 6 skew points are nothing, while 16 points, which is still 20% of ATM vol is the more likely scenario.
sle: is my interpretation of your statement "if vol level and skew are correlated, you are not normalizing them correctly" now correct?
njrookie
