### CFA Practice Question

There are 985 practice questions for this topic.

### CFA Practice Question

Excess kurtosis is a problem for investment researchers using normal distributions because ______

A. historical returns are better modeled with leptokurtic distributions.
B. it is difficult to calculate.
C. the likelihood of extreme outcomes will be underestimated.

Historical returns are better modeled with fat-tailed (or platykurtic) distributions, not leptokurtic ones. If a researcher uses a normal distribution to model a fat-tailed distribution, the estimated probability of an extreme outcome will be underestimated. Thus, the frequency of market crashes will be underestimated (an underestimation of volatility risk).

User Comment
surob I thought Leptokurtis is fat-tailored. It has larger tails. I am confused. Can anyone please explain?
jallado0 try this link surob http://mathworld/wolfarm.com/leptokurtic.html , i hope it makes it clear for u
bobert If there is excess kurtosis it means bigger tails (leptokurtosis) because the top is pulled higher than would be on a normal distribution. Therefore if there are going to be extreme outcomes, they will be underestimated using the normal because it has shorter tails than a distribution with leptokurtosis.

You are using a normal to estimate what is really leptokurtic. There will be more returns centered around the mean but there will also be more extreme values at the tails.

Hope that helps.
Sumit14 Normal Curve is Normal
Leptokurtosis - pull the normal curve up from the mean so that the whole curve moves up along with the tails. this will make the Leptokurtosis curve thinner at the center.
Platokurtosis - push the normal curve down from the mean so that the whole curve moves down along with the tails. this will make the Platokurtosis curve thicker at the center.
rfvo To further elaborate.....Positive values of kurtosis do not indicate a distribution that has fat tails. Positive values of excess kurtosis (kurtosis > 3) indicate fat tails.
tschorsch try this image