Leptokurtosis: Difference between revisions
From ACT Wiki
Jump to navigationJump to search
imported>Doug Williamson (Added links) |
imported>Doug Williamson (Classify page.) |
||
| Line 3: | Line 3: | ||
It means a more ‘pointy-headed’ and ‘fat tailed’ observed distribution, compared with the distributions predicted by the normal and lognormal models. | It means a more ‘pointy-headed’ and ‘fat tailed’ observed distribution, compared with the distributions predicted by the normal and lognormal models. | ||
Importantly there is a fatter downside tail (‘left tail’) in the observed data. | |||
Importantly, there is a fatter downside tail (‘left tail’) in the observed data. | |||
In other words the observed frequency of large negative returns (or results) is greater than predicted - for example - by the lognormal model of the distribution assumed in the Black Scholes option pricing model. | In other words the observed frequency of large negative returns (or results) is greater than predicted - for example - by the lognormal model of the distribution assumed in the Black Scholes option pricing model. | ||
Because of leptokurtosis, Value at Risk models which use a normal frequency distribution will understate the Value at Risk. | Because of leptokurtosis, Value at Risk models which use a normal frequency distribution will understate the Value at Risk. | ||
| Line 18: | Line 20: | ||
* [[Tail]] | * [[Tail]] | ||
* [[Value at risk]] | * [[Value at risk]] | ||
[[Category:The_business_context]] | |||
Latest revision as of 17:36, 1 July 2022
Leptokurtosis is observed in many financial distributions.
It means a more ‘pointy-headed’ and ‘fat tailed’ observed distribution, compared with the distributions predicted by the normal and lognormal models.
Importantly, there is a fatter downside tail (‘left tail’) in the observed data.
In other words the observed frequency of large negative returns (or results) is greater than predicted - for example - by the lognormal model of the distribution assumed in the Black Scholes option pricing model.
Because of leptokurtosis, Value at Risk models which use a normal frequency distribution will understate the Value at Risk.
