Skip to content

Mean shift versus dispersion change in normal risk

A five-unit mean shift changes both summaries by five; a five-unit dispersion increase changes them by different amounts.

On this page

Use the original 95% baseline

Start from mean loss 0 and dispersion 10. VaR is 16.448536 and ES 20.627128 loss units under the declared normal law.

Change one parameter at a time

Original parameter comparison
RecordMeanDispersionVaRES
Baseline01016.44853620.627128
Mean shift51021.44853625.627128
Dispersion increase01524.67280430.940692

Identify the reason for each change

The mean shift adds five to every loss percentile and its selected-tail mean. Increasing dispersion stretches the centered loss law: the quantile rises by about 8.224268 and ES by about 10.313564.

Retain a controlled model comparison

These are separate assumption records. A real position change may affect location, dispersion, shape and dependence together. The table isolates the entered model parameters rather than claiming an investment intervention changes only one.

Further references