Mean-Variance Model
From Open Risk Manual
Definition
A Mean-Variance Model is any of the underlying analytic portfolio models that underpin wikipedia:Modern portfolio theory. A mean-variance model is built on a limited set of parameters namely, mean, standard deviation and correlations. [1]
Advantages
The small and intuitive number of parameters makes the calibration of these models relatively straightforward. Mean-variance models produce reasonable return projections, have nice mathematical properties and are easy to use, interpret and understand.
Issues and Challenges
Mean-variance models present problems when
- dealing with business applications that require analysis over multiperiod time horizons
- when the analysis requires direct access to economic variables
- when the impact of scenarios is highly non-linear, exposing the limitations of simple distributional assumptions
References
- ↑ Economic Scenario Generators, A Practical Guide, Society of Acturaries