Random portfolios are a general tool that undoubtedly have a number of applications yet to be discovered.
Random portfolios are a close analogue to the statistical bootstrap and to random permutation tests. Wherever there are portfolios with constraints there is likely to be an opportunity to use random portfolios to learn about dispersal and/or to compare with “what would happen by chance”.
Random portfolios are also a source of realistic portfolios that may be input to models.
Uses of random portfolios that have been discussed in the blog include:
- alpha decay
- the variability of strategies
- calibrating realized efficient frontiers
- study what it means to have a portfolio beta of 1 (and again)
- indices need not be representative
- investigating portfolio diversity
- understanding portfolio optimization
- effect of risk fraction versus weight constraints, part 1 and part 2