Screening New Portfolio Strategies by Return Correlation

Summary

The document asks how to tell whether a proposed quantitative strategy adds independent exposure to a portfolio of existing strategies. It contrasts a basic comparison of long, neutral, and short position signs with a portfolio selection approach based on return correlation. The cited example says a strategy may be considered when its returns have low correlation, within roughly plus or minus 30 percent, to the returns of strategies already held, alongside other performance criteria.

This gives a simple diversification screen, but it does not explain how to estimate correlation, assess changing relationships, or measure overlap in positions and risk. Nor does it provide evidence that the cited threshold is generally suitable. The question of whether to add, replace, or discard similar strategies remains unanswered; the brief response offers a screening idea rather than a full portfolio construction method.

Key ideas

• Return correlation can serve as an initial screen for whether a new strategy diversifies an existing portfolio.
• The cited example uses a low-correlation threshold alongside other selection criteria.
• The document does not explain how to choose between adding, replacing, or rejecting similar strategies.

Tags

# Adding a new strategy to an existing portfolio


# Adding a new strategy to an existing portfolio












I wanted some help in looking for suitable articles/literature. Suppose an investor has a bunch (bouquet?) of quantitative strategies already generating trading signals for him. If he comes up with a new strategy, how can he test if it overleverages the securities in his portfolio by generating essentially the same trading signals or is independent of his existing strategies? A rudimentary method might involve comparing the sign of closing and intraday positions of each pair of strategies (long, neutral, or short), but are there any better ways of achieving the same? Further, if a strategy in the portfolio seems similar to the new one, how do you decide if you should swap them, include both, remove both, or discard the new one altogether?

## Answer by Attack68 (score 1)

https://quant.stackexchange.com/a/40139

This isn't literature but Quantopian (a crowd sourced hedge fund) selects new strategies if their returns have a low correlation (+- 30%) with other strategies in their portfolio (as well as other performance criteria!). - https://www.quantopian.com/allocation

Shown in full with attribution under the source's licence. Licence: CC BY-SA 4.0 (Stack Exchange)

This summary was written by Stratmill's research agent from the original; it is not a copy of the source.